Datasets:

TaylorAI

/

pubmed_commercial

like 12

PMC005xxxxxx/PMC5002103.txt

==== Front Gut PathogGut PathogGut Pathogens1757-4749BioMed Central London 12110.1186/s13099-016-0121-5ResearchAntimicrobial resistance of Campylobacter isolates from small scale and backyard chicken in Kenya Nguyen Tuan Ngoc Minh tuan.nguyen@fli.deminhtuannn@hvu.edu.vn 123Hotzel Helmut helmut.hotzel@fli.de 2Njeru John john.njeru@fli.de 24Mwituria Joyce JMwituria@kemri.org 4El-Adawy Hosny hosny.eladawy@fli.de 25Tomaso Herbert herbert.tomaso@fli.de 2Neubauer Heinrich heinrich.neubauer@fli.de 2Hafez Hafez M. hafez@vetmed.fu-berlin.de 31 Hung Vuong University, Phu Tho, Vietnam 2 Friedrich-Loeffler-Institut, Institute of Bacterial Infections and Zoonoses, Jena, Germany 3 Institute of Poultry Diseases, Free University Berlin, Berlin, Germany 4 Centre for Microbiology Research (CMR), Kenya Medical Research Institute, Nairobi, Kenya 5 Department of Poultry Diseases, Faculty of Veterinary Medicine, Kafrelsheikh University, Kafrelsheikh, Egypt 27 8 2016 27 8 2016 2016 8 1 399 6 2016 12 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 Thermophilic Campylobacter species are a major cause of bacterial foodborne diarrhoea in humans worldwide. Poultry and their products are the predominant source for human campylobacteriosis. Resistance of Campylobacter to antibiotics is increasing worldwide, but little is known about the antibiotic resistance in Campylobacter isolated from chicken in Kenya. In this study, 35 suspected Campylobacter strains isolated from faeces and cloacal swabs of chicken were tested for their susceptibility to seven antibiotics using a broth microdilution assay and molecular biological investigations. Results Overall, DNA of thermophilic Campylobacter was identified in 53 samples by PCR (34 C. jejuni, 18 C. coli and one mix of both species) but only 35 Campylobacter isolates (31 C. jejuni and 4 C. coli) could be re-cultivated after transportation to Germany. Isolates were tested for their susceptibility to antibiotics using a broth microdilution assay. Additionally, molecular biological detection of antibiotic resistance genes was carried out. C. jejuni isolates showed a high rate of resistance to nalidixic acid, tetracycline and ciprofloxacin of 77.4, 71.0 and 71.0 %, respectively. Low resistance (25.8 %) was detected for gentamicin and chloramphenicol. Multidrug resistance in C. jejuni could be detected in 19 (61.3 %) isolates. Resistance pattern of C. coli isolates was comparable. Resistance to ciprofloxacin was confirmed by MAMA–PCR and PCR–RFLP in all phenotypically resistant isolates. The tet(O) gene was detected only in 54.5 % of tetracycline resistant C. jejuni isolates. The tet(A) gene, which is also responsible for tetracycline resistance, was found in 90.3 % of C. jejuni and in all C. coli isolates. Thirteen phenotypically erythromycin-resistant isolates could not be characterised by using PCR–RFLP and MAMA–PCR. Conclusions To the best of our knowledge, this study is the first report about resistance to antibiotics in thermophilic Campylobacter originating from chicken in Kenya. Campylobacter spp. show a high level of resistance to ciprofloxacin, nalidixic acid and tetracycline but also a remarkable one to chloramphenicol and gentamicin and they are multidrug resistant. Resistance to antibiotics is a global public health concern. In Kenya, resistance surveillance needs further attention in the future. Efforts to establish at least a National Laboratory with facilities for performing phenotypic and genotypic characterization of thermophilic Campylobacter is highly recommended. Keywords CampylobacterAntibiotic resistanceMicrodilutionChickenKenyaissue-copyright-statement© The Author(s) 2016 ==== Body Background Thermophilic Campylobacter (C.) species have become the most frequent cause of bacterial gastroenteritis in humans worldwide [1]. Campylobacteriosis exceed the total number of those caused by Salmonella, Shigella, and Escherichia coli O157:H7 in humans, recently [2]. Campylobacter infections are normally self-limiting in adults but can cause diarrhoea or even mortality in children in developing as well as in developed countries [3, 4]. A study from western Kenya showed that 20 % of patients with diarrhoea were infected by Campylobacter but in the group of children below 5 years Campylobacter was represented with 42 % [5]. Commercial poultry and free-living birds are natural reservoirs of thermophilic campylobacters. The organism has been isolated from numerous bird species, including Columbiformes and domestic and free-living Galliformes and Anseriformes. C. jejuni has been found in all areas of commercial poultry production [6, 7]. Prevalence rates in poultry, especially in slaughter-age broiler flocks, could reach 100 % on some farms. Although, Campylobacter is insignificant for poultry health, it is a predominant cause of foodborne gastroenteritis in humans worldwide, and contaminated poultry meat is recognized as the main source of human infections [7, 8]. In general, the knowledge about Campylobacter in Kenya is limited. Most of the published reports describe prevalence and antibiotic resistance in Campylobacter of human origin [4, 9–11]. Other reports gave information about Campylobacter as cause of foodborne diseases [12] and contamination of raw chicken and beef from butcheries and markets in Nairobi [13]. Information on thermophilic Campylobacter of animal origin from Kenya is lacking. Resistance against antibiotics in bacteria is of public health concern. Most commonly used drugs in treatment of campylobacteriosis in humans are erythromycin, fluoroquinolones or tetracycline [14]. Although, this antimicrobial treatment is usually not necessary, however the misuse of antibiotics is widespread in Kenya [5]. Attention on resistance of Campylobacter is raising and warning has been launched not to misuse antibiotics such as macrolides, fluoroquinolones or alternative drugs [15]. Kenyan Campylobacter isolates from humans showed a high resistance rate against erythromycin (52 %), but only low resistance to ciprofloxacin, tetracycline and nalidixic acid with 6, 18 and 26 % in the past, respectively [5]. Clinical breakpoints of Campylobacter susceptibility based on epidemiological cut-off (ECOFF) values were recommended. EUCAST MIC distributions incorporate human and veterinary clinical data from several sources worldwide [16]. The method of choice for testing antibiotic susceptibility and determination of minimum inhibitory concentration (MIC) values of Campylobacter isolates is the broth microdilution assay [17, 18]. In addition to phenotypical determination of antibiotic resistance, genetic analysis of resistance determinants in Campylobacter can be carried out. A replacement of threonine by isoleucine at amino acid 86 in the gyrA gene [19, 20] and a mutation at position 2074 and 2075 on the 23S rRNA gene are the main mechanisms for fluoroquinolone and erythromycin resistance, respectively [21]. Presence of tet(O) and/or tet(A) genes is responsible for tetracycline resistance [22]. A mismatch amplification mutation assay (MAMA–PCR) can be used for detection of the mutations in gyrA and 23S rRNA genes in C. coli and C. jejuni responsible for ciprofloxacin and erythromycin resistance, respectively [21, 23]. PCR-restriction fragment length polymorphism (PCR–RFLP) technique [24] is available for detection of erythromycin resistance as well as specific PCR assays for tet(O) and tet(A) genes. These methods allow the investigation of antibiotic resistance of Campylobacter even in samples from which no Campylobacter could be isolated. To the best of our knowledge there is no report available about antibiotic resistance of thermophilic Campylobacter species isolated from chicken in Kenya. MICs and results of molecular assays on the resistance of recent Kenyan C. coli and C. jejuni are presented. Methods Sample collection and Campylobacter isolation In total, 35 geographically different native breed layer flocks were sampled. The chickens were housed in backyards and homesteads of small scale farmers from the outskirts of Thika, a town 40 km northeastern of Nairobi, Kenya. Farmers kept between 10 and 1000 layers. The birds were fed on commercially formulated ration from different sources and sometimes supplied with the leftover and residual food. All the manufactures used antibiotics as part of the ingredients in the feed. During the rearing of these chickens, antibiotics were used for prevention and treatment of diseases without any instructions. Ten to 30 cloacal swabs and faecal samples were collected from each flock according to flock size. Campylobacter were isolated in Kenya Medical Research Institute, Nairobi according to the guidelines of ISO 10272-1 [25]. The isolates were preserved in 1.5 ml Eppendorf tubes filled with skimmed milk medium for 1-week transportation from Kenya to Friedrich–Loeffler-Institut, Jena, Germany for further laboratory analysis. Campylobacter strains were re-cultivated on both Mueller–Hinton agar and CCDA (Oxoid GmbH, Wesel, Germany) under microaerophilic conditions (5 % O2, 10 % CO2, and 85 % N2) at 37 °C for 48–72 h. DNA extraction DNA from viable bacteria was extracted using the High Pure PCR Template Preparation Kit™ (Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. Skimmed milk samples of Campylobacter that could not be re-cultivated were treated with phenol–chloroform to extract DNA. Briefly, 500 µl of skimmed milk medium was boiled for 5 min. After cooling the liquid was mixed with 500 µl buffer-saturated phenol (Carl Roth GmbH, Karlsruhe, Germany) and centrifuged for 5 min. at 13,400 rpm (miniSpin, Eppendorf, Hamburg, Germany). 500 µl chloroform/isoamyl alcohol (24:1 vol/vol) was added to the aqueous phase, mixed and centrifuged for 5 min. at 13,400 rpm. DNA from the aqueous phase was precipitated by mixing with 0.6 volume of isopropanol at room temperature. After centrifugation, the supernatant was discarded and the DNA was air dried and finally dissolved in 50 µl 10 mM Tris (Carl Roth GmbH). Multiplex PCR for identification of Campylobacter species A mPCR assay was used to identify thermophilic Campylobacter species (C. jejuni, C. coli, and C. lari) as described by El-Adawy et al. [26]. Briefly, the PCR was performed in a 50-μl reaction mixture containing 5.0 μl of 10 × Taq reaction buffer complete (Jena Bioscience GmbH, Jena, Germany), 2.0 μl of dNTP mix (2 mM each; Carl Roth GmbH), 2.0 μl of each primer (Jena Bioscience GmbH), and 0.2 μl of Taq Pol thermostable DNA polymerase (Jena Bioscience GmbH). Amplification reactions were carried out in a TRIO Thermoblock cycler (Biometra, Göttingen, Germany) using the following programme: one cycle of 1 min at 96 °C was followed by 35 cycles each consisting of 60 s at 95 °C of denaturation, 90 s at 59 °C of annealing, and 60 s at 72 °C of elongation. The PCR was terminated after a final extension step of 5 min. at 72 °C. Amplification generated 857, 589, 522, and 462 base pair DNA fragments specific for the genus Campylobacter and the species C. jejuni, C. lari, and C. coli, respectively. For analysis, 20 μl of PCR products were subjected to electrophoresis in a 1.5 % agarose gel for 1 h, stained with ethidium bromide (0.5 μg/ml), and visualized under UV light. Results were documented using BioImage system GeneGenius (Syngene, Synoptics Ltd., Cambridge, UK). Reference strains C. jejuni DSM 4688, C. coli DSM 4689, and C. lari DSM 11375 (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany) were used as positive controls. Antimicrobial susceptibility testing and determination of MICs The antimicrobial susceptibility of C. jejuni and C. coli isolates was tested against seven antibiotic agents (chloramphenicol, erythromycin, ciprofloxacin, nalidixic acid, gentamicin, streptomycin, and tetracycline) using the Sensititre™ Campylobacter plates—EUCAMP (Trek Diagnostic Systems Ltd., East Grinstead, UK). The MIC values were detected using different concentration ranges as previously described [23]. Briefly, Campylobacter isolates grown on Mueller–Hinton agar (Oxoid GmbH) supplemented with 10 % bovine blood under microaerophilic conditions were suspended in NaCl solution (0.9 %) to obtain a turbidity corresponding to a McFarland standard of 0.5 (Dr. Lange, CADAS photometer 30, Berlin, Germany). One-hundred and fifty milliliters of the above suspension were diluted with 10 ml Mueller–Hinton broth (Oxoid GmbH) resulting in a concentration of approximately 106–107 colony forming units (cfu)/ml. One hundred milliliters of the inoculum was filled in each well of the plate; the plates were sealed and incubated at 37 °C for 24 h under microaerophilic conditions. Results were read either visually or photometrically (Tecan Deutschland GmbH, Crailsheim, Germany) using the computer program easyWIN fitting (version V6.1, 2000). C. jejuni DSM 4688 and C. coli DSM 4689 (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH) were included in each batch of broth microdilution assay for quality control. The lowest concentration of antibiotics that prevents visible growth of the microorganism is defined as the MIC. Molecular biological detection of antibiotic resistance determinants Extracted Campylobacter DNA from all samples and strains was used for molecular biological determination of selected antibiotic resistance determinants by PCR. Erythromycin resistance Detection of mutations at positions 2074 and 2075 in domain V of the 23S rRNA gene, which mediates resistance to erythromycin, was carried out by MAMA–PCR and PCR–RFLP as described previously [21, 24]. Genes responsible for resistance of erythromycin and ciprofloxacin were tested at two loci using MAMA–PCR and PCR–RFLP. Primers and their sequences are given in Table 1.Table 1 List of primers and primer sequences used for detection of antimicrobial resistance genes Antibiotic Method Primer Sequence (5’–3’) Amplicon length (bp) Reference Erythromycin MAMA–PCRa 23SRNA-F ERY2075-R ERY2074-R TTA GCT AAT GTT GCC CGT ACC G TAG TAA AGG TCC ACG GGG TCG C AGT AAA GGT CCA CGG GGT CTG G 485 [17] PCR–RFLPa F2-campy-23S R2-campy-23S AAT TGA TGG GGT TAG CAT TAG C CAA CAA TGG CTC ATA TAC AAC TGG 316 [20] Ciprofloxacin MAMA–PCRb CampyMAMAgyrA1 CampyMAMAgyrA5 TTT TTA GCA AAG ATT CTG AT CAA AGC ATC ATA AAC TGC AA 265 [19] MAMA–PCRc GZgyrACcoli3F CanpyMAMAgyrA8 TAT GAG CGA TAT TAT CGG TC TAA GGC ATC GTA AAC AGC CA 192 [24] Tetracycline tet(O) PCR DMT 1 DMT 2 GGC GTT TTG TTT ATG TGC G ATG GAC AAC CCG ACA GAA GC 559 [25] tet(A) PCR Tet(A)-F Tet(A)-R GTG AAA CCC AAC ATA CCC C GAA GGC AAG CAG GAT GTA G 888 [18] tet(A) PCR tet-A-1 tet-A-2 GCT CAC GTT GAC GCA GGA AAG ATC GTC ATT GTC CGT TAC 486 This study a23S rRNA gene mutation b gyrA gene mutation Campylobacter jejuni c gyrA gene mutation Campylobacter coli Ciprofloxacin resistance A single point mutation (Thr-86-Ile) in the quinolone resistance-determining region (QRDR) of gyrA gene was defined as source of high-level resistance to fluoroquinolones [23]. MAMA–PCR for C. jejuni isolates was carried out as described previously [27], for C. coli a procedure according to Zirnstein et al. [28] was used. Primer details are given in Table 1. Tetracycline resistance Primers DMT1 and DMT2 (Jena Bioscience GmbH) were used for the detection of the tet(O) gene which is strongly associated with tetracycline resistance in C. jejuni and C. coli as described previously [29]. As a second gene locus associated with tetracycline resistance the presence of tet(A) was examined by a previously described PCR assay [22]. An alternative, in-house validated PCR assay was created based on tet(A) sequences (GenBank acc. no. JX891463 and JX891464)). Briefly, primers tet-A-1 and tet-A-2 (Table 1; Jena Bioscience GmbH) were used with the following PCR programme: An initial denaturation at 96 °C for 60 s was followed by 35 cycles of denaturation (96 °C for 15 s), annealing (49 °C for 60 s) and extension (72 °C for 30 s). PCR was terminated by final extension at 72 °C for 60 s. The PCR resulted in a 486 bp product. All PCR products were analyzed by electrophoresis on 1.5 % agarose gels, staining with ethidium bromide and visualization under UV light. DNA sequencing PCR products obtained by tet(A) PCRs were sequenced by cycle sequencing with BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Darmstadt, Germany) according to the instructions of the manufacturer. In addition to the amplification primers of the Tet(A)-F/R fragment tet-A-A (5’-AAT TTT CTT CAA ATA AGG-3’) and tet-A-B (5’-GTC ATT CTT ATA TTA AGT GG-3’) were used as sequencing primers. Sequencing products were analyzed with an ABI PRISM 3130 genetic analyzer. MALDI-TOF mass spectrometry Cultured bacteria were suspended in 300 µl of bi-distilled water and mixed with 900 ml of ethanol (Carl Roth GmbH). Further treatment of samples and measurement were described by El-Ashker et al. [30]. Results Identification of bacteria In total, 58 isolates suspected as Campylobacter were recovered from faeces and cloacal swabs of chicken flocks in Kenya. After storage in skimmed milk medium and transportation to Germany only 40 of these isolates could be re-cultivated. Four C. coli and 31 C. jejuni were identified by mPCR (Table 2). Five other isolates were identified by MALDI-TOF mass spectrometry as members of genera Bacillus, Staphylococcus, Ochrobactrum as well as two Bordetella isolates.Table 2 Results of cultivation and multiplex PCR identification of Campylobacter isolates Cultivation mPCR identification of Campylobacter Total n (%) C. jejuni C. coli Not identified C. coli/C. jejuni Positive (n) 31 4 5 0 40 (69.0) Negative (n) 3 14 0 1 18 (31.0) Eighteen skimmed milk tubes contained Campylobacter DNA [14 C. coli, 3 C. jejuni and one sample harboured both C. coli and C. jejuni (Table 2)]. Antimicrobial susceptibility profiles and multidrug resistance The results of antimicrobial susceptibility testing of C. jejuni and C. coli isolates and the rate of resistance to seven antimicrobial agents are given in Tables 3 and 4, respectively. The C. jejuni isolates showed a high rate of resistance to nalidixic acid, tetracycline and ciprofloxacin with 77.4, 71.0 and 71.0 %, respectively. Low resistance rates were detected for gentamicin and chloramphenicol, both with 25.8 % of the isolates. For the low number of C. coli isolates (n = 4) a similar pattern was observed. Only two isolates were susceptible to all tested antimicrobial agents, one isolate was resistant to all tested antibiotics.Table 3 Results of MIC determination and resistance rates of Kenyan Campylobacter jejuni isolates Boldface in italic type indicates the number of resistant isolates. A thick black line indicates the break point between clinically sensitive and resistant strains R* resistance rate Table 4 Results of MIC determination and resistance rates of Kenyan Campylobacter coli isolates Boldface in italic type indicates the number of resistant isolates. A thick black line indicates the break point between clinically sensitive and resistant strains R* resistance rate The multidrug resistance profiles of 31 C. jejuni isolates are shown in Table 5. Multidrug resistance to three or more classes of antibiotics was found in 19 isolates (61.3 %) and was observed in a range between 5.3 and 26.3 %.Table 5 Multidrug resistance profiles of 19 Campylobacter jejuni isolates Antibiotic resistance profile No. of resistant isolates (%) TET, CIP, NAL, ERY, GEN, STR, CHL 1 (5.3) TET, CIP, NAL, ERY, STR, CHL 1 (5.3) TET, CIP, NAL, ERY, GEN, STR 3 (15.8) TET, CIP, NAL, GEN, STR, CHL 1 (5.3) TET, CIP, NAL, ERY, STR 2 (10.5) TET, CIP, NAL, ERY, CHL 1 (5.3) TET, CIP, NAL, ERY 2 (10.5) TET, CIP, NAL, STR 1 (5.3) TET, CIP, NAL 5 (26.3) CIP, NAL, ERY, STR, CHL 1 (5.3) CIP, NAL, STR 1 (5.3) TET tetracycline, CIP ciprofloxacin, NAL nalidixic acid, ERY erythromycin, GEN gentamicin, STR streptomycin, CHL chloramphenicol Molecular biological detection of antibiotic resistance determinants DNA of 35 viable Campylobacter isolates and of 18 non-growing samples was investigated by PCR to detect antibiotic resistance. Mismatch amplification mutation assay (MAMA–PCR) was used to characterize a gyrA gene mutation associated with ciprofloxacin resistance as well as mutations in 23S rRNA genes as cause of erythromycin resistance. The molecular biological detection of resistance to ciprofloxacin in both C. coli and C. jejuni was also confirmed by change of amino acid 86 from threonine to isoleucine in the gyrA gene. Additional to the ciprofloxacin resistant Campylobacter isolates (Tables 3, 4), three Campylobacter DNAs were detected harbouring the gyrA gene mutation. The results were confirmed by PCR–RFLP according to Vacher et al. [24]. Mutations at positions 2074 and 2075 of the 23S rRNA genes in 13 phenotypically erythromycin-resistant isolates could neither be detected by using PCR–RFLP nor MAMA–PCR. The tet(O) gene which is mainly responsible for tetracycline-resistance was detected by PCR in 12 out of 22 resistant C. jejuni isolates (54.5 %) and in all tetracycline resistant C. coli isolates. tet(O) gene was not detected in DNA extracted from the non-growing samples. Additionally, a newly developed PCR assay was used for the detection of the tet(A) gene. tet(A) was detected in 28 out of 31 C. jejuni (90.3 %) and in all 4 C. coli isolates. In 3 out of 14 non-growing samples which harboured C. coli DNA, tet(A) gene could be found as well as in the one sample where both C. jejuni and C. coli were detected. Discussion The antimicrobial susceptibility patterns among Campylobacter isolates originating from chicken in Kenya were analyzed according to the guidelines of CLSI for Enterobacteriaceae which had been guided by previous reports [17, 31–34]. Clinical breakpoints for interpretation of MIC values of C. jejuni and C. coli from chicken are available [31, 35]. In this study a commercially available broth microdilution assay was used for the determination of MIC values for seven antibiotics. The assay already proved to be suitable in previous investigations [21, 27, 34, 36, 37]. In this study, only 40 out of 58 suspected Campylobacter samples could be re-cultivated. The storage conditions (temperature, microaerophilic atmosphere) using skimmed milk medium were possibly not ideal. However, it had been demonstrated that C. jejuni can survive up to 14 days at 1 °C or 2.5 days at 20 °C in sterile skimmed milk [38, 39]. Alternative storage of Campylobacter cultures using transport medium (for example Amies medium) or cryovials is recommended for future investigations. In 53 out of 58 collected samples, Campylobacter DNA was identified by mPCR assay [26]. The majority of the cultures proved to be C. jejuni (88.6 %) which is in agreement with previous studies in chicken [6, 7]. In 18 DNA extracts of non-viable samples, 3 C. jejuni (16.7 %), one mixed population of C. jejuni/C. coli and 14 C. coli were identified. These findings are in agreement with those of a previous study that found a longer viability of C. jejuni in comparison to C. coli in biological milieu [40]. It may be possible that some C. coli isolates had been lost during the shipment period. In summary, C. jejuni was identified much more often than C. coli (64.2 %) by mPCR investigation. In agreement with other studies, the findings highlighted the usefulness of mPCR as a reliable, sensitive, time and cost saving method for identification of thermophilic Campylobacter [26]. The antibiotic susceptibility of 35 Campylobacter isolates from Kenyan chicken was investigated using European Committee on Antimicrobial Susceptibility Testing and epidemiological cut-off values (EUCAST–ECOFFS) [16]. A broth microdilution assay was used as a standardized, easy, and reliable method for the determination of MIC of seven antibiotics [17, 31–34]. High resistance rates were obtained for ciprofloxacin, nalidixic acid and tetracycline with more than 70 % which is in agreement with a recent European Food Safety Authority (EFSA) report [41]. These results are in contrast to those of Brooks et al. [5] who reported resistance rates for Campylobacter recovered from humans with diarrhoea in Western Kenya for ciprofloxacin, nalidixic acid and tetracycline with 6, 26 and 18 % in 2006, respectively. The general high rates of resistance in the chicken isolates may be caused by availability and uncontrolled use of antibiotics by small farmers [42]. Resistance to chloramphenicol is remarkable with 25.8 % in this investigation. Use of chloramphenicol is banned in animal breeding in Europe for more than 20 years, but still it is often used in many third world countries [43]. It is easy to obtain antibiotics over-the-counter and other unregulated venues and injudicious use promotes the development of resistance to antimicrobial agents. Resistance to gentamicin in the isolates obtained from chicken was low in this study (25.8 %), but Campylobacter isolated from broilers and turkeys were totally susceptible to gentamicin [37, 41, 44, 45]. Erythromycin resistance rates found in this study correspond to those of similar studies elsewhere [41, 44, 45]. Multidrug resistance was detected in 61.3 % of the Campylobacter isolates. Eleven different combinations were found (Table 5). Frequent, resistance to ciprofloxacin, nalidixic acid and tetracycline was identified (17 out of 19 multidrug resistant isolates) which is in agreement with previous investigation using Vietnamese Campylobacter isolates [45]. However, EFSA [41] reported low level of multidrug resistance in C. jejuni from broilers of the member states of the EU. The emerging of antibiotic resistance has been attributed to the overuse and misuse of antimicrobial agents in both the developed and developing world. Antibiotics are widely used as growth supplements in livestock and to prevent infections [46]. The emerging of multidrug resistance may reflect acquisition of different resistance determinants on the same DNA molecule or single determinants, such as multidrug pumps, that specify efflux activity against different antimicrobial agents [47]. The mechanisms of genetic resistance might be chromosomal or plasmid-borne, and represent a combination of endogenous and acquired genes. In general, mechanisms of antibiotic resistance as modification of the antibiotic by aminoglycoside-modifying enzymes (AphA, AadE, Sat), enzymatic inactivation of the antibiotic by β-lactamase and modification of the DNA gyrase target, mutations in 23S rRNA genes were included for aminoglycosides, beta-lactams, fluoroquinolones, macrolides and tetracyclines, respectively [48, 49]. The multidrug efflux pump CmeABC has been involved in the resistance mechanisms of C. jejuni and C. coli to tetracyclines, fluoroquinolones, macrolides and beta-lactams [49]. Molecular biological methods were used for detection of antibiotic resistance determinants either using DNA isolated from cultures or that of non-cultured bacteria [27, 37]. All isolates of this study which were resistant to ciprofloxacin carried a mutation of the amino acid 86 of the gyrA resulting in a change from threonine to isoleucine. This mutation was detected also in 3 DNA samples extracted from skimmed milk. The MAMA–PCR protocol allowed the detection of the gyrA mutation and PCR–RFLP was confirming the mutation from (ACA to ATA) of amino acid 86. This result was in agreement with previous reports showing that both methods are simple, reliable, rapid tools that can be used as screening methods [27, 37]. In Campylobacter, resistance to erythromycin is chromosomally encoded by an alteration of the 23S rRNA gene. High level resistance to erythromycin is caused by mutations at position 2074 and/or 2075 of the domain V of this gene. In this study the mutations were neither detected by MAMA–PCR nor by PCR–RFLP. The tet(O) gene is known to be responsible for tetracycline resistance in Campylobacter isolates [29]. In this study, only 54.5 % of the tetracycline resistant isolates harboured the tet(O) gene. The tet(A) gene also plays role in resistance to tetracycline [22]. The efflux gene tet(A) is coding for an approximately 46 kDa membrane-bound efflux protein for membrane-associated proteins and is involved in the export of tetracycline from the cell [50]. In this study, using the recommended primers for tet(A) amplification [18] PCR products of 696 bp instead of 888 bp were obtained. DNA sequencing of amplicons and database search resulted in 99.0 % homology to a partial putative integral membrane protein and a putative periplasmic protein. Hence, a new PCR assay based on tet(A) gene sequences for C. jejuni (acc. no. JX891464) and C. coli (acc. no. JX891463) was developed. Parameters such as limit of detection, limit of quantification, PCR efficiency and specificity were considered during an in-house validation process. Amplicon length was 486 and the amplicons were sequenced to confirm the identity. The tet(A) gene was much more frequently identified in the Kenyan Campylobacter isolates than tet(O) (35 vs 13). To the best of our knowledge this is the first report on the status of antibiotic susceptibility of thermophilic Campylobacter from chicken in Kenya. High level of resistance to ciprofloxacin, erythromycin and nalidixic acid as well as multidrug resistance was detected previously in Kenya. In Kenya, this problem is reported to be caused by the increasing rate of unregulated over-the-counter sale without prescriptions of these antibiotics, mainly to humans self-treatment of suspected infections and to a lesser extent for use in animals [51]. These findings also demonstrate the potential for antibiotic-resistant bacteria to spread through the food chain from animals treated with antibiotics for humans. Such misuse and overuse may have resulted in the selection of resistant mutants or acquisition of antibiotic resistance genes from other organisms through the process of genetic exchange. It is recommendable that a long-term local surveillance programme is adopted for monitoring changes in resistance among Campylobacter isolates. Efforts to establish at least a National Laboratory with facilities for performing phenotyping and genotyping methods is highly recommended. Emphasis should be given on educational advertising to reduce the input of antibiotics in animal breeding to minimize the potential hazard for humans. Authors’ contributions TNMN, HH, JN, HE, HT and HMH participated in the conception and design of the study. TNMN, HH, JN, JM and HE performed farm and laboratory work. TNMN, HH, JN, HE, HT, HN and HMH analysed the data and wrote the manuscript. All authors contributed to the analysis and supported the manuscript discussion. All authors read and approved the final manuscript. Acknowledgements The authors thank Byrgit Hofmann and Ronald Ng’etich for excellent technical assistance. Competing interests The authors declare that they have no competing interests. Availability of data and material The data supporting the findings of this study are contained within the manuscript. Ethics approval and consent to participate This study was carried out in strict accordance with the recommendations of the National Guidelines on the Care and Use of Animals in Research, Education and Training in Kenya, Consortium for National Health Research (CNHR) which complies with the international laws and regulation regarding ethical considerations, transport, housing and experimental use of animals in research. Funding This study was supported by Friedrich–Loeffler-Institut, Institute of Bacterial Infections and Zoonoses, Germany and Vietnam Ministry of Agriculture and Rural Development. ==== Refs References 1. Man SM The clinical importance of emerging Campylobacter species Nat Rev Gastroenterol Hepatol 2011 8 669 685 10.1038/nrgastro.2011.191 22025030 2. European Food Safety Authority (EFSA). Manual for reporting of food-borne outbreaks in accordance with Directive 2003/99/EC from the year 2010/2011. European Food Safety Authority (EFSA). Parma. 2011; 138. 3. Coker AO Isokpehi RD Thomas BN Amisu KO Obi CL Human campylobacteriosis in developing countries Emerg Infect Dis 2002 8 237 244 10.3201/eid0803.010233 11927019 4. Saidi SM Iijima Y Sang WK Mwangudza AK Oundo JO Taga K Epidemiological study on infectious diarrheal diseases in children in a coastal rural area of Kenya Microbiol Immunol 1997 41 773 778 10.1111/j.1348-0421.1997.tb01925.x 9403500 5. Brooks JT Ochieng JB Kumar L Okoth G Shapiro RL Wells JG Surveillance for bacterial diarrhea and antimicrobial resistance in rural Western Kenya, 1997-2003 Clin Infect Dis 2006 43 393 401 10.1086/505866 16838225 6. Agunos A Waddell L Léger D Taboada E A systematic review characterizing on-farm sources of Campylobacter spp. for broiler chickens PLoS ONE 2014 9 8 e104905 10.1371/journal.pone.0104905 25171228 7. Sahin O Kassem II Shen Z Lin J Rajashekara G Zhang Q Campylobacter in poultry: ecology and potential interventions Avian Dis 2015 59 185 200 10.1637/11072-032315-Review 26473668 8. Hermans D Pasmans F Messens W Martel A Van Immerseel F Rasschaert G Poultry as a host for the zoonotic pathogen Campylobacter jejuni Vector Borne Zoonotic Dis 2012 12 89 98 10.1089/vbz.2011.0676 22133236 9. Joyce T McGuigan KG Elmore-Meegan M Conroy RM Prevalence of enteropathogens in stools of rural Maasai children under five years of age in the Maasailand region of the Kenyan Rift Valley East Afr Med J 1996 73 59 62 8625866 10. Shapiro RL Kumar L Phillips-Howard P Wells JG Adcock P Brooks J Antimicrobial-resistant bacterial diarrhea in rural western Kenya J Infect Dis 2001 183 1701 1704 10.1086/320710 11343224 11. Swierczewski BE Odundo EA Koech MC Ndonye JN Kirera RK Odhiambo CP Surveillance for enteric pathogens in a case–control study of acute diarrhea in Western Kenya Trans R Soc Trop Med Hyg 2013 107 83 90 10.1093/trstmh/trs022 23222955 12. Ombui JN Kagiko MM Arimi SM Foodborne diseases in Kenya East Afr Med J 2001 78 40 44 11320765 13. Osano O Arimi SM Retail poultry and beef as sources of Campylobacter jejuni East Afr Med J 1999 76 141 143 10442113 14. Blaser B Engberg J Nachamkin I Szymanski CM Blaser MJ Clinical aspects of Campylobacter jejuni and Campylobacter coli infections Campylobacter 2008 Washington DC American Society for Microbiology 99 121 15. EFSA: Summary Report on antimicrobial resistance in zoonotic and indicator bacteria from humans, animals and food. EFSA J.2015;13(2):4036. doi:10.2903/j.efsa.2015.4036. 16. EUCAST. Breakpoint tables for interpretation of MICs and zone diameters. http://www.eucast.org, Version 3.1. 2013. 17. Ge B Wang F Sjölund-Karlsson M McDermott PF Antimicrobial resistance in Campylobacter : susceptibility testing methods and resistance trends J Microbiol Methods 2013 95 57 67 10.1016/j.mimet.2013.06.021 23827324 18. Luber P Bartelt E Genschow E Wagner J Hahn H Comparison of broth microdilution, E test, and agar dilution methods for antibiotic susceptibility testing of Campylobacter jejuni and Campylobacter coli J Clin Microbiol 2003 41 1062 1068 10.1128/JCM.41.3.1062-1068.2003 12624030 19. Wang Y Huang WM Taylor DE Cloning and nucleotide sequence of the Campylobacter jejuni gyr A gene and characterization of quinolone resistance mutations Antimicrob Agents Chemother 1993 37 457 463 10.1128/AAC.37.3.457 8384814 20. Alonso R Mateo E Girbau C Churruca E Martinez I Fernandez-Astorga A PCR-restriction fragment length polymorphism assay for detection of gyr A mutations associated with fluoroquinolone resistance in Campylobacter coli Antimicrob Agents Chemother 2004 48 4886 4888 10.1128/AAC.48.12.4886-4888.2004 15561873 21. Alonso R Mateo E Churruca E Martinez I Girbau C Fernandez-Astorga A MAMA–PCR assay for the detection of point mutations associated with high-level erythromycin resistance in Campylobacter jejuni and Campylobacter coli strains J Microbiol Methods 2005 63 99 103 10.1016/j.mimet.2005.03.013 15927294 22. Abdi-Hachesoo B Khoshbakht R Sharifiyazdi H Tabatabaei M Hosseinzadeh S Asasi K Tetracycline resistance genes in Campylobacter jejuni and C. coli isolated from poultry carcasses Jundishapur J Microbiol 2014 7 e12129 10.5812/jjm.12129 25485062 23. Zirnstein G Li Y Swaminathan B Angulo F Ciprofloxacin resistance in Campylobacter jejuni isolates: detection of gyr A resistance mutations by mismatch amplification mutation assay PCR and DNA sequence analysis J Clin Microbiol 1999 37 3276 3280 10488192 24. Vacher S Menard A Bernard E Megraud F PCR-restriction fragment length polymorphism analysis for detection of point mutations associated with macrolide resistance in Campylobacter spp Antimicrob Agents Chemother 2003 47 1125 1128 10.1128/AAC.47.3.1125-1128.2003 12604552 25. International Standards Organization. [ISO] 10272-1. Microbiology of food and animal feeding stuffs-horizontal method for detection and enumeration of Campylobacter spp. Part 1: detection method. 2006. 26. El-Adawy H Hotzel H Tomaso H Neubauer H Hafez HM Elucidation of colonization time and prevalence of thermophilic Campylobacter species during turkey rearing using multiplex polymerase chain reaction Poult Sci 2012 91 454 459 10.3382/ps.2010-01810 22252360 27. El-Adawy H Hotzel H Düpre S Tomaso H Neubauer H Hafez HM Determination of antimicrobial sensitivities of Campylobacter jejuni isolated from commercial turkey farms in Germany Avian Dis 2012 56 685 692 10.1637/10135-031912-Reg.1 23397839 28. Zirnstein G Helsel L Li Y Swaminathan B Besser J Characterization of gyr A mutations associated with fluoroquinolone resistance in Campylobacter coli by DNA sequence analysis and MAMA–PCR FEMS Microbiol Lett 2000 190 1 7 10.1111/j.1574-6968.2000.tb09253.x 10981681 29. Mazi W Senok AC Al-Mahmeed A Arzese A Bindayna K Botta G Trends in antibiotic sensitivity pattern and molecular detection of tet (O)-mediated tetracycline resistance in Campylobacter jejuni isolates from human and poultry sources Jpn J Infect Dis 2008 61 82 84 18219143 30. El-Ashker M Gwida M Tomaso H Monecke S Ehricht R El-Gohary F Staphylococci in cattle and buffaloes with mastitis in Dakahlia Governorate, Egypt J Dairy Sci 2015 98 7450 7459 10.3168/jds.2015-9432 26364099 31. Clinical and Laboratory Standards Institute (CLSI). Performance standards for antimicrobial susceptibility testing; 20th informal supplement M100–S20. 2010; 30(1). 32. Hariharan H Sharma S Chikweto A Matthew V DeAllie C Antimicrobial drug resistance as determined by the E-test in Campylobacter jejuni , C. coli , and C. lari isolates from the ceca of broiler and layer chickens in Grenada Comp Immunol Microbiol Infect Dis 2009 32 21 28 10.1016/j.cimid.2008.01.010 18329712 33. Luangtongkum T Morishita TY El-Tayeb AB Ison AJ Zhang Q Comparison of antimicrobial susceptibility testing of Campylobacter spp. by the agar dilution and the agar disk diffusion methods J Clin Microbiol 2007 45 590 594 10.1128/JCM.00986-06 17122005 34. Luber P Wagner J Hahn H Bartelt E Antimicrobial resistance in Campylobacter jejuni and Campylobacter coli strains isolated in 1991 and 2001–2002 from poultry and humans in Berlin Germany Antimicrob Agents Chemother 2003 47 3825 3830 10.1128/AAC.47.12.3825-3830.2003 14638490 35. The European Committee on Antimicrobial Susceptibility Testing (EUCAST). Breakpoint tables for interpretation of MICs and zone diameters. Version 4.0. 2014. 36. Clinical and Laboratory Standards Institute (CLSI). Performance standards for antimicrobial disk and dilution susceptibility tests for bacterial isolates from animals; approved standards-3rd ed. M31–A3 2008; 28(8). 37. El-Adawy H Ahmed MFE Hotzel H Tomaso H Tenhagen BA Hartung J Antimicrobial susceptibilities of Campylobacter jejuni and Campylobacter coli recovered from organic turkey farms in Germany Poult Sci 2015 94 2831 2837 10.3382/ps/pev259 26371330 38. Christopher FM Smith GC Vanderzant C Effect of temperature and pH on the survival of Campylobacter fetus J Food Prot 1982 45 253 259 39. Doyle MP Roman DJ Prevalence and survival of Campylobacter jejuni in unpasteurized milk Appl Environ Microbiol 1982 44 1154 1158 6897503 40. Blaser MJ Hardesty HL Powers B Wang WL Survival of Campylobacter fetus subsp. jejuni in biological milieus J Clin Microbiol 1980 11 309 313 6892819 41. European Food Safety Authority (EFSA)/European Centre for Disease Prevention and Control (ECDC). EU summary report an antimicrobial resistance in zoonotic and indicator bacteria from humans, animals and food in 2013. EFSA J. 2015; 13:4036. 42. Center for Disease Dynamics, Economics & Policy. 2015. State of the World´s antibiotics. 2015. CDDEP: Washington DC, 38–48. 43. European Food Safety Authority (EFSA). Scientific opinion on chloramphenicol in food and feed. EFSA J. 2014; 12:3907. 44. Luangtongkum T Morishita TY Ison AJ Huang S McDermott PF Zhang Q Effect of conventional and organic production practices on the prevalence and antimicrobial resistance of Campylobacter spp. in poultry Appl Environ Microbiol 2006 72 3600 3607 10.1128/AEM.72.5.3600-3607.2006 16672508 45. Nguyen TNM Hotzel H El-Adawy H Tran HT Le MTH Tomaso H Genotyping and antibiotic resistance of thermophilic Campylobacter isolated from chicken and pig meat in Vietnam Gut Pathogens 2016 8 19 10.1186/s13099-016-0100-x 27175218 46. Ventola CL The antibiotic resistance crisis: part 1: causes and threats Pharm Ther 2015 40 277 283 47. Levy SB Factors impacting on the problem of antibiotic resistance J Antimicrob Chemother 2002 49 25 30 10.1093/jac/49.1.25 11751763 48. Zhang Q Plummer P Nachamkin I Szymanski C Blaser M Mechanisms of antibiotic resistance in Campylobacter Campylobacter 2008 3 Washington DC ASM Press 263 276 49. Iovine NM Resistance mechanisms in Campylobacter jejuni Virulence 2013 4 230 240 10.4161/viru.23753 23406779 50. Chopra I Roberts M Tetracyline antibiotics: mode of action, applications, molecular biology, and epidemiology of bacterial resistance Microbiol Mol Rev 2001 62 232 260 10.1128/MMBR.65.2.232-260.2001 51. Kariuki S Revathi G Kariuki N Kiiru J Mwiruria J Hart CA Characterisation of community acquired non-typhoidal Salmonella from bacteraemia and diarrhoeal infection in children admitted to hospital in Nairobi Kenya BMC Microbiol 2006 6 101 10.1186/1471-2180-6-101 17173674

PMC005xxxxxx/PMC5002104.txt

==== Front BMC Vet ResBMC Vet. ResBMC Veterinary Research1746-6148BioMed Central London 80610.1186/s12917-016-0806-5Case ReportSurviving and fatal Elephant Endotheliotropic Herpesvirus-1A infections in juvenile Asian elephants – lessons learned and recommendations on anti-herpesviral therapy Dastjerdi Akbar +44 20 8026 9394akbar.dastjerdi@apha.gsi.gov.uk 1Seilern-Moy Katharina katharina.seilern-moy@apha.gsi.gov.uk 12Darpel Karin k.darpel@surrey.ac.uk 2Steinbach Falko falko.steinbach@apha.gsi.gov.uk 12Molenaar Fieke fieke.molenaar@zsl.org 31 Animal and Plant Health Agency-Weybridge, New Haw, Addlestone, Surrey, KT15 3NB UK 2 School of Veterinary Medicine, Faculty of Health and Medical Sciences, University of Surrey, Guildford, GU2 7XH UK 3 ZSL Whipsnade Zoo, Dunstable, Bedfordshire LU6 2LF UK 27 8 2016 27 8 2016 2016 12 1 1785 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 Elephant Endotheliotropic Herpesviruses (EEHVs) can cause acute haemorrhagic disease in young Asian elephants (Elephas maximus) and clinical EEHV infections account for the majority of their fatalities. The anti-herpesviral drug famciclovir (FCV) has been used routinely to treat viraemic at-risk elephants, but thus far without proven efficacy. This paper presents clinical and virological investigations of two EEHV-1A infected elephants treated with FCV, and discusses anti-herpesvirus therapies of viraemic elephants. Cases presentations Two 1.5 year old male Asian elephants at a zoological collection in the UK developed clinical EEHV-1A infections. Case 1 showed signs of myalgia for the duration of 24 hours before returning back to normal. EEHV-1A DNAemia was confirmed on the day of clinical signs and continued to be present for 18 days in total. Trunk shedding of the virus commenced 10 days after detection of initial DNAemia. Case 2 tested positive for EEHV-1A DNAemia in a routine blood screening sample in the absence of clinical signs. The blood viral load increased exponentially leading up to fatal clinical disease seven days after initial detection of DNAemia. Both calves were treated with 15 mg/kg FCV per rectum on detection of DNAemia and penciclovir, the FCV metabolite, could be detected in the blood at assumed therapeutic levels. The early indicators for clinical disease were a marked absolute and relative drop in white blood cells, particularly monocytes prior to the detection of viraemia. The most prognostic haematological parameter at later stages of the disease was the platelet count showing a continuous sharp decline throughout, followed by a dramatic drop at the time of death. Conclusions The EEHV-1A viraemic animals investigated here further highlight the ongoing threat posed by these viruses to juvenile Asian elephants. The findings call into question the efficacy of rectal FCV in clinical cases and direct towards the use of alternative anti-herpesvirus drugs and complementary treatments such as plasma infusions if no improvement in either viral load or the above-mentioned blood parameters are observed in the initial days of viraemia despite anti-herpesvirus therapy. Keywords EEHV-1elephantid herpesvirusEEHV-haemorrhagic diseaseEEHV-HDfamciclovirganciclovirissue-copyright-statement© The Author(s) 2016 ==== Body Background Elephant Endotheliotropic Herpesviruses (EEHVs) or Elephantid Herpesviruses (ElHVs) belong to the genus Proboscivirus of the family Herpesviridae [1, 2] and fall into 7 genotypes [3–9]. Phylogenetically, the viruses are most closely, although distantly, related to human cytomegalovirus (HCMV) and three human Roseolo viruses (human herpesvirus-6A, -6B, and -7), all members of the subfamily Betaherpesvirinae [1]. EEHVs, in particular EEHV-1, are the cause of acute haemorrhagic disease (EEHV-HD) in young Asian elephants (Elephas maximus), and clinical EEHV infections account for most of the fatalities in juvenile Asian elephants, both in their natural habitat and in captivity [7–13]. EEHV-1 induced acute haemorrhagic disease is manifested initially by sudden onset of non-specific clinical signs, which often develop into haemorrhagic disease and sudden death within 1 to 7 days [9, 12]. It has also been documented that EEHV viraemia is detectable at least several days before the onset of clinical signs [13, 14]. To enable early therapeutic intervention in response to detection of EEHV viraemia, some zoological collections have implemented regular blood sampling and laboratory testing of juvenile Asian elephants. Conferring from herpesviruses’ properties, it could be assumed that once an elephant is infected it remains a latent carrier for life. Screenings of elephant herds have revealed widespread prevalence of the viruses and their periodical reactivation and shedding through biological fluids, especially trunk secretions [13–19]. Routes of EEHVs transmission to naïve individuals are still unclear, however, considering other herpesviruses’ transmission routes, contact with infected bodily secretions is a very likely possibility and vertical transmission cannot be excluded. EEHVs, similar to viruses in the subfamily Alphaherpesvirinae and Gammaherpesvirinae, comprise of a genome with a thymidine kinase (TK) and a serine/threonine protein kinase (PK) [7, 20], which encode enzymes crucial for processing many anti-herpesviral prodrugs into their active derivatives. Based on approaches in treatment of other herpesviruses with TK and PK genes, three nucleoside analogues [famciclovir (FCV) and more recently ganciclovir (GCV) and aciclovir (ACV)] have been administered to treat viraemic elephant calves, with the aim to avert progression to clinical disease, but as yet without proven clinical efficacy. It is known that several elephant calves have survived laboratory confirmed EEHV infections. These elephant calves were treated with FCV at doses of 5.5-8.0 mg/kg twice daily (BID) rectally [9, 18, 19, 21–24] or with GCV at 5 mg/kg BID intravenously followed by oral administration [14, 25]. One elephant calf with advanced EEHV-1 clinical signs has been treated successfully with ACV at 12mg/kg BID intravenously for 15 days along with penicillin G and supportive treatment including vitamin C, physiological saline and 5 % Dextrose [26]. Apart from the latter case, it seems likely that many of surviving elephant calves represent relatively mild cases, in view of length of time taken to reach the highest viral load, level of viraemia at this point and recovery of blood parameters. It is therefore likely that these affected individuals might have recovered without drug intervention [9]. FCV, chemically known as 2-[2-(2-amino-9H-purin-9-yl)ethyl]-1,3-propanediol diacetate, is a synthetic acyclic guanosine analogue that is rapidly de-acetylated and oxidized in the body to its metabolite penciclovir (PCV). Activation of PCV requires phosphorylation in the virally infected cells by viral and cellular kinases. The drug has a half-life of approximately three hours and its catabolite is excreted renally [27]. PCV interacts with herpesvirus DNA polymerase, the enzyme responsible for creating DNA molecules through assembly of nucleotides (building blocks of DNA), and thus terminates DNA synthesis. A recent study showed that when administered to healthy young Asian elephants both orally or per rectum, FCV is absorbed and metabolised to PCV with plasma levels considered therapeutic in humans [27, 28]. ACV, another guanosine analogue with a biological half-life of 2-4 hours, is still used widely for the treatment and prevention of clinical Herpes Simplex Virus (HSV) and Varicella Zoster Virus (VZV) diseases in humans. Activation of ACV within cells via the viral and cellular kinases yields ACV–triphosphate, which interacts adversely with herpesvirus DNA synthesis [29]. GCV is another guanosine analogue that is activated via viral and cellular kinases. Its oral prodrug valganciclovir (VGCV) is the most commonly prescribed drug for the prevention and therapy of HCMV in immunocompromised patients [29]. This paper describes the investigations into two EEHV-1A viraemic elephants from a UK zoological collection that were both treated with FCV. It also discusses findings of the investigations in relation to anti-herpes viral therapies in EEHV viraemic elephants. Case presentations Virus screening Blood samples of EDTA whole blood and blood swabs were collected under trained behaviour from an ear vein after aseptic preparation of the medial aspect of the ear pinna. For EDTA whole blood samples, the animal was placed in lateral recumbency and the lowermost ear was sampled, using a standard 21G 1-inch needle and a Vacutainer® blood collection system. For blood swabs, a dry 6-inch foam tipped swab (Sigma Aldrich, UK) was used for collection of a drop of blood after puncture of an ear vein with an insulin needle, with the elephant in standing position. Blood samples were kept at room temperature before being shipped to the laboratory, where they were stored at 4–8 °C until being processed. Under trained behaviour, conjunctival swabs and trunk swabs were collected. Conjunctival swabs were collected using a sterile dry swab for each eye, running these on the inside of the lower palpebra. Trunk secretions from the mucosa of each naris were collected on a dry 6-inch foam tipped swab at a depth of approximately 8–10 cm, under trained behaviour, and stored at −20 °C prior to further testing. Blood smears were made of freshly collected EDTA blood samples, air dried and stained with Diff-Quick (LabPak, UK) for an in-house relative white blood cell (WBC) differential count. Nucleic acid was extracted from 140 μl EDTA whole blood using the QIAamp® Viral RNA Mini kit (Qiagen, UK) according to the manufacturer’s protocol. For blood swabs, 560 μl buffer AVL from the kit were directly added to the foam tipped swabs, left for 15 minutes at room temperature, vortexed and spun briefly, and processed as per the manufacturer’s instructions. To process trunk swab samples for nucleic acid extraction, 300 μl PBS and 6–8 glass beads (3 mm in diameter) were added to the swab inside a 2 ml Eppendorf tube, vortexed briefly and incubated for 30 minutes at room temperature. The tube content was then vortexed and nucleic acid was extracted following the EDTA whole blood protocol using 140 μl of the cleared suspension. The qPCR used in this study was performed as described [16] using the QuantiFast® Pathogen PCR + IC kit (Qiagen). Briefly, 2 μl extracted DNA was added to 0.1 μl each forward and reverse primer (100 pmol/μl), 0.05 μl hydrolysis probe (100 pmol/μl), 5 μl 5× QuantiFast Pathogen Master Mix, 0.5 μl 50× High-ROX Dye solution, and 17.25 μl RNase-Free Water. The qPCR was performed on a MX3005P machine (Agilent Technologies, UK) using the cycling conditions of one cycle of 95 °C for five minutes followed by 45 cycles of 95 °C for 15 seconds, 60 °C for 30 seconds, and signal acquisition at the end of each cycle. The conventional PCRs were carried out as described [30] using several primer pairs to amplify partial sequences of polymerase (POL), glycoprotein L (gL) and gM genes. The DNA standard used was a HPLC purified 100 bases synthetic oligonucleotide (Eurofins Genomics, UK), identical to the sequence of the qPCR target, diluted to provide serial copy numbers from 106 to 101 copies per PCR test. The synthetic oligonucleotide dilutions were run in triplicate in the PCRs to create a standard curve for EEHV genome quantification. Supportive therapy and drug administration Rectal fluids were provided after manual removal of all faeces from the rectum. Several litres of hand-warm water were then deposited into the distal colon at approximately one meter depth, through a standard garden hose pipe. Intravenous (i/v) access was acquired after antiseptic preparation of the medial aspect of the ear pinna, using an 18G 2.5-inch indwelling catheter (Milacath®, Mila International, USA) in an ear vein. The catheter was secured using Durapore® tape and superglue. Standard giving sets were used for i/v fluids and filtered giving sets for the administration of blood and blood components (Baxter, UK). Plasma was collected from adult herdmates under trained behaviour, in lateral recumbency from either an ear vein or a medial hind leg vein, in one litre citrate blood collection bags (Crusse, Denmark), stored for a minimum of eight hours at 5 °C and separated using manual pressure. The plasma was then stored at 5 °C and warmed to room temperature prior to administration. Anti-herpes viral treatment consisted of either FCV or GCV. FCV (Novartis Pharmaceuticals, UK) 250 and/or 500 mg tablets were crushed using a pestle and mortar and suspended in hand warm water. Using a 60 ml syringe and a meter long tube, the suspension was administered in the distal colon after removal of all faeces as described above. GCV (Cymevene® 500 mg, Roche, UK) was diluted in one litre of physiological saline attached to a giving set and an i/v catheter, and administered slowly over a period of one hour. For measurement of plasma levels of GCV and PCV, the FCV metabolite, whole blood samples obtained approximately 6-10 hours following drugs administration, were centrifuged at 800 × g for 10 minutes and plasma was collected and stored at −20 °C for later analysis. Plasma samples were analysed at the College of Veterinary Medicine, University of Tennessee, USA using a reverse phase HPLC method with solid phase extraction (SPE). The compounds were separated on an Atlantis T3 (4.6 × 250 mm) column with a mobile phase of 10 mM ammonium phosphate pH 2.9 and acetonitrile (95:5). The flow rate was 1.2 ml/min and fluorescence was measured at excitation 253 nm and emission 360 nm. Standard curves for plasma analysis were prepared by fortifying untreated, pooled elephant plasma with the drug to produce a linear concentration range of 10-5000 ng/ml. Average recovery was 100 % while intra and inter-assay variability ranged from 0.5 to 5.1 %. The lower limit of quantification was 10 ng/ml. Case 1 In 2013, 1.5 year old male Asian elephant calf ‘Scott’ was presented with unwillingness to bend either foreleg, swinging both legs forward when walking. Its ears were drooping, the head was down and the calf was off solid food. An EDTA whole blood sample taken on the day tested qPCR positive for EEHV-1 DNA. Prior to this event, routine blood samples and trunk swabs tested negative for EEHV-1. Using a qPCR assay, the blood viral load was determined as 8.16*103 viral genome copies (vgc)/ml (Fig. 1). Genomic analysis of this EEHV-1 by sequencing of PCR amplicons of partial viral POL, gL and gM genes identified the virus as EEHV-1A genotype. The sequences comparison with those of two EEHV-1A fatalities in 2009 (Riddle and Betts) at the collection revealed several nucleotide substitutions in the POL, gL and gM gene segments, indicating a genomically different virus from those in Riddle and Betts (Fig. 2).Fig. 1 EEHV-1A DNA levels in blood and trunk secretions of Case 1. Clinical suspicion of EEHV infection was confirmed by detection of EEHV-1A DNA in blood. The animal showed clinical signs suggestive of active EEHV-1A infection on Day 0 and Day 1. EEHV-1A DNA load in EDTA whole blood (solid line) and trunk secretions (dashed line) was measured using a qPCR Fig. 2 Nucleotide sequence comparison of the EEHV-1A detected. Partial nucleotide sequences of DNA polymerase [U38/POL (194 bp)], glycoprotein L [U82/gL (381 bp)] and glycoprotein M [U72/gM (500 bp)] from Case 1 and Case 2 were compared with those of two previous EEHV-1A fatalities (Riddle and Betts) from the same zoological collection. Sequences were aligned using the MegAlign 13 software of the DNASTAR Lasergene 13 package (DNASTAR Inc. Madison, USA). Identical nucleotides to those of Case 1 are denoted by dots in the alignments Treatment with 15 mg/kg FCV three times daily (TID) per rectum was instigated eight hours after these initial clinical signs. The first two treatments, six hours apart, consisted of ground FCV tablets suspended in ultrasound gel (Healthlife®, Barclays-Swan Ltd, UK); however, most of the drug was expelled at the next defaecation. Twelve hours after the initial presentation, the calf started to voluntarily bend his right foreleg again, and by the next morning, despite some stiffness in the shoulders, the legs were in normal use and its appetite had returned. The dose of FCV was therefore reduced to 12 mg/kg TID (Day 1); this time dissolved in hand warm water and placed 1.5 meters deep into the rectum. All clinical signs were resolved by Day 2, when the blood viral load peaked at 3.61*104 vgc/ml, after which it started to decrease gradually. The FCV dose was further reduced to 8 mg/kg BID on Day 7. Full haematology at this point showed a relatively low monocyte count (5.94*109/L, 26 % of WBC), but was normal (51 % of WBC) when tested three weeks later (Table 1a). No virus was detectable in the blood from Day 18 onwards (Fig. 1), and FCV treatment ceased on Day 21. Viral DNA excretion through eye (data not shown) and trunk secretions commenced on Day 10 and continued for at least two months, with a peak at around Day 48 of 4.66*103 vgc/test reaction. No further samples were tested beyond Day 57.Table 1 Haematological parameters of Case 1 and Case 2 measured during clinical EEHV-1A infection A Days WBC (*10e9/L) Neutrophils (*10e9/L) Lymphocytes (*10e9/L) Monocytes (*10e9/L) Eosinophils (*10e9/L) Basophils (*10e9/L) 7 22.85 7.77 8.91 5.94 (26 %) 0.23 0 23 13.69 5.75 2.87 6.16 (45 %) 0 0 213 15.91 4.61 5.73 5.25 (33 %) 0.32 0 Physiological mean (% of WBC) 15.51 3.77 (25.5 %) 5.02 (23.65 %) 6.46 (50 %) 0.24 (1.55 %) 0.02 (0.10 %) Days RBC (*10e12/L) HGB (g/dL) HCT (%) MCV (fL) MCH (pg) MCHV (g/dL) RDW (%) PLT (*10e9/L) PCV (%) 7 3.54 13.4 37.7 107 37.8 35.5 14.3 470 36 23 5.06 19.2 55.1 109 37.9 34.8 14.5 968 55 213 3.62 13.3 39.2 108 36.8 33.9 13.9 996 37 Physiological (mean) 3.55 14.28 40.60 114.25 40.19 35.17 14.33 617.65 40.80 B Days EEHV-1A (vgc/ml) WBC (*10e9/L) Neutrophils (*10e9/L) Lymphocytes (*10e9/L) Monocytes (*10e9/L) Eosinophils (*10e9/L) Basophils (*10e9/L) -33 5.0 × 102 23.08 4.62 5.77 12.46 (54 %) 0 0 -29 ND 25.34 4.56 7.6 13.18 (52) % 0 0 -5 1.2 × 105 6.18 3.46 1.85 0.8 (13 %) 0.06 0 -3 2.7 × 106 7.35 5.22 1.32 0.81 (11 %) 0 0 0 7.2 × 106 13.25 8.22 2.65 2.25 (17 %) 0.13 0 1 1.9 × 107 12.925 7.565 2 2.975 (23 %) 0 0.195 Physiological mean (% of WBC) 15.51 3.77 (25.5 %) 5.02 (23.65 %) 6.46 (50 %) 0.24 (1.55 %) 0.02 (0.10 %) Days RBC (*10e12/L) HGB (g/dL) HCT (%) MCV (fL) MCH (pg) MCHV (g/dL) RDW (%) PLT (*10e9/L) PCV (%) -33 3.65 14.6 43.2 118.3 39.9 33.7 13.9 950 42 -29 3.58 14.5 41.4 115.5 40.4 35 38.1 760 38 -5 3.48 14.2 41.8 120.4 41 34 14 560 37 -3 3.36 14.2 38.4 114.4 42.4 37.1 13.7 183 38 0 2.89 12.4 32.9 113.7 42.9 37.8 13.8 98 34 1 3.07 12.95 35.05 114.2 42.2 36.9 13.6 13 36.5 Physiological mean 3.55 14.28 40.60 114.25 40.19 35.17 14.33 617.65 40.80 The haematological parameters for Case 1 (A) were only available for Days 7, 23 and 213 following suspicion of clinical EEHV-1A infection. The haematological parameters and viral load for Case 2 (B) were measured on several occasions from 33 days prior to the observation of signs indicative of clinical EEHV-1A infection (Day 0) to the Day 1 when the elephant succumbed to the infection at midnight. Normal platelet range of live juveniles in the collection was established, during a period of seven months from 20 blood samples of the elephants, to be at 372-878*10e9/L (mean of 625*10e9/L). The rest of the physiological parameters were from 24 measurements from four individuals under eight years of age (age range at time of sampling: 1.5-7y). Parameter values with a substantial change are highlighted in bold. Percentages of monocytes in the white blood cells are also given in the table vgc/ml, viral genome copies per millilitre of EDTA whole blood; WBC, white blood cell count; RBC, red blood cell count; HGB, haemoglobin; HCT, haematocrit; MCV, mean cell volume; MCH, mean corpuscular haemoglobin; MCHV, mean corpuscular haemoglobin volume; RDW, red blood cell distribution width; PLT, platelet count; PCV, packed cell volume; ND, not detected As a pre-emptive measure, the calf’s two siblings, a 3-year old male and a 4-year old female, were treated simultaneously with 12 mg/kg FCV TID from Day 0 to Day 6, which was tapered to 8 mg/kg BID from Day 7 to Day 12, after which treatment ceased. Neither sibling was positive for EEHV-1 DNA in blood or bodily secretions throughout this period. Plasma PCV levels, measured at trough levels 8-10 hours following the latest FCV administrations in both Case 1 (Fig. 3) and its male sibling (data not shown), ranged between 68 and 934 ng/ml. Levels measured at the dosing regimens of 15 mg/kg TID and 12 mg/kg TID reached values considered therapeutic in humans with genital herpes (>100 ng/ml) [27].Fig. 3 Penciclovir, the metabolite of famciclovir, plasma levels in 2 juvenile Asian elephants with EEHV-1A DNAemia. Famciclovir (FCV) was rectally administered in Case 1 at 15 mg/kg twice daily (BID) on observation of clinical signs indicative of EEHV-1A infection (Day 0) and continued at the same dose rate until Day 1, when it was reduced to 12 mg/kg thrice daily (TID). This was followed by further reduction on Day 3 to 8 mg/kg BID and continued unchanged until Day 18. In Case 2 FCV was administered at 15 mg/kg TID for two days following the detection of EEHV-1A DNA in the blood (Day -5), five days prior to observation of EEHV-1A clinical signs, and continued at the same dosage till the last dose of FCV at 9:00 am on Day 0; blood was only collected at 17:00 pm on that day. The death of Case 2 was at midnight on Day 1; seven days after first detection of EEHV-1 DNA in the blood. Dashed line indicates the minimum level of PCV considered therapeutic in humans with genital herpes (>100 ng/ml) [27] Case 2 In 2015, a blood swab collected during routine weekly sampling of 1.5 year old male calf ‘Max’ tested positive for EEHV-1 DNA by qPCR. This result was confirmed by a subsequent EDTA whole blood sample that tested qPCR positive at 5.0*102 vgc/ml of blood. At this point monocytes were abundant (12.46*109/L, 54 % of WBC) and the platelet count was 950*109/L. (Table 1b, Day -33). A follow-up blood sample four days later was EEHV-1 DNA negative by qPCR and the calf remained negative on four subsequent blood swabs taken during a three-week period. On week 4 and one day prior to the calf being diagnosed again with EEHV-1 DNAemia the calf was presented with a swollen right temporal gland, which was painful on palpation and from which purulent discharge could be expressed. The gland was treated with pevidine (Sanofi, UK) flushes and topical amoxicillin/clavulanic acid application following bacteriological culture and sensitivity testing. After this positive EEHV qPCR result from a blood swab, an EDTA whole blood sample also tested positive, with a viral load of 1.2*105 vgc/ml, the monocyte count had markedly reduced to 0.8*109/L (13 % of WBC) and the platelet count almost halved to 560*109 /L, in the absence of clinical signs (Day -5). PCR results for this sample became available two days after the actual sampling date and treatment was immediately instigated at 15 mg/kg FCV TID per rectum. Blood samples were then taken regularly for EEHV-1 qPCR, viral quantification and haematology. Genomic analysis of this EEHV-1 by sequencing of PCR amplicons of partial viral POL, gL and gM genes identified the virus as EEHV-1A genotype. However, comparing these nucleotide sequences to those of Case 1, it was revealed that the virus was genomically slightly different, but both cases were more dissimilar to Riddle and Betts (Fig. 2). Nucleotide sequences of the partial segments of the three genes for Riddle and Betts were identical. This was not unexpected as the two fatalities occurred within two weeks of each other, but longer sequence reads would be necessary to verify this further. Viral load increased exponentially as time progressed, during which the monocyte count stabilised around 2.54*109/L (17 % of WBC), but platelet counts showed a continuous decline (Table 1b). Lymphocyte count has also shown substantial reduction on the detection of DNAemia, but other haematological parameters such as neutrophil and basophil counts remained unchanged. PCV plasma levels as high as 934 ng/ml were attained during the first days of treatment (Fig. 3). Five days into the DNAemia, the calf started to show signs of lethargy. At this point, PCV plasma levels had dropped to 303 ng/ml, about a third of its previous peak. The decision was made to start i/v fluids, consisting of Harmann’s (15 ml/kg/h, Vetivex 11, Dechra, UK) supplemented with Duphalyte (1 ml/kg BW, Pfizer, UK), and complement the treatment from solely FCV per rectum to i/v plasma and GCV. The calf first displayed clinical signs of haemorrhagic disease six days into the DNAemia: oedema of the head and ventral abdomen, tongue petechiae and progressive signs of cyanosis. The platelet count had also reduced drastically at this point to 98*109/L and viral load had increased to 7.2*106 vgc/ml of blood. Intravenous cannulas were in place in both ears for administration of 1 ml/kg plasma TID, fluids and antibiotics (2 mg/kg marbofloxacin, Marbocyl®, Vetoquinol). Due to logistics, GCV (5 mg/ml) could not be administered until 16:00h that day, but achieved plasma levels of 2529 ng/ml following i/v injection. Clinical signs progressed despite treatment and the platelet count further reduced to 13*109/L and viral load increased to 1.9*107 vgc/ml of blood. The calf subsequently died at midnight, just into the next day as a result of hypoxaemia and hypovolaemia. Discussion The early clinical signs of EEHV-HD are non-pathognomonic and often manifested as lethargy, anorexia and lameness. As a consequence of generalized capillary endothelial cell damage, these early clinical signs are likely to be followed by more severe and generally terminal signs [5, 8, 12]. Routine blood testing of at-risk Asian elephants for EEHV viraemia has been suggested as part of a monitoring scheme to predict and potentially prevent fatal disease, as the viral DNA can be detected at least several days prior to any clinical signs [13, 14]. This routine monitoring was proven to be valuable in this study, where Case 2 continuous EEHV-1 DNAemia was detected five days prior to the earliest clinical signs, although unfortunately this early initiated treatment was not able to prevent fatality. In addition to detection and quantification of viral DNA through real-time qPCR, drop in the monocyte/neutrophil ratio, and absolute reduction of the platelet count, make both the monocytes and platelets the most useful prognostic haematological parameters of EEHV-HD, although the exact physiological process for these changes is as yet unknown. It is possible that the monocytes are subjected to immediate assault of the virus [19], and that the viral damage to the endothelial cells beyond the adaptive nature of the host response leads to the development of disseminated intravascular coagulation (DIC) and hence critical loss of platelets. In Case 2, the platelet count showed a gradual decline well before the fatal progressive viraemia, reaching a dangerously low level before death. Archived data from three previous EEHV fatalities at the same collection show a similar critically low platelet count following clinical EEHV-1 manifestation (Table 2). Similar findings have also been recently reported from Asian elephants affected by EEHV-1B, 4 and 5 viraemia [13, 18, 19]. It is noteworthy that mean haematological parameters vary substantially in Asian elephants’ blood, e.g. a reported range of 80-975*109/L for platelets [31–35]. Monocytes are usually the most abundant white blood cells in the majority of juveniles. In these cases, the relative drop of monocytes, compared to neutrophils, in the early stages of the disease could easily be identified on straightforward in-house white blood cell differentials. There are, however, individuals in which the monocytes are on average less abundant and hence this early indicator will be less beneficial in these cases. It is therefore advisable to establish individual baselines for at least the abovementioned haematological parameters for each at-risk elephant in advance, and that clinicians familiarise themselves with in-house blood smear differentials.Table 2 Haematological parameters of three additional fatal cases of EEHV-1A and EEHV-1B infections from the same zoological collection Animals (date of sampling) Virus type WBC (*10e9/L) Neutrophils (*10e9/L) Lymphocytes (*10e9/L) Monocytes (*10e9/L) Eosinophils (*10e9/L) Basophils (*10e9/L) Emilia (18/12/2006) EEHV-1B 17 11.85 3.21 1.9 (11.2 %) 0.03 0.02 Riddle (01/05/2009) EEHV-1A 2.71 1.49 0.84 0.27 (10 %) 0 0 Betts (15/05/2009) EEHV-1A 5.1 3.11 1.02 0.97 (19 %) 0 0 Physiological mean (% of WBC) 15.51 3.77 (25.5 %) 5.02 (23.65 %) 6.46 (50 %) 0.24 (1.55 %) 0.02 (0.10 %) Animals (date of sampling) RBC (*10e12/L) HGB (g/dL) HCT (%) MCV (fL) MCH (pg) MCHV (g/dL) RDW (%) PLT (*10e9/L) PCV (%) Emilia (18/12/2006) 3.05 12.8 39.2 122 42 34.4 15.7 34.6 NA Riddle (01/05/2009) 3.55 13 37.7 106 36.7 34.6 17.4 25.6 35 Betts (15/05/2009) 3.16 12.9 55.1 112 40.8 36.5 17.3 71.5 NA Physiological (mean) 3.55 14.28 40.60 114.25 40.19 35.17 14.33 617.65 40.80 Blood samples were taken following the observation of EEHV-1 infection clinical signs. WBC, lymphocyte, monocyte and platelet counts (highlighted in bold) consistently show substantial drop from their physiological level in the two fatalities caused by EEHV-1A. Percentages of monocytes in the white blood cells are also given in the table WBC, white blood cell count; RBC, red blood cell count; HGB, haemoglobin; MCV, mean cell volume; MCH, mean corpuscular haemoglobin; MCHV, mean corpuscular haemoglobin volume; RDW, red blood cell distribution width; PLT, platelet count; PCV, packed cell volume; NA, not available Monitoring kinetics of EEHV-1A load in the two elephants revealed rapid increase in viral load in the elephant that succumbed to the disease (Case 2) despite continuous anti-herpesviral treatment up to the EEHV-1 clinical signs. From at least three days prior to onset of clinical signs, PCV reached plasma levels suggested as therapeutic for treatment of diseases caused by alphaherpesviruses infections in humans [27], but here it had failed to control virus replication as the virus load continued to increase. Considering the approximate three hour half-life of PCV [27], the peak levels could have been up to three times higher than measured, as blood samples were taken 6-8 hours following the FCV administration. The efficacy of FCV treatment to curb progression to EEHV-HD in clinical EEHV has always been debatable, as compromised cellular function at enteric level could affect absorption. This was indicated by the measured PCV level in Case 2, which dropped remarkably 48 hours prior to the onset of outwardly visible clinical signs. Factors such as drug expulsion and individual differences in drug metabolism are also considered possible causes of this drop [24]. These findings therefore favour those anti-herpesviral drugs that could be administered intravenously, to bypass the intestinal absorption step. Additionally, the available sequences for EEHV TK and PK have only 25 % amino acid identity with their orthologues in HSV and HCMV, target viruses for these drugs [9] and indirect surrogate in vitro assays have shown mild sensitivity for only the PK against PCV [36]. Taking into account the EEHV viraemia sequel following FCV administration in current and previous studies, the efficacy of FCV remains doubtful. Therefore the use of other, i/v administered and potentially more effective, anti- herpesviral therapies should be explored, especially in cases where pre-EEHV-HD viraemia, established through routine blood PCR monitoring, does not appear to be affected by FCV within the first days of treatment. In the absence of comprehensive data on a highly effective and safe anti-herpesviral drug, alternative drugs at present could include i/v administered nucleoside analogues such as GCV and ACV. The second-line drugs could be the pyrophosphate analogue foscarnet and the nucleotide analogue cidofovir, but these drugs all have their own safety issues. The most common adverse effect during GCV therapy in humans has been haematological toxicity, but this appears to be readily reversible on discontinuation of the drug [37]. It should also be noted that ACV is a hundred times more potent than FCV in inhibiting herpesviruses replication, but it suffers from a shorter half-life and intracellular concentration [38]. This drug has been recommended for clinical situations in humans where a rapid decrease in viral load is desirable, and could be of similar therapeutic value in EEHV clinical infections. The renal toxicity of the drug should also be considered in its therapy. Frequent measurement of EEHV load in blood samples is therefore of paramount value to establish the efficacy of ongoing therapy. Complementary therapies that are likely to be of value include supportive therapies e.g. maintaining hydration and actions to prevent DIC and shock, such as i/v plasma administration from EEHV antibody positive adult donors, as a source of both antibodies and platelets, and a broad spectrum antibiotic to combat potential concurrent bacterial infections that might possibly be at the root of triggering the viral replication, predisposing to the infection or arise as a secondary infection. Intravenous access appears to be of significant value for supportive treatment and anti-viral drug administration, and is straightforward in trained elephants that are used to being handled. The EEHV Advisory Group convened to debate this in 2014 and currently recommends minimum standards of care for all at-risk elephants irrespective of management systems, which include a section on Calf Training for intravenous access (www.eehvinfo.org). In elephants that have not attained this stage of training at the time of infection, repeated standing sedations using butorphanol and detomidine are possible, as used successfully in other institutions (Lauren Howard, personal communication). As far as antibodies are concerned, Case 1 had been positive on an indirect ELISA for glycoprotein B (gB) about 9 months prior to the clinical signs and survived the viraemia [17]. On the other hand, Case 2 was also highly positive on the same gB ELISA (van den Doel, pers. communication) at the time of EEHV infection, but succumbed to the virus. For other closely related betaherpesvirus infections (e.g. HCMV), antibodies to the gB protein seemed to enhance protection by curbing at least the dissemination of the virus in the blood [29]. However, considering the fatality in Case 2 despite high anti-gB antibody, sole measurement of gB antibody may not be an indicator of protection against EEHV, as it is currently unknown which viral proteins are targeted by neutralising antibodies. The possibility also exists that early IgG antibodies to EEHV show a low avidity for the antigen and hence an avidity test is required to determine the value of antibodies against EEHV [39]. Also, in the view that herpesviruses (EEHVs and elephant gamma herpesviruses [EGHVs]) are ubiquitous in Asian elephant populations, the notion that the above mentioned ELISA is cross reactive with gB of similar viruses or other unknown agents cannot be excluded [9]. In this study, a detectable EEHV-1 DNAemia was correlated with the disease symptoms only when the blood viral load was at around 8.16*103 vgc/ml of blood for Case 1 and 7.2*106 vgc/ml for Case 2. A previous report states that a threshold of around 1.0*104 vgc/ml of blood had coincided with apparent illness [14]. Clinical signs appeared at a much higher viral load in Case 2, emphasising a daily rapid increase in viral load as a more favourable determinant of initiating anti-herpesviral therapy, its success and disease prognosis. Our results also further endorse that DNAemia precedes virus shedding in trunk secretions of EEHV-1A infected animals, similar to those reported previously [13, 14, 18, 19]. Although there is no evidence that infectious particles are shed at a level sufficient to infect an in-contact animal, the potential risk these carrier elephants may pose following a clinical EEHV case should be considered. Conclusions The EEHV viraemic cases investigated here further highlight the ongoing threat posed by the virus to the juvenile Asian elephants, potentially leading to gradual disappearance of their free-living population and an inability to maintain a viable captive population to safeguard the species from extinction. Therefore, effective control of future EEHV infections is fundamental to improve the survival rate of affected animals. Routine EEHV testing combined with viral quantification, WBC, monocyte and platelet counts following confirmed DNAemia in at-risk elephants should deliver an early warning of the clinical disease and prognosis, providing enough time to evaluate treatment opportunities to avert a fatal outcome. This investigation further questions the efficacy of the rectally administered FCV in controlling the viraemia and urges towards the exploration of alternative therapies. These therapies could include preferably i/v administered anti-herpes viral drugs, complementary plasma transfusions and antibiotics against potential secondary infections. Abbreviations BIDBis in die (twice a day) DNADeoxyribonucleic acid EDTAEthylenediaminetetraacetic acid HPLCHigh performance liquid chromatography i/vIntravenous PBSPhosphate buffered saline PCRPolymerase chain reaction qPCRQuantitative PCR TIDTer in die (thrice a day) Acknowledgements We are grateful to Nic Masters, Jo Dodds and Karla Berry for veterinary support and to Lee Sambrook and his excellent team of elephant keepers for sample collection, which has made this study possible. We also thank Dr. Sherry Cox at the College of Veterinary Medicine, the University of Tennessee, USA for the penciclovir and ganciclovir measurements. Provision of the gB ELISA results by Petra van den Doel from Erasmus Medical Center, The Netherlands is highly appreciated. Many thanks also to Christopher Finnegan, APHA for performing some of the qPCRs. The funding bodies played no role in the design of this study, analysis and interpretation of data; in the writing of the manuscript; or in the decision to submit the manuscript for publication. Funding Funding for this work was partly provided by the Zoological Society of London, Chester Zoo and Rotterdam Zoo. Availability of data and material All the data are presented in the main paper and accompanying tables and figures. Authors’ contribution KSM performed the laboratory investigation. FM undertook the clinical work on the elephants. AD, KSM and FM led the manuscript preparation. KSM, AD, FM, KD, FS contributed to the interrogation of results and manuscript preparation. All authors read and approved the final manuscript. Competing interests The author(s) declare that they have no competing interests. Ethics approval and consent to participate All samples used for this study were for clinical purposes under the Veterinary Surgeons Act 1966, United Kingdom. Veterinary interventions such as treatment, chemical restraint and blood sample collection for the purpose of health screening or disease treatment fall under the Veterinary Surgeons Act 1966 and do not require further ethical approval, as they are intended to maintain the well-being of animals. ==== Refs References 1. Davison AJ Eberle R Ehlers B Hayward GS McGeoch DJ Minson AC The order Herpesvirales Archives of Virology 2009 154 171 177 10.1007/s00705-008-0278-4 19066710 2. Richman LK Hayward GS Fowler ME Miller RE Elephant herpesviruses Zoo and wild animal medicine 2012 7 496 502 3. Richman LK Montali RJ Garber RL Kennedy MA Lehnhardt J Hildebrandt T Novel endotheliotropic herpesviruses fatal for Asian and African elephants Science 1999 283 1171 1176 10.1126/science.283.5405.1171 10024244 4. Fickel J Richmann LK Montali RJ Schaftenaar W Goritz F Hildebrandt TB A variant of the endotheliotrophic herpesvirus in Asian elephants (Elephas maximus ) in European zoos Vet Microbiol 2001 82 2001 10.1016/S0378-1135(01)00363-7 5. Garner MM Helmick K Ochsenreiter J Richman LK Latimer E Wise AG Clinico-pathological features of fatal disease attributed to new variants of endotheliotropic herpesvirus in two Asian elephants (Elephas maximus ) Vet Pathol 2009 46 97 104 10.1354/vp.46-1-97 19112123 6. Latimer E Zong JC Heaggans SY Richman LK Hayward GS Detection and evaluation of novel herpesviruses in routine and pathological samples from Asian and African elephants: identification of two new probosciviruses (EEHV5 and EEHV6) and two new gammaherpesviruses (EGHV3B and EGHV5) Vet Microbiol 2011 147 1-2 28 41 10.1016/j.vetmic.2010.05.042 20579821 7. Wilkie GS Davidson AJ Wason M Kerr K Sanderson S Bouts T Complete genome sequences of elephant endotheliotropic herpesviruses 1A and 1B determined directly from fatal cases J Virol 2013 87 12 6700 6712 10.1128/JVI.00655-13 23552421 8. Wilkie GS Davidson AJ Kerr K Stidworthy MF Redrobe S Steinbach F First fatality associated with elephant endotheliotropic herpesvirus 5 in an Asian elephant: Pathological findings and complete viral genome sequence Sci Rep 2014 4 6299 10.1038/srep06299 25199796 9. Long SY Latimer EM Hayward GS Review of elephant endotheliotropic herpesviruses and acute haemorrhagic disease ILAR J 2015 56 3 283 96 10.1093/ilar/ilv041 26912715 10. Zachariah A Zong JC Long SY Latimer EM Heaggans SY Richman LK Fatal herpesvirus haemorrhagic disease in wild and orphan Asian elephants in Southern India J Wildl Dis 2013 49 2 381 393 10.7589/2012-07-193 23568914 11. Zong JC Latimer E Heaggans SY Richman LK Hayward GS Pathogenesis and molecular epidemiology of fatal elephant endotheliotropic disease associated with the expanding Proboscivirus genus of the Betaherpesvirinae 2007 Orlando International Elephant Conservation and Research Symposium 23 35 12. Richman LK Montali RJ Cambre RC Schmitt D Hardy D Hildebrandt TB Clinical and pathological findings of a newly recognised disease of elephants caused by endotheliotropic herpesviruses J Wildlife Dis 2000 36 1 12 10.7589/0090-3558-36.1.1 13. Atkins L Zong JC Tan J Mejia A Heaggans SY Nofs SA Elephant endotheliotropic herpesvirus 5, a newly recognized elephant herpesvirus associated with clinical and subclinical infections in captive Asian elephants (Elephas maximus ) J Zoo Wildl Med 2013 44 1 136 43 10.1638/1042-7260-44.1.136 23505714 14. Stanton JJ Zong JC Eng C Howard L Flanagan J Stevens M Kinetics of viral loads and genotypic analysis of elephant endotheliotropic herpesvirus-1 infection in captive Asian elephants (Elephas maximus ) J Zoo Wildlife Med 2013 44 1 42 54 10.1638/1042-7260-44.1.42 15. Stanton JJ Zong JC Latimer E Tan J Herron A Hayward GS Detection of pathogenic elephant endotheliotropic herpesvirus in routine trunk washes from healthy adult Asian elephants (Elephas maximus ) by use of real-time quantitative polymerase chain reaction assay Am J Vet Res 2010 71 925 933 10.2460/ajvr.71.8.925 20673092 16. Hardman K Dastjerdi A Gurrala R Routh A Banks M Steinbach F Detection of elephant endotheliotropic herpesvirus type 1 in asymptomatic elephants using TaqMan real-time PCR Vet Rec 2012 170 8 205 10.1136/vr.100270 22186378 17. van den Doel P Prieto VR van Rossum-Fikkert SE Schaftenaar W Latimer E Howard L A novel antigen capture ELISA for the specific detection of IgG antibodies to elephant endotheliotropic herpesvirus BMC Vet Res 2015 11 203 10.1186/s12917-015-0522-6 26268467 18. Fuery A Tan J Peng R Flanagan JP Tocidlowski ME Howard LL Ling PD Clinical infection of two captive Asian elephants (Elephas maximus ) with elephant endotheliotropic herpesvirus 1B J Zoo Wildl Med 2016 47 1 319 24 10.1638/2015-0074.1 27010294 19. Fuery A Browning GR Tan J Long S Hayward GS Cox SK Clinical infection of captive Asian elephants (Elephas maximus ) with elephant endotheliotropic herpesvirus 4 J Zoo Wildl Med 2016 47 1 311 8 10.1638/2015-0072.1 27010293 20. Ehlers B Dural G Marschall M Schregel V Goltz M Hentschke J Endotheliotropic elephant herpesvirus, the first betaherpesvirus with a thymidine kinase gene J. Gen. Virol. 2006 87 2781 2789 10.1099/vir.0.81977-0 16963736 21. Schmitt DL Hardy DA Montali RJ Richman LK Lindsay WA Isaza R Use of famciclovir for the treatment of endotheliotropic herpesvirus infections in Asian elephants (Elephas maximus ) J Zoo Wildl Med 2000 31 518 522 10.1638/1042-7260(2000)031[0518:UOFFTT]2.0.CO;2 11428400 22. Schaftenaar W, Mensink JMCH, de Boer AM, Hildebrandt TB, Fickel J. Successful treatment of a subadult Asian elephant bull (Elephas maximus) infected with the endotheliotropic elephant herpes virus. Proceedings. 40th Int Symp Dis Zoo Wild Anim 2001, 141–146. 23. Tocidlowski ME, Marin M, Flanagan JP, Howard LL. Medical monitoring and treatment of clinical EEHV viraemia in juvenile Asian elephants (Elephas maximus) at the Houston zoo. Proceedings. 10th International Elephant Endotheliotropic Herpesvirus (EEHV) Workshop 2015, Houston, USA. Page 38 24. Howard LL, Cox S. Penciclovir levels in Asian elephants undergoing famciclovir treatment for clinical EEHV viremia at the Houston Zoo. Proceedings. 10th International Elephant Endotheliotropic Herpesvirus (EEHV) Workshop 2015, Houston, USA. Pages 33-34 25. Padilla, LR. Use of ganciclovir in Asian elephants for treatment of EEHV infection. Proceedings. 10th International Elephant Endotheliotropic Herpesvirus (EEHV) Workshop 2015, Houston, USA. Page 32 26. Boonprasert K, Mahasawangku1 S, Angkawanish T, Jansittiwej S, Langkaphin W, Sombutputorn P, et al. Endotheliotropic herpesvirus treatments in a wild orphan baby Asian elephant (Elephas maximus). Proceedings. 10th International Elephant Endotheliotropic Herpesvirus (EEHV) Workshop 2015, Houston, USA. Page 35 27. Gill KS Wood MJ The clinical pharmacokinetics of famciclovir Clin Pharmacokinet 1996 31 1 8 10.2165/00003088-199631010-00001 8827396 28. Brock AP Isaza R Hunter RP Richman LK Montali RJ Schmitt DL Estimates of the pharmacokinetics of famciclovir and its active metabolite penciclovir in young Asian elephants (Elephas maximus ) Am J Vet Res 2010 73 12 1996 1999 10.2460/ajvr.73.12.1996 23176429 29. Field HJ Vere Hodge RA Recent developments in anti-herpesvirus drugs Br Med Bull 2013 106 213 49 10.1093/bmb/ldt011 23596085 30. Seilern-Moy K Darpel K Steinbach F Dastjerdi A Distribution and load of elephant endotheliotropic herpesviruses in tissues from associated fatalities of Asian elephants Virus Res. 2016 220 91 96 10.1016/j.virusres.2016.04.012 27102836 31. Silva ID Kuruwita VY Haematology, plasma, and serum biochemistry values in domesticated elephants (Elephas maximus ceylonicus ) in Sri Lanka Gajah 1994 12 47 32. Niemuller C Gentry PA Liptrap RM Longitudinal study of haematological and biochemical constituents in blood of the Asian elephants (Elephas maximus ) Comp Biochem Physiol 1990 96A 131 134 10.1016/0300-9629(90)90053-U 33. Lewis JH Comparative haematology: Studies on elephants, Elephas maximus Comp Biochem Physiol 1974 49A 175 181 10.1016/0300-9629(74)90553-2 34. Brown IRF White PT Elephant blood haematology and chemistry Comp Biochem Physiol 1980 65B 1 12 35. Wiedner E. Proboscidea. In: Fowler’s zoo and wildlife medicine, 8th Edition, Eds Fowler ME & Miller RE. 2015, 524 36. Ackermann M, Schetle N. In vitro investigation into antiviral efficacy and EEHV. Proceedings. 10th International Elephant Endotheliotropic Herpesvirus (EEHV) Workshop 2015, Houston, USA. Pages 31 37. Markham A Faulds D Ganciclovir. An update of its therapeutic use in cytomegalovirus infection Drugs 1994 48 3 455 84 10.2165/00003495-199448030-00009 7527763 38. Cassady KA Whitely RJ New therapeutic approaches to the alphaherpesvirus infections J Antimicrob Chemother 1997 39 2 119 128 10.1093/jac/39.2.119 9069530 39. Prince HE Lapé-Nixon M Role of cytomegalovirus (CMV) IgG avidity testing in diagnosing primary CMV infection during pregnancy Clin Vaccine Immunol 2014 21 10 1377 84 10.1128/CVI.00487-14 25165026

PMC005xxxxxx/PMC5002105.txt

==== Front BMC CancerBMC CancerBMC Cancer1471-2407BioMed Central London 273110.1186/s12885-016-2731-1Research ArticleChanges in the in vitro activity of platinum drugs when administered in two aliquots Al-Eisawi Zaynab Zaynab.al-eisawi@sydney.edu.au 16Beale Philip Philip.Beale@sswahs.nsw.gov.au 2Chan Charles Charles.chan@sswahs.nsw.gov.au 3Yu Jun Qing Jun.Yu@sydney.edu.au 1Proschogo Nicholas n.proschogo@chem.usyd.edu.au 4Molloy Mark mark.molloy@mq.edu.au 5Huq Fazlul +61 2 9351 9522Fazlul.Huq@sydney.edu.au 171 Discipline of Biomedical Science, Sydney Medical School, University of Sydney, Sydney, NSW 2141 Australia 2 Sydney Cancer Centre, Concord Hospital, Sydney, NSW 2139 Australia 3 Department of Pathology, Concord Hospital, Sydney, NSW 2139 Australia 4 Mass Spectrometry Unit, School of Chemistry, University of Sydney, Sydney, NSW 2006 Australia 5 Australian Proteome Analysis Facility, Macquarie University, Sydney, NSW 2109 Australia 6 Department of Medical Laboratory Sciences, Faculty of Allied Health Science, Hashemite University, Zarqa, Hashemite Kingdom of Jordan 7 Discipline of Biomedical Science, School of Medical Sciences, Sydney Medical School, The University of Sydney, Cumberland Campus C42, 75 East Street, Lidcombe, NSW 1825 Australia 26 8 2016 26 8 2016 2016 16 1 6888 8 2014 28 6 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 management of ovarian cancer remains a challenge. Because of the lack of early symptoms, it is often diagnosed at a late stage when it is likely to have metastasized beyond ovaries. Currently, platinum based chemotherapy is the primary treatment for the disease. However acquired drug resistance remains an on-going problem. As cisplatin brings about apoptosis by intrinsic and extrinsic pathways, this study aimed to determine changes in activity of platinum drugs when administered in two aliquots as against a bolus and sought to determine association with changes in GSH, speciation of platinum drugs and changes in protein expression. Methods The efficacy of administering cisplatin, carboplatin and oxaliplatin in two aliquots with a time gap was investigated in ovarian A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines. The cellular accumulation of platinum, level of platinum − DNA binding and cellular glutathione level were determined, and proteomic studies were carried out to identify key proteins associated with platinum resistance in ovarian A2780cisR cancer cell line. Results Much greater cell kill was observed with solutions left standing at room temperature than with freshly prepared solutions, indicating that the increase in activity on ageing was related to speciation of the drug in solution. Proteomic studies identified 72 proteins that were differentially expressed in A2780 and A2780cisR cell lines; 22 of them were restored back to normal levels as a result of synergistic treatments, indicating their relevance in enhanced drug action. Conclusions The proteins identified are relevant to several different cellular functions including invasion and metastasis, cell cycle regulation and proliferation, metabolic and biosynthesis processes, stress-related proteins and molecular chaperones, mRNA processing, cellular organization/cytoskeleton, cellular communication and signal transduction. This highlights the multifactorial nature of platinum resistance in which many different proteins with diverse functions play key roles. This means multiple strategies can be harnessed to overcome platinum resistance in ovarian cancer. The results of the studies can be significant both from fundamental and clinical view points. Keywords Cisplatin (CS)Carboplatin (CB)Oxaliplatin (OX)Drug combinationSynergismDrug uptakeDrug resistanceAging effectissue-copyright-statement© The Author(s) 2016 ==== Body Background Platinum-based drugs cisplatin (CS), carboplatin (CB), and oxaliplatin (OX) are routinely used in the clinic to treat various cancers including testicular, ovarian, lung, bladder, colon, head and neck cancers [1]. However, efficacy is limited by dose limiting toxicities and acquired drug resistance [2] that may arise due to decreased cellular accumulation of platinum drugs, inactivation by conjugation with glutathione or sequestration involving metallothionein, enhanced tolerance to platinum-DNA adducts and enhanced DNA repair mechanisms [2, 3]. Cancer related pathways are bound to be highly complex often involving both intrinsic and extrinsic pathways [4]. As applied to cell death caused by CS, it was suggested that depending on the status of the cell, different pathways would become more significant at different time points. We proposed that the administration of first aliquot of CS would place cancer cells under increased oxidative stress caused by depletion of cellular thiols due to their binding with the drug [5] and if so when the second aliquot was administered after a brief time period (2 to 4 h), depleted glutathione level would allow more of the drug to bind with DNA resulting into increased apoptosis. Thus, the sequenced administration of CS in two aliquots with a small time gap could be looked upon as being the combination of two drugs with somewhat different mechanisms of action [5]. The present study aimed to determine the efficacy of the administration of CS, CB and OX (Fig. 1) in two aliquots with time gaps of 2, 4, 8, 24 h in ovarian tumour models. We also sought to determine whether the use of ‘aged solutions’ of the drugs (where the solutions for the second aliquots were left standing at room temperature for the duration of the time gap) had a similar or greater effect on cell kill. The rationale behind using both fresh and aged solutions was to determine the effect of hydrolysis of platinum drugs on the combined drug action. Although platinum − DNA binding is believed to be an essential step in CS induced apoptosis, the programmed cell death is brought about downstream by multiple proteins. Thus, the study also aimed to determine changes in expression of key proteins associated with drug resistance in ovarian cancer cell lines.Fig. 1 Chemical structures of cisplatin, carboplatin and oxaliplatin Methods Materials CB and OX were obtained from Sigma Aldrich, Sydney, Australia. CS was synthesized according to previously described method [6]. Foetal calf serum (FCS), RPMI-1640, 200 mM L-glutamine, and 5.6 % sodium bicarbonate were obtained from Trace Biosciences Pty Ltd Australia. DNA extraction kit JETQUICK Blood DNA Spin Kit/50 was obtained from Astral Scientific Pty Ltd, Sydney, Australia. GSH/GSSG-Glo™ assay kit was obtained from Promega, Sydney, Australia. Other chemicals were obtained mostly from Sigma-Aldrich, Sydney, Australia. Ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines were gifts from Ms. Mei Zhang, Royal Prince Alfred Hospital, Sydney, Australia. Stock solutions of platinum drugs were prepared to a final concentration of 1 mM; CS was first dissolved in DMF then made up in milli-Q water to a final ratio of 1:4 DMF to milli-Q water, whereas CB and OX were prepared in milli-Q water only. Stock solutions were then filtered to insure sterility. Cell culture Human ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines (Table 1) were seeded in 25 cm2 tissue culture flasks in an incubator at 37 °C in a humidified atmosphere consisting of 5 % CO2 and 95 % air. The cells were maintained in logarithmic growth phase in complete medium consisting of RPMI 1640, 10 % heat inactivated FCS, 20 mM Hepes, 0.11 % bicarbonate, and 2 mM glutamine without antibiotics [7]. Each cell line was seeded at a density of 4–6 × 103 cells/well in flat-bottomed 96-well culture plate in 10 % FCS/RPMI 1640 culture medium. The plate was incubated for 24 h at 37 °C in a humidified atmosphere allowing cells to attach.Table 1 Human ovarian cancer cell lines used in this study Cell line Phenotype A2780 Untreated ovarian tumour A2780cisR CS resistant ovarian tumour A2780ZD0473R ZD0473a resistant ovarian tumour SKOV-3 Oestrogen receptor positive ovarian tumour aZD0473 (also known as JM473 and AMD 473) is a sterically hindered platinum complex with a cis-geometry like cisplatin Cytotoxicity assay MTT reduction assay was carried out to determine cytotoxicity of CS, CB and OX administered as a bolus and in two aliquots with a time gap. Stock solutions of drugs were subjected to serial dilutions to give final concentrations ranging from 0.16 to 250 μM. The dilutions were performed using 10 % RMPI-1640 medium without serum as the vehicle and were added to equal volumes of cell culture in triplicate wells and then cells were left to incubate for 72 h. These treatments were carried out to determine IC50 values i.e. drug concentrations required for 50 % cell kill. In treatment in two aliquots with a time a gap, cells were treated with solutions of CS, CB and OX at three different concentrations based on their IC50 values. The first aliquot administered at time zero was that of a freshly prepared solution (denoted as ‘fresh’) whereas the second aliquot administered at 2 h (0/2 h), 4 h (0/4 h), 8 h (0/8 h) or 24 h (0/24 h) was using either freshly prepared solution (fresh) so that the combination was termed fresh/fresh or aged solution left at room temperature for the period of the time gap (denoted as ‘aged’) so that the combination was termed fresh/aged. Cells in drug free medium served as control. The rationale behind doing experiments with both fresh and aged solutions was to determine the effect of hydrolysis of platinum drugs on the combined drug action. The period of drug treatment was 72 h counted from the time of administration of the first dose. Cell growth inhibition was determined using the MTT reduction assay. Combination index values (CIs) were used as measures of synergism, additiveness or antagonism calculated using the program CalcuSyn [8–10] and previously described method [11]. Platinum accumulation and platinum-DNA binding Cellular accumulation of platinum and platinum − DNA binding levels in A2780 and A2780cisR cell lines were determined as applied to administration of CS in two aliquots with a time gap of 2 and 4 h and at a final concentration of 50 mM, for both fresh/fresh and fresh/aged combinations. The drug was added to culture plates containing exponentially growing A2780 and A2780cisR cells in 10 ml 10 % FCS/RPMI 1640 culture medium (cell density = 5 × 106 cells/ml). The cells containing the drug were incubated for 24 h at the end of which cell monolayers were collected and cell suspensions (10 ml) were transferred to centrifuge tube and spun at 3500 rpm for 2 min at 4 °C. The cells were washed twice with ice-cold phosphate-buffered saline (PBS) and the pellets were stored at −20 °C until assayed. At least three independent experiments were performed. Cellular accumulation Following drug incubation the cell pellets were suspended in 0.5 ml 1 % triton-X, held on ice then sonicated. Total intracellular platinum contents were determined by graphite furnace atomic absorption spectrophotometry. Drug–DNA binding High molecular weight DNA from cell pellets were isolated using JETQUICK Blood DNA Spin Kit/50 (Astral Scientific, Australia) according to the modified protocol of Bowtell [12]. Platinum contents of the samples were determined by graphite furnace AAS. A260/A280 ratios were found to be between 1.75 and 1.8 for all samples, indicating high purity of the DNA. Cellular glutathione As a measure of cellular health and the redox state of the cells, the levels of total glutathione (GSH and GSSG) and oxidised glutathione (GSSG) in A2780 and A2780cisR cell lines were determined as applied to treatments with CS and CB administered in two aliquots with a time gap of 4 h. Drugs made in 10 % RMPI-1640 serum free medium were added to equal volumes of cell culture wells of a white wall clear bottom 94 well plate containing exponentially growing A2780 and A2780cisR cells (cell density = 12 × 103 cells/well). Cells were left to incubate for 24 h. The medium was aspirated out of the treatment wells with minimal disturbance of the cell pellets and cells were washed with 200 μl of PBS, then the levels of glutathione were determined using the GSH/GSSG-Glo™ Assay kit (Promega, Australia). The plate was read in a LUMIstar Omega luminometer (BMG LABTECH, USA). Mass spectral analysis Mass spectrometry was used to explore hydrolysis of OX rather than that of CS as OX and its hydrolysis products were more sensitive to mass spectral measurements than those of CS. Solution of OX first made in milli-Q water, was diluted (1:2) with methanol, cell culture medium, or pH adjusted (neutral) milli-Q water. The solutions were injected by syringe pump (flow rate 150 ul/h) into a Bruker Apex Qe 7 T Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTICR) in positive ion electrospray ionization mode. The instrument was optimized and externally mass calibrated before use. The presence and evolution of OX species was monitored over a period of 4 h, with measurements made every hour. Proteomics Proteomic studies were carried to determine the proteins that were differentially expressed in the parent A2780 and cisplatin-resistant A2780cisR cell lines but were restored back due to treatment with CS in two aliquots. Ovarian cancer A2780 and A2780cisR cell lines were cultured in 50 cm2 petri dishes to produce at least a million cells per dish. Cells were treated with solutions of CS (at IC50) administered as a bolus and in two aliquots with 2 h time gap using both aged and fresh solutions (CS + CS (2/0 h) aged/aged and CS + CS (2/0 h) fresh/fresh). Untreated control cells were also included. The period of incubation with the drugs was 24 h. Following drug treatment, cell pellets were collected, rinsed with ice-cold PBS and centrifuged at 3500 rpm for 2 min at 4 °C. The pellets were lysed in a cell lysis solution containing 2 M thiourea, 8 M urea, 4 % CHAPS, 65 mM dithiothreitol (BIORAD, Australia). Isoelectric focusing (IEF) of the sample containing 200 μg of proteins was performed using 11 cm, pH 3–10 non-linear ReadyStrip™ IPG Strip in Protean i12 IEF cell unit (BIORAD, Australia) rehydrated in 8 M urea, 2 M thiourea, 4 % CHAPS, 60 mM dithiothreitol, 0.2 % carrier ampholyte, 0.0002 % bromophenol blue and deStreak (BIORAD, Australia). Two equilibration steps of the IPG were performed in SDS equilibration buffer containing SDS, 6 M urea, 50 % glycerol, 1.5 M Tris HCl (pH 8.8), and bromophenol blue with the first containing 0.5 g dithiothreitol and the second 0.5 g iodoacetamide. Protein concentration was determined using Bio-Rad Protein Assay (BIO-RAD, Australia). SDS-PAGE was performed using 4–20 % SDS Criterion™ TGX™ pre-cast gels in a Criterion Dodeca™ cell separation unit (BIO-RAD, Australia) at constant 200 V for 100 min in a Tris-glycine-HCl buffer system. The gels were stained with Bio-Safe Coomassie Stain (BIO-RAD, Australia) for 60 min. At least a duplicate of gels containing protein spots from the same sample were used for analysis. The gel images obtained by ChemiDoc™ MP Imaging system (BIO-RAD, Australia) were analysed for protein spots using Melanie version 7.0 software (GeneBio, Switzeland). A 2.0 fold change in the expression of a protein across the matched groups was used as the cut-off for differential expression. Analysis of variance (ANOVA), a statistical tool used to detect differences between experimental group means, was performed using a target significance level of 0.05. MALDI-TOF/TOF MS Protein spots were excised from preparative 2-D gels stained with Bio-Safe Coomassie Stain (BIO-RAD, Australia). The spots were destained with 120 μl of (50 % acetonitrile (ACN)/50 mM NH4HCO3) solution and heated at 37 °C for 30 min with mild shaking. The solution was then discarded. The gels were treated with 25 μl ACN and left to dry for 15 min. The solution was discarded then the spots were left to dry with the lid left open in the oven at 37 °C for 15 min followed by cooling at 4 °C. The spots were digested with 10 μl trypsin for 10 min on ice. The trypsin supernatants were placed in 96-well plate at 4 °C followed by 10 μl addition of 25 mM NH4HCO3 for overnight digestion at 37 °C. The resulting peptides were extracted with 0.1 % trifluoroacetic acid (TFA) then extracted and concentrated by C18 zip-tips (Millipore, μ-C18, P10 size) on Xcise (Proteome Systems). A 1 μl aliquot was manually spotted onto a MALDI AnchorChip plate with 1 μl of matrix (CHCA, 1 mg/mL in 90 % v/v ACN, 0.1 % TFA) and left to dry in air. Matrix assisted laser desorption ionisation mass spectrometry (MALDI-MS) was performed with 4800 plus MALDI TOF/TOF Analyser (AB Sciex). A neodymium-doped yttrium aluminum garnet (Nd:YAG) laser (355 nm) was used to irradiate the sample. Spectra were acquired in reflectron MS scan mode in the mass range of 700–4000 Da. The instrument was then switched to MS/MS (TOF-TOF) mode where the eight strongest peptides from the MS scan were isolated and fragmented by collision induced dissociation (CID), then re-accelerated to measure their masses and intensities. A near point calibration was applied and would give a typical mass accuracy of 50 ppm or less. The data on peptides masses were analysed using database search program Mascot (Matrix Science Ltd, London, UK). The peak lists were searched against Homo sapiens entries in the SwissProt database. The protein identification was undertaken at Australian Proteome Analysis Facility (APAF) the infrastructure provided by the Australian Government through the National Collaborative Research Infrastructure Strategy (NCRIS). Results Cytotoxicity Figure 2 shows the cell survival fractions versus concentration plots for CS, CB, and OX as applied to the human ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines. The IC50 values of CS, CB and OX are presented in Table 2. As expected, the values were higher in the resistant A2780cisR, A2780ZD0473R and SKOV-3 cell lines with OX having the largest value in SKOV-3. The parent A2780 cell line was most sensitive to the drugs, A2780ZD0473R was least sensitive to CB and SKOV-3 was least sensitive to OX.Fig. 2 Cell growth inhibition following increasing concentrations of platinum drugs. Cell survival fractions of ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines following treatment with increasing concentrations of a CS, b CB and c OX for 72 h were determined using MTT assay and spectrophotometric measurement. Error bars represent the standard deviation (where straight lines or curves containing the dot points are meaningless) Table 2 Summary of the IC50 values (μM) and resistance factors (RF) for CS, CB and OX as applied to the ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines. IC50 is the drug concentration required for 50 % cell kill and RF is the ratio of the IC50 value in the resistant A2780cisR and A2780ZD0473R cell lines over that in the responsive parent A2780 cell line A2780 A2780cisR RF A2780ZD0473R RF SKOV-3 CS 0.5 ± 0.03 8.2 ± 0.6 17.7 6.0 ± 0.5 13.0 10.2 ± 0.5 CB 14.0 ± 1.4 48.9 ± 3.9 3.5 64.6 ± 3.2 4.6 43.4 ± 3.9 OX 0.2 ± 0.01 0.4 ± 0.02 1.9 0.4 ± 0.03 2.1 43.6 ± 3.0 Administration in two aliquots Figure 1 gave the combination index (CI) values applying to administration of CS, CB and OX in two aliquots with a time gap of 2, 4, 8, or 24 h to the ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines where CI values of <1, =1 and >1 indicated respectively synergism, additivity and antagonism in combined action. Cells were treated with solutions of CS, CB and OX at three different concentrations based on their IC50 values. When CS was administered in two aliquots (fresh/aged), the killing of A2780 cells was most pronounced when the time gap was 8 h; all other time gaps were found to be additive to antagonistic. In A2780cisR cell line, much pronounced cell kill was observed for time gaps of 2, 4 and 8 h and antagonism was observed when it was increased to 24 h. In A2780ZD0473R cell line also, extremely pronounced cell kill was observed for time gaps of 2, 4 and 8 h and reduced cell kill was observed when this was increased to 24 h. In SKOV-3 cell line, synergistic cell kill was observed only for the 24 h time gap. When CB was administered in two aliquots (fresh/aged) in A2780 cell line synergistic kill was observed when the time gap was 4, 8 or 24 h. It was most pronounced when the time gap was 4 and 8 h but antagonistic when the time gap was 2 h. In A2780cisR cell line, pronounced cell kill was observed for the time gaps of 2, 4 and 8 h and reduced cell kill was observed when it was 24 h. In A2780ZD0473R cell line also, extremely pronounced cell kill was observed for time gaps of 2, 4 and 8 h and reduced cell kill was observed when this was increased to 24 h. In SKOV-3 cell line, antagonism was observed for all time gaps. When OX was administered in two aliquots (fresh/aged), cell kill in A2780 cells was pronounced when the time gap was 4, 8 and 24 h and reduced cell kill was observed when the time gap was 2 h. In A2780cisR and A2780ZD0473R, all the time gaps (especially 4 and 8 h) produced pronounced cell kill. When OX was administered in two aliquots (fresh/aged) to SKOV-3 cell line, cell kill was pronounced when the time gap was 2 h, additive when it was 4 h and reduced cell kill was observed when it was increased to 8 and 24 h. The results indicate the administration of CS, CB, and OX in two aliquots, with the first aliquot fresh and the second aliquot aged, generally caused enhanced cell kill especially in the resistant A2780cisR and A2780ZD0473R cell lines. In contrast, administration of CS, CB and OX in two aliquots with both being fresh, produced mostly reduced cell kill in A2780, A2780cisR and A2780ZD0473R cell lines. In SKOV-3 cell line, enhanced cell kill resulted when the time gap was 2, 4, and 8 h and reduced cell kill resulted when it was 24 h (Fig. 3).Fig. 3 Combination Index (CI) values following the administration of a CS, b CB and c OX in two aliquots with a time gap of 2, 4, 8, or 24 h as applied to the ovarian cancer A2780, A2780cisR, A2780ZD0473R and SKOV-3 cell lines using both fresh/fresh and fresh/aged combinations. CI values were calculated following 72 h treatments. CI values of <1, =1 and >1 indicate respectively synergism, additivity and antagonism in combined drug action Platinum accumulation To determine whether the enhanced cell kill was associated with an increase in platinum accumulation, the levels of platinum accumulation following the administration of CS in two aliquots with a time gap of 2 and 4 h in A2780 and A2780cisR cell lines were determined, using both fresh and aged solutions (Fig. 4). It was found that both fresh/fresh and fresh/aged combinations were associated with significantly higher platinum accumulations in A2780 cell line. The highest platinum accumulation resulted from fresh/aged combination with 4 h time gap. In the resistant A2780cisR cell line, only the aged solution resulted in increased platinum accumulation.Fig. 4 Cellular platinum accumulation in ovarian cancer A2780 and A2780cisR cell lines resulting from administration of CS in two aliquots using both fresh/fresh and fresh/aged combinations of CS with a time gap of 2 and 4 h. Cells were treated with the drugs for 24 h followed by collection, lysis and finally Pt was determined using AAS. Data was statistically analyzed using the paired Student’s t test: * p < 0.05 indicates significant difference from control. Error bars represent the standard deviation Platinum-DNA binding As a key step in the antitumour action of platinum drugs is their binding with DNA (that initiates downstream processes in the cell cycle leading to apoptosis), platinum − DNA binding levels in A2780 and A2780cisR cell lines were determined applying to administration of solutions of CS in two aliquots with a time gap of 2 and 4 h, using both fresh/fresh and fresh/aged combinations. Figure 2 gave platinum − DNA binding levels in ovarian cancer A2780 and A2780cisR cell lines as applied to the administration of CS in two aliquots with a time gap of 2 and 4 h. The platinum − DNA binding levels in A2780cisR cell line were generally found to be greater from administration in two aliquots than from the bolus for both 2 and 4 h time gaps, as applied to both fresh/fresh and fresh/aged combinations but more so for the latter. In the parent A2780 cell line, only the fresh/aged combination resulted in greater Pt − DNA binding level (for both 2 and 4 time gaps) than the bolus (Fig. 5).Fig. 5 Platinum − DNA binding in ovarian cancer A2780 and A2780cisR cell lines as applied to the administration of CS in two aliquots using both fresh/fresh and fresh/aged combinations of CS with a time gap of 2 and 4 h. Cells were treated with the drugs for 24 h followed by collection, DNA extraction and finally pt detection using AAS. Data was statistically analyzed using the paired Student’s t test: * p < 0.05 indicates significant difference from control. Error bars represent the standard deviation Cellular glutathione Since oxidative stress is like a double edged sword in cancer that can lead to both programmed cell death and cell survival, the effect of nature of administration of CS and CB on the cellular glutathione level was also investigated. Specifically, the effect of administration of the drug as a bolus and in two aliquots applying to both fresh/fresh and fresh/aged solutions of CB and CS was investigated. The levels of total glutathione (GSH plus GSSG) and the oxidized glutathione (GSSG) in A2780 and A2780cisR cell lines were determined (Fig. 6a, b). In line with reported result [13–15], total glutathione was found to be higher in the CS-resistant A2780cisR cell line than in the parent A2780 cell line before and after drug treatment. There was a significant decrease in total glutathione level following treatment of A2780 and A2780cisR cells with solutions of CS and CB. This was found to be true for both bolus administration and that in two aliquots but more so for the bolus. The results indicate that bolus administration of CS and CB produced a greater oxidative stress in the cells than that in two aliquots. As noted earlier, there was an increase in cell kill due to the administration of CS and CB in two aliquots using fresh/aged combinations.Fig. 6 Levels of a total glutathione (GSH plus GSSG) and b oxidized glutathione (GSSG) in relative luminescence units (RLU × 104) in A2780 and A2780cisR cells before and after their treatments with solutions of CS and CB administered as a bolus and in two aliquots with time gaps of 2 and 4 h using fresh/fresh and fresh/aged combinations. Cells were treated for 24 h and glutathione content was determined using GSH/GSSG-Glo Assay kit. Data was statistically analyzed using the paired Student’s t test: * p < 0.05 indicates significant difference from control. Error bars represent the standard deviation Platinum speciation With the idea that the greater cell kill resulting from administration of CS, CB and OX in two aliquots using fresh/aged combinations could be due to speciation of the drugs (resulting in the production of more cytotoxic species), mass spectral analysis of solutions of OX in cell culture media were performed. Figure 7 gave magnified mass spectrum showing major peaks of OX dissolved in cell culture medium. The major peaks observed in the mass spectra are given in Table 3.Fig. 7 Magnified mass spectrum showing major peaks of OX dissolved in cell culture medium. The 543.1 peak is the base peak due to the cell culture medium Table 3 Major peaks in the mass spectra of aged solution of OX in cell media m/z No. of possibilities Formula Relative intensity (×106) Suggested structure 398 1 Pt(C6H14N2)(C2O4) 16 420.0493 4 [Pt(C6H14N2)(C2O4) + Na]+ 4.6 436.0233 2 Pt(C6H14N2)2 – 3H 7.2 478.0080 14 [Pt(C6H14N2)(C8H11NO3)]+ 8 680.1303 67 [Pt(C6H14N2)(C10H17N3O6S)(CH3CH2SH) + 2H]+ 26 738.0889 157 [Pt(C6H14N2)(C10H17N3O6S)(C3H7NO2S)]+ 9 795.0480 277 [Pt2(C6H14N2)2(C2O4)2 + H]+ 21 940.2129 377 [Pt(C6H14N2)(C10H17N3O6S)2O]+ 19 Peak with m/z = 420.0493 is believed to be due to the molecular ion [Pt(DACH)(C2O4) + Na]+ where Na+ is from the culture media. Peak with m/z = 436.0233 is believed to be due to the molecular ion [Pt(DACH)(C2O4) + K]+ where K+ is from the culture media. Peak with m/z = 478.0080 is due to Pt(DACH)(Pyr) (where Pyr = pyridoxine (vitamin B6)), believed to be formed in the cell culture media. The peak with m/z = 680.1303 may be due to Pt(DACH)(SCH3)(GSH) formed in the culture media and that with m/z = 795.0480 is believed to be due to dimeric species consisting of two Pt(DACH) units catenated by two oxalate ligands. Finally, the peak with m/z = 940.2129 is believed to be due to Pt(DACH)(GSH)2. Proteomics As stated earlier, proteomic studies were carried out to identify the proteins differentially expressed in the resistant A2780cisR cell line as compared to the levels found in parent A2780 cell line. 2-D gels resolved over 390 proteins of which 72 were found to be differentially expressed in A2780cisR cell line as compared to the parent A2780 cell line (Fig. 8). Administration of CS in two aliquots with a time gap was found to restore the expression of at least 22 proteins to the levels found in the parent cell line, of which 12 were down-regulated and 10 up-regulated prior to drug treatment. A summary of the proteins, their possible functions, and associations with neoplasia are given in Table 4. The proteins belong to seven groups based on cellular functions namely: invasion and metastasis, cell cycle regulation and proliferation, metabolic and biosynthesis processes, stress-related proteins and chaperones, mRNA processing proteins, cellular organization/cytoskeleton, cellular communication and signal transduction (Fig. 9). The proteins are believed to be associated with platinum resistance in ovarian cancer. A more detailed description of their functions as applied to platinum resistance in ovarian cancer is given in the discussion.Fig. 8 2-DE pattern of whole-cell proteins in A2780 cell line. The 2-D gel was stained with coomasiee brilliant blue (11 cm, pH 3–10 non-linear, 4–20 % SDS-PAGE, 200 μg proteins). The protein spots differentially expressed in A2780cisR compared to the parent A2780 cell line identified in this study are marked with their spot number ID Table 4 In total, 22 proteins with differential expression (2-fold increase or decrease, ANOVA p < 0.05) between A2780cisR and the parent A2780 cell line due to treatment with CS in two aliquots with a time gap of 2 h using fresh/fresh and fresh/aged combinations were identified by MALDI-TOF/TOF Spot ID Full name MALDI location Expression in A2780cisR Post treatment expression Function Tumour association CS CS + CS fresh aged 3 COX5A Cytochrome c oxidase subunit 5A, mitochondrial Score: 63 Mass: 16752 pI: 6.30 Coverage: 11 % Mitochondria Down Regulated PR PR OR Ribosome biogenesis; mitochondrial respiratory chain Breast [77]; cervix [78]; colon [77]; gastric [31]; Kidney [79]; Lung [80–82]; Nasopharyngeal [32]; Oesophageal [77]; Ovarian [77]; Prostate [77]; Thyroid [83] 19 COF2 Cofilin‐2 Score: 127 Mass: 18725 pI: 7.66 Coverage: 39 % Cytoplasm, cytoskeleton Down Regulated R R OR Actin polymerization Pancreatic [84]; Prostate [85] 34 1433G 14‐3‐3 protein gamma Score: 315 Mass: 28285 pI: 4.80 Coverage: 44 % Cytoplasm Down Regulated PR PR OR Adapter protein Breast [86] 41 MARE1 Microtubule‐associated protein RP/EB family member 1 Score: 62 Mass: 29980 pI: 5.02 Coverage: 29 % Cytoplasm Down Regulated - R OR Microtubule cytoskeleton dynamics; cell migration. Colon [87]; lung [88]; gastric [89]; oesophageal [90]; 50 NPM Nucleophosmin Score: 68 Mass: 32555 pI: 4.64 Coverage: 26 % Nucleoplasm Down Regulated - R OR Chaperone; ribosome biogenesis; p53 and ARF regulation Bladder [91]; breast [92]; Colon [93]; gastric [94]; haematopoietic [40]; ovarian [95]; prostate [96] 51 ANXA1 Annexin A1 Score: 369 Mass: 38690 pI: 6.57 Coverage: 53 % Nucleus, cytoplasm Down Regulated R R OR Calcium/phospholipid-binding protein breast [97]; gastric [98]; HNSC [36]; lung [99]; oesophageal [100, 101]; prostate [100, 102]; 62 RSSA 40S ribosomal protein SA Score: 311 Mass: 32833 pI: 4.79 Coverage: 34 % Cell membrane, cytoplasm, nucleus Up Regulated R R - laminin receptor; fate determination; tissue morphogenesis Breast [103]; cervical [104]; colon [54, 55, 105]; melanoma [106] 65 ACTB Actin, cytoplasmic 1 Score: 767 Mass: 41710 pI: 5.29 Coverage: 53 % Cytoplasm Up Regulated PR R PR cell motility Colon [107]; gastric [108]; liver [109]; sarcoma [110] 70 CALU Calumenin Score: 141 Mass: 37084 pI: 4.47 Coverage: 14 % Sarcoplasmic, endoplasmic reticulum Up Regulated - OR PR Vitamin K-dependent carboxylation Colon [111]; gastric [112]; glioblastoma [113]; prostate [114]; 72 HNRPF Heterogeneous nuclear ribonucleoprotein F Score: 342 Mass: 45643 pI: 5.38 Coverage: 33 % Nucleus Down Regulated R OR R processing of pre-mRNAs; alternative splicing events Gastric [115]; thyroid [116] 76 ENOA Alpha‐enolase Score: 471 Mass: 47139 pI: 7.01 Coverage: 47 % Cytoplasm Down Regulated - PR R Glycolysis; hypoxia tolerance; tumour suppressor Breast [117]; lung [118]; nasopharyngeal [119]; 85 ATPA ATP synthase subunit alpha, mitochondrial Score: 391 Mass: 59714 pI: 9.16 Coverage: 29 % Mitochondria Up Regulated - OR OR Production of ATP from ADP Breast [120]; colon [121]; gastric [122]; leukemia [123]; thyroid [124]; 92 PDIA3 Protein disulfide-isomerase A3 Score: 525 Mass: 56747 pI: 5.98 Coverage: 54 % Endoplasmic reticulum Down Regulated PR PR R endopeptidase; electron carrier; Lung [125]; ovarian [48]; Prostate [46]; retinoblastoma [126] 100 HNRPK Heterogeneous nuclear ribonucleoprotein K Score: 123 Mass: 50944 pI: 5.39 Coverage: 26 % Cytoplasm, nucleus Up Regulated - OR - Pre-mRNA-binding proteins; p53 response to DNA damage Breast [127]; colon [128]; Leukemia [129, 130]; lung [131]; Oropharyngeal [132]; pancreatic [133]; 103 CH60 60 kDa heat shock protein, mitochondrial Score: 762 Mass: 61016 pI: 5.70 Coverage: 39 % Mitochondria Up Regulated PR OR PR macromolecular assembly; stress working chaperone Bronchus [134]; Colon [135]; leukemia [136]; prostate [137] 104 TCPQ T‐complex protein 1 subunit theta Score: 200 Mass: 59583 pI: 5.42 Coverage: 24 % Cytoplasm Down Regulated R OR R chaperone Colon [138]; liver [139]; 127 GRP75 Stress‐70 protein, mitochondrial Score: 839 Mass: 73635 pI: 5.87 Coverage: 33 % Mitochondria, nucleus Down Regulated R PR R Cell proliferation; cellular aging; chaperone Brain, Breast, colon, kidney, lung, ovarian [140] 139 GLU2B Glucosidase 2 subunit beta Score: 80 Mass: 59388 pI: 4.33 Coverage: 11 % Endoplasmic reticulum Up Regulated OR OR - Regulatory subunit of glucosidase II Colon [138, 141]; sarcoma [142]; 149 HSP74 Stress‐70 protein, mitochondrial Score: 187 Mass: 94271 pI: 5.11 Coverage: 10 % Cytoplasm Down Regulated - OR PR Stress response; cell proliferation, differentiation Breast [143] 170 HNRPC Heterogeneous nuclear ribonucleoproteins Score: 142 Mass: 33650 pI: 4.95 Coverage: 8 % Nucleus Up Regulated - R OR mRNA processing and translation Breast [144]; colon [145]; lung [146]; pancreatic [147]; 215 CALR Calreticulin Score: 246 Mass: 48112 pI: 4.29 Coverage: 17 % Endoplasmic, sarcoplasmic reticulum Up Regulated - R OR Calcium-binding chaperone Bladder [69]; breast [148]; colon [149]; gastric [150]; glioblastoma [69]; liver [151]; ovarian [69]; pancreas [152]; prostate [153] 246 VIME Vimentin Score: 539 Mass: 53619 pI: 5.06 Coverage: 81 % Cytoplasm Up Regulated P OR - Filaments attached to nucleus and endoplasmic reticulum Breast [154]; gastric [155]; lung [156]; pancreas [157] (−) indicates that the spot was either not visible on the gel or that the treatment had no effect on the expression of the protein R Restored, PR Partially Restored, OR Over Restored Fig. 9 Grouping of proteins and enzymes based on the cellular functions that were found to be differentially expressed in the resistant A2780cisR cell line as compared to the parent A2780 cell line and that have undergone further changes in expression after treatment with the administration of CS in two aliquots Discussion In this study, the efficacy of administering CS, CB and OX in two aliquots with a time gap was investigated with the idea that results may provide information about the processes that take place during platinum-based chemotherapy practiced in the clinic. In particular, the results might provide valuable insight into the molecular aspect of administering the same drug in cycles. As noted earlier [11], among the three platinum drugs, OX was most active against the ovarian cancer A2780, A2780cisR and A2780ZD0473R cell lines but had the lowest activity against SKOV-3 cell line. The higher activity of OX as compared to CS against A2780, A2780cisR and A2780ZD0473R cell lines, could be due to differences in both the leaving groups and the carrier ligands in OX and CS (oxalate in OX as against chloride in CS and trans-R,R-diaminocyclohexane abbreviated as DACH in OX as against ammonia in CS). This difference allows several conformational differences to exist in the intrastrand 1,2-(GpG) adducts formed by CS and OX [16, 17]. Whereas the CS crosslink preferentially undergoes hydrogen bonding with the 5’ side of adduct (that causes a greater structural distortion to the base pair at the 5’ end), OX does so with the 3’ end of the intrastrand crosslink. Also, in the case of OX there is a strong hydrogen bond between the NH2 group of the DACH ligand and the O6 oxygen atom of the 3’ guanine nucleobase. It has been suggested that the conformational differences between OX − DNA and CS − DNA adducts may be responsible for differences in their protein recognition and cellular processing [16]. The low activity of OX against SKOV-3 cell line may lie in the p53-null status of the cells [18, 19]. As applied to the administration of CS, CB and OX in two aliquots with a time gap, experiments were done using both fresh/fresh and fresh/aged solutions. The purpose behind using both fresh and aged solutions was to determine the effect of speciation of activity of platinum drugs. Hydrolysis of platinum drugs produces highly reactive mono- and di-aquo species that can deprotonate and polymerise to produce species with multiple metal centres. The presence of the species is believed to alter the nature of interaction with DNA (as well as that with other cellular platinophiles). For example, multi-nuclear platinum species may carry a greater net positive charge making them more easily attracted to the negatively charged DNA. Furthermore, unlike CS that binds predominantly to one strand of DNA, multinuclear species may bind more significantly to both the strands of DNA [20, 21]. When CS, CB and OX were administered in two aliquots with a time gap than as a bolus, it was found that there was a greater cell kill with fresh/aged combination. The extent of cell kill was dependent on the duration of the time gap with 8 h time gap producing most pronounced cell kill in the parent A2780 cell line. In A2780cisR and A2780ZD0473R resistant cell lines, time gaps of 2, 4, and 8 h all caused pronounced cell death. In contrast, in SKOV-3 cell line, 24 h time gap was additive in action whereas 2 to 8 h time gaps were antagonistic. Additiveness to antagonism observed with fresh/fresh combinations and synergism seen with fresh/aged combinations, suggest that the increased cell kill associated with the latter may be due to the formation of more cytotoxic species resulting from hydrolysis, deprotonation and polymerisation reactions and the rates of these processes are considered to be greater at the ambient temperature than under the frozen condition. In line with the idea, it was found that when solutions of CS were left standing at room temperature, the cell killing effect increased with time (data not shown) as was observed by others (Yachnin [21]). To gain a better understanding of the speciation of platinum drugs in solution in terms of hydrolysis and formation of more cytotoxic species, limited mass spectral measurements with solutions of OX were carried out. The results are discussed latter in the paper. Also whether the enhanced cell kill due to the administration of platinum drugs in two aliquots with a time gap was associated with a corresponding increase in cellular accumulation of platinum and consequently a greater level of platinum − DNA binding or it was due to speciation of platinum drugs in cell culture medium at the ambient temperature, the cellular platinum accumulation and platinum − DNA binding levels associated with the administration of CS in two aliquots with a time gap (2 and 4 h) were determined, as applied to both fresh/fresh and fresh/aged combinations. It may be noted that although platinum − DNA binding is a necessary step in the programmed cell death due to platinum drugs, it is not sufficient as the programmed cell death is actually carried out by downstream processes in the cell cycle in which many proteins are involved. In any case, the drugs must enter the cells before they can bind with DNA or be deactivated due to binding with cellular platinophiles [3, 22]. Figures 4 and 5 gave the cellular accumulation and platinum − DNA binding levels resulting from administration of CS in two aliquots with time gaps of 2 and 4 h in A2780 and A2780cisR cells, as applied to both fresh/fresh and fresh/aged combinations. It was found that fresh/aged combination produced greater platinum accumulation than the fresh/fresh combination suggesting that there were either greater influx or reduced efflux or both. Since the species formed are likely to be positively charged, they are more likely to be transported by organic cationic transporters rather than by passive diffusion and copper transporter 1 that is also known to transport CS into the cell (CTR1) [23]. As applied to platinum − DNA binding level, it was found that in the parent A2780 cell line administration in two aliquots with a time gap resulted in a lower platinum − DNA binding level with fresh/fresh combination than the bolus but a greater level from fresh/aged combination. In the CS-resistant A2780cisR cell line, both fresh/fresh and fresh/aged combinations resulted in significantly greater platinum − DNA binding levels than the bolus. The results indicate that (regardless of the ageing status of the solution) the administration of platinum drugs in two aliquots with a time gap may be better able to overcome mechanism of platinum resistance in the CS-resistant cell line, giving support to the merit of cycle regimens used in the clinic. Cellular glutathione levels were also determined as a measure of changes in redox status of the cell. Figure 6a and b respectively gave the levels of total glutathione (GSH and GSSG) and oxidized glutathione (GSSG) in A2780 and A2780cisR cell lines following their treatments with solutions of CS and CB administered as a bolus and in two aliquots with a 4 h time gap. Although the level of total glutathione in both A2780 and A2780cisR cell lines after drug treatments was lower than the levels before treatment (irrespective of whether the drug was administered as a bolus or in two aliquots), the decrease was greater for the bolus than the administrations in two aliquots. CB was found to be more effective than CS in lowering the level of GSH in the CS-resistant A2780cisR cell line although a kinetic study suggested that rates of reaction CS and OX with GSH were 5-fold greater than that of CB [24]. A smaller decrease in GSH level observed in both A2780 and A2780cisR cells after treatment with CS and CB given as a bolus than in two aliquots (with a time gap), rebuts the hypothesis that the first aliquot would decrease glutathione level enabling more of the second aliquot of platinum to bind to DNA. This means that the greater cell kill resulting from administration of platinum drugs in two aliquots with a time gap can be due to speciation of platinum drugs rather than changes in cellular glutathione level. Thus, other changes such as differential expression of proteins may be responsible for enhanced cell kill. The level of oxidised glutathione (GSSG) is also found to decrease following treatment of cells with CS and CB suggesting enhanced efflux of conjugated products through MRPs. Although in this study the mass spectral analysis of platinum speciation has been attempted primarily to find out about speciation of platinums on ageing i.e. when left standing at ambient temperature, it is important to note that platinum speciation especially in solution in the biological matrix gained increasing focus when it became clear that (1) the active drugs were the hydrolysed products rather than the intact molecules and that (2) the drugs are inactivated due to conjugation with proteins and peptides (such as glutathione) [25]. When solutions of OX made in buffered cell culture media were left standing at room temperature, a number of peaks were observed indicating that OX in solution underwent hydrolysis followed by further speciation. It should however be noted that many possible matches in terms of structure could be found for a given peak in the mass spectrum (having defined mass to charge ratio, mass error of less than 2 ppm and isotopic splitting pattern). For example the peak with m/z = 420 had four possible matches whereas that with m/z = 940 had 377 matches. Notwithstanding the multitudes of structural possibilities for the observed peaks, it can be seen that OX in solution had underwent hydrolysis and that the products of hydrolysis conjugated with components present in the cell culture media. A detailed description of the peaks in terms of suggested structures is given as follows. It should be noted when the drug enters the cell, other compounds are likely to be formed due to binding with the cellular constituents. For example, formation of inactive complexes such as Pt(DACH)(GSH), Pt(DACH)(Cys) and Pt(DACH)(Met) due to binding of the Pt(DACH) unit with cellular constituents GSH, Cys and methaionine (Met) have been reported [26]. The adducts may be inactive if they are inert towards binding with DNA. However, formation of such adducts within the cell may lead to oxidative stress and the resulting side effects due to depletion of cellular thiols [27]. A more definitive statement about the cytotoxicity of adducts can be made when they are isolated using a suitable method of separation e.g. reverse phase HPLC. It is possible that dimeric species Pt2(C6H14N2)2(C2O4)2 can be a potent DNA binder with enhanced cytotoxicity. It is appropriate to note that many more dimeric and trimeric species were formed in the aged solution of OX in mQ water i.e. in the absence of culture media than in its presence. One important difference between speciation of OX in the buffered cell culture media and in solution in mQ water was that a greater number of hydrolysed products e.g. hydroxy, aqua and oxalate bridged dimeric, trimeric, tetrameric and even pentameric species were formed in the aged solution of OX in milli-Q water. The paucity of such multimeric species in the cell culture media, may indicate that complexation with constituents of the media had served to hinder the polymerisation reactions. It is also possible that the background noise in the spectra might have served to mask their presence. Finally, the presence of dimeric species was also reported in the aged solution of cisplatin [28]. Proteomics Proteomics involving 2-D gel electrophoresis and mass spectrometry were employed to identify key proteins associated with drug resistance in A2780cisR ovarian cancer cell line. It was based on the idea that proteins associated with drug resistance would undergo marked changes in expression in the resistant A2780cisR cell line as compared to that in the parent A2780 cell line. It was also thought that the proteins in question might be restored back to normalcy after treatment with drugs in two aliquots that caused enhanced cell kill (although the involvement of multiple pathways both in apoptosis and cell survival may mean not all of the proteins need to be concurrently targeted to bring about the cell death). However, the difficulty in the extraction of hydrophobic proteins including transmembrane proteins means that the identification of proteins such as CTR1 and other platinum influx transporters may remain elusive. Furthermore, proteomic results provide only a static picture at a selected time point (or points) whereas the cells being dynamic in nature would undergo continual changes. Other drawbacks of proteomics include difficulty in extraction of proteins that are low in abundance and inability to provide a complete picture at the levels of organs and organisms [29]. Notwithstanding these limitations, in this study 390 proteins were identified of which 72 underwent significant changes in expression in the resistant A2780cisR cells compared to the levels found in the parent A2780 cells. Administration of CS in two aliquots with a time gap restored the expressions of at least 22 proteins to levels comparable to those found in the parent cell line, of which 12 were down-regulated and 10 up-regulated prior to drug treatment. A summary of the proteins, their possible functions, and associations with neoplasia were given in Table 4. One of the proteins undergoing differential expression is mitochondrial cytochrome c oxidase subunit 5A (COX5A) which is one of the 13 subunits that make up cytochrome c oxidase (COX), the terminal enzyme of the mitochondrial electron transport chain. It was down-regulated in untreated A2780cisR cell line as compared to the parent A2780 cell line, over-restored due to treatment with aged solution of CS administered in two aliquots with 2 h time gap but only partially-restored when the same treatment was given with fresh solution. Although the three largest subunits COX1, COX2 and COX3 form the catalytic core of cytochrome oxidase, COX5A coupled to COX4 is essential for the assembly of the entire unit [30] so that a decrease in the expression of COX5A may be an indicator of under performance of the entire enzyme. COX5A was found to be down-regulated in other cancers such as nasopharyngeal and gastric carcinomas [31, 32]. Thus, the up-regulation of COX5A in A2780cisR cells following treatment with solutions of CS administered in two aliquots with a time gap can be a factor responsible for the pronounced cell kill. Cofilin-2 (COF-2) was also down-regulated in A2780cisR cell line as compared to the parent A2780 cell line. Administration of fresh solution of CS in two aliquots with a 2 h time gap caused its full restoration whereas that with the aged solution led to over-restoration. The results indicate that COF-2, which is the major component of intranuclear and cytoplasmic actin rods that plays a critical role in the regulation of actin filament dynamics in eukaryotes [33], may be associated with platinum resistance in ovarian cancer and that the employed treatments have been able to overcome the mechanism of resistance applying to the protein. Another protein found to be down-regulated in A2780cisR as compared to that in A2780 cell line was 14-3-3γ. Present abundantly in the cytoplasm, 14-3-3 proteins participate in a wide variety of activities including DNA repair, apoptosis, the onset of cell differentiation and senescence, and the coordination of cell adhesion, motility, intracellular signalling and cell cycle control [34]. Administration of CS in two aliquots with 2 h time gap caused its partial restoration when fresh solutions were used but led to over-restoration with aged solutions. The results indicate that 14-3-3γ may be a key protein associated with platinum resistance in ovarian cancer. Another protein of interest is MARE1 that belongs to the microtubule-associated protein RP/EB family. It is a prototypic member of microtubule plus-end tracking proteins (+TIPs) that control microtubule dynamics and are associated with different cellular structures. MARE1 was down-regulated in A2780cisR ovarian cancer cell line as compared to A2780 cell line. The administration of CS in two aliquots with 2 h time gap led to its full restoration. Although down-regulation of MARE1 inhibits the formation of stable microtubule, the role of MARE1 in inducing chromosomal instability remains unclear [35]. The next identified protein was Annexin A1 that belongs to the annexin family of proteins that have been implicated in many molecular and cellular processes, including modulation of phospholipase A2 and kinase activities in signal transduction, maintenance of cytoskeleton and extracellular matrix integrity, tissue growth and differentiation, inflammation, and blood coagulation [36–38]. ANXA1 was found to be down-regulated in the CS-resistant A2780cisR cell line as compared to the parent A2780 cell line. The administration of CS in two aliquots with 2 h time gap fully restored ANXA1. The results indicate that ANXA1 may be playing an important role in drug resistance in ovarian cancer and that the administered combinations have been able to overcome associated mechanism of resistance. Another protein found to be down-regulated in the CS-resistant A2780cisR cell line, was nucleophosmin (NPM), also known as nucleolar phosphoprotein B23. The administration of CS in two aliquots caused its full restoration when fresh solutions were used but caused its over-restoration with the aged solution. The over-expression of the protein in A2780cisR cells due to treatment with the aged solution can be a reason why a greater cell kill was produced from the aged solution than fresh counterpart. A number of studies suggest that NPM is involved in cancer pathogenesis. In mice, inactivation of NPM in the germ line leads to a host of developmental defects that cause embryonic lethality at mid-gestation. Haploid-insufficiency of NPM leads to unrestricted centrosome duplication and genomic instability with mice developing myelodysplasia with an acceleration of oncogenesis [39]. Moreover, disruption of the NPM gene by translocation is frequently found in human hematopoietic malignancies [40]. The fact that NPM contributes to oncogenesis by activating the oncogenic potential of the fused protein partner, suggests that the down-regulation of NPM may also indicate the under-regulation of the tumour suppressors p53, Rb and ARF [41, 42]. Alpha-enolase (ENOA) was also down-regulated in A2780cisR cell line as compared to the parent A2780 cell line. The protein acts as a transcriptional repressor and possibly functions as a tumour suppressor. Partial restoration occurred when A2780cisR cells were treated with fresh solutions of CS administered in two aliquots with 2 h time gap whereas the same with aged solutions caused its full restoration. The results indicate that the treatments in two aliquots using aged solutions of CS were able to overcome mechanism of resistance associated with ENOA. Another protein found to be down-regulated in CS-resistant A2780cisR cell line as compared to its CS sensitive counterpart was PDIA3. Partial and over-restoration of PDIA3 occurred when cells were treated respectively with fresh and aged solutions of CS administered in two aliquots. Although the expression of PDIA was found to be up-regulated in some cancers such as breast, it was found to be down-regulated in other cancers such as gastric and prostate cancers [43–46]. Besides its role as a chaperone, it was suggested that PDIA3 might be functioning as a pro-apoptotic protein in prostate cancer; a decrease of caspase activity was related to due to down-regulation of PDIA3 in prostate cancer cell lines [46]. Down-regulation of PDIA3 might be playing a role in the late onset of prostate cancer progression. Down-regulation of PDIA3 also correlated with increased tumour invasion and advanced stage of gastric cancer. Hence PDIA3 has been proposed to be a negative prognostic marker [45]. In addition to its role in the ER stress pathway, PDIA3 has also gained attention due to its association with the major histocompatibility complex (MHC) class I pathway. In PDIA3 deficient mice, MHC I is impaired and negatively influences presentation of antigenic peptides helping tumours to escape from immune surveillance by cytotoxic T cells [47]. Although PDIA3 was found to be over-expressed in YDOV-139 ovarian cancer cell line [48], in this study, CS-resistant ovarian cancer cell line showed decreased levels of the protein compared to the parent cell line suggesting its involvement in resistance to cisplatin. As PDIA3 is reported to be up-regulated by hypoxia [49], low levels of PDIA3 found in A2780cisR cells indicate that the cells may not be under hypoxic stress. Calumenin (CALU) was also up-regulated in the resistant A2780cisR as compared to that in A2780 cell line. It is a ubiquitous calcium-binding protein localized in the endoplasmic reticulum and involved in such functions as protein folding and sorting [50]. Although the exact role of CALU is yet to be elucidated, the functions of the calcium-binding family are well understood. They have been associated with resistance to chemotherapeutic drugs [51, 52]. In this study, administration of fresh solution of CS in two aliquots with a 2 h time gap caused its over-restoration whereas that with aged solution of CS led to its partial-restoration. The results indicate that CALU may be associated with drug resistance and that the employed drug combinations have been able to overcome the associated mechanism. Another protein that was up-regulated in the A2780cisR as compared to the parent A2780 cell line was the 40S ribosomal protein SA (RSSA). The administration of CS in two aliquots using fresh solution fully restored its expression. The elevated expression of RSSA in the resistant A2780cisR cell line and its restoration back to normalcy due to synergistic treatments give support to the idea that the receptor is associated with metastasis and drug resistance. Laminin has been implicated in a wide variety of biological processes including cell adhesion, differentiation, migration, signalling, neurite outgrowth and metastasis. The protein also serves as a major adhesion substrate for invasive cancer cells. Indeed, there is a direct correlation between the ability of malignant cells to attach to laminin and their metastatic potential [53]. Thus, over-expression of RSSA was observed in many cancers indicating its potential role in tumour progression. Highly metastatic cancer cells are found to express at their surface significantly more laminin receptors than do their much less metastatic or benign counterparts [54]. Breast, cervical, colorectal and gastric carcinomas are found to express high levels of RSSA [54, 55]. Over-expression of this receptor is not restricted to epithelial tumours. For example, melanomas and lymphomas also display increased expression of the receptor [56, 57]. Next protein that was up-regulated in A2780cisR as compared to A2780 cell line was ACTB that is a member of the actin family. Actins are essential for a large range of cell functions including cell division, migration, junction formation, chromatin remodelling, transcriptional regulation, vesicle trafficking, and cell shape regulation [58]. Administration of CS in two aliquots using fresh solution caused its full restoration in A2780cisR cell line but partial restoration when aged solutions were used. The results indicate the employed treatments were able to overcome the associated mechanism of resistance. Three HNRNP proteins identified in this study were HNRPC, HNRPF and HNRPK. HNRPF was down-regulated whereas HNRPC and HNRPK were up-regulated in the A2780cisR cell line compared to the parent A2780 cell line. Although the proteins were differentially expressed in the resistant cell line, there does not appear to be a direct relationship between HNRNPs and cellular resistance to platinum drugs. ATPA and ATPB were up-regulated in the CS-resistant A2780cisR cell line as compared to the parent A2780 cell line. Administration of both fresh and aged solutions of CS in two aliquots with 2 h time gap caused over-restoration of ATPB. ATP synthase has been implicated in angiogenesis, cellular immunity, cholesterol uptake and cellular pH regulation [59]. A novel approach by Juan et al. targeted the deregulation and over-expression of ATP synthase in breast cancer using an ATP synthase inhibitor [60]. It may be noted that the spot corresponding to ATPA was positioned in a cluster of spots that were closely placed and were of poor quality after treatment so that some uncertainty remained about ATPA. Further experiments would be needed to ascertain changes in expression of ATPA. However, ATP synthase appears to be associated with drug resistance and that the employed treatments were able to completely inhibit the expression of ATPB. The protein coded as T-complex protein 1 subunit theta (TCPQ/CCTθ) was down-regulated in A2780cisR as compared to the level found in parent A2780 cell line. Although evidence is emerging about the diverse roles played by the complex, relatively little is known about the functional divergence of the individual subunits, and how this may relate their role in tumour development and progression. Treatment with solutions of CS given in two aliquots with 2 h time gap caused its over-restoration with fresh solution and partial restoration with aged solution. Another protein glucosidase 2 subunit beta coded as GLU2B was up-regulated in the A2780cisR cell line as compared to the A2780 cell line. Though the exact function of GLU2B in cell differentiation is yet to be determined, it may be influencing glycosylation process of newly synthesized proteins and may act as a regulator of distinct developmental processes [61]. It was reported by Otto Warburg about 70 years ago that tumour cells exhibited an altered metabolism, characterized by increased glucose uptake and elevated glycolysis [62]. Indeed, an increase in the rate of glycolysis is one of the metabolic alterations found in most cancer cells [63, 64]. GLU2B was over-restored in A2780cisR cell line due to the administration of fresh solution of CS given in two aliquots with a time gap but the same with aged solutions had no effect on the expression of GLU2B. The results suggest that GLU2 may be associated with drug resistance in ovarian cancer although the enhanced cell kill associated with aged solutions cannot be related to the protein. Calreticulin (CALR) that is a multifunctional protein that acts as a major calcium-binding protein in the lumen of the endoplasmic reticulum [65], was up-regulated in the CS-resistant A2780cisR cell line as compared to the parent A2780 cell. It is involved in a wide variety of cellular processes including modulation of calcium signals, storage and buffering of calcium, regulation of steroid-dependent gene expression via direct interaction with steroid receptors, cell adhesion via direct binding to integrin α, a chaperone in protein folding, autoimmune response and long-term neuromodulations [66]. Down-regulation of CALR by antisense was found to increase sensitivity of neuroblastoma × glioma NG-108–15 cells to cytotoxic calcium overload [67]. In contrast, up-regulation of CALR has been shown to protect HeLa cells from apoptosis [66]. Nakajo et al. reported that the expression of CALR markedly decreased before the apoptosis event in human leukemia HL-60 cells [68]. Furthermore, increased expression of CALR was a poor prognostic factor in diverse tumours including neuroblastoma, bladder cancer, and non-Hodgkin’s lymphoma [69]. Interestingly, CALR is thought to function as a pro-phagocytic signal highly expressed on the surface of several human cancers, but minimally expressed on normal cell counterparts line [69]. The protein was fully-restored due to treatment with fresh solutions of CS given in two aliquots with a time gap whereas treatment with the aged solutions caused its over-restoration. The results can be seen to indicate that CALR may be associated with platinum resistance in ovarian cancer. Next protein that was found to be up-regulated in CS-resistant A2780cisR cell line as compared to the parent A2780 cell line was vimentin (VIME). VIME is ubiquitously expressed in normal mesenchymal cells and known to maintain cellular integrity and provide resistance against stress. Increased expression of VIME has been reported in many epithelial cancers including melanoma, prostate, gastric, oesophageal, hepatocellular, pancreatic and breast carcinomas [70–73]. The over-expression of VIME in cancer is extensively reported to correlate with increased tumour growth, invasion, metastasis and poor prognosis [74]. In contrast, down-regulation of VIME inhibits carcinoma cell migration and adhesion [75]. VIME has gained much importance as a marker of epithelial-mesenchymal transition (EMT); a cellular reprogramming process in which the epithelial cells acquire a mesenchymal phenotype that renders the cells to significantly alter their shape and show increased motility [76]. It was over-restored due to treatment with fresh solutions of CS administered in two aliquots with a time gap whereas the same with the aged solutions had no significant effect on its expression. The results suggest that VIME may be associated with platinum resistance and that the increase in its expression may be associated with poor prognosis in platinum refractory ovarian cancer. A number of heat shock proteins (HSP) were also found to be differentially expressed in A2780cisR cell line as compared to A2780 cell line. The first was chaperonin 60 kDa (CH60) also known as HSP60. It was up-regulated in A2780cisR cell line as compared to A2780 cell line. The protein was restored back when the cells were treated with solutions of CS administered in two aliquots – over-restored in the case of fresh solution and partially restored in the case of aged solution. The results indicate that the over-expression of HSP60 may provide the cells with survival advantage by making them more tolerant against drugs and that the employed drug treatments were able to overcome the associated mechanism of resistance. Three HSPs belonging to the HSP-70 family (HSP-70s) namely HSP7C, HSP74 and GRP75 were also identified in the study. HSP74 was down regulated in A2780cisR cell line as compared to the parent A2780 cell line. It was over-restored due to treatment with fresh solution of CS administered in two aliquots and partially restored when the same was given using aged solution. HSP7C did not show any change in expression in A2780cisR cell line as compared to A2780 cell line. The results of the study show that whereas HSP74 may be associated with CS resistance in the tested ovarian cancer cell lines, HSP7C may not be so. GRP75 was also slightly down-regulated in A2780cisR as compared to A2780 cell line. It was partially and fully-restored when the cells were treated with fresh and aged solutions of CS administered in two aliquots respectively. The expression pattern of GRP75 suggests that the protein may not be a major player in CS resistance. In summary belongingness of the 22 identified proteins to various functional groupings such as invasion and metastasis, cell cycle regulation and proliferation, metabolic and biosynthesis processes, stress-related proteins and chaperones, mRNA processing, cellular organization/cytoskeleton, cellular communication and signal transduction highlights that platinum resistance is multifactorial in nature in which many proteins with diverse functions may be playing key roles; inevitably the loss of control of functions can endow tumour cells with the ability to escape programmed cell death and proliferate without control. The results also indicate that multiple strategies can be gainfully employed to overcome the resistance. Putting into context With the idea that the effect of administration of platinum drugs in two aliquots with a time gap may induce changes in multiple parameters, in addition to changes in the combined drug action, cellular accumulation of platinum, level of platinum − DNA binding, cellular glutathione level and changes in protein expression were also determined. Whereas the combined drug action was quantified for all aliquoted administrations, the cellular accumulation of platinum, level of platinum − DNA, cellular glutathione levels and proteomic studies were carried out for a subset of the experiments. It was thought that a careful consideration of the results for the subset might lead to more meaningful conclusions. Table 5 provides a summary of all the results for the subset.Table 5 Summary of results following treatment of A2780cisR cell line with CS administered in two aliquots with 2 h time gap Cytotoxic effect Cellular accumulation DNAbinding Proteomicsa Fresh Aged Fresh Aged Fresh Aged Fresh Aged CS + CS 2 h Antagonistic Synergistic Unchanged Increased Increased Greater increase COF2 MARE1 NPM ANXA1 RSSA ACTB CALU HNRPF ATPA HNRPK CH60 TCPQ GLU2B HSP74 HNRPC CALR VIME (D,R), (D,R), (D,R), (D,R), (U,R), (U,R), (U,OR), (D,R), (U,OR), (U,OR), (U,OR), (D,OR), (U,OR), (D,OR), (U,R), (U,R), (U,OR). COX5A COF2 1433γ MARE1 NPM ANXA1 HNRPF ENOA ATPA PDIA3 TCPQ GRP75 HNRPC CALR (D,OR), (D,OR), (D,OR), (D,OR), (D,OR), (D,OR), (D,OR), (D,R), (U,OR), (D,R), (D,R), (D,R), (U,R), (U,OR). aThe proteins listed have been fully/over restored in A2780cisR due to the selected combinations as compared to the parent A2780 cell line. Keys: U = up-regulated, D = down-regulated, OR = over-restored, R = fully-restored Although treatment with fresh solution was found to be less synergistic to antagonistic (whereas that with aged solution was more synergistic), both the cellular accumulation of platinum and the level of platinum − DNA binding were elevated due to treatment of A2780cisR cells with the fresh solution of CS in two aliquots. The failure of the increased cellular accumulation of platinum and more importantly that of the increased level of platinum − DNA binding to translate into enhanced cell death highlights the fact that although platinum − DNA binding can be a necessary step towards programmed cell death, it is not sufficient as apoptosis is brought about by downstream processes in the cell cycle in which many proteins may be playing key roles. The presence of crosstalk between pro-apoptotic and anti-apoptotic pathways can also be seen to complicate the situation. For the antagonistic administration of CS in two aliquots using fresh solution, the following proteins: ACTB, CALU, CH60, GLU2B, HSP74 and VIME in treated A2780cisR cells were fully or over restored to the levels found in the parent A2780 cell line. In contrast, for the administration of CS in two aliquots using aged solution which was synergistic in action, the following proteins: COX5A, 1433γ, ENOA, ATPA, PDIA3 and GRP75 were fully or over restored in treated A2780cisR cells as compared to the levels found in untreated A2780 cells. This is illustrated in the Venn diagram below (Fig. 9) where the proteins listed in red namely COX5A, 1433γ, ENOA, PDIA3 and GRP75 are considered to be characteristic of incubation of cells with aged solution CS in two aliquots with 2 h time gap (Fig. 10).Fig. 10 Venn diagram listing the proteins expressed in A2780cisR cells that were restored or over restored compared to the levels found in A2780 cells after treatment of cells with aged (synergistic) and fresh (antagonistic) solutions of CS in two aliquots with 2 h time gap COX5A, 1433G, ENOA, PDIA3 and GRP75 that are characteristic of synergistic administration of aged solution CS in two aliquots with 2 h time gap were initially down-regulated in the resistant cell line but up-regulated following the drug treatment indicating that these proteins may be playing a pro-apoptotic role so that their down-regulation served to dampen cell death. In contrast, the proteins: MARE1, ANXA1, RSSA, ACTB, CALU, CH60 and HSP7C that were restored due to treatment with both synergistic and antagonistic combinations, may not be so critically associated with synergistic drug action in spite of them being a hallmark of cancer cell biology and platinum drug resistance. The results of the present study can be seen to confirm that synergistic administration of drugs may provide a means of overcoming drug resistance due to the reestablishment of cellular control functions. Conclusion The results of the present study show that when platinum drugs are administered in two aliquots with a time gap, a greater cell kill is produced from treatment with aged solutions than that with fresh ones. A smaller decrease in cellular GSH level in both A2780 and A2780cisR cells after treatment with CS and CB given in two aliquots than as a bolus, indicates that the increased activity resulting from administration in two aliquots cannot be due to changes in GSH. The increased activity on ageing is believed to be related to speciation of the drug in solution. Proteomic studies have identified 72 proteins that were differentially expressed in A2780 and A2780cisR cell lines, 22 of them were restored back to the levels found in the parent cell line as a result of synergistic treatments, indicating their relevance in synergistic drug action. Among them COX5A, 1433G, ENOA, PDIA3 and GRP75 that were down-regulated in the resistant A2780cisR cell line as compared to that in parent A2780 cell line but up-regulated after synergistic treatments, are considered to play a more critical role in bringing about apoptotic cell death. In contrast, MARE1, ANXA1, RSSA, ACTB, CALU, CH60 and HSP7C which were restored due to treatment with both synergistic and antagonistic combinations, may not be so critically involved in apoptosis or escape from it, in spite of them being a hallmark of cancer cell biology and platinum drug resistance. Finally, it should be stated that a major limitation of the study is that it gives a static picture. However, cells are dynamic in which changes (especially for proteins) are constant. Abbreviations ACNAcetonitrile ACTBActin, cytoplasmic 1 ANXA1Annexin A1 ATPAATP synthase subunit alpha, mitochondrial CALRCalreticulin CALUCalumenin CBCarboplatin CBDCACyclobutanedicarboxylate CH6060 kDa heat shock protein, mitochondrial CICombination index CIDCollision induced dissociation COF2Cofilin-2 COX5ACytochrome c oxidase subunit 5A, mitochondrial CSCisplatin CTR1Copper transporter 1 CuCopper DACHDiaminocyclohexane ENOAAlpha-enolase FCSFoetal calf serum GLU2BGlucosidase 2 subunit beta GRP75Stress 70 protein, mitochondrial GSHGlutathione GSSGOxidised glutathione HCLHydrochloric Acid HisHistidine HNRPHeterogeneous nuclear ribonucleoproteins HSP74Stress-70 protein, mitochondrial IEFIsoelectric focusing IkBInhibitor of Kappa B MALDI-MSMatrix assisted laser desorption ionisation mass spectrometry MARE1Associated protein MTT3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide Nd:YAGNeodymium-doped yttrium aluminum garnet NF-kBNuclear factor Kappa B NPMNucleophosmin OXOxaliplatin MAPRE1Microtubule-associated protein RP/EB family member 1 MetMethionine MRPMultidrug resistance PDIA3Protein disulfide-isomerase A3 PBSPhosphate buffered saline PtPlatinum PyrPyridoxine (vitamin B6) ROSReactive oxygen species RSSA40S ribosomal protein SA SDSSodium dodecyl sulfate TCPQT complex protein 1 subunit theta TFATrifluoroacetic acid VIMEVimentin 1433G14-3-3 protein gamma Acknowledgements Part of the proteomic work was undertaken at APAF the Infrastructure provided by the Australian Government through the National Collaborative Research Infrastructure Strategy (NCRIS). Zaynab Al-Eisawi is grateful to the Discipline of Biomedical Science, Sydney Medical School, The University of Sydney for the Discipline of Biomedical Science PhD Research Scholarship. Funding This research is partly supported by Biomedical Science Research Initiative Grant and Cancer Research Donation Fund. Availability of data and materials The datasets supporting the conclusions of this article are included within the article. Authors’ contributions ZA developed the methodology, acquired and interpreted data, and drafted the manuscript. PB, CC and JQY aided in study design. FH designed the study, developed the methodology, interpreted data, edited the manuscript, and oversaw the study. NP contributed to mass spectral analysis and read the manuscript. MM contributed to proteomics and read the manuscript. Finally, all authors have read and approved final the 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. Wang D Lippard SJ Cellular processing of platinum anticancer drugs Nat Rev Drug Discov 2005 4 307 320 10.1038/nrd1691 15789122 2. Siddik ZH Cisplatin: mode of cytotoxic action and molecular basis of resistance Oncogene 2003 22 7265 7279 10.1038/sj.onc.1206933 14576837 3. Muggia FM Los G Platinum resistance: laboratory findings and clinical implications Stem Cells 1993 11 182 193 10.1002/stem.5530110304 8318904 4. Wu YJ Muldoon LL Neuwelt EA The Chemoprotective Agent N-Acetylcysteine Blocks Cisplatin-Induced Apoptosis through Caspase Signaling Pathway J Pharmacol Exp Ther 2005 312 424 431 10.1124/jpet.104.075119 15496615 5. Al-Eisawi Z Beale P Chan C Yu JQ Huq F Modulation of Cisplatin Cytotoxicity due to its Combination with Bortezomib and the Nature of its Administration Anticancer Res 2011 31 2757 2762 21868517 6. Dhara SC Rapid method for the synthesis of cis[Pt(NH3 )2 Cl2 ] Indian J Chem 1970 8 193 194 7. Mosmann T Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays J Immunol Methods 1983 65 55 63 10.1016/0022-1759(83)90303-4 6606682 8. Chou TC Drug Combination Studies and Their Synergy Quantification Using the Chou-Talalay Method Cancer Res 2010 70 440 446 10.1158/0008-5472.CAN-09-1947 20068163 9. Chou TC Theoretical Basis, Experimental Design, and Computerized Simulation of Synergism and Antagonism in Drug Combination Studies Pharmacol Rev 2006 58 621 681 10.1124/pr.58.3.10 16968952 10. Chou TC Talalay P Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors Adv Enzyme Regul 1984 22 27 55 10.1016/0065-2571(84)90007-4 6382953 11. Al-Eisawi Z Beale P Chan C Yu J Huq F Carboplatin and oxaliplatin in sequenced combination with bortezomib in ovarian tumour models J Ovarian Res 2013 6 78 10.1186/1757-2215-6-78 24209693 12. Bowtell DDL Rapid isolation of eukaryotic DNA Anal Biochem 1987 162 463 465 10.1016/0003-2697(87)90421-0 3037944 13. Balendiran GK Dabur R Fraser D The role of glutathione in cancer Cell Biochem Funct 2004 22 343 352 10.1002/cbf.1149 15386533 14. Godwin AK Meister A O’Dwyer PJ Huang CS Hamilton TC Anderson ME High resistance to cisplatin in human ovarian cancer cell lines is associated with marked increase of glutathione synthesis Proc Natl Acad Sci 1992 89 3070 3074 10.1073/pnas.89.7.3070 1348364 15. Rabik CA Dolan ME Molecular mechanisms of resistance and toxicity associated with platinating agents Cancer Treat Rev 2007 33 9 23 10.1016/j.ctrv.2006.09.006 17084534 16. Sharma S Gong P Temple B Bhattacharyya D Dokholyan NV Chaney SG Molecular Dynamic Simulations of Cisplatin- and Oxaliplatin-d(GG) Intrastand Cross-links Reveal Differences in their Conformational Dynamics J Mol Biol 2007 373 1123 1140 10.1016/j.jmb.2007.07.079 17900616 17. Wu Y Bhattacharyya D King CL Baskerville-Abraham I Huh SH Boysen G Solution Structures of a DNA Dodecamer Duplex with and without a Cisplatin 1,2-d(GG) Intrastrand Cross-Link: Comparison with the Same DNA Duplex Containing an Oxaliplatin 1,2-d(GG) Intrastrand Cross-Link Biochemistry (Mosc) 2007 46 6477 6487 10.1021/bi062291f 18. Vogelstein B Lane D Levine AJ Surfing the p53 network Nature 2000 408 307 310 10.1038/35042675 11099028 19. Yuan Z Cao K Lin C Li L Liu HY Zhao XY The p53 upregulated modulator of apoptosis (PUMA) chemosensitizes intrinsically resistant ovarian cancer cells to cisplatin by lowering the threshold set by Bcl-xL and Mcl-1 Mol Med 2011 17 1262 1274 10.2119/molmed.2011.00176 21863213 20. Kelland LR Farrell NP Platinum-Based Drugs in Cancer Therapy 2000 1 New Jersey Humana Press 21. Yachnin JR Wallin I Lewensohn R Sirzén F Ehrsson H The kinetics and cytotoxicity of cisplatin and its monohydrated complex Cancer Lett 1998 132 175 180 10.1016/S0304-3835(98)00175-X 10397471 22. Zisowsky J Koegel S Leyers S Devarakonda K Kassack MU Osmak M Relevance of drug uptake and efflux for cisplatin sensitivity of tumor cells Biochem Pharmacol 2007 73 298 307 10.1016/j.bcp.2006.10.003 17097621 23. Zhang S Lovejoy KS Shima JE Lagpacan LL Shu Y Lapuk A Organic Cation Transporters Are Determinants of Oxaliplatin Cytotoxicity Cancer Res 2006 66 8847 8857 10.1158/0008-5472.CAN-06-0769 16951202 24. Hagrman D Goodisman J Souid A-K Kinetic Study on the Reactions of Platinum Drugs with Glutathione J Pharmacol Exp Ther 2004 308 658 666 10.1124/jpet.103.059410 14610218 25. Michalke B Platinum speciation used for elucidating activation or inhibition of Pt-containing anti-cancer drugs J Trace Elem Med Biol 2010 24 69 77 10.1016/j.jtemb.2010.01.006 20413063 26. Di Francesco AM Ruggiero A Riccardi R Cellular and molecular aspects of drugs of the future: oxaliplatin Cell Mol Life Sci 2002 59 1914 1927 10.1007/PL00012514 12530522 27. Townsend DM Marto JA Deng M MacDonald TJ Hanigan MH High Pressure Liquid Chromatography and Mass Spectrometry Characterization of the Nephrotoxic Biotransformation Products of Cisplatin Drug Metab Dispos 2003 31 705 713 10.1124/dmd.31.6.705 12756201 28. Cui M Mester Z Electrospray ionization mass spectrometry coupled to liquid chromatography for detection of cisplatin and its hydrated complexes Rapid Commun Mass Spectrom 2003 17 1517 1527 10.1002/rcm.1030 12845575 29. Klein JB Thongboonkerd V Proteomics in Nephrology 2004 Switzerland Karger 30. Fornuskova D Stiburek L Wenchich L Vinsova K Hansikova H Zeman J Novel insights into the assembly and function of human nuclear-encoded cytochrome c oxidase subunits 4, 5a, 6a, 7a and 7b Biochem J 2010 428 363 374 10.1042/BJ20091714 20307258 31. Nishigaki R Osaki M Hiratsuka M Toda T Murakami K Jeang K-T Proteomic identification of differentially-expressed genes in human gastric carcinomas Proteomics 2005 5 3205 3213 10.1002/pmic.200401307 16003825 32. Liu J Zhan X Li M Li G Zhang P Xiao Z Mitochondrial proteomics of nasopharyngeal carcinoma metastasis BMC Med Genomics 2012 5 62 10.1186/1755-8794-5-62 23217164 33. Agrawal PB, Joshi M, Savic T, Chen Z, Beggs AH. Normal myofibrillar development followed by progressive sarcomeric disruption with actin accumulations in a mouse Cfl2 knockout demonstrates requirement of cofilin-2 for muscle maintenance. Hum Mol Genet. 2012;21. 34. Fu H Subramanian RR Masters SC 14-3-3 Proteins: Structure, Function, and Regulation Annu Rev Pharmacol Toxicol 2000 40 617 647 10.1146/annurev.pharmtox.40.1.617 10836149 35. Wen Y Eng CH Schmoranzer J Cabrera-Poch N Morris EJS Chen M EB1 and APC bind to mDia to stabilize microtubules downstream of Rho and promote cell migration Nat Cell Biol 2004 6 820 830 10.1038/ncb1160 15311282 36. Garcia Pedrero JM Fernandez MP Morgan RO Herrero Zapatero A Gonzalez MV Suarez Nieto C Annexin A1 Down-Regulation in Head and Neck Cancer Is Associated with Epithelial Differentiation Status Am J Pathol 2004 164 73 79 10.1016/S0002-9440(10)63098-2 14695321 37. Mulla A Christian HC Solito E Mendoza N Morris JF Buckingham JC Expression, subcellular localization and phosphorylation status of annexins 1 and 5 in human pituitary adenomas and a growth hormone-secreting carcinoma Clin Endocrinol (Oxf) 2004 60 107 119 10.1111/j.1365-2265.2004.01936.x 14678296 38. Fernandez MP Morgan RO Fernandez MR Carcedo MT The gene encoding human annexin V has a TATA-less promoter with a high G + C content Gene 1994 149 253 260 10.1016/0378-1119(94)90157-0 7958998 39. Grisendi S Bernardi R Rossi M Cheng K Khandker L Manova K Role of nucleophosmin in embryonic development and tumorigenesis Nature 2005 437 147 153 10.1038/nature03915 16007073 40. Lim MJ Wang XW Nucleophosmin and human cancer Cancer Detect Prev 2006 30 481 490 10.1016/j.cdp.2006.10.008 17113241 41. Bertwistle D Sugimoto M Sherr CJ Physical and Functional Interactions of the Arf Tumor Suppressor Protein with Nucleophosmin/B23 Mol Cell Biol 2004 24 985 996 10.1128/MCB.24.3.985-996.2004 14729947 42. Kurki S Peltonen K Latonen L Kiviharju TM Ojala PM Meek D Nucleolar protein NPM interacts with HDM2 and protects tumor suppressor protein p53 from HDM2-mediated degradation Cancer Cell 2004 5 465 475 10.1016/S1535-6108(04)00110-2 15144954 43. Updike MS Sawdy JC Wang L-S Liu S Huang Y-W Ye W Primary cultured human breast epithelial cells up-regulate protein disulfide isomerase in response to zeranol Anticancer Res 2007 27 407 410 17352260 44. Lee HH Lim CA Cheong YT Singh M Gam LH Comparison of Protein Expression Profiles of Different Stages of Lymph Nodes Metastasis in Breast Cancer Int J Biol Sci 2012 8 353 362 10.7150/ijbs.3157 22393307 45. Leys CM Nomura S Lafleur BJ Ferrone S Kaminishi M Montgomery E Expression and prognostic significance of prothymosin-α and ERp57 in human gastric cancer Surgery 2007 141 41 50 10.1016/j.surg.2006.05.009 17188166 46. Pressinotti N Klocker H Schafer G Luu V-D Ruschhaupt M Kuner R Differential expression of apoptotic genes PDIA3 and MAP3K5 distinguishes between low- and high-risk prostate cancer Mol Cancer 2009 8 130 10.1186/1476-4598-8-130 20035634 47. Garbi N Tanaka S Momburg F Hammerling GJ Impaired assembly of the major histocompatibility complex class I peptide-loading complex in mice deficient in the oxidoreductase ERp57 Nat Immunol 2006 7 93 102 10.1038/ni1288 16311600 48. Chay D Cho H Lim BJ Kang ES Oh YJ Choi SM ER-60 (PDIA3) is highly expressed in a newly established serous ovarian cancer cell line, YDOV-139 Int J Oncol 2010 37 399 412 20596667 49. Graven KK Molvar C Roncarati JS Klahn BD Lowrey S Farber HW Identification of protein disulfide isomerase as an endothelial hypoxic stress protein Am J Physiol Lung Cell Mol Physiol 2002 282 L996 L1003 10.1152/ajplung.00359.2001 11943664 50. Yabe D Taniwaki M Nakamura T Kanazawa N Tashiro K Honjo T Human Calumenin Gene (CALU): cDNA Isolation and Chromosomal Mapping to 7q32 Genomics 1998 49 331 333 10.1006/geno.1998.5245 9598325 51. Takazawa K Tsuchiya H Yoshimichi U Kanazawa Y Sadao I Tomita K Expression analysis for the identification of genes involved in acquired resistance to cisplatin in osteosarcoma cells Cancer Genomics Proteomics 2006 3 373 382 52. Hegde R Thimmaiah P Yerigeri MC Krishnegowda G Thimmaiah KN Houghton PJ Anti-calmodulin acridone derivatives modulate vinblastine resistance in multidrug resistant (MDR) cancer cells Eur J Med Chem 2004 39 161 177 10.1016/j.ejmech.2003.12.001 14987825 53. Ménard S Castronovo V Tagliabue E Sobel ME New insights into the metastasis-associated 67 kD laminin receptor J Cell Biochem 1997 67 155 165 10.1002/(SICI)1097-4644(19971101)67:2<155::AID-JCB1>3.0.CO;2-W 9328821 54. Cioce V Castronovo V Shmookler BM Garbisa S Grigioni WF Liotta LA Increased Expression of the Laminin Receptor in Human Colon Cancer J Natl Cancer Inst 1991 83 29 36 10.1093/jnci/83.1.29 1824600 55. Castronovo V Campo E Brûle FA Claysmith AP Cioce V Liu F-T Inverse Modulation of Steady-State Messenger RNA Levels of Two Non-Integrin Laminin-Binding Proteins in Human Colon Carcinoma J Natl Cancer Inst 1992 84 1161 1169 10.1093/jnci/84.15.1161 1386115 56. Carbone A Gloghini A Colombatti A Castronovo V Menard S Expression of the monomeric 67-kd laminin-binding protein in human lymphomas as defined by MLuC5 monoclonal antibody and paraffin section immunohistochemistry Hum Pathol 1995 26 541 546 10.1016/0046-8177(95)90251-1 7750938 57. Vacca A Ribatti D Roncali L Lospalluti M Serio G Carrel S Melanocyte tumor progression is associated with changes in angiogenesis and expression of the 67-kilodalton laminin receptor Cancer 1993 72 455 461 10.1002/1097-0142(19930715)72:2<455::AID-CNCR2820720222>3.0.CO;2-8 7686442 58. Perrin BJ Ervasti JM The actin gene family: function follows isoform Cytoskeleton 2010 67 630 634 10.1002/cm.20475 20737541 59. Chi SL Pizzo SV Cell surface F1Fo ATP synthase: A new paradigm? Ann Med 2006 38 429 438 10.1080/07853890600928698 17008306 60. Juan HF Chang HY Huang TC Hsu CH Kuo WH Chang KJ Method of treating cancer using ATP synthase inhibitors. US patent no. 514456 National Taiwan University 2010 61. Hoffrogge R Beyer S Hübner R Mikkat S Mix E Scharf C 2-DE profiling of GDNF overexpression-related proteome changes in differentiating ST14A rat progenitor cells Proteomics 2007 7 33 46 10.1002/pmic.200600614 17146836 62. Warburg O The Metabolism of Tumours: Investigations from the Kaiser Wilhelm Institute for Biology Br J Surg 1930 19 168 63. Lee HC Yin PH Lin JC Wu CC Chen CY Wu CW Mitochondrial Genome Instability and mtDNA Depletion in Human Cancers Ann N Y Acad Sci 2005 1042 109 122 10.1196/annals.1338.011 15965052 64. Young C Anderson S Sugar and fat - that’s where it’s at: metabolic changes in tumors Breast Cancer Res 2008 10 202 10.1186/bcr1852 18304378 65. Michalak M Milner RE Burns K Opas M Calreticulin Biochem J 1992 285 681 692 10.1042/bj2850681 1497605 66. Zhu N Pewitt EB Cai X Cohn EB Lang S Chen R Calreticulin: An Intracellular Ca++ − Binding Protein Abundantly Expressed and Regulated by Androgen in Prostatic Epithelial Cells Endocrinology 1998 139 4337 4344 9751517 67. Liu N Fine RE Simons E Johnson RJ Decreasing calreticulin expression lowers the Ca2+ response to bradykinin and increases sensitivity to ionomycin in NG-108-15 cells J Biol Chem 1994 269 28635 28639 7961812 68. Nakajo S Okamoto M Masuda Y Sakai I Ohsawa S Nakaya K Geranylgeraniol Causes a Decrease in Levels of Calreticulin and Tyrosine Phosphorylation of a 36-kDa Protein Prior to the Appearance of Apoptotic Features in HL-60 Cells Biochem Biophys Res Commun 1996 226 741 745 10.1006/bbrc.1996.1423 8831684 69. Chao MP Jaiswal S Weissman-Tsukamoto R Alizadeh AA Gentles AJ Volkmer J Calreticulin Is the Dominant Pro-Phagocytic Signal on Multiple Human Cancers and Is Counterbalanced by CD47 Sci Transl Med 2010 2 63ra94 10.1126/scitranslmed.3001375 21178137 70. Wu M Bai X Xu G Wei J Zhu T Zhang Y Proteome analysis of human androgen-independent prostate cancer cell lines: Variable metastatic potentials correlated with vimentin expression Proteomics 2007 7 1973 1983 10.1002/pmic.200600643 17566973 71. Takemura K Hirayama R Hirokawa K Inagaki M Tsujimura K Esaki Y Expression of Vimentin in Gastric Cancer: A Possible Indicator for Prognosis Pathobiology 1994 62 149 154 10.1159/000163895 7945921 72. Jin H Morohashi S Sato F Kudo Y Akasaka H Tsutsumi S Vimentin expression of esophageal squamous cell carcinoma and its aggressive potential for lymph node metastasis Biomed Res 2010 31 105 112 10.2220/biomedres.31.105 20460738 73. Gilles C Polette M Mestdagt M Nawrocki-Raby B Ruggeri P Birembaut P Transactivation of Vimentin by β-Catenin in Human Breast Cancer Cells Cancer Res 2003 63 2658 2664 12750294 74. Satelli A Li S Vimentin in cancer and its potential as a molecular target for cancer therapy Cell Mol Life Sci 2011 68 3033 3046 10.1007/s00018-011-0735-1 21637948 75. Mcinroy L Määttä A Down-regulation of vimentin expression inhibits carcinoma cell migration and adhesion Biochem Biophys Res Commun 2007 360 109 114 10.1016/j.bbrc.2007.06.036 17585878 76. Thiery JP Epithelial-mesenchymal transitions in tumour progression Nat Rev Cancer 2002 2 442 454 10.1038/nrc822 12189386 77. Herrmann PC Gillespie JW Charboneau L Bichsel VE Paweletz CP Calvert VS Mitochondrial proteome: Altered cytochrome c oxidase subunit levels in prostate cancer Proteomics 2003 3 1801 1810 10.1002/pmic.200300461 12973739 78. De LC-HE Perez-Plasencia C Perez-Cardenas E Gonzalez-Fierro A Trejo-Becerril C Chavez-Blanco A Transcriptional changes induced by epigenetic therapy with hydralazine and magnesium valproate in cervical carcinoma Oncol Rep 2011 25 399 407 21152880 79. Yusenko MV Ruppert T Kovacs G Analysis of differentially expressed mitochondrial proteins in chromophobe renal cell carcinomas and renal oncocytomas by 2-D gel electrophoresis Int J Biol Sci 2010 6 213 224 10.7150/ijbs.6.213 20440404 80. inventor Identification of cytochrome oxidase COX4-2 as an early lung cancer diagnostic biomarker. Patent no. US patent classification 2006 81. Chen WL Kuo KT Chou TY Chen CL Wang CH Wei YH The role of cytochrome c oxidase subunit Va in non-small cell lung carcinoma cells: association with migration, invasion and prediction of distant metastasis BMC Cancer 2012 12 273 10.1186/1471-2407-12-273 22748147 82. Telang S Nelson K Siow D Yalcin A Thornburg J Imbert-Fernandez Y Cytochrome c oxidase is activated by the oncoprotein Ras and is required for A549 lung adenocarcinoma growth Mol Cancer 2012 11 60 10.1186/1476-4598-11-60 22917272 83. Baris O Savagner F Nasser V Loriod B Granjeaud S Guyetant S Transcriptional Profiling Reveals Coordinated Up-Regulation of Oxidative Metabolism Genes in Thyroid Oncocytic Tumors J Clin Endocrinol Metab 2004 89 994 1005 10.1210/jc.2003-031238 14764826 84. Wang Y Kuramitsu Y Ueno T Suzuki N Yoshino S Iizuka N Differential expression of up-regulated cofilin-1 and down-regulated cofilin-2 characteristic of pancreatic cancer tissues Oncol Rep 2011 26 1595 1599 21894436 85. Zhu B Fukada K Zhu H Kyprianou N Prohibitin and cofilin are intracellular effectors of transforming growth factor β signaling in human prostate cancer cells Cancer Res 2006 66 8640 8647 10.1158/0008-5472.CAN-06-1443 16951178 86. Song Y Yang Z Ke Z Yao Y Hu X Sun Y Expression of 14-3-3γ in patients with breast cancer: Correlation with clinicopathological features and prognosis Cancer Epidemiol 2012 36 533 536 10.1016/j.canep.2012.05.003 22658894 87. Ladd JJ Busald T Johnson MM Zhang Q Pitteri SJ Wang H Increased Plasma Levels of the APC-Interacting Protein MAPRE1, LRG1, and IGFBP2 Preceding a Diagnosis of Colorectal Cancer in Women Cancer Prev Res (Phila Pa) 2012 5 655 664 10.1158/1940-6207.CAPR-11-0412 88. Fujii K Kondo T Yokoo H Yamada T Iwatsuki K Hirohashi S Proteomic study of human hepatocellular carcinoma using two-dimensional difference gel electrophoresis with saturation cysteine dye Proteomics 2005 5 1411 1422 10.1002/pmic.200401004 15751005 89. Kim K Lee H-C Park J-L Kim M Kim S-Y Noh S-M Epigenetic regulation of microRNA-10b and targeting of oncogenic MAPRE1 in gastric cancer Epigenetics 2011 6 740 751 10.4161/epi.6.6.15874 21562367 90. Wang Y Zhou X Zhu H Liu S Zhou C Zhang G Overexpression of EB1 in human esophageal squamous cell carcinoma (ESCC) may promote cellular growth by activating β-catenin/TCF pathway Oncogene 2005 24 6637 6645 10.1038/sj.onc.1208819 16007168 91. Tsui KH Cheng AJ Chang PL Pan TL Yung BYM Association of nucleophosmin/B23 mRNA expression with clinical outcome in patients with bladder carcinoma Urology 2004 64 839 844 10.1016/j.urology.2004.05.020 15491744 92. Skaar TC Prasad SC Sharareh S Lippman ME Brünner N Clarke R Two-dimensional gel electrophoresis analyses identify nucleophosmin as an estrogen regulated protein associated with acquired estrogen-independence in human breast cancer cells J Steroid Biochem Mol Biol 1998 67 391 402 10.1016/S0960-0760(98)00142-3 10030688 93. Nozawa Y Van Belzen N Van Der Made ACJ Dinjens WNM Bosman FT Expression of Nucleophosmin/B23 in Normal and Neoplastic Colorectal Mucosa J Pathol 1996 178 48 52 10.1002/(SICI)1096-9896(199601)178:1<48::AID-PATH432>3.0.CO;2-Y 8778315 94. Tanaka M Sasaki H Kino I Sugimura T Terada M Genes Preferentially Expressed in Embryo Stomach are Predominantly Expressed in Gastric Cancer Cancer Res 1992 52 3372 3377 1596895 95. Shields LB Gerçel-Taylor C Yashar CM Wan TC Katsanis WA Spinnato JA Induction of immune responses to ovarian tumor antigens by multiparity J Soc Gynecol Investig 1997 4 298 304 10.1016/S1071-5576(97)00068-3 9408885 96. Subong ENP Shue MJ Epstein JI Briggman JV Chan PK Partin AW Monoclonal antibody to prostate cancer nuclear matrix protein (PRO:4–216) recognizes nucleophosmin/B23 Prostate 1999 39 298 304 10.1002/(SICI)1097-0045(19990601)39:4<298::AID-PROS11>3.0.CO;2-M 10344220 97. Ahn SH Sawada H Ro JY Nicolson GL Differential expression of annexin I in human mammary ductal epithelial cells in normal and benign and malignant breast tissues Clin Exp Metastasis 1997 15 151 156 10.1023/A:1018452810915 9062391 98. Yu G Wang J Chen Y Wang X Pan J Li Q Tissue microarray analysis reveals strong clinical evidence for a close association between loss of annexin A1 expression and nodal metastasis in gastric cancer Clin Exp Metastasis 2008 25 695 702 10.1007/s10585-008-9178-y 18535914 99. Masaki T Tokuda M Ohnishi M Watanabe S Fujimura T Miyamoto K Enhanced expression of the protein kinase substrate annexin in human hepatocellular carcinoma Hepatology 1996 24 72 81 8707286 100. Paweletz CP Ornstein DK Roth MJ Bichsel VE Gillespie JW Calvert VS Loss of annexin 1 correlates with early onset of tumorigenesis in esophageal and prostate carcinoma Cancer Res 2000 60 6293 6297 11103786 101. Xia SH Hu LP Hu H Ying WT Xu X Cai Y Three isoforms of annexin I are preferentially expressed in normal esophageal epithelia but down-regulated in esophageal squamous cell carcinomas Oncogene 2002 21 6641 6648 10.1038/sj.onc.1205818 12242662 102. Kang JS Calvo BF Maygarden SJ Caskey LS Mohler JL Ornstein DK Dysregulation of annexin I protein expression in high-grade prostatic intraepithelial neoplasia and prostate cancer Clin Cancer Res 2002 8 117 123 11801547 103. Viacava P Naccarato AG Collecchi P Ménard S Castronovo V Bevilacqua G The Spectrum 67-kD Laminin Receptor Expression in Breast Carcinoma Progression J Pathol 1997 182 36 44 10.1002/(SICI)1096-9896(199705)182:1<36::AID-PATH802>3.0.CO;2-W 9227340 104. Al-Saleh W Delvenne P Van Den Brule FA Menard S Boniver J Castronovo V Expression of the 67 kD laminin receptor in human cervical preneoplastic and neoplastic squamous epithelial lesions: an immunohistochemical study J Pathol 1997 181 287 293 10.1002/(SICI)1096-9896(199703)181:3<287::AID-PATH762>3.0.CO;2-W 9155714 105. Sanjuán X Fernández PL Miquel R Muñoz J Castronovo V Ménard S Overexpression of the 67-kD Laminin Receptor Correlates with Tumour Progression in Human Colorectal Carcinoma J Pathol 1996 179 376 380 10.1002/(SICI)1096-9896(199608)179:4<376::AID-PATH591>3.0.CO;2-V 8869283 106. Kumazoe M Sugihara K Tsukamoto S Huang Y Tsurudome Y Suzuki T 67-kDa laminin receptor increases cGMP to induce cancer-selective apoptosis J Clin Invest 2013 123 787 799 23348740 107. Nowak D Skwarek-Maruszewska A Zemanek-Zboch M Malicka-Blaszkiewicz M Beta-actin in human colon adenocarcinoma cell lines with different metastatic potential Acta Biochim Pol 2005 52 461 468 15940343 108. Xu J Zhang Z Chen J Liu F Bai L Overexpression of β-actin is closely associated with metastasis of gastric cancer Hepatogastroenterology 2013 60 620 623 23635433 109. Popow A Nowak D Malicka-Blaszkiewicz M Actin cytoskeleton and beta-actin expression in correlation with higher invasiveness of selected hepatoma Morris 5123 cells J Physiol Pharmacol 2006 57 Suppl 7 111 123 17228099 110. Le PU Nguyen TN Drolet-Savoie P Leclerc N Nabi IR Increased β-Actin Expression in an Invasive Moloney Sarcoma Virus-transformed MDCK Cell Variant Concentrates to the Tips of Multiple Pseudopodia Cancer Res 1998 58 1631 1635 9563473 111. Volmer MW Stühler K Zapatka M Schöneck A Klein-Scory S Schmiegel W Differential proteome analysis of conditioned media to detect Smad4 regulated secreted biomarkers in colon cancer Proteomics 2005 5 2587 2601 10.1002/pmic.200401188 15912508 112. Kim B Bang S Lee S Kim S Jung Y Lee C Expression Profiling and Subtype-Specific Expression of Stomach Cancer Cancer Res 2003 63 8248 8255 14678982 113. Zhang R Tremblay T-L McDermid A Thibault P Stanimirovic D Identification of differentially expressed proteins in human glioblastoma cell lines and tumors Glia 2003 42 194 208 10.1002/glia.10222 12655603 114. Deprimo SE Diehn M Nelson JB Reiter RE Matese J Fero M Transcriptional programs activated by exposure of human prostate cancer cells to androgen Genome Biol 2002 3 1 10.1186/gb-2002-3-7-research0032 115. Yang Y-X Sun X-F Cheng A-L Zhang G-Y Yi H Sun Y Increased expression of HSP27 linked to vincristine resistance in human gastric cancer cell line J Cancer Res Clin Oncol 2009 135 181 189 10.1007/s00432-008-0460-9 18758817 116. Pines A Bivi N Vascotto C Romanello M D’Ambrosio C Scaloni A Nucleotide receptors stimulation by extracellular ATP controls Hsp90 expression through APE1/Ref-1 in thyroid cancer cells: A novel tumorigenic pathway J Cell Physiol 2006 209 44 55 10.1002/jcp.20704 16741950 117. Tu SH Chang CC Chen CS Tam KW Wang YJ Lee CH Increased expression of enolase α in human breast cancer confers tamoxifen resistance in human breast cancer cells Breast Cancer Res Treat 2010 121 539 553 10.1007/s10549-009-0492-0 19655245 118. Chang GC Liu KJ Hsieh CL Hu TS Charoenfuprasert S Liu HK Identification of α-Enolase as an Autoantigen in Lung Cancer: Its Overexpression Is Associated with Clinical Outcomes Clin Cancer Res 2006 12 5746 5754 10.1158/1078-0432.CCR-06-0324 17020980 119. Cheng C Long X Li X Xie M Guo M The expressions of alpha-enolase in the nasopharyngeal cancer tissue J Clin otorhinolaryngol Head Neck Surg 2011 25 554 556 120. Duffy MJ Predictive Markers in Breast and Other Cancers: A Review Clin Chem 2005 51 494 503 10.1373/clinchem.2004.046227 15637130 121. Seth R Keeley J Abu-Ali G Crook S Jackson D Ilyas M The putative tumour modifier gene ATP5A1 is not mutated in human colorectal cancer cell lines but expression levels correlate with TP53 mutations and chromosomal instability J Clin Pathol 2009 62 598 603 10.1136/jcp.2009.064436 19261598 122. Kim NS Hahn Y Oh JH Lee JY Oh KJ Kim JM Gene cataloging and expression profiling in human gastric cancer cells by expressed sequence tags Genomics 2004 83 1024 1045 10.1016/j.ygeno.2003.12.002 15177556 123. Hofmann WK de Vos S Elashoff D Gschaidmeier H Hoelzer D Koeffler HP Relation between resistance of Philadelphia-chromosome-positive acute lymphoblastic leukaemia to the tyrosine kinase inhibitor STI571 and gene-expression profiles: a gene-expression study Lancet 2002 359 481 486 10.1016/S0140-6736(02)07678-X 11853794 124. Onda M Emi M Yoshida A Miyamoto S Akaishi J Asaka S Comprehensive gene expression profiling of anaplastic thyroid cancers with cDNA microarray of 25 344 genes Endocr Relat Cancer 2004 11 843 854 10.1677/erc.1.00818 15613457 125. Teramoto R Minagawa H Honda M Miyazaki K Tabuse Y Kamijo K Protein expression profile characteristic to hepatocellular carcinoma revealed by 2D-DIGE with supervised learning Biochim Biophys Acta Protein Proteomics 2008 1784 764 772 10.1016/j.bbapap.2008.02.011 126. Chambery A Farina A Di Maro A Rossi M Abbondanza C Moncharmont B Proteomic Analysis of MCF-7 Cell Lines Expressing the Zinc-Finger or the Proline-Rich Domain of Retinoblastoma-Interacting-Zinc-Finger Protein J Proteome Res 2006 5 1176 1185 10.1021/pr0504743 16674107 127. Nagai MA da Rós N Neto MM de Faria Junior SR Brentani MM Hirata R Gene expression profiles in breast tumors regarding the presence or absence of estrogen and progesterone receptors Int J Cancer 2004 111 892 899 10.1002/ijc.20329 15300801 128. Kwong KY Bloom GC Yang I Boulware D Coppola D Haseman J Synchronous global assessment of gene and protein expression in colorectal cancer progression Genomics 2005 86 142 158 10.1016/j.ygeno.2005.03.012 15951154 129. Notari M Neviani P Santhanam R Blaser BW Chang J-S Galietta A A MAPK/HNRPK pathway controls BCR/ABL oncogenic potential by regulating MYC mRNA translation Blood 2006 107 2507 2516 10.1182/blood-2005-09-3732 16293596 130. Perrotti D Neviani P From mRNA Metabolism to Cancer Therapy: Chronic Myelogenous Leukemia Shows the Way Clin Cancer Res 2007 13 1638 1642 10.1158/1078-0432.CCR-06-2320 17363515 131. Keenan J Murphy L Henry M Meleady P Clynes M Proteomic analysis of multidrug-resistance mechanisms in adriamycin-resistant variants of DLKP, a squamous lung cancer cell line Proteomics 2009 9 1556 1566 10.1002/pmic.200800633 19242932 132. Lohavanichbutr P Houck J Fan W Genomewide gene expression profiles of hpv-positive and hpv-negative oropharyngeal cancer: Potential implications for treatment choices Arch Otolaryngol Head Neck Surg 2009 135 180 188 10.1001/archoto.2008.540 19221247 133. Alldinger I Dittert D Peiper M Fusco A Chiappetta G Staub E Gene expression analysis of pancreatic cell lines reveals genes overexpressed in pancreatic cancer Pancreatology 2005 5 370 379 10.1159/000086537 15983444 134. Cappello F Di Stefano A David S Rappa F Anzalone R La Rocca G Hsp60 and Hsp10 down-regulation predicts bronchial epithelial carcinogenesis in smokers with chronic obstructive pulmonary disease Cancer 2006 107 2417 2424 10.1002/cncr.22265 17048249 135. Cappello F David S Rappa F Bucchieri F Marasa L Bartolotta T The expression of HSP60 and HSP10 in large bowel carcinomas with lymph node metastase BMC Cancer 2005 5 139 10.1186/1471-2407-5-139 16253146 136. Thomas X Campos L Mounier C Cornillon J Flandrin P Le Q-H Expression of heat-shock proteins is associated with major adverse prognostic factors in acute myeloid leukemia Leuk Res 2005 29 1049 1058 10.1016/j.leukres.2005.02.010 16038731 137. Tang D Abdul Khaleque MD Jones EL Theriault JR Li C Wong WH Expression of heat shock proteins and heat shock protein messenger ribonucleic acid in human prostate carcinoma in vitro and in tumors in vivo Cell Stress Chaperones 2005 10 46 58 10.1379/CSC-44R.1 15832947 138. Bauer KM Lambert PA Hummon AB Comparative label-free LC-MS/MS analysis of colorectal adenocarcinoma and metastatic cells treated with 5-fluorouracil Proteomics 2012 12 1928 1937 10.1002/pmic.201200041 22623418 139. Farooq M Hozzein WN Elsayed EA Taha NA Wadaan MAM Identification of Histone Deacetylase 1 Protein Complexes in Liver Cancer Cells Asian Pac J Cancer Prev 2013 14 915 921 10.7314/APJCP.2013.14.2.915 23621261 140. Wadhwa R Takano S Kaur K Deocaris CC Pereira-Smith OM Reddel RR Upregulation of mortalin/mthsp70/Grp75 contributes to human carcinogenesis Int J Cancer 2006 118 2973 2980 10.1002/ijc.21773 16425258 141. Petrova DT Asif AR Armstrong VW Dimova I Toshev S Yaramov N Expression of chloride intracellular channel protein 1 (CLIC1) and tumor protein D52 (TPD52) as potential biomarkers for colorectal cancer Clin Biochem 2008 41 1224 1236 10.1016/j.clinbiochem.2008.07.012 18710659 142. Suehara Y Kikuta K Nakayama R Tochigi N Seki K Ichikawa H GST-P1 as a histological biomarker of synovial sarcoma revealed by proteomics PROTEOMICS – Clinical Applications 2009 3 623 634 10.1002/prca.200800211 143. Pucci-Minafra I Cancemi P Albanese NN Di Cara G Marabeti MR Marrazzo A New Protein Clustering of Breast Cancer Tissue Proteomics Using Actin Content as a Cellularity Indicator J Proteome Res 2008 7 1412 1418 10.1021/pr700748m 18284189 144. Semaan SM Wang X Stewart PA Marshall AG Sang Q-XA Differential phosphopeptide expression in a benign breast tissue, and triple-negative primary and metastatic breast cancer tissues from the same African-American woman by LC-LTQ/FT-ICR mass spectrometry Biochem Biophys Res Commun 2011 412 127 131 10.1016/j.bbrc.2011.07.057 21802409 145. Michael MZ O’ Connor SM Van Holst Pellekaan NG Young GP James RJ Reduced Accumulation of Specific MicroRNAs in Colorectal Neoplasia Mol Cancer Res 2003 1 882 891 14573789 146. Li C Tan YX Zhou H Ding SJ Li SJ Ma D Proteomic analysis of hepatitis B virus-associated hepatocellular carcinoma: Identification of potential tumor markers Proteomics 2005 5 1125 1139 10.1002/pmic.200401141 15759316 147. Stephenson AJ Smith A Kattan MW Satagopan J Reuter VE Scardino PT Integration of gene expression profiling and clinical variables to predict prostate carcinoma recurrence after radical prostatectomy Cancer 2005 104 290 298 10.1002/cncr.21157 15948174 148. Erić A Juranić Z Milovanović Z Marković I Inić M Stanojević-Bakić N Effects of Humoral Immunity and Calreticulin Overexpression on Postoperative Course in Breast Cancer Pathol Oncol Res 2009 15 89 90 10.1007/s12253-008-9112-2 18925427 149. Toquet C Jarry A Bou-Hanna C Bach K Denis MG Mosnier JF Altered Calreticulin expression in human colon cancer: Maintenance of Calreticulin expression is associated with mucinous differentiation Oncol Rep 2007 17 1101 1107 17390051 150. Chen CN Chang CC Su TE Hsu WM Jeng YM Ho MC Identification of Calreticulin as a Prognosis Marker and Angiogenic Regulator in Human Gastric Cancer Ann Surg Oncol 2009 16 524 533 10.1245/s10434-008-0243-1 19050968 151. Yoon GS Lee H Jung Y Yu E Moon HB Song K Nuclear Matrix of Calreticulin in Hepatocellular Carcinoma Cancer Res 2000 60 1117 1120 10706133 152. Hong SH Misek DE Wang H Puravs E Giordano TJ Greenson JK An Autoantibody-Mediated Immune Response to Calreticulin Isoforms in Pancreatic Cancer Cancer Res 2004 64 5504 5510 10.1158/0008-5472.CAN-04-0077 15289361 153. Alaiya A Roblick U Egevad L Carlsson A Franzén B Volz D Polypeptide expression in prostate hyperplasia and prostate adenocarcinoma Anal Cell Pathol 2000 21 1 9 10.1155/2000/351963 11254220 154. Thomas PA Kirschmann DA Cerhan JR Folberg R Seftor EA Sellers TA Association between Keratin and Vimentin Expression, Malignant Phenotype, and Survival in Postmenopausal Breast Cancer Patients Clin Cancer Res 1999 5 2698 2703 10537332 155. Fuyuhiro Y Yashiro M Noda S Kashiwagi S Matsuoka J Doi Y Clinical Significance of Vimentin-positive Gastric Cancer Cells Anticancer Res 2010 30 5239 5243 21187520 156. Dauphin M Barbe C Lemaire S Nawrocki-Raby B Lagonotte E Delepine G Vimentin expression predicts the occurrence of metastases in non small cell lung carcinomas Lung Cancer 2013 81 117 122 10.1016/j.lungcan.2013.03.011 23562674 157. Handra-Luca A Hong SM Walter K Wolfgang C Hruban R Goggins M Tumour epithelial vimentin expression and outcome of pancreatic ductal adenocarcinomas Br J Cancer 2011 104 1296 1302 10.1038/bjc.2011.93 21448168

PMC005xxxxxx/PMC5002106.txt

==== Front Environ HealthEnviron HealthEnvironmental Health1476-069XBioMed Central London 17210.1186/s12940-016-0172-6ResearchThe contribution of motor vehicle emissions to ambient fine particulate matter public health impacts in New York City: a health burden assessment Kheirbek Iyad ikheirbe@health.nyc.gov 1Haney Jay Jay.Haney@icfi.com 2Douglas Sharon Sharon.Douglas@icfi.com 2Ito Kazuhiko kito1@health.nyc.gov 1Matte Thomas tmatte@health.nyc.gov 11 New York City Department of Health and Mental Hygiene, Bureau of Environmental Surveillance and Policy, 125 Worth Street, Third Flr. CN-34E, New York, NY 10013 USA 2 ICF International, 101 Lucas Valley Road, Suite 260, San Rafael, CA 94903 USA 26 8 2016 26 8 2016 2016 15 1 8916 3 2016 12 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 On-road vehicles are an important source of fine particulate matter (PM2.5) in cities, but spatially varying traffic emissions and vulnerable populations make it difficult to assess impacts to inform policy and the public. Methods We estimated PM2.5-attributable mortality and morbidity from on-road vehicle generated air pollution in the New York City (NYC) region using high-spatial-resolution emissions estimates, air quality modeling, and local health incidence data to evaluate variations in impacts by vehicle class, neighborhood, and area socioeconomic status. We developed multiple ‘zero-out’ emission scenarios focused on regional and local cars, trucks, and buses in the NYC region. We simulated PM2.5 concentrations using the Community Multi-scale Air Quality Model at a 1-km spatial resolution over NYC and combined modeled estimates with monitored data from 2010 to 2012. We applied health impact functions and local health data to quantify the PM2.5-attributable health burden on NYC residents within 42 city neighborhoods. Results We estimate that all on-road mobile sources in the NYC region contribute to 320 (95 % Confidence Interval (CI): 220–420) deaths and 870 (95 % CI: 440–1280) hospitalizations and emergency department visits annually within NYC due to PM2.5 exposures, accounting for 5850 (95 % CI: 4020–7620) years of life lost. Trucks and buses within NYC accounted for the largest share of on-road mobile-attributable ambient PM2.5, contributing up to 14.9 % of annual average levels across 1-km grid cells, and were associated with 170 (95 % CI: 110–220) PM2.5-attributable deaths each year. These contributions were not evenly distributed, with high poverty neighborhoods experiencing a larger share of the exposure and health burden than low poverty neighborhoods. Conclusion Reducing motor vehicle emissions, especially from trucks and buses, could produce significant health benefits and reduce disparities in impacts. Our high-spatial-resolution modeling approach could improve assessment of on-road vehicle health impacts in other cities. Electronic supplementary material The online version of this article (doi:10.1186/s12940-016-0172-6) contains supplementary material, which is available to authorized users. Keywords Fine particulate matter (PM2.5)Community multiscale air quality model (CMAQ)BenMAPTrafficHealth impact assessmentAir quality managementNew York City tax levy fundsissue-copyright-statement© The Author(s) 2016 ==== Body Background Fine particulate matter (PM2.5) is a common air pollutant that has been associated with multiple adverse health outcomes [1]. Despite declines in PM2.5 concentrations in New York City (NYC), recent estimates suggest ambient levels contribute to large numbers of avoidable premature deaths and diseases [2], and studies have shown a significant association between traffic-related air pollution and premature mortality [3]. Other studies have shown increased risk of respiratory and cardiovascular disease associated with close residential proximity to traffic pollution [4, 5]. Air quality public health impact analyses have emerged as an important approach for estimating the public health toll of air pollution, comparing its risks to other public health threats, and evaluating strategies and regulations designed to reduce exposures. Assessing source-specific contributions to health burdens can help prioritize strategies that offer the maximum benefit and minimize inequalities [6]. Typically, air quality and health modeling analyses performed for regulatory decision making or policy research are conducted at coarse spatial resolutions (e.g. 12-km, 36-km, county-level) [7–9]. However, analyses at these spatial scales do not allow researchers and policymakers to examine relationships between population health susceptibility and air pollution exposures, both of which can spatially vary substantively within a city at smaller scales. To address these limitations in the regulatory methodology, new modeling approaches are needed to combine information on within city disparities in both exposures and health. While regulatory efforts have reduced emissions, on-road mobile sources continue to contribute to ambient levels of multiple air pollutants in NYC. Local source apportionment analyses conducted using data from the early 2000s suggested that 16–39 % of ambient PM2.5 concentrations in NYC were attributable to traffic sources [10, 11]. More recently, saturation sampling and land-use regression (LUR) modeling have demonstrated that traffic emissions density is an important contributor to within-city spatial variation in PM2.5, nitrogen dioxide (NO2) and black carbon levels in NYC [12]. While these studies provide useful information on the relative importance of local source sectors, source apportionment analyses using monitor data are limited by the locations of monitors, while LUR models that use surrogate indicators of emissions do not account for dispersion and chemical transformation processes and therefore may not be well-suited to quantify source contributions to ambient levels. Deterministic models of emissions, dispersion and chemical transformation processes can estimate exposure increments from individual sources, and recent developments in methods using high resolution modeling in urban areas can better represent spatial gradients across neighborhoods with wide variation in baseline health incidence [13, 14]. NYC, with high densities of populations living near emissions sources, also has the highest density of primary PM2.5 emissions among large US cities [15]. Wide variation in baseline health rates exist across the city, strongly associated with area-based poverty concentration [16]. Recent sustainability planning efforts in NYC have focused on reducing PM2.5 levels overall while shrinking disparities in exposures [17]. However, to date, limited data exist on the health burden from on-road mobile source emissions, the relative importance of regional as compared to local sources, or the differential contributions of differing vehicle classes. This information provides valuable context for developing and prioritizing local policy for cities. To evaluate the extent and variation of PM2.5-attributable mortality and morbidity due to emissions of on-road mobile source primary PM2.5 and PM2.5 precursors in the region, we applied a local-scale air quality and health modeling framework to the five counties of NYC and the 28-county NYC metropolitan region. We estimated separately the PM2.5-attributable burden from emissions from all motor vehicle traffic in the region and within NYC, trucks and buses within NYC, cars within NYC, and on-road mobile sources in the region outside of NYC. We then explored the disparity in air quality and public health burden across neighborhoods of differing poverty status. Methods We built an air quality and public health modeling framework for on-road mobile sources that included emissions inventory development and spatial allocation, meteorological and air quality modeling, combining modeling results and monitored data, and health impact calculations by modifying a prior framework used in an evaluation of heating fuel conversions in buildings [14, 18] (Fig. 1).Fig. 1 Data inputs and models for estimating the PM2.5-attributable public health burden from motor vehicles Emissions inventory preparation To characterize baseline conditions we built inventories from EPA’s 2008 National Emissions Inventory (NEI) modeling platform [19]. Described in detail elsewhere, we prepared emissions for three nested grids centered over NYC at 15-km national-scale, 5-km regional-scale, and 1-km local-scale horizontal resolution [18]. We replaced emissions for the on-road mobile source and building heating sectors in the 2008 NEI with more recent, refined local data to better reflect their spatial patterns. We estimated on-road mobile source emissions using the most recently available county-level data from EPA’s 2011 National Emissions Inventory [20]. County-level emission estimates were spatially allocated to road links in proportion to modeled, link-level vehicle miles traveled from the 2005 New York State Metropolitan Transportation Council (NYMTC) Best Practices Model (BPM) [21]. Despite the relatively older time frame of the NYMTC BPM model, it provided the most recently available modeled counts for cars, trucks, and buses for links within the 28 counties in the NYC region and we assumed that relative spatial patterns in traffic density were reasonably stable between 2005 and 2011. To improve the spatial accuracy within the five NYC counties, the NYMTC shapefile was spatially aligned to the TeleAtlas street segment database within ArcGIS 9.2 Data & Maps. For grid cells within the 1-km and 5-km modeling grids, we calculated on-road mobile source emissions of total volatile organic compounds (VOC), oxides of nitrogen (NOx), carbon monoxide (CO), sulfur dioxide (SO2), ammonia (NH3), primary PM2.5, and PM2.5 and VOC species profiles. Emissions were allocated do grid cells by first computing at each roadway link the vehicle miles traveled (VMT) by vehicle class (car, truck, bus) by multiplying the annual NYMTC vehicle-specific counts by the length of the segment. We then created ratios, by vehicle type, of the VMTs on each link to the total VMTs in the county. Second, we downloaded the on-road mobile source portion of the 2011 EPA NEI (V1) [20] and matched the Source Classification Code (SCC) subcategories to NYMTC car, truck, and bus categories. All light-duty and heavy-duty gasoline and diesel truck SCC codes were placed in the ‘truck’ category, while light-duty gasoline and diesel vehicles and motorcycles were included in the ‘car’ category. The heavy duty diesel bus SCC codes were included in the ‘bus’ category. Third, we estimated annual link-level emissions for each pollutant and vehicle type by multiplying the county-level emissions by the ratio of the VMTs on each link to the total VMTs in the county. Fourth, we created emissions totals for each pollutant/vehicle type in each 1-km and 5-km grid cell by summing the emissions from links that fell within each grid cell. For links that crossed multiple grid cells, emissions were apportioned based on the fraction of the link’s length included in each grid cell. Finally, because NYMTC does not include VMT breakdowns for categories within ‘car,’ ‘truck’, and ‘bus,’ we approximated these by extracting the county-level VMT data from EPA’s 2008 VMT database [22], then calculated the VMT fractions for gasoline, light- and heavy-duty diesel vehicles. This was then used to assign VOC and PM2.5 speciation profiles by estimating the VOC and PM2.5 emissions for gasoline, light- and heavy-duty diesel vehicles using the VMT fractions for each vehicle type, and then assigning the corresponding PM2.5 and VOC speciation profiles to the each of the categories. To more accurately represent current building boiler emissions in NYC overall and the within the city, we updated the 2008 NEI for Nos. 2, 4, and 6 heating oil boilers using local permit data reflecting emissions as of 2015. These methods are described in detail elsewhere [14]. Briefly, emissions from Nos. 2, 4, and 6 boilers were calculated using EPA emissions factors [23] and NYC Department of Environmental Protection (NYCDEP) permit data that identify the location and heat throughput of the boiler. No. 4 emissions factors were adjusted to account for NYC-regulated 1500 ppm sulfur content, while No.2 emissions factors assumed a 15 ppm sulfur content, consistent with New York State-wide limits [24]. As many buildings are undergoing conversions of Nos. 4 and 6 boilers to comply with recent regulations [25], we reviewed the permit database as of December 2014, and assigned each building an annual emissions value based on the fuel they were using at that time and spatially allocated these emissions based on boiler address. To estimate emissions from No.2 boilers below the permitting threshold (350,000 Btu), we used NEI emissions not accounted for in the permits, allocating to buildings using surrogate data on building area and the county-specific percent of buildings using No.2 heating oil. We prepared CMAQ-ready emissions by merging estimates for biogenic sources and all anthropogenic sectors with the updated on-road mobile source inventory and fuel oil boiler inventories. These emissions were processed using the Sparse Matrix Operator Kernel Emissions processor software (version 3.1) to create the air quality modeling input for the base case. We created three additional inventories reflecting removal of specific source categories: zeroing out all motor vehicle emissions within NYC (Sc1), zeroing out truck and bus emissions within NYC (Sc2), and zeroing out all motor vehicles in the 23 counties that surround the five NYC counties (Sc3). Air quality modeling Detailed discussion of the application and evaluation of the meteorological and air quality modeling system has been presented elsewhere [18]. In short, meteorological fields for all grids were developed for 2008 using the Weather Research and Forecasting Model (WRF). Air quality modeling was conducted using the Community Multi-Scale Air Quality Model (CMAQ) version 5.0. Annual CMAQ simulations were conducted separately for the base case and each of the three zero-out scenarios and we utilized the daily simulated PM2.5 mass and species concentrations from the 1-km grid cells over NYC for subsequent health burden analyses. Health burden analysis Exposure estimates at a 1-km resolution were developed using EPA’s Modeled Attainment Test Software (MATS) [26]. MATS combines the CMAQ modeled output with monitored PM2.5 mass and speciation data to create combined spatial surfaces, providing exposure estimates that use the monitor data but leverages the CMAQ simulated values to better estimate spatial gradients and surface response to changes in emissions. We developed 3 year, quarterly average estimates based on 2010–2012 EPA federal reference monitors (FRM) and speciation trends network (STN) monitors and the daily CMAQ modeling. We computed the change in number of health events due to changes in PM2.5 concentrations between the base case and each of the three scenarios (Sc1, Sc2, and Sc3) using health impact functions [27, 28]. We applied risk functions for all-cause mortality from chronic exposure among those above 30 years of age [29], emergency department visits for asthma from acute exposure among all age groups (seasonally-specific risk estimates) [30], hospitalizations for all respiratory outcomes from acute exposure among those above 20 years of age (seasonally specific risk estimates for populations above 65 years of age) [31, 32], and hospitalizations for all cardiovascular outcomes from acute exposure among those above 40 years of age (seasonally specific risk estimates) [33]. Risk functions were chosen based on those determined to be most relevant to current New York City populations by selecting those published in peer-reviewed scientific journals and favoring those conducted in New York City when possible. We utilized NYC-specific risk functions for PM2.5-attributable emergency department visits for asthma and hospitalizations for cardiovascular disease. When local studies were not available, we used recent large, multi-city studies or those included in EPA risk analyses [34]. Baseline health data were obtained for 2009–2011 from the NYC Department of Health and Mental Hygiene Bureau of Vital Statistics and the New York Statewide Planning and Research Cooperative System, summarized across 22 age and sex groups within each of 42 United Hospital Fund (UHF) neighborhoods (zip code aggregates). Additional details on risk estimate selection and baseline health data is described elsewhere [27]. Population data for the same age/sex/neighborhood groups were calculated based on the US Census Bureau Population Estimate program [35]. We estimated 3 year, quarterly average health impacts of each of the scenarios within each of the 42 neighborhoods and summed the quarterly estimates to produce annual burdens. All health impact calculations were performed on a quarterly basis using EPA’s Benefits Mapping and Analysis Program (BenMAP) version 4.067 [36]. Further detail on our methodological choices for estimating the public health burden of PM2.5 on NYC residents can be found elsewhere [27]. To estimate years of life expectancy lost (YLL) we calculated life expectancy for 5 year age groupings using the city-wide, baseline mortality rates and standard abridged life table methods from the Centers for Disease Control and Prevention [37]. Years of life lost due to exposures associated with each scenario were calculated by multiplying the number of deaths in each age group attributable to the change in PM2.5 by the remaining life expectancy, then summing across all ages. We first report the impacts on a citywide basis of removing all traffic in the 28-county region (adding Sc1 heath impacts to Sc3 health impacts), all traffic within NYC (Sc1), trucks and buses within NYC (Sc2), cars within NYC (subtracting Sc2 health impacts from Sc1 health impacts) and traffic from sources within the region but outside of NYC (Sc3). We explore correlations between on-road mobile source category contributions to ambient PM2.5 and neighborhood poverty then grouped neighborhoods based on percent of population residing under the federal poverty threshold (Low: 0–10 %, Medium: 10–20 %, High: 20–30 %, and Very High >30 %), calculated using the 2008–2012 American Community Survey. We report gradients in PM2.5 concentrations, rates of PM2.5-attributable health outcomes, and percent contribution to the total number of health events by neighborhood poverty level. Results Emissions from motor vehicles within NYC produced 1817 tons of primary PM2.5, 43,934 tons of NOx, 20,613 tons of total VOCs, and 336 tons of SO2, annually, accounting for 17.5, 38.3, 21.9, and 4.6 % of all local pollutant emissions, respectively. Of the primary PM2.5 emissions produced by motor vehicles, the majority are produced by trucks and buses, accounting for 12.8 % of all local primary PM2.5 emissions. Based on the CMAQ model alone, primary PM2.5 concentrations attributable to truck and bus emissions within NYC contributed to an average of 27 % of total PM2.5 concentrations from all on-road mobile sources in the region. Secondarily generated PM2.5 from truck and bus precursor emissions within NYC accounted for an average of 12 % of total PM2.5 concentrations from all on-road mobile sources in the region (Additional file 1: Table S1). Cars within NYC contributed to an average of 10 and 25 % of total PM2.5 concentrations from all on-road mobile sources in the region due to primary and secondarily formed PM2.5, respectively (Additional file 1: Table S1). Based on link level NYMTC estimates within NYC, cars contributed 94 % of city VMTs while trucks and buses accounted for 6 %. Based on the combined model and monitor exposure surface, we estimate that traffic in the 28-county area contributed 0.38 to 2.60 μg/m3 across 1-km grid cells within NYC, accounting for 3.9 to 22.7 % of ambient PM2.5 levels (Fig. 2). Trucks and buses within NYC showed the largest within city contributions to ambient levels, accounting for 0.0 to 1.71 μg/m3 of PM2.5, or 0.0 to 14.9 % of PM2.5 concentrations. Emissions from cars within NYC and regional traffic (outside NYC) showed less of a contribution, with regional traffic mainly impacting grid cells along the edges of the city.Fig. 2 Estimated PM2.5 levels in the Base Case and contributions to ambient levels from on-road mobile source categories (1-km resolution) We estimate that, each year, emissions from on-road mobile sources within the five NYC counties contribute to 260 (95 % CI: 180, 340) PM2.5-attributable deaths from chronic PM2.5 exposure and 720 (95 % CI: 380, 1050) PM2.5-attributable emergency room visits and hospital admissions due to respiratory and cardiovascular outcomes from acute exposure (Table 1). Among these, emissions from buses and trucks account for the largest share of the city-wide burden, contributing to 170 (95 % CI: 110, 220) PM2.5-attributable deaths each year while cars contributed to 100 (95 % CI: 70, 120) PM2.5-attributable deaths each year. On-road mobile sources in the metropolitan area outside of the five NYC counties contribute to an additional 60 (95 % CI: 40, 80) PM2.5-attributable deaths each year and 150 (95 % CI: 70, 220) PM2.5-attributable morbidity outcomes each year. Overall, we estimate PM2.5 exposures from on-road mobile sources in the metropolitan region contribute to 320 (95 % CI: 220, 420) PM2.5-attributable deaths each year within NYC, contributing to 5850 (95 % CI: 4020, 7680) life years lost annually. The confidence intervals reported here only reflect those from the risk estimates derived from the epidemiologic studies and do not account for uncertainties in the other analysis steps.Table 1 City-wide PM2.5-attributable health burdens of on-road mobile source emissions Count (95 % CI), percent of all events (95 % CI), percent of PM2.5-attributable events (95 % CI) All motor vehicles in metropolitan region (Sc1 health impacts plus Sc3 health impacts) All motor vehicles in NYC (Sc1 health impacts) Buses and trucks in NYC (Sc2 health impacts) Cars in NYC (Sc1 health impacts minus Sc2 health impacts) All motor vehicles outside NYC (Sc3 impacts) Emergency Room Visits, Respiratory (All Ages, acute exposure) 660 (380, 940), 0.76 % (0.44 %, 1.1 %), 13.11 % (7.5 %, 18.6 %) 550 (320, 780), 0.64 % (0.37 %, 0.90 %), 10.94 % (6.3 %, 15.5 %) 360 (210, 510), 0.42 % (0.24 %, 0.59 %), 7.19 % (4.17 %, 10.12 %) 190 (100, 270), 0.22 % (0.12 %, 0.31 %), 3.75 % (1.98 %, 5.36 %) 110 (60, 160), 0.13 % (0.07 %, 0.19 %), 2.17 % (1.19 %, 3.17 %) Hospital Admissions, Cardiovascular (Ages 40 and above, acute exposure) 90 (20, 150), 0.14 % (0.03 %, 0.25 %), 13.32 % (3.1 %, 23.0 %) 70 (20, 120), 0.12 % (0.03 %, 0.20 %, 10.94 % (3.1 %, 18.4 %) 40 (10, 80), 0.07 % (0.02 %, 0.13 %), 6.84 % (1.53 %, 12.27 %) 30 (10, 50), 0.04 % (0.02 %, 0.08 %), 4.09 % (1.53 %, 7.67 %) 20 (4, 30), 0.03 % (0.01 %, 0.05 %), 2.41 % (0.61 %, 4.60 %) Hospital Admissions, Respiratory (Ages 20 and above, acute exposure) 120 (40, 190), 0.27 % (0.09 %, 0.45 %), 12.96 % (4.5 %, 21.3 %) 100 (40, 150), 0.22 % (0.09 %, 0.35 %), 10.68 % (4.48 %, 16.8 %) 60 (20, 100), 0.14 % (0.05 %, 0.24 %, 6.80 % (2.24 %, 11.21 %) 30 (10, 60), 0.08 % (0.02 %, 0.14 %), 3.88 % (1.12 %, 6.73 %) 20 (10, 30), 0.05 % (0.02 %, 0.07 %), 2.28 % (1.12 %, 3.36 %) Premature Mortality (Ages 30 and above, chronic exposure) 320 (220, 420), 0.68 % (0.47 %, 0.89 %), 13.22 % (9.14 %, 17.44 %) 260 (180, 340), 0.55 % (0.38 %, 0.72 %), 10.81 % (7.48 %, 14.12 %) 170 (110, 220), 0.35 % (0.23 %, 0.47 %), 6.86 % (4.57 %, 9.14 %) 100 (70, 120), 0.2 % (0.15 %, 0.26 %), 3.95 % (2.91 %, 4.98 %) 60 (40, 80), 0.12 % (0.09 %, 0.17 %), 2.41 % (1.67 %, 3.32 %) Years of Life Lost (Ages 30 and above, chronic exposure) 5850 (4020, 7680) 4800 (3300, 6300) 3050 (2090, 4000) 1750 (1210, 2300) 1050 (720, 1380) We observed only a weak relationship between baseline PM2.5 concentrations and neighborhood poverty status, due to variable levels of PM2.5 across high income neighborhoods (Fig. 3). Affluent neighborhoods in NYC include many densely developed areas in Manhattan with high source density as well as more suburban neighborhoods with fewer emissions in Staten Island and Queens. However, there is a stronger relationship between on-road mobile-source-attributable PM2.5 and neighborhood poverty. This relationship is consistent for both bus/truck-attributable PM2.5 and car-attributable PM2.5, although a steeper gradient is found for bus/truck-attributable PM2.5 (average absolute difference in impact between low and high poverty neighborhoods of 0.36 μg/m3 for trucks/buses and 0.22 μg/m3 for cars), reflecting high densities of truck traffic in low-income neighborhoods.Fig. 3 Correlations of estimated baseline PM2.5 concentrations and contributions from on-road mobile sources with neighborhood poverty metrics There are large disparities in PM2.5-attributable health outcomes across neighborhoods with variable poverty status (Table 2). Across all source categories, higher mobile source PM2.5-attributable rates of morbidity and mortality are found in high poverty neighborhoods as compared to low poverty neighborhoods. This is due to the large disparity in the underlying rates of morbidity and mortality and higher on-road mobile source impacts on PM2.5 concentrations. The widest disparities are found for PM2.5-attributable emergency department visits for asthma. On-road mobile sources in the region contribute to rates of PM2.5-attributable asthma emergency department visits that are 8.3 times higher in the very high poverty neighborhoods relative to low poverty neighborhoods, due to high source density and relatively high asthma morbidity rates in these communities. The percent of incidences due to on-road mobile sources, which reflect the impacts of sources on neighborhood PM2.5 levels, also showed disparities across neighborhoods of varying poverty status, with higher percentages in lower income neighborhoods across all source categories, except regional traffic emissions outside of NYC. Regional traffic emissions did not produce large gradients in the percent of incidences across neighborhoods of varying poverty status due to relatively even impacts on PM2.5 concentrations across the city, with slightly higher impacts on PM2.5 concentrations in some grid cells along the edges of the City in higher income neighborhoods of Manhattan and the northern Bronx.Table 2 Distribution of PM2.5-attributable health outcomes due to on-road mobile sources by area poverty Metric Source sector Low poverty Medium poverty High poverty Very high poverty (n = 6) (n = 19) (n = 10) (n = 7) Impacts on PM2.5 Concentrations (μg/m3, percent of ambient concentrations) All on-road mobile sources in region 1.09 (10.9 %) 1.14 (11.2 %) 1.21 (11.6 %) 1.64 (14.0 %) All on-road mobile sources in NYC 0.81 (8.1 %) 0.94 (9.2 %) 0.97 (9.3 %) 1.39 (11.9 %) Trucks and buses in NYC 0.51 (5.1 %) 0.60 (5.9 %) 0.62 (6.0 %) 0.87 (7.5 %) Cars in NYC 0.30 (3.0 %) 0.33 (3.3 %) 0.35 (3.3 %) 0.52 (4.5 %) All on-road mobile sources outside NYC 0.28 (2.8 %) 0.21 (2.0 %) 0.24 (2.3 %) 0.25 (2.1 %) Impacts on Mortality among residents above 30 years of age PM2.5-attributable rate per 100,000 residents (95 % CI), Percent of all events (95 % CI), Percent of PM2.5-attributable events (95 % CI) All on-road mobile sources in region 5.27 (3.62, 6.92) 0.58 % (0.40 %, 0.76 %) 11.8 % (8.1 %, 15.4 %) 5.86 (4.03, 7.69) 0.63 % (0.43 %, 0.82 %) 12.6 % (8.7 %, 16.6 %) 7.36 (5.06, 9.66) 0.7 % (0.48 %, 0.92 %) 13.7 % (9.4 %, 17.9 %) 8.98 (6.17, 11.78) 0.88 % (0.60 %, 1.15 %) 15.3 % (10.5 %, 20.1 %) All on-road mobile sources in NYC 3.92 (2.69, 5.14) 0.43 % (0.30 %, 0.57 %) 8.7 % (6 %, 11.5 %) 4.81 (3.31, 6.32) 0.52 % (0.35 %, 0.68 %) 10.4 % (7.1 %, 13.6 %) 6.03 (4.14, 7.92) 0.57 % (0.39 %, 0.75 %) 11.2 % (7.7 %, 14.7 %) 7.6 (5.22, 9.97) 0.74 % (0.51 %, 0.97 %) 12.9 % (8.9 %, 17 %) Trucks and buses in NYC 2.51 (1.72, 3.29) 0.28 % (0.19 %, 0.36 %) 5.6 % (3.8 %, 7.3 %) 3.07 (2.11, 4.03) 0.33 % (0.23 %, 0.43 %) 6.6 % (4.5 %, 8.7 %) 3.85 (2.65, 5.06) 0.37 % (0.25 %, 0.48 %) 7.2 % (4.9 %, 9.4 %) 4.73 (3.25, 6.20) 0.46 % (0.32 %, 0.61 %) 8 % (5.5 %, 10.6 %) Cars in NYC 1.41 (0.97, 1.85) 0.16 % (0.11 %, 0.20 %) 3.1 % (2.2 %, 4.1 %) 1.75 (1.20, 2.29) 0.19 % (0.13 %, 0.25 %) 3.8 % (2.6 %, 4.9 %) 2.18 (1.50, 2.86) 0.21 % (0.14 %, 0.27 %) 4 % (2.8 %, 5.3 %) 2.87 (1.98, 3.77) 0.28 % (0.19 %, 0.37 %) 4.9 % (3.4 %, 6.4 %) All on-road mobile sources outside NYC 1.35 (0.93, 1.78) 0.15 % (0.10 %, 0.20 %) 3.0 % (2.1 %, 4.0 %) 1.05 (0.72, 1.37) 0.11 % (0.08 %, 0.15 %) 2.3 % (1.5 %, 3.0 %) 1.33 (0.91, 1.74) 0.13 % (0.09 %, 0.17 %) 2.5 % (1.7 %, 3.2 %) 1.38 (0.95, 1.81) 0.13 % (0.09 %, 0.18 %) 2.4 % (1.6 %, 3.1 %) Impacts on Emergency Department Visits for Asthma among all residents (PM2.5-attributable rate per 100,000 residents (95 % CI), Percent of all events (95 % CI) Percent of PM2.5-attributable events (95 % CI) All on-road mobile sources in region 2.39 (1.4, 3.39) 0.64 % (0.37 %, 0.9 %) 10.8 % (6.3 %, 15.3 %) 4.64 (2.71, 6.58) 0.7 % (0.41 %, 1 %) 12 % (7 %, 17 %) 9.54 (5.51, 13.56) 0.76 % (0.44 %, 1.09 %) 13.2 % (7.6 %, 18.7 %) 19.97 (11.37, 28.56) 0.83 % (0.47 %, 1.18 %) 14.2 % (8.1 %, 20.3 %) All on-road mobile sources in NYC 1.79 (1.04, 2.53) 0.48 % (0.28 %, 0.67 %) 8.1 % (4.7 %, 11.4 %) 3.87 (2.26, 5.47) 0.59 % (0.34 %, 0.83 %) 10 % (5.8 %, 14.1 %) 7.89 (4.57, 11.19) 0.63 % (0.37 %, 0.9 %) 10.9 % (6.3 %, 15.5 %) 16.96 (9.66, 24.23) 0.7 % (0.4 %, 1 %) 12.1 % (6.9 %, 17.3 %) Trucks and buses in NYC 1.17 (0.7, 1.65) 0.31 % (0.19 %, 0.44 %) 5.3 % (3.2 %, 7.5 %) 2.58 (1.54, 3.62) 0.39 % (0.23 %, 0.55 %) 6.7 % (4 %, 9.4 %) 5.21 (3.09, 7.32) 0.42 % (0.25 %, 0.59 %) 7.2 % (4.3 %, 10.1 %) 10.98 (6.42, 15.54) 0.45 % (0.27 %, 0.64 %) 7.8 % (4.6 %, 11.1 %) Cars in NYC 0.61 (0.34, 0.88) 0.16 % (0.09 %, 0.23 %) 2.8 % (1.5 %, 4 %) 1.29 (0.72, 1.85) 0.19 % (0.11 %, 0.28 %) 3.3 % (1.8 %, 4.8 %) 2.68 (1.48, 3.87) 0.21 % (0.12 %, 0.31 %) 3.7 % (2 %, 5.3 %) 5.97 (3.25, 8.69) 0.25 % (0.13 %, 0.36 %) 4.3 % (2.3 %, 6.2 %) All on-road mobile sources outside NYC 0.61 (0.36, 0.86) 0.16 % (0.1 %, 0.23 %) 2.7 % (1.6 %, 3.9 %) 0.78 (0.45, 1.1) 0.12 % (0.07 %, 0.17 %) 2 % (1.2 %, 2.8 %) 1.65 (0.94, 2.36) 0.13 % (0.08 %, 0.19 %) 2.3 % (1.3 %, 3.3 %) 3.02 (1.7, 4.33) 0.12 % (0.07 %, 0.18 %) 2.1 % (1.2 %, 3.1 %) Impacts on Hospitalizations for Cardiovascular Disease among residents over 40 years of age. (PM2.5-attributable rate per 100,000 residents (95 % CI), Percent of all events (95 % CI) Percent of PM2.5-attributable events (95 % CI) All on-road mobile sources in region 1.59 (0.39, 2.78) 0.12 % (0.03 %, 0.21 %) 11.5 % (2.8 %, 20.3 %) 2.12 (0.52, 3.71) 0.13 % (0.03 %, 0.23 %) 12.7 % (3.1 %, 22.3 %) 2.74 (0.68, 4.8) 0.15 % (0.04 %, 0.26 %) 13.7 % (3.4 %, 24 %) 3.81 (0.96, 6.66) 0.18 % (0.05 %, 0.32 %) 15.3 % (3.8 %, 26.7 %) All on-road mobile sources in NYC 1.15 (0.28, 2.02) 0.09 % (0.02 %, 0.15 %) 8.4 % (2.1 %, 14.7 %) 1.75 (0.43, 3.07) 0.11 % (0.03 %, 0.19 %) 10.5 % (2.6 %, 18.4 %) 2.25 (0.56, 3.94) 0.12 % (0.03 %, 0.22 %) 11.2 % (2.8 %, 19.7 %) 3.23 (0.81, 5.64) 0.15 % (0.04 %, 0.27 %) 12.9 % (3.2 %, 22.6 %) Trucks and buses in NYC 0.72 (0.17, 1.28) 0.05 % (0.01 %, 0.09 %) 5.3 % (1.2 %, 9.3 %) 1.1 (0.26, 1.94) 0.07 % (0.02 %, 0.12 %) 6.6 % (1.6 %, 11.6 %) 1.42 (0.34, 2.5) 0.08 % (0.02 %, 0.14 %) 7.1 % (1.7 %, 12.5 %) 1.99 (0.48, 3.49) 0.1 % (0.02 %, 0.17 %) 8 % (1.9 %, 14 %) Cars in NYC 0.43 (0.11, 0.75) 0.03 % (0.01 %, 0.06 %) 3.1 % (0.8 %, 5.4 %) 0.65 (0.17, 1.13) 0.04 % (0.01 %, 0.07 %) 3.9 % (1 %, 6.8 %) 0.83 (0.22, 1.45) 0.05 % (0.01 %, 0.08 %) 4.2 % (1.1 %, 7.2 %) 1.24 (0.33, 2.15) 0.06 % (0.02 %, 0.1 %) 5 % (1.3 %, 8.6 %) All on-road mobile sources outside NYC 0.43 (0.1, 0.76) 0.03 % (0.01 %, 0.06 %) 3.2 % (0.8 %, 5.5 %) 0.37 (0.09, 0.65) 0.02 % (0.01 %, 0.04 %) 2.2 % (0.6 %, 3.9 %) 0.49 (0.12, 0.86) 0.03 % (0.01 %, 0.05 %) 2.5 % (0.6 %, 4.3 %) 0.58 (0.15, 1.01) 0.03 % (0.01 %, 0.05 %) 2.3 % (0.6 %, 4.1 %) Impacts on Hospitalizations for Respiratory Disease among residents above 20 years of age (PM2.5-attributable rate per 100,000 residents(95 % CI), Percent of all events (95 % CI) Percent of PM2.5-attributable events (95 % CI) All on-road mobile sources in region 1.13 (0.42, 1.85) 0.55 (0.22, 0.89) 11 % (4.1 %, 18 %) 1.44 (0.55, 2.34) 0.25 % (0.09 %, 0.4 %) 12.1 % (4.6 %, 19.6 %) 2.04 (0.8, 3.29) 0.28 % (0.11 %, 0.46 %) 13.4 % (5.2 %, 21.5 %) 3.58 (1.43, 5.71) 0.33 % (0.13 %, 0.52 %) 14.7 % (5.9 %, 23.5 %) All on-road mobile sources in NYC 0.84 (0.31, 1.37) 0.16 % (0.06 %, 0.27 %) 8.1 % (3 %, 13.3 %) 1.19 (0.45, 1.93) 0.2 % (0.08 %, 0.33 %) 10 % (3.8 %, 16.1 %) 1.69 (0.66, 2.71) 0.23 % (0.09 %, 0.38 %) 11 % (4.3 %, 17.7 %) 3.02 (1.21, 4.83) 0.28 % (0.11 %, 0.44 %) 12.4 % (5 %, 19.8 %) Trucks and buses in NYC 0.54 (0.2, 0.88) 0.11 % (0.04 %, 0.17 %) 5.3 % (1.9 %, 8.6 %) 0.77 (0.29, 1.24) 0.13 % (0.05 %, 0.21 %) 6.4 % (2.4 %, 10.4 %) 1.08 (0.42, 1.74) 0.15 % (0.06 %, 0.24 %) 7.1 % (2.8 %, 11.4 %) 1.87 (0.75, 2.99) 0.17 % (0.07 %, 0.27 %) 7.7 % (3.1 %, 12.3 %) Cars in NYC 0.3 (0.11, 0.48) 0.06 % (0.02 %, 0.1 %) 2.9 % (1.1 %, 4.7 %) 0.42 (0.16, 0.69) 0.07 % (0.03 %, 0.12 %) 3.5 % (1.3 %, 5.8 %) 0.6 (0.23, 0.97) 0.08 % (0.03 %, 0.13 %) 3.9 % (1.5 %, 6.3 %) 1.15 (0.46, 1.84) 0.11 % (0.04 %, 0.17 %) 4.7 % (1.9 %, 7.5 %) All on-road mobile sources outside NYC 0.3 (0.11, 0.48) 0.06 % (0.02 %, 0.09 %) 2.9 % (1.1 %, 4.7 %) 0.25 (0.09, 0.41) 0.04 % (0.02 %, 0.07 %) 2.1 % (0.8 %, 3.4 %) 0.36 (0.14, 0.58) 0.05 % (0.02 %, 0.08 %) 2.4 % (0.9 %, 3.8 %) 0.55 (0.22, 0.89) 0.05 % (0.02 %, 0.08 %) 2.3 % (0.9 %, 3.6 %) Baseline Outcome Rates (rate per 100,000 residents) All-cause mortality (ages 30 and above) 907.5 934.2 1051.1 1024.6 Emergency department visits for asthma (all ages) 374.9 659.5 1248.4 2416.1 Hospitalizations for cardiovascular disease (ages 40 and above) 1354.9 1589.3 1824.5 2089.1 Hospitalizations for respiratory disease (ages 20 and above) 508.9 589.0 722.8 1089.0 Discussion In this study, we applied a high-spatial-resolution modeling framework to assess the impacts of on-road mobile source generated primary PM2.5 and PM2.5 precursors on NYC populations. We estimated that over 300 deaths each year in the five counties of NYC are due to PM2.5 exposures related to motor vehicle emissions in the 28-county region, contributing to 5850 YLL annually. These exposures also contribute to almost 900 emergency department visits and hospitalizations for respiratory and cardiovascular disease annually. Overall, on-road mobile sources contribute to 0.7 % of all deaths in NYC each year and 13.2 % of PM2.5-attributable deaths, with the largest share of this impact due to emissions from trucks and buses on NYC roadways. Within NYC, we observed wide variation in incremental ambient PM2.5 contributions from traffic across 1-km grid cells. The largest impacts on air quality levels and health outcomes are found in the highest poverty areas of the city due to overlapping patterns of traffic density (particularly truck traffic) and higher underlying baseline incidence of morbidity. Comparative analysis of traffic types demonstrated that trucks and buses, despite a much lower share of overall VMT within the city, contribute to the largest share of the on-road mobile source air quality burden with the majority of the primary PM2.5 emissions coming from heavy duty diesel trucks and buses. We found that these sources contribute to 7.5 % of the ambient levels of PM2.5 in high poverty neighborhoods and up to 0.6 % of all deaths in the most affected neighborhood. Traffic from counties in the region outside of NYC showed less of an impact on local PM2.5 concentrations, which were evenly distributed across neighborhoods of varying poverty status. Prior work in other cities and nationally has also explored the air quality and public health impacts of traffic. A nationwide analysis indicted that all mobile sources (including all non-road, aircraft, locomotive, marine vessels, and ocean-going vessels) could contribute to 17,000 PM2.5-attributable deaths in 2016 [6]. Other research has pointed to the importance of traffic-related PM2.5 on mortality, suggesting that in 2005 traffic emissions contributed to 3000 PM2.5-attributable deaths nationally [8]. Applying a similar air quality and health modeling framework as was used in this analysis, additional research has suggested significant regional benefits to eliminating motor vehicle trips [38]. Prior natural experiments on removal of traffic in urban areas during events have shown some associated improvements in air quality, although the benefits are often pollutant specific and vary based on the situation being studied, particularly when evaluating pollutants with strong regional contributions such as PM2.5 [39, 40]. Source apportionment analyses conducted using data from the early 2000s from a few monitoring locations in the region suggested that 16–39 % of ambient PM2.5 concentrations in NYC are attributable to traffic sources [10, 11]. These estimates are higher than those found in this analysis, potentially due to the limitations in the numbers of monitors used in the source-apportionment studies (where monitoring sites are skewed to high emissions locations) and newer traffic emissions estimates that reflect lower emissions from on-road mobile sources in more recent years. To our knowledge this is the first analysis in this region that explicitly examines impacts of differing types of vehicles at a high spatial resolution across neighborhoods within an urban area, which provides valuable insight when exploring effective emissions control strategies. This analysis also provides a new perspective on variation in PM2.5 exposures across populations of differing socioeconomic status (SES). Prior work has found that higher SES communities in NYC experience higher overall PM2.5 and NO2 exposures, due to the confluence of building and traffic sources in high-income areas, a pattern that is unusual among major metropolitan areas where lower SES areas often experience higher pollutant exposures [41, 42]. In contrast to the pattern for total PM2.5, on-road mobile source-attributable PM2.5 concentrations are higher in low-income neighborhoods of the city, indicating that efforts to reduce exposures in these burdened communities should be focused on on-road mobile source-related programs. We find that measures to reduce emissions from heavy-duty vehicles within NYC should be prioritized, particularly those traveling roadways in neighborhoods with high densities of susceptible populations and low income residents. Studies conducted in other cities have shown success implementing congestion charging schemes or low-emissions zones that target the most polluting trucks and buses [43, 44], with differing observations on the distribution of benefits by socioeconomic status, depending on the location evaluated [45, 46]. In designing congestion mitigation schemes, this analysis suggests a focus on NYC as a whole and on heavy-duty diesel vehicles would yield significantly greater health benefits, as opposed to focusing on vehicles in the most congested urban core. Measures to reduce VMTs and emissions from trucks and buses within the city may need to address trips from all types of vehicles originating inside and outside of the city. For example, in the Hunt’s Point section of the South Bronx, an area with high burdens of PM2.5-attributable morbidity and mortality from truck emissions, an estimated 57 % of trucks servicing the meat and produce market (one of the largest food distribution centers in the world) came from outside of the city [47]. Other surveys have suggested 20 % of car miles traveled in NYC are from trips originating outside of the city [48]. While direct emissions from cars have less of an impact on air quality and health compared to heavy duty diesel vehicles, car trips contribute to congestion, which increases diesel emissions on routes shared with trucks and buses. While this study offers new insights and methods for assessing PM2.5-attributable health impacts, there are some limitations. The confidence intervals described in our results reflect only the confidence intervals reported in the risk estimates derived from the epidemiologic studies and do not account for uncertainties in the other steps of the analysis. EPA’s inventory estimates are subject to uncertainties in emissions factors, vehicle mix, and activity. Despite likely simulating spatial gradients in emissions better than other commonly used surrogates such as road density, they may not fully account for higher emissions in low speed stop-and-go traffic within the congested urban core. Future work would benefit for more precise estimates of emissions at a neighborhood-level. The meteorological and air quality simulations also carry uncertainties common in these types of studies. Prior evaluation of the base case modeling, however, showed that the WRF and CMAQ models performed within recommended bias and precision benchmarks [18]. A strength of our study is that it employs 1-km PM2.5 exposure modeling, a higher resolution than prior studies of this type and thus better accounts for within-city variations in susceptibility. This provides new methods and insight into how source-specific impacts can vary within an urban area and among populations of differing socioeconomic status. Despite this, 1-km resolution may not fully capture microscale, near-roadway exposures that can vary within several hundred meters of the roadway [49, 50]. Our health impact estimates include common limitations described elsewhere [28] some of which have been addressed by using neighborhood-level health outcome data. We have utilized epidemiological studies that assume uniform relative risk across all neighborhoods with varying traffic density. Emerging research has suggested stronger associations between asthma morbidity and air pollutant exposures in higher traffic areas (implying that PM2.5 emissions from traffic may be more toxic), and such effect modification research is a field of ongoing study [51]. Similarly, while our analyses applied risk estimates based on total PM2.5 exposures, recent analyses of the ACS cohort has suggested higher chronic mortality risk associated with PM2.5 with higher sulfur content [52]. As more evidence accumulates we will evaluate the sensitivity of our burden estimates to varying risk functions, and future work will evaluate how variations in neighborhood-level risk contribute to disparities in impacts across the City. While this analysis has leveraged associations between PM2.5 and excess emergency department visits and hospitalizations, studies have shown that air pollution exposures can also contribute to new cases of asthma [53], suggesting morbidity estimates in this analysis are conservative. Finally, our analysis focused only on the impacts of on-road mobile source emissions on PM2.5 associated mortality and select cardio-respiratory outcomes and does not account for the wide range of additional negative effects of motor vehicle traffic and congestion, including health effects associated with other pollutants and noise, contributions to greenhouse gas emissions, risk of pedestrian and other injury, and time wasted. Conclusion Local scale air quality and public health modeling can provide valuable information on the contribution of sources to pollution-attributable health and disparity within an urban area. In this study, we presented a methodology for assessing the public health impacts of traffic in cities, and evaluating these impacts across populations with varying underlying health and socioeconomic status. In applying these methods in NYC, we found that air pollutant emissions from on-road mobile sources contribute to hundreds of preventable PM2.5-attributable deaths, hospitalizations, and emergency department visits among residents of NYC, with disproportionate impacts in high poverty neighborhoods, indicating that increased policy efforts should focus on the most polluting vehicles in these neighborhoods. Additional file Additional file 1: Table S1. Contributions of source categories to primary and secondary PM2.5 levels at grid cells within New York City. (DOCX 13 kb) Abbreviations BenMAPEnvironmental Benefits Mapping and Analysis Program BPMBest Practices Model CIConfidence interval CMAQCommunity Multi-scale Air Quality Model COCarbon monoxide DEPNew York City Department of Environmental Protection EPAUS Environmental Protection Agency FRMFederal reference monitors LURLand use regression NEINational Emissions Inventory NH3Ammonia NO2Nitrogen dioxide NOxOxides of nitrogen NYCNew York City NYMTCNew York State Metropolitan Transportation Council PM2.5Fine particulate matter SCCSource classification code SESSocioeconomic status SO2sulfur dioxide STNSpeciation trends network UHFUnited Hospital Fund VMTvehicle miles traveled VOCVolatile organic compounds WRFWeather research forecasting model YLLYears of life expectancy lost We thank Sarah Johnson of NYC DOHMH for assisting in the spatial allocation of mobile source emissions and Kazue Anan of NYC DOHMH for help with figures. We also appreciate the assistance of Wenhui Li and Gil Maduro of the NYC DOHMH Office of Vital Statistics in developing the life tables used in this analysis. Funding This work was supported by New York City tax levy funds. Availability of data and materials Emissions data used in this analysis are publically available via https://www.epa.gov/air-emissions-inventories. Authors are not authorized to share baseline health data or traffic data, but can be accessed through formal requests to the New York City Department of Health’s Office of Vital Statistics, the New York State Statewide Planning and Research Cooperative System, and the New York Metropolitan Transportation Council. Air pollution monitoring data is publically available through https://www.epa.gov/aqs. Authors’ contributions IK contributed to study design, data analysis, and led drafting and editing the manuscript. JH and SD led air quality modeling analyses and contributed to study design, data analysis and editing the manuscript. KI contributed to study design, interpreting results and drafting and editing the manuscript. TM oversaw the method development and data analysis and contributed to drafting and editing the manuscript. All authors participated in interpretation of the results and 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. USEPA. Integrated science assessment for particulate matter. US Environmental Protection Agency. 2009. https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=216546. Accessed 9 Aug 2016. 2. Kheirbek I, Johnson S, Ito K, Matte T, Kass D, Caputo Jr S, Mahnovski S, CH S, Licata A, Eisl H, et al. New York City Trends in Air Pollution and its Health Consequences. New York City Department of Health and Mental Hygiene. 2013. https://www1.nyc.gov/assets/doh/downloads/pdf/environmental/air-quality-report-2013.pdf. Accessed 9 Aug 2016. 3. HEI. Traffic-Related Air Pollution: A Critical Review of the Literature on Emissions, Exposure, and Health Effects HEI Panel on the Health Effects of Traffic-Related Air Pollution. Health Effects Institute. 2010. http://pubs.healtheffects.org/getfile.php?u=553. Accessed 9 Aug 2016. 4. Brugge D Durant JL Rioux C Near-highway pollutants in motor vehicle exhaust: a review of epidemiologic evidence of cardiac and pulmonary health risks Environ Health 2007 6 23 10.1186/1476-069X-6-23 17688699 5. Lin S Munsie JP Hwang S-A Fitzgerald E Cayo MR Childhood asthma hospitalization and residential exposure to state route traffic Environ Res 2002 88 73 81 10.1006/enrs.2001.4303 11908931 6. Fann N Fulcher CM Baker K The recent and future health burden of air pollution apportioned across US sectors Environ Sci Technol 2013 47 3580 3589 10.1021/es304831q 23506413 7. Grabow ML Spak SN Holloway T Stone B Mednick AC Patz JA Air quality and exercise-related health benefits from reduced car travel in the midwestern United States Environ Health Perspect 2012 120 68 76 10.1289/ehp.1103440 22049372 8. Levy JI Buonocore JJ Von Stackelberg K Evaluation of the public health impacts of traffic congestion: a health risk assessment Environ Health 2010 9 65 10.1186/1476-069X-9-65 20979626 9. USEPA. Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards. US Environmental Protection Agency. 2012. https://www3.epa.gov/otaq/climate/documents/420r12016.pdf. Accessed 9 Aug 2016. 10. Ito K Xue N Thurston G Spatial variation of PM2.5 chemical species and source-apportioned mass concentrations in New York City Atmos Environ 2004 38 31 11. Lall R Thurston GD Identifying and quantifying transported vs. local sources of New York City fine particulate matter air pollution Atmos Environ 2006 40 333 346 10.1016/j.atmosenv.2006.04.068 12. Clougherty JE Kheirbek I Eisl HM Ross Z Pezeshki G Gorczynski JE Johnson S Markowitz S Kass D Matte T Intra-urban spatial variability in wintertime street-level concentrations of multiple combustion-related air pollutants: the New York City Community Air Survey (NYCCAS) J Expo Sci Environ Epidemiol 2013 23 3 10.1038/jes.2012.125 13. Cook R Isakov V Touma JS Benjey W Thurman J Kinnee E Ensley D Resolving local-scale emissions for modeling air quality near roadways J Air Waste Manag Assoc 2008 58 3 18236789 14. Kheirbek I Haney J Douglas S Ito K Caputo S Jr Matte T The public health benefits of reducing fine particulate matter through conversion to cleaner heating fuels in New York City Environ Sci Technol 2014 48 13573 13582 10.1021/es503587p 25365783 15. NYC. PlaNYC: Update April 2011. A Greener, Greater New York. The City of New York. 2011. http://www.nyc.gov/html/planyc/downloads/pdf/publications/planyc_2011_planyc_full_report.pdf. Accessed 9 Aug 2016. 16. Environment & Health Data Portal. New York City Department of Health. http://a816-dohbesp.nyc.gov/IndicatorPublic/PublicTracking.aspx. Accessed 9 Aug 2016. 17. NYC. One New York: The Plan for a Strong and Just City. The City of New York. 2015. http://www1.nyc.gov/html/onenyc/index.html. Accessed 9 Aug 2016. 18. ICF International. High-Resolution Air Quality Modeling of New York City to Assess the Effects of Changes in Fuels for Boilers and Power Generation. ICF International. 2014. www.nyccleanheat.org/sites/default/files/public/NYC_High_Resolution_AQM_Final_23June2014.pdf. Accessed 9 Aug 2016. 19. USEPA. 2008 National Emissions Inventory. United States Environmental Protection Agency. https://www.epa.gov/air-emissions-inventories/2008-national-emissions-inventory-nei-data. Accessed 9 Aug 2016. 20. USEPA. 2011 National Emissions Inventory. United States Environmental Protection Agency. https://www.epa.gov/air-emissions-inventories/2011-national-emissions-inventory-nei-data. Accessed 9 Aug 2016. 21. New York Metropolitan Transportation Council. Best Practices Model. 2005. https://www.nymtc.org/Data-and-Modeling/New-York-Best-Practice-Model-NYBPM. Accessed 9 Aug 2016. 22. USEPA. 2008 Modeling Platform. ftp://ftp.epa.gov/EmisInventory/2007v5. Accessed 9 Aug 2016. 23. USEPA. AP-42, Fifth Edition, Volume I. Chapter 1: External Combustion Sources. US Environmental Protection Agency. 2010. https://www3.epa.gov/ttn/chief/ap42/ch01/index.html. Accessed 9 Aug 2016. 24. New York State. New York Codes, Rules and Regulations Subpart 225-1, Fuel Composition and Use; Sulfur Limitations. New York State Department of Environmental Protection. 2012. https://govt.westlaw.com/nycrr/Browse/Home/NewYork/NewYorkCodesRulesandRegulations?guid=Ib74b1990b5a011dda0a4e17826ebc834&originationContext=documenttoc&transitionType=Default&contextData=(sc.Default). Accessed 9 Aug 2016. 25. NYC Clean Heat Program. http://www.nyccleanheat.org/. Accessed 9 Aug 2016. 26. Abt Associates. Modeled Attainment Test Software. Abt Associates Inc. 2010 https://www3.epa.gov/scram001/modelingapps_mats.htm. Accessed 9 Aug 2016. 27. Kheirbek I Wheeler K Walters S Kass D Matte T PM2.5 and ozone health impacts and disparities in New York City: sensitivity to spatial and temporal resolution Air Qual Atmos Health 2013 6 473 486 10.1007/s11869-012-0185-4 23710262 28. Hubbell B Fann N Levy JI Methodological considerations in developing local-scale health impact assessments: balancing national, regional, and local data Air Qual Atmos Health 2009 2 99 110 10.1007/s11869-009-0037-z 29. Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi Y, Turner MC, Pope III CA, Thurston G, Calle EE. Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality. Health Effects Institute. 2009. http://www.healtheffects.org/Pubs/RR140-Krewski.pdf. Accessed 9 Aug 2016. 30. Ito K Thurston GD Silverman RA Characterization of PM2.5, gaseous pollutants, and meteorological interactions in the context of time-series health effects models J Expo Sci Environ Epidemiol 2007 17 Suppl 2 45 60 10.1038/sj.jes.7500627 31. Moolgavkar SH Air pollution and hospital admissions for chronic obstructive pulmonary disease in three metropolitan areas in the United States Inhal Toxicol 2000 12 Suppl 4 75 90 10.1080/089583700750019512 12881887 32. Zanobetti A Franklin M Koutrakis P Schwartz J Fine particulate air pollution and its components in association with cause-specific emergency admissions Environ Health 2009 8 58 10.1186/1476-069X-8-58 20025755 33. Ito K Mathes R Ross Z Nádas A Thurston G Matte T Fine particulate matter constituents associated with cardiovascular hospitalizations and mortality in New York City Environ Health Perspect 2011 119 4 34. US EPA. Quantitative health risk assessment for particulate matter. US Environmental Protection Agency: Office of Air Quality Planning and Standards. 2010. https://www3.epa.gov/ttn/naaqs/standards/pm/data/PM_RA_FINAL_June_2010.pdf. Accessed 9 Aug 2016. 35. New York City Department of Health and Mental Hygiene. Epiquery: NYC Interactive Health Data System - NYCDOHMH neighborhood population estimates, modified from the US Census Bureau vintage population estimates. http://www1.nyc.gov/site/doh/data/health-tools/nyc-population.page. Accessed 9 Aug 2016. 36. USEPA. Environmental Benefits Mapping and Analysis Program (BenMAP) http://www.epa.gov/benmap. Accessed 9 Aug 2016. 37. Arias E. United States Life Tabels, 2010. US Department of Health and Human Services Centers for Disease Control and Prevention. 2010. http://www.cdc.gov/nchs/data/nvsr/nvsr63/nvsr63_07.pdf. Accessed 9 Aug 2016. 38. Grabow ML Spak SN Holloway T Stone B Jr Mednick AC Patz JA Air quality and exercise-related health benefits from reduced car travel in the midwestern United States Environ Health Perspect 2012 120 1 39. Friedman MS Powell KE Hutwagner L Graham LM Teague WG Impact of changes in transportation and commuting behaviors during the 1996 Summer Olympic Games in Atlanta on air quality and childhood asthma JAMA 2001 285 897 905 10.1001/jama.285.7.897 11180733 40. Hong A Schweitzer L Yang W Marr LC Impact of temporary freeway closure on regional air quality: a lesson from Carmageddon in Los Angeles, United States Environ Sci Technol 2015 49 3211 3218 10.1021/es505185c 25629730 41. Shmool JL Bobb JF Ito K Elston B Savitz DA Ross Z Matte TD Johnson S Dominici F Clougherty JE Area-level socioeconomic deprivation, nitrogen dioxide exposure, and term birth weight in New York City Environ Res 2015 142 624 632 10.1016/j.envres.2015.08.019 26318257 42. Hajat A Diez-Roux AV Adar SD Auchincloss AH Lovasi GS O’Neill MS Sheppard L Kaufman JD Air pollution and individual and neighborhood socioeconomic status: evidence from the Multi-Ethnic Study of Atherosclerosis (MESA) Environ Health Perspect 2013 121 1325 1333 24076625 43. Johansson C Burman L Forsberg B The effects of congestions tax on air quality and health Atmos Environ 2009 43 4843 4854 10.1016/j.atmosenv.2008.09.015 44. Jones AM Harrison RM Barratt B Fuller G A large reduction in airborne particle number concentrations at the time of the introduction of “sulphur free” diesel and the London Low Emission Zone Atmos Environ 2012 50 129 138 10.1016/j.atmosenv.2011.12.050 45. Cesaroni G Boogaard H Jonkers S Porta D Badaloni C Cattani G Forastiere F Hoek G Health benefits of traffic-related air pollution reduction in different socioeconomic groups: the effect of low-emission zoning in Rome Occup Environ Med 2012 69 133 139 10.1136/oem.2010.063750 21821870 46. Tonne C Beevers S Armstrong B Kelly F Wilkinson P Air pollution and mortality benefits of the London Congestion Charge: spatial and socioeconomic inequalities Occup Environ Med 2008 65 620 627 10.1136/oem.2007.036533 18308748 47. URS. Hunts Point Truck Study. URS/Goodkind & O’Dea, Inc. https://www.dot.ny.gov/regional-offices/region11/projects/project-repository/bese/pdf/hp_p1.pdf. Accessed 9 Aug 2016. 48. Schaller Consulting. CITYinFLUX: Understanding and Untangling Traffic and Transportation in New York City. Schaller Consulting. 2007. http://www.schallerconsult.com/pub/cityinflux.pdf. Accessed 9 Aug 2016. 49. Greco SL Wilson AM Hanna SR Levy JI Factors influencing mobile source particulate matter emissions-to-exposure relationships in the Boston urban area Environ Sci Technol 2007 41 7675 7682 10.1021/es062213f 18075073 50. Karner AA Eisinger DS Niemeier DA Near-roadway air quality: synthesizing the findings from real-world data Environ Sci Technol 2010 44 5334 5344 10.1021/es100008x 20560612 51. Delfino RJ Wu J Tjoa T Gullesserian SK Nickerson B Gillen DL Asthma morbidity and ambient air pollution: effect modification by residential traffic-related air pollution Epidemiol 2014 25 48 57 10.1097/EDE.0000000000000016 52. Thurston GD, Ito K, Lall R, Burnett RT, Turner MC, Krewski D, Shi Y, Jerrett M, Gapstur S, Driver R, et al. NPACT Study 4: Mortality and Long-Term Exposure to PM2.5 and its Components in the American Cancer Society’s Cancer Prevention Study II Cohort. In: National Particle Component Toxicity (NPACT) Initiative: Integrated Epidemiologic and Toxicologic Studies of the Health Effects of Particulate Matter Components. Health Effect Institute. 2013. http://pubs.healtheffects.org/getfile.php?u=934. Accessed 9 Aug 2016. 53. McConnell R Islam T Shankardass K Jerrett M Lurmann F Gilliland F Gauderman J Avol E Künzli N Yao L Childhood incident asthma and traffic-related air pollution at home and school Environ Health Perspect 2010 118 1021 1026 10.1289/ehp.0901232 20371422

PMC005xxxxxx/PMC5002107.txt

==== Front BMC Pregnancy ChildbirthBMC Pregnancy ChildbirthBMC Pregnancy and Childbirth1471-2393BioMed Central London 103510.1186/s12884-016-1035-4DebateA common monitoring framework for ending preventable maternal mortality, 2015–2030: phase I of a multi-step process Moran Allisyn C. amoran@usaid.gov 1Jolivet R. Rima rjolivet@hsph.harvard.edu 2Chou Doris choud@who.int 3Dalglish Sarah L. sarah.dalglish@gmail.com 3Hill Kathleen Kathleen.Hill@jhpiego.org 4Ramsey Kate kramsey@msh.org 5Rawlins Barbara Barbara.Rawlins@jhpiego.org 4Say Lale sayl@who.org 31 US Agency for International Development, Global Health Bureau, Washington, DC USA 2 Maternal Health Task Force, Boston, MA USA 3 World Health Organization, Geneva, Switzerland 4 Maternal and Child Survival Program, Jhpiego, Washington, DC USA 5 Averting Maternal Death and Disability Program, Columbia University, New York, NY USA 26 8 2016 26 8 2016 2016 16 1 2506 2 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 While global maternal mortality declined 44 % between 1990 and 2015, the majority of countries fell short of attaining Millennium Development Goal targets. The Sustainable Development Goals (SDGs), adopted in late 2015, include a target to reduce national maternal mortality ratios (MMR) to achieve a global average of 70 per 100,000 live births by 2030. A comprehensive paper outlining Strategies toward Ending Preventable Maternal Mortality (EPMM) was launched in February 2015 to support achievement of the SDG global targets. To date, there has not been consensus on a set of core metrics to track progress toward the overall global maternal mortality target, which has made it difficult to systematically monitor maternal health status and programs over time. Findings The World Health Organization (WHO), Maternal Health Taskforce (MHTF), and the US Agency for International Development (USAID) along with its flagship Maternal and Child Survival Program (MCSP), facilitated a consultative process to seek consensus on maternal health indicators for global monitoring and reporting by all countries. Consensus was reached on 12 indicators and four priority areas for further indicator development and testing. These indicators are being harmonized with the Every Newborn Action Plan core metrics for a joint global maternal newborn monitoring framework. Next steps include a similar process to agree upon indicators to monitor social, political and economic determinants of maternal health and survival highlighted in the EPMM strategies. Conclusion This process provides a foundation for the maternal health community to work collaboratively to track progress on core global indicators. It is important that actors continue to work together through transparent and participatory processes to track progress to end preventable maternal mortality and achieve the SDG maternal mortality targets. Electronic supplementary material The online version of this article (doi:10.1186/s12884-016-1035-4) contains supplementary material, which is available to authorized users. Keywords Maternal mortalityIndicatorsMonitoring and evaluationhttp://dx.doi.org/10.13039/100000200United States Agency for International DevelopmentAID-OAA-A-14-00028http://dx.doi.org/10.13039/100004423World Health OrganizationHuman Reproduction Programmeissue-copyright-statement© The Author(s) 2016 ==== Body Background In September 2015, the era of the Millennium Development Goals (MDGs) came to an end, and while the worldwide maternal mortality ratio (MMR) declined by 44 % from 1990 to 2015, the global target of a 75 % reduction was not reached. In fact, the majority of countries fell short of attaining MDG5 targets for maternal mortality reduction. Maternal mortality remains unacceptably high with approximately 303,000 maternal deaths occurring each year, with the largest burden in Sub-Saharan Africa and Asia [1]. Beginning in January 2013, the Ending Preventable Maternal Mortality (EPMM) Working Group, led by the World Health Organization (WHO) with support from partner organizations, achieved multi-stakeholder consensus on goals for maternal health and survival from 2015 to 2030. The EPMM targets for maternal mortality reduction at the global and country levels are:By 2030, the global average maternal mortality ratio (MMR) should be less than 70 maternal deaths per 100,000 live births By 2030, every country should reduce its maternal mortality ratio by at least two thirds from 2010 baseline, and no country should have a ratio higher than 140 deaths per 100,000 live births (twice the global target) All countries are tasked with achieving equity in MMR among sub-populations A comprehensive paper outlining the Strategies toward Ending Preventable Maternal Mortality (EPMM) was released by WHO in February 2015 [2]. This was followed by a synthesis of priorities for ending preventable maternal and newborn deaths and stillbirths in a series of technical papers in support of the updated United Nations Global Strategy for Women’s, Children’s, and Adolescent’s Health (2016–2030) [3]. The EPMM global MMR target was incorporated into the Sustainable Development Goals (SDGs) adopted by member states and launched in late 2015 [4]. The SDGs form the basis of a new global development agenda, which is broad and comprehensive, covering a wide range of social, economic and environmental goals. The EPMM global target was also included in the updated UN Global Strategy. To achieve the ambitious SDG and Global Strategy MMR targets, it will be imperative to accelerate coverage of quality essential services and to address the underlying social, political and economic determinants of maternal health across all settings. To support implementation of EPMM strategies and to track progress toward global and country MMR targets, consensus on priority, methodologically robust maternal health (MH) indicators is urgently needed. To this end, in May 2015, the EPMM Working Group specified a plan to develop a comprehensive maternal health global monitoring and reporting framework in two phases:Phase 1: Consensus on a core set of priority, methodologically robust maternal health indicators with direct relevance for reducing preventable mortality (proximal to causes of death) for global monitoring and reporting by all countries Phase 2: Consensus on complementary indicators to track progress toward addressing the social, political and economic determinants of maternal health and survival (distal to causes of death) highlighted in the global EPMM strategy for testing by countries as part of national health plans and monitoring frameworks This paper describes the process and results of Phase I – to develop a core set of maternal health indicators for global monitoring and reporting by all countries. This core set of maternal health indicators will be harmonized with Every Newborn Action Plan (ENAP) core metrics to create a joint global maternal and newborn monitoring framework [5]. Methods The World Health Organization, Maternal Health Taskforce (MHTF), and the US Agency for International Development (USAID) along with its flagship Maternal and Child Survival Program (MCSP), facilitated an iterative consultative process to reach consensus on a core set of maternal health indicators for global monitoring and reporting under Phase I on behalf of the EPMM Working Group. Key phases of the process led by an EPMM-designated steering committee included: (1) structured mapping exercise to identify indicators for monitoring and evaluation of maternal health; (2) development of key principles, criteria, and consultative process for selection of EPMM phase I indicators; (3) selection of experts/stakeholders to participate in consultative process in line with agreed criteria; (4) facilitation of a series of virtual consultations via webinar, phone and in writing to rank and prioritize indicators based on a set of pre-determined questions and criteria; and (5) an in-person one-day technical meeting with identified experts/stakeholders to reach consensus on a core set of MH indicators for global monitoring (Fig. 1).Fig. 1 Process for Consensus Building, August-October, 2015 The EPMM-designated steering committee included six persons selected to represent a range of expertise, including technical, policy and measurement experts, who were tasked with overseeing the process of defining a core set of global maternal health indicators (Table 1). The Steering Committee identified four principles to guide development and prioritization of a set of core MH indicators for global monitoring and reporting. First, the scope was limited to priority indicators appropriate for global monitoring and reporting by all countries. It was agreed that the primary purpose of the indicators would be to track progress toward maternal mortality reduction at the global level, with no more than 10 to 12 indicators in total. It was acknowledged that complementary (and sometimes overlapping) indicators would be needed to support effective program management at sub-national and service delivery levels but that these indicators were beyond the scope of the EPMM global indicator exercise. During the process, indicators for other levels of monitoring and reporting were systematically captured and recorded for future research and development, building on numerous ongoing indicator development efforts. A second guiding principle was to limit Phase I indicators to measures directly linked to the proximal causes of maternal death to mirror the core set of newborn health metrics outlined in the Every Newborn Action Plan. More distal determinants, such as social, economic, political, and health systems factors, will be developed under Phase II of the EPMM process.Table 1 Steering Committee Membersa Name Organization Job title Training, area of expertise Allisyn Moran, PhD MHS US Agency for International Development Senior Maternal Health Advisor Monitoring and evaluation, research, maternal and child health R. Rima Jolivet, CNM, DrPH Maternal Health Task Force Maternal Health Technical Director Certified nurse-midwife, public health, maternal health system strengthening, quality improvement Kathleen Hill, MD Maternal and Child Survival Program, Jhpiego Team Lead, Maternal Health Family physician, public health, service delivery, program implementation, quality improvement Barbara Rawlins, MPH Maternal and Child Survival Program, Jhpiego Team Lead, Monitoring and Evaluation Monitoring and evaluation, research, maternal and child health Lale Say, MD World Health Organization, Department of Reproductive Health and Research Coordinator, adolescents and at-risk populations Reproductive Health and Research Physician, monitoring and evaluation, research, program strengthening Sarah Dalglish, PhD World Health Organization, Department of Reproductive Health and Research Consultant Political Economy, International Relations, Health Politics and Policy Kate Ramsey, MPH, DrPH (c) (position at the time of this project) Averting Maternal Death and Disability Program, University of Columbia Senior Research Officer Public health, maternal and newborn health service delivery, health systems, implementation research aNo financial conflicts declared by any members of the Steering Committee The third agreed principle was to use the same organizational categories of these ENAP metrics, to facilitate a harmonized global maternal and newborn monitoring framework. Thus it was agreed to retain the ENAP metrics organizational categories of impact, coverage and input indicators. Finally, it was assumed that indicators would be reported at least every 5 years to track progress toward global targets. In parallel, the steering committee finalized indicator selection criteria, drawing from criteria used in other indicator prioritization activities (e.g. WHO 100 core indicators [6]). Selection criteria included: (1) relevance for directly reducing preventable mortality; (2) validity (measures what it is supposed to measure); (3) feasibility (feasible to measure routinely through household surveys, health facility assessments, routine information systems); (4) degree of data availability across countries; and (5) strengthens or complements existing monitoring frameworks (MDGs, SDGs; ENAP; and other global efforts (e.g. Global Strategy for Women's, Children's, and Adolescent's Health; WHO Quality of Care Initiative, etc.)). The WHO indicator mapping exercise (phase one above) included searching online databases, reviewing program documents and other global monitoring frameworks, and discussions with implementing partners and donor agencies. An initial expert consultation was conducted in May 2015. Applying these principles and selection criteria, the Steering Committee compiled the indicators prioritized during this consultation, as well as priority indicators for tracking SDGs, and other indicator mapping efforts. This list was refined and updated, which resulted in a final list of 20 to 25 indicators. In the second phase a series of virtual consultations was conducted with selected expert participants. Key stakeholders and experts were identified through review of participant lists from other indicator/measurement consultations, discussions with the broader EPMM Working Group, and other relevant stakeholders. The focus was on identifying participants who work directly in measurement of maternal health including clinicians, researchers, program implementers, global policy makers and representatives from Ministries of Health, UN agencies, and donors. (Table 2). Based on the structured indicator review and input from measurement experts, the Steering Committee developed key questions related to the underlying principles and indicator selection criteria to guide discussion and feedback during a series of virtual consultations. In each virtual technical review session, the indicators were discussed, guided by the key questions. Feedback from previous sessions was also shared, so learnings were incorporated and refined during successive phases. Written feedback was solicited from those unable to participate in planned consultations.Table 2 List of participant organizations Organization Department, Location Specialty Phase of process EPMM core group Steering committee Virtual 1 Virtual 2 In-person meeting Written comments Invited, unable to participate UN agencies  UNFPA MN Technical Division Programs, M&E, Clinical X X X X  UNICEF Division of Data, Research and Policy Epidemiology, Research  World Health Organization MCA, RHR Epidemiology, M&E, Clinical X X X X X Programs  Averting Maternal Death and Disability Program (AMDD)a Columbia University SPH Research, Programs, M&E X X X X  Family Care Internationalb Advocacy, Programs X X  ICF Macroc Demographic and Health Surveys (DHS) MH Measurement X  Improving Coverage Measurementd JHU, Institute for International Programs Epidemiology, Research X X  Maternal and Child Survival Programe Jhpiego M&E, Research, Program, Policy, Clinical X X X X X  Maternal Health Task Forcef Harvard, SPH , Advocacy, Clinical X X X X X  MEASURE Evaluationg University of North Carolina, Chapel Hill M&E, Programs X X  Population Councilh Evaluation and Research, Programs X  White Ribbon Alliancei Advocacy X Universities, Research Institutes  icddr,b Dhaka, Bangladesh Research, Surveillance, Programs, Clinical X X  Ifakara Health Institute Dar es Salaam, Tanzania Research, Surveillance, Programs, Clinical X  Johns Hopkins University Baltimore, MD, USA Research X  London School of Hygiene and Tropical Medicine London, UK Research X  Makerere University Kampala, Uganda Research, Surveillance, Programs, Clinical X  University of Aberdeen Scotland Research X  University of Heidelberg Research Programs X X  University of Ouagadougou Ouagadougou, Burkina Faso Research, Programs, Clinical X  University of Southampton Southampton, UK Research X X Ministries of Heatlh  Department of Health, South Africa Policy, Programs, M&E, Clinical X  Ministry of Health, Kenya Policy, Programs, M&E, Clinical X  Ministry of Health, Nigeria Policy, Programs, M&E, Clinical X Consultants  Ethiopia Research, M&E X  Ghana Clinical, Programs, Research X  Kazakhstan Clinical, Research maternal death and response X Donors  Bill & Melinda Gates Foundation Seattle, WA USA Clinical, M&E, Research X X  Children’s Investment Fund Foundation London, UK M&E, Research X  US Agency for International Development Washington, DC USA Programs, M&E, Research, Clinical X X X X X Note: The “Clinical” specialty refers to OB/GYN, general practitioners, midwives, and other medical professions; M&E refers to monitoring and evaluation aAMDD: https://www.mailman.columbia.edu/research/averting-maternal-death-and-disability-amdd bFCI: http://www.familycareintl.org/en/home cDHS: http://www.dhsprogram.com/ dICM: http://www.jhsph.edu/research/centers-and-institutes/institute-for-international-programs/current-projects/improving-coverage-measurements-for-mnch/ eMCSP: http://www.mcsprogram.org/ fMHTF: https://www.mhtf.org/ gMEASURE: http://www.cpc.unc.edu/measure/ hPopCouncil: http://www.popcouncil.org/ iWRA: http://whiteribbonalliance.org/ To culminate the consultation process, an in-person meeting was held in Washington, DC in mid-September 2015, bringing together a broader group of experts. The meeting was convened by the Maternal and Child Survival Program with financial support from the US Agency for International Development (USAID). Some participants were supported through funding from other institutions and development partners. This in-person consultation was guided by the same key questions addressed in previous virtual session. During the in-person meeting, participants considered the results from the first rounds of feedback to deliberate and reach consensus on a final core set of maternal health indicators for global monitoring and reporting. The in-person one-day consultation also addressed commonly perceived gaps in maternal health indicators, proposals for additional indicators and other issues raised by participants. Inputs from all steps of the consultation (virtual, written and in-person) were compiled in a matrix summarizing feedback and main discussion points (Additional file 1). A report of the in-person consultation highlighting the main discussion points was also developed. Results In total, 45 experts from 11 countries participated in one or more stages of the structured consultative process to reach consensus on a core set of maternal health indicators for global monitoring and reporting. The experts agreed on 12 maternal health indicators - three impact, seven coverage, and two input indicators (Table 3). Five indicators overlap with the Every Newborn Action Plan core metrics, six indicators overlap with the WHO list of 100 core indicators, and one indicator overlaps with the WHO 2013 Quality of MNCH care indicators [5–7] (Table 3; Additional file 2). The definitions, disaggregators, and data sources to accompany the Core Maternal Health Indicators for Global Monitoring and Reporting have been finalized (Table 4). Four areas were prioritized for further research and testing, namely content of antenatal care, content of postpartum care, respectful maternity care, and Met Need for Emergency Obstetric Care (EmOC).Table 3 Core Maternal Health Indicators for Global Monitoring and Reporting Indicator Priority areas for indicator development Impact 1. Maternal mortality ratioa c 2. Maternal cause of death (direct/indirect) based on ICD-MM 3. Adolescent birth ratec COVERAGE: care for all women and girls 4. Four or more antenatal care visitsc Content of antenatal careb 5. Skilled attendant at birtha c Content of postnatal care 6. Institutional delivery Respectful maternity careb 7. Early postnatal/postpartum care for woman and baby (within 2 days of birth)a c 8. Met need for family planningc 9. Uterotonic immediately after birth for prevention of post-partum hemorrhage (among facility births)b COVERAGE: care for women and girls with complications 10. Caesarean section ratea Met need for Emergency Obstetric Care INPUTS: counting 11. Maternal death registration INPUT: Availability of care 12. Availability of functional Emergency Obstetric Care facilitiesa aENAP indicator bLink to WHO Quality of Care metrics cLink to WHO 100 Core indicators NOTES: - WHO will propose a definition for maternal death registration - Availability of functional emergency obstetric care facilities requires additional definition - The current definition will be used in the short term, with ongoing efforts to improve definition (both numerator and denominator) - Countries should continue to monitor met need for emergency obstetric are and update definitions once they are finalized based on ongoing work - Additional priority indicators – Efforts will link with ongoing efforts such as WHO Antenatal Care Guideline revision process, WHO Quality of Care Initiative, Global Strategy for Women’s, Children’s, and Adolescent Health - Currently, antenatal and postnatal care are measured through the number of contacts with the health system (as defined as “visits”), but the quality or content of these visits are not assessed. Two of the priority areas moving forward include defining core content for these antenatal and postnatal visits as well as measures to assess those the core content areas - Content of antenatal care could include: blood pressure, testing and treatment of infectious disease, counseling on danger signs, testing for HIV/AIDS, prevention of malaria during pregnancy, birth planning, etc. Content of postpartum care could include: monitoring bleeding, counseling for family planning, observing breastfeeding, counseling and assessment of postpartum depression, etc Table 4 Core Maternal health indicators for global monitoring and reporting – definitions and data sources Indicator Definition Numerator Denominator Disaggregation Data Source(s) Impact Maternal Mortality Ratioa c Death from any cause related to or aggravated by pregnancy or its management (excluding accidental or incidental causes) during pregnancy, childbirth or within 42 days of termination of pregnancy, per 100,000 live births for a specified time period Number of maternal deaths Per 100,000 live births in the same period Sub-national HH surveys, CRVS, admin, modeled estimates. RAMOS, Confidential Inquiries, Census Maternal cause of death (proportion) Deaths from any cause related or aggravated by pregnancy or its management (excluding accidental or incidental causes) during pregnancy, childbirth or within 42 days of termination of pregnancy for a specified time period (using ICD-MM) Number of maternal deaths by cause Total number of maternal deaths in the same period Cause, Indirect/Direct HH surveys, CRVS, admin, modeled estimates, RAMOS, Confidential Inquiries Adolescent birth ratec Number of births to women 15 to 19 years of age per 1,000 women within specified time period Number of births to women 15 to 19 years of age Per 1,000 women 15 to 19 years of age in the same period Sub-national HH surveys, census, CRVS COVERAGE: care for all women and girls Four or more ANC visitsc Proportion of pregnant women and girls who have made at least four antenatal care visits within specified time period Number of women and girls who received ANC at least four times during pregnancy Total number of women and girls with a live birth in the same period Wealth quintile Residence Age Sub-national HH surveys Skilled attendant at birtha c Proportion of births attended by skilled health personnel (doctor, nurse, midwife) within a specified time period Number of live births attended by skilled health personnel Total number of live births in the same period Wealth quintile Residence Age Sub-national HH surveys Institutional delivery Proportion of births in a health facility within a specified time period Number of live births in a health facility Total number of live births in the same period Wealth quintile Residence Age Sub-national HH surveys Oxytocin immediately after birth for prevention of post-partum hemorrhageb Proportion of women and girls who gave birth in a facility receiving oxytocin immediately after birth within a specified time period Number of women and girls who gave birth in a facility who received oxytocin immediately after birth Total number women and girls with a facility birth in same period Sub-national Age HMIS, Facility records Early postnatal/postpartum care for women and babies (within 2 days of birth)a c Proportion of women/girls with a recent birth and their babies who received postnatal care within two days of birth (regardless of place of delivery) within a specified time period Number of women/girls and their babies who received postnatal care within two days of childbirth Total number of women/girls with a last live birth in the same period Wealth quintile Residence Age Sub-national HH surveys Met need for family planningc Proportion of women and girls, either married or in a union, who have their need for contraception satisfied within a specified time period Number of women and girls who have their need for contraception satisfied Total number of women and girls, married or in union, in need of contraception in same period Wealth quintile Residence Age Sub-national Postpartum/non-postpartum HH surveys COVERAGE: care for women and girls with complications Caesarean section ratea Proportion of women and girls with a live birth delivered by caesarean section within a specified time period Numbers deliveries by caesarean section Total number of live births to women and girls in same period Wealth quintile Residence Age Sub-national HH surveys INPUT: counting Maternal death registration, including cause of deatha Proportion of maternal deaths registered with cause of death specified based on ICD-MM codes within a specified time period Number of maternal deaths registered with cause of death specified based on ICD-MM codes Total estimated number of maternal deaths in the same time period Sub-national Health facilities, CRVS, census INPUT: Availability of care Availability of functional EmOC facilitiesa (per population) At least five emergency obstetric care facilities (per 500,000 people) including at least one comprehensive and the rest basic emergency obstetric care facilities. Number of obstetric care facilities that provided EmOC signal functionsd in the last three months Per 500,000 population Sub-national, Facility level HF surveys, routine facility monitoring, census or other population data source aENAP indicator bLink to WHO Quality of Care metrics cLink to WHO 100 Core Indicators dSignal functions; Basic: 1) parenteral antibiotics; 2) uterotonic drugs; 3) parenteral anticonvulsants for preeclampsia and eclampsia; 4) manual removal of placenta; 5) remove retained products (e.g. manual vacuum extraction, dilation and curettage); 6) perform assisted vaginal delivery (e.g. vacuum extraction, forceps delivery); and 7) basic neonatal resuscitation (e.g., with bag and mask); Comprehensive: All seven basic plus: 8) perform surgery (e.g., caesarean section); and 9) perform blood transfusion The group was unanimous in its consensus to deliver this core set of maternal health indicators to WHO for further member state consultation and deliberation through global processes, including harmonization with core metrics from the ENAP and consideration as part of a combined monitoring framework for maternal and newborn health within the Indicator Framework of the Global Strategy for Women’s, Children’s and Adolescents’ Health (2016–2030). Consensus was reached through extensive discussion of all proposed indicators (see Table 5). The final core list of indicators focuses primarily on routine care for all women; however, ending preventable maternal death requires appropriate management of maternal complications. The group discussed two potential composite indicators related to essential services for obstetric complications- Unmet Obstetric Need (UON: the proportion of women who receive surgery/major obstetric intervention among all women who experience a severe complication that is an indication for surgery/major obstetric intervention) and Met Need for Emergency Obstetric Care (the proportion of women who are expected to experience a complication who receive treatment) [8, 9]. Neither indicator met all the selection criteria for inclusion, due to issues related to feasibility and validity. For example, Met Need for EmOC assumes 15 % of women will experience a complication; this denominator has not been validated and may vary by setting. UON assumes 1 to 2 % of women will experience a severe complication, which has been validated in several settings. However the UON numerator is based on obstetric complications requiring surgical interventions which can be variable even for a single complication (e.g. pre-eclampsia/eclampsia (PEE) or postpartum hemorrhage (PPH) can often be managed without surgical intervention.) [9]. The group agreed that Met Need for EmOC was more appropriate for global monitoring, but that further refinement of the denominator would be needed for inclusion in a global monitoring framework. It was agreed that countries that are currently collecting Met Need for EmOC should continue to monitor this indicator at national and sub-national levels until the indicator has been updated.Table 5 Maternal Health Indicators considered for global monitoring and reporting, by indicator selection criteria Indicator Criteria (√ = YES, ~ = Some, X = No) Notes Relevance Validity Feasibility Availability Complements MH Monitoring FWa Impact  Maternal mortality ratio √ √ ~ ~ √  Maternal cause of death √ ~ ~ ~ √ - Global Strategy indicator – need to strengthen information systems to routinely collect   Case fatality rate √ ~ ~ ~ X - More feasible to collect at sub-national or service delivery levels  Adolescent birth rate √ √ √ √ √ COVERAGE: care for all women and girls  4 or more antenatal care visits √ √ √ √ √  Blood pressure screening during antenatal care √ √ √ √ X - May focus ANC on only BP screening  Full course of iron/folate during pregnancy √ X √ √ X - May focus ANC on only iron folate  Skilled attendant at birth √ ~ √ √ √ - SDG indicator  Institutional delivery √ √ √ √ X - Complements skilled attendant at birth  Oxytocin within 1 minute of birth (facility births) √ ~ ~ ~ X - Indicates quality of delivery care - Feasible to strengthen routine health information systems to collect - Timing is challenging - Change wording to “uterotonic immediately after birth”  Companion at birth √ ~ X X X - Respectful maternity care important, but this is only one element - May focus respectful maternity care only on companion at birth - Not feasible for many high-volume facilities - Not included in large-scale household surveys  Early postpartum care for woman √ ~ √ √ √ - Postnatal/postpartum care within 2 days of birth, regardless of place of delivery  Met Need for family planning √ √ √ √ √ COVERAGE: Care for women and girls with complications  Caesarean section rate among poor (bottom two quintiles) √ √ ~ ~ X  Caesarean section rate √ √ √ √ √  Met need for EmOC √ X X X X  UnMet need for EmOC √ X X X X INPUT: counting  Birth registration X ~ ~ ~ √ - Not relevant for maternal mortality – included in ENAP core metrics  Death registration, including cause of death √ ~ ~ ~ √ INPUT: Availability of care  Availability of EmOC per 500,000 population √ ~ ~ ~ X - Essential to include an indicator on emergency obstetric care – indicator to be refined and updated Indicators in italics NOT included in final core list NOTE: √ = YES; ~ = Some; X = No aSDGs, WHO 100 Core indicators, WHO Quality of Care metrics, Every Newborn Action Plan (ENAP) There was widespread agreement on the need to refine the skilled birth attendant (SBA) indicator to increase its validity. WHO and UNICEF are working in this area to review and revise the current metadata and enhance measurement (personal communication, Chou and Amouzou), and the indicator met other inclusion criteria. Therefore, SBA was included in the core MH indicator list, with institutional delivery added as a supplemental indicator. There were also important discussions about how to interpret caesarean section rates, given the global ongoing debates about the appropriate population-based proportion and rising rates of voluntary caesarean sections. It was agreed that it will be particularly important for this indicator to be disaggregated by residence and socio-economic quintile, and increasingly to capture indications [10]. The group discussed a measure to indicate the availability of functional EmOC facilities. Some participants felt this indicator needed further refinement prior to inclusion, especially to re-define the denominator to be based on expected pregnancies/births instead of population. Ultimately, the group agreed that this indicator met the criteria for global monitoring and that the definition would be updated moving forward. Particular areas of contention included whether measures of facility-level quality of care and coverage are useful and feasible for global monitoring and reporting (e.g. measures of specific content of antenatal, labor and delivery, and postpartum care). While many participants agreed that such measures are critically needed for program improvement and effectiveness at country level, there was a lack of consensus on whether or not quality measures should be included in a global monitoring framework as well as active debate about which indicator would adequately capture all elements of care. A single indicator for routine labor and delivery care—provision of uterotonic immediately after birth—was selected. In addition, the group strongly agreed on the importance of a high-level indicator of provision of respectful maternity care (RMC) and elimination of mistreatment in childbirth [11]. There was agreement that such metrics need further development and testing prior to inclusion for global monitoring and reporting. The development of indicators related to quality of antenatal care, postnatal care, RMC and Met Need for EmOC were flagged as an important priority for further development and testing. The group acknowledged ongoing work by WHO to define quality of care indicators for maternal and newborn care in health facilities as part of the WHO Quality of Care framework. Discussion “What gets measured gets done.” This simple phrase captures the synergistic relationship between measurement and action [12], which is critical for achieving ambitious targets outlined in EPMM Strategy documents, the SDGs, and the UN Global Strategy for Women’s, Children’s, and Adolescent’s Health. This paper describes the results and process of defining core maternal health indicators for global monitoring and reporting as part of the EPMM strategy. Agreement on a core set of global maternal health indicators is important for galvanizing action and accountability to reduce preventable maternal mortality as well as to highlight technical areas needing further development and indicator testing. WHO will propose the agreed indicators for further member state consultation through global processes, including harmonization with Every Newborn Action Plan core metrics into a combined global monitoring framework for maternal and newborn health. All participants pledged their support for these processes. The MDG5 goals and targets raised the profile of global maternal mortality reduction, galvanizing political will and resource allocation, which contributed to accelerated progress in mortality reduction. Between 1990 and 2000, MMR declined at an estimated average annual rate of 1.2 % globally, compared with an average of 3.0 % between 2005 and 2015 after the launch of the Millennium Development Goals in 2000 [1]. But accelerating further declines and ensuring equity will now require different strategies and concerted efforts to reach vulnerable and at-risk populations. A common global monitoring framework that includes key coverage and outcome indicators will be crucial to monitor whether strategies are working as intended to track progress toward global targets. The core maternal health indicators are the result of a systematic sequential process to prioritize a small number of impact, coverage and input measures for global monitoring. These global indicators are not intended to address the measurement needs and priorities for improvement at all levels of the system across all contexts. Other complementary efforts are underway to address maternal and newborn health measurement needs related to processes and outcomes at national, subnational and service delivery levels. The core global level indicators complement these other measurement efforts and contribute to policy decision-making, resource mobilization, program planning, and aim to accountability for achieving preventable maternal deaths across the globe. There has been promising progress in defining, testing and using indicators to monitor maternal health programs at subnational and service-delivery levels [13–15]. The global indicator on availability of functional EmOC facilities will build on this existing body of work to incorporate elements of basic maternity care, newborn care, human resources, travel time, and others. In addition, WHO is leading an initiative to define standards and metrics for quality of facility-based care around the time of birth. The WHO quality of care framework and associated metrics prioritize measures for use by managers and providers to drive improvements in maternal and newborn care at the point of service [16]. In collaboration with the ENAP metrics improvement plan, work is underway to validate several facility-based indicators, including the Phase I EPMM indicator uterotonic use immediately after birth. This validation work will include assessing whether the woman was aware that a uterotonic was given after birth to prevent postpartum hemorrhage. There are also many efforts underway to develop and test innovative methods to generate higher quality data to more accurately track facility-level quality of care metrics routinely. Innovative approaches around combining population-based surveys and health facility surveys are in progress [17, 18] as well as unique methods for routine monitoring of signal functions for obstetric and newborn care through quarterly monitoring visits (personal communication, UNFPA). Furthermore, there are efforts underway to use geographic information systems (GIS) to capture access to care and visualize data at sub-national levels to improve decision making [19]. As these methods are tested and refined, they will greatly contribute to more robust and accurate data to track program implementation and outcomes, which can be utilized at sub-national and service delivery levels to improve quality of care, and ultimately feed into global level monitoring and reporting. To allow for evolving improvements in measurement strategies and data quality, the global monitoring framework must be dynamic and flexible to incorporate new developments and indicators. There were some limitations. First, because the goal was to reach expert consensus on technically sound indicators for monitoring progress toward global maternal health goals, the timeframe was dictated by time-bound global political processes, including the launch of the SDGs and the Global Strategy. As a result, it was not possible to conduct a lengthy process utilized in other indicator prioritization exercises or research studies. However, every attempt was made to follow a systematic and transparent process, including a series of virtual consultations with experts from diverse backgrounds based on pre-determined principles, indicator selection criteria, and standard questions, which culminated with an in-person meeting. This process provided a transparent mechanism for consensus-building, within the time constraints. In total, over 45 stakeholders participated in the process; however, there were limited participants from the country level. WHO will conduct further consultations with member states through global processes, and the core MH indicators from this process have been introduced into consultations regarding indicator selection for the UN Global Strategy for Women’s, Children’s and Adolescent’s Health. Second, to meet the aim of a core set of indicators acceptable for global monitoring and reporting by all countries, the issues of feasibility, such as current data availability and quality, were by necessity balanced against the desire to “push the envelope” and drive accelerated health information system improvements. The group attempted to deal with this tension openly in the discussions and some indicators that will require further refinement and testing were included for this reason while others were not. For example, maternal cause of death is included as a core indicator, although further work is needed to standardize diagnosis, coding, and promote recording of all deaths. Finally, monitoring the core set of indicators for maternal health cannot be implemented without attention to data systems and efforts to ensure adequate data quality. We must support countries in developing and implementing plans to routinely collect these indicators through existing data systems and surveys, such as national health management information systems, household survey programs such as Demographic and Health Surveys (DHS) or Multiple Indicator Cluster Surveys (MICS), and health facility assessments such as the Service Availability and Readiness Assessment (SARA), the Service Provision Assessment (SPA), and not create parallel structures. The recently launched Health Data Collaborative is working to coordinate investments and build capacity at country level to collect indicators required for SDG monitoring. The global Strategies toward Ending Preventable Maternal Mortality [2] highlights an overall call to improve metrics, measurement systems and data quality. Representing one among several essential components of that goal, a core set of global indicators is key to galvanizing action and prioritizing attention to maternal health and survival, driving investment in the improvement of services, delivery systems and other proximal determinants, and providing a mechanism for accountability to stakeholders. The next phase of this process aims to develop consensus on measures to address the more distal causes of maternal survival and health, drawing from the key objectives in EPMM strategic framework that focus on the underlying social, political, and economic determinants of maternal health and survival, providing a pathway for countries to operationalize the EPMM strategic objectives and measure change over time. The SDGs are intended to realize the rights of all and to achieve gender equality and the empowerment of women and girls, balancing the dimensions of sustainable development: the economic, social, and environmental [20]. The UN Secretary General’s Global Strategy for Women’s, Children’s, Adolescent’s Health addresses health across sectors and SDGs to move beyond the focus on survival to thriving and transformation. The Indicator and Monitoring Framework for the Global Strategy describes 16 key indicators considered the minimal subset of a total of 60 indicators to provide assessment of progress towards the implementation of the Global Strategy objectives [21]. The EPMM Phase I indicators maternal mortality ratio and adolescent birth rate are included within this minimum dataset, however expanded monitoring of outcomes related to the key themes of EPMM will help to fully realize and address priority issues in the SDGs, including equity, human resources, gender, and human rights. Taken together, the development of the EPMM Phase I and II indicators will track progress and address a comprehensive set of issues related to not only “survival” but also “thrive and transform”; together they track progress. Conclusion Within the global maternal health community, there has been a lack of consensus on core metrics for monitoring maternal health status, health system performance and program implementation necessary to achieve EPMM targets. This is largely due to the complex nature of maternal health programs that must address a wide range of health determinants and deliver a package of services across a continuum within the context of differing political priorities, health system structures, and in coordination with non-health sector actors. However, the lack of robust measures, functioning measurement systems, and data of reliable quality have stymied efforts to accurately track whether countries are meeting global health targets. It is important that members of the global and maternal health communities work together through participatory processes to track progress at all levels (global, regional, national, subnational, and facility) to end preventable maternal mortality. This paper describes a core set of maternal health indicators for global monitoring and reporting for all countries, developed through an inclusive, iterative, and consultative process led by WHO on behalf of the EPMM Working Group. Monitoring these core indicators will be critical to support achievement of the global targets outlined in the SDGs and ultimately to end preventable maternal mortality. Additional files Additional file 1: Summary of feedback from stakeholders on proposed EPMM Phase 1 indicators. (XLSX 69 kb) Additional file 2: Core Maternal Health Indicators for Global Monitoring and Reporting and linkages to Every Newborn Action Plan (ENAP) indicators. (DOCX 23 kb) Abbreviations AMDDAverting maternal death and disability program EmOCEmergency obstetric care ENAPEvery newborn action plan EPMMEnding preventable maternal mortality MCSPMaternal and child survival program MDGMillennium development goal MHMaternal health MHTFMaternal health task force MMRMaternal mortality ratio PEEPre-eclampsia/eclampsia PPHPostpartum hemorrhage RMCRespectful maternity care SBASkilled birth attendant SDGSustainable development goals USAIDUnited States agency for international development WHOWorld health organization Acknowledgements The authors would like to acknowledge Ann Beth Moller (WHO) for assistance with the indicator mapping exercise, as well as all participants in the webinars, consultations or provided textual comments throughout the process. We also acknowledge and thank Mary Ellen Stanton (USAID) and Ana Langer (MHTF) for their review and comments which greatly strengthened the manuscript. Funding This work was supported by partners in the EPMM Working Group, including the US Agency for International Development, WHO/HRP, and the Bill & Melinda Gates Foundation through support to the Maternal Health Task Force. Authors’ contributions All authors contributed to the concept and process of the indicator finalization. ACM and RRJ drafted the manuscript with inputs from DC, SLD, KH, KR, BR, and LS. 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. Declarations The authors alone are responsible for the views expressed in this manuscript and they do not necessarily represent the views, decisions or policies of the institutions with which they are affiliated. ==== Refs References 1. WHO, UNICEF, UNFPA, World Bank Group and United Nations Population Division Trends in Maternal Mortality: 1990 to 2015 2015 Geneva World Health Organization 2. WHO Strategies toward Ending Preventable Maternal Mortality 2015 3. Chou D Ending preventable maternal and newborn mortality and stillbirths BMJ 2015 351 h4255 10.1136/bmj.h4255 26371222 4. https://sustainabledevelopment.un.org/topics accessed 20 Nov 2015. 5. Moxon SG Ruysen H Kerber KJ Amouzou A Fournier S Grove J Moran AC Vaz LM Blencowe H Conroy N Gülmezoglu A Vogel JP Rawlins B Sayed R Hill K Vivio D Qazi SA Sitrin D Seale AC Wall S Jacobs T Ruiz Peláez J Guenther T Coffey PS Dawson P Marchant T Waiswa P Deorari A Enweronu-Laryea C Arifeen S Lee AC Mathai M Lawn JE Count every newborn: a measurement improvement roadmap for coverage data BMC Pregnancy Childbirth 2015 15 Suppl 2 S8 10.1186/1471-2393-15-S2-S8 26391444 6. World Health Organization Global Reference List of 100 Core Health Indicators 2015 Geneva World Health Organization 7. World Health Organization and Partnership for Maternal, Newborn and Child Health Consultation on Improving measurement of the quality of maternal, newborn and child care 2014 Geneva World Health Organization 8. World Health Organization, UNFPA, UNICEF, and AMDD Monitoring emergency obstetric care a handbook 2009 Geneva World Health Organization 9. Unmet Need for Major Obstetric Interventions Network. Tackling Unmet Need for Major Obstetric Interventions. Accessed at: http://uonn.org/en/home on 28 Apr 2016 10. World Health Organization. WHO Statement of Caesarean Section Geneva 2015 Switzerland World Health Organization 11. Vogel JP Bohren MA Tuncalp O Oladapo OT Adanu RM Balde MD Maung TM Fawole B Abu-Bonsaffoh K Dako-Gyeke P Maya ET Camara MC Diallo AB Diallo S Wai KT Myint T Olutayo L Titiloye M Alu F Idris H Gulmezoglu MA WHO Research Group on the Treatment of Women During Childbirth How women are treated during facility-based childbirth: development and validation and measurement tools in four countries – phase 1 formative research study protocol Reprod Health 2015 12 60 10.1186/s12978-015-0047-2 26198988 12. World Health Organization WHO guidance for measuring maternal mortality from a census 2013 Geneva World Health Organization 13. Collender G Gabrysch S Campbell OM Reducing maternal mortality: better monitoring, indicators and benchmarks needed to improve emergency obstetric care. Research summary for policy makers Trop Med Int Health 2012 17 6 694 6 10.1111/j.1365-3156.2012.02983.x 22512353 14. Brenner S De Allegri M Gabrysch S Chinkhumba J Sarker M Muula AS The quality of clinical maternal and neonatal healthcare – a strategy for identifying ‘routine care signal functions’ PLoS One 2015 15 10(4) e0123968 10.1371/journal.pone.0123968 25875252 15. Gabrysch S Civitelli G Edmond KM Mathai M Ali M Bhutta ZA Campbell OM New signal functions to measure the ability of health facilities to provide routine and emergency newborn care PLoS Med 2012 9 11 e1001340 10.1371/journal.pmed.1001340 23152724 16. Tuncalp Ӧ, Were WM, MacLennan C, Oladapo OT, Gulmezoglu AM, Bahl R, Daelmans B, Mathai M, Say L, Kristensen F, Temmerman M, Bustreo F. Quality of care for pregnant women and newborns—the WHO vision. BJOG. 2015; doi: 10.1111/1471-0528.13451 17. Bryce J Arnold F Blanc A Hancioglu A Newby H Requejo J Measuring Coverage in MNCH: New Findings, New Strategies, and Recommendations for Action PLoS Med 2013 10 5 e1001423 10.1371/journal.pmed.1001423 23667340 18. Marchant T Tilley-Gyado RD Tessema T Singh K Gautham M Umar N Berhanu D Cousens S Armstrong Schellenberg JR Adding content to contacts: measurement of high quality contacts for maternal and newborn health in Ethiopia, north east Nigeria, and Uttar Pradesh, India PLoS One 2015 22 10(5) e0126840 10.1371/journal.pone.0126840 26000829 19. Ebener S Guerra-Arias M Campbell J Tatem AJ Moran AC Amoako Johnson F Fogstad H Stenberg K Neal S Bailey P Porter R Matthews Z The geography of maternal and newborn health: the state of the art Int J Health Geography 2015 14 19 10.1186/s12942-015-0012-x 20. United Nations Sustainable Development Goals. http://www.un.org/sustainabledevelopment/sustainable-development-goals/. Accessed on 18 May 2016. 21. World Health Organization. Indicator and monitoring framework for the Global Strategy for Women’s, Children’s and Adolescent’s Health (2016-2030). Geneva, Switzerland: World Health Organization. 2016.

PMC005xxxxxx/PMC5002108.txt

==== Front BMC Evol BiolBMC Evol. BiolBMC Evolutionary Biology1471-2148BioMed Central London 73710.1186/s12862-016-0737-6Research ArticleTransgenerational plasticity following a dual pathogen and stress challenge in fruit flies http://orcid.org/0000-0002-8854-3928Nystrand M. +61-(0)3-9902 4346magdalena.nystrand@monash.edu Cassidy E. J. elizabethjanecassidy@gmail.com Dowling D. K. damian.dowling@monash.edu School of Biological Sciences, Monash University, Clayton, VIC 3800 Australia 27 8 2016 27 8 2016 2016 16 1 1712 5 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 Phenotypic plasticity operates across generations, when the parental environment affects phenotypic expression in the offspring. Recent studies in invertebrates have reported transgenerational plasticity in phenotypic responses of offspring when the mothers had been previously exposed to either live or heat-killed pathogens. Understanding whether this plasticity is adaptive requires a factorial design in which both mothers and their offspring are subjected to either the pathogen challenge or a control, in experimentally matched and mismatched combinations. Most prior studies exploring the capacity for pathogen-mediated transgenerational plasticity have, however, failed to adopt such a design. Furthermore, it is currently poorly understood whether the magnitude or direction of pathogen-mediated transgenerational responses will be sensitive to environmental heterogeneity. Here, we explored the transgenerational consequences of a dual pathogen and stress challenge administered in the maternal generation in the fruit fly, Drosophila melanogaster. Prospective mothers were assigned to a non-infectious pathogen treatment consisting of an injection with heat-killed bacteria or a procedural control, and a stress treatment consisting of sleep deprivation or control. Their daughters and sons were similarly assigned to the same pathogen treatment, prior to measurement of their reproductive success. Results We observed transgenerational interactions involving pathogen treatments of mothers and their offspring, on the reproductive success of daughters but not sons. These interactions were unaffected by sleep deprivation. Conclusions The direction of the transgenerational effects was not consistent with that predicted under a scenario of adaptive transgenerational plasticity. Instead, they were indicative of expectations based on terminal investment. Electronic supplementary material The online version of this article (doi:10.1186/s12862-016-0737-6) contains supplementary material, which is available to authorized users. Keywords Anticipatory effectsMaternal effectsMaternal stressParental effectsPhenotypic plasticityTransgenerational effectsTransgenerational plasticityAustralian Research Council (ARC)DP110104965Monash University Margaret Clayton Women in Research Postdoctoral Fellowshipissue-copyright-statement© The Author(s) 2016 ==== Body Background Maternal effects occur when the maternal phenotype, or the environment in which the mother resides, shapes the expression of the offspring phenotype [1–3]. Hence, maternal effects can be traced to a diverse range of factors, from variation in maternal behavior to physiology, and can result in the differential transfer of nutrients, hormones, antibodies or epigenetic markers to offspring [1, 4–6]. Because of classic parent-offspring conflict over resources, maternal effects can have a negative influence on offspring fitness [3]. However, maternal effects can also be adaptive if they act to directly augment offspring fitness [1, 3, 7]. Such ‘positive’ maternal effects provide examples of adaptive transgenerational plasticity, sometimes referred to as ‘anticipatory maternal effects’ [1, 3, 8]. In particular, maternal effects are adaptive when an environmental challenge experienced by a mother prior to reproduction confers a relative increase in fitness to her subsequent offspring when they are themselves faced with a similar environmental challenge to that previously faced by their mother [3, 8]. As such, conclusive evidence for adaptive transgenerational plasticity can only be provided by studies that not only experimentally manipulate parental environments pre-reproduction, but also subsequently manipulate environments of the offspring, thereby either matching or mis-matching maternal and offspring environmental conditions [3, 8]. Such designs have successfully been adopted in a range of taxa; a classic example includes a study of seed beetles (Stator limbatus), in which eggs that were deposited in the same environment as that experienced by the mother had greater survival than eggs that were deposited in a novel environment [9]. Likewise, in a study of toxicant-resistance in bryozoans (Bugula neritina), offspring that were exposed to the same pollutant as their mothers had higher survival than offspring that were not [10], and a study of American bellflowers (Campanulastrum americanum), showed that offspring had higher fitness when their light environment was matched to that experienced by their mothers [11]. However, there are also a range of studies that have failed to find evidence for adaptive transgenerational plasticity [3, 12–14], and indeed, a recent meta-analysis concluded that there was only weak evidence for adaptive transgenerational plasticity across natural populations of plants and animals [8]. Numerous studies conducted in invertebrates have shown that pathogen challenges administered to females prior to reproduction can influence phenotypic expression in their offspring. In several cases, application of a maternal pathogen challenge has been associated with increased robustness of the offspring phenotype, including increased survival [15–17], reproductive success [18, 19] or augmented immune responses e.g. [15, 16, 20–22]. Because such effects have been shown to augment proxies of offspring fitness, it has been commonly inferred that they provide an adaptive benefit to offspring. Furthermore, because several studies have shown that such maternal pathogen challenges can affect the expression of offspring immune components (e.g. offspring phenoloxidase [PO] or antimicrobial activity), researchers have commonly concluded that such findings represent cases of ‘transgenerational immune priming’, and such priming effects have been suggested to be functionally similar to the transfer of acquired immunity in vertebrates (e.g. the actual transfer of antibodies) [4, 18, 23]. Yet, most of the studies exploring pathogen-mediated transgenerational effects in invertebrates suffer from two limitations. First, while there is evidence that mothers can enhance offspring immune function, there is no evidence that mothers transmit actual immune factors [24] in a manner similar to that of vertebrate adaptive immune priming (e.g. antibody transfer) [4, 5, 25]. Hence, any analogues between invertebrate ‘immune priming’ and the vertebrate adaptive immune system are contentious and in need of further research to specifically address the underlying mechanisms. This is because pathogen-induced transgenerational effects on immune expression could simply be a result of general condition-dependence, or be mediated by epigenetic changes to the DNA [1, 4, 24, 26–28]. Alternatively, maternally induced offspring immune augmentation may be caused by mothers transferring the pathogen per se (or parts thereof), hence triggering the offspring’s own immune system, as suggested in some recent studies that have demonstrated maternal transfer of bacterial fragments to eggs [24, 29]. However, the lack of precise mechanistic explanations is a relatively minor problem from an evolutionary standpoint. Of broader significance is the second limitation, which relates to the experimental design; most prior studies have failed to implement designs with the experimental power to detect whether transgenerational patterns were adaptive. It is for this exact reason that a recent meta-analysis surveying the evidence and effect sizes associated with adaptive trans-generational immunity, chose not to incorporate studies examining transgenerational consequences of pathogen-mediated challenges as part of their base of evidence [8]. We elaborate below. The adaptive value of transgenerational immune effects, to both mother and offspring, will inevitably depend on the benefits that such effects bring to offspring produced by pathogen-exposed mothers relative to the costs imposed on the mother from investing in offspring immune-protection − a cost-benefit equation that could, for example, be heavily shaped by maternal condition-dependence if pathogen-challenged mothers are in poorer condition [30]. A prediction based on life history theory is that offspring of pathogen-treated mothers would exhibit higher relative fitness if they themselves are exposed to the same pathogen in their lifetime, but lower relative fitness otherwise (due to the costs associated with the pathogen challenge in the maternal generation). That is, from a life-history perspective, assessing the benefits of a maternal pathogen challenge across matched and mis-matched contexts is key to interpretation. Yet, only a handful of previous invertebrate studies have utilized experimental and analytical approaches that can fully dissect the adaptive nature of pathogen-mediated transgenerational plasticity on offspring fitness [15, 16, 18, 19, 31, 32], by harnessing factorial designs in which maternal and offspring immune environments are in fact matched or mis-matched [3, 8]. Furthermore, while it is well established that the magnitude and direction of maternal effects can be substantially altered depending on internal (e.g. maternal age, size) or external (e.g. environmental condition, diet) environment, e.g. [10, 33, 34], studies addressing such context-dependence in relation to pathogen-mediated maternal effects are limited, and most research comes from subjects that are assayed under benign and largely stress-free laboratory environments [35–37]. It thus remains unclear whether cases of adaptive pathogen-mediated transgenerational plasticity will be upheld when mothers are faced with multifaceted stresses similar to those likely to be experienced under natural conditions [35, 37]. Here, we set out to test whether interactions between a pathogen-challenge combined with a stress challenge in the maternal generation interact with a pathogen challenge in the offspring generation, to shape offspring reproductive success in fruit flies (Drosophila melanogaster). The pathogen challenge was a non-infectious challenge, in that the bacteria used (Escherichia coli mixed with Micrococcus luteus) had been heat-killed prior to use, thus activating maternal immune genes but without causing direct pathogenesis [19, 38]. We limited our assay to measurement of one key life-history trait in the offspring ─ reproductive success. This is a trait that has previously been shown to be sensitive to a similar heat-killed pathogen challenge in this species [19, 39]. We used sleep deprivation as a measure of maternal stress, because it is well-studied in Drosophila. Variation in sleep has been shown to interact with a number of physiological and behavioural traits in Drosophila, such immune function, metabolism, learning and memory function [40–48]. Moreover, sleep patterns are known to influence the expression of a number of immune-related genes [41, 44, 49, 50]. Of key interest to our study is prior evidence suggesting that flies experiencing sleep deprivation exhibit greater resistance to bacterial infection relative to controls [41, 51], suggesting that sleep-stress might play an important, but hitherto unexplored role in augmenting the strength of pathogen-mediated adaptive transgenerational effects. The experiment was designed to test two main predictions. The first prediction was that interactions between maternal and offspring pathogen challenges on offspring reproductive success would conform to a pattern expected under adaptive transgenerational plasticity, with matched combinations of maternal and offspring pathogen exposure outperforming mis-matched combinations (i.e. pathogen-challenged offspring would perform better if their mothers had also been pathogen-challenged). The second prediction revolved around whether maternal sleep deprivation prior to reproduction would act to augment signatures of pathogen-mediated adaptive transgenerational effects or not, in female and male offspring. Methods Fly population The flies used in this study originated from a natural population collected from three different localities in Coffs Harbour (New South Wales, Australia), in February 2010 [52]. Each of a total of 60 wild-caught, non-virgin females, contributed 10 sons and 10 daughters to form a mass-bred laboratory population, which was then maintained in pairs of 20 adults across twelve 10-dram vials, under standardized rearing conditions of 12 L: 12 D cycle, 25 °C, 6 ml of potato-dextrose-agar-yeast substrate and ad libitum access to live yeast per vial. The population was propagated using flies that were allowed a 20 h period to oviposit when 4 days of adult age. The eggs of each vial were then trimmed to between 80 and 100 per vial. Upon their eclosion into adulthood, offspring were admixed with offspring from all other vials prior to their separation and transfer into 12 new vials, each comprised of 20 pairs of adults. At the onset of the experiment, the laboratory population had been maintained under these conditions in the laboratory for four years (~100 generations of evolution). Thus, the population effectively represents a “laboratory island population”, in which the laboratory environment (and specifically the rearing protocol outlined above) now represents the natural environment in which the population has evolved, and in which most quantitative traits would have had ample time to reach their new evolutionary optima [53]. Admittedly, this laboratory population is much simpler than a natural population, and lacks the multifaceted stresses that natural populations will face. But, it serves as an excellent platform on which to add and assess the effects of multiple stresses, enabling us to assess the transgenerational consequences of dual sleep and pathogen challenges, in the context of the environment in which this population has actually adapted over tens of generations. Recent work on this laboratory population has confirmed abundant levels of quantitative genetic variance for life history trait expression, and strong signatures of transgenerational inheritance mediated by sexual conflict [52, 54, 55]. Prior to the experiment in June 2014, a replicate copy of this population was collected, and the population size expanded to 60 vials per generation, maintained using the same protocols and conditions as described above. Experimental details Pathogen treatment To manipulate maternal and offspring perception of the local pathogenic environment, we challenged flies with a mix of heat-inactivated bacteria composed of the Gram-positive Micrococcus luteus and the Gram-negative Escherichia coli. There are two main benefits of challenging flies with heat-killed bacteria. Firstly, it allows us to detect the effects a primary immune challenge, whilst excluding the confounding negative effects associated with invading, replicating pathogens [20, 56]. Secondly, and crucially, it ensures that transgenerational effects on offspring fitness are not simply artefacts of the transmission of a live bacterial infection from mother to offspring [24]. Moreover, the rationale behind administering a mix of Gram-positive and Gram-negative bacteria was grounded in the fact that they activate different immune pathways; Gram-positive bacteria primarily stimulate the Toll pathway, and Gram-negative bacteria primarily stimulate the Imd pathway [38, 57]. Both these pathways are used in the defense against bacteria and fungi, by regulating the production of antimicrobial peptides (AMPs). By triggering both pathways, we speculate that the perceived immune insult may be enhanced, which might possibly generate a stronger response (by additive effects, or by interactive crosstalk) and cause a longer lasting cost in the flies [38, 57, 58]. Virgin females (N = 210) were collected from the experimental population, and transferred to vials of 9–14 individuals per vial (x̄ = 11.40 ± SE 0.18), with ad libitum access to live yeast provided on the surface of the potato-dextrose-agar-yeast substrate. At three days of adult age, these females were assigned to a maternal pathogen treatment, which consisted of two levels: an injection of heat-killed bacteria or a procedural control (using the nano-injector “Nanoject”, Drummond Scientific Company, Broomall, PA, USA). Half of the females were assigned to the control, and received a microinjection of 41.1 nl of phosphate-buffered saline (PBS, Sigma Aldrich table P4417, pH 7.4) into their abdomens. The other half received the pathogen treatment, which consisted of a cocktail of equal volume of a Gram-positive bacteria (M. luteus strain, A204, OD600 = 0.1, corresponding to ~ 1.1 x 106 CFU per fly) and a Gram-negative bacteria (E. coli, strain K12, OD600 = 1.0, corresponding to ~ 27.5 x 106 CFU per fly), both provided as heat-killed (verified by colony growth test, supplied by Micromon, Monash University, Australia), and diluted in PBS. While Drosophila may encounter both E. coli and M. luteus in the wild [59], it is unclear to what extent these bacterial species may influence Drosophila fitness. In fact, little is still known about which species of bacteria that D. melanogaster are most likely to come across in their natural environment, and one of the few studies to address host-pathogen associations in wild populations have found it to be both diverse and varied across populations (looking at D. melanogaster from the East-coast in the USA) [60]. In contrast, another study exploring the pathogenic species composition across Drosophila populations and species throughout the world, found that it was made up of four predominant bacterial families. However, similar to the study by Corby-Harris and Promislow (2008), the sampling distribution in this study was also limited, and while sampling 14 wild Drosophila species and two closely related species, only three were in fact D. melanogaster populations – all of which, again, were from the USA [59]. Thus, the host-pathogen constellations across species and regions are still largely unknown. Therefore, rather than exactly attempting to mimic natural (bacterial) conditions, we chose to use pathogen species which have proven effects on Drosophila immune function and fitness; both bacteria have been shown to have the capacity to induce an immune response in adult D. melanogaster (in their live or heat-killed form) [38, 39, 57, 61], and both are capable of influencing reproductive success [19, 39]. The concentration of E. coli and M. luteus used in the pathogen treatment was based on previous work [19], which had adopted a pilot experiment to test for the effect of different bacterial concentrations on the reproductive success of male and female flies (Additional file 1: Figure S1). Specifically, the final concentration used per pathogen were chosen because they had the largest, and most similar, effects on depressing reproductive success of females. Finally, the injected volumes were based on established protocols, which has recently also been verified to be within the optimal range for bacterial injections in Drosophila [62]. Following the injection, flies were returned to fresh vials, in the absence of live yeast. Sleep deprivation treatment During the night (starting midnight) following the injections (pathogen treatment vs. control), half of the females of each pathogen treatment level were subjected to a sleep deprivation treatment, and the other half to a control. Specifically, the vials of flies assigned to the sleep deprivation were placed on a shaker (Biosan Multi RS-60 rotator, Biosan Medical-Biological Research & Technologies, Riga, Latvia), which was programmed to a cycle starting with a set speed of 3 RPM for 4 min, followed by a 12 s rotation at an angle between -60 and +60°, and finishing with 1° vibrations lasting 1 s. This cycle was repeated continuously for 8 h, at 25 °C in the complete dark (until 8 am). The control females were exposed to the same environment (25 °C, dark), but were not placed on the shakers and hence were not deprived of sleep. The sleep deprivation protocol is based on well-established methods for Drosophila, in which sleep is defined as any bout of inactivity that lasts for longer than five continuous minutes [63, 64]. The final protocol adopted in our study was optimized via a pilot study, which demonstrated that the treatment was sufficient to induce sleep-deprivation (Additional file 1: Table S1, Figure S2-3). Briefly, flies subjected to this treatment compensated by sleeping more in the first four hours of daylight, compared with flies subjected to the control (Additional file 1: Table S1. Figure S2). Moreover, flies compensated by sleeping for a substantially longer period of time following the application of the sleep deprivation treatment during night-time hours (i.e. when flies would generally rest) compared to when the sleep-deprivation treatment was applied during daylight hours (Additional file 1: Table S1, Figure S3). Furthermore, the large difference in sleep compensation when the treatment was applied during night-time relative to day-time hours verifies that the response is not merely reflecting a general response to mechanical stress (in which case the flies should compensate equally much regardless of time of day) [64]. Hence, taken together, the results from our sleep treatment pilot study suggest that the sleep deprivation protocol adopted in here was sufficient to disrupt sleep patterns of the flies (more details in the Additional file 1: “Supplementary methods”). Production of focal offspring, offspring pathogen treatments and reproductive assays On the day following sleep-deprivation (24 h post-injection), an equal number of four days old males from the lab population were added to vials containing virgin females of a sex ratio of 1:1. Flies were allowed to mate for 4 h, after which females were transferred to individual vials for 22 h, to enable ovipositioning. These vials were supplemented with exactly 5 μl of a live yeast slurry (which consisted of 1.2 g yeast dissolved in 10 ml purified water). At the completion of these 22 h, the eggs of each vial were then carefully transferred into fresh vials, containing 6 ml of food substrate, at a density of approximately 25 eggs per vial (x̄ = 22.27 ± SE 0.19, minimum 12 eggs), and females were discarded. Nine days later, approximately 30 daughters (x̄ = 29 ± SE 0.4) and 30 sons (x̄ = 34 ± SE 1.7) were collected as virgins from each combination of maternal pathogen and sleep deprivation treatments across each experimental block and placed into three sex-specific vials (x̄ females = 3.3 ± SE 0.3, x̄ males = 3.3 ± SE 0.3) per experimental block. Once these offspring had reached three days of adult age, males and females were assigned to the same pathogen treatment protocol (cocktail or control) as had been applied to the maternal generation, in all possible maternal sleep × pathogen treatment combinations (Fig. 1). In total, approximately 200 focal offspring per maternal sleep – and pathogen treatment combination were injected and then placed in individual, non-yeasted vials for recovery. A four day old virgin fly of the opposite sex was added to each of the vials of focal offspring for 4 h, to enable mating. Following mating, the male flies of each vial were discarded (i.e. tester males in the female assay, and the focal males in the male assay).Fig. 1 Experimental design. Mothers were first assigned to a pathogen treatment (pathogen-challenged or control) and then allocated to one of two groups, that were either exposed to a night of sleep-deprivation or undisrupted sleep. Two daughters and two sons from of each of these mothers were then exposed to the pathogen treatment (one of each to the challenge and one of each to the control), prior to an assay of their reproductive performance A previous study on this population has demonstrated that reproductive output across the first 4 days of ovipositing (starting at age four to five) is tightly correlated with output across the first 10 days [19]. Thus, the females of each vial were retained and allowed to oviposit for the next 96 h, with each female transferred to a new vial every 24 h. All eclosing offspring of these females (the grand-offspring of the treated mothers) were counted 13 days later. Focal female reproductive success was therefore the total number of offspring produced across a 4-day ovipositioning window following a 4 h mating opportunity. Similarly, focal male reproductive success was the total number of offspring produced per male, after a 4 h mating opportunity with a solitary tester female, who was then provided with a 4-day ovipositioning window. The experiment was repeated over seven full blocks, all of which included all possible treatment and sexes. This generated a final sample of 798 focal male offspring (i.e. the sons of the treated mothers) and 813 focal female offspring (i.e. the daughters), who were each screened for their reproductive success. Note that the initial sample size was larger (854 for females, 848 for males), but was reduced due to mortality (N = 57) or flies escaping (N = 36) throughout the experiment. We also collected offspring mortality data post-injections, for the 48 period immediately following offspring injections, and the 24 h period immediately following the maternal injections. This data showed that mortality 24 h post-injection did not differ between the pathogen treatment and the control treatment (Dead: NPBS = 31, NBac = 36, Ntotal = 1031, χ2 = 0.37, p = 0.54). Likewise, there were no differences in offspring mortality in the 48 h period post-injection for flies subjected to the pathogen challenge relative to the control (Dead females: NPBS = 5, NBac = 0, total females including escapees = 854; Dead males: NPBS = 15, NBac = 11, χ2 = 0.62, p = 0.56, total males including escapees = 848). There were also no differences in post-injection mortality among the different treatments in female offspring mortality when looking across the entire four day period of ovipositing (i.e., over a 96 h period post-injections, Female deaths: NPBS = 12, NBac = 14, χ2 = 0.15, p = 0.85). Statistical analysis All data were analysed using mixed models in R v. 3.1.1 [65]. The response variable was offspring reproductive success (i.e. number of grand-offspring to sleep-treated × pathogen-treated mothers), and fixed factors were maternal sleep treatment (deprived or control), maternal pathogen treatment (heat-killed bacteria mix or control PBS), and offspring pathogen treatment (heat-killed bacteria mix or control PBS). We also fitted random effects that fully accounted for the structure of the data, i.e. experimental block, vial identity, and maternal identity. The reproductive success of male and female offspring were analysed in separate models, because the assayed trait was not directly comparable across sexes. The female trait is a gauge of reproductive fecundity over four days in early life following a 4 h mating opportunity. The male trait is a gauge of the ability of a male to mate with a female over a 4 h mating window, and measures his subsequent fertility resulting from that mating. The models were overdispersed, and also zero-inflated (as initially indicated by visual inspection of frequency distribution). We confirmed zero-inflation by 1) comparing all possible models in a fixed framework (i.e. block entered as a factor), which allowed us to apply a Vuong’s test [66] to model assessment, and 2) using a full mixed model, then simulating 95 % confidence intervals (C.I.) around the number of 0 values expected from a true Poisson model corrected for overdispersion. Based on the outcome of these tests, we applied a zero-inflated model, fitted with a negative binomial distribution (to relax the assumption of equal means and variance; NB1 fit, in which variance is calculated as Øμ). This was done with the package glmmADMB [67], which utilises a model that allows the zeros to be a mix of structural and sampling data points [68]. In our case, some of the zeros may have been generated because the females did not mate, and others may have arisen from females that had mated but that were infertile themselves, or that mated with infertile males. Both models were type II, and fixed effects were estimated using maximum likelihood. All interactions up to second order were tested. The effect of higher order interactions were assessed using log-likelihood ratio tests, comparing model deviance between the full model and the reduced model initially, and thereafter, between the progressively reduced models. Hence, model reduction was conducted by removing the least significant interactions (p > 0.05) one at a time in a stepwise manner. We present the final model, in which all non-significant interactions has been dropped (Table 1). We also confirmed that the final reduced model had a higher level of empirical support by comparing AIC values between both the null and the full model (AIC difference >2). Raw data has been deposited in Dryad [69].Table 1 Effect of pathogen- and sleep treatments on (a) female and (b) male offspring reproductive performance a) Fixed effects df LRT Pr (>χ2)  Maternal sleep treatment 1 2.78 0.0955  Maternal pathogen treatment 1 0.00 1.0000  Offspring pathogen treatment 1 0.12 0.7290 Maternal pathogen treatment × Offspring pathogen treatment 1 5.54 0.0186 Random effects Variance  Parental vial (block) 0.0066  Maternal identity (Parental vial (block)) 0.0192 b) Fixed effects df LRT Pr (>χ2)  Maternal sleep treatment 1 0.02 0.8875  Maternal pathogen treatment 1 0.70 0.4028  Offspring pathogen treatment 1 0.34 0.5598 Random effects Variance  Parental vial (block) 0.0048  Maternal identity (Parental vial (block)) 0.0418 Log-likelihood ratios (LRT) and their associated p-values were generated by comparing the full and reduced models in a stepwise manner, by sequentially removing non-significant terms. Final fit was confirmed by comparing AIC values between the null, reduced, and the full model (AIC > 2). Displayed are final models (significant effects emboldened) Results Reproductive success of the focal female offspring was influenced by an interaction between the maternal and offspring pathogen treatments (Table 1). Female offspring exhibited lower reproductive success when both they, and their mothers, had been administered the control treatment, than when they had received the pathogen treatment but their mothers had received the control (Fig. 2a). The maternal sleep deprivation treatment had no effect on reproductive success of females across generations (Table 1a).Fig. 2 Total number of eclosing offspring (raw means ± SE) produced by a) pathogen- and control-treated female offspring produced by pathogen- or control-treated mothers, and b) pathogen- and control-treated male offspring produced by pathogen- or control-treated mothers In contrast, male offspring reproductive success was not influenced by the maternal pathogen challenge, maternal sleep treatment, or by the offspring pathogen challenge, nor by any of the interactions between these factors (Table 1b, Fig. 2b). Discussion We screened for pathogen-mediated adaptive transgenerational plasticity in a population of D. melanogaster, and investigated whether such effects were modified by an additional maternal stress administered via sleep deprivation. We found that transgenerational interactions, tied to the maternal and offspring pathogen treatments, influenced the reproductive success of female, but not male, offspring. The patterns observed, however, were not consistent with those predicted under an adaptive scenario, whereby pathogen-challenged offspring would be expected to benefit if their mothers had previously faced the same pathogen-challenge and therefore would provide offspring with information on their future (pathogenic) environment. While pathogen-challenged daughters exhibited higher reproductive success than their control counterparts, these patterns were only apparent amongst daughters of control-treated mothers. Thus, pathogen-challenged mothers did not confer any detectable adaptive benefits to their offspring when it came to the reproductive capacity of offspring following an identical pathogen challenge. Rather, the pattern of higher reproductive success amongst pathogen-challenged daughters born to control-treated mothers is possibly more consistent with a process of terminal investment. Furthermore, the reported effects were not in any way altered by the addition of a second maternal stress, imposed by sleep deprivation. A key prediction of our study was that we would confirm evidence for pathogen-associated transgenerational plasticity, consistent with the results of a previous study in D. melanogaster [19]. In that study, both mothers and fathers of two different ages, and their daughters, were subjected to a pathogen-challenge, consisting only of the Gram-positive bacterium M. luteus, or a procedural control. While control-treated daughters produced by control-treated mothers exhibited the highest reproductive success, pathogen-challenged daughters performed better if their mothers had likewise received the pathogen-challenge than if their mothers had received the control. However, the results we present here are discordant with those previously observed, with pathogen-challenged female offspring born to control-treated mothers exhibiting higher reproductive success than maternal-offspring combinations in which both parties received the control. A key difference between the present study and the study of Nystrand and Dowling (2014) is that here we used a mix of both heat-killed Gram-negative and Gram-positive bacteria, in an attempt to maximize any potential effects on the immune system by increasing the chances that both the Imd and Toll pathways would be activated [38]. We note, however, that while there is evidence that these immune pathways can act synergistically [38, 70], this does not necessarily translate to a stronger, additive immune response. Nevertheless, it is conceivable that a more complex pathogen challenge might incur a higher cost to the host. In addition, this complexity could have been further reinforced if the sleep deprivation treatment, indirectly, accentuated the stress to the immune system,by activating the highly-conserved JAK-stat pathway. The JAK-stat signaling pathway, apart from being involved in development and immune function in Drosophila (primarily activated by septic injury) [38, 71–73], has recently also been implied to play a role in controlling circadian rhythm by driving rest-activity rhythms [74]. It is also possible that the primary difference between this study and that of Nystrand and Dowling (2014b) is simply a result of dose, given that the double challenge administered here also meant that we injected a much higher concentration of bacteria. Indeed, Nystrand and Dowling (2014a) previously recorded a pronounced dose-dependent effect on reproductive success in Drosophila [56]. Hence, we speculate that the increased effects associated with a double dose of bacteria (i.e. E. coli and M. luteus) in the current study, might have been sufficient to invoke a response more consistent with a scenario of ‘terminal investment’, and as a result, override any adaptive maternal pathogen-driven fitness effects. The fact that the lowest reproductive success was attributable to control-treated female offspring who were produced by control-treated mothers supports this interpretation. Under a scenario of terminal investment, an individual perceiving an imminent threat to survival should reallocate their resources to current reproduction at the expense of somatic maintenance and survival [75–78]. To this point, we have only discussed the effects of reproductive success in the focal female offspring. We found no pathogen- or sleep-mediated transgenerational effects on the reproductive success of sons. This provides some evidence that the effects of pathogen-mediated transgenerational interactions on offspring fitness are sex-specific, and may disproportionately affect the reproductive success of females. This contention, however, requires further testing. While this is the first study to specifically explore whether pathogen-mediated transgenerational effects adaptively augment the reproductive success of males in D. melanogaster (by using a factorial design in which the pathogen treatments of mothers and their offspring were matched or mismatched across offspring sexes), we note that the reproductive assay of male reproductive success adopted here might less accurately reflect male reproductive potential in nature relative to the female assay. We assayed male reproductive success based on a male’s ability to successfully mate with one virgin female, and in the event of doing so, the number of offspring that female went on to produce. Thus, male reproductive success in this assay is based on male pre-copulatory capacity to engage a female in copulation, as well as post-copulatory variation in the quality of the male ejaculate, such as proteomic variation in the constitution of his reproductive proteins, and the number and quality of his sperm transferred during copulation [79, 80]. While female reproductive success is limited by the total number of ova she can oviposit over a given period of time (in this assay, 4 d), male reproductive success can increase as a function of the number of females he has access to [76]. Furthermore, male outcomes are likely to change when assayed under conditions of sperm competition with rival males [80] – a scenario that aligns with the natural mating system of the species. Hence, it will be worthwhile to further investigate the potential for adaptive transgenerational effects manifesting in males, using assays that gauge male fertility outcomes under scenarios of multiple matings or reproductive competition. A key goal of our study was to determine whether any adaptive maternal effects mediated by a non-infectious pathogen challenge would be further modified in response to an additional stressor. We used sleep deprivation as an inducer of stress, because it is known to upregulate a cluster of immune genes involved in anti-inflammatory responses in D. melanogaster [41]. In fact, recent work has established that many of the physiological responses and genes involved in sleep, stress and immunity are related or the same [41, 47, 49, 81–87]. We found no direct transgenerational effects of sleep deprivation on the reproductive success of offspring, and nor did sleep deprivation interact with the maternal or offspring pathogen treatment. We note that our maternal sleep deprivation treatment was applied on the night immediately prior to the mothers reproducing. Thus, it did not represent a chronic sleep stress treatment over many nights, but rather a short-term but intense bout of sleep disruption. We also note that the treated individuals in our study had access to ad libitum food, which introduces the potential for compensatory behavior by the flies to offset the effects of the sleep deprivation. In particular, dietary conditions have been shown to affect the outcome of sleep-deprivation in experiments in both humans [88, 89] and rats [90]. Finally, the reported effects on reproductive success are unlikely to be attributable to the effects of selection leading to differential mortality among either the pathogen-treated mothers or pathogen-treated daughters relative to the control-treated counterparts. Mortality levels following the pathogen treatment were invariably low, and no higher amongst mothers or offspring subjected to the pathogen challenge relative to the controls. Thus, our results plausibly reflect true transgenerational interactions tied to maternal effects. Conclusions We documented transgenerational interactions, involving a maternal and an offspring challenge with a non-infectious pathogen, on reproductive success. The observed effects were detected in daughters only, and were not influenced by an additional stress of sleep deprivation in the maternal generation. Furthermore, the transgenerational interactions were not consistent with an interpretation of pathogen-mediated adaptive plasticity, but do provide some preliminary support for a response mediated via terminal investment. Indeed, when considering that this study used a double challenge (whether caused by a more complex response induced by stimulating both main immune pathways, or the result of a simple dose-effect introduced by the administration of higher concentrations of bacteria), relative to a previous study that utilized only a Gram-negative bacterium, this raises the possibility that there might well be a switch-point above which terminal investment effects override those of any potential pathogen-mediated adaptive transgenerational effects. These are research avenues that are amenable to further experimental enquiry, using experimental designs that incorporate dose-dependence and multiple pathogens, of varying perceived pathogenicity (heat-killed and live). Moreover, further exploration of the mechanisms that underpin the pathogen-mediated transgenerational interactions presented here would be valuable, to disentangle whether the observed effects are mediated by transfer of immune factors, a controversial premise [91–93]; or alternatively via classic maternal effects, either condition-dependent and involving reallocation of maternal resources following a challenge with a pathogen, or via epigenetic effects. Such avenues provide a natural extension of this research and an exciting framework for further study into the regulation of pathogen-mediated transgenerational effects, and how such effects may be altered by context-dependence. Additional file Additional file 1: Table S1. Result of pilot experiment exploring sleep disruptive treatment. Figure S1 a & b. Results of pilot experiments testing the effect of different concentrations of two different heat-killed pathogens, including procedural controls (PBS) and naïve controls, on reproductive success. Figure S2 and S3. Pilot experiment exploring the average number of minutes slept post pathogen-challenge (Figure S2), and post sleep-treatment during the night vs. the day (Figure S3). Supplementary methods - Potential influence of the sleep treatment on maternal mating behaviour, and reproductive success, including concomitant effects on offspring reproductive success. (DOC 187 kb) Acknowledgements We thank Winston Yee for maintenance of laboratory stocks. Funding This work was supported by both the Australian Research Council (Australian Postdoctoral Fellowship and Discovery Project, DP110104965) and a Monash University Margaret Clayton Women in Research Postdoctoral Fellowship to M.N. Availability of data and materials The data sets supporting the results of this article are available in the Dryad repository, http://dx.10.5061/dryad.8m68k. Authors’ contributions MN designed and coordinated the study, conducted the statistical analysis, and wrote the manuscript; EC carried out lab work, and contributed to design; DKD helped design the study and write the manuscript. All authors read and approved the final manuscript Competing interests None of the authors have any competing interests. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. ==== Refs References 1. Mousseau TA Fox CW Maternal Effects As Adaptations 1998 USA Oxford University Press 2. Badyaev AV Uller T Parental effects in ecology and evolution: mechanisms, processes and implications Philos Trans R Soc Lond B Biol Sci 2009 364 1520 1169 1177 10.1098/rstb.2008.0302 19324619 3. Marshall D Uller T When is a maternal effect adaptive? Oikos 2007 116 12 1957 1963 10.1111/j.2007.0030-1299.16203.x 4. Hasselquist D Nilsson J-Å Maternal transfer of antibodies in vertebrates: trans-generational effects on offspring immunity Proc R Soc Lond B Biol Sci 2009 364 1513 51 60 10.1098/rstb.2008.0137 5. Grindstaff JL Brodie ED Ketterson ED Immune function across generations: integrating mechanism and evolutionary process in maternal antibody transmission Proc R Soc Lond B Biol Sci 2003 270 1531 2309 2319 10.1098/rspb.2003.2485 6. Mousseau TA Fox CW The adaptive significance of maternal effects Trends Ecol Evol 1998 13 10 403 407 10.1016/S0169-5347(98)01472-4 21238360 7. Wolf JB Wade MJ What are maternal effects (and what are they not)? Philos Trans R Soc Lond Ser B: Biol Sci 2009 364 1520 1107 1115 10.1098/rstb.2008.0238 19324615 8. Uller T Nakagawa S English S Weak evidence for anticipatory parental effects in plants and animals J Evol Biol 2013 26 10 2161 2170 10.1111/jeb.12212 23937440 9. Fox CW Thakar MS Mousseau TA Egg size plasticity in a seed beetle: an adaptive maternal effect Am Nat 1997 149 149 163 10.1086/285983 10. Marshall DJ Transgenerational plasticity in the sea: context-dependent maternal effects across the life history Ecology 2008 89 2 418 427 10.1890/07-0449.1 18409431 11. Galloway LF Etterson JR Transgenerational Plasticity Is Adaptive in the Wild Science 2007 318 5853 1134 1136 10.1126/science.1148766 18006745 12. Rossiter M Incidence and consequences of inherited environmental effects Annu Rev Ecol Syst 1996 27 1 451 476 10.1146/annurev.ecolsys.27.1.451 13. Bernardo J Maternal Effects in Animal Ecology Am Zool 1996 36 2 83 105 10.1093/icb/36.2.83 14. Räsänen K Kruuk LEB Maternal effects and evolution at ecological time-scales Funct Ecol 2007 21 3 408 421 10.1111/j.1365-2435.2007.01246.x 15. Moret Y 'Trans-generational immune priming': specific enhancement of the antimicrobial immune response in the mealworm beetle, Tenebrio molitor Proc R Soc Lond Ser B: Biol Sci 2006 273 1592 1399 1405 10.1098/rspb.2006.3465 16. Roth O Joop G Eggert H Hilbert J Daniel J Schmid-Hempel P Kurtz J Paternally derived immune priming for offspring in the red flour beetle, Tribolium castaneum J Anim Ecol 2010 79 2 403 413 10.1111/j.1365-2656.2009.01617.x 19840170 17. Hernández López J Schuehly W Crailsheim K Riessberger-Gallé U Trans-generational immune priming in honeybees Proc R Soc Lond B Biol Sci 2014 281 1785 20140454 10.1098/rspb.2014.0454 18. Little TJ O'Connor B Colegrave N Watt K Read AF Maternal Transfer of Strain-Specific Immunity in an Invertebrate Curr Biol 2003 13 6 489 492 10.1016/S0960-9822(03)00163-5 12646131 19. Nystrand M Dowling DK Transgenerational interactions involving parental age and immune status affect female reproductive success in Drosophila melanogaster Proc R Soc Lond Ser B: Biol Sci 2014 281 1794 20141242 10.1098/rspb.2014.1242 20. Sadd BM Kleinlogel Y Schmid-Hempel R Schmid-Hempel P Trans-generational immune priming in a social insect Biol Lett 2005 1 4 386 388 10.1098/rsbl.2005.0369 17148213 21. Tidbury HJ Pedersen AB Boots M Within and transgenerational immune priming in an insect to a DNA virus Proc R Soc Lond Ser B: Biol Sci 2011 278 1707 871 876 10.1098/rspb.2010.1517 22. Zanchi C Troussard J-P Martinaud G Moreau J Moret Y Differential expression and costs between maternally and paternally derived immune priming for offspring in an insect J Anim Ecol 2011 80 6 1174 1183 10.1111/j.1365-2656.2011.01872.x 21644979 23. Little TJ Kraaijeveld AR Ecological and evolutionary implications of immunological priming in invertebrates Trends Ecol Evol 2004 19 2 58 60 10.1016/j.tree.2003.11.011 16701227 24. Freitak D Schmidtberg H Dickel F Lochnit G Vogel H Vilcinskas A The maternal transfer of bacteria can mediate trans-generational immune priming in insects Virulence 2014 5 4 547 554 10.4161/viru.28367 24603099 25. Grindstaff JL Maternal antibodies reduce costs of an immune response during development J Exp Biol 2008 211 5 654 660 10.1242/jeb.012344 18281327 26. Mousseau TA Dingle H Maternal Effects in Insect Life Histories Annu Rev Entomol 1991 36 1 511 534 10.1146/annurev.en.36.010191.002455 27. Curley JP Mashoodh R Champagne FA Epigenetics and the origins of paternal effects Horm Behav 2011 59 3 306 314 10.1016/j.yhbeh.2010.06.018 20620140 28. Jablonka E Raz G Transgenerational Epigenetic Inheritance: Prevalence, Mechanisms, and Implications for the Study of Heredity and Evolution Q Rev Biol 2009 84 2 131 176 10.1086/598822 19606595 29. Salmela H Amdam GV Freitak D Transfer of Immunity from Mother to Offspring Is Mediated via Egg-Yolk Protein Vitellogenin PLoS Path 2015 11 7 e1005015 10.1371/journal.ppat.1005015 30. Sadd BM Schmid-Hempel P A distinct infection cost associated with trans-generational priming of antibacterial immunity in bumble-bees Biol Lett 2009 5 6 798 801 10.1098/rsbl.2009.0458 19605389 31. Prior NH Washington CN Housley JM Hall SR Duffy MA Caceres CE Maternal effects and epidemiological traits in a planktonic host-parasite system Evol Ecol Res 2011 13 4 401 413 32. Sadd BM Schmid-Hempel P Facultative but persistent trans-generational immunity via the mother's eggs in bumblebees Curr Biol 2007 17 24 R1046 R1047 10.1016/j.cub.2007.11.007 18088585 33. Plaistow SJ Benton TG The influence of context-dependent maternal effects on population dynamics: an experimental test Philos Trans R Soc Lond B Biol Sci 2009 364 1520 1049 1058 10.1098/rstb.2008.0251 19324610 34. Agrawal AA Transgenerational consequences of plant responses to herbivory: an adaptive maternal effect? Am Nat 2001 157 5 555 569 10.1086/319932 18707262 35. Lazzaro BP Little TJ Immunity in a variable world Philos Trans R Soc Lond B Biol Sci 2009 364 1513 15 26 10.1098/rstb.2008.0141 18926975 36. Triggs AM Knell RJ Parental diet has strong transgenerational effects on offspring immunity Funct Ecol 2012 26 6 1409 1417 10.1111/j.1365-2435.2012.02051.x 37. Eggert H Diddens-de Buhr MF Kurtz J A temperature shock can lead to trans-generational immune priming in the Red Flour Beetle, Tribolium castaneum Ecol Evol 2015 5 6 1318 1326 10.1002/ece3.1443 25859336 38. Lemaitre B Hoffmann J The Host Defense of Drosophila melanogaster Annu Rev Immunol 2007 25 1 697 743 10.1146/annurev.immunol.25.022106.141615 17201680 39. Zerofsky M Harel E Silverman N Tatar M Aging of the innate immune response in Drosophila melanogaster Aging Cell 2005 4 2 103 108 10.1111/j.1474-9728.2005.00147.x 15771614 40. Svetec N Zhao L Saelao P Chiu JC Begun DJ Evidence that natural selection maintains genetic variation for sleep in Drosophila melanogaster BMC Evol Biol 2015 15 41 10.1186/s12862-015-0316-2 25887180 41. Williams J Sathyanarayanan S Hendricks J Sehgal A Interaction between sleep and the immune response in Drosophila : a role for the NFkappaB relish Sleep 2007 30 389 400 17520783 42. Koh K Evans JM Hendricks JC Sehgal A A Drosophila model for age-associated changes in sleep:wake cycles Proc Natl Acad Sci USA 2006 103 37 13843 13847 10.1073/pnas.0605903103 16938867 43. Bushey D, Cirelli C. From Genetics to Structure to Function: Exploring Sleep in Drosophila. Int Rev Neurobiol. 2011;(99):213–44. 44. Kuo T-H Pike D Beizaeipour Z Williams J Sleep triggered by an immune response in Drosophila is regulated by the circadian clock and requires the NFkappaB Relish BMC Neurosci 2010 11 1 17 10.1186/1471-2202-11-17 20144235 45. Cirelli C The genetic and molecular regulation of sleep: from fruit flies to humans Nat Rev Neurosci 2009 10 8 549 560 10.1038/nrn2683 19617891 46. Cirelli C Tononi G Is Sleep Essential? PLoS Biol 2008 6 8 e216 10.1371/journal.pbio.0060216 18752355 47. Shaw PJ Tononi G Greenspan RJ Robinson DF Stress response genes protect against lethal effects of sleep deprivation in Drosophila Nature 2002 417 6886 287 291 10.1038/417287a 12015603 48. Li X Yu F Guo A Sleep Deprivation Specifically Impairs Short-term Olfactory Memory in Drosophila Sleep 2009 32 11 1417 1424 19928381 49. Kuo T-H Williams JA Increased Sleep Promotes Survival during a Bacterial Infection in Drosophila Sleep 2014 37 6 1077 1086 24882902 50. Cirelli C LaVaute TM Tononi G Sleep and wakefulness modulate gene expression in Drosophila J Neurochem 2005 94 5 1411 1419 10.1111/j.1471-4159.2005.03291.x 16001966 51. Kuo TH Williams JA Acute sleep deprivation enhances post-infection sleep and promotes survival during bacterial infection in Drosophila Sleep 2014 37 5 859 869 24790264 52. Williams BR Van Heerwaarden B Dowling DK SgrÒ CM A multivariate test of evolutionary constraints for thermal tolerance in Drosophila melanogaster J Evol Biol 2012 25 7 1415 1426 10.1111/j.1420-9101.2012.02536.x 22587877 53. Rice W Linder J Friberg U Lew T Morrow E Stewart A Inter-locus antagonistic coevolution as an engine of speciation: Assessment with hemiclonal analysis Proc Natl Acad Sci U S A 2005 102 6527 6534 10.1073/pnas.0501889102 15851669 54. Dowling DK Williams BR Garcia-Gonzalez F Maternal sexual interactions affect offspring survival and ageing J Evol Biol 2014 27 1 88 97 10.1111/jeb.12276 24215553 55. Garcia-Gonzalez F, Dowling DK: Transgenerational effects of sexual interactions and sexual conflict: non-sires boost the fecundity of females in the following generation. Biol Lett 2015, 11(3).doi: 10.1098/rsbl.2015.0067. 56. Nystrand M Dowling DK Dose-dependent effects of an immune challenge at both ultimate and proximate levels in Drosophila melanogaster J Evol Biol 2014 27 5 876 888 10.1111/jeb.12364 24731072 57. Nehme NT Quintin J Cho JH Lee J Lafarge M-C Kocks C Ferrandon D Relative Roles of the Cellular and Humoral Responses in the Drosophila Host Defense against Three Gram-Positive Bacterial Infections PLoS One 2011 6 3 e14743 10.1371/journal.pone.0014743 21390224 58. McKean K Yourth C Lazzaro B Clark A The evolutionary costs of immunological maintenance and deployment BMC Evol Biol 2008 8 1 76 10.1186/1471-2148-8-76 18315877 59. Chandler JA Morgan Lang J Bhatnagar S Eisen JA Kopp A Bacterial Communities of Diverse Drosophila Species: Ecological Context of a Host–Microbe Model System PLoS Genet 2011 7 9 e1002272 10.1371/journal.pgen.1002272 21966276 60. Corby-Harris V Promislow DEL Host ecology shapes geographical variation for resistance to bacterial infection in Drosophila melanogaster J Anim Ecol 2008 77 4 768 776 10.1111/j.1365-2656.2008.01399.x 18489569 61. Elrod-Erickson M Mishra S Schneider D Interactions between the cellular and humoral immune responses in Drosophila Curr Biol 2000 10 13 781 784 10.1016/S0960-9822(00)00569-8 10898983 62. Khalil S Jacobson E Chambers MC Lazzaro BP Systemic Bacterial Infection and Immune Defense Phenotypes in Drosophila Melanogaster J Vis Exp 2015 99 52613 63. Shaw P Cirelli C Greenspan R Tononi G Correlates of sleep and waking in Drosophila melanogaster Science 2000 287 1834 1837 10.1126/science.287.5459.1834 10710313 64. Cirelli C Searching for sleep mutants of Drosophila melanogaster Bioessays 2003 25 10 940 949 10.1002/bies.10333 14505361 65. R Development Core Team R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing 2012 Vienna R Foundation for Statistical Computing 66. Vuong QH Likelihood Ratio Tests for Model Selection and Non-Nested Hypotheses Econometrica 1989 57 2 307 333 10.2307/1912557 67. Skaug HJ, Fournier DA, Nielsen A, Magnusson A, Bolker BM: Package glmmADMB: Generalized linear mixed models using AD Model Builder. R package version 0.7.3. http://r-forge.r-project.org/projects/glmmadmb/. 2012. 68. Hu M-C Pavlicova M Nunes EV Zero-inflated and Hurdle Models of Count Data with Extra Zeros: Examples from an HIV-Risk Reduction Intervention Trial Am J Drug Alcohol Abuse 2011 37 5 367 375 10.3109/00952990.2011.597280 21854279 69. Nystrand M Cassidy EJ Dowling D Data from: Transgenerational plasticity following a dual pathogen and stress challenge in fruit flies Dryad Digital Repository. http://dx. 2016 70. Valanne S Wang J-H Rämet M The Drosophila Toll Signaling Pathway J Immunol 2011 186 2 649 656 10.4049/jimmunol.1002302 21209287 71. Agaisse H Perrimon N The roles of JAK/STAT signaling in Drosophila immune responses Immunol Rev 2004 198 72 82 10.1111/j.0105-2896.2004.0133.x 15199955 72. Arbouzova NI Zeidler MP JAK/STAT signalling in Drosophila : insights into conserved regulatory and cellular functions Development 2006 133 14 2605 2616 10.1242/dev.02411 16794031 73. Wang H Chen X He T Zhou Y Luo H Evidence for Tissue-Specific JAK/STAT Target Genes in Drosophila Optic Lobe Development Genetics 2013 195 4 1291 1306 10.1534/genetics.113.155945 24077308 74. Luo W Sehgal A Regulation of circadian behavioral output via a MicroRNA-JAK/STAT circuit Cell 2012 148 4 765 779 10.1016/j.cell.2011.12.024 22305007 75. Clutton-Brock TH Reproductive Effort and Terminal Investment in Iteroparous Animals Am Nat 1984 123 2 212 229 10.1086/284198 76. Bateman AJ Intra-sexual selection in Drosophila Heredity (Edinb) 1948 2 3 349 368 10.1038/hdy.1948.21 18103134 77. Roff DA. Life History Evolution. Sunderland: Sinauer; 2002 78. Williams GC Natural Selection, the Costs of Reproduction, and a Refinement of Lack's Principle Am Nat 1966 100 916 687 690 10.1086/282461 79. Simmons LW: Sperm Competition and Its Evolutionary Consequences in the Insects. New Jersey: Princeton University Press; 2001. 80. Yee WKW Sutton KL Dowling DK In vivo male fertility is affected by naturally occurring mitochondrial haplotypes Curr Biol 2013 23 2 R55 R56 10.1016/j.cub.2012.12.002 23347935 81. Meerlo P Sgoifo A Suchecki D Restricted and disrupted sleep: Effects on autonomic function, neuroendocrine stress systems and stress responsivity Sleep Med Rev 2008 12 3 197 210 10.1016/j.smrv.2007.07.007 18222099 82. Harbison ST Mackay TFC Anholt RRH Understanding the neurogenetics of sleep: progress from Drosophila Trends Genet 2009 25 6 262 269 10.1016/j.tig.2009.04.003 19446357 83. Zimmerman J Naidoo N Raizen D Pack A Conservation of sleep: insights from non-mammalian model systems Trends Neurosci 2008 31 371 376 10.1016/j.tins.2008.05.001 18538867 84. Adamo SA Why should an immune response activate the stress response? Insights from the insects (the cricket Gryllus texensis ) Brain, Behav Immun 2010 24 2 194 200 10.1016/j.bbi.2009.08.003 19679179 85. Maier SF Bi-directional immune–brain communication: Implications for understanding stress, pain, and cognition Brain Behav Immun 2003 17 2 69 85 10.1016/S0889-1591(03)00032-1 12676570 86. Elenkov IJ Chrousos GP Stress System – Organization, Physiology and Immunoregulation Neuroimmunomodulation 2006 13 5-6 257 267 10.1159/000104853 17709947 87. Badyaev AV Stress-induced variation in evolution: from behavioural plasticity to genetic assimilation Proc R Soc Lond Ser B: Biol Sci 2005 272 1566 877 886 10.1098/rspb.2004.3045 88. Fang Z Spaeth AM Ma N Zhu S Hu S Goel N Detre JA Dinges DF Rao H Altered salience network connectivity predicts macronutrient intake after sleep deprivation Sci Rep 2015 5 8215 10.1038/srep08215 25645575 89. Markwald RR Melanson EL Smith MR Higgins J Perreault L Eckel RH Wright KP Impact of insufficient sleep on total daily energy expenditure, food intake, and weight gain Proc Natl Acad Sci USA 2013 110 14 5695 5700 10.1073/pnas.1216951110 23479616 90. Thimgan MS Gottschalk L Toedebusch C McLeland J Rechtschaffen A Gilliland-Roberts M Duntley SP Shaw PJ Cross-Translational Studies in Human and Drosophila Identify Markers of Sleep Loss PLoS One 2013 8 4 e61016 10.1371/journal.pone.0061016 23637783 91. Hauton C Smith VJ Adaptive immunity in invertebrates: A straw house without a mechanistic foundation Bioessays 2007 29 11 1138 1146 10.1002/bies.20650 17935208 92. Little TJ Colegrave N Sadd BM Schmid-Hempel P Studying immunity at the whole organism level Bioessays 2008 30 4 404 405 10.1002/bies.20738 18348252 93. Rowley AF Powell A Invertebrate Immune Systems–Specific, Quasi-Specific, or Nonspecific? J Immunol 2007 179 11 7209 7214 10.4049/jimmunol.179.11.7209 18025161

PMC005xxxxxx/PMC5002109.txt

==== Front BMC GenomicsBMC GenomicsBMC Genomics1471-2164BioMed Central London 302910.1186/s12864-016-3029-zResearch ArticleGlobal profiling of alternative RNA splicing events provides insights into molecular differences between various types of hepatocellular carcinoma Tremblay Marie-Pier 1Armero Victoria E. S. 1Allaire Andréa 1Boudreault Simon 1Martenon-Brodeur Camille 1Durand Mathieu 2Lapointe Elvy 2Thibault Philippe 2Tremblay-Létourneau Maude 1Perreault Jean-Pierre 1Scott Michelle S. 1http://orcid.org/0000-0001-7822-0933Bisaillon Martin Martin.Bisaillon@USherbrooke.ca 11 Département de biochimie, Pavillon de recherche appliquée sur le cancer, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, 3201 Jean-Mignault, Sherbrooke, QC J1E 4K8 Canada 2 Plateforme RNomique, Université de Sherbrooke, Sherbrooke, QC J1E 4K8 Canada 26 8 2016 26 8 2016 2016 17 1 68331 5 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 Dysregulations in alternative splicing (AS) patterns have been associated with many human diseases including cancer. In the present study, alterations to the global RNA splicing landscape of cellular genes were investigated in a large-scale screen from 377 liver tissue samples using high-throughput RNA sequencing data. Results Our study identifies modifications in the AS patterns of transcripts encoded by more than 2500 genes such as tumor suppressor genes, transcription factors, and kinases. These findings provide insights into the molecular differences between various types of hepatocellular carcinoma (HCC). Our analysis allowed the identification of 761 unique transcripts for which AS is misregulated in HBV-associated HCC, while 68 are unique to HCV-associated HCC, 54 to HBV&HCV-associated HCC, and 299 to virus-free HCC. Moreover, we demonstrate that the expression pattern of the RNA splicing factor hnRNPC in HCC tissues significantly correlates with patient survival. We also show that the expression of the HBx protein from HBV leads to modifications in the AS profiles of cellular genes. Finally, using RNA interference and a reverse transcription-PCR screening platform, we examined the implications of cellular proteins involved in the splicing of transcripts involved in apoptosis and demonstrate the potential contribution of these proteins in AS control. Conclusions This study provides the first comprehensive portrait of global changes in the RNA splicing signatures that occur in hepatocellular carcinoma. Moreover, these data allowed us to identify unique signatures of genes for which AS is misregulated in the different types of HCC. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-3029-z) contains supplementary material, which is available to authorized users. Keywords Liver cancerHepatitis B virusHepatitis C virusRNA splicingGene dysregulationCentre de recherche du Centre hospitalier universitaire de SherbrookeFaculté de médecine et des sciences de la santé de lUniversité de Sherbrookeissue-copyright-statement© The Author(s) 2016 ==== Body Background Hepatocellular carcinoma (HCC) represents the fifth most prevalent tumor type and the third leading cause of cancer-related deaths around the world [1]. Chronic infection with hepatitis B virus (HBV) and hepatitis C virus (HCV) are the major causes of HCC worldwide [2]. These two hepatotropic viruses represent the most important risk factors for the development of HCC, with an estimated 80 % of all HCC cases globally [3]. HBV and HCV induce carcinogenesis via different molecular mechanisms. HBV infection causes HCC via both indirect and direct pathways [4]. HBV infection promotes chronic injury to liver cells, with constant necro-inflammation and regeneration activity [5]. As a consequence, an increase in hepatocyte turnover is observed which leads to the accumulation of potential critical mutations with subsequent malignant transformation and clonal expansion, leading to HCC [6]. Moreover, HBV infection can directly lead to HCC by integrating the viral genomic DNA into the host genome [7]. HBV integration can lead to numerous mutagenic consequences, including deletions, translocation, large inverted duplications, and amplifications, resulting in chromosomal instability [8]. In contrast, HCV does not integrate into the hepatocyte genome. The central hypothesis for HCV carcinogenesis is that it occurs via indirect pathways through chronic inflammation and hepatocellular injury. This is supported by the fact that cirrhosis is frequently a pre-requisite for HCV-induced HCC [4]. However, HCC has been reported to develop in a very small proportion of non-cirrhotic patients, suggesting a more direct effect from HCV itself [9]. Accordingly, various HCV proteins have been shown to have a role in the development of HCC development in various experimental models [10]. Alterations in RNA splicing of a small number of cellular genes have been observed in HCC, and although still limited, the available data suggest that splicing defects may play a role in hepatocarcinogenesis [11]. Higher eukaryotes use alternative splicing (AS) to change the composition of transcripts encoded by a gene through selection of different exons to be included in the mature mRNA, thus producing a variability at the proteomic level [12]. AS therefore diversifies the cellular proteome because it results in multiple transcripts and, as a consequence, various proteins are created from a single gene. These different proteins produced from the same gene can frequently promote different or even opposite biological effects. As a result, the relative abundance of each isoform may have consequences on cellular functions. An example of this is clearly demonstrated by the BCL-x gene. AS of the Bcl-x transcript results in two different mature mRNAs: Bcl-x(L), which has anti-apoptotic effects, and Bcl-x(S) which promotes apoptosis [13]. Therefore, it is not surprising that AS is tightly regulated and that variations in splicing patterns have been associated with many human diseases including cancer [14]. During the last few years, cancer-specific splice variants and cancer-associated changes in the relative levels of spliced isoforms of genes with a recognized role in carcinogenesis have been observed [15]. AS alterations can confer selective advantages to tumors, such as angiogenesis, proliferation, cell invasion and avoidance of apoptosis [16]. Recent evidences indicate that some of these can be used as prognostic or diagnostic biomarkers, and the development of strategies to correct and/or inhibit pathological splicing events will be key in order to develop future therapeutic approaches [17]. Various studies have shown alterations in the AS patterns of a limited number of cellular genes in HCC [11]. One interesting case of dysregulated AS occurring in HCC is the overexpression of a splice variant of DNA methyltransferase 3b (DNMT3b), namely DNMT3b4 [18]. An elevation of the ratio of DNMT3b4 to DNMT3b3 mRNA is correlated with DNA hypomethylation on pericentromeric satellite regions which may induce chromosomal instability and is considered an early event during hepatocarcinogenesis [18]. Other examples of aberrantly-spliced genes detected in HCC include the serine/threonine kinase aurora kinase B (AURKB), the E3 ubiquitin ligase, the p53-antagonistic protein MDM2, the cell surface adhesion cadherin 17 protein, and the Hugl-1, Klf6, and p73 tumor suppressors [11]. These examples illustrate the oncogenic potential of aberrantly spliced isoforms in HCC tissues. In the present study, alterations to the global cellular AS landscape of more than 377 liver tissue samples from The Cancer Genome Atlas (TCGA) were investigated using high-throughput RNA sequencing data. This study provides a comprehensive portrait of global changes in the RNA splicing signatures that occur in HCC. We identify modifications in the AS patterns of transcripts encoded by more than 2500 genes such as transcription factors, tumor suppressor genes, kinases, and splicing factors. These findings allowed the identification of unique gene signatures for which AS is misregulated in different types of HCC. Results Modification of the cellular AS landscape in HCC Alterations to the global RNA splicing landscape of more than 377 liver tissue samples (Fig. 1a) from The Cancer Genome Atlas (TCGA) were investigated using high-throughput RNA sequencing data. The overview of all the analyses performed in this study is outlined in Fig. 1b. RNA-seq data were collected and mapped to the reference genome, followed by transcript assembly and analysis of RNA isoform abundance. We initially focused our study on HBV-associated HCC (THBV, Tumors with HBV) and HCV-associated HCC (THCV, Tumors with HCV), and we compared the global splicing patterns with normal tissues (NNoV, Normal, no virus). In order to identify the cellular AS patterns that are altered in HCC, we evaluated the modifications in splice abundances by quantifying all the primary transcripts that generate two or more isoforms. Alternative splicing events (ASEs) were detected and quantified using the percent-spliced-in (PSI) metric based on isoform expression (transcript-per-million, TPM) for the long and the short isoforms (see Experimental Procedures). The lists of ASEs (42,000–45,000 ASEs) were filtered to keep only data with at least two replicates for both types of HCC and normal hepatic tissue. The ASEs that differed between HCC tissues and normal hepatic tissue with a P-value of less than 0.05 were conserved. To ensure higher stringency, the ASEs were further filtered with a cutoff Q-value (false-discovery rate) of less than 0.05. From these events, only those with a difference higher than 10 % in |Delta PSI| were considered biologically relevant. Using such an approach, we identified 3250 primary transcripts belonging to 2051 genes for which the AS pattern was significantly modified in patients with HBV-associated HCC (Q ≤0.05, |Delta PSI| ≥10) (Fig. 1c). Moreover, the AS patterns of 1380 primary transcripts belonging to 907 genes were significantly modified in patients with HCV-associated HCC. The complete list of the differential ASEs associated with HBV- and HCV-associated HCC is provided in Additional file 1. The list includes six of the previously identified aberrantly spliced isoforms found in HCC in previous studies (DNMT3b, AURKB, UBE3B, MDM2, KLF6, and TP73). RNA-Seq data were not available for the two other previously identified examples of aberrantly-spliced genes detected in HCC (CDH17 and HUGL1).Fig. 1 Alternative splicing of HBV- and HCV-associated hepatocellular carcinoma. a Classification of the TCGA RNA sequencing data for HCC and healthy tissues. b Overview of the strategy used to identify the changes in splicing for each type of HCC. c The splicing events list for HBV-associated HCC (THBV/NNoV) and HCV-associated HCC (THCV/NNoV) was filtered to keep only significant ASEs Our analysis revealed that the expression of a vast number of cellular transcripts was modified in patients with HBV- and HCV-associated HCCs as compared to normal healthy tissues (Fig. 2a–b). In the case of HBV-associated HCC, 361 transcripts had |Delta PSI| values higher than 30 % (Additional file 1: Figure S1A), while 200 transcripts had |Delta PSI| values greater than 30 % for HCV-associated HCC. To identify cellular pathways for the differential ASEs associated with HBV- and HCV-associated HCC, we performed a functional enrichment analysis of the genes for which AS is modified. In both cases, the analysis revealed significantly enriched (Q ≤0.05) terms in both metabolic process (28.6 and 27.4 % for HBV- and HCV-associated HCC, respectively) and developmental process (6.6 and 7.1 % for HBV- and HCV-associated HCC, respectively) (Additional file 1: Figure S2A). Indeed, a significant number of the modified ASEs are encoded by genes with important roles in metabolism. For instance, alterations in the splicing patterns of BCAT2, which encodes a branched chain aminotransferase found in mitochondria, DAP3, which helps in protein synthesis within the mitochondrion, and ECHDC2, which is involved in fatty acid biosynthesis, were found to be present both in HBV- and HCV-associated HCC. A large number of these proteins appear to interact either directly or indirectly with each other (Additional file 1: Figure S2B). Through manual curation of functional annotations, we also identified numerous modifications in the AS patterns of genes potentially involved in oncogenic transformation. We identified many transcription factors, receptors, tumor suppressors, and kinases for which AS is significantly modified in HBV- and HCV-associated HCC (Additional file 1: Table S1). For example, HBV-associated HCC displayed alterations in the AS patterns of 138 tumor suppressors, 156 transcription factors, 73 kinases, and 132 receptors. Some of the dysregulated ASEs that were identified include KRAS, a member of the small GTPase superfamily implicated in various pathologies (such as lung adenocarcinoma, pancreatic cancer and colorectal carcinoma), MDM2, a nuclear-localized E3 ubiquitin ligase which can promote tumor formation by targeting tumor suppressor proteins (such as p53) for proteasomal degradation, and BRCA1 which encodes a nuclear phosphoprotein that plays a role in maintaining genomic stability, also acts as a tumor suppressor and is frequently mutated in breast and ovarian cancers. Interestingly, some of these modifications are found both in HBV- and HCV-associated HCC while others are unique to the respective virus-associated HCC (Fig. 3a–b and Additional file 1: Figure S3A–B). It should be noted that we found no evidence for correlation between the presence of one or more splice variants and clinical parameters (such as patient survival).Fig. 2 Global profiling of alternative splicing event modifications in HBV- and HCV-associated HCC. a Alternative splicing event modifications in HBV-associated HCC (THBV/NNoV) and HCV-associated HCC (THCV/NNoV) are presented with their associated changes in gene expression. The graphs represent the relation between the difference of splicing for each hepatitis-associated HCC compared with healthy non-viral tissues, and the difference in gene expression for each of these alternative splicing event modifications. b Heatmap representations of isoform ratios (PSI values) for all tissues analyzed in the current study. The first heat map shows PSI values for HBV-associated HCC. HBV-associated HCC tissues (THBV) are shown in red, and the comparative healthy tissues are shown in green (NNoV). The second heat map shows PSI values for HCV-associated HCC. HCV-associated HCC tissues (THCV) are shown in blue, and the comparative healthy tissues are shown in green (NNoV) Fig. 3 AS modifications in transcripts encoded by tumor suppressors and oncogenes in HBV- and HCV-associated HCC. a Alterations in the AS patterns of transcripts encoded by tumor suppressors in HBV- and HCV-associated HCC. Some transcripts have multiple ASEs that are modified. Red bars indicate negative Delta PSI values, and blue bars represent positive Delta PSI values. b Alterations in the AS patterns of transcripts encoded by oncogenes in HBV- and HCV-associated HCC. The tumor suppressors and oncogenes were selected based on the TSGene [64] and allOnco databases (http://www.bushmanlab.org/links/genelists) Characterization of the ASEs that are modified in HCC We next compared the profiles of the ASEs that are altered in HBV- and HCV-associated HCC. The vast majority of the ASEs that are modified in HCC were found at a level of one splicing event per gene (Additional file 1: Figure S1B). Analysis of the correlation between AS and gene expression indicated stable expression of most of the transcripts that are differentially spliced (Fig. 2a). Overall, we observed no clear association between the expression level and the cancer-specific splicing pattern, demonstrating that there is no relationship between the regulation of transcription and the control of alternative splicing in these cancer-specific genes. Among the transcripts for which AS was significantly affected in HBV- and HCV-associated HCC, many ASEs we documented affect known protein domains. Additional file 1: Table S2 displays the consequences of the differentially spliced transcripts on protein function for 60 transcripts that are differentially spliced in HCC. Among the differential ASEs, at least two ASEs resulted in the loss of predicted nuclear localization signals (NLS). These are RINT1, which plays a role in cell cycle checkpoint control and is essential for telomere length control, and EHD3, a paralog of EHD1, a protein which is thought to play a role in the endocytosis of IGF1 receptors. Twenty other ASEs were predicted to disrupt known functionally critical protein domains. Most of these genes encode for proteins involved in metabolic processes. For instance, some of the dysregulated ASEs lead to protein domain loss in HADHA. This gene encodes the alpha subunit of the mitochondrial trifunctional protein, an enzyme which catalyzes the final three steps of mitochondrial beta-oxidation of long chain fatty acids, and ALDH2, the second enzyme of the major oxidative pathway of alcohol metabolism. Experimental validation on HCC tissues We next sought to experimentally validate the results that were obtained from the RNA-Seq data. We used PCR analysis to experimentally validate the differential ASEs that were observed through RNA-seq studies. Using liver cDNA arrays from both healthy and cancerous tissues, we designed specific primers to allow detection of the various ASEs by PCR. Examples of differentially spliced transcripts are presented in Fig. 4a illustrating the modifications in isoform usage in transcripts encoded by three different genes (OSBPL6, VPS13A, ZNF692). In all three cases, the experimental Delta PSI values were similar to the values obtained from RNA-Seq analysis (Fig. 4b). The experimental Delta PSI values ranged from −17.9 (for the VPS13A transcript) to 15.6 (for the ZNF692 transcript). Overall, our results demonstrated that changes in AS levels revealed through transcriptome sequencing could also be detected by PCR analysis (Fig. 4b).Fig. 4 Validation of ASEs dysregulated in HCC. a Overview of the two isoforms encoded by OSBPL6, VPS13A and ZNF692 genes. Exons are depicted in red and the intervening introns are shown as thin black lines (not to scale). The primers used to detect the isoforms by RT-PCR assays are shown in gray and the sizes of the expected amplicons are also indicated. b cDNA obtained from tissues were analyzed by PCR using specific primers to detect both isoforms of the transcripts encoded by the OSBPL6, VPS13A and ZNF692 genes. Capillary electrophoregrams of the PCR reactions are shown. The positions and the amplitude of the detected amplicons are highlighted by red boxes. The positions of the internal markers are also indicated. The Delta PSI values for both the RNA-Seq and experimental validation assays are shown on the right Comparison between different types of HCC The large size of the samples analyzed by high-throughput RNA sequencing allowed us to obtain data for the alterations to the global RNA splicing landscape of two other types of HCC, namely HCC associated with the co-presence of both HBV and HCV (THBV&HCV) and virus-free HCC (TNoV, Tumor no virus). Because of their shared modes of transmission, co-infection of HBV and HCV is not uncommon, particularly in countries with a high prevalence of HBV or HCV [19]. HBV and HCV co-infection results in more severe liver pathologies and in an enlarged risk of HCC than monoinfection [19]. While the number of HCC tissues associated with the presence of both HBV/HCV was relatively small in our samples collection (n = 3), our statistical analysis allowed us to identify 220 transcripts for which AS is modified in comparison to normal tissues (Additional file 1: Figures S1C, S4A, Tables S3–S6). Remarkably, 79 genes for which AS was modified were found to be common between the three types of virus-associated HCC (Fig. 5a). This raised the possibility that these common ASEs might be related to a general characteristic of cancerous tissues. We therefore analyzed the cellular AS patterns that are altered in virus-free HCC of 131 patients (TNoV; tumor, no virus), and we evaluated the changes in relative splice abundances. Using such an approach, we identified 1517 genes for which the AS pattern was significantly modified in virus-free HCC (Q ≤0.05, |Delta PSI| ≥10) (Additional file 1: Figures S1C, S4B, Tables S3-S6). Further analysis revealed that 75 of the 79 previously identified genes for which AS was modified in the three types of virus-associated HCC were also identified in virus-free HCC, thereby indicating that the modifications in AS found in those genes are associated with carcinogenic tissues. It should be noted that the AS of the four remaining genes were also found to be modified in virus-free HCC although they were not retained following the statistical analysis i.e. the Q-values were slightly higher than 0.05 and/or the |Delta PSI| were slightly lower than 10. Overall, our analysis allowed us to identify 761 unique transcripts for which AS is misregulated in HBV-associated HCC, while 68 are specific to HCV-associated HCC, 54 to HBV&HCV-associated HCC, and 299 to virus-free HCC (Fig. 5b, Additional file 1: Tables S7–S11).Fig. 5 Comparison between HBV-, HCV-, HBV&HCV-, and non-viral-associated HCC. a The list displays the common differentially spliced transcripts for the various types of HCC with the corresponding Delta PSI values, the associated gene expression (in Log2), and the associated biological processes. b Comparison of the genes with dysregulated ASEs between HBV-, HCV-, HBV&HCV-, and virus-free HCC Expression levels and AS of RNA splicing factors The precise mechanism which leads to a modification of the AS landscape in HCC is unknown at the moment. Changes in splice site choice generally arise from variations in the assembly of the spliceosome or by altering the binding of splicing factors to the RNA transcripts [20]. Although splicing is controlled by a large set of splicing factors, dysregulated expression of individual splicing factors has been shown to frequently result in aberrant splicing [21]. Taking these findings into consideration, we thus monitored both the expression levels and the changes in splicing patterns of RNA encoding for splicing factors and spliceosomal proteins. As shown in Fig. 6a, the expression of some splicing factors and spliceosomal proteins is indeed affected in HCC. The expression of 94 proteins involved in splicing was modulated by more than 2-fold in HBV-associated HCC. Interestingly, the expression of ESRP1, a splicing factor known to regulate diverse types of alternative splicing events [22], was increased by more than 21-fold in HBV-associated HCC. Very similar expression profiles were also found in HCV-, HBV&HCV-associated HCC and virus-free HCC (Additional file 1: Figure S5). Immunonohistochemical staining confirmed the over-expression of various splicing factors in HCC tissues, as compared to normal liver tissues, including ESRP1, CWC27, and DDX41 (Fig. 6b).Fig. 6 RNA splicing factors in HCC. a Iris Graph representing the expression profile of splicing factors for HBV-associated HCC. Differences in gene expression levels are shown on a logarithmic color scale (Log2), from red (negative changes in expression) to blue (increase in gene expression). b Expression of splicing factors in both normal liver tissues and hepatocellular carcinomas. These images were extracted from the Human Protein Atlas database, according to its academic usage permission (see Ref. [65] and www.proteinatlas.org). The images show the results of immunohistochemical staining using specific antibodies (complete list and experimental protocol found at www.proteinatlas.org), followed by detection with horseradish peroxidase. Immunohistochemical staining was performed on both normal and HCC tissues for three different splicing factors (ESRP1, CWC27, and DDX41). c Misregulation of splicing factors alternative splicing in HBV- (THBV/NNoV, yellow), HCV- (THCV/NNoV, blue), HBV&HCV- (THBV&HCV/NNoV, green) and virus-free HCC (TNoV/NNoV, red). The delta PSI values are represented in red (negative delta PSI values) and in blue (positive delta PSI values). d Expression levels of hnRNPC in patients with or without tumors (left panel). Kaplan–Meier overall survival curve (right panel) for HCC patients expressing high (red) or low (blue) levels of hnRNPC. HNRNPC transcript level was negatively correlated with overall survival Modifications to the AS patterns of transcripts encoding proteins involved in splicing (splicing factors and proteins of the spliceosome) was also investigated. Our study identified many splicing factors that were differentially spliced upon in HCC (Fig. 6c). In the case of HBV-associated HCC, the AS patterns of 26 splicing factors were significantly modified. One example is RBFOX2, which encodes an RNA-binding protein that is thought to be a key regulator of alternative exon splicing. Other examples include CELF2, which regulates pre-mRNA alternative splicing and may also be involved in mRNA editing, and translation, and MBNL1, which can also regulate splicing on specific pre-mRNA targets. Modifications to both the expression level and AS of splicing factors likely contribute to the observed changes in the cellular AS landscape of HCC. We further investigated the implication of splicing factor levels on AS by using RNA interference and a reverse transcription-PCR screening platform to examine the roles of splicing factors in different cell lines [23]. We selected various splicing factors which were over- or under-expressed in HBV- and HCV-associated HCC. These splicing factors were each targeted with specific siRNAs, and we carried out a loss-of-function study in various cell lines (PC-3, SKOV3, NIH:OVCAR-3, MDA-MB-231, MCF7). In the case of HBV-associated HCC, specific siRNAs were designed against splicing factors U2AF2, SF3A2, RBM8A, RBM4, PRPF4B, NOVA1, KHSRP, HNRPU, HNRPL, HNRPH1, HNRPC, and HNRPA1. Depletion efficiencies were evaluated 96 h posttransfection by Western blotting and/or quantitative RT-PCR assays. All assays indicated that depletion had been achieved (data not shown). We selected 96 transcripts that belong to a subset of apoptotic genes because the functional consequences of alternative splicing on apoptosis have previously been clearly established [24]. Five of these transcripts (APP, AXIN1, F3, FGFR4, and POLB) were of particular interest since their AS was altered in HBV-associated HCC. The results are depicted in Fig. 7 and show that the impact of the individual knockdowns varies considerably. For instance, the knockdown of SF3A2 has a definitive impact on the AS pattern of the BCL2L1, C11ORF4, DRF1, FANCA, FN1.3, POLB, PTPN13, SHC1, and TNFRSF10B transcripts in all cell lines tested. However, the same knockdown has no impact on the AS profile of numerous transcripts such as BCMP11, CCL4, CTNNA1, FGFR1, FGFR2, FN1.2, GAPD, GATA3, GNB3, HNRPAB, HSC20, PAXIP1, PTK2, PTK2B, RAD52, RSN, SHMT1, STM1, and TLK1 transcripts. We conclude that the expression of individual splicing factors has an impact on AS profiles, but cannot explain all the changes observed. The impact of the individual knockdowns on AS was also analyzed for splicing factors which were over- or under-expressed in HCV-associated HCC (Additional file 1: Figure S6). As observed previously [23], our results demonstrate that targets for individual RNA splicing factors can vary in different cellular contexts.Fig. 7 Modifications to AS of 96 transcripts in response to knockdown of splicing factors with specific siRNAs. Using specific siRNAs, twelve splicing factors (U2AF2, SF3A2, RBM8A, RBM4, PRPF4B, NOVA1, KHSRP, HNRPU, HNRPL, HNRPH1, HNRPC, and HNRPA1) were individually knocked-down in different cell lines to assess their implication in splicing of 96 differents transcripts. Individual knockdowns and ASEs are shown to indicate which knockdowns caused a shift in alternative splicing in various cell lines (PC-3, SKOV3, NIH:OVCAR-3, MDA-MB-231, MCF7). Each column represents a distinct knockdown performed with specific siRNAs. The changes in PSI values are displayed. The map depicts the changes in PSI values in a color-coded scale. White areas indicate no shifts. Asterisks indicate transcripts for which AS was altered in HBV-associated HCC Expression pattern of hnRNPC significantly correlates with patient survival We next investigated the relationship between the mRNA expression levels of proteins involved in RNA splicing and patient survival. A substantial number of HCC biomarkers with potential prognostic significance have been identified in recent decades and have shown promising results [25]. Kaplan-Meier analysis was performed to evaluate overall survival for patients as a function of the mRNA expression levels of proteins involved in splicing. Of all the proteins involved in splicing for which the expression level was modulated by more than 2-fold in HCC, the expression level of hnRNPC, a regulator of mRNA splicing, was the only one showing correlation with patient survival (Fig. 6d). A high level of expression of hnRNPC was associated with reduced patient survival (p < 0.02 by Cox regression assay). No significant correlations were found between high-levels of hnRNPC expression and clinicopathological features, including age, gender, and tumor stage. Interestingly, the previous depletion assays (Fig. 7) indicated that a decrease in hnRNP C expression affected 20 % (19/96) of the ASEs tested in at least two cell lines, thereby demonstrating the importance of this protein in the regulation of AS. Expression of HBx alters cellular AS Although changes in splicing patterns are features of cancer development/progression, it whether these changes are the source or the consequence of a malignant phenotype remains unknown. We therefore investigated the attractive possibility that the presence of a viral protein could contribute to modifications in AS of virus-associated HCC. We focused on the HBx protein of HBV since currently available evidence supports a role for this protein in the pathogenesis of HBV-associated HCC. HBx has been shown to promote cell cycle progression, deactivate negative growth regulators, and prevent the expression of tumor suppressor genes and senescence-related factors [26]. The HBx protein (harboring a HA epitope) was expressed in HEK293T cells and detected by SDS-PAGE followed with anti-HA immunoblotting (Fig. 8a). HEK293T cells were chosen because of their high transfection efficiency and their extensive use in RNA splicing studies [27–30]. We next relied on a high-throughput reverse transcription-PCR (RT-PCR)-based platform for splicing annotation, to examine AS in 1200 cancer-associated genes [31]. Once again, the relative abundance of ASEs was expressed as PSI values. Our results demonstrated that the AS patterns of 56 cancer-related genes which were modified in HBV-associated HCC were also significantly modified upon HBx expression (Fig. 8b). An example of modified AS pattern modified upon HBx expression is provided in Fig. 8c.Fig. 8 Involvement of the HBx protein from HBV in AS. a Immunoblotting analysis using anti-HA antibody for the detection of HBx-HA protein from the cell lysates isolated after 5 days of transfection with pLenti6V5A-HBx. Control untransfected cells (T(−)) were also used in this assay. b List of ASEs common to HBV-associated HCC and HBx-expressing cells. c Example of an ASE modified following the expression of HBx. Overview of the two isoforms encoded by ENO3 gene. Exons are depicted in red and the intervening introns are shown as thin black lines (not to scale). The primers used to detect the isoforms by RT-PCR assays are shown in gray and the sizes of the expected amplicons are also indicated (top panel). RT-PCR reactions were performed on control cells (T(−)) and cells expressing HBx (HBx) using specific primers to detect both isoforms of the transcripts encoded by the ENO3 gene. Capillary electrophoresis was performed and an image of the detected reaction products is shown (lower panel). The positions of the expected amplicons are indicated by arrows. d Mass spectrometry analysis of cellular proteins interacting with HBx. The average H/L ratios of the HBx affinity purification-mass spectrometry experiments were plotted versus the total intensities. The red dots indicate cellular proteins involved in RNA splicing The observation that the expression of HBx leads to modifications in the AS profile of cellular genes led us to investigate the ability of HBx to interact with cellular proteins involved in splicing. We used a quantitative proteomics-based approach relying on stable isotope labeling of amino acids in cell culture (SILAC) coupled to mass spectrometry (LC-MS/MS) to identify cellular proteins that interact with HBx. We used SILAC based on three different sets of conditions in order to detect cellular proteins that bind to HBx in the absence of RNA/DNA (see Methods). Cell lysates were incubated with anti-HA antibodies and bound proteins were precipitated, washed, eluted, trypsinized, and subjected to MS/MS. Potential contaminants and non-specific binders were considered using the CRAPOME database (Contaminant Repository for Affinity Purification and Mass Spectrometry) [32]. Using such a strategy, our quantitative mass spectrometry analysis demonstrated that SF1, a splicing factor which is necessary in the early phases of spliceosome assembly, was highly associated with HBx (Fig. 8d). No other cellular protein involved in RNA splicing was strongly associated with HBx. The role of this interaction in the regulation of AS in HBV-associated HCC will need to be further investigated in future studies. Discussion In the current study, we present the first large-scale screen of HCC–associated ASEs. The study demonstrated that the AS patterns of numerous cellular transcripts are modified in HCC. Moreover, these data allowed us to identify unique signatures of genes for which AS is misregulated in the different types of HCC. Analysis of the AS landscape of the various types of HCC revealed numerous potential HCC–specific markers, which significantly increases the number of biomarkers that can presently be identified by current expression profiling approaches. We believe that our study could lead to the identification of molecular candidates for diagnostics and/or targetable pathways in HCC. Currently, α-fetoprotein (AFP) together with pathology and iconography detection are routinely used in early clinical diagnosis for liver cancer [33]. However, the widely used marker AFP does not always yield satisfactory results in the early diagnosis of HCC, particularly in the case of AFP-negative HCC, thereby limiting the universality of its application [33]. The use of other HCC markers (such as GPC3 and TGF-β1) is presently being explored [34], which is expected to improve early diagnostic rate. Undoubtedly, future developments in molecular genetics and proteomic analysis should lead to the identification of other useful HCC-specific markers. The current identification of unique signatures for genes in which AS is misregulated in the different types of HCC constitutes a step toward that goal. In order to meet the augmented requirements of proliferation, cancer cells frequently display important changes in pathways of energy metabolism and nutrient uptake [35]. Oncogenic transformation often results in an increase in both bioenergetic potential and the nutrient uptake in a cell-autonomous fashion [36]. The fact that metabolic enzymes such as succinate dehydrogenase (SDH) and fumarase (FH) function as tumor suppressors in human cells suggests that metabolic dysregulation can be an initiating event in cancer [36]. The recent discovery of activating mutations in isocitrate dehydrogenase (IDH1) in glioblastoma demonstrates that an activating mutation in a metabolic enzyme are selected during carcinogenesis [37]. The emergence of metabolic enzymes acting as important regulators of cancer cell growth highly suggests that metabolic control is a critical factor in carcinogenesis. In the present study, many transcripts that were differentially spliced in HCC samples are known to play critical roles in cell metabolism. For example, the expression of the full-length ALDH2 transcript was found to be significantly reduced and replaced by a shorter form that is predicted to encode a truncated Aldh2 protein with a loss of a portion of the aldehyde dehydrogenase domain. Aldehyde dehydrogenase is a key enzyme involved in the major oxidative pathway of alcohol metabolism. Interestingly, the enlarged exposure to acetaldehyde in patients with a catalytically inactive form of Aldh2 has previously been shown to confer an increase susceptibility to many types of cancer, including esophageal cancer [38]. Another example of a gene involved in metabolism for which AS is modified in HCC is HADHA, which encodes the alpha subunit of the mitochondrial trifunctional protein catalyzing the 3-hydroxyacyl-CoA dehydrogenase and enoyl-CoA hydratase activities. These steps are involved in mitochondrial beta-oxidation of long chain fatty acids. Our results showed that the expression of the full-length HADHA transcript was found to be significantly reduced and replaced by a shorter form that is predicted to encode a truncated Hadha protein with a loss of a portion of the enoyl-CoA domain. Interestingly, a link between HADHA and cancer has previously been observed in breast cancer where the HADHA gene was significantly under-expressed in cancerous tissues, especially in tumors with estrogen receptor-negative status [39]. In comparison with normal tissues, a substantial number of RNA splicing factors have been found to be upregulated or downregulated in various cancers. As changes in the concentration of these factors have been shown to modify the selection of splice sites [40], it is predicted that the abnormally expressed splicing factors found in cancerous cells induce the production of mRNA isoforms that are either nonexistent or less abundant in normal cells. A major challenge will be to determine if these modifications contribute directly to carcinogenesis, or whether they constitute one of the many processes that are modified in cancer cells. The demonstration that overexpression of the Serine/Arginine-Rich Splicing Factor 1 (SRSF1, also known as SF2/ASF) can trigger malignant transformation suggests that the abnormal expression of proteins involved in splicing can contribute to carcinogenesis [41]. In addition, the expression SRSF1 has been shown to be increased in several tumor types [42] and fibroblasts overexpressing SRSF1 caused tumor formation when injected into mice [41]. In the current study, many changes in the expression levels and/or splicing patterns of splicing factors were detected in HCC samples. Interestingly, our data demonstrated that the transcript encoded by ESRP1 was found to be significantly over-expressed in all types of HCC investigated (32-fold increase in HBV-associated HCC). How the altered expression dynamics of ESRP1 (and other splicing regulators) contribute to AS homeostasis remains to be examined. Interestingly, a previous report identified ASEs regulated by Esrp1 using RNA silencing technology in a human epithelial cell line [43]. Using such a strategy, the authors identified 148 alternative splicing events in a total of 134 different genes that were regulated by Esrp1 [43]. In the present study, the splicing of 32 of these Esrp1-regulated transcripts are also found to be differentially spliced in HBV-associated HCC suggesting, at least in part, that some of the observed splicing alterations observed in HCC could be related to Esrp1 over-expression. In the case of HCV-associated HCC and virus-free HCC, the splicing of 17 and 20 of these Esrp1-regulated transcripts were also found to be differentially spliced, respectively. AS has been shown to add an additional layer of regulation that functions in concert with transcriptional alterations during the epithelial-to-mesenchymal transition (EMT) [43]. EMT is an important and reversible process during the development/differentiation of multiple organs and has been implicated in cancer progression and metastasis [44]. It involves reorganization of the actin cytoskeleton, loss of cell–cell adhesion, the acquisition of increased cell motility, and changes in the transcriptional profile of epithelial cells (such as downregulation of E-cadherin and upregulation of vimentin and N-cadherin [45]. Moreover, previous studies have shown that a downregulation of epithelial-specific splicing regulators (ESRP1 and ESRP2) occurs in cells which undergo EMT [46]. Down-regulation of ESRP1 and ESRP2 has also been closely associated with a motile phenotype of cancer cells [47]. EMT is a crucial event in HCC progression which causes an increase in malignancy of hepatocytes associated with tumor cell invasion and metastasis [48]. For HCC, intrahepatic metastasis is most frequent as the hepatic microenvironment displays a dense vasculature and is the competent milieu for HCC cells in the metastasis cascade [48]. In the current study, expression of ESRP1 was significantly increased in HCC as evidenced from both RNA-Seq and immunohistochemical analyses. This finding was unexpected since downregulation of epithelial-specific splicing regulators is usually associated with cells undergoing EMT [46]. The exact significance of this result is not understood at the moment. In HCC, EMT likely occurs at the leading edge of tumor tissues under the particular control of extrinsic factors derived from the tumor microenvironment in a paracrine fashion [46]. As such, EMT is traditionally associated with highly metastatic HCC. Moreover, it should be noted that both the epithelial and mesenchymal cell phenotypes are not static. Indeed, both cell types have the capacity to undergo important cellular transformations (such as EMT and MET: mesenchymal-to-epithelial transition) that can alter both cellular morphology and behavior [43]. The MET-EMT cycle, along with differences between primary HCC and metastatic HCC, might also explain why no significant evidence for gene over-expression of traditional mesenchymal markers (such as vimentin or N-cadherin) or repression of epithelial ones (e.g., E-cadherin) was found in tissues from HCC patients in the TCGA database. Tumor markers based on the expression of specific proteins that can be used in the prognosis of HCC have been identified during the past few years, although consensus has still not been reached [25]. More recently, an important number of profiles based on the expression levels of mRNAs have been used as predictors of prognosis in various cancers [49]. For instance, the mRNAs levels of hnRNP K, a member of the hnRNP family of proteins which are involved in transcription, RNA splicing, and translation, has previously been shown to be aberrantly increased in numerous cancers [50, 51]. It was reported that high-level hnRNP K expression is correlated with poorer overall survival among patients with nasopharyngeal carcinoma (NPC) and prostate cancer [50, 51]. High levels of expression of another member of the family, namely hnRNP D, has also been linked to reduced survival in oral cancer [52]. In the current study, a high level of mRNA expression of hnRNPC was also associated with reduced patient survival in HCC. Clearly, additional studies will need to be performed to see if such an observation can eventually have any clinical application. Nonetheless, validation of potential tumor markers such as hnRNPC will continue to be a major focus of HCC research in the next few years. Indeed, improvements in early diagnosis are still needed since only 30 % of patients with HCC are candidates for potentially curative treatments at the time of diagnosis [53]. We believe that some RNA splicing isoforms markers could eventually be used as biomarkers. Ultimately, these could serve as diagnostic or prognostic tool by detecting their presence in cancer cells or in the fluids of patients. Large-scale studies are starting to reveal the extent of modifications that occur in AS in different types of cancer [14]. Although the precise mechanism which leads to a modification of the AS landscape in HCC is unknown at the moment, we demonstrated that the expression of the HBV HBx protein can result in modifications in the AS profiles of cellular genes. Our results showed that the AS patterns of 56 cancer-related genes which were modified in HBV-associated HCC were also significantly altered upon HBx expression, thereby suggesting a potential mechanism by which a viral infection can alter cellular AS profiles. Many direct and/or indirect molecular mechanisms by which the expression of a viral protein such as HBx can lead to a modification of the cellular AS landscape can be envisaged. However, the demonstration that HBx can interact with splicing factor SF1 is clearly a step towards identifying the steps required for the extensive modifications of AS observed in HCC. Interestingly, strategies to modulate AS by splice-switching oligonucleotides in order to correct aberrant ASEs, and/or to induce expression of therapeutic splice variants are now being developed [17, 54]. It is tempting to speculate that such a strategy could be applied to HCC. The current identification of extensive changes in the cellular AS landscape in HCC likely represents a first step toward the development of anticancerous agents based on the AS modifications identified in HCC. The development of highly specific molecular tools, such as splice-switching oligonucleotides, which could precisely alter the proportion of splice variants will also be critical to assess the function of these splice variants. Conclusions The present analysis provided a comprehensive portrait of global changes in the RNA splicing signatures that occur in HCC. These data also allowed the identification of unique signatures of genes for which AS is misregulated in different types of HCC which could lead to the identification of new molecular candidates for diagnostics and/or targetable pathways in HCC. Methods RNA-seq data analysis RNA-Seq samples were provided by the CGHub data portal (https://cghub.ucsc.edu/). As only BAM files were available at the time the data was obtained, a custom script was used to extract read information from the given alignment files and generate valid FASTQ files for re-alignment (two FASTQ files with the reads in the same order such that the aligner can align them in pairs). Sequence reads were aligned on a transcriptome reference sequence database (UCSCGene Hg19 using Bowtie v2 aligner (default parameters) and associated gene isoforms quantified in transcript-per-million (TPM) using RSEM for each sequenced sample [55]. Alternative splicing events were automatically detected and quantified using the percent-spliced-in (PSI, Ψ) metric based on isoform expression TPM for the long (L) and the short (S) forms using equation below: Ψ=LL+S Genes with one single isoform or no Human Genome Organisation (HUGO) ID were not considered for further analysis. Ethics, consent, and permissions Not applicable. RNA-Seq data were obtained from the TCGA Portal (https://tcga-data.nci.nih.gov/tcga/). Gene expression analysis The gene list was initially filtered to keep only data present in at least two replicates for both virus-associated HCC and normal hepatic tissue. The amount of transcript in each sample was calculated in transcripts per million (TPM) from the transcript-estimated read counts which were provided by TCGA and the isoform lengths from the UCSC (June 2011) annotation. To ensure higher reproducibility, only genes with expression levels higher than one TPM in either dataset were conserved. Fold changes in base 2 logarithm were then calculated between both virus-associated HCC and normal hepatic tissue average TPM. Q-values were calculated to take into account multiple statistical hypothesis testing and results under 0.05 were considered significant. Alternative splicing analysis The alternative splicing event (ASE) list was filtered to keep only data present in at least two replicates for both virus-associated HCC and normal hepatic tissue. Events with a P-value less than 0.05 were conserved. To ensure higher stringency, the ASEs were further filtered with a cutoff Q-value of less than 0.05. From these events, only those with a difference greater than 10 % in PSI were considered biologically relevant. Statistical analysis Welch’s t-test (Student’s t-test with unequal sample sizes and unequal variances) was calculated through the GSL library (http://www.gnu.org/software/gsl/) integrated to Perl system analysis for gene expression and alternative splicing data. Also, false discovery rates were calculated with the Q-value package in R (https://cran.r-project.org/src/contrib/Archive/qvalue/) based on Storey and Tibshirani [56]. For all other analyses, Graph Pad Prism version 6.05 was used to run statistical analysis. Gene ontology analysis Gene ontology analysis with PANTHER was performed using the database for annotation, visualization, and integrated discovery (DAVID) [57]. String networks Using the STRING database (version 10) [58], genes were submitted for generation of protein-protein interaction network from the Homo sapiens interactome. High-resolution evidence views were created and saved. Functional ASE prediction Using the FAST-BD or EASANA suite, the splicing patterns of genes of interest were visualized. DNA sequences of representative transcripts presenting short and long isoforms were downloaded and translated into proteins using ExPASy translation tool [59]. Predicted proteins were then compared using Multalin (truncation and frameshift event) [60], PFAM (loss or appearance of a functional domain) [61], and NLS Mapper (loss or gain of nuclear localization signal) [62]. PCR validation Liver Cancer cDNA Arrays form Origene TissueScan plates (Rockville, MD) were assessed for the expression of the various transcripts using the manufacturer’s protocol. The plates contained cDNAs from 8 normal and 39 liver cancer tissues. All forward and reverse primers were individually resuspended to 20–100 μM in Tris-EDTA buffer (IDT) and diluted as a primer pair to 1 μM in RNase DNase-free water (IDT). PCR reactions were performed in 10 μl in 96 well plates on a CFX-96 thermocycler (BioRad). The following cycling conditions were used: 3 min at 95 °C; 50 cycles: 15 s at 95 °C, 30 s at 60 °C, 30 s at 72 °C. For every PCR run, control reactions performed in the absence of template were performed for each primer pair and these were consistently negative. The amplified products were analyzed by automated chip-based microcapillary electrophoresis on Caliper LC-90 instruments (Caliper LifeSciences). Amplicon sizing and relative quantitation were performed by the manufacturer’s software. Splicing factors knockdown siRNAs, Western blot analysis, RT-PCR assays and bioinformatic analysis were performed as described before [23]. Immunoprecipitation and mass spectrometry HEK293T cells were grown in light (DMEM-R0K0), medium (DMEM-R6K4) or heavy media (DMEM-R10K8) containing the heavy isotopes of arginine and lysine supplemented with 10 % fetal bovine serum. Cells were transfected with pLenti6V5A-HBx to express the HBV HBx protein for five days. Untransfected control cells were also grown for 5 days. Cells were harvested separately, washed three times with PBS, and treated (heavy) or not (light and medium) with 20 mg/ml RNase A (Invitrogen) and 200 units DNase RQ1 (Promega) for two hours. Immunoprecipitation was performed using HA beads (Roche 11815016001) following the manufacturer’s instructions. LC-MS/MS was performed as described previously [63]. Additional file Additional file 1: Table S1. Protein families encoded by transcripts that are differentially spliced in various types of HCC. Table S2. Bioinformatical prediction of functional changes caused by some of ASEs identified. Table S3. List of tumor suppressors for which AS is dysregulated in various types of HCC. Table S4. List of oncogenes for which AS is dysregulated in various types of HCC. Table S5. List of kinases for which AS is dysregulated in various types of HCC. Table S6. List of transcription factors for which AS is dysregulated in various types of HCC. Table S7. List of genes for which AS is dysregulated in all types of HCC. Table S8. List of genes uniquely dysregulated in HBV-associated HCC. Table S9. List of genes uniquely dysregulated in HCV-associated HCC. Table S10. List of genes uniquely dysregulated in HBV&HCV-associated HCC. Table S11. List of genes uniquely dysregulated in virus-free HCC. Figure S1. Characterization of splicing mysregulation in HCC. Figure S2. Characterization of ASEs that are modified in HBV- and HCV-associated HCC. Figure S3. AS modifications in transcripts encoded by kinases and transcriptions factores in HBV- and HCV-associated HCC. Figure S4. Global profiling of ASE modifications in both HBV&HCV-associated HCC and virus-free-associated HCC. Figure S5. RNA splicing factors in HCC. Figure S6. Modifications to AS of 96 transcripts in response to knockdown of splicing factors with specific siRNAs (PDF 6675 kb) Abbreviations ASAlternative splicing ASEAlternative splicing event HBVHepatitis B virus HCCHepatocellular carcinoma HCVHepatitis C virus NNoVNormal tissues, no virus PSIPercent-spliced-in TCGAThe Cancer Genome Atlas THBVTumors with HBV THBV&HCVTumors with both HBV and HCV THCVTumors with HCV TNoVVirus-free gastric hepatic tissues TPMTranscript-per-million Acknowledgements The authors would like to thank François-Michel Boisvert and Dominique Lévesque for mass spectrometry experiments and analysis. Funding This work was supported by grants from the Centre de recherche du Centre hospitalier universitaire de Sherbrooke (M.B. and J.P.P) and the Faculté de médecine et des sciences de la santé de l’Université de Sherbrooke (M.B. and J.P.P). J.P.P holds the Université de Sherbrooke Research Chair in Structure and Genomic RNA. M.B. is a Chercheur-boursier Sénior from the Fonds de recherche du Québec-Santé (FRQ-S). M.P.T. holds a graduate training award from the Faculté de médecine et des sciences de la santé de l’Université de Sherbrooke. Availability of data and materials Primary and processed data are deposited at the Data Coordinating Center (https://tcga-data.nci.nih.gov/tcga/tcgaDownload.jsp); primary sequence files are deposited in CGHub (https://cghub.ucsc.edu/). Authors’ contributions MPT contributed to experimental design and analysis of the data, performed validation and co-immunoprecipitation experiments, and helped in writing the manuscript; VESA AA, SB, and CMB assisted in data analysis and the generation of figures; EL and MD performed PCR validations; PT performed informatics analyses; MTL, MS, and JPP contributed to experimental design and analysis; MB contributed to experimental design, data analysis and wrote the manuscript with contributions from all the authors. All authors approved the 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. RNA-Seq data were obtained from the TCGA Portal (https://tcga-data.nci.nih.gov/tcga/). ==== Refs References 1. Rahib L Smith BD Aizenberg R Rosenzweig AB Fleshman JM Matrisian LM Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States Cancer Res 2014 74 2913 21 10.1158/0008-5472.CAN-14-0155 24840647 2. El-Serag HB Epidemiology of viral hepatitis and hepatocellular carcinoma Gastroenterology 2012 142 1264 1273 10.1053/j.gastro.2011.12.061 22537432 3. Tsai WL Chung RT Viral hepatocarcinogenesis Oncogene 2010 29 2309 24 10.1038/onc.2010.36 20228847 4. Fung J Lai CL Yuen MF Hepatitis B and C virus-related carcinogenesis Clin Microbiol Infect 2009 15 964 70 10.1111/j.1469-0691.2009.03035.x 19874379 5. Chisari FV Isogawa M Wieland SF Pathogenesis of hepatitis B virus infection Pathol Biol 2010 58 258 66 10.1016/j.patbio.2009.11.001 20116937 6. Rogler CE Chisari FV Cellular and molecular mechanisms of hepatocarcinogenesis Sem Liver Dis 1992 12 265 78 10.1055/s-2007-1007398 7. Murakami Y Saigo K Takashima H Minami M Okanoue T Bréchot C Paterlini-Bréchot P Large scaled analysis of hepatitis B virus (HBV) DNA integration in HBV related hepatocellular carcinomas Gut 2005 54 1162 68 10.1136/gut.2004.054452 16009689 8. Jiang Z Jhunjhunwala S Liu J Haverty PM Kennemer MI Guan Y Lee W Carnevali P Stinson J Johnson S The effects of hepatitis B virus integration into the genomes of hepatocellular carcinoma patients Genome Res 2012 22 593 601 10.1101/gr.133926.111 22267523 9. Bralet MP Régimbeau JM Pineau P Dubois S Loas G Degos F Valla D Belghiti J Degott C Terris B Hepatocellular carcinoma occurring in nonfbrotic liver: epidemiologic and histopathologic analysis of 80 French cases Hepatology 2000 32 200 204 10.1053/jhep.2000.9033 10915724 10. Heindryckx F Colle I Vlierberghe HV Experimental mouse models for hepatocellular carcinoma research Int J Exp Pathol 2009 90 367 86 10.1111/j.1365-2613.2009.00656.x 19659896 11. Berasain C Goñi S Castillo J Latasa MU Prieto J Ávila MA Impairment of pre-mRNA splicing in liver disease: Mechanisms and consequences World J Gastroenterol 2010 16 3091 3102 10.3748/wjg.v16.i25.3091 20593494 12. Black DL Mechanisms of alternative pre-messenger RNA splicing Annu Rev Biochem 2003 72 291 336 10.1146/annurev.biochem.72.121801.161720 12626338 13. Boise LH González-García M Postema CE Ding L Lindsten T Turka LA Mao X Nuñez G Thompson CB Bcl-x, a bcl-2-related gene that functions as a dominant regulator of apoptotic cell death Cell 1993 74 597 608 10.1016/0092-8674(93)90508-N 8358789 14. Faustino NA Cooper TA Pre-mRNA splicing and human disease Genes Dev 2003 17 419 37 10.1101/gad.1048803 12600935 15. Bonomi S Gallo S Catillo M Pignataro D Biamonti G Ghigna C Oncogenic alternative splicing switches: role in cancer progression and prospects for therapy Int J Cell Biol 2013 2013 962038 10.1155/2013/962038 24285959 16. Bechara EG Sebestyén E Bernardis I Eyras E Valcárcel J RBM5, 6, and 10 differentially regulate NUMB alternative splicing to control cancer cell proliferation Mol Cell 2013 52 720 33 10.1016/j.molcel.2013.11.010 24332178 17. Brosseau JP Lucier JF Lamarche AA Shkreta L Gendron D Lapointe E Thibault P Paquet E Perreault JP Abou Elela S Redirecting splicing with bifunctional oligonucleotides Nucleic Acids Res 2014 42 e40 10.1093/nar/gkt1287 24375754 18. Saito Y Kanai Y Sakamoto M Saito H Ishii H Hirohashi S Overexpression of a splice variant of DNA methyltransferase 3b, DNMT3b4, associated with DNA hypomethylation on pericentromeric satellite regions during human hepatocarcinogenesis Proc Natl Acad Sci U S A 2002 99 10060 65 10.1073/pnas.152121799 12110732 19. Liu Z Hou J Hepatitis B Virus (HBV) and Hepatitis C Virus (HCV) Dual Infection Int J Med Sci 2006 3 57 62 10.7150/ijms.3.57 16614744 20. Grosso AR Martins S Carmo-Fonseca M The emerging role of splicing factors in cancer EMBO Rep 2008 9 1087 93 10.1038/embor.2008.189 18846105 21. Fackenthal JD Godley LA Aberrant RNA splicing and its functional consequences in cancer cells Dis Model Mech 2008 1 37 42 10.1242/dmm.000331 19048051 22. Warzecha CC Shen S Xing Y Carstens RP The epithelial splicing factors ESRP1 and ESRP2 positively and negatively regulate diverse types of alternative splicing events RNA Biol 2009 6 546 62 10.4161/rna.6.5.9606 19829082 23. Venables JP Koh CS Froehlich U Lapointe E Couture S Inkel L Bramard A Paquet ER Watier V Durand M Lucier JF Gervais-Bird J Tremblay K Prinos P Klinck R Elela SA Chabot B Multiple and specific mRNA processing targets for the major human hnRNP proteins Mol Cell Biol 2008 28 6033 43 10.1128/MCB.00726-08 18644864 24. Schwerk C Schulze-Osthoff K Regulation of apoptosis by alternative pre-mRNA splicing Mol Cell 2005 19 1 13 10.1016/j.molcel.2005.05.026 15989960 25. Qin LX Tang ZY The prognostic molecular markers in hepatocellular carcinoma World J Gastroenterol 2002 8 385 392 10.3748/wjg.v8.i3.385 12046056 26. Geng M Xin X Bi LQ Zhou LT Liu XH Molecular mechanism of hepatitis B virus X protein function in hepatocarcinogenesis World J Gastroenterol 2015 21 10732 38 10.3748/wjg.v21.i38.10732 26478665 27. Stamm S Smith C Luhrmann R Alternative Pre-mRNA Splicing: Theory and Protocols 2012 Germany Wiley-Blackwell 28. Pawellek A McElroy S Samatov T Mitchell L Woodland A Ryder U Gray D Lührmann R Lamond AI Identification of small molecule inhibitors of pre-mRNA splicing J Biol Chem 2014 289 34683 98 10.1074/jbc.M114.590976 25281741 29. Convertini P Shen M Potter PM Palacios G Lagisetti C de la Grange P Sudemycin E influences alternative splicing and changes chromatin modifications Nucleic Acids Res 2014 42 4947 61 10.1093/nar/gku151 24623796 30. Liu X Biswas S Berg MG Antapli CM Xie F Wang Q Tang MC Tang GL Zhang L Dreyfuss G Cheng YQ Genomics-guided discovery of thailanstatins A, B, and C As pre-mRNA splicing inhibitors and antiproliferative agents from Burkholderia thailandensis MSMB43 J Nat Prod 2013 76 685 93 10.1021/np300913h 23517093 31. Klinck R Bramard A Inkel L Dufresne-Martin G Gervais-Bird J Madden R Paquet ER Koh C Venables JP Prinos P Multiple alternative splicing markers for ovarian cancer Cancer Res 2015 68 657 63 10.1158/0008-5472.CAN-07-2580 18245464 32. Mellacheruvu D Wright Z Couzens AL Lambert JP St-Denis NA Li T Miteva YV Hauri S Sardiu ME Low TY The CRAPome: a contaminant repository for affinity purification-mass spectrometry data Nat Methods 2013 10 730 6 10.1038/nmeth.2557 23921808 33. Zhao YJ Ju Q Li GC Tumor markers for hepatocellular carcinoma Mol Clin Oncol 2013 1 593 98 24649215 34. Shirakawa H Kuronuma T Nishimura Y Hasebe T Nakano M Gotohda N Takahashi S Nakagohri T Konishi M Kobayashi N Glypican-3 is a useful diagnostic marker for a component of hepatocellular carcinoma in human liver cancer Int J Oncol 2009 34 649 56 19212669 35. Garber K Energy deregulation: Licensing tumors to grow Science 2006 312 1158 59 10.1126/science.312.5777.1158 16728625 36. Jones RG Thompson CB Tumor suppressors and cell metabolism: a recipe for cancer growth Genes Dev 2009 23 537 48 10.1101/gad.1756509 19270154 37. Parsons DW Jones S Zhang X Lin JC Leary RJ Angenendt P Mankoo P Carter H Siu IM Gallia GL An integrated genomic analysis of human glioblastoma multiforme Science 2008 321 1807 12 10.1126/science.1164382 18772396 38. Lee CH Wu DC Wu IC Goan YG Lee JM Chou SH Chan TF Huang HL Hung YH Huang MC Genetic modulation of ADH1B and ALDH2 polymorphisms with regard to alcohol and tobacco consumption for younger aged esophageal squamous cell carcinoma diagnosis Int J Cancer 2009 125 1134 42 10.1002/ijc.24357 19449376 39. Mamtani M Kulkarni H Association of HADHA expression with the risk of breast cancer: targeted subset analysis and meta-analysis of microarray data BMC Res Notes 2012 5 25 10.1186/1756-0500-5-25 22240105 40. Matlin AJ Clark F Smith CW Understanding alternative splicing: towards a cellular code Nat Rev Mol Cell Biol 2005 6 386 98 10.1038/nrm1645 15956978 41. Karni R de Stanchina E Lowe SW Sinha R Mu D Krainer AR The gene encoding the splicing factor SF2/ASF is a proto-oncogene Nat Struct Mol Biol 2007 14 185 193 10.1038/nsmb1209 17310252 42. Anczuków O Rosenberg AZ Akerman M Das S Zhan L Karni R Muthuswamy SK Krainer AR The splicing factor SRSF1 regulates apoptosis and proliferation to promote mammary epithelial cell transformation Nat Struct Mol Biol 2012 19 220 28 10.1038/nsmb.2207 22245967 43. Warzecha CC Carstens RP Complex changes in alternative pre-mRNA splicing play a central role in the epithelial-to-mesenchymal transition (EMT) Semin Cancer Biol 2012 22 417 27 10.1016/j.semcancer.2012.04.003 22548723 44. Thiery JP Acloque H Huang RY Nieto MA Epithelial-mesenchymal transitions in development and disease Cell 2009 139 871 90 10.1016/j.cell.2009.11.007 19945376 45. LaGamba D Nawshad A Hay ED Microarray analysis of gene expression during epithelial-mesenchymal transformation Dev Dyn 2005 234 132 142 10.1002/dvdy.20489 16010672 46. Warzecha CC Sato T Nabet B Hogenesch JB Carstens RP ESRP1 and ESRP2 are epithelial cell-type-specific regulators of FGFR2 splicing Mol Cell 2009 33 591 601 10.1016/j.molcel.2009.01.025 19285943 47. Ishii H Saitoh M Sakamoto K Kondo T Katoh R Tanaka S Motizuki M Masuyama K Miyazawa K Epithelial splicing regulatory proteins 1 (ESRP1) and 2 (ESRP2) suppress cancer cell motility via different mechanisms J Biol Chem 2014 289 27386 99 10.1074/jbc.M114.589432 25143390 48. van Zijl F Zulehner G Petz M Schneller D Kornauth C Hau M Machat G Grubinger M Huber H Mikulits W Epithelial-mesenchymal transition in hepatocellular carcinoma Future Oncol 2009 5 1169 79 10.2217/fon.09.91 19852728 49. Mehta S Shelling A Muthukaruppan A Lasham A Blenkiron C Laking G Print C Predictive and prognostic molecular markers for cancer medicine Ther Adv Med Oncol 2010 2 125 48 10.1177/1758834009360519 21789130 50. Barboro P Repaci E Rubagotti A Salvi S Boccardo S Spina B Truini M Introini C Puppo P Ferrari N Heterogeneous nuclear ribonucleoprotein K: altered pattern of expression associated with diagnosis and prognosis of prostate cancer Br J Cancer 2009 100 1608 16 10.1038/sj.bjc.6605057 19401687 51. Chen LC Hsueh C Tsang NM Liang Y Chang KP Hao SP Yu JS Chang YS Heterogeneous ribonucleoprotein K and thymidine phosphorylase are independent prognostic and therapeutic markers for nasopharyngeal carcinoma Clin Cancer Res 2008 14 3807 13 10.1158/1078-0432.CCR-08-0155 18559600 52. Kumar M Matta A Masui O Srivastava G Kaur J Thakar A Shukla NK RoyChoudhury A Sharma M Walfish PG Nuclear heterogeneous nuclear ribonucleoprotein D is associated with poor prognosis and interactome analysis reveals its novel binding partners in oral cancer J Transl Med 2015 13 285 10.1186/s12967-015-0637-3 26318153 53. Bruix J Llovet JM Prognostic prediction and treatment strategy in hepatocellular carcinoma Hepatology 2002 35 519 24 10.1053/jhep.2002.32089 11870363 54. Bauman JA Kole R Modulation of RNA splicing as a potential treatment for cancer Bioeng Bugs 2011 2 125 28 10.4161/bbug.2.3.15165 21637003 55. Langmead B Salzberg S Fast gapped-read alignment with Bowtie 2 Nature Met 2012 9 357 35 10.1038/nmeth.1923 56. Storey JD Tibshirani R Statistical significance for genomewide studies Proc Natl Acad Sci U S A 2003 100 9440 9445 10.1073/pnas.1530509100 12883005 57. Jiao X Sherman BT Huang DW Stephens R Baseler MW Lane HC Lempicki RA DAVID-WS: a stateful web service to facilitate gene/protein list analysis Bioinformatics 2012 28 1805 06 10.1093/bioinformatics/bts251 22543366 58. Szklarczyk D Franceschini A Wyder S Forslund K Heller D Huerta-Cepas J Simonovic M Roth A Santos A Tsafou KP STRING v10: protein–protein interaction networks, integrated over the tree of life Nucleic Acids Res 2015 43 D447 D452 10.1093/nar/gku1003 25352553 59. Artimo P Jonnalagedda M Arnold K Baratin D Csardi G de Castro E Duvaud S Flegel V Fortier A Gasteiger E ExPASy: SIB bioinformatics resource portal Nucleic Acids Res 2012 40 W597 W603 10.1093/nar/gks400 22661580 60. Corpet F Multiple sequence alignment with hierarchical clustering Nucleic Acids Res 1988 16 10881 10890 10.1093/nar/16.22.10881 2849754 61. Finn RD Bateman A Clements J Coggill P Eberhardt RY Eddy SR Heger A Hetherington K Holm L Mistry J The Pfam protein families database Nucleic Acids Res 2014 42 D222 D230 10.1093/nar/gkt1223 24288371 62. Kosugi S Hasebe M Tomita M Yanagawa H Systematic identification of cell cycle-dependent yeast nucleocytoplasmic shuttling proteins by prediction of composite motifs Proc Natl Acad Sci U S A 2009 106 10171 76 10.1073/pnas.0900604106 19520826 63. Drissi M Dubois ML Douziech M Boisvert FM Quantitative Proteomics Reveals Dynamic Interactions of the Minichromosome Maintenance Complex (MCM) in the Cellular Response to Etoposide Induced DNA Damage Mol Cell Proteomics 2015 14 2002 13 10.1074/mcp.M115.048991 25963833 64. Zhao M, Kim P, Mitra R, Zhao J, Zhao Z. TSGene 2.0: an updated literature-based knowledgebase for tumor suppressor genes. Nucleic Acids Res. 2016;44(D1):D1023–31. 65. Berglund L Björling E Oksvold P Fagerberg L Asplund A Szigyarto CA Persson A Ottosson J Wernérus H Nilsson P Lundberg E Sivertsson A A genecentric Human Protein Atlas for expression profiles based on antibodies Mol Cell Proteomics 2008 10 2019 27 10.1074/mcp.R800013-MCP200 18669619

PMC005xxxxxx/PMC5002110.txt

==== Front Biomed Eng OnlineBiomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 2675915912010.1186/s12938-015-0120-7ResearchComputer-aided detection (CADe) and diagnosis (CADx) system for lung cancer with likelihood of malignancy Firmino Macedo jose.macedo@ifrn.edu.br 1Angelo Giovani giovani.angelo@gmail.com 2Morais Higor higor.morais@ifrn.edu.br 1Dantas Marcel R. ribeirodantasdm@gmail.com 2Valentim Ricardo ricardo.valentim@ufrnet.br 21 Department of Information and Computer Science, Federal Institute of Rio Grande do Norte (IFRN), Natal, Brazil 2 Laboratory of Technological Innovation in Healthcare, University Hospital Onofre Lopes (HUOL), Natal, Brazil 6 1 2016 6 1 2016 2016 15 1 223 9 2015 22 12 2015 © Firmino et al. 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 CADe and CADx systems for the detection and diagnosis of lung cancer have been important areas of research in recent decades. However, these areas are being worked on separately. CADe systems do not present the radiological characteristics of tumors, and CADx systems do not detect nodules and do not have good levels of automation. As a result, these systems are not yet widely used in clinical settings. Methods The purpose of this article is to develop a new system for detection and diagnosis of pulmonary nodules on CT images, grouping them into a single system for the identification and characterization of the nodules to improve the level of automation. The article also presents as contributions: the use of Watershed and Histogram of oriented Gradients (HOG) techniques for distinguishing the possible nodules from other structures and feature extraction for pulmonary nodules, respectively. For the diagnosis, it is based on the likelihood of malignancy allowing more aid in the decision making by the radiologists. A rule-based classifier and Support Vector Machine (SVM) have been used to eliminate false positives. Results The database used in this research consisted of 420 cases obtained randomly from LIDC-IDRI. The segmentation method achieved an accuracy of 97 % and the detection system showed a sensitivity of 94.4 % with 7.04 false positives per case. Different types of nodules (isolated, juxtapleural, juxtavascular and ground-glass) with diameters between 3 mm and 30 mm have been detected. For the diagnosis of malignancy our system presented ROC curves with areas of: 0.91 for nodules highly unlikely of being malignant, 0.80 for nodules moderately unlikely of being malignant, 0.72 for nodules with indeterminate malignancy, 0.67 for nodules moderately suspicious of being malignant and 0.83 for nodules highly suspicious of being malignant. Conclusions From our preliminary results, we believe that our system is promising for clinical applications assisting radiologists in the detection and diagnosis of lung cancer. Keywords Computer-aided detection systemLung cancer diagnosis Medical image analysisDetection of pulmonary nodulesLikelihood of malignancyCADe and CADxissue-copyright-statement© The Author(s) 2016 ==== Body Background Lung cancer is a disease characterized by the appearance and uncontrolled proliferation of abnormal lung cells. This disease is one of the main causes of mortality worldwide, with approximately 1.59 million deaths per year [1]. The detection in initial stages is considered the most effective way to improve survival of patients, in which case, the 5-year survival rate is approximately 54 % [2]. On the other hand, when the pathology is detected in advanced stages the survival rate for 5 years is only 4 % [2]. Currently, computed tomography (CT) is the imaging modality most suitable for examinations for early detection of lung cancer. CT provides images with high spatial resolution, high temporal resolution and high resolution of contrast of anatomical structures of the chest. This way, it is possible to display small nodules that could hardly be viewed on conventional radiography [3]. According to Awai et al. [4], the detection rate for lung cancer using CT is 2.6–10 times greater than using analog radiography. However, CT generates a large number of medical images which combined to the workload of radiologists could result in inaccurate detection (failure to detect cancer) or misinterpretation (inability to properly diagnose a tumor). Consequently, computer systems become indispensable to assist radiologists in their decision making. A CAD (Computer-Aided Detection and Diagnosis) system is a class of computer systems that aim to assist in the detection and/or diagnosis of diseases through a “second opinion” [5]. The goal of CAD systems is to improve the accuracy of radiologists with a reduction of time in the interpretation of images. CAD systems are classified into two groups: Computer-Aided Detection (CADe) systems and Computer-Aided Diagnosis (CADx) systems. CADe are systems geared for the location of lesions in medical images. Moreover, CADx systems perform the characterization of the lesions, for example, the distinction between benign and malignant tumors. A CADe system for detection of pulmonary nodules usually consists of four main stages: segmentation of the lungs, detection of the candidate nodules, characteristics analysis and elimination of false positives. The segmentation of lung images serves to separate the region in study from other organs and tissues in radiological images. With the segmented images, a search is performed aiming to find abnormal structures present in the lungs, which may be nodules. Then, the characteristics of the possible detected nodules are extracted. The main characteristics that are commonly used for detection of pulmonary nodules are: intensity values of pixels and morphological and texture analysis. Finally, the candidate nodules are classified into nodules or non-nodules by the classifier. This final stage is very important, because it determines the final performance by removing non-nodules and maintaining real nodules. There are several studies in the literature that propose CADe systems for the detection of pulmonary nodules, among them Armato et al. [6]. They have developed a CADe system that used Linear Discriminant Analysis and had a sensitivity of 70 % with 9.6 FP per case. This system has been validated with 187 nodules (solitary pulmonary nodules and juxtapleural nodules). Suzuki et al. [7] developed a pattern recognition technique based on an artificial neural network for a CADe system, named MTANN, and obtained a sensitivity of 80.3 % with 4.8 FP per case, being tested with 121 nodules (solitary, juxtapleural, juxtavascular and ground-glass nodules). Messay et al. [8] presented a CADe system that used a FLD classifier (Fisher Linear Discriminant) and obtained a sensitivity of 82.66 % with 3 FP per case being validated with 143 nodules (solitary, juxtapleural, juxtavascular and ground-glass nodules). Tan et al. [9] developed a CADe system that used a neural classifier and obtained a sensitivity of 87.5 % with an average of 4 FP by case, being tested with 574 nodules (solitary, juxtapleural and juxtavascular nodules). Cascio et al. [10] showed an CADe system that made use of a neural classifier and obtained a sensitivity of 97 % with 6.1 FP per case being validated with 148 nodules (solitary and juxtapleural nodules). Teramoto and Fujita [11] proposed a CADe system that used cylindrical filters and SVM and obtained a sensitivity of 80 % with 4.2 FP per case, being validated with 103 nodules (solitary, juxtapleural, juxtavascular and ground-glass nodules). Han et al. [12] used the Hierarchical Vector Quantization (VQ) method and SVM and obtained a sensitivity of 82.7 % with 4 FP per scan, being tested with 490 nodules (solitary, juxtapleural and ground-glass nodules). Erdal and Aybars [13] proposed a CADe system that provides automatic detection of juxtapleural nodule using the GLMR classifier and image processing techniques and obtained an accuracy of 92.91 %, being tested with 124 juxtapleural nodules. They contributed providing seven new features (five shape-based and two both shape and texture based) extracted from nodule candidates that improve the detection performance. The research showed that 98.6 % of the lung nodules found have remained stable or have become smaller during 2 years of observation, and only 1.4 % of the nodules are actually cancer (malignant structures) [14]. To provide further guidance of malignancy, researchers have been developing systems to aid the diagnosis (CADx). In general, CADx systems extract the characteristics of the images and use a classifier to measure the malignancy. Usually CADx systems for diagnosis of lung cancer are assessed through the ROC curve, more precisely through the area under the ROC curve (Az). A CADx system has been proposed by Shah et al. [15], where they selected 31 characteristics and used logistic regression as classifier, reaching a value of Az 0.92 in distinguishing between 19 malignant nodules and 16 benign nodules, all solitary nodules. Way et al. [16] developed a CADx system that used morphological characteristics, intensity values and surface characteristics. Using the LDC classifier (Linear Discriminant Classifier), Way et al. obtained a Az of 0.857 on the classification of 124 malignant nodules and 132 benign nodules in 152 patients. Suzuki et al. [17] developed a neural network named MTANN to distinguish between benign and malignant nodules. Suzuki et al. reached a value of 0.88, being tested with 76 malignant nodules and 413 benign nodules. Lee et al. [18] developed a supervised learning system that made use of genetic algorithms and Linear Discriminant Analysis (LDA) for the analysis of 216 characteristics of the images and clinical history of patients. They obtained a Az of 0.889 when evaluated 62 malignant nodules and 63 benign nodules, all solitary nodules. Orozco et al. [19] used 11 characteristics calculated from the wavelet transform and SVM as classifier. They obtained a Az of 0.805 being tested with 23 malignant nodules and 22 non-nodules. CADe and CADx systems for lung cancer have been important areas of research in recent decades. However, these research areas are being worked on separately. According to Fraioli et al. [20] one of the main problems of CADe systems for pulmonary nodules is that they detect the nodules but do not characterize them. Thus, computer systems that only detect nodules are not enough for clinical application. Currently, CADe systems do not present the radiological characteristics of the tumor, resulting in lack of information for radiologists, and CADx systems do not detect tumor and do not have good levels of automation [21]. This way, the new CAD should incorporate into a single software a system for detection (CADe) as well as diagnosis (CADx) [20, 22]. In this article, we propose a new approach to CAD systems encompassing detection of different types of nodules and the determination of the likelihood of malignancy of the nodules through computed tomography scans. Our contribution is to improve the level of automation by performing both detection and diagnosis, with little user intervention, and allowing for a satisfactory accuracy for use in clinical and hospital settings. In addition, we present, for the academic community, the use of the Watershed and HOG techniques for distinguishing the lung structures and feature extraction for pulmonary nodules, respectively. The proposed CAD system and the experimental results of its validation are described below. Methods This section presents the materials used in this research and the proposal of a new CAD system for detection and characterization of pulmonary nodules on CT images. Materials The database used in this research consisted of 420 cases, randomly obtained from LIDC-IDRI (Lung Image Database Consortium) [23]. The LIDC-IDRI database is publicly available in the Cancer Imaging Archive (TCIA), and currently contains 1010 CT scans of the chest, collected in different equipment and different configuration parameters (for example, slice thickness, pixel size and total number of slices). The nodule size ranges from 3 mm to 30 mm and can be primary lung cancer, metastatic disease, benign nodule or indeterminate nature. All nodules were evaluated by four experienced radiologists, that through software tools extracted regions of the nodules and described their likelihood of malignancy. The probability was divided into five degree: highly unlikely, moderately unlikely, indeterminate, moderately suspicious and highly suspicious. More details about the database, such as methods and protocols used to acquire image data and the process of annotation of the lesions can be found in Armato et al. [23]. The nodules obtained in our study were diagnosed by consensus by at least two radiologists. 1109 nodules were used, as follows: 196 highly unlikely of being malignant, 254 moderately unlikely of being malignant, 323 of undetermined malignancy, 167 moderately suspicious of being malignant and 169 highly suspicious of being malignant (as shown in Fig. 1). We use the original DICOM images with 16 bit resolution and evaluations from LIDC-IDRI radiologists for the training and test of the supervised classifier. The database tested consisted of solitary, juxtapleural, juxtavascular, small and ground-glass nodules.Fig. 1 Relationship between each category of malignancy with the number of nodules and the number of patients with those nodules In the research was used exams of 420 patients having between 1 and 8 pulmonary nodules. Among them, 31 with nodules highly unlikely of being malignant, 98 with nodules moderately unlikely of being malignant, 139 with nodules with undetermined malignacy, 64 with nodules moderately suspicious of being malignant and 47 with nodules highly suspicious of being malignant (as shown in Fig. 1). Proposed CAD system The proposed CAD system consists of five stages: 3D segmentation of the lungs in CT images, 3D segmentation of the internal structures of the lungs, detection of candidates nodules, elimination of false positives and the calculation of the likelihood of malignancy. During the entire stage of image processing images with 16 bit resolution were used. Below, the details are presented, step by step. 3D segmentation of lungs The segmentation of the lung images can be defined as a process of delineating the spatial extent of the lungs that appear in images of the thorax. This process is possible in CT images because the attenuation values generated for the image reflect the density of the various tissues. The attenuation is typically expressed as the relative attenuation coefficient, called Hounsfield unit (HU) [24]. Thus, a new semi-automatic method is proposed for segmenting lung CT images combining region growing algorithm and morphological filters, as shown in Fig. 2.Fig. 2 Diagram of the semi-automatic method proposed for the segmentation of lung images in Computed Tomography scans of the chest At the beginning of the segmentation process, the user must inform two points (called seeds) in the image that corresponds to pixels that are inside the right and left lung. In sequence, a pre-processing filter is assigned, called Curvature Flow, to eliminate noise in the image. This filter is an algorithm of finite differences proposed by Sethian [25] and implemented by the Insight Segmentation and Registration Toolkit (ITK) [26]. ITK is an open-source, cross-platform system that provides an extensive suite of software tools for image analysis. Then it uses a segmentation algorithm based on regions growing, called Connected Threshold of the ITK toolkit [26]. This algorithm groups the neighboring voxels according to their intensity within a threshold. As boundaries of similarity, the following values were used: −1000 HU and −200 HU. These values were chosen because they encompass the lung tissue, pulmonary vessels and the air within the lungs [27]. In the resulting images of the grouping, the appearing of small structures that are not grouped, including juxtapleural nodules, is common. Juxtapleural nodules are the ones attached to the pleural surface. This segmentation problem is caused due to the fact that nodules have Hounsfield Units similar to the pleura [13]. In order to include these structures, a 3D morphological closing filter with twelve units in radius, by the ITK toolkit, was used to perform a binary dilation followed by an erosion [28]. As a result, it creates a binary mask with voxels. Inside the lungs, the value is one and outside the lungs, the value is zero. This mask is used to determine the volume of the segmented lung, i.e. the voxel number inside the lung. If the volume is close to the statistic average found in lung volumes (equal to 3,545,668 voxels), the mask will be applied in the original image and the segmented lung will be shown to the user. If the volume of the mask is of high variance (over 60 %), a new attempt to group voxels will be held. If the volume of the mask corresponds to the statistic average and if the user accepts the segmentation, the process ends. Otherwise, if the user does not accept the segmentation, the user must inform other seed points to restart the process. 3D segmentation of internal structures of lungs After the segmentation process of the lungs, the segmentation of internal organs is performed. In this segmentation, the internal structures (e.g., trachea, bronchi and pulmonary vessels) are separated aiming to distinguish pulmonary nodules, in case there is any. In this part, the Watershed transform was used, proposed by Vincent and Soille [29, 30] and implemented by ITK Toolkit [26]. This method defines a function f(x, y, z) to group a set of voxels that are local minima. The proposed method used the function f(x, y, z) for calculating the magnitude of the gradient shown in Eqs. 1 and 2. 1 f(x,y,z)=∂J∂x2+∂J∂y2+∂J∂z2 2 Jw=I⊙∂∂w⊙G where: I is the original 3D image and G is a 3D Gaussian function. Through this segmentation, it is possible to group the tissues that have similar intensities allowing lung structures to be separated, especially pulmonary nodules from other structures. Figure 3 shows an example of a reconstruction of the segmented lungs and of the pulmonary structures segmented by the Watershed transform with the function to calculate the gradient magnitude.Fig. 3 3D reconstruction: a of the lungs and b of lung structures segmented by the proposed method Detection of candidate pulmonary nodules The diagnosis of lung cancer usually begins with the identification of an abnormality in radiological tests. These abnormalities are very variable and depend on their location and relationship with the bronchi and vessels. However, the most common radiological patterns are: collapse, consolidation (mass), pleural effusion and different combinations [31]. Initially, they are rounded, but, according to growth, they tend to lose this shape and take more irregular settings and ill-defined contours. The nodules can be classified into: small nodules, nodules attached to vessels (called juxtavascular), nodules attached to the wall of the lung (called juxtapleural) and ground-glass nodules [21]. Small nodules represent nodules with a diameter smaller than 5 mm. Juxtavascular nodules refer to nodules that are connected to blood vessels, while juxtapleural nodules refer to cases in which they are connected to the parenchyma wall or to the diaphragm. Ground-glass nodules refer to a type of nodule where the intensity value of the pixels are significantly lower than those of solid nodules [32]. Figure 4 shows examples of the following types of nodules: ground-glass nodules, juxtapleural nodules, small nodules and juxtavascular nodules.Fig. 4 Examples of different types of lung nodules. a ground-glass nodule with irregular shape, b solid juxtapleural nodule in ovoid shape, c solid spherical nodule with 4 mm in diameter and d solid juxtavascular nodule in ovoid shape To analyze the internal structures that have been segmented and separate the possible nodules, a rule-based classifier has been used. The first rule, referred to as Roundness was applied to the segmented structure aiming to detect spherical or semi-spherical objects. Its calculation is shown in Eq. 3. Whenever the Roundness exceeds a threshold, the segmented object is considered to be a non-nodule and disposed from later stages. The second rule, called Elongation, aims to detect cylindrical structures. Its calculation is shown in Eq. 4. Whenever the Elongation is less than a threshold, the object is considered to be a non-nodule. The third rule is based on image texture by calculating the Energy. The Energy, calculated through the co-occurrence matrix, expresses the uniformity of texture on the image aiming to eliminate regions that do not contain nodules. The energy is calculated by Eq. 5 and if the measured value is less than a threshold, the object is not considered to be a nodule. 3 (R1)Roundness=An(r)a<8.3×104 4 (R2)Elongation=MPmaxMPmin<6.8×104 5 (R3)Energy=∑1ns∑levels-1i,j=0Pi,j2<3.3 where: An is the area of a hypersphere (of radius r) that has the same nodule volume, a is the area of the nodule, MPmax is the largest main image moment MPmin is the smallest main image moment, ns is the number of slices where the nodule appears, levels corresponds to the maximum intensity value in the gray scale of the image and Pi,j is the histogram of co-occurrence of gray level of the image. More details about Roundness, Elongation and Energy, such as definitions and parameters can be found in Lehmann [28] and Hall-Beyer [33]. The rule-based classifier has been used to quickly remove some structures that are easily distinguishable as False Positives (e.g., bronchi, trachea, pulmonary vessels) so as to eliminate the influence of such structures on the subsequent stages. In this work, the rules were designed based on knowledge gained from radiologists and statistical study of the morphological characteristics, intensity and texture of lung nodules and non-nodules at CT scans available in LIDC-IDRI. The rules were defined with relatively tenuous criteria so that they are not specific to the data set used in this work. Elimination of false positives At this stage we will eliminate remaining false positives (FPs) while preserving true positives. In the context of CADe systems, the false positive term means lesions that are identified by the CAD algorithm, but are not nodules. Typical false positives were: vessels with sharp curvature, thick vessels with bifurcations, stains generated by respiratory or cardiac motion and scarring on the parenchymal tissue (parenchymal tissue). Typically, FPs are removed by classification algorithms. For this, candidate nodules are segmented and their features are extracted. The job of the classifier is to determine boundaries for the separation of classes (i.e. nodules and non-nodules) based on the extracted features. The method of feature extraction of nodules used was the Histogram of Oriented Gradient (HOG) [34] of the Skimage library [35]. Skimage is a collection of algorithms for image processing and computer vision available free of charge and free of restrictions. The basic idea of this method is that the appearance and shape of objects present in images can be characterized by the distribution of the intensity and direction of the gradients of pixels. The HOG subdivides the images into small regions (called cells) and for each pixel inside the cell its gradient is calculated through Eq. 6. Thereafter, for each cell it computes a histogram of the gradients. The gradient indicates the direction of maximum intensity variation near the pixel, catching contours, silhouettes and some information about the texture. To improve accuracy, the cells histograms are normalized by their grouping with the neighboring cells histograms. This grouping of cells is called a block and this normalization results in better invariance to the changes in lighting and shading. 6 ∇f=∂f∂xX+∂f∂yY where: ∂f∂x is the gradient in the direction X and ∂f∂y is the gradient in the direction Y. The HOG was calculated for each slice of the object. Then a resulting histogram was generated by grouping the HOG of each slice. The result presented high dimensionality with feature vectors with dimensions between 77 and 2,380,848 for each candidate nodule. In this way, the Principal Component Analysis (PCA) was applied to reduce the dimensionality, so that data can be handled and stored more efficiently. PCA is a mathematical method, proposed by Hotelling [36], which uses orthogonal transformation to convert a set of variables, possibly correlated, to a set of values of linearly uncorrelated variables called principal components. The reduction of the dimension of the data consists in obtaining the main components from the ordering of the extracted eigenvalues of the covariance matrix of the original data [37]. The method proposed by Thomas P Minka [38] allows the use of PCA by defining a minimum percentage of variance to be maintained without the need to determine in advance the number of components. With this, PCA was used keeping 80 % of the variance of the original data. After the PCA, the feature vector of candidate nodules had a dimension of 73. The PCA algorithm used in this study was implemented by the Sklearn library [39]. Sklearn is a simple and efficient open source tool for data mining and data analysis. Finally, the Support Vector Machines (SVM) classifier is used to analyze the results of the PCA. The SVM is a technique based on the Statistical Learning Theory of the type supervised training [40], able to generalize problems of binary classification from a data set. Its operation is given through nonlinear functions (called kernels) that map input vectors in a high dimensional space (called feature space) [41]. In the proposed method the SVM classifier was chosen for elimination of False Positives because it provides the best results when compared to other classifiers. This comparison will be presented in the "Results and discussion" sections. Compute the likelihood of malignancy Once detected, the next stage is to determine the likelihood of malignancy of the nodules. Regarding the texture, nodules with ground-glass opacity are malignant in 59–73 % of cases, while solid nodules have a probability of 7–9 %. Regarding its shape, nodules of irregular shape have a higher likelihood of malignancy compared to round nodules. Calcification patterns of nodules are also useful to determine the malignancy. Calcification patterns for benign nodules have been described as: central, popcorn, solid and laminated. Benign nodules often have well-defined smooth edges. Spiculated nodules and nodules with irregular or lobular margins are more often malignant nodules. According McNitt-Gray et al. [42] the probability of malignancy is related to the age of the patient, whether he is a smoker and the features of shape and appearance of nodules. These features are: calcification patterns, internal structure, margin, shape, texture and presence of lobulation and spiculation. Thus, the proposed CAD system will use the concepts of McNitt-Gray et al. to determine the likelihood of malignancy considering patients older than 60 years old and known to smoke. The likelihood of malignancy is divided into five degree: highly unlikely, moderately unlikely, indeterminate, moderately suspicious and highly suspicious. The use of five degree was suggested by experienced radiologists interviewed. The radiologist should see the highlighted nodules and their 3D reconstruction, as shown in Fig. 5. For reconstruction, the VTK library was used [43]. As a result, radiologists must report seven features of texture, shape and appearance of nodules. These features are: calcification patterns, internal structure, Lobulation, Margin, Sphericity, Spiculation and Texture. The features and their respective values are shown in Table 1.Fig. 5 Examples of a nodule detected by the system. a A highlighted nodule in the various slices of the test, b 3D reconstruction of the nodule found Table 1 Features of texture, shape and appearance of nodules and their values that radiologists should tell the system Features Values Calcification Popcorn, laminated, solid, non-central, central, and absent Internal structure Soft tissue, fluid, fat, and air Lobulation Marked, intermediate, and none Margin Poorly defined, intermediate, and sharp Sphericity Linear, ovoid, and round Spiculation Marked, intermediate, and none Texture Non-solid, Part solid/(mixed), and solid With the values provided by radiologists, the system uses an SVM classifier, trained previously, to determine the likelihood of malignancy. The SVM classifier was chosen because it provides the best results when compared to other classifiers tested. Based on the likelihood of malignancy, the radiologist may have more information to take the decision on the treatment and monitoring of patients. Results and discussion Lung cancer is responsible for over 1.59 million deaths each year. CAD systems are being developed to assist radiologists in the detection and diagnosis in order to decrease this rate. These systems must provide acceleration in the diagnosis, reducing errors and improving the quantitative evaluation. Currently, the main criteria used to evaluate the detection of CAD systems (CADe) are: false positive (FP) rate and sensitivity. FP is when the system presents positively to a nodule when in reality it does not exist. The sensitivity (S) is the relation between true positives and the false negative, given by Eq. 7. 7 S=TP(TP+FN) where: S is the sensitivity, TP is the true positive rate and FN is the false negative rate. TP represents the results that the system presented positively to a sample that actually had the disease. FN represents the negative results when the sample had the disease. The Receiver Operator Characteristics (ROC) is the performance measure most widely used to evaluate the diagnosis of CAD systems (CADx). The ROC provides an index related to the accuracy and effectiveness through the relationship between the probability of true positives and the probability of false positives. The ROC curve is created by plotting the true positive (TP) rate on the Y axis, and the false positive (FP) rate on the X axis at various threshold settings. After the computation of the ROC curve, the area under the curve (Az) should be calculated using the Eq. 8. Sklearn library [39] was used to generate the ROC curve and calculate the Az. 8 Az=∫-∞∞TPR(T)FPR′(T)dT where: TPR is true positive rate and FPR is false positive rate. To validate the CAD system 420 CT scans were used. Computation time of the proposed system was approximately 12 min per case using a notebook with Intel Core i7-4500U CPU 1.80 GHz × 4. The segmentation stage had an accuracy of 97 % and is not effective in cases of severe pathologies, which alter the opacity of the lung outlines. Validate of classifier The method to validate the ability of generalization of the classifier was the 10-fold Cross Validation and Leave-one-out. The effectiveness of the SVM is verified by comparing with FLD (Fisher’s Linear Discriminant) and Gaussian Naive Bayes. The SVM used was the C-Support Vector Classification (SVC) with radial kernel implemented by the Sklearn library. SVC is one implementation of SVM with multiclass support that performs a one-vs-one approach [44]. If k is the number of classes, then k(k-1)/2 classifiers are constructed and each one trains data from two classes. When training an SVM with the Radial Basis Function kernel, two parameters must be considered: C and gamma. The parameters used were C=5 and 9 gamma=1n=177indetection,17indiagnosis. where: n is the number of features. Other values were tested but these showed better results. Fisher’s Linear Discriminant (FLD) is a supervised classifier that projects high-dimensional data onto one-dimensional space for classification [45]. The FLD used was implemented by the Sklearn library with Singular Value Decomposition solver. Naive Bayes methods is a supervised classifier based on applying Bayes’ theorem with MAP (maximum a posteriori) estimation [46]. The Naive Baye used was the Gaussian Naive Bayes implemented by Sklearn. Following the results of validation tests will be presented. 10-fold cross validation In the 10-fold Cross Validation method the original database is randomly separated in k mutually exclusive subsets and of the same size. These K subsets, K-1 are used to train the classifier and 1 subset is used for validation tests. This procedure is repeated until all k subsets are used for the tests. The result of this process is the average of performance in all tests. The aim of repeating multiple times is to increase the reliability of the estimate of the accuracy of the classifier. In the tests performed, we used k=10. This method is commonly used in validation of CAD systems [7–10]. The results are shown in Table 2.Table 2 Performance comparison of classifiers for detection and diagnosis using 10-fold cross validation Classifier Sensitivity (%) FP Az FLD 89.2  6.89 0.91, 0.81, 0.69, 0.70, 0.81 Naive Bayes 93.9 7.03 0.89, 0.82, 0.67, 0.70, 0.83 SVM 94.4 7.04 0.91, 0.80, 0.72, 0.67, 0.83 As result, the CAD system with SVM performs better. The sensitivity found for detection was of 94.4 % with a FP rate of 7.04 per case. The area values of the ROC curve found with SVM were between 0.91 for the nodules with highly unlikely malignancy, 0.80 for nodules with moderately unlikely malignancy, 0.72 for nodules with indeterminate malignancy, 0.67 for the nodules suspected of moderately malignancy and 0.83 for highly suspected malignant nodules. Leave-one-out validation The Leave-one-out Validation is a statistical technique used to determine, during training, the generalization capability of classifiers [40]. The dataset should be divided randomly into two distinct sets, one for training (used to train) and one for validation (used to validate). This method also is used in validation of CAD systems [47, 48]. In the tests we used: the training subset containing 294 CT scans (with 215 cancerous and 79 non-cancerous) and the validation subset containing 126 CT scans (with 89 cancerous and 37 non-cancerous). The effectiveness of the SVM is verified by comparing with FLD (Fisher’s Linear Discriminant) and Gaussian Naive Bayes, both implemented by the Sklearn library. The results are shown in Table 3.Table 3 Performance comparison of classifiers for detection and diagnosis using leave-one-out validation Classifier Sensitivity ( %) FP Az FLD 88.99 7.47 0.93, 0.79, 0.68, 0.72, 0.80 Naive Bayes 90.47 7.44 0.92, 0.80, 0.66, 0.70, 0.85 SVM 93.9 7.21 0.93, 0.78, 0.69, 0.67, 0.85 Through the Table 3, it can be inferred that the performance of the SVM is the best. The sensitivity found for detection was of 93.9 % with a FP rate of 7.21 per case. The area values of the ROC curve found with SVM were between 0.93 for the nodules with highly unlikely malignancy, 0.78 for nodules with moderately unlikely malignancy, 0.69 for nodules with indeterminate malignancy, 0.67 for the nodules suspected of moderately malignancy and 0.85 for highly suspected malignant nodules. Comparing the performance A relative comparison of our detection method using SVM with other mentioned in the literature was performed and shown in Table 4. In this Table, each line represents a CADe system for the detection of pulmonary nodules on CT images. For each system it presents the sensitivity obtained, numbers of false positives and number of nodules used for validation. Based on this comparison, it can be inferred that the performance of the proposed system is among the best in sensitivity and has been tested with a larger amount of nodules.Table 4 Performance comparison of pulmonary nodule detection methods by sensitivity, FP and the number of nodules used in the validation Methods Sensitivity (%) FP N of nodules Armato et al. [6] 70  9.6/case 187 Suzuki et al. [7] 80.3 4.8/case 121 Messay and Rogers [8] 82.66 3/case 143 Tan et al. [9] 87.5 4/case 574 Cascio et al. [10] 97 6.1/case 148 Teramoto e Fujita [11] 80 4.2/case 103 Han et al. [12] 82.7 4/scan 490 Our method 94.4 % 7.04/case 1109 The Fig. 6 shows the results obtained by the system, with SVM method, in the diagnosis of malignancy in terms of ROC curves. A literature review was conducted to identify the results of other CADx systems for the diagnosis of lung cancer and compare with the proposed CADx system. A summary of the results obtained by other CADx systems and the proposed method is shown in Table 5.Fig. 6 ROC curve for the distinction of classes: Highly Unlikely, Moderately Unlikely, Indeterminate, Moderately Suspicious and Highly Suspicious, obtained in the 10-fold Cross Validation with SVM Table 5 Performance comparison of diagnostic methods of pulmonary nodules by ROC curve Methods Az Classes Shah et al. [15] 0.92 Malignant and benign Way et al. [16] 0.857 Malignant and benign Suzuki et al. [17] 0.88 Malignant and benign Lee et al. [18] 0.889 Malignant and benign Orozco et al. [19] 0.805 Malignant and benign Our method 0.91, 0.80, 0.72, 0.67 and 0.83 Likelihood of malignancy In Table 5, each row represents a published method followed by the area under the ROC curve obtained by the system and the type of classification performed. Based on this comparison, it can be inferred that the performance of our system is at the same level of the results shown in the mentioned documents. However, our system presents as advantage the diagnosis based in likelihood of malignancy through the subdivision into five degree, allowing more aid in the decision making by radiologists. Accordingly, we believe that our system is clinically useful for the detection and diagnosis of pulmonary nodules, because it performed well in the detection, in the diagnosis and it has a good level of automation. The experimental results on the set of independent data show the generalization of the proposed method. However, the system does not detect lung nodules smaller than 3 mm and should not be used in cases in which there is the presence of severe pathologies, which alter the opacity of lung outlines. Conclusion A new CAD system has been proposed for the detection and diagnosis of pulmonary nodules in CT images of the chest, grouping in a single system both identification and characterization of nodules. For this, the use of Histogram of Oriented Gradient (HOG) was proposed for characterization of nodules and the use of the Watershed technique to segment lung internal structures to separate the possible nodules from other structures. Besides, a rule-based classifier and a Support Vector Machine (SVM) have been used to eliminate false positives. The segmentation method achieved an accuracy of 97 %, though not being effective in cases where severe pathologies occurred, altering the opacity of lung outlines. The detection system showed a sensitivity of 94.4 % with 7.04 false positive per case for the detection of nodules with diameters between 3 mm and 30 mm and of different types (isolated, juxtapleural, juxtavascular and ground-glass nodules). For the diagnosis of malignancy, our system presented ROC curves with areas of: 0.91 for nodules highly unlikely of being malignant, 0.80 for nodules moderately unlikely of being malignant, 0.72 for nodules with indeterminate malignancy, 0.67 for nodules moderately suspicious of being malignant and 0.83 for nodules highly suspicious of being malignant. The diagnosis based on likelihood of malignancy enables greater aid in the decision making made by radiologists. From our preliminary results, we believe that our system is promising for clinical applications assisting radiologists in the detection and diagnosis of lung cancer. However, we are still working to enable the segmentation of the lungs of patients with diseases that alter the opacity and outlines of lungs and to automatically generate seed points. Image processing techniques are being analyzed in order to generate the characteristics of nodules (Calcification patterns, internal structure, Lobulation, Margin, Sphericity, Spiculation and Texture) directly from the images allowing greater automation of the system. As future work, we plan to conduct a clinical trial of the proposed method and verify their performance in a real environment, analyze other method of feature extraction of nodules and conduct another study to find the optimal values of SVM classifier. Macedo Firmino and Giovani Angelo contributed equally Authors’ contributions MF and GA: review of the literature, organizing, responsible for coding the algorithms, performs the tests and validations, and preparing the manuscript. HM, MRD and RV: manuscript review, modification, editing, revision, and supervise all the process. All authors read and approved the final manuscript. Acknowledgements The authors acknowledge the National Cancer Institute and the Foundation for the National Institutes of Health and their critical role in the creation of the free publicly available LIDC/IDRI Database used in this study. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Organization WH. International Agency for Research on Cancer. GLOBOCAN 2012: Estimated Cancer Incidence, Mortality and Prevalence Worldwide in 2012. http://globocan.iarc.fr/Pages/fact_sheets_cancer.aspx?cancer=lung. Accessed 21 Jul 2015. 2. U.S.N.I. of Health. SEER Cancer Statistics Review, 1975-2011. http://seer.cancer.gov/archive/csr/1975_2011/. Accessed: 21 Jul 2015. 3. Li Q Recent progress in computer-aided diagnosis of lung nodules on thin-section CT Comput Med Imaging Graph 2007 31 4–5 248 57 10.1016/j.compmedimag.2007.02.005 17369020 4. Kazuo A Kohei M Akio O Masanori K Haruo H Shinichi H Yasumasa N Pulmonary nodules at chest ct: effect of computer-aided diagnosis on radiologists’ detection performance Radiology 2004 230 347 352 10.1148/radiol.2302030049 14752180 5. Suzuki K A review of computer-aided diagnosis in thoracic and colonic imaging Quantitative Imag Med Surg 2012 2 3 163 176 6. Armato SG Gieger ML Moran CJ Blackburn JT Doi K Macmahan H Computerized detection of pulmonary nodules on CT scans Radiographics 1999 19 5 1303 1311 10.1148/radiographics.19.5.g99se181303 10489181 7. Suzuki K, Armato III SG, Li F, Sone S, Doi K. Massive training artificial neural network (mtann) for reduction of false positives in computerized detection of lung nodules in low-dose computed tomography. Medical Physics. 2003;30(7):1602–17. 8. Messay T Hardie RC Rogers SK A new computationally efficient CAD system for pulmonary nodule detection in CT imagery Med Image Anal 2010 14 3 390 406 10.1016/j.media.2010.02.004 20346728 9. Tan M Deklerck R Jansen B Bister M Cornelis J A novel computer-aided lung nodule detection system for ct images Med Phys 2011 38 10 5630 5645 10.1118/1.3633941 21992380 10. Cascio D Magro R Fauci F Iacomi M Raso G Automatic detection of lung nodules in ct datasets based on stable 3d mass-spring models Comput Biol Med 2012 42 11 1098 1109 10.1016/j.compbiomed.2012.09.002 23020972 11. Teramoto A Fujita H Fast lung nodule detection in chest ct images using cylindrical nodule-enhancement filter Int J Comp Assist Radiol Surg 2013 8 2 193 205 10.1007/s11548-012-0767-5 12. Han H Li L Han F Song B Moore W Liang Z Fast and adaptive detection of pulmonary nodules in thoracic ct images using a hierarchical vector quantization scheme IEEE J Biomed Health Inform 2015 19 2 648 659 10.1109/JBHI.2014.2328870 25486657 13. Taşcı E, Uğur A. Shape and texture based novel features for automated juxtapleural nodule detection in lung cts. J Med Syst 2015;39(5). 14. Swensen SJ Jett JR Hartman TE Midthun DE Sloan JA Sykes A-M Aughenbaugh GL Clemens MA Lung cancer screening with ct: Mayo clinic experience Radiology 2003 226 3 756 761 10.1148/radiol.2263020036 12601181 15. Shah SK McNitt-Gray MF Rogers SR Goldin JG Suh RD Sayre JW Petkovska I Kim HJ Aberle DR Computer aided characterization of the solitary pulmonary nodule using volumetric and contrast enhancement features1 Acad Radiol 2005 12 10 1310 1319 10.1016/j.acra.2005.06.005 16179208 16. Way TW Sahiner B Chan H-P Hadjiiski L Cascade PN Chughtai A Bogot N Kazerooni E Computer-aided diagnosis of pulmonary nodules on ct scans: Improvement of classification performance with nodule surface features Med Phys 2009 36 7 3086 3098 10.1118/1.3140589 19673208 17. Suzuki K Li F Sone S Doi K Computer-aided diagnostic scheme for distinction between benign and malignant nodules in thoracic low-dose ct by use of massive training artificial neural network IEEE Trans Med Imag 2005 24 9 1138 1150 10.1109/TMI.2005.852048 18. Lee MC, Boroczky L, Sungur-Stasik K, Cann AD, Borczuk AC, Kawut SM, Powell CA. Computer-aided diagnosis of pulmonary nodules using a two-step approach for feature selection and classifier ensemble construction. Artif Intel Med. 2010; 50(1):43–53 (Knowledge Discovery and Computer-Based Decision Support in Biomedicine) 19. Orozco H, Villegas O, Sánchez V, Domínguez H, Alfaro M. Automated system for lung nodules classification based on wavelet feature descriptor and support vector machine. BioMed Eng. 2015;14(9). 20. Fraioli F Serra G Passariello R Cad (computed-aided detection) and cadx (computer aided diagnosis) systems in identifying and characterising lung nodules on chest ct: overview of research, developments and new prospects La radiologia medica 2010 115 3 385 402 10.1007/s11547-010-0507-2 20077046 21. El-Baz A Beache GM Gimel’farb G Suzuki K Okada K Elnakib A Soliman A Abdollahi B computer-aided diagnosis systems for lung cancer: challenges and methodologies Int J Biomed Imaging 2013 2013 1 46 22. Firmino M, Morais AH, Mendoca RM, Dantas MR, Hekis HR, Valentim R. Computer-aided detection system for lung cancer in computed tomography scans: review and future prospects. BioMed Eng. 2014; 13(41). 23. Sg A The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): a completed reference database of lung nodules on CT scans Med Phys 2011 38 2 915 931 10.1118/1.3528204 21452728 24. Preim B, Bartz D. Chapter 03—acquisition of medical image data. In: Bartz BP (ed.) Visualization in Medicine. The Morgan Kaufmann Series in Computer Graphics, Morgan Kaufmann, Burlington. 2007. p. 35–64. 25. Sethian JA Level set methods and fast marching methods: evolving interfaces in computational geometry, fluid mechanics, computer vision, and materials science, Cambridge monographs on applied and computational mathematics 1999 Cambridge Cambridge University Press 26. Lowekamp BC, Chen DT, Ibanez L, Blezek D. The design of simpleitk. Front Neuroinform. 2013;7:45. 27. Dougherty G Digital image processing for medical applications 2009 1 USA Cambridge University Press 28. Lehmann G. Label object representation and manipulation with ITK. Insight J 2007; 08 29. Vincent L Soille P Watersheds in digital spaces: an efficient algorithm based on immersion simulations IEEE Trans Pat Anal Mach Intel 1991 13 6 583 598 10.1109/34.87344 30. Beare R, Lehmann G. The watershed transform in itk—discussion and new developments. 2006. 31. Sharma CP, Behera D, Aggarwal AN, Gupta D, Jindal SK. Radiographic patterns in lung cancer. Indian J Chest Dis Allied Sci 2002; 41(1):25–30. 32. Min JH Lee HY Lee KS Han J Park K Ahn M-J Lee S-J Stepwise evolution from a focal pure pulmonary ground-glass opacity nodule into an invasive lung adenocarcinoma: An observation for more than 10 years Lung Cancer 2010 69 1 123 126 10.1016/j.lungcan.2010.04.022 20478641 33. Hall-Beyer M. The GLCM Tutorial Home Page. Current Version: 2.10. http://www.fp.ucalgary.ca/mhallbey/tutorial.htm 34. Dalal N, Triggs B. Histograms of oriented gradients for human detection. In: Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference On, vol. 1, 2005. p. 886–8931. 35. van der Walt S Schönberger JL Nunez-Iglesias J Boulogne F Warner JD Yager N Gouillart E Yu T The scikit-image contributors: scikit-image: image processing in Python Peer J 2014 2 453 10.7717/peerj.453 36. Hotelling H Analysis of a complex of statistical variables into principal components J Educ Psych 1933 24 1 48 10.1037/h0071325 37. Buciu I Gacsadi A Directional features for automatic tumor classification of mammogram images Biomed Signal Process Control 2011 6 4 370 378 10.1016/j.bspc.2010.10.003 38. Minka TP Automatic choice of dimensionality for PCA Adv Neural Inform Process Syst 2001 15 598 604 39. Pedregosa F Varoquaux G Gramfort A Michel V Thirion B Grisel O Blondel M Prettenhofer P Weiss R Dubourg V Vanderplas J Passos A Cournapeau D Brucher M Perrot M Duchesnay E Scikit-learn: machine learning in Python J Mach Learn Res 2011 12 2825 2830 40. Haykin S Neural networks: a comprehensive foundation 1998 2 Upper Saddle River Prentice Hall PTR 41. Cortes C Vapnik V Support-vector networks Mach Learn 1995 20 3 273 297 42. McNitt-Gray MF The lung image database consortium (lidc) data collection process for nodule detection and annotation Acad Radiol 2007 14 12 1464 1474 10.1016/j.acra.2007.07.021 18035276 43. Schroeder W Martin KM Lorensen WE The Visualization Toolkit 2006 4 Clifton Park Kitware Inc 44. Knerr S Personnaz L Dreyfus G Soulie F Hérault J Single-layer learning revisited: a stepwise procedure for building and training a neural network Neurocomputing 1990 Berlin Heidelberg Springer 41 50 45. Hastie T Tibshirani R Friedman J The elements of statistical learning data mining, inference, and prediction 2008 2 Stanford Springer 46. Zhang H Barr V Markov Z The Optimality of Naive Bayes FLAIRS Conference 2004 Miami Beach, Florida AAAI Press 47. Gurcan MN Sahiner B Petrick N Chan HP Kazerooni EA Cascade PN Hadjiiski L Lung nodule detection on thoracic computed tomography images: preliminary evaluation of a computer-aided diagnosis system Med Phys 2002 28 2552 2558 10.1118/1.1515762 12462722 48. Matsumoto S, Kundel HL, Gee JC, Gefter WB, Hatabu H. Pulmonary nodule detection in CT images with quantized convergence index filter. Med Image Anal. 2006;10(3):343–352. (Special Issue on The Second International Workshop on Biomedical Image Registration (WBIR’03))

PMC005xxxxxx/PMC5002111.txt

==== Front BMC Health Serv ResBMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 170610.1186/s12913-016-1706-xResearch ArticlePsychometric properties of the Nursing Home Survey on Patient Safety Culture in Norwegian nursing homes http://orcid.org/0000-0001-7770-344XCappelen Kathrine Kathrine.Cappelen@hit.no 1Aase Karina karina.aase@uis.no 2Storm Marianne marianne.storm@uis.no 2Hetland Jørn Joern.Hetland@psysp.uib.no 3Harris Anette Anette.Harris@uib.no 31 Centre for Caring Research South, Faculty of Health and Social Sciences, University College of Southeast Norway, Porsgrunn, Norway 2 Department of Health Studies, Faculty of Social Sciences, University of Stavanger, Stavanger, Norway 3 Department of Psychosocial Science, Faculty of Psychology, University of Bergen, Bergen, Norway 27 8 2016 27 8 2016 2016 16 1 44616 7 2015 24 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 Developing a culture where staff are actively aware of how to prevent adverse events is a challenge. The use of survey tools to assess the status of patient safety culture seems to be acceptable as an early step in improving patient safety. The Nursing Home Survey on Patient Safety Culture (NHSOPSC) includes 12 dimensions and is specifically developed for nursing homes. In this study, we describe a Norwegian version of the NHSOPSC and assess its psychometric properties when tested on a sample of healthcare staff in nursing homes. Methods The NHSOPSC was translated into Norwegian and pilot tested before being distributed to 12 nursing homes in Norway. Of the 671 healthcare staff invited, 466 (69 %) answered the questionnaire. SPSS 23.0 was used for descriptive data analysis and estimating internal consistency (Cronbach’s alpha). The dimensional structure of the questionnaire was tested by confirmatory factor analysis (CFA) using Mplus (version 7.2). Results The CFA testing of the original 12-factor solution suggested that some modifications were needed because of the high correlations between three of the latent factors. A subsequent analysis resulted in a final ten-factor solution. The final model showed acceptable fit to the data (root mean square error of approximation = 0.060, 90 % confidence interval: 0.057–0.063, comparative fit index = 0.934, Tucker-Lewis index = 0.926, χ2 = 2058.33, df = 765, p < 0.001) and acceptable factor loadings ranging from 0.402 to 0.891. Moreover, moderate-to-strong correlations ranging from 0.455 to 0.812 were found between the ten latent factors. Finally, moderate-to-high correlations were found between the ten latent factors and an overall rating of patient safety in the nursing home. Conclusions Factor analysis indicated that a modified ten-factor model fitted the data set in a Norwegian community healthcare context with acceptable goodness-of-fit values and could be recommended as a useful tool to assess staff perceptions of patient safety issues in Norwegian nursing homes. Electronic supplementary material The online version of this article (doi:10.1186/s12913-016-1706-x) contains supplementary material, which is available to authorized users. Keywords Nursing homesPatient safetyPerceptionQuality improvementSafety culturePsychometricshttp://dx.doi.org/10.13039/501100004190Norsk Sykepleierforbund15/0011Cappelen Kathrine http://dx.doi.org/10.13039/501100005416Norges Forskningsråd204637Cappelen Kathrine issue-copyright-statement© The Author(s) 2016 ==== Body Background Proven methods and useful tools for measuring quality of care and patient safety in nursing homes appear to be lacking, although studies have shown that patient safety culture is poorly developed and patients may be at risk of harm [1]. Adverse events also seem to be common in long-term care, including nursing homes [2, 3]. One of the most discussed areas within the field of patient safety is how to develop a culture where staff are actively aware of the potential for adverse events [3–6]. Safety culture may be considered part of the organizational culture; it creates norms and influences staff attitudes and behavior. The Agency for Healthcare Research and Quality (AHRQ) refers to the following definition of safety culture, which is also applicable to nursing homes: “The safety culture of an organization is the product of individual and group values, attitudes, perceptions, competencies, and patterns of behavior that determine the commitment to, and the style and proficiency of, an organization’s health and safety management. Organizations with a positive safety culture are characterized by communications founded on mutual trust, by shared perceptions of the importance of safety, and by confidence in the efficacy of preventive measures.” [7] p. 23. Safety culture is a complex phenomenon, which makes it difficult to operationalize [4, 8]. An extensive range of safety culture properties have been identified and organized into dimensions or so called subcultures, such as leadership, teamwork, evidence-based patient care, communication, learning, just culture, and patient centered care, which form a safety culture framework [8]. These dimensions may represent ways of conceptualizing safety culture. Halligan and Zecevic [4] found that the most frequently cited dimensions were leadership commitment to safety, open communication founded on trust, organizational learning, a nonpunitive approach to adverse event reporting and analysis, teamwork, and shared belief in the importance of safety [4]. Studies from both hospitals and nursing homes point at the association between patient safety culture and clinical outcomes such as safe care [1, 3, 9, 10]. Major efforts have been carried out to improve patient safety, such as campaigns, programs and adverse event reporting systems [11]. In 2011, the Norwegian Ministry of Health and Care Services launched the campaign, “In Safe Hands,” which continued as a national program for patient safety with increased attention on primary care, including nursing homes. Assessing patient safety culture is a major strategy in the ongoing program. Several validated instruments for mapping safety culture in healthcare organizations are accessible, including the Safety Attitude Questionnaire and Hospital Survey of Patient Safety Culture (HSOPSC) [6, 12, 13]. In 2004, the AHRQ released the HSOPSC, and it has since been translated and tested in different countries [14]. Although it was originally developed in the United States, 44 different European studies conducted in 20 different countries have used the HSOPSC [15]. A Swedish version of the HSOPSC has been tested on a large sample in both hospitals and primary healthcare [16]. In 2008, in a response to interest shown by nursing homes in an instrument similar to the HSOPSC, the AHRQ released the Nursing Home Survey on Patient Safety Culture (NHSOPSC). The NHSOPSC is designed especially for nursing home staff and explores their perceptions of patient safety culture [17]. According to the AHRQ, NHSOPSC can be used in the following ways: “As a diagnostic tool to assess the status of patient safety culture in a nursing home, as an intervention to raise staff awareness about patient safety issues, as a mechanism to evaluate the impact of patient safety culture improvement initiatives, and as a way to track changes in patient safety culture over time.” [17] p. 1. The NHSOPSC is in line with Schein’s definition of organizational culture defined as: “(a) a pattern of basic assumptions, (b) invented, discovered, or developed by a given group, (c) as it learns to cope with its problems of external adaptation and internal integration, (d) that has worked well enough to be considered valid and, therefore (e) is to be taught to new members as the (f) correct way to perceive, think, and feel in relation to those problems.” [18] p. 123. The central core is the basic assumptions and shared perceptions, primarily unconscious and representing learned responses based on experiences. Over time, this may develop into a common practice, “this is how we do it here”. Even though the NHSOPSC has been available for some years, its use in Europe is still limited. The only available validation study in a European context is a Swiss version (translated into German) based on data from nine nursing homes (n = 477). The Swiss version gave a nine-dimensional fit as opposed to the 12 original dimensions. The study concluded that the survey’s dimensionality needed further clarification, mainly to distinguish items addressing the unit-level from those at the facility level [19]. Hence, there is a need for further studies to test the original version of NHSOPSC. The Swiss study applied exploratory factor analyses, often resulting in a reduction in number of factors due to the use of a data reduction method. The aim of this study is therefore to test the psychometric properties of the Norwegian version of the original NHSOPSC by asking for staff perceptions of patient safety culture in a sample of Norwegian nursing homes. Based on previous research and the original structures and theoretical domains our hypothesis was that the original factor structure of the NHSOPSC could be replicated in a Norwegian context [1, 2, 17]. Methods Design, setting, and recruitment The study was designed as a cross-sectional survey examining staff perceptions of patient safety in nursing homes using a Norwegian version of the NHSOPSC. The study has been reported according to the STROBE checklist for cross-sectional studies (Additional file 1). Healthcare staff from 12 different nursing homes in southern and western Norway, including urban and rural districts, received the survey. The nursing homes varied in size and organization. Norwegian nursing homes offer advanced care such as long- and short-term care, subacute and acute care, rehabilitation, specialized care for patients with dementia and cognitive impairments, and palliative care. Inclusion criteria were staff, defined as healthcare workers with a minimum of 30 % part-time position and experience with the nursing home policies and day-to-day activities. Staff working in both nursing homes and assisted living or home care had to work at least a 30 % part-time position in the nursing home. All respondents had to be able to read and understand Norwegian. Data collection A total of 671 questionnaires were distributed as paper versions to 12 nursing homes in southern and western Norway. An informative letter presenting the study with a return envelope accompanied the questionnaires. The survey was conducted in June–September, 2013. Questionnaire The NHSOPSC includes 43 survey items measuring 12 dimensions (Table 1). The questionnaire is divided into four sections (A, B, C, and D) in addition to overall ratings of the nursing home, and background variables [17].Table 1 The original patient safety culture dimensions of the NHSOPSC used in the Norwegian nursing home study Patient safety culture dimensions Items 1. Teamwork 4 2. Staffing 4 3. Compliance with procedures 3 4. Training and skills 3 5. Nonpunitive response to mistakes 4 6. Handoffs 5 7. Feedback and communication about incidents 4 8. Communication openness 3 9. Supervisor expectations and actions promoting patient safety 3 10. Overall perceptions of patient safety 3 11. Management support for patient safety 3 12. Organizational learning 4 Nine of the 12 dimensions in NHSOPSC are similar to the HSOPSC, but the included items are different. All items in NHSOPSC are rated on Likert-type scales from 1 to 5 with verbal anchors. “Does not apply or don’t know” was included as a response category. Overall rating questions considered as outcomes consist of a single statement, “I would tell friends that this is a safe nursing home for their family” (Yes, Maybe, No) and a graded overall rating, “Please give this nursing home an overall rating on patient safety” (scale from 1 to 5). Translation procedures and content validity The US version of the NHSOPSC was first translated to Norwegian by a translation team, and then back-translated by an independent and experienced translator fluent in English and Norwegian according to recommended guidelines [20]. The translation team members were fluent in both languages and included an occupational therapist with a background in community healthcare and nursing homes and experienced in translating surveys, a research assistant with experience in surveys and epidemiology, an assistant professor and researcher within elderly care, and a coordinator. The draft translation was pretested in a group of healthcare staff working in nursing homes and community healthcare representing different professions such as nurses responsible for quality work, an apprentice, nurse managers, a medical doctor, and an assistant (n = 7). The pretest emphasized whether or not the items were relevant and understandable for the users. The final review was accepted by an external researcher who had also translated the hospital version (HSOPSC) of the instrument into Norwegian [21]. The Norwegian version retained 43 items. One item, B3, “We have all the information we need when residents are transferred from hospital,” was split into B3a, “We get the medical information when patients are transferred from hospital” and B3b, “We get the nursing report when patients are transferred from hospital.” The AHRQ guidelines [20] were adjusted to a Norwegian healthcare context and the role of nursing homes. Consequently, the concept of resident safety was replaced by patient safety, which is in line with the legal definition of a patient in Norway [22]. Furthermore, in the Norwegian version, the focus is on the “the nursing home as a whole” referred to as the facility level, using terms like “our nursing home” and “your nursing home.” The sample of nursing homes differs in both size and organization. Statistical analysis Response rate and variability Response rate and variability were examined with frequency analysis. Response variability was considered low when 90 % or more of the respondents chose the answer “agree/strongly agree” or “most of the time/always”. Chi-square for independency was used to examine if the number of “does not apply or don’t know” and number of missing values differed significantly based on respondents’ background variables. The analyses were performed with SPSS (version 23.0; IBM SPSS, Armonk, NY). P-value < 0.05 was considered significant. Confirmatory factor analysis (internal structure) Structural equation modeling was used to test how the original 12-factor solution fitted the Norwegian data. The means and variance adjusted weighted least square estimator appropriate for ordinal data were used. Missing data, including the response option “does not apply or don’t know”, were handled by pairwise deletion, which is the default when using this estimator. The best-fit model was assessed with the comparative fit index (CFI) [23], the Tucker-Lewis index (TLI) [24] and the root mean square error of approximation (RMSEA) [25]). CFI level above 0.95, TLI level above 0.90, and a RMSEA value less than 0.060 is recommended as a good fit [26]. Factor loadings were expected to be above 0.30 [27]. Internal consistency Cronbach’s alpha coefficients were used to measure internal consistency of the NHSOPSC instrument and its dimensions. Homogeneity was considered good if the alpha values were between 0.70 and 0.90 and acceptable if the alpha values were >0.60 [15]. Validity based on relation to other variables (overall rating questions) The two overall rating questions had substantially different connotations; therefore, they were treated as complementary questions. The first question, “I would tell friends that this is a safe nursing home for their family,” was explored using descriptive statistics (n, %) in SPSS (version 23.0; IBM SPSS, Armonk, NY). The correlation between the second overall rating question “Please give this nursing home an overall rating on patient safety,” and the modified Norwegian ten-factor model was assessed in Mplus (version 7.2; Muthén and Muthén, Los Angeles, CA). P-value < 0.05 was considered significant. Results A total of 671 participants were invited of which 466 (69 %) returned the Norwegian NHSOPSC questionnaire. The respondents were managers (n = 29), healthcare workers with a minimum of a bachelor degree (n = 181), healthcare workers with upper secondary school education (n = 226), assistants (n = 12), and others (n = 9). Furthermore, 70 % of the respondents worked more than 25 h per week. Background variables are presented in Table 2.Table 2 Background variables of the respondents in the Norwegian nursing home study (n = 466) Background variables Number Percent Staff position or background (n = 457)  Managers including leaders at first-line level 29 6.3  Healthcare workers with a minimum of bachelor degree 181 39.6  Healthcare workers, upper secondary school 226 49.5  Assistants 12 2.6  Others 9 2.0 Number and years in nursing home (n = 457)  <1 year 29 6.3  1–5 years 114 24.9  6–10 years 105 23.0  11–15 years 105 23.0  16–20 years 44 9.6  >21 years 60 13.1 Work hours per week (n = 454)  <15 h 7 1.5  16–24 h 121 26.7  25–35.5 h 253 55.7  >35.5 h 73 16.1 Work shift (most often) (n = 447)  Daytime 303 67.8  Afternoon 85 19.0  Nighttime 59 13.2 Working directly with patients most of the time (n = 458)  Yes 436 95.2  No 22 4.8 Response rate and variability The response rates per item and the number of “does not apply or don’t know” responses, and the number of missing responses are presented in Table 3. Missing values range from 3 to 36. The highest missing values were related to the items; “We receive medical information when patients are transferred from hospital” (B3a) and “We receive nursing report when patients are transferred from hospital” (B3b). The responses “does not apply or don’t know”, also treated as missing values in the analysis, range from 0 (0 %) to 88 (19 %).Table 3 Dimensions/items with corresponding mean and standard deviation (SD), response rate (n %), responses to “Does not apply, Do not know” category, and missing values Dimensions/Items Mean (SD) Response rate n (%) Does not apply or don’t know Missing 1. Teamwork  A1. Staff in this nursing home treat each other with respect 4.31 (0.79) 463 (69 %) 0 3  A2. Staff support one another in this nursing home 4.18 (0.79) 459 (68 %) 0 7  A5. Staff feel like they are part of a team 4.10 (0.82) 463 (69 %) 0 3  A9. When someone gets really busy in this nursing home, other staff help out 3.71 (0.85) 459 (68 %) 1 6 2. Staffing  A3. We have enough staff to handle the workload 2.97 (0.89) 456 (68 %) 3 7  A8 (R). Staff have to hurry because they have too much work to do 2.55 (0.92) 458 (68 %) 2 6  A16. Patients need are met during shift changes 3.80 (0.81) 450 (67 %) 9 7  A17 (R). It is hard to keep patients safe because so many staff quit their jobs 4.11 (0.85) 427 (64 %) 30 9 3. Compliance with procedures  A4. Staff follow standard procedures to care for patients 4.00 (0.79) 455 (68 %) 6 5  A6 (R). Staff use shortcuts to get their work done faster 3.38 (0.92) 450 (67 %) 11 5  A14 (R). To make work easier, staff often ignore procedures 3.82 (0.81) 444 (66 %) 17 5 4. Training and skills  A7. Staff get the training they need in this nursing home 3.63 (0.86) 460 (69 %) 3 3  A11. Staff have enough training on how to handle difficult patients 3.21 (0.86) 450 (67 %) 8 8  A13. Staff understand the training they get in this nursing home 3.89 (0.73) 437 (65 %) 20 9 5. Nonpunitive response to mistakes  A10 (R). Staff are blamed when a patient is harmed 4.13 (0.77) 427 (63 %) 30 9  A12 (R). Staff are afraid to report their mistakes 3.76 (0.83) 435 (65 %) 26 5  A15. Staff are treated fairly when they make mistakes 3.95 (0.79) 417 (62 %) 40 9  A18. Staff feel safe reporting their mistakes 3.97 (0.74) 447 (67 %) 14 5 6. Handoffs  B1. Staff are told what they need to know before taking care of a patient for the first time 4.01 (0.72) 458 (68 %) 4 4  B2. Staff are told when there is a change in a patients’ care plan 3.79 (0.78) 447 (67 %) 13 6  B3a. We have all the information we need when patients are transferred from the hospital (medical information) 4.09 (0.80) 386 (58 %) 44 36  B3b. We have all the information we need when patients are transferred from the hospital (nursing report) 4.24 (0.74) 381 (57 %) 55 30  B10. Staff are given all the information they need to take care of patients 4.24 (0.63) 458 (68 %) 1 7 7. Feedback and communication about incidents  B4. When staff report something that could harm a patient, someone takes care of it 4.25 (0.69) 434 (65 %) 27 5  B5. In this nursing home, we talk about ways to keep patients from happening again 3.98 (0.78) 453 (68 %) 3 10  B6. Staff tell someone if they see something that might harm a patient 4.42 (0.58) 454 (68 %) 6 6  B8. In this nursing home, we discuss ways to keep patients safe from harm 4.07 (0.69) 450 (67 %) 6 10 8. Communication openness  B7. Staff ideas and suggestions are valued in this nursing home 3.85 (0.74) 457 (68 %) 1 8  B9 (R). Staff opinions are ignored in this nursing home 3.82 (0.83) 448 (67 %) 8 10  B11. It is easy for staff to speak up about problems in this nursing home 3.90 (0.85) 446 (66 %) 12 8 9. Supervisor expectations and actions promoting patient safety  C1. My supervisor listen to staff ideas and suggestions about patient safety 4.22 (0.76) 453 (68 %) 6 7  C2. My supervisor says a good word to staff who follow the right procedures 4.19 (0.78) 452 (67 %) 6 8  C3. My supervisor pays attention to patient safety problems in this nursing home 4.38 (0.66) 450 (67 %) 9 7 10. Overall perceptions of patient safety  D1. Patients are well cared for in this nursing home 4.33 (0.73) 462 (69 %) 1 3  D6. This nursing home does a good job keeping patients safe 3.97 (0.64) 449 (67 %) 11 6  D8. This nursing home is a safe place for patients 4.30 (0.64) 461 (69 %) 1 4 11. Management support for patient safety  D2. Management asks staff how the nursing home can improve patient safety 3.58 (0.96) 423 (63 %) 33 10  D7. Management listen to staff ideas and suggestions to improve patient safety 3.80 (0.83) 443 (66 %) 19 4  D9. Management often walks around the nursing home to check on patients care 2.96 (1.14) 412 (61 %) 45 9 12. Organizational learning  D3 (R). This nursing home lets the same mistakes happen again and again 3.70 (0.85) 433 (65 %) 27 6  D4. It is easy to changes to improve patient safety in this nursing home 3.59 (0.79) 439 (65 %) 18 9  D5. This nursing home is always doing things to improve patient safety 3.79 (0.74) 444 (66 %) 18 4  D10. When this nursing home makes changes to improve patient safety, it checks to see if the changes worked 3.61 (0.84) 365 (54 %) 88 13 Item-response categories: Items A1–A18, items C1–C3, and items D1–D10: 1 = strongly disagree; 2 = disagree; 3 = neither agree nor disagree; 4 = agree; 5 = strongly agree; item B1–B11: 1 = never; 2 = rarely; 3 = sometimes; 4 = most of the time; and 5 = always R reverse coded The items in the NHSOPSC instrument are rated on scales from 1 to 5. The results showed that all five response alternatives had been used for 36 of the items, while the response alternatives 2–5 had been used for eight of the items: “Staff are told what they need to know before taking care of a patient for the first time” (B1), “We receive nursing report when patients are transferred from hospital (B3b), “When staff report something that could harm a patient, someone takes care of it” (B4), “Staff tell someone if they see something that may harm a patient” (B6), “Staff are given all the information they need to take care for patients” (B10), “This nursing home does a good job keeping patients safe” (D6), “This nursing home is a safe place for patients” (D8) and the overall rating question “Please give this nursing home an overall rating on patient safety” (E2). The response variability was below 90 % for all items except “Staff tell someone if they see something that may harm a patient” (B6), “Staff are given all the information they need to take care for patients” (B10), “My supervisor pays attention to patient safety problems in this nursing home” (C3) and “This nursing home is a safe place for patients” (D8) in which the response variability was 96, 91, 93 and 91 % respectively. Some items have a relatively high response rate to the “does not apply or don’t know” category. Results revealed that mainly staff with lower education, staff having worked less than a year or part time workers had missing responses or responded “does not apply or don’t know”. Significant differences in responses “does not apply or don’t know” according to the background variables are listed in Appendix A (Additional file 2) and significant differences in missing responses according to the background variables are listed in Appendix B (Additional file 3). Confirmatory factor analysis (internal structure) In accordance with the expected dimensionality of the NHSOPSC instrument, we first tested a confirmatory factor analysis (CFA) with 12 latent factors and their respective indicators (n = 347). Factors were allowed to correlate in the model. However, the initial CFA showed a negative definite matrix for some of the latent factors. The analysis indicated that this was caused by an almost perfect correlation between the latent factors “Organizational learning,” “Overall perceptions of patient safety” (r = 0.94), and “Management support for patient safety” (r = 0.94). These high correlations indicate that these latent factors could be merged into one factor labeled “Management and organization learning.” The results were in accordance with a Swiss validation of the NHSOPSC instrument [19] and in accordance with the organizational structure in Norwegian nursing homes. Accordingly, we therefore tested a ten-factor model where the three factors were merged. The model fit was acceptable (RMSEA = 0.063, 90 % confidence interval [CI]: 0.060–0.066, CFI = 0.921, TLI = 0.912, χ2 = 2317.67, df = 815, p < 0.001). However, the standardized factor loadings for two of the indicators, B3a and B3b, were <0.40 (0.158 and 0.345, respectively). Therefore, we tested a modified model omitting these indicators from the respective factor (F6). The modified ten-factor model showed a better fit (RMSEA = 0.060, 90 % CI: 0.057–0.063, CFI = 0.934, TLI = 0.926, χ2 = 2058.33, df = 765, p < 0.001). In the final model, the factor loadings varied between 0.402 and 0.891 and loaded significantly (p < 0.001) on the latent factors for all indicators (Table 4). The correlations between the 10 latent factors were moderate to strong and varied between 0.455 and 0.812 (Table 5).Table 4 Confirmatory factor analysis of the Norwegian NHSOPSC ten-factor model with corresponding factor loadings and reliability (Cronbach’s alpha) (n = 347) Factors (Cronbach’s alpha) Items Factor loading F1 Teamwork (α = 0.79)  A1 0.883  A2 0.885  A5 0.859  A9 0.538 F2 Staffing (α = 0.55)  A3 0.556  A8 (R) 0.402  A16 0.653  A17 (R) 0.573 F3 Compliance with procedures (α = 0.58)  A4 0.781  A6 (R) 0.447  A14 (R) 0.648 F4 Training and skills (α = 0.67)  A7 0.737  A11 0.662  A13 0.715 F5 Nonpunitive response to mistakes (α = 0.65)  A10 (R) 0.433  A12 (R) 0.558  A15 0.767  A.18 0.834 F6 Handoffs (α = 0.74)  B1 0.758  B2 0.709  B10 0.826 F7 Feedback and communication about incidents (α = 0.74)  B4 0.752  B5 0.675  B6 0.689  B8 0.800 F8 Communication openness (α = 0.74)  B7 0.778  B9 (R) 0.673  B11 0.812 F9 Supervisor expectations and actions promoting patient safety (α = 0.84)  C1 0.891  C2 0.865  C3 0.857 F10 Management and organizational learning (new factor) (α = 0.90)  D1 0.733  D2 0.764  D3 (R) 0.705  D4 0.759  D5 0.816  D6 0.869  D7 0.818  D8 0.820  D9 0.589  D10 0.733 The new factor management and organizational learning includes “overall perception of safety,” “management support for patient safety,” and “organizational learning.” Item-response categories: Items A1–A18, items C1–C3, and items D1–D10: 1 = strongly disagree; 2 = disagree; 3 = neither agree nor disagree; 4 = agree; 5 = strongly agree; item B1–B11: 1 = never; 2 = rarely; 3 = sometimes; 4 = most of the time; and 5 = always R reverse coded items Table 5 Correlation between the 10 latent factors in the Norwegian NHSOPSC measured with Mplus (n = 347). All p-values < 0.001 F1 F2 F3 F4 F5 F6 F7 F8 F9 F1 1 F2 .609 1 F3 .642 .675 1 F4 .675 .782 .758 1 F5 .603 .664 .647 .663 1 F6 .491 .634 .613 .755 .455 1 F7 .535 .609 .614 .649 .595 .812 1 F8 .685 .705 .635 .709 .715 .725 .784 1 F9 .591 .510 .506 .602 .632 .521 .543 .770 1 F10 .593 .716 .650 .715 .553 .730 .751 .779 .627 F1 = Teamwork, F2 = Staffing, F3 = Compliance with procedures, F4 = Training and skills, F5 = Nonpunitive response to mistakes, F6 = Handoffs, F7 = Feedback and communication about incidents, F8 = Communication openness, F9 = Supervisor expectations and actions promoting patient safety, F10 = Management and organizational learning (new factor) Internal consistency Results of the internal consistency analyses of the Norwegian NHSOPSC ten-factor model showed that eight of the ten factors reached an acceptable level of internal consistency with Cronbach’s alpha values >0.60. Two factors showed low values: “Staffing” (0.55) and “Compliance with procedures” (0.58) (Table 4). Validity based on relation to other variables (overall rating questions) Correlation analyses performed in Mplus showed moderate-to-strong correlation between all factors in the modified ten-factor model and the overall rating question, “Please give this nursing home an overall rating of patient safety” (E2), with a range between 0.478 and 0.824 (Table 6).Table 6 Correlations between the ten-factor model of the Norwegian NHSOPSC and the outcome measure item “Please give this nursing home an overall rating on patient safety” (E2) measured with Mplus (n = 337) Factors E2 F1 Teamwork .498** F2 Staffing .664** F3 Compliance with procedures .527** F4 Training and skills .598** F5 Nonpunitive response to mistakes .520** F6 Handoffs .606** F7 Feedback and communication about incidents .608** F8 Communication openness .673** F9 Supervisor expectations .478** F10 Management and organizational learning (new dimension) .824** **p < 0.01 The overall rating question, “I would tell friends that this is a safe nursing home for their family” (E1), documented that most of the respondents answered yes (86.6 %, n = 388), some answered maybe (12.7 %, n = 57), and only a few (0.7 %, n = 3) answered no. Discussion Various safety culture tools are available to measure safety culture in healthcare organizations, among them surveys [28, 29]. We have described the results of a validation study using the Norwegian version of the NHSOPSC in a nursing home context. As such, it represents the first survey-based study of patient safety culture in Norwegian nursing homes. The response rate was satisfactory (69 %) compared with similar validation studies from Switzerland (66 %) [19], Norway (55 %) [21], and the United Kingdom (37 %) [14]. Regarding the number of factors, the CFA results from the Norwegian NHSOPSC study were in contrast to the original US version, but in accordance with the results from a Swiss study where the number of latent factors was reduced [17, 19]. The Swiss study concluded with nine factors, while our study concluded with ten factors. In the Swiss study, two factors “Overall perceptions of resident safety” and “Organizational Learning” were merged into one factor [19], while the same two factors and one additional factor “Management support for patient safety” were merged in the modified Norwegian model. Corresponding reductions of factors have been observed in similar validation studies of safety culture in hospitals where surveys are translated and tested in different contexts [14]. The distinction between facility and unit level within the nursing home was not applicable for the sample of nursing homes in this study because of differences in size and organizational models in the Norwegian setting. Therefore, the Norwegian version of the NHSOPSC focuses on the nursing home at facility level, referred to as “our nursing home,” “this nursing home,” and “your nursing home” in the questionnaire. Supervisors represent leaders at the lowest level conducting “day-to-day leadership” in nursing homes. The items B3a and B3b did not contribute to the confirmed ten-factor model. During the translation and pretesting process, the original item, “We have all the information we need when residents are transferred from the hospital,” was split into two items related to medical information and nursing reports. By asking more specifically for information needed from the hospital (medical report, nursing report), these items no longer contributed to the model for measuring patient culture in nursing homes. This means that receiving medical information and nursing reports is still essential for patient safety but does not represent a cultural issue for nursing homes. Therefore, future studies should adhere to the original item B3 as one item. After removing the indicators B3a and B3b and reducing the number of factors, confirmatory analysis indicated that a ten-factor model best fitted the data set in a Norwegian community healthcare context with acceptable goodness-of-fit values. In the Norwegian ten-factor model, the latent factors “Staffing” and “Compliance with procedures,” showed alpha values below 0.60. This might be explained by the number of items included in the latent factors besides some items showing a low loading. The factor “Compliance with procedures” included only three items, and one, “Staff use shortcuts to get their work done faster” (A6), had a low loading (0.447). Furthermore, the factor “Staffing” included four items, and one, “Staff have to hurry because they have too much work to do” (A8), had a low loading (0.402). In this survey, both “Staffing” and “Compliance with procedures” are regarded as dimensions, although they are usually not among the most cited dimensions within the field of patient safety culture [4, 8]. Despite questionable internal consistency, we found no reason to remove these dimensions and related items from the questionnaire. These factors display important opinions and attitudes related to staffing and compliance with procedures that influence patient safety in nursing homes. Another argument for keeping the items in the instrument intact is to ensure the possibility for cross-country comparisons. The AHRQ has established comparative databases serving as important input for benchmarking and improvement efforts [30]. Finally, reliability measured with Cronbach’s alpha is not a recommended method when using latent factors [31]. A more appropriate method has been described by Raykov and Penev [32], but this method is only applicable for continuous variables so far. Both the original NHSOPSC and the Norwegian version include only two overall rating questions (E1, E2). The correlations revealed a medium-to-large strength between the ten factors and the question, “Please give this nursing home an overall rating on resident safety” (E2). All ten factors reached statistical significance. The question, “I would tell friends that this is a safe nursing home for their family” (E1), revealed that more than 86 % of the respondents answered, “Yes”, indicating that it corresponds with the overall rating on patient safety. The Norwegian study shows relatively high scores. In comparison, an average of 76 % answered “Yes” to the same question in the US reference database including 263 nursing homes [30]. Presuming that there might be an association between safety culture and safe care for patients in nursing homes, more specific outcome measures related to adverse events would be of interest. Limitations The prevalence of missing values in the current study was high; however, they were mainly linked to the category, “does not apply or don’t know”. We argue that this should be considered a valid answer by many of the respondents with lower education having worked less than a year or working part time, having in common that they may not yet hold sufficient information to be able to respond. For example the item, “When this nursing home makes changes to improve patient safety, it checks to see if the changes worked” (D10) revealed significant differences related to the background variable “number of years in the nursing home”. Another limitation of the study is that we have not considered the possible cluster effect of unit versus facility levels in our analysis. However, a two-level confirmatory analysis was performed for the within-level factors with individuals (level 1) nested within units (level 2), accounting for the correlation of ratings within units (results not reported). The model showed poor fit, which was probably because of the relatively small sample size. However, the analysis showed interclass correlation variations between 0.005 and 0.0045, indicating that less than 5 % of the dimension variance was explained by unit/facility levels. Therefore, the unit/facility-level dimension was unessential in the Norwegian data set. Challenges related to translation and use of surveys outside the geographical and healthcare context in which they are developed must be considered [14]. Organizational culture and the role of nursing homes in healthcare systems will differ across country-specific contexts because of current healthcare policies, reforms, and financial systems [33]. Nevertheless, questions related to patient safety cultures and staff awareness of adverse events will most likely have common denominators at a cross-national level. This requires some consistency regarding terminology and a minimum of shared items. In this study staff’s opinions are treated as identical with staff’s perceptions and relate to how “reality” is interpreted. By asking for opinions, the survey also tries to map attitudes, but this raises several challenges related to the ability to measure attitudes. Several studies suggest the existence of an association between patient safety attitudes, staff behaviour and patient outcomes [1, 34]. However, the relationship between attitudes and behaviour seems to be complex and influenced by social factors such as management and staff support [35]. We have therefore limited our analysis to focus mainly on perceptions. By conducting a NHSOPSC survey in nursing homes, patient safety issues will inevitably be put on the agenda. Therefore, the survey could be considered an intervention to raise awareness of patient safety issues among nursing home staff [17]. Discussing survey results may help nursing homes identify strengths and quality-improvement areas. Presenting the results may be a good starting point, but lasting changes depend on continuous work and survey follow-up [30]. Conclusions The study results indicated that a ten-factor model of the NHSOPSC instrument fitted the data set in a Norwegian community healthcare context with acceptable goodness-of-fit values. The NHSOPSC survey instrument seems to include the most-frequently theoretically cited dimensions related to safety culture as a concept. A validation study including a strategic selection of respondents from a larger sample of nursing homes will be required to assess safety culture at the unit/facility level. Additional files Additional file 1: STROBE checklist for cross-sectional studies. (DOC 84 kb) Additional file 2: Appendix A including tables 7, 8, 9 and 10. Percentages of participants responding “does not apply or don’t know” according to staff position or background. (DOCX 18 kb) Additional file 3: Appendix B including tables 11, 12, 13 and 14. Percentages of “Missing” according to staff position or background. (DOCX 21 kb) Abbreviations AHRQAgency for Healthcare Research and Quality CFAConfirmatory factor analysis CFIComparative fit index CIConfidence interval HSOPSCHospital Survey on Patient Safety Culture NHSOPSCNursing Home Survey on Patient Safety Culture RMSARoot mean square error of approximation TLITucker-Lewis index The authors would like to thank the management and staff of the participating nursing homes for all assistance in the collection of data, and Nina Konglevoll for technical assistance. We also thank PhD Espen Olsen for his supervision in the translation of the NHSOPSC, and Professor Solveig Hauge for supervision in designing the study and for contributing in the translation process. Funding The study received funding from the Norwegian Nurses Organisation and the Norwegian Research Council (grant agreement no. 204637). The funders had no part in the design of the study, collection, analysis, interpretation of data and in writing the manuscript. Availability of data and materials The datasets analysed in the current study are available by request from the corresponding author. Data are stored at Uni Research Health, Uni Research, Bergen, Norway, and the University of Stavanger, Department of Health Studies, Norway. Authors’ contributions KC was responsible for designing the study, developing the Norwegian version of the NHSOPSC, data collection, and statistical analysis, interpretation of data, and writing the first draft of the manuscript. KA supervised the design of the study, contributed to interpretation of data, and critically revised the manuscript. MS participated in data collection, and drafted and revised the manuscript. JH supervised the statistical analysis, and contributed to interpretation of data and revision of the manuscript. AH was responsible for statistical analyses, and contributed to interpretation of data and revision of the manuscript. All authors read and approved the final manuscript. Authors’ information KC, RN, Cand. polit., Assistant professor, 1Centre for Caring Research South, Faculty of Health and Social Science, University College of Southeast Norway. Research interests: Patient safety and leadership. KA, PhD in Work Environment and Safety, Professor at the Department of Health Studies, Faculty of Social Sciences, University of Stavanger, Norway. Head of the research group “Quality and Safety in Health Care Systems.” Research interests: Patient safety, organizational/cultural factors, transitional care. MS, PhD in Management, Associate professor at Department of Health Studies, Faculty of Social Sciences, University of Stavanger, Norway. Research interests: Patient and user involvement, organizational culture and learning, patient safety, and intervention research. JH, PhD in Psychology. Associate professor Department of Psychosocial Science, Faculty of Psychology, University of Bergen, Norway. Research interests: Work and organizational psychology, leadership and health. AH, PhD in Psychology, Associate professor at the Faculty of Psychology, University of Bergen, Norway. Research interests: Work psychology. 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 Norwegian Social Science Data Services (Ref. No. 2012/32450) and the Norwegian Regional Committees for Medical and Health Research Ethics (Ref. No. 2011/1978). Participation was based on written informed consent. ==== Refs References 1. Thomas KS Hyer K Castle NG Branch LG Andel R Weech-Maldonado R Patient Safety Culture and the Association with Safe Resident Care in Nursing Homes Gerontologist 2012 52 6 802 811 10.1093/geront/gns007 22383542 2. Castle NG Wagner LM Ferguson JC Handler SM Nursing home deficiency citations for safety J Aging Soc Policy 2011 23 1 34 57 10.1080/08959420.2011.532011 21207305 3. Wagner LM Capezuti E Rice JC Nurses’ Perceptions of Safety Culture in Long-Term Care Settings J Nurs Scholarsh 2009 41 2 184 192 10.1111/j.1547-5069.2009.01270.x 19538703 4. Halligan M Zecevic A Safety culture in healthcare: a review of concepts, dimensions, measures and progress BMJ Qual Saf 2011 20 4 338 343 10.1136/bmjqs.2010.040964 21303770 5. Botwinick L Bisognano M Haraden C Leadership Guide to Patient Safety IHI Innovation Series whitepaper. Edited by IHI 2006 Cambridge Institute of Healthcare Improvement 6. Waterson P Patient Safety Culture: Theory, Methods and Application. Wey court East, Union Road, Farnham, Surrey, GU9/PT 2014 England Ashgate Publishing Limited 7. Advisory Committee for the Safety of Nuclear Installations Organising for Safety Sheffield: HSE Books: Human Factors Study Group Third Report 1993 8. Sammer CE Lykens K Singh KP Mains DA Lackan NA What is patient safety culture? A review of the literature J Nurs Scholarsh 2010 42 2 156 165 10.1111/j.1547-5069.2009.01330.x 20618600 9. El-Jardali F Dimassi H Jamal D Jaafar M Hemadeh N Predictors and outcomes of patient safety culture in hospitals BMC Health Serv Res 2011 11 45 10.1186/1472-6963-11-45 21349179 10. Ross J Patient Safety Outcomes: The Importance of Understanding the Organizational Culture and Safety Climate J PeriAnesthesia Nurs 2011 26 5 347 348 10.1016/j.jopan.2011.08.001 11. Saunes IS, Ringard Å. What´s done to improve patient safety? Initiatives in seven countries. Report No. 17 - 2013. Oslo: The Norwegian Knowledge Centre for Health Services; 2013. 12. Sexton J Helmreich R Neilands T Rowan K Vella K Boyden J Roberts P Thomas E The Safety Attitudes Questionnaire: psychometric properties, benchmarking data, and emerging research BMC Health Serv Res 2006 6 1 44 10.1186/1472-6963-6-44 16584553 13. Sorra JS Nieva VF Hospital Survey on Patient Safety Culture. (Prepared by Westat, under Contract No. 290-96-0004) 2004 Rockville Agency for Healthcare Research and Quality 14. Waterson P Griffiths P Stride C Murphy J Hignett S Psychometric properties of the Hospital Survey on Patient Safety Culture: findings from the UK Qual Saf Health Care 2010 19 5 e2 20211960 15. Manser AHT Waterson P The Use of the Hospital Survey on Patient Safety Culture in Europe Patient Safety Culture–Theory, Methods and Application. edn 2014 Wey Court East, Union Road, Farnham Ashgate Publishing Limited 229 261 16. Hedskold M Pukk-Harenstam K Berg E Lindh M Soop M Ovretveit J Sachs MA Psychometric properties of the hospital survey on patient safety culture, HSOPSC, applied on a large Swedish health care sample BMC Health Serv Res 2013 13 1 332 10.1186/1472-6963-13-332 23964867 17. Sorra J Franklin M Streagle S Nursing Home Survey on Patient Safety Culture. (Prepared by Westat, under Contract No. 233-02-0087). AHRQ Publication No. 08-0060 2008 Rockville Agency for Healthcare Research and Quality 18. Schein EH Organizational Culture Am Psychol 1990 45 109 119 10.1037/0003-066X.45.2.109 19. Zúñiga F Schwappach D De Geest S Schwendimann R Psychometric properties of the Swiss version of the Nursing Home Survey on Patient Safety Culture Saf Sci 2013 55 88 118 10.1016/j.ssci.2012.12.010 20. Translation Guidelines for the AHRQ Surveys on Patient Safety Culture. http://www.ahrq.gov/sites/default/files/wysiwyg/professionals/quality-patient-safety/patientsafetyculture/resources/transguide.pdf. Accessed 7 Feb 2012. 21. Olsen E Reliability and validity of the Hospital Survey on Patient Safety Culture at a Norwegian hospital. Quality and Safety Improvement Research: Methods and Research Practice from the International Quality Improvement Research Network (QIRN) 2008 173 186 22. Ministry of Health and Care Services. Pasient–og brukerettighetsloven § 1–3. 2012. https://lovdata.no/dokument/NL/lov/1999-07-02-63. 23. Bentler P Comparative fit indexes in structural models Psychol Bull 1990 107 2 238 246 10.1037/0033-2909.107.2.238 2320703 24. Tucker LR Lewis C A reliability coefficient for maximum likehood factor analysis Psychometrica 1973 38 1 10 10.1007/BF02291170 25. Steiger JH Structural model evaluation approach Multivar Behav Res 1990 25 173 180 10.1207/s15327906mbr2502_4 26. Hu L Bentler PM Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives Struct Equa Model: A Multidisciplinary Journal 1999 6 1 1 55 10.1080/10705519909540118 27. Kline P An easy guide to factor analysis 2002 London Routledge 28. Carthey J. Improving Patient Safety Culture in Healthcare Organisations. In: Waterson P, editor. Patient Safety Culture, Theory, Methods and Application. Farnham: Ashgate Publishing Limited; 2014. p.139–58. 29. Flin R Burns C Mearns K Yule S Robertson E Measuring safety climate in health care Qual Saf Health Care 2006 15 109 115 10.1136/qshc.2005.014761 16585110 30. Sorra J, Famolaro T, Yount N, et al. Nursing Home Survey on Patient Safety Culture 2014 User Comparative Database Report. (Prepared by Westat, Rockville, MD, under Contract No. HHSA290201300003C). Rockville, MD: Agency for Healthcare Research and Quality; AHRQ Publication No. 15-0004. 2014. http://www.ahrq.gov/professionals/quality-patient-safety/patientsafetyculture/nursing-home/2014/index.html. 31. Raykov T Reliability if deleted, not ‘alpha if deleted’: evaluation of scale reliability following component deletion Br J Math Stat Psychol 2007 60 Pt 2 201 216 10.1348/000711006X115954 17971267 32. Raykov T Penev S Estimation of maximal reliability for multiple-component instruments in multilevel designs Br J Math Stat Psychol 2009 62 Pt 1 129 142 10.1348/000711007X255345 18001517 33. Braithwaite J, Matsuama Y, Mannion R, Johnson J (ed.). Healhtcare Reform, Quality and Safety. Farnham: Ashgate; 2015. 34. Ulrich B, Kear T. Patient Safety and Patient Safety Culture: Foundations of Excellent Health Care delivery. Nephrol Nurs J. 2014;41(5):447-56, 505. 35. Dixon-Woods M, Baker R, Charles K, Dawson J, Jerzembek G, Martin G, McCarthy I, McKee L, Minion J, Ozieranski P, et al. Culture and behaviour in the English National Health Service: overview of lessons from a large multimethod study. BMJ Qual Saf. 2013;0:1-10.

PMC005xxxxxx/PMC5002112.txt

==== Front BMC ObesBMC ObesBMC obesity2052-9538BioMed Central London 11610.1186/s40608-016-0116-2Research ArticleA comparison of snack serving sizes to USDA guidelines in healthy weight and overweight minority preschool children enrolled in Head Start Charvet Andrea achar065@fiu.edu 1Brogan Hartlieb Kathryn hartlieb@fiu.edu 1Yeh Yulyu dx7053@wayne.edu 2Jen K.-L. Catherine cjen@wayne.edu 21 Department of Dietetics and Nutrition, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA 2 Department of Nutrition and Food Science, College of Liberal Arts and Sciences, Wayne State University, 3009 Science Hall, Detroit, MI 48202 USA 27 8 2016 27 8 2016 2016 3 1 3626 9 2015 14 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 Obesity disproportionately affects children from low-income families and those from racial and ethnic minorities. The relationship between snacking and weight status remains unclear, although snacking is known to be an important eating episode for energy and nutrient intake particularly in young children. The purpose of this pilot study was to examine the snack intake of minority preschool children enrolled in the Head Start Program in four centers in Detroit, Michigan, and investigate differences by child weight status. Methods This secondary data analysis used snack time food observation and anthropometric data from a convenience sample of 55 African American children (44 % girls, mean age = 3.8 years). Snack intake data was obtained over a mean of 5 days through direct observation of children by dietetic interns, and later converted into food group servings according to the United States Department of Agriculture (USDA) meal patterns and averaged for each child. Height and weight measurements were systematically collected and BMI-for-age percentiles were used to classify children into weight categories. One sample, paired samples and independent samples t-tests were performed to test for differences within and between means. Results Based on BMI-for-age percentiles, 72.7 % of the sample was under/healthy weight and 27.3 % was overweight/obese. Average (mean ± SD) intake of milk (0.76 ± 0.34) and overall fruits/vegetables (0.77 ± 0.34) was significantly lower than one USDA serving, while average intake of grains and breads (2.04 ± 0.89), meat/meat alternates (2.20 ± 1.89) and other foods (1.43 ± 1.08) was significantly higher than one USDA serving (p ≤ 0.05). Children ate more when offered canned versus fresh fruits (0.93 ± 0.57 vs. 0.65 ± 0.37, p = 0.007). Except for a significantly higher milk intake in the overweight/obese group compared to the under/healthy weight group (0.86 ± 0.48 vs. 0.72 ± 0.27, p = 0.021], no relationship was found between snack food intake and weight category. Only in the overweight/obese group was the intake of milk and fresh fruits not significantly different than one USDA serving. Conclusions Findings suggest that regardless of weight status low-income minority preschool children are consuming larger serving sizes when offered less healthy versus healthier snack foods. Continued efforts should be made to provide healthful snack foods at preschool settings to prevent obesity and promote healthier food habits. Keywords SnackServing sizeUSDAOverweightChildrenMinorityHead Starthttp://dx.doi.org/10.13039/100006710Wayne State Universityissue-copyright-statement© The Author(s) 2016 ==== Body Background The incidence of childhood obesity increased dramatically in the past 30 years, and has become a major concern in children’s health outcomes. Data from the 2011–2012 National Health and Nutrition Examination Survey (NHANES) show that 22.8 % of all children ages 2 to 5 years old are overweight or obese [1]. Despite a leveling off or slight decrease in childhood obesity rates in some populations [1–3], the obesity rates remain high particularly among children from low-income families and those from racial and ethnic minorities. Among African American children aged 2–5 years, despite a decrease in combined overweight and obesity rates between 2007–2008 and 2011–2012, obesity alone rates remained unchanged at approximately 11.4 % [1, 4]. Childhood obesity has serious psychosocial and physical consequences. Overweight and obese children are more likely to become targets of discrimination, which can lead to depression [5, 6]. Overweight or obese children have a significantly higher risk of becoming overweight or obese adults when compared to healthy weight children [7–9]. Obesity is a risk factor for many chronic diseases, including diabetes, coronary heart disease, dyslipidemias, asthma, certain cancers, and arthritis [10–12]. Preschool children from minority families are at a higher risk for elevated cardiovascular disease biomarkers [13]. For example, Brogan et al. reported that at least 40 % of 3–5 year-old children in a sample composed of 92.5 % African Americans presented borderline or high-risk levels of high-density lipoprotein and triglycerides [14]. Obesity in childhood also generates a large economic burden. Children who are diagnosed with obesity have significantly higher healthcare costs when compared to healthy weight children [15]. Furthermore, if obesity and overweight prevalence in the U.S. population continues to increase at current rates, the total healthcare costs are estimated to more than double during each subsequent decade [16]. The high incidence and serious consequence of childhood obesity at an early age, especially among minorities, highlights the importance of strategies targeting a healthy weight on minority groups in the preschool years. While the causes of obesity include environmental and individual factors [2], this paper will focus on snacking, which is likely to play an important role in the development of overweight and obesity [17, 18]. Snacks can serve as a potential way to introduce new foods and to offer nutrients not consumed at other meals. Snacking has been defined in different ways [19]. Given the range of definitions, for the purpose of this study a snack was defined as a meal served between lunch and dismissal from childcare at Head Start centers. Healthy snacking between the three main meals has been suggested as a way to improve the quality of the food intake in children even though the relationship between snacking behaviors and weight status in children remains unclear [20, 21]. A literature review by Larson and Story addressing the implications of snacking on weight status found mixed results, the majority of cross-sectional and longitudinal studies either found no relationship between snacking behavior and weight status or found that children consuming snacks between meals were less likely to be obese [21]. Maffeis and colleagues assessed snack intake in 8- to 10-year old children and its relationship to body size and reported that obese and overweight children are not eating significantly more snacks than normal weight children, but are eating significantly more energy dense snacks, with a preference to those with a salty taste [20]. Findings from a study by Evans et al. suggest that the number of snacks is positively associated with a better diet quality as evidenced by an increased Healthy Eating Index (HEI) 2005 score in elementary school-age children [22]. Snacking frequency and the contribution of energy from snacks to total daily energy intake has been increasing over the years. Cross et al. evaluated snacking patterns among adults and children in the U.S. and reported that 92.6 % of elementary school aged children snacked once a day or more [23]. Between the years of 1977 and 1996 the percent of preschool aged children snacking at least once every day increased from 79 to 94 %, and the mean number of snacks per day increased significantly from 1.73 to 2.29, with the proportion of energy from snacks increasing from 19 to 24 % [24]. New data from the United States Department of Agriculture (USDA) reports that in 2011–2012 snacks contributed to approximately 30 % of total daily energy intake in preschool aged children [25]. Given the high prevalence of snacking among young children and its potential impact on diet quality, snacking behaviors continue to be an important area of study. Children ages six and younger spend, on average, 25 h per week in child-care settings, with approximately two out of three preschool-aged children being enrolled in some type of non-parental childcare program [26]. It has been suggested that childcare experiences may have a significant influence on eating habits, and consequently on weight status in childhood [27]. Head Start is a Federal program that provides preschool services in center-based settings for low-income children between the ages of 0 and 5 years [28], therefore it was a good setting for examining snacking in low-income minority children. Understanding differences in snacking based on weight status may provide valuable data for obesity prevention interventions. The purpose of this pilot study was two-fold, first to examine snack intake of minority preschool children enrolled in the Head Start Program in four centers in Detroit, Michigan, in relationship to USDA serving size guidelines for reimbursable snacks; and second, to investigate differences in snacking by child weight status. It was hypothesized that snack intake would be significantly different between children who are underweight and healthy weight and children who are overweight and obese. Methods Setting and participants This secondary data analysis used de-identified snack time food observation and anthropometric data collected at baseline from the intervention study “Healthy Kids Healthy Lives” [29], during the fall of 2008. The Wayne State University Institutional Review Board approved the original study protocol and written informed consent was obtained from all caregivers. The study was conducted through a Head Start system with 15 sites in the city of Detroit, Michigan. A convenience sample of 60 African American children aged 3 to 5 years were recruited from four of the Head Start sites. To be included in the study children were eligible to participate in the Head Start program, which include meeting federal poverty guidelines to evaluate family income, as well as consenting to participate and having attended preschool on food observation days. Children with special needs were excluded from study enrollment as this may impact their snack intake and body weight. The Institutional Review Board at Florida International University approved the current study. Snack intake Snack intake was obtained from the children through direct observation by dietetic interns (DI) twice per week during their community nutrition internship rotation, between October 14th and December 12th, 2008. Attendance of children to the Head Start center is variable. Snack intake observation was recorded for all children present the days the DI’s were at the centers. Meals served at the Head Start centers are family-style, meaning children select their own portions and serve themselves [30]. Prior to completing the food observations, the DIs reviewed visual portion size materials. Groups of 3 to 8 children sat in semicircular tables with a Head Start teacher leading the family-style snack period. The DIs sat in an unobtrusive location during the snacking episodes at the Head Start sites and recorded on paper the amount of each snack food the child consumed. Portions consumed were determined by visual estimates of volume since weighting foods was not feasible at the sites. Head Start centers provide meals and snacks free of charge. In order for the centers to receive reimbursement from the U.S. government, they are required to follow certain guidelines established by the USDA for the meals they serve. These guidelines require that meals and snacks include a minimum number of age-appropriate servings from 4 food groups: milk, fruits and vegetables, grains and breads, and meat and meat alternates. A reimbursable snack must include at least two of the four groups [31]. Table 1 provides a summary of the food groups and serving sizes per USDA regulations for children ages 3 through 5 years, along with examples of snack foods served across the four Head Start centers observed for this study. Because the snack foods served at the four observed centers were similar, they are grouped together when reported.Table 1 USDA reimbursable snack pattern for child carea, ages 3–5, with examples of snacks served Food group Minimum serving size Examples of snacks served at the Head Start centers Milk  Fluid milk 1/2 cupb 2 % reduced fat milk Vegetables/Fruits  Full strength juice or 1/2 cup Orange slices, grapes, applesauce, canned pineapple, fruit cocktail, banana, canned peaches, carrots  Fruit and/or vegetable 1/2 cup Grains/Breadc  Bread or 1/2 slice Pumpkin bread, whole-wheat bagel, raisin bran muffin, blueberry muffin, oatmeal cookie, peanut butter cookie, whole-wheat crackers, animal crackers, Rice Krispies, Honey Nut Cheerios, Raisin Bran, cupcake, yellow cake, tortilla chips  Cornbread, biscuits, rolls, muffins or 1/2 serving  Cold dry cereal or 1/4 cup  Hot cooked cereal or 1/4 cup  Cooked pasta or noodles or grains 1/4 cup Meat/Meat Alternate  Meat or poultry or fishd or 1/2 oz Sliced turkey, turkey and beans chili, cheddar cheese, cottage cheese, scrambled eggs, peanut butter, Light & Fit yogurt, vanilla yogurt, strawberry yogurt  Alternate protein product or 1/2 oz  Cheese or 1/2 oz  Egge or 1/2 egg  Cooked dry beans or peas or 1/8 cup  Peanut or other nut or seed butter or 1 Tbsp.  Nuts and/or seeds or 1/2 oz  Yogurtf 2 oz or 1/4 cup aAdapted from the U.S. Department of Agriculture (USDA), Food and Nutrition Service (FNS). Federal regulations, part 226-Child and Adult Care Food Program [33] bA cup means a standard measuring cup cBreads and grains must be made from whole-grain or enriched meal or flour. Cereal must be whole-grain or enriched or fortified dA serving consists of the edible portion of cooked lean meat or poultry or fish eOne-half egg meets the required minimum amount (1 oz or less) of meat alternate fYogurt may be plain or flavored, unsweetened or sweetened Certain food items do not meet the requirement for any of the four food groups, but may still be used. The USDA Food Buying Guide for Child Nutrition Programs uses the term “other foods” and includes such items since they are frequently used as condiments and seasonings, to round out the meal, to improve acceptability, and to satisfy children’s appetites [32]. Other foods served at the Head Start sites included items such as cream cheese, cheese dip, and ice cream. Anthropometric measurements Weight and height measurements were collected from all participating children before beginning the study. Trained graduate student data collectors weighed children without coats and shoes using a portable digital scale (Tanita Model BC551). Two measures were obtained from each child. If the two readings were within 0.1 kg, an average was taken and used as the body weight. If the two readings differed by more than 0.1 kg, a third reading was taken and readings were averaged. Heights were obtained using a portable stadiometer (Seca 214, Seca North America East, Hanover, MD). Two measurements without shoes were recorded. If the two readings were within 0.2 cm of each other the readings were averaged and recorded. If the difference was more than 0.2 cm, a third measurement was taken before averaging. Weight and height measurements were used to calculate body mass index (BMI). BMI-for-age percentiles were used to categorize the child as underweight (<5 %), healthy weight (5–84.9 %), overweight (85–94.9 %) and obese (≥95 %), following the Centers for Disease Control and Prevention (CDC) guidelines for BMI-for-age weight status categories [33]. Details on the method of calculating BMI and its expression as BMI-for-age percentiles can be found on the CDC website [33]. Data analysis For the purpose of this study, snack intake was converted into food group servings based on the USDA meal patterns as the minimum amount of food that can be served to comply with the licensing standards for adequate nutrition for children between 3 and 5 years of age [34]. To assist in determining serving sizes for foods not included on the USDA meal patterns the authors contacted the U.S. Department of Agriculture Food and Nutrition Services and the Bureau of Child Care Food Program. Per recommendations the USDA Food Buying Guide for Child Nutrition Programs was used [32]. Food groups included milk, fruits and vegetables, grains and breads, meat and meat alternates, and other foods. In order to explore snack food intake quality, food groups were further categorized by their healthfulness. The fruits and vegetables food group was subdivided into fresh fruits and canned fruits as indicated by the measurement. In the instances that fresh vs. canned fruit was not specified by the DI (66 % of fruit intake records) pineapple, pears, peaches and apricots were assigned canned fruit, while bananas, apples, grapes and oranges were assigned fresh fruit. The grains and bread food group was subdivided into low sugar/low fat grains and breads and high sugar/high fat grains and breads. Low sugar/low fat grains and breads included breakfast cereal, wheat crackers, wheat bagels, and whole wheat breads. High sugar/high fat grains and breads included cookies, graham and animal crackers, cupcakes and cakes. Though this study did not measure nutrient quality, it can be assumed that fresh fruits and low sugar/low fat grains and breads tend to be healthier than canned fruits and high sugar/high fat grains and breads. Average snack intake for each food component was calculated per individual. For each child, an absent day was not included in the data analysis. Meanwhile, a day the child was present at the Head Start center but did not consume a certain food component that was offered, intake was computed as “zero” and that day was counted for data analysis. BMI categories were grouped into dichotomized variables to allow for comparisons by weight status. Children who were underweight and healthy weight were grouped into a variable labeled underweight/healthy weight and children who were overweight and obese were grouped into a variable labeled overweight/obese. Data was entered into the computer and analyzed using the Statistical Package for Social Sciences version 21 software (SPSS, IBM Corp, Armonk, NY). A significance level of 5 % was used on all statistical tests performed. One sample t-tests were conducted to test for differences comparing to USDA serving sizes. Paired samples t-tests were conducted to test for differences within fresh fruits and canned fruits, and low sugar/low fat grains and breads and high sugar/high fat grains and breads. To test for equality of means for intake of each food component by weight category, independent samples t-tests were performed. Results Demographics From the 60 children recruited, a total of 55 African American children enrolled in the Head Start program were included in the study. Two children enrolled were excluded from the study due to missing anthropometric data, and three enrolled children were excluded due to missing snack intake data for the days they were present at the Head Start center. Study participants had a mean age of 46 months (3.8 years) ± 7.95, with 31 boys (56 %) and 24 girls (44 %). According to the BMI-for-age percentiles five children were underweight (9.1 %), 35 were at healthy weight (63.6 %), seven were overweight (12.7 %), and eight were obese (14.5 %). When weight categories were grouped, 72.7 % (n = 40) children were underweight/healthy weight, and 27.3 % (n = 15) were overweight/obese. Snack quantity and quality Snack time observation data recorded for the days each child was present at the Head Start site varied from 1 to 18 days of data record, because of children’s variability in attendance. On average a child was present for 5 days (median = 4 days, SD = 3.90). Milk (2 %) was the only food group offered daily. Fruits or vegetables were offered 13 out of 18 days (fresh fruits 5 days, canned fruits 8 days, vegetables 1 day, offered together with a fruit), grains and breads offered 11 days (low sugar/low fat 6 days, high sugar/high fat 5 days), meat and meat alternates 5 days and other foods were offered three out of the 18 days. Table 2 reports the children’s mean snack intake of each food group compared to one USDA serving. The mean snack intake of the milk and the fruits and vegetables group was significantly lower than one USDA serving (p < 0.001). However, the mean snack intake of grains and breads (p < 0.001), meats and meat alternates (p = 0.001), and other foods (p = 0.047) was significantly higher than one USDA serving. When the canned and fresh fruit categories were examined separately, only the mean snack intake of fresh fruits was significantly lower than one USDA serving (p < 0.001). Canned fruits mean snack intake was not significantly different from one USDA serving (p = 0.124). Mean snack intake of low sugar/low fat and high sugar/high fat grains and breads were both greater than one USDA serving.Table 2 Intake (mean ± SD) per snacking episode in number of servings, compared to one USDA serving Food group na Mean intake ± SD p-value Milk 55 0.76 ± 0.34 <0.000** Fruits and Vegetables Overall 53 0.77 ± 0.34 <0.000**  Fresh fruits 40 0.67 ± 0.34 <0.000**  Canned fruits 43 0.87 ± 0.53 0.124 Grains and Breads Overall 51 2.04 ± 0.89 <0.000**  Low sugar/low fatb 44 1.75 ± 0.71 <0.000**  High sugar/high fatc 42 2.12 ± 1.10 <0.000** Meat and Meat Alternatesd 34 2.20 ± 1.89 0.001** Other Foodse 27 1.43 ± 1.08 0.047* Statistics are based on all observations with valid data an represents the total number of children analyzed per food group. Milk was the only food group offered daily, for which all participating children had a record of intake bLow sugar/low fat grains and breads included breakfast cereal, wheat crackers, wheat bagels, whole wheat breads cHigh sugar/high fat grains and breads included cookies, graham and animal crackers, cupcakes and cakes dMeat and meat alternates included cheeses, yogurt, eggs, deli meats, and peanut butter eOther foods included cream cheese, cheese dip, and ice cream * 95 % significance level ** 95 % significance level When comparing child snack intake by healthy and less healthy categories, paired samples test showed that children had significantly higher fruit intake when offered canned fruits vs. fresh fruits (0.93 ± 0.57 vs. 0.65 ± 0.37 servings respectively, p = 0.007). No difference was found when comparing mean snack intake of low sugar/low fat grains and breads with high sugar/high fat grains and breads (1.72 ± 0.71 vs. 1.83 ± 0.89 servings respectively, p = 0.562), consumption of both categories was significantly higher than one USDA serving. Snack intake by child weight status Independent samples test results showed that the only food group for which intake differed significantly between the overweight/obese and underweight/healthy weight categories was milk. The overweight/obese children consumed more milk than the underweight/healthy weight children (0.86 ± 0.48 vs. 0.72 ± 0.27, p = 0.021). The milk and fresh fruits intake of overweight/obese children was not significantly different from one USDA serving in single sample t-test analysis, while the underweight/healthy weight group’s intake was significantly lower than one USDA serving. Table 3 presents the data for snack intake by child weight status.Table 3 Intake (mean ± SD) by weight category per snacking episode in number of servings, and comparison to one USDA serving Food group Under/healthy weight Overweight/obese p-valueb na Mean intake ± SD p-valuec n Mean intake ± SD p-valuec Milk 40 0.72 ± 0.27 <0.000** 15 0.86 ± 0.48 0.288 0.021* Fruits & Vegetables Overall 39 0.77 ± 0.33 <0.000** 14 0.76 ± 0.37 0.032* 0.644  Fresh fruits 32 0.64 ± 0.35 <0.000** 8 0.78 ± 0.32 0.092 0.860  Canned fruits 30 0.90 ± 0.54 0.327 13 0.80 ± 0.52 0.206 0.761 Grains & Breads Overall 37 2.06 ± 0.95 <0.000** 14 1.97 ± 0.75 <0.000** 0.658  Low sugar/low fatd 30 1.68 ± 0.70 <0.000** 14 1.91 ± 0.74 0.001** 0.649  High sugar/high fate 33 2.14 ± 1.17 <0.000** 9 2.01 ± 0.82 0.006** 0.296 Meat and Meat Alternatesf 24 2.29 ± 2.14 0.007** 10 1.97 ± 1.13 0.025* 0.100 Other Foodsg 19 1.27 ± 1.06 0.277 8 1.81 ± 1.10 0.075 0.954 Statistics are based on all observations with valid data an represents the total number of children analyzed per food group. Milk was the only food group offered daily, for which all participating children had a record of intake b p-value is referent to differences in mean intake between the under/healthy weight group and the overweight/obese group c p-value is referent to differences in mean intake from one USDA serving dLow sugar/low fat grains and breads included breakfast cereal, wheat crackers, wheat bagels, whole wheat breads eHigh sugar/high fat grains and breads included cookies, graham and animal crackers, cupcakes and cakes fMeat and meat alternates included cheeses, yogurt, eggs, deli meats, and peanut butter gOther foods included cream cheese, cheese dip, and ice cream Discussion The first aim of this study was to examine the snack intake of minority preschool children enrolled in the Head Start program in relationship to USDA serving size guidelines for reimbursable snacks. The minority preschool children in our population consumed significantly less than one USDA snack serving of milk and fruits and vegetables, while eating significantly more than one serving of grains and breads, meat and meat alternates, and other foods. We also observed a higher intake of canned fruits over fresh fruits. The literature suggests that the intake pattern of our participants, if consistent throughout other meals during the day, could be obesity promoting. An inverse association exists between fruit and vegetable intake and body weight, with a potential effect in reducing the risk for childhood obesity [35, 36]. Adequate milk intake is also associated with a reduced risk for overweight and obesity [37, 38]. On the other hand, foods belonging to the grains and breads, meat and meat alternates, and other foods groups tend to be higher in calories, and many times in fats and added sugars, than milk and fruits and vegetables. The observed elevated intake when compared to USDA guidelines of snack foods that tend to be higher in calories is consistent with other research. For example, Ford et al. while examining changes in dietary intake among 2- to 6-year-old children from 1989 to 2008 found a significant increase in total daily energy intake by 109 kcal in parallel with an increase in foods high in added sugars, solid fats, and sodium in the preschooler diet [39]. Poti and Popkin examined trends in energy intake in children aged 2 to 18 years old by eating location and food source and found an increase in total daily energy intake by 179 kcal from 1977 to 2006, combined with an increase in foods eaten away from home, which have been associated with a higher energy density and lower nutritional quality [40]. Portion size for energy dense foods has also increased over time, and has been related to excess energy intake by children of all ages [41–43]. Therefore an elevated intake of energy dense foods, combined with larger portion sizes, may negatively affect diet quality and contribute excess calories placing children at risk for excess weight in the future. The second aim of this study was to investigate differences in snacking by child weight status. Our sample had a higher prevalence of overweight/obesity (27.3 %) compared to the reported national level of 22.8 % for this age group [1]. Results from this study did not identify any significant relationship between the observed snack intake and child weight status. Other factors not included in this analysis such as foods from other meals, which contribute to approximately 70 % of total daily energy intake in children ages 2 to 5 years [25] may be influencing body weight in this sample, The Head Start sites included in this study are located in low-income neighborhoods, presenting high unemployment and crime rates, which may also have contributed to the increased prevalence of obesity. Lower socioeconomic status neighborhoods are at a higher risk of obesity [44, 45]. Overall our findings are consistent with previous studies, which also could not establish a relationship in preschool children between portion size, healthfulness, and obesity, although researchers report this behavior presents itself as obesity promoting [6, 46]. A significantly higher intake of milk by overweight/obese children when compared to the underweight and healthy weight group was observed. However, neither of the two groups exceeded the intake of one USDA serving for milk. Dairy products are the most important source of calcium in children’s diets and studies with children demonstrated that a higher intake of dairy products is negatively associated to body weight [37, 38, 47]. Our findings did not support the findings from other studies. While overweight/obese children’s intake of milk was not significantly different from one USDA serving, underweight/healthy weight children’s milk intake stayed significantly below it. Milk served at the Head Start centers included in our sample was 2 % reduced fat. Studies showing a protective effect of milk against overweight and obesity usually compare low fat versus whole milk. This may be a reason why our study did not find similar results. Milk remains an important food component on the diets of children and there is a need to meet recommendations for milk intake. Findings also suggest that Head Start children are consuming significantly more than one USDA serving of grains and breads, whether offered healthy or less healthy options. Data supports the need to encourage the consumption of healthier grains and breads for snacks at preschool settings serving minority low-income children. The literature promotes offering only healthier options of such food items in preschool menus. [48]. Food preferences develop mostly during early childhood [39], making this group a potential target for nutrition interventions aimed at promoting healthy eating habits. Skinner et al. demonstrated that newly tasted foods are significantly more likely to be accepted between children 2 to 4 years old than by children 8 years old [49]. Other studies confirm the same finding: the earlier a food is introduced, the more likely it is that the child will like and consume such food item [50–52]. The observed higher intake of less healthy snacks when such were offered instead of healthier snack options, independently of weight category, suggests that there is a need to encourage preschool centers to offer a variety of healthy food items for snacks, including fresh fruits and vegetables and low fat or fat-free milk. The convenience small sample size for this study should not be overlooked as a limitation, however little is known about the snacking habits of minority preschool children when eating at the preschool sites. Moreover, the homogeneity of the sample concerning socioeconomic status, ethnicity, education level, and neighborhood surroundings reduce the generalizability of the results yet provide insight to a vulnerable population. As a strength, in the present study snack intake was recorded through direct observation by trained dietetic interns, reducing the likelihood of over or under reporting common when dietary recalls are used. It should be pointed out that this study only considered one eating episode throughout the entire day. Other meals served at the Head Start centers were not observed, and no information was collected about the snack quality and quantity when the preschoolers were at home. This deserves further investigation. Conclusions Findings from this study suggest that low-income minority preschool children are consuming larger serving sizes when offered less healthy versus healthier snack food items while eating at preschool. This snacking pattern did not show any significant relationship to the current weight status of the children. Future studies with larger sample sizes examining snacking differences in overweight/obese and underweight/healthy weight preschool children are encouraged. Studies considering more eating episodes throughout the day to give a better understanding of the children’s intake are needed. Efforts to provide healthful snack foods at the preschool setting will promote continued exposure to healthy foods. Over time this may lead to an increased acceptance of such food items and prevention of obesity. Abbreviations BMIBody mass index CDCCenters for Disease Control and Prevention DIDietetic interns NHANESNational Health and Nutrition Examination Survey USDAUnited States Department of Agriculture Acknowledgements Jessica A. Michaloski, RD, LD–Public Health Nutrition Consultant, Florida Department of Health, Bureau of Child Care Food Program, Division of Community Health Promotion, Tallahassee, Florida, and Laura Carroll, Food and Nutrition Service Intern, for providing clarification about serving sizes for reimbursable meals under USDA federal regulations. Elaine McLaughlin, Center for Nutrition Policy and Promotion, USDA, for her dedication in helping find the right person to answer the questions. Funding A.C. was supported by NIH/NIGMS R25 GM061347. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The parent study was supported by Wayne State University Research Enhancement Fund. Availability of data and materials The authors can be contacted regarding inquiries to the data set and data sharing. Authors’ contributions AC performed data extraction, analysis and interpretation, contributed to study design and drafted the manuscript. KBH made substantial intellectual contribution to study design, data analysis and interpretation, acquisition of data, and critically revised the manuscript for content throughout the writing stages. YY and KLCJ made substantial contribution to study conception and were responsible 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 This study has been performed in accordance with the Declaration of Helsinki. The Wayne State University Institutional Review Board approved the original study protocol and written informed consent was obtained from all caregivers. The Institutional Review Board at Florida International University approved the current study. ==== Refs References 1. Ogden CL Carroll MD Kit BK Flegal KM Prevalence of childhood and adult obesity in the United States, 2011–2012 JAMA 2014 311 8 806 814 10.1001/jama.2014.732 24570244 2. Swinburn BA Sacks G Hall KD The global obesity pandemic: Shaped by global drivers and local environments Lancet 2011 378 9793 804 814 10.1016/S0140-6736(11)60813-1 21872749 3. Robert Wood Johnson Foundation, Issue Brief, June 2016. Declining childhood obesity rates. http://www.rwjf.org/en/library/research/2016/06/declining-childhood-obesity-rates.html. Accessed 17 Aug 2016. 4. Ogden CL Carroll MD Curtin LR Lamb MM Flegal KM Prevalence of high body mass index in US children and adolescents, 2007–2008 JAMA 2010 303 3 242 249 10.1001/jama.2009.2012 20071470 5. Campbell K Crawford D Family food environments as determinants of preschool-aged children’s eating behaviors: Implications for obesity prevention policy. A review Aust J Nutr Diet 2001 58 19 25 6. Kuhl ES Clifford LM Stark LJ Obesity in preschoolers: Behavioral correlates and directions for treatment Obesity (Silver Spring, Md) 2012 20 1 3 29 10.1038/oby.2011.201 7. Serdula MK Ivery D Coates RJ Freedman DS Williamson DF Byers T Do obese children become obese adults? A review of the literature Prev Med 1993 22 2 167 177 10.1006/pmed.1993.1014 8483856 8. Eriksson J Forsén T Tuomilehto J Osmond C Barker D Size at birth, childhood growth and obesity in adult life Int J Obes Relat Metab Disord: J Int Assoc Study Obes 2001 25 5 735 740 10.1038/sj.ijo.0801602 9. Freedman DS Khan LK Serdula MK Dietz WH Srinivasan SR Berenson GS The relation of childhood BMI to adult adiposity: The Bogalusa heart study Pediatrics 2005 115 1 22 27 15629977 10. Kelsey MM Zaepfel A Bjornstad P Nadeau KJ Age-related consequences of childhood obesity Gerontology 2014 60 3 222 228 10.1159/000356023 24434909 11. Manger WM Manger LS Minno AM Obesity prevention in young schoolchildren: Results of a pilot study J Sch Health 2012 82 10 462 468 10.1111/j.1746-1561.2012.00723.x 22954165 12. Baker JL Olsen LW Sørensen TIA Childhood body-mass index and the risk of coronary heart disease in adulthood N Engl J Med 2007 357 23 2329 2337 10.1056/NEJMoa072515 18057335 13. Messiah SE Arheart KL Natale RA Hlaing WM Lipshultz SE Miller TL BMI, waist circumference, and selected cardiovascular disease risk factors among preschool-age children Obesity (Silver Spring, Md) 2012 20 9 1942 1949 10.1038/oby.2011.353 14. Brogan K Danford C Yeh Y Jen KC Cardiovascular disease risk factors are elevated in urban minority children enrolled in head start Child Obesity (Print) 2014 10 3 207 213 10.1089/chi.2013.0146 15. Hampl SE Carroll CA Simon SD Sharma V Resource utilization and expenditures for overweight and obese children Arch Pediatr Adolesc Med 2007 161 1 11 14 10.1001/archpedi.161.1.11 17199061 16. Wang Y Beydoun MA Liang L Caballero B Kumanyika SK Will all Americans become overweight or obese? Estimating the progression and cost of the US obesity epidemic Obesity (Silver Spring, Md) 2008 16 10 2323 2330 10.1038/oby.2008.351 17. Nicklas TA Morales M Linares A Children’s meal patterns have changed over a 21-year period: The Bogalusa heart study J Am Diet Assoc 2004 104 5 753 761 10.1016/j.jada.2004.02.030 15127060 18. Savige G MacFarlane A Ball K Worsley A Crawford D Snacking behaviours of adolescents and their association with skipping meals Int J Behav Nutr Phys Act 2007 4 36 10.1186/1479-5868-4-36 17868479 19. Patro B Szajewska H Meal patterns and childhood obesity Curr Opin Clin Nutr Metab Care 2010 13 3 300 304 10.1097/MCO.0b013e32833681a2 20075720 20. Maffeis C Grezzani A Perrone L Del Giudice EM Saggese G Tatò L Could the savory taste of snacks be a further risk factor for overweight in children? J Pediatr Gastroenterol Nutr 2008 46 4 429 437 10.1097/MPG.0b013e318163b850 18367957 21. Larson N Story M A review of snacking patterns among children and adolescents: What are the implications of snacking for weight status? Child Obes 2013 9 2 104 115 10.1089/chi.2012.0108 23470091 22. Evans EW Jacques PF Dallal GE Sacheck J Must A The role of eating frequencyon total energy intake and diet quality in a low-income, racially diverse sample of school children Public Health Nutr 2015 18 3 474 481 10.1017/S1368980014000470 24780506 23. Cross AT Babicz D Cushman LF Snacking patterns among 1,800 adults and children J Am Diet Assoc 1994 94 12 1398 1403 10.1016/0002-8223(94)92542-9 7963190 24. Jahns L Siega-Riz AM Popkin BM The increasing prevalence of snacking among US children from 1977 to 1996 J Pediatr 2001 138 4 493 498 10.1067/mpd.2001.112162 11295711 25. US Department of Agriculture, Agricultural Research Service. 2014. Snacks: Percentage of selected nutrients contributed by food and beverages consumed at snack occasions, by gender and age, What we eat in America, NHANES 2011–2012. http://www.ars.usda.gov/SP2UserFiles/Place/80400530/pdf/1112/Table_25_SNK_GEN_11.pdf. Updated October 2, 2014. Accessed 17 Aug 2016. 26. Erinosho TO Hales DP McWilliams CP Emunah J Ward DS Nutrition policies at child-care centers and impact on role modeling of healthy eating behaviors of caregivers J Acad Nutr Diet 2012 112 1 119 124 10.1016/j.jada.2011.08.048 22709641 27. Larson N Ward DS Neelon SB Story M What role can child-care settings play in obesity prevention? A review of the evidence and call for research efforts J Am Diet Assoc 2011 111 9 1343 1362 10.1016/j.jada.2011.06.007 21872698 28. Child and adult care food program (CACFP) child day care centers. U.S. Department of Agriculture, Food and Nutrition Service Web site. http://www.fns.usda.gov/cacfp/child-day-care-centers. Updated January 27, 2014. Accessed 17 Aug 2016. 29. Shen B Lee T Janisse H Brogan K Jen C African-American preschoolers’ physical activity levels in Head Start Res Q Exerc Sport 2012 83 2 168 174 10.1080/02701367.2012.10599847 22808702 30. Benjamin Neelon SE Briley ME Position of the American Dietetic Association: Benchmarks for nutrition in child care J Am Diet Assoc 2011 111 4 607 615 10.1016/j.jada.2011.02.016 21443997 31. U.S. Department of Agriculture (USDA), Food and Nutrition Service (FNS). Federal regulations, part 226-child and adult care food program. http://www.floridahealth.gov/programs-and-services/childrens-health/child-care-food-program/guidance/_documents/regulations.pdf. Accessed 17 Aug 2016. 32. U.S. Department of Agriculture (USDA), Food and Nutrition Service (FNS). Food buying guide for child nutrition programs. http://www.fns.usda.gov/tn/foodbuying-guide-child-nutrition-programs. Updated 2014. Accessed 17 Aug 2016. 33. Centers for Disease Control and Prevention (CDC). About BMI for children and teens. http://www.cdc.gov/healthyweight/assessing/bmi/childrens_BMI/about_childrens_BMI.html. Updated 2014. Accessed 17 Aug 2016. 34. U.S. Department of Agriculture (USDA), Food and Nutrition Service (FNS). Child care meal pattern. http://www.fns.usda.gov/sites/default/files/Child_Meals.pdf. Accessed 17 Aug 2016. 35. Vernarelli JA Mitchell DC Hartman TJ Rolls BJ Dietary energy density is associated with body weight status and vegetable intake in U.S. children J Nutr 2011 141 12 2204 2210 10.3945/jn.111.146092 22049295 36. Alinia S Hels O Tetens I The potential association between fruit intake and body weight-a review Obes Rev: Off J Int Assoc Study Obes 2009 10 6 639 647 10.1111/j.1467-789X.2009.00582.x 37. Barba G Troiano E Russo P Venezia A Siani A Inverse association between body mass and frequency of milk consumption in children Br J Nutr 2005 93 1 15 19 10.1079/BJN20041300 15705220 38. Carruth BR Skinner JD The role of dietary calcium and other nutrients in moderating body fat in preschool children Int J Obes Relat Metab Disord: J Int Assoc Study Obes 2001 25 4 559 566 10.1038/sj.ijo.0801562 39. Ford CN Slining MM Popkin BM Trends in dietary intake among US 2- to 6-year-old children, 1989–2008 J Acad Nutr Diet 2013 113 1 35 42 10.1016/j.jand.2012.08.022 23260722 40. Poti JM Popkin BM Trends in energy intake among US children by eating location and food source, 1977–2006 J Am Diet Assoc 2011 111 8 1156 1164 10.1016/j.jada.2011.05.007 21802561 41. Piernas C Popkin BM Increased portion sizes from energy-dense foods affect total energy intake at eating occasions in US children and adolescents: Patterns and trends by age group and sociodemographic characteristics, 1977–2006 Am J Clin Nutr 2011 94 5 1324 1332 10.3945/ajcn.110.008466 21918222 42. Piernas C Popkin BM Food portion patterns and trends among U.S. children and the relationship to total eating occasion size, 1977–2006 J Nutr 2011 141 6 1159 1164 10.3945/jn.111.138727 21525258 43. McConahy KL Smiciklas-Wright H Mitchell DC Picciano MF Portion size of common foods predicts energy intake among preschool-aged children J Am Diet Assoc 2004 104 6 975 979 10.1016/j.jada.2004.03.027 15175599 44. Rahman T Cushing RA Jackson RJ Contributions of built environment to childhood obesity Mt Sinai J Med 2011 78 1 49 57 10.1002/msj.20235 21259262 45. Story M Kaphingst KM Robinson-O’Brien R Glanz K Creating healthy food and eating environments: Policy and environmental approaches Annu Rev Public Health 2008 29 253 272 10.1146/annurev.publhealth.29.020907.090926 18031223 46. Pourshahidi LK Kerr MA McCaffrey TA Livingstone MBE Influencing and modifying children’s energy intake: The role of portion size and energy density Proc Nutr Soc 2014 73 3 397 406 10.1017/S0029665114000615 24886909 47. Skinner JD Bounds W Carruth BR Ziegler P Longitudinal calcium intake is negatively related to children’s body fat indexes J Am Diet Assoc 2003 103 12 1626 1631 10.1016/j.jada.2003.09.018 14647089 48. Fox MK Condon E Briefel RR Reidy KC Deming DM Food consumption patterns of young preschoolers: Are they starting off on the right path? J Am Diet Assoc 2010 110 12 Suppl S52 S59 10.1016/j.jada.2010.09.002 21092769 49. Skinner JD Carruth BR Wendy B Ziegler PJ Children’s food preferences: A longitudinal analysis J Am Diet Assoc 2002 102 11 1638 1647 10.1016/S0002-8223(02)90349-4 12449287 50. Cooke L The importance of exposure for healthy eating in childhood: A review J Hum Nutr Diet: Off J Br Diet Assoc 2007 20 4 294 301 10.1111/j.1365-277X.2007.00804.x 51. Harris G Development of taste and food preferences in children Curr Opin Clin Nutr Metab Care 2008 11 3 315 319 10.1097/MCO.0b013e3282f9e228 18403930 52. Birch LL Fisher JO Development of eating behaviors among children and adolescents Pediatrics 1998 101 3 Pt 2 539 549 12224660

PMC005xxxxxx/PMC5002113.txt

==== Front BMC Med EducBMC Med EducBMC Medical Education1472-6920BioMed Central London 73710.1186/s12909-016-0737-7Research ArticleParental satisfaction of U.S. physicians: associated factors and comparison with the general U.S. working population Shanafelt Tait D. shanafelt.tait@mayo.edu 13Hasan Omar 2Hayes Sharonne 1Sinsky Christine A. 2Satele Daniel 14Sloan Jeff 14West Colin P. 13Dyrbye Lotte N. 131 Mayo Clinic, Rochester, MN USA 2 American Medical Association, Chicago, IL USA 3 Department of Internal Medicine, 200 First Street, Rochester, MN 55905 USA 4 Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, MN USA 27 8 2016 27 8 2016 2016 16 1 2285 5 2016 12 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 Physicians work considerably longer hours and are less satisfied with work-life balance than U.S. workers in other fields. There is, however, minimal data on physicians’ parental satisfaction. Methods To evaluate differences in parental satisfaction among physicians and workers in other fields, we surveyed U.S. physicians as well as a probability-based sample of the general U.S. working population between August 2014–October 2014. Parental satisfaction and the perceived impact of career on relationships with children were evaluated. Results Among 6880 responding physicians (cooperation rate 19.2 %), 5582 (81.1 %) had children. Overall, physicians were satisfied in their relationships with their children, with 4782 (85.9 %) indicating that they were either very satisfied [n = 2738; (49.2 %)] or satisfied [n = 2044 (36.7 %)]. In contrast, less than half believed their career had made either a major [n = 1212; (21.8 %)] or minor positive [n = 1260; (22.7 %)] impact on their relationship with their children, with a slightly larger proportion indicating a major (n = 2071 [37.2 %]) or minor (n = 501 [9 %]) negative impact. Women physicians were less likely to believe their career had made a positive impact as were younger physicians. Hours worked/week inversely correlated with the belief that career had made a positive impact on relationships with children. Both men (OR: 2.75; p < 0.0001) and women (OR: 4.33; p < 0.0001) physicians were significantly more likely to report that their career had a negative impact on relationships with their children than the sex-matched U.S. working population. Conclusions U.S. physicians report generally high satisfaction in their relationships with their children. Despite their high satisfaction, physicians have a more negative perception of the impact of their career on relationships with their children than U.S. workers in general. Electronic supplementary material The online version of this article (doi:10.1186/s12909-016-0737-7) contains supplementary material, which is available to authorized users. Keywords PhysiciansDoctorsParentsSatisfactionChildrenCareerWork-life integrationissue-copyright-statement© The Author(s) 2016 ==== Body Background Being a physician is a demanding and rewarding endeavor. Physicians undergo long and arduous training, work dramatically more hours than workers in other fields, and are at high risk for professional burnout [1]. As a counterbalance to these factors, physicians engage in intellectually stimulating work, are well compensated relative to many professions, and often develop meaningful and fulfilling relationships with patients [2, 3]. The net effect of these positive and negative forces on each physician is unique and can vary over the course of their career [4–6]. The impact of a career in medicine on physicians’ personal relationships has been a subject of interest in recent years. This dimension has taken on new complexity as the proportion of women physicians and two-physician couples have increased. Despite the stereotypes that physicians neglect their partners and have high divorce rates, recent studies in the U.S. suggest divorce rates among physicians are lower than both the population in general as well as other professionals [7]. Recent studies also suggest high marital satisfaction among spouses/partners of U.S. physicians and indicate that the amount of time spent together each day is the dominant driver of relationship satisfaction among spouses/partners of physicians [8]. While these studies have provided insights regarding the marital satisfaction of physicians, one aspect that has not been well studied is the impact of a career in medicine on physicians’ relationships with their children. The largest study exploring this dimension to date included only 415 physicians, excluded single parents, and was derived from a survey conducted >25 years ago [9]. That study found that physicians were less satisfied in their relationships with their children than their relationships with their spouse, with comparable parental satisfaction among both women and men physicians. On multivariate analysis, greater parental satisfaction was found among physicians less than age 45 and among those whose spouse was supportive of their career, as well as those married to an individual working as a professional or as a homemaker [9]. Physicians working in an employed practice model also had higher parental satisfaction than those working in solo or group practice. A number of societal shifts have occurred since the time of that study, including an increase in the proportion of physicians who are women, a higher prevalence of single parents and a greater number of adults in two-career relationships [10]. Many changes have also occurred in the practice structure of medicine with a majority of physicians now functioning as employees of large medical organizations [11]. There are no data comparing the parental experience of physicians to that of the general U.S. population. Here we report the results of a large study evaluating the parental satisfaction of U.S. physicians along with comparison to the general U.S. population. Methods As previously reported [12], we surveyed a national sample of U.S. physicians from all specialties between August 2014-October 2014 using the AMA’s Physician Master File, a nearly complete record of all physicians in the United States. Participation was voluntary and results were anonymous. In accordance with established methodology, the 35,922 physicians who opened at least one invitation e-mail were considered to have received the invitation to participate in the study [13]. A probability-based sample of employed U.S. adults was also surveyed for comparison to physicians in October 2014. The population survey was conducted using Knowledge Panel®, a probability-based panel designed to be representative of the U.S. population. Participants in the KnowledgePanel® are initially chosen scientifically by a random selection of telephone numbers and residential addresses. Persons in selected households are then invited by telephone or by mail to participate. Additional technical information is available at: http://www.gfk.com/products-a-z/us/knowledgepanelr-north-america/. The Mayo Clinic Institutional Review Board approved this study. The initial results of this study, including the personal and professional characteristics of participating physicians, assessment of physician well-being (e.g. burnout, symptoms of depression, suicidal ideation, satisfaction with work/life balance, career satisfaction), and comparison of physician well-being to that of working U.S. adults have been previously reported [12]. Parental satisfaction Participating physicians were asked whether they had any children. Those who indicated they had children were asked the age of their youngest child and questions regarding their parental satisfaction (questions provided in Additional file 1). One item asked physicians to rate their overall satisfaction with the relationship they have with their children (response options: very satisfied, satisfied, neither satisfied nor dissatisfied, dissatisfied, very dissatisfied). The second item asked physicians to rate the impact their career had made on their relationship with their children (response options: major positive impact, minor positive impact, no impact, minor negative impact, major negative impact). These items were formatted to be similar to previously used items evaluating marital satisfaction [1, 8, 9]. The second question was also asked of working U.S. adults. Statistical analysis Standard descriptive summary statistics were used to characterize physician and control samples. Associations between variables were evaluated using the Kruskal-Wallis test or Chi Square test as appropriate. All tests were two-sided with Type I error rates of 0.05. Multivariate analysis of differences across specialties was performed using logistic regression. Pooled multivariate logistic regression analysis of physicians with respect to the impact of their career on their relationship with their children was also performed to identify personal and professional factors associated with the dependent variable. For comparisons with population controls, physician data was restricted to responders between the ages of 29 and 65 years of age who were not retired in order to match the age and employment status of the U.S. population sample. All analyses were done using SAS version 9 (SAS Institute, Inc., Cary, North Carolina). Results Of the 35,922 physicians who received an invitation to participate, 6880 (19.2 %) completed surveys [12]. Among the 6880 responding physicians, 5582 (81.1 %) indicated that they had children. The personal and professional characteristics of physicians with children, as well as all responding physicians, are shown in Table 1.Table 1 Characteristics of physicians with children relative to all physician responders Characteristics 2014 Responders N = 6880 Parents only N = 5582 Gender  Male 4497(67.5 %) 3986 (71.9 %)  Female 2162(32.5 %) 1556 (28.1 %)  Missing 221 40 Age in Years  Median 56 57  </= 35 466 (7.0 %) 244 (4.4 %)  36–45 1212 (18.2 %) 919 (16.6 %)  46–55 1484 (22.3 %) 1240 (22.4 %)  56–65 2206 (33.1 %) 1937 (35.0 %)  >65 1289 (19.4 %) 1201 (21.7 %)  missing 223 41 Relationship Status  Missing 209 32  Single 722 (10.8 %) 303 (5.5 %)  Married 5573 (83.5 %) 5004 (90.2 %)  Partnered 274 (4.1 %) 156 (2.8 %)  Widowed 102 (1.5 %) 87 (1.6 %) Age Youngest Child  <5 740 (11.2 %) 740 (13.3 %)  5–12 873 (13.2 %) 873 (15.7 %)  13–18 781 (11.8 %) 781 (14.1 %)  19–22 737 (11.1 %) 737 (13.3 %)  22+ 2420 (36.5 %) 2420 (43.6 %)  Missing 255 31 Hours Worked Per Week  <40 h 1172(17.4 %) 990 (18.1 %)  40–49 h 1340(19.9 %) 1066 (19.5 %)  50–59 h 1667(24.7 %) 1350 (24.6 %)  60–69 h 1526(22.6 %) 1238 (22.6 %)  70–79 h 535(7.9 %) 436 (8.0 %)  ≥80 h 509(7.5 %) 399 (7.3 %)  Missing 131 103 # Nights on Call Per Week  Median(IQR) 1(0–3) 1 (0–3) Primary Practice Setting  Private practice 3605(52.6 %) 3048 (54.8 %)  Academic Medical Center 1625(23.7 %) 1251 (22.5 %)  Veterans hospital 104(1.5 %) 74 (1.3 %)  Active military practice 58(0.8 %) 37 (0.7 %)  Not in practice or retired 160(2.3 %) 136 (2.4 %)  Other 1303(19 %) 1020 (18.3 %)  Missing 205 16 Level of satisfaction relationship with children  Very satisfied 2738 (49.2 %)  Satisfied 2044 (36.7 %)  Neither satisfied/dissatisfied 341 (6.1 %)  Dissatisfied 372 (6.7 %)  Very dissatisfied 73 (1.3 %)  No children/missing 14 Impact of career on relationship with children  Major positive impact 1212 (21.8 %)  Minor positive impact 1260 (22.7 %)  No impact 517 (9.3 %)  Minor negative impact 2071 (37.2 %)  Major negative impact 501 (9.0 %)  No children/missing 21 Parental satisfaction of U.S. physicians Overall, physicians were satisfied in their relationships with their children, with 4782 (85.9 %) indicating that they were either very satisfied [n = 2738; (49.2 %)] or satisfied [n = 2044; (36.7 %)]. Slightly less than half of physicians believed their career had made either a major positive [n = 1212; (21.8 %)] or a minor positive [n = 1260; (22.7 %)] impact on their relationships with their children, while fewer believed their career had made a minor [n = 2071; (37.2 %)] or major negative [501; (9 %)] impact, respectively. Small differences in satisfaction in relationships with their children were observed by gender, age, hours worked/week, nights on call/week, specialty, practice setting, or method of compensation (Additional file 1 Figure S1). Larger associations were observed between these personal and professional characteristics and physicians’ perception of the impact of their career on relationships with their children (Fig. 1a-f). Women physicians were more likely to report a negative impact of their career on relationships with their children than their male colleagues (women: 51.7 % vs. men: 41.1 %; p ˂ 0.001; Fig. 1a). Younger physicians were also less likely to believe their career had made a positive impact than their older colleagues (Fig. 1b). Physicians whose relationship status was single or partnered were less likely to be satisfied than those who were married (Fig. 1c). Hours worked/week inversely correlated with the belief that career had made a positive impact on relationships with their children, with an apparent dose effect (Fig. 1d). Those in a salary-plus-bonus model were less likely to believe their career had made a positive impact on relationships with their children than those in either a pure salaried position or a pure incentive-based system (Fig. 1e).Fig. 1 Physicians’ Perception of the Impact of Their Career On Relationships With Their Children. a. Impact of Career on Relationship with Children by Gender. b. Impact of Career on Relationship with Children by Age. c. Impact of Career on Relationship with Children by Relationship Status. d. Impact of Career on Relationship with Children by Hours worked/wk. e. Impact of Career on Relationship with Children by method of compensation. f. Impact of Career on Relationship with Children by Specialty. g. Impact of Career on Relationship with Children by practice setting The association between specialty and practice setting and physicians’ belief that their career had made a positive impact on relationships with their children is shown in Fig. 1f and g. Specialties least likely to believe that their career had made a positive impact on relationships with their children included obstetrics and gynecology, internal medicine subspecialties, and anesthesiology, while those most likely to believe that their career had made a positive impact included ophthalmologists, general pediatricians, and psychiatrists. The results of multivariate analysis including the variables age, gender, relationship status, hours/week, specialty, practice setting, and method of compensation are shown in Table 2. Women physicians and physicians who were older were more likely to be satisfied in their relationships with their children. Physicians whose relationship status was single or partnered were less likely to be satisfied than those who were married. Physicians who were in an academic or “other” practice setting were more likely to be satisfied than those in private practice, while those who were not in practice or retired were less likely to be satisfied. Physicians in a salaried model were less likely to be satisfied in their relationship with their children than those in a pure incentive-based compensation model. Hours worked/week also remained strongly associated with parental satisfaction in the multivariate analysis, where each additional hour worked/week reduced the likelihood of satisfaction by 1 %.Table 2 Multivariate analysis to identify factors associated with satisfaction in relationship with children Variable OR (95 % CI) P Age (for each year older) 1.01 (1.00–1.02) 0.0070 Female (vs. Male) 1.29 (1.07–1.56) 0.0089 Partnered (vs. Married) 0.46 (0.31–0.68) 0.0001 Single (vs. Married) 0.60 (0.46–0.77) <0.0001 Academic practice vs. private practice 1.50 (1.20–1.88) 0.0004 Not in practice/retired (vs. private practice) 0.54 (0.31–0.94) 0.0293 Other practice setting (vs. private practice) 1.31 (1.04–1.66) 0.0244 Salaried w no incentive pay (vs. incentive pay only) 0.72 (0.58–0.90) 0.0036 Hours worked per week (for each additional hour) 0.99 (0.98–0.99) <0.0001 Factors in model: age, gender, relationship status, hours/week, specialty, practice setting, method of compensation We also conducted a multivariate analysis including the same variables to identify factors associated with physicians’ perception that their career had made a positive impact on relationships with their children (Table 3). Increasing age (odds ratio [OR] = 1.039 [95 % confidence interval [CI]: 1.034-1.045]; p ˂ 0.0001) was associated with an increased likelihood of viewing career as a positive influence, while the number of hours worked/week (OR = 0.990 [95 % CI: 0.986-0.993]; p ˂ 0.0001) was inversely associated with the perception that career had made a positive impact. Physicians whose relationship status was single or partnered were less likely to view their career as a positive influence on relationship with children than those who were married. Physicians who specialized in general pediatrics and ophthalmology were more likely to believe that their career had made a positive impact than those in family medicine (used as the reference specialty).Table 3 Multivariate analysis to identify factors associated with perception that career has had positive impact on relationship w/children in physicians Variable OR (95 % CI) P Age (for each year older) 1.039 (1.034–1.045) <0.0001 Partnered (vs. Married) 0.693 (0.485–0.989) 0.0436 Single (vs. Married) 0.596 (0.461–0.769) <0.0001 Hours worked per week (for each additional hour) 0.990 (0.986–0.993) <0.0001 General Pediatrics as specialty (vs. Family Med) 1.657 (1.213–2.264) 0.0015 Ophthalmology as specialty (vs. Family Med) 1.432 (1.000–2.048) 0.0497 Factors in model: age, gender, relationship status, hours/week, specialty, practice setting, compensation Physicians compared to U.S. working population Next, we compared perceptions of the impact of career on relationships with children among physicians to the general U.S. working population (n = 3991; 74.2 %). Both men (physician 48.7 %; population 25.3 %; OR = 2.795 [95 % CI: 2.474-3.157]; p ˂ 0.001) and women (physician 54.0 %; population 21.3 %; OR = 4.327 [95 % CI: 3.715-5.041]; p ˂ 0.001) physicians were significantly more likely to report that their career had a negative impact on relationships with their children than U.S. workers (Fig. 2a). The strong relationship between increasing age and the perception that career had a positive impact on relationships with children was of greater magnitude among physicians than the general population for both men and women (Fig. 2b). Although the proportion of men and women physicians who believed that career had a positive impact on their relationships with children was lower for physicians than the age and sex-matched general population 45 and younger, it was similar to or higher than the age-matched general population for those age 46–65.Fig. 2 Perceived Impact of Career On Relationships With Their Children: Comparison of Physicians to U.S. Workers in General. a Impact of Career on Relationship with Children by Gender. b Impact of Career on Relationship with Children by Age In the pooled (physicians and U.S. workers) multivariate analysis, the relationship between advancing age and a more favorable view of the impact of career on relationship with children persisted, as did the inverse association with hours worked/week. Single parents were less likely to view their career as a positive influence on relationship with children than those who were married. The effect of education on the perceived impact of career suggested a beneficial effect to higher levels of education. When compared to individuals who were high school graduates, a stepwise effect on likelihood of satisfaction was observed among individuals with an associate’s degree, bachelor’s degree, master’s degree, or professional/doctoral degree in a field other than medicine (i.e. greater likelihood of satisfaction with higher educational attainment). The beneficial effects of education also held true for physicians relative to high school graduates; however, the effect size of a doctoral degree in medicine degree appeared roughly equivalent to that of a bachelor’s degree, and was lower than that observed with a master’s degree or a professional/doctoral degree in a field outside medicine (Table 4).Table 4 Predictors of career having positive impact on relationship with children in physicians and population norms Variable OR (95 % CI) P Age (for each year older) 1.026 (1.021–1.031) <0.0001 Hours worked per week (for each additional hour) 0.986 (0.983–0.989) <0.001 Single (vs. Married) 0.702 (0.601–0.820) <0.0001 Some college, no degree (vs. High School Graduate) 1.394 (1.131–1.717) 0.0018 Associates Degree (vs. High School Graduate) 1.581 (1.247–2.004) 0.0002 Bachelors Degree (vs. High School Graduate) 2.144 (1.759–2.613) <0.0001 Masters Degree (vs. High School Graduate) 2.570 (2.040–3.238) <0.0001 Professional/Doctorate Degree (vs. High School Graduate) 2.868 (2.068–3.976) <0.0001 Physicians (vs. High School Graduate) 2.117 (1.787–2.508) <0.0001 Factors in model: age, gender, hours/week, relationship status, highest level of education Discussion There is minimal information available regarding physicians’ parental satisfaction and perception of the impact of their career on their children. To our knowledge, this is the first national study to assess physicians’ parental satisfaction in the last 25 years, and the only study to evaluate a comparison cohort of working adults from the general U.S. population. Generally, physicians reported high levels of parental satisfaction. Despite the high satisfaction with their relationships with their children, nearly half of physicians reported their career had adversely impacted those relationships. Not surprisingly, hours worked/week had an inverse relationship with the likelihood of perceiving that career had a positive impact. Although women physicians were more likely to report a negative impact of career on relationship with children in univariate analysis, this difference did not persist on multivariate analysis adjusting for age and other factors. Collectively, these results suggest that, while physicians view their career as a potential barrier to the relationships they desire to have with their children, most are satisfied with relationships they develop nonetheless. Both men and women physicians were dramatically more likely than U.S. workers in general to believe their career had a negative impact on their relationship with their children. Physicians also had a more polarized view of the impact of their career than U.S. workers; <10 % of physicians reported their career had “no impact” on their relationship with their children as compared to 35 %-40 % of U.S. workers. Notably, in multivariate analysis of both physicians and U.S. workers, advancing age was associated with a higher likelihood of viewing career as having a positive impact, while the number of hours worked/week had an inverse association with the perceived impact of career on relationships with children. We are unable to determine the potential reason for this association with age given the cross-sectional nature of the study. A variety of factors may contribute to the more favorable view of the impact of career on parenting with age including: i) a generational effect, ii) a change in the perceived impact of career for those having children later, iii) a change in perception of the impact of career on parenting as the parent and children get older, iv) differences in role conflict over the course of a career, and v) other unmeasured interacting variables factors (e.g. whether the individual is in a two career relationship and the professional characteristics of their partner). In both physicians and workers in other fields, the highest level of education achieved was also related to the perceived impact of career on relationships with children independent of gender. Relative to high school graduates, a stepwise greater likelihood of reporting that career had a positive impact was observed for those with an associate’s degree, bachelor’s degree, master’s degree, and professional/doctoral degree in a field other than medicine. In contrast to this clear stepwise, incremental association between level of education and the perception that career had made a positive impact among non-physicians, the magnitude of effect of doctoral degree in medicine was more similar to that of a bachelor’s degree and lower than a master’s or professional/doctoral degree in other fields. Collectively, these findings regarding the relationship between work hours, age and level of education with parental satisfaction are thought provoking, particularly for women. They may argue against the notion that pursuing an advanced or professional degree has an adverse impact on relationship with children and may also suggest that the perceived impact of career on relationships with children becomes more favorable with age. The multivariate analyses of physicians also provide some potential insights for physicians attempting to mitigate a negative effect of their career on their relationships with their children and seeking to cultivate parental satisfaction. First, they suggest that these individuals be acutely aware of the total number of hours worked more so than other variables. Second, for the most part, specialty is irrelevant with respect to satisfaction in relationship with children. Third, practice setting does appear to be related to satisfaction in relationship with children (higher for those in academic practice settings than private practice) but not impact of career on relationship with children. At a minimum, it would seem physician parents need to rigorously monitor work hours, periodically evaluate if they are spending too much time at work, and make sure their specialty and practice setting allow them adequate flexibility to tailor work hours to meet both personal and professional priorities. In this regard, it should be noted that reducing professional work hours may be a helpful strategy for many physicians [14–17], and that the available evidence suggests that working part time does not adversely impact quality of care [8, 9] or patient satisfaction [15, 18–21]. How do these results relate to previous reports? In 1988, Warde and colleagues surveyed 656 married physicians with children in southern California (single parents excluded) [9]. Parental satisfaction among physicians was somewhat lower than marital satisfaction in this cohort. Lower parental satisfaction was observed among older physicians and those experiencing greater “role conflict”, defined as “frustration with the competing demands of career, marriage, and family”. A second report from this cohort indicated greater role conflict was present among women physicians and younger physicians [22]. Among physicians younger than 45, 51 % of men and 87 % of women reported they had made career changes for their children. The most common career changes reported were decreasing work hours, a change in practice type, and interrupting career for their children [22]. More recent studies have also found work-home conflicts and how they are resolved to be powerful contributors to physician burnout [23–25]. Childrearing can be particularly challenging for women physicians due to cultural and societal parenting expectations [26–28]. In 1988, Levinson and colleagues reported the results of a seminal study of 862 women physicians working as faculty members of academic internal medicine practices who were under the age of 50 [29]. They found women physicians were more likely to delay childbearing compared to U.S. women and took relatively short maternity leaves [29]. Collectively, 87 % of academic women physicians reported that they relied on a paid employee to assist in the care of their children, and 68 % reported childrearing had slowed their career progress [29]. At the time of this study, few institutions provided job-sharing opportunities for these faculty members. Although a variety of initiatives to address some of these challenges specific to women physicians have been reported or proposed in recent years [30–33], much more progress is needed given the rapidly growing number of women physicians. Our study is subject to a number of limitations. First, approximately one in five of physicians surveyed responded. As previously reported, the demographic characteristics of responders were generally consistent with U.S. physicians overall [12]. Nonetheless, how representative participants are of all physicians is unknown. Second, our data are cross-sectional in nature, and we are unable to assess longitudinal changes over time. Third, a number of variables that may interact with parental satisfaction (e.g. characteristics of partner and their career, marital discord, access to family support, access and affordability of day care) were not assessed. Fourth, the study describes the experience of physicians in the U.S. and its generalizability of physicians in other countries is unknown. Finally, although they are similar to previous questions evaluating parental satisfaction in other studies of physicians, the questions we used to evaluate parental satisfaction are not standardized instruments. The fact that we collected simultaneous information using one of these items in a sample of the general U.S. working population for comparison may, to some extent, mitigate this limitation. Our study also has a number of important strengths. Participants in this study represent a large and diverse sample of U.S. physicians from all specialties and practice types. Unlike previous studies [9, 29], we did not limit the sample to physicians of a single gender, specialty, or practice setting, and we did not exclude single parents. The comparison of physicians to the general U.S. working population is also a unique contribution of this study. Conclusions U.S. physicians report generally high satisfaction in their relationships with their children. Despite their high satisfaction, physicians have a more negative perception regarding the impact of their career on relationships with their children than other U.S. workers. The number of hours worked/week and specialty choice were strongly associated with the perceived impact of physicians’ career on their relationships with their children. Additional studies evaluating how physicians can promote healthy relationships with their children and the characteristics of the practice environment that facilitate healthy relationships would be useful. Additional file Additional file 1: Survey Questions Related to Parental Satisfaction. Figure S1. Physicians’ Satisfaction with Relationship with Children. A. Satisfaction with Relationship with Children by Gender. B. Satisfaction with Relationship with Children by Age. C. Satisfaction with Relationship with Children by Relationship status. D. Satisfaction with Relationship with Children by Hours worked/wk. E. Satisfaction Relationship with Children by nights on call/week. F. Satisfaction Relationship with Children by Specialty. G. Satisfaction Relationship with Children by practice setting. H. Satisfaction Relationship with Children by method compensation. Table S1. Demographic Characteristics of Population Sample of Parents. (DOCX 73 kb) Abbreviations CIConfidence interval OROdds ratio USUnited States Acknowledgements None. Funding Funding for this study was provided by the Mayo Clinic Program on Physician Well-being. Availability of data and materials data is shared property of the American Medical Association and Mayo Clinic and will not be made public at this time because continued analysis is ongoing. Authors’ contributions Concept and design: TS, OH, CW, LD. Acquisition of data and interpretation: TS, OH, SH, CS, DS, JS, CW, LD. Drafting and critical revising the manuscript: TS, OH, SH, CS, DS, JS, CW, LD. Final approval of the manuscript: TS, OH, SH, CS, DS, JS, CW, LD. All authors read and approved the final manuscript. Competing interest The authors declare they have no competing interests. Consent for publication not applicable. Ethics approval and consent to participate The Mayo Clinic Institutional Review Board reviewed and approved the study. Participation was elective and all results anonymous. As approved by the Institutional Review Board reviewed consent was implied by completion of the survey. ==== Refs References 1. Shanafelt TD Boone S Tan L Burnout and satisfaction with work-life balance among US physicians relative to the general US population Arch Intern Med 2012 172 18 1377 1385 10.1001/archinternmed.2012.3199 22911330 2. Horowitz C Suchman A Jr WB Frankel R What do doctors find meaningful about their work? Ann Intern Med 2003 138 772 10.7326/0003-4819-138-9-200305060-00028 12729445 3. Clever LH Some things have not changed Ann Intern Med 2000 132 1 85 89 10.7326/0003-4819-132-1-200001040-00014 10627257 4. Shanafelt T Sloan J Habermann T The Well-Being of Physicians Am J Med 2003 I114 513 517 10.1016/S0002-9343(03)00117-7 12727590 5. Linzer M Visser MR Oort FJ Smets EM McMurray JE de Haes HC Predicting and preventing physician burnout: results from the United States and the Netherlands Am J Med 2001 111 2 170 175 10.1016/S0002-9343(01)00814-2 11498074 6. Dyrbye LN Varkey P Boone SL Satele DV Sloan JA Shanafelt TD Physician satisfaction and burnout at different career stages Mayo Clin Proc 2013 88 12 1358 1367 10.1016/j.mayocp.2013.07.016 24290109 7. Ly DP Seabury SA Jena AB Divorce among physicians and other healthcare professionals in the United States: analysis of census survey data BMJ 2015 350 h706 10.1136/bmj.h706 25694110 8. Shanafelt TD Boone SL Dyrbye LN The medical marriage: a national survey of the spouses/partners of US physicians. Mayo Clin Proc 2013 88 3 216 225 10.1016/j.mayocp.2012.11.021 23489448 9. Warde CM Moonesinghe K Allen W Gelberg L Marital and parental satisfaction of married physicians with children J Gen Intern Med 1999 14 3 157 165 10.1046/j.1525-1497.1999.00307.x 10203621 10. AAMC. Current Status of the U.S. Physician Workforce. 2015; http://aamcdiversityfactsandfigures.org/section-ii-current-status-of-us-physician-workforce/. Accessed 9/2/2015. 11. Hawkins M. 2012 Review of Physician Recruiting Incentives. 2012; http://www.merritthawkins.com/uploadedFiles/MerrittHawkins/Pdf/mha2012incentivesurveyPDF.pdf. Accessed May 5, 2014 12. Shanafelt TD Hasan O Dyrbye LN Changes in Burnout and Satisfaction With Work-Life Balance in Physicians and the General US Working Population Between 2011 and 2014 Mayo Clin Proc 2015 90 12 1600 1613 10.1016/j.mayocp.2015.08.023 26653297 13. American Association for Public Opinion Research (AAPOR). Standard Definitions, Final Dispositions of Case Codes, and Outcome Rates for Surveys; Revised 2016. http://www.aapor.org/AAPOR_Main/media/publications/Standard-Definitions20169theditionfinal.pdf. Accessed 22 Aug 2016. 14. Murray A Safran DG Rogers WH Inui T Chang H Montgomery JE Part-time physicians. Physician workload and patient-based assessments of primary care performance Arch Fam Med 2000 9 4 327 332 10.1001/archfami.9.4.327 10776360 15. Mechaber HF Levine RB Manwell LB Part-time physicians…prevalent, connected, and satisfied J Gen Intern Med 2008 23 3 300 303 10.1007/s11606-008-0514-3 18214623 16. McMurray JE Heiligers PJ Shugerman RP Part-time medical practice: where is it headed? Am J Med 2005 118 1 87 92 10.1016/j.amjmed.2004.11.005 15639215 17. Cull WL O'Connor KG Olson LM Part-time work among pediatricians expands Pediatrics 2010 125 1 152 157 10.1542/peds.2009-0767 20008433 18. Levine RB Harrison RA Mechaber HF Phillips C Gallagher TH Professional characteristics and job satisfaction among SGIM members: a comparison of part-time and full-time physician members J Gen Intern Med 2008 23 8 1218 1221 10.1007/s11606-008-0635-8 18483832 19. Fein OT Garfield R Impact of physicians’ part-time status on inpatients’ use of medical care and their satisfaction with physicians in an academic group practice Acad Med 1991 66 11 694 698 10.1097/00001888-199111000-00014 1747182 20. Parkerton PH Wagner EH Smith DG Straley HL Effect of part-time practice on patient outcomes J Gen Intern Med 2003 18 9 717 724 10.1046/j.1525-1497.2003.20401.x 12950480 21. Panattoni L Stone A Chung S Tai-Seale M Patients report better satisfaction with part-time primary care physicians, despite less continuity of care and access J Gen Intern Med 2015 30 3 327 333 10.1007/s11606-014-3104-6 25416600 22. Warde C Allen W Gelberg L Physician role conflict and resulting career changes. Gender and generational differences J Gen Intern Med 1996 11 12 729 735 10.1007/BF02598986 9016419 23. Dyrbye LN Freischlag J Kaups KL Work-home conflicts have a substantial impact on career decisions that affect the adequacy of the surgical workforce Arch Surg 2012 147 10 933 939 10.1001/archsurg.2012.835 23117833 24. Dyrbye LN Sotile W Boone S A survey of U.S. physicians and their partners regarding the impact of work-home conflict J Gen Intern Med 2014 29 1 155 161 10.1007/s11606-013-2581-3 24043567 25. Dyrbye LN Shanafelt TD Balch CM Satele D Freischlag J Physicians married or partnered to physicians: a comparative study in the American College of Surgeons J Am Coll Surg 2010 211 5 663 671 10.1016/j.jamcollsurg.2010.03.032 21035046 26. Harrison RA Gregg JL A time for change: an exploration of attitudes toward part-time work in academia among women internists and their division chiefs Acad Med 2009 84 1 80 86 10.1097/ACM.0b013e3181900ebd 19116482 27. Poppas A Cummings J Dorbala S Douglas PS Foster E Limacher MC Survey results: a decade of change in professional life in cardiology: a 2008 report of the ACC women in cardiology council J Am Coll Cardiol 2008 52 25 2215 2226 10.1016/j.jacc.2008.09.008 19095147 28. Carr PL Ash AS Friedman RH Relation of family responsibilities and gender to the productivity and career satisfaction of medical faculty Ann Intern Med 1998 129 7 532 538 10.7326/0003-4819-129-7-199810010-00004 9758572 29. Levinson W Tolle SW Lewis C Women in academic medicine. Combining career and family N Engl J Med 1989 321 22 1511 1517 10.1056/NEJM198911303212205 2811971 30. Villablanca AC Beckett L Nettiksimmons J Howell LP Career flexibility and family-friendly policies: an NIH-funded study to enhance women's careers in biomedical sciences J Womens Health (Larchmt) 2011 20 10 1485 1496 10.1089/jwh.2011.2737 21859346 31. Valantine H Sandborg CI Changing the culture of academic medicine to eliminate the gender leadership gap: 50/50 by 2020 Acad Med 2013 88 10 1411 1413 10.1097/ACM.0b013e3182a34952 23969359 32. Howell LP Beckett LA Nettiksimmons J Villablanca AC Generational and gender perspectives on career flexibility: ensuring the faculty workforce of the future Am J Med 2012 125 7 719 728 10.1016/j.amjmed.2012.03.013 22727238 33. Strong EA De Castro R Sambuco D Work-life balance in academic medicine: narratives of physician-researchers and their mentors J Gen Intern Med 2013 28 12 1596 1603 10.1007/s11606-013-2521-2 23765289

PMC005xxxxxx/PMC5002114.txt

==== Front Ir Vet JIr Vet JIrish Veterinary Journal0368-07622046-0481BioMed Central London 275706167110.1186/s13620-016-0071-8Case ReportGastric inflammatory pseudotumour secondary to Actinomyces hordeovulneris infection in a cat Pietra Marco marco.pietra@unibo.it Zanoni Renato Giulio renatogiulio.zanoni@unibo.it Peli Angelo angelo.peli@unibo.it Brunetti Barbara b.brunetti@unibo.it Linta Nikolina nikolina.linta@libero.it Capitani Ombretta ombretta.capitani@unibo.it Spinella Giuseppe giuseppe.spinella@unibo.it Department of Veterinary Medical Sciences, School of Agriculture and Veterinary Medicine, Alma Mater Studiorum - University of Bologna, Ozzano dell’Emilia (BO), 40064 Italy 26 8 2016 26 8 2016 2016 69 1216 3 2016 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 The authors report the first case of feline gastric actinomycosis associated with infection by Actinomyces hordeovulneris. Case presentation A 4-year-old, neutered male, semi-feral European cat, with a 1 year history of chronic vomiting, was referred to the clinic. Abdominal ultrasound examination identified a hypoechoic focal transmural thickening with loss of normal wall layering and hyperechoic speckles at the gastric body. Initial gastroscopic examination showed a tumour-like gastric mass with an ulcerated depression at the level of the greater curvature. Histologic examination of endoscopic biopsy specimens was consistent with a severe lymphoplasmacytic gastritis. After 2 months, due to persistence of abdominal discomfort, surgical exploration and intraoperative sampling of gross abnormalities was recommended. Full thickness gastric wall biopsies, and fine needle aspiration of the gastric thickening and gastric lymph node, were performed. Histopathological examination identified a transmural pyogranulomatous gastritis. Aspirate samples of the gastric wall cultured positive, with colony morphology, biochemical testing and PCR of the 16 s rRNA gene compatible with Actinomyces hordeovulneris. After 4 months of treatment with cefovecin (8 mg/kg subcutaneously every 14 days), the vomiting completely resolved, as well as the ultrasonographic gastric alteration. Conclusion This case report of feline gastric actinomycosis, caused by Actinomyces hordeovulneris, suggests that gastric bacterial infection should be considered in cases of focal gastric wall thickening associated with chronic vomiting in the cat, which may otherwise closely resemble neoplastic disease. Once a diagnosis of actinomycosis was obtained, a correct treatment with antibiotic therapy can resolve it. Keywords CatGastritisPseudotumourActinomyces hordeovulnerisissue-copyright-statement© The Author(s) 2016 ==== Body Background Primary gastric actinomycosis is a rare gastric wall infection caused by Actinomyces species; it is rarely reported in people and dogs and is characterized by the formation of multiple gastric abscesses, draining sinuses with abundant granulation, and dense fibrous tissue [1]. To the authors’ knowledge, there have been no reports of gastric Actinomyces infection in cats, and the authors believe this condition is possibly under-diagnosed due to its resemblance to neoplastic disease. In this case report, the authors describe the unusual presentation of an intramural gastric infection with Actinomyces hordeovulneris in a cat. Here we describe the clinical, sonographic, and endoscopic pattern that mimics a malignant tumour and describe the diagnostic procedures required in this case for diagnosis. Case presentation A 4-year-old, 6-kg, neutered male, semi-feral European cat that had been regularly vaccinated (against feline viral rhinotracheitis, calicivirus, and panleukopenia), and dewormed was referred to the Bologna University Veterinary Teaching Hospital for a diagnostic evaluation to determine the cause of chronic vomiting. Vomiting had first been observed approximately 1 year previously, was initially intermittent, and had gradually increased without concurrent decrease in appetite or body weight. An abdominal ultrasound examination, performed by another veterinarian 6 months after the onset of clinical signs, had shown only a moderate enlargement of jejunal lymph nodes without modification of their echogenicity or shape, and without change of the echo structure of other abdominal organs. No laboratory analyses had been performed at that time. The owner reported a partial response to long-term treatment with ranitidine, which had been continued up until the time of referral. However, within the month prior to referral, vomiting had worsened to once daily and occurred most often in the morning, without association to feeding. On clinical examination, the cat was bright and had a body condition score of 6/9. Rectal temperature was within reference interval (38.5 °C), mucous membranes were congested, and no modification in size or consistency of peripheral lymph nodes was noted. Respiratory and cardiovascular examinations were unremarkable. On abdominal palpation, an immobile, firm and painful mass, approximately 4 cm in diameter, was identified in the cranial abdomen. Haematological and serum biochemical tests revealed only a mild leukopenia (white blood cell count 4.46 × 109/L, reference interval 5.0–19.0 × 109/L) associated with lymphocytopenia (lymphocyte count 0.31 × 109/L, reference interval 1.5–7.0 × 109/L). Results of faecal flotation were negative, and total thyroxine concentration was within reference interval (15 nmol/L, reference range 5–40 nmol/L). The cat was negative for serum feline leukaemia virus antigen1, but was positive for serum feline immunodeficiency virus antibodies1. Abdominal ultrasound showed a hypoechoic transmural thickening (approximately 9 mm) at the gastric body, associated with loss of normal wall layering and hyperechoic speckles possibly consistent with focal fibrosis (Fig. 1). A moderate volume of anechoic fluid was seen in the gastric lumen. Gastric and portal lymph nodes were mildly enlarged and hypoechoic, but with normal shape. No other ultrasound findings were reported. Thoracic radiographs were unremarkable.Fig. 1 Initial abdominal ultrasound (cat). Transmural thickening (about 9 mm) of the gastric body wall with loss of normal wall layering Endoscopic examination of the stomach showed a gastric mass with an ulcerated depression at the level of the greater curvature between the fundus and gastric body (Fig. 2). The endoscopic appearance of the duodenum was normal. Histological examination of biopsy specimens obtained from the base and edges of the region of ulceration was not compatible with a neoplastic process, and a healing ulcer associated with severe lymphoplasmacytic gastritis was diagnosed.Fig. 2 Initial endoscopy of the stomach (cat). Tumour-like gastric mass with an ulcerated depression at the level of the greater curvature between the fundus and gastric body The cat was treated with marbofloxacin (5 mg/kg PO every 24 h for 10 day) and rabeprazole sodium (0.8 mg/kg PO every 24 h for 30 day) and fed a hypoallergenic diet. After initial clinical improvement, vomiting frequency subsequently increased gradually over the following 2 months to once daily, with a concurrent reduction of body weight (5.5 kg). The cat was again presented, and examination identified abdominal discomfort, along with the palpable mass in the epigastric region that had been previously noted. Ultrasound examination identified similar changes to that seen previously with a focal hypoechoic wall thickening and loss of normal layering on the gastric greater curvature (increased from 9 to 11.8 mm from the previous ultrasound exam) extending for approximately 3 cm from the fundus to the gastric body; the mass was accompanied by marked enlargement of gastric (76 × 129 mm) and pancreatic-duodenal (57 × 97 mm) lymph nodes with an ovoid shape. At this presentation, given progression of disease and that previously findings may not have been representative, endoscopy together with celiotomy was recommended to allow direct observation of both sides of the gastric wall, and to allow surgical full thickness biopsy and aspiration of gastric lesion. In addition an aspiration of the associated lymphadenopathy was advised. Endoscopic examination identified a large whitish, thickened mucosal area, the central region of which appeared to be covered by fibrin at the level of the greater curvature in front of the pyloric antrum, ascribable to pseudomembranous plaque, without evidence of an ulcer (Fig. 3). Based upon the appearance of the lesion, endoscopic biopsies were not performed.Fig. 3 Second endoscopy of the stomach (cat). Performed 2 months after the initial endoscopy, shows the presence of a large whitish thickened mucosal area with a central part covered by fibrin at the level of the greater curvature in front of the pyloric antrum, indicating a healing phase ulcer At exploratory coeliotomy, evaluation of the gastric body identified a poorly delineated lesion at the greater curvature without macroscopic serosal alterations or adhesions with the omentum or other organs. Palpation of the gastric lesion confirmed a poorly delineated thickening of the gastric wall, with a tough and fibrous consistency. Enlarged gastric and pancreatic-duodenal lymph nodes were observed. Inspection of the other abdominal organs did not identify any gross pathological change. Two 14G full thickness core biopsies of gastric wall (stored in formalin for histologic procedures), and a fine needle aspiration of the gastric thickening (for a bacteriological examination) were performed. At biopsy, the consistency of the gastric wall appeared firm. A fine needle aspiration of the gastric lymph node (smeared on slides for a cytological examination) was obtained. Histopathology of the gastric biopsy identified severe, chronic transmural pyogranulomatous gastritis. Cytological examination of the gastric lymph node indicated non-specific reactivity, with no evidence of neoplastic disease. On bacteriological examination, after 72 h of incubation in a capnophilic and anaerobic atmosphere, white colonies had grown on Columbia agar with 5 % sheep blood. The colonies had a molar tooth-like appearance and, after 7 days of incubation, showed slight haemolysis. The isolate was identified as A. hordeovulneris by chemical tests and 16S rRNA gene sequence, and based upon sensitivity testing was susceptible to penicillin, amoxicillin/clavulanate, cephalothin, cefovecin, tetracycline, enrofloxacin, and trimethoprim/sulfamethoxazole. Following isolation of A. hordeovulneris, further staining of biopsy material was performed including Gram-staining and ZN-staining, in order to potentially identify organisms not previously observed in histopathologic sections. Gram-positive filamentous organisms were visible in multiple microscopic fields consistent with Actinomyces spp. (Fig. 4).Fig. 4 Gastric wall. Gram staining showing filamentous Gram-positive bacteria (arrow) On the basis of these findings, treatment chosen was amoxicillin/clavulanate (12.5 mg/kg PO every 12 h for 4 months) and clebopride (5 μg/kg PO every 8 h for 2 months) with the purpose of increasing gastric empting. After 1 week, amoxicillin/clavulanate was replaced by cefovecin (8 mg/kg SC every 14 days for 4 months) due to a lack of cooperation of the patient in taking medication orally. The frequency of vomiting was reduced in the following week, and subsequently resolved completely. On follow-up clinical examination, performed 4 months after surgery, 18 months after onset of clinical signs, body weight was increased again to 6.50 kg, and the only abnormal findings were mild abdominal tenderness and thickness of several intestinal loops. A complete blood count and biochemistry profile identified again a mild leukopenia (white blood cell count 4.04 × 109/L, reference interval 5.0–19.0 × 109/L) and lymphocytopenia (lymphocytes 0.27 × 109/L, reference interval 1.5–7.0 × 109/L). Abdominal ultrasound identified complete resolution of the gastric lesion with normal thickness and almost normal layering of gastric wall (Fig. 5). The gastric lymph node was mildly enlarged (2.9 cm) and appeared hypoechoic with maintained shape. Unlike the previous ultrasound examinations, two focal asymmetric hypoechoic thickenings with loss of normal layering of the jejunal wall were observed. In one loop the wall measured 6.7 mm, for a length of approximately 3.2 cm (Fig. 6), and in the other the wall measured 3.1 mm, extending for a length of approximately 2 cm. A suggested surgical biopsy of jejunal wall thickening was declined by the owner.Fig 5 Third abdominal ultrasound (cat). Performed 4 months post-surgery, shows a complete healing of the gastric lesion with normal thickness and normal layering of the gastric wall Fig. 6 Third abdominal ultrasound (cat). Performed 4 months post-surgery, evidences a focal asymmetric hypoechoic thickening with loss of normal layering of the jejunal wall Only ultrasound-guided aspirate of one of the intestinal lesions was performed and a cytological diagnosis of possible lymphoma was obtained. Screening thoracic radiography was therefore performed to evaluate extent and stage of disease, and was unremarkable. Because it can be difficult in the cat to differentiate, by cytomorphology, lymphoma from inflammatory lymphoid cell infiltration of the gastrointestinal wall, a clonal T-cell receptor and B-cell receptor gene rearrangement analysis in feline lymphoid cells using GeneScan analysis was performed [2, 3]. In particular, formalin-fixed, paraffin-embedded gastric tissue (sample 1), stained smears of gastric lymph node cytological samples (sample 2) collected during the exploratory laparotomy, and stained smears of small intestinal aspirated on ultrasound and already diagnosed as suspected lymphoma (sample 3) were analysed. Polyclonal patterns for T-cell and B-cell receptor were identified in the samples 1 and 2, suggestive of reactive change, however a monoclonal pattern for T-cell receptor and polyclonal pattern for B-cell receptor were identified in sample 3, potentially compatible with a T-cell lymphoma. Due to the semi-feral, aggressive nature of the cat, the owner declined multimodal chemotherapy [4], and chose treatment with methylprednisolone acetate injectable suspension (5 mg/kg SC every 30 day). At the time of the last telephone inquiry, 14 months after initial presentation (26 months after clinical sign onset), the owner reported that the cat remained free of clinical signs, showed no vomiting, and continued to have a good appetite and body condition. This case report describes the diagnosis and treatment of a gastric pseudotumour due to Actinomyces hordeovulneris. The lesion is so called in human medicine because the chronic inflammatory process is able to induce granulomatous lesions mimicking a malignant tumour on clinical, ultrasonographic, and gross examination [5]. Actinomyces species are Gram-positive, branching, filamentous bacteria, and are common inhabitants of the feline oral cavity and gastrointestinal tract [6]. In cats Actinomyces have been associated with some cases of ocular [7], nasal [8], intracranial [9], pleural [10], peritoneal [11, 12], ileocolic junction [13] subcutaneous [14] and lymph nodal [15] disease. To the authors’ knowledge, no report of focal gastric wall infection by Actinomyces has been previously described in the cat. Gastric and disseminated abdominal disease has been reported in a dog [16], and rare clinical cases of intramural gastric actinomycosis have been reported in people [1, 5, 17, 18]. Primary gastric actinomycosis is generally infrequent because the bacteria growth is inhibited by the acidic environment of the gastric lumen, and therefore the possibility of intramural translocation is unlikely [1, 18]. It is proposed in previous reports that gastric infection could be related to pharmaceutically-induced hypochlorhydria or to previous gastrectomy, both of which may subsequently increase gastric pH and thereby promote proliferation and opportunistic gastric wall infection [19]. In the case described herein, a possible reason for the development of gastric wall infection by A. hordeovulneris could be due to previous chronic therapy with ranitidine, a drug that is potentially capable to reduce hydrochloric acid secretion. Although ranitidine administered intravenously has been demonstrated to reduce gastric acidity in cats [20], administration of oral ranitidine may not result in similar effects [21]. Actinomyces spp infection has also been identified to occur following mucosal trauma, which may result from penetration via foreign material carrying the bacteria; this was not confirmed in our case but is a possibility. Interestingly co-colonization of A. hordeovulneris with other bacteria was not observed in the case described, as has generally been observed in previous reports in people and animals [13] and is thought to enable an anaerobic environment appropriate for growth of actinomycetes. However, confirmation of A. hordeovulneris involvement in the lesion was late with respect to the onset of clinical signs, and followed administration of marbofloxacin, therefore co-colonization may have been missed. We stress that marbofloxacin has to be considerate as a second-line antibiotic, and it should be limited to infections with identified susceptible bacteria. In the case presented here, marbofloxacin was administered following the first endoscopy due to the unclear history and chronic course of disease. Common clinical signs reported both in canine and human gastric actinomycosis are chronic fever, emaciation, and vomiting [17]; also reported in humans is epigastric pain and gastrointestinal bleeding. Chronic vomiting and an irregular mobile mass in the mid-abdomen that was painful on palpation, without evident weight loss and fever, were the main clinical signs detected in the case reported here, similar to those reported by Sharman et al. [13]. Persistent leukopenia was identified in the current case, in comparison to the finding of leucocytosis, hyperglobulinaemia and hypoalbuminaemia reported by Sharman et al. [13] which are more typical findings of chronic inflammatory disease. This condition could be explained, in our case, by co-infection with the retrovirus FIV, which is able to sometimes induce constant leukopenia, as already reported by Shelton et al. [22]. Equally, for the current case leucopenia could reflect consumption due to the presence of a gastric granuloma, although given its persisteance despite clearance of the actinomycetoma. Histopathology is required to differentiate actinomycetomas from neoplastic disease, given that clinical, imaging and endoscopic findings are non-specific and mimick that of neoplasia. In previous studies, gastric actinomycosis was diagnosed directly on histological analysis of a resected specimen by evidence of colony gram-positive filamentous rods with a surrounding layer of neutrophils (macroscopically so-called “sulphur granules”) [1, 13, 16, 17, 19]. In our case, sulphur granules were not present, and histopathologic examination did not allow identification of bacterial colonization within the pyogranulomatous inflammation using standard staining techniques. Involvement of filamentous bacteria was only confirmed following retrospective gram-staining once positive culture results were obtained. This finding leads the Authors to suggest that staining techniques directed towards identifying infectious agents should be performed where similar conditions are suspected, for example where pyogranulomatous inflammation is detected. Even so, in our case, an aetiologic diagnosis also required bacterial culture of the gastric layer aspirate, followed by identification by PCR of the organism isolated. The therapeutic approach of treating gastric actinomycosis in humans is controversial because while some authors [1] recommend surgical resection of the involved gastric tissue followed by administration of antibiotics (ampicillin, cephalosporin, tetracycline, macrolide or clindamycin) for a prolonged period (up to 12 months) with a favourable outcome, others indicate exclusive treatment with antibiotics without surgical debridement [16]. In the present case report, we observed complete restoration of the original condition of the gastric layer with 4 months’ treatment with antibiotics. It should also be noted that concurrent FIV infection could be a factor that contributed to both gastric infection by A. hordeovulneris and the development of the suspected intestinal lymphoma. The authors would note, that the differentiation of intestinal lymphoma from gastrointestinal inflammation should follow a stepwise diagnostic algorithm that first uses histologic assessment, followed by immunophenotyping and then PCR to determine clonality of lymphocytes [23]. In our case it was not possible to follow the correct diagnostic algorithm as the owner declined intestinal biopsy, allowing only aspiration for cytology. Therefore only cytology together with assessment for clonal T- and B-cell receptor rearrangement contributed to the diagnosis of suspected lymphoma. As the cell population harvested by not be entirely representative using this methodology, the results of clonality analysis might also not be representative in this case. In addition one known pitfall of clonality analysis is that chronic antigenic stimulation, such as may occur with IBD, can also result in an expanded clonal population of lymphocytes and false positive results. The diagnosis of lymphoma reached here should therefore be interpreted with caution, especially given the long-term response to minimal therapy. Conclusions Gastric actinomycosis remains a diagnostic challenge due to the rarity of the infection and the similarity of lesions with neoplastic disease. In any case, especially in cats with predisposing factors such as conditions able to induce immunodeficiency, clinicians should suggest the execution of bacterial isolation from the gastric wall for a definitive diagnosis. 1 Snap® FIV/FeLV Combo Plus, IDEXX Laboratories, Milano (Italy). Acknowledgements The authors acknowledge Misty R. Bailey, ELS(D), for editing of the manuscript. Funding Not applicable. Availability of data and material Not applicable. Authors’ contributions MP drafted the manuscript and performed endoscopic examinations. RGZ performed bacteriological examination and genetic analysis based on 16S rDNA gene sequence. PA reported clinical findings. BB supported the histopathological diagnosis. NL performed ultrasound examinations, OC supervised the manuscript, and GS performed surgery. 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. Lee SH Kim HJ Kim HJ Chung IK Kim HS Park SH Primary gastric actinomycosis diagnosed by endoscopic biopsy: case report Gastrointest Endosc 2004 69 586 9 10.1016/S0016-5107(04)00009-4 2. Mochizuki H Nakamura K Sato H Goto-Koshino Y Sato M Takahashi M GeneScan analysis to detect clonality of T-cell receptor γ gene rearrangement in feline lymphoid neoplasms Vet Immunol Immunopathol 2012 145 402 9 10.1016/j.vetimm.2011.12.015 22261505 3. Daniaux LA Laurenson MP Marks SL Moore PF Taylor SL Chen RX Ultrasonographic thickening of the muscularis propria in feline small intestinal small cell T-cell lymphoma and inflammatory bowel disease J Feline Med Surg 2014 16 89 98 10.1177/1098612X13498596 23900499 4. Kiselow MA Rassnick KM McDonough SP Goldstein RE Simpson KW Weinkle TK Outcome of cats with low-grade lymphocytic lymphoma: 41 cases (1995–2005) J Am Vet Med Assoc 2008 232 405 10 10.2460/javma.232.3.405 18241108 5. Evans J Chan C Gluch L Fielding I Eckstein R Inflammatory pseudotumour secondary to Actinomyces infection Aust N Z J Surg 1999 69 467 9 10.1046/j.1440-1622.1999.01602.x 10392897 6. Love DN Vekselstein R Collings S The obligate and facultatively anaerobic bacterial flora of the normal feline gingival margin Vet Microbiol 1990 22 267 75 10.1016/0378-1135(90)90114-B 2353448 7. Westermeyer HD Ward DA Whittemore JC Lyons JA Actinomyces endogenous endophthalmitis in a cat following multiple dental extractions Vet Ophthalmol 2013 16 459 63 10.1111/vop.12023 23281798 8. Henderson SM Bradley K Day MJ Tasker S Caney SMA Hotston Moore A Investigation of nasal disease in the cat-a retrospective study of 77 cases J Feline Med Surg 2004 6 245 57 10.1016/j.jfms.2003.08.005 15265480 9. Barrs VR Nicoll RG Churchery RK Beckz JA Intracranial empyema: literature review and two novel cases in cats J Small Anim Pract 2007 48 449 54 10.1111/j.1748-5827.2006.00307.x 17543019 10. Barrs VR Allan GS Martin P Beatty JA Malik R Feline pyothorax: a retrospective study of 27 cases in Australia J Feline Med Surg 2005 7 211 22 10.1016/j.jfms.2004.12.004 16055006 11. Scott PC Taylor TK Gilmore JF Hart AT Suppurative peritonitis in cats associated with anaerobic bacteria Aust Vet J 1984 61 367 8 10.1111/j.1751-0813.1984.tb07159.x 6529399 12. Kawamura N Shimada A Morita T Murakami S Azuma R Fujiwara M Intraperitoneal actinomycosis in a cat Vet Rec 2005 157 593 4 10.1136/vr.157.19.593 16272549 13. Sharman MJ Goh CS Kuipers von Lande RG Hodgson JL Intra-abdominal actinomycetoma in a cat J Feline Med Surg 2009 11 701 5 10.1016/j.jfms.2008.10.007 19181557 14. Love DN Jones RF Bailey M Johnson RS Isolation and characterisation of bacteria from abscesses in the subcutis of cats J Med Microbiol 1979 12 207 12 10.1099/00222615-12-2-207 458840 15. Murakami S Yamanishi MW Azuma R Lymph node abscess due to Actinomyces viscosus in a cat J Vet Med Sci 1997 59 1079 80 10.1292/jvms.59.1079 9409531 16. Edwards DF Nyland TG Weigel JP Thoracic, abdominal, and vertebral actinomycosis. Diagnosis and long-term therapy in three dogs J Vet Intern Med 1988 2 184 91 10.1111/j.1939-1676.1988.tb00315.x 3230558 17. Tajima S Waki M Ohata A Koda K Maruyama Y Xanthogranulomatous gastritis associated with actinomycosis: report of a case presenting as a large submucosal mass Int J Clin Exp Pathol 2015 8 1013 8 25755811 18. Skoutelis A Panagopoulos C Kalfarentzos F Bassaris H Intramural gastric actinomycosis South Med J 1995 88 647 50 10.1097/00007611-199506000-00010 7777882 19. Fernández-Aceñero MJ Silvestre V Fernández-Roldán R Cortés L García-Blanch G Gastric actinomycosis: a rare complication after gastric bypass for morbid obesity Obes Surg 2004 14 1012 5 10.1381/0960892041719572 15329195 20. Fändriks L Jönson C Effects of acute administration of omeprazole or ranitidine on basal and vagally stimulated gastric acid secretion and alkalinization of the duodenum in anaesthetized cats Acta Physiol Scand 1990 138 2 181 6 10.1111/j.1748-1716.1990.tb08831.x 2316379 21. Šutalo S Ruetten M Hartnack S Reusch CE Kook PH The effect of orally administered ranitidine and once-daily or twice-daily orally administered omeprazole on intragastric pH in cats J Vet Intern Med 2015 29 840 6 10.1111/jvim.12580 25966746 22. Shelton GH Abkowitz JL Linenberger ML Russell RG Grant CK Chronic leukopenia associated with feline immunodeficiency virus infection in a cat J Am Vet Med Assoc 1989 194 253 5 2537275 23. Kiupel M Smedley RC Pfent C Xie Y Xue Y Wise AG Diagnostic algorithm to differentiate lymphoma from inflammation in feline small intestinal biopsy samples Vet Pathol 2011 48 212 22 10.1177/0300985810389479 21149848

PMC005xxxxxx/PMC5002115.txt

==== Front World J Surg OncolWorld J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 98510.1186/s12957-016-0985-3ResearchThe expression of microRNA 574-3p as a predictor of postoperative outcome in patients with esophageal squamous cell carcinoma Okumura Tomoyuki +81-76-434-7331okumura@med.u-toyama.ac.jp 1Kojima Hirohumi hirofumi@med.u-toyama.ac.jp 1Miwa Takeshi tmiwa@med.u-toyama.ac.jp 1Sekine Shinichi sekky0921@kyi.biglobe.ne.jp 1Hashimoto Isaya yasai831138@yahoo.co.jp 1Hojo Shozo shozo@med.u-toyama.ac.jp 1Nagata Takuya naga0103@med.u-toyama.ac.jp 1Shimada Yutaka yshimada@pharm.kyoto-u.ac.jp 121 Department of Surgery and Science, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, 2630 Sugitani, Toyama, 930-0194 Japan 2 Department of Nanobio Drug Discovery, Graduate School of Pharmaceutical Sciences, Kyoto University, Kyoto, Japan 26 8 2016 26 8 2016 2016 14 1 22827 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 Despite advances in radical esophagectomies and adjuvant therapy, the postoperative prognosis in esophageal squamous cell carcinoma (ESCC) patients remains poor. The aim of this study was to identify a molecular signature to predict postoperative favorable outcomes in patients with ESCC. Methods As a training data set, total RNA was extracted from formalin-fixed paraffin-embedded samples of surgically removed specimens from 19 ESCC patients who underwent curative esophagectomy. The expression of microRNA (miRNA) was detected using a miRNA oligo chip on which 885 genes were mounted. As a validation data set, we obtained frozen samples of surgically resected tumors from 12 independent ESCC patients and the expression of miR-574-3p was detected by quantitative real-time PCR. Results Our microarray analysis in the training set patients identified three miRNAs (miR-574-3p, miR-106b, and miR-1303) and five miRNAs (miR-1203, miR-1909, miR-204, miR-371-3p, miR-886-3p) which were differentially expressed between the patients with (n = 14) and without (n = 5) postoperative tumor relapse (p < 0.01 and p < 0.05, respectively). Higher expression of miR-574-3p, which showed the most significant association with non-relapse (p = 0.001), was associated with favorable overall survival (p = 0.016). Real-time PCR experiments on the validation set patients confirmed that higher expression of miR-574-3p was associated with non-tumor relapse (p = 0.029) and better overall survival (p = 0.004). Conclusions Our results suggest that the aberrant expression of the miRNAs identified in this study plays key roles in the progression of ESCC. miR-574-3p was suggested to have a tumor suppressor effect, and thus, to be a predictor of postoperative outcome in patients with ESCC. Electronic supplementary material The online version of this article (doi:10.1186/s12957-016-0985-3) contains supplementary material, which is available to authorized users. Keywords Esophageal cancerSquamous cell carcinomamiRNAPrognostic markerSurgeryhttp://dx.doi.org/10.13039/501100003478Ministry of Health, Labour and Welfare15K1008815K10089Okumura Tomoyuki issue-copyright-statement© The Author(s) 2016 ==== Body Background Recent progress in radical esophagectomies and perioperative adjuvant therapy has facilitated an improved postoperative prognosis, even in patients with advanced esophageal squamous cell carcinoma (ESCC) [1]. However, many of the patients still exhibit postoperative tumor recurrence with a 5-year progression-free survival rate of about 40 % [2, 3]. Therefore, innovative strategies to predict early postoperative tumor relapse may provide us with a more accurate diagnosis to determine operative indications. MicroRNAs (miRNAs) are small, single-stranded, non-coding RNAs that play a key role in the initiation and progression of cancer through the post-transcriptional regulation of gene expression [4]. They have also been reported to be potential biomarkers for therapeutic effects and prognoses in cancer patients [5]. In the present study, we investigated the expression profiles of miRNAs in surgically removed ESCC specimens to identify relapse-associated miRNAs. Methods Patients and surgical specimens We selected eighty-three consecutive ESCC patients who underwent curative surgery in Toyama University Hospital between 1991 and 2008, excluding 16 cases who died of other causes within 5 years after surgery. Then, we obtained formalin-fixed paraffin-embedded (FFPE) samples of the tumors and their normal counterparts, with adequate tissue size from 36 cases as a training data set. In addition, we obtained frozen samples of surgically resected tumors and their normal counterparts, which had been kept at −80 °C, from 12 independent ESCC patients who received curative surgery in our hospital between 2008 and 2010 to use as a validation data set. The observation period was calculated from the date of surgery until the patient succumbed to the disease or the last follow-up contact. All cases were staged according to the International Union Against Cancer TNM Classification 7th edition [6]. RNA extraction from FFPE specimens Sections (10 μm) were prepared from each FFPE specimen. Paraffin was removed by xylene treatment, and the tissues were then washed with ethanol twice to remove the xylene. The tissues were treated with proteinase K at 37 °C overnight. Following centrifugation, the supernatant was processed with a silica-based spin column (Toray Industries, Japan) in order to obtain purified total RNA. The degrees of RNA cross-linking and RNA degradation were analyzed by electrophoresis using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). miRNA assays using microarray miRNA profiling was examined using a Toray 3D-Gene® miRNA oligo chip (Toray Industries), on which 885 genes were mounted. The detailed procedure for this experiment has been described previously [7]. The expression level of each miRNA was normalized using the median signal strength for the entire gene in each chip. RNA extraction from frozen specimens and quantitative RT-PCR analysis for microRNA s Frozen specimens were obtained from surgically removed esophagus with Lugol’s solution staining. Tumor samples were cut from main part of the tumor, and normal tissues from mucosal epithelia with positive Lugol staining, more than 5 cm distant from the tumor edge. The sections were also confirmed by a board-certified pathologist. Total RNA was extracted from frozen specimens using the TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) according to a standard protocol. cDNA was prepared from total RNA samples using the Taq Man microRNA reverse transcription kit on the ABI Prism 7000 real-time PCR system, according to the manufacturer’s instructions (Applied Biosystems). Predesigned Taq Man microRNA assays for hsa-miR-574-3p (Assay ID 002349) and RNU6B (Assay ID 001093) were purchased from Applied Biosystems. qRT-PCR was performed using a Taq Man universal PCR master mix, according to the manufacturer’s protocol (Applied Biosystems). The microRNA quantities were analyzed in duplicate and normalized against U6B as an internal control. The tumor-to-normal ratio (T/N ratio) was calculated based on the miRNA expression levels in each tumor and the corresponding normal counterpart. Statistical analysis All analyses were carried out with JMP 9.0 software (SAS Institute Inc., Cary, NC, USA). The average expression level (T/N ratio) of each miRNA was calculated and log2-transformed. Differences in miRNA expression levels between two variables were analyzed by the Student’s t test and logistic regression analysis. Hierarchical clustering analysis was performed with non-supervised Ward’s method. The relationship between the expression of miRNAs and various clinicopathological factors were assessed using the chi-squared test. The Kaplan-Meier method was used to estimate patient survival. Differences in postoperative outcome for the expression of the selected miRNAs were analyzed using the log-rank test. p < 0.01 was used for significance in microarray analysis with the Student’s t test and logistic regression analysis, while p < 0.05 was used for significance in the chi-squared test and log-rank test. We constructed receiver operating characteristic (ROC) curves and calculated the area under the curve (AUC) to evaluate the specificity and sensitivity as a postoperative prognostic predictor. Results Patient information and tumor characteristics For a training data set, in 36 cases from which we extracted total RNA from archival FFPE samples, we obtained high-quality RNA of both tumors and corresponding normal esophageal epithelia from 19 patients. The clinical information and tumor characteristics of these patients are shown in Table 1. Among the 19 patients, five patients achieved 5-year disease-free survival with an observation period (average ± SD) of 138.7 ± 45.1 months, while 14 patients died of tumor relapse at an observation period (average ± SD) of 14 ± 9.2 months after surgery. These two groups were referred to as the non-relapse group (n = 5) and relapse group (n = 14). All patients in the relapse group and three of the five patients in the non-relapse group were males. There were no statistical differences between the relapse and non-relapse group in the other clinicopathological characteristics such as age, tumor location, histological grade, pathologic tumor depth, lymph node metastasis, distant metastasis, TNM stage, lymphatic vessel invasion, or venous invasion. All 19 patients underwent radical surgery with no residual tumors (R0). None of the five patients in the non-relapse group received preoperative adjuvant therapy, while three of the 14 (21.4 %) in the relapse group received preoperative chemo- or chemo-radio therapy. One of the five (20.0 %) patients in the non-relapse group and seven of the 14 (50.0 %) patients in the relapse group received postoperative chemo- or chemo-radio therapy. The postoperative observation period (average ± SD) in the non-relapse and relapse groups were 138.7 ± 45.1 and 14 ± 9.2 months, respectively (p < 0.001). In the relapse group, the duration between surgery and tumor relapse (average ± SD) was 10.0 ± 8.4 months. The sites of tumor relapse were; liver: 5, lymph node: 2, local recurrence: 1, liver and lymph node: 4, lymph nodes and local recurrence: 2.Table 1 Clinicopathological characteristics of the 19 patients in the training data set Non-relapse Relapse p value Case number 5 14 Age (mean ± SD) 68.4 ± 10.7 61 ± 9.6 0.111 Gender  Male 3 14 0.012  Female 2 0 Tumor location  Upper/middle thoracic 1 7 0.244  Lower thoracic 4 7  Histological grade  Well/moderate 4 13 0.421  Poorly 1 1 Pathologic tumor depth  T1-2 3 3 0.111  T3-4 2 11 Lymph node metastasis  N0 2 3 0.418  N1-3 3 11 Distant metastasis  M0 5 14 –  M1 0 0 TNM stage (IUAC TNM Classification 7th edition)  1–2 3 4 0.211  3–4 2 10 Lymphatic vessel invasion  Negative 2 3 0.418  Positive 3 11 Venous invasion  Negative 3 4 0.211  Positive 2 10 Curability  R0 5 14 –  R1 0 0  Preoperative chemotherapy 0 2 –  Preoperative chemo radiation therapy (CRT) 0 1  Postoperative chemotherapy 0 2 –  Postoperative CRT 1 5  Observation period (month, mean ± SD) 138.7 ± 45.1 14 ± 9.2 <0.001  Duration between surgery and tumor relapse – 10 ± 8.4 – Type of tumor relapse  Liver (H) 0 5  Lymph node (Ly) 0 2  Local (L) 0 1  H + Ly 0 4  Ly + L 0 2 For the validation data set, we obtained high-quality RNA from both tumors and corresponding normal esophageal epithelia from 12 ESCC patients. Eight patients were male and four patients were female, and the average age (average ± SD) was 63.6 ± 9.9 years. All 12 patients underwent radical surgery with no residual tumors (R0). Eight out of 12 patients achieved a 5-year disease-free survival with an observation period (average ± SD) of 64.6 ± 18.3 months, while four patients died of tumor relapse at an observation period (average ± SD) of 51.2 ± 27.4 months after surgery (p = 0.049, Table 4). These two groups were referred to as the relapse group (n = 4) and non-relapse group (n = 8). The expression of miRNAs in FFPE samples of ESCC tumors detected by microarray In the 885 miRNAs assessed using the miRNA oligo chip, the signals of 320 miRNAs were detected in all examined samples. Comparisons between the average expression levels of miRNAs in the 19 ESCC tumors and those in the 19 corresponding normal tissues revealed that the expression of 10 miRNAs (miR-16, miR-93, miR-200c, miR-15b, miR-25-3p, miR-34a, miR-181a, miR-107, miR-103a, and miR-151a) were more than twofold higher, while that of another 10 miRNAs (miR-133b, miR-513, miR-1224, miR-30c, miR-1236, miR-378a, miR-550a, miR-675, miR-149, and miR-1973) were more than twofold lower in the tumors than in their normal counterparts with statistical difference (Fig. 1a). The relationship between the expression of these 20 miRNAs in the tumors (T/N ratio) and clinicopathological features of the tumors were summarized in Additional file 1: Table S1. The higher expression of miR-200c correlated with progressed TNM Stage (p = 0.038) and postoperative tumor relapse (p = 0.038). On the other hand, the expression of the other 19 miRNAs in tumors did not show correlation with clinicopathological characteristics of the patients. Hierarchical clustering based on all 320 detected miRNAs did not show any relationship between clusters and postoperative tumor relapse (Fig. 1b).Fig. 1 Expression of miRNAs in ESCC. a Twenty miRNAs that were expressed in ESCC tumors with more than a twofold difference compared with corresponding normal counter parts. b Cluster analysis. The diagram shows the results of the hierarchical clustering of miRNAs and 19 samples. Columns: 320 miRNAs that were detected in all 19 cases. Red represents a higher expression level; green represents a lower expression level. Rows: samples. Cases 1–5 (highlighted in red) were non-relapse cases. Cases 6–19 were relapse cases Differentially expressed miRNAs between patients with and without postoperative tumor relapse When the expression (T/N ratio) of miRNAs was compared between relapse (n = 14) and non-relapse (n = 5) cases using two types of statistical method (t test and logistic regression analysis), the lower expression of miR-574-3p and miR-106b, as well as higher expression of miR-1303, correlated with postoperative tumor relapse with statistical significance (p < 0.01, Table 2). The Box-and-Whisker plots for these three miRNAs were shown in Fig. 2a.Table 2 Expression of miRNAs associated with postoperative tumor relapse Non-Relapse (n = 5) Relapse (n = 14) t test Logistic regression analysis Average SD Average SD p value p value miR-574-3p 0.87 0.18 0.67 0.05 0.001 0.009 miR-106b 1.18 0.26 0.87 0.16 0.004 0.005 miR-1303 1.11 0.17 1.53 0.24 0.007 0.001 miR-1203 1.34 0.70 0.73 0.24 0.006 0.011 miR-1909 0.92 0.17 0.71 0.13 0.009 0.016 miR-204 1.10 0.35 0.76 0.23 0.016 0.033 miR-371-3p 1.27 0.31 0.94 0.25 0.021 0.019 miR-886-3p 0.90 0.49 1.49 0.55 0.037 0.033 Fig. 2 The relationship between the expression of miRNAs (miR-574-3p, miR-106b, and miR-1303) and patient outcome in the training cohort. a The expression of miR-574-3p, miR-106b, and miR-1303 in non-relapse and relapse groups. b The association between the expression of  miRNAs (miR-574-3p, miR-106b, and miR-1303) and patient prognosis In addition, the lower expression of other four miRNAs (miR-1203, miR-1909, miR-204, miR-371-3p), and higher expression of miR-886-3p, correlated with postoperative tumor relapse (p < 0.05, Table 2). Based on a ROC curve analysis to differentiate patients with relapse from patients without relapse, the cut-off value, the largest AUC value, sensitivity, and specificity for the expression of the top three miRNAs are shown in Table 3.Table 3 Cut-off value for predicting relapse analyzed using receiver-operating characteristic (ROC) curves Name Cut-off AUC Sensitivity Specificity miR-574-3p 0.70 0.89 0.93 0.80 miR-106b 1.00 0.89 0.79 0.80 miR-1303 1.20 0.94 0.86 1.00 Based on the cut-off value listed in Table 3, the overall survival rate of the patients with high expression of miR-574-3p, low expression of miR-1303 were significantly better (Fig. 2b, d). On the other hand, the expression of miR-106b did not correlate with the overall survival rate of the patients (Fig. 2c). The expression of miR-574-3p, miR-106b and miR-1303 in frozen samples of ESCC detected by quantitative RT-PCR analysis Compared with the corresponding normal esophageal mucosa, the expression of miR-574-3p, miR-106b and miR-1303 were upregulated in 8/12 (66.7 %), 11/12 (91.7 %), and 7/12 (58.3 %) of the validation set patients (Fig. 3a–c). When the expression (T/N ratio) of miRNAs was compared between relapse (n = 4) and non-relapse (n = 8) cases, higher expression of miR-574-3p was inversely correlated with tumor relapse (p = 0.029), while the expression of miR-106b and miR-1303 did not show correlation with tumor relapse (Table 4). The Box-and-Whisker plots for these three miRNAs are shown in Fig. 4a.Fig. 3 The expression of miRNAs (miR-574-3p, miR-106b and miR-1303) in ESCC patients in the validation cohort. a The expression (T/N ratio) of miR-574-3p in frozen samples of surgically removed ESCC specimens detected by RT-PCR. b The expression (T/N ratio) of miR-106b in frozen samples of surgically removed ESCC specimens detected by RT-PCR. c The expression (T/N ratio) of miR-1303 in frozen samples of surgically removed ESCC specimens detected by RT-PCR Table 4 Clinicopathological characteristics of the 12 patients in the validation set Non-relapse Relapse p value Case number 8 4 Age (mean ± SD) 63.0 ± 10.3 66.7 ± 8.0 0.425 Gender  Male 4 4 0.08  Female 4 0 Tumor location  Upper/middle thoracic 5 2 0.67  Lower thoracic 3 2 Histological grade  Well/moderate 7 2 0.16  Poorly 1 2 Pathologic tumor depth  T1-2 2 1 1.00  T3-4 6 3  Lymph node metastasis  N0 3 1 0.66  N1-3 5 3 Distant metastasis  M0 8 4 -  M1 0 0 TNM stage  1–2 3 1 0.66  3–4 5 3 Lymphatic vessel invasion  Negative 3 1 0.66  Positive 5 3 Venous invasion  Negative 2 1 1.00  Positive 6 3 Curability  R0 8 4 –  R1 0 0  Preoperative chemotherapy 5 3 – Preoperative CRT 0 0  Postoperative chemotherapy 2 3 –  Postoperative CRT 0 0  Observation period (month, mean ± SD) 64.6 ± 18.3 51.1 ± 33.6 0.049  Duration between surgery and tumor relapse – 8.8 ± 2.9 – Type of tumor relapse  Liver (H) 0 2  Lymph node (Ly) 0 2  miR-574-3p (<0.7) 1 3 0.029  miR-574-3p (≥0.7) 7 1  miR-106b (<1.00) 1 1 0.576  miR-106b (≥1.00) 7 3  miR-1303 (<1.20) 5 2 0.576  miR-1303 (≥1.20) 3 2 Fig. 4 The relationship between the expression of miRNAs (miR-574-3p, miR-106b, and miR-1303) and patient outcome in the validation cohort. a The expression of miR-574-3p, miR-106b and miR-1303 in non-relapse and relapse groups. b The association between miR-574-3 expression and patient prognosis. c The association between miR-106b expression and patient prognosis. d The association between miR-1303 expression and patient prognosis The overall survival rate of the patients with high expression of miR-574-3p was significantly better than that of those with low expression (p = 0.004), while the expression of miR-106b and miR-1303 did not show correlation with overall survival rate of the patients (Fig. 4b–d). Discussion In the 320 miRNAs assessed using the miRNA oligo chip, the expression of 20 miRNAs were aberrantly expressed in the tumors compared with their normal counterparts, with more than a twofold difference. Among these 20 miRNAs, 13 miRNAs, such as miR-16 [8], miR-93 [9], miR-200c [10], miR-25-3p [11], miR-34a [12], miR-181a [13], miR-107 [14], miR-103a [15], miR-151a [16], miR-149 [17], miR-550a [18], miR-378a [19], and miR-30c [20], have been associated with the malignant potential of esophageal cancer. In addition, the other seven miRNAs, such as miR-15b-5p [21], miR-1973 [22], miR-675-3p [23], miR-1236-3p [24], miR-1224-3p [25], miR-513a-5p [26], and miR-133b [27], have been associated with the malignant potential of other types of cancer. These results indicated that the archival FFPE samples were successfully used to identify differentially expressed miRNAs in this study, and suggested the crucial roles of these miRNAs in the development and progression of ESCC. The expression profiles of miRNAs have been shown to differentiate molecular subtypes in several types of cancer, such as pancreatic cancer [28] and small cell carcinoma of the esophagus [29]. A previous report from our laboratory has demonstrated that hierarchical clustering based on all 410 detected miRNAs showed two discrete clusters in primary small cell carcinoma of the esophagus, and these two clusters were identical to the two patients groups, i.e., patients with and without postoperative tumor relapse [29]. However, in this present study, hierarchical clustering based on all 320 detected miRNAs did not show a relationship between clusters and postoperative tumor relapse, indicating the need to select a specific gene set to predict postoperative outcomes. All the 15 miRNAs that were significantly expressed differentially between the relapse and non-relapse groups (p < 0.05) have been reported to be either oncogenes [18, 29–36], or tumor suppressors [37–42] in various type of tumors. In the list of miRNAs, the higher expression of miR-574-3p, which showed the most significant inverse association with postoperative tumor relapse (p = 0.001), was linked with a significantly better overall survival rate of the patients in the training set. In the validation set of the patients, higher expression of miR-574-3p was inversely associated with tumor relapse (p = 0.029) and a better overall survival rate (p = 0.004), further suggesting that the aberrant expression of the miRNAs identified in this study plays key roles in the postoperative outcome of ESCC. miR-574-3p has been reported to be a tumor suppressor miRNA in various cancers [43]. In gastric cancer, reduced expression of miR-574-3p in tumors and inhibition of cell proliferation, migration, and invasion in miR-574-3p-transfected cancer cells have been reported [37]. In bladder cancer cells, miR-574-3p has been reported to target mesoderm development candidate 1 (MESDC1) miRNA, subsequently inhibiting cell proliferation, migration and invasion ability, and induced cell apoptosis [44]. In prostate cancer, miR-574-3p was reported to regulate “Wnt signaling” to reduce cell proliferation by targeting epidermal growth factor receptor (EGFR) expression [43]. On the other hand, this is the first report of a correlation between the down-regulation of miR-574-3p expression and poor prognosis in patients with ESCC, suggesting the tumor suppressor effects of miR-574-3p in ESCC. Although molecular targets of miR-574-3p in ESCC have yet to be investigated, the Wnt signaling pathway has been reported to play an important role in progression, metastasis, and invasion in ESCC [45]. Overexpression of EGFR was also associated with aggressive biological behaviors in ESCC [46]. Accordingly, it is possible that miR-574-3p targets the Wnt signaling pathway and/or EGFR to suppress malignant features in ESCC as well. Investigations to reveal the biological role of miR-574-3p, such as transfection of ESCC cell lines with miR-574-3p expression vector to assess the regulation of malignant phenotype, may provide us with the basis of the prognostic significance of miR-574-3p. It may also provide us with molecular targets to develop novel diagnostic and/or therapeutic strategies. In this study, we obtained formalin-fixed paraffin-embedded (FFPE) samples of the tumors and their normal counterparts, with adequate tissue size from 36 cases and extracted high-quality RNA from 19 patients. After these sample collection with technical limitation, clinicopathological characteristics of the 19 patients summarized in Table 1 showed distinctive features compared to general population of ESCC patients. For example, higher incidence of postoperative tumor relapse (14/19, 73.7 %) was seen and the major type (9/14, 64.3 %) of the tumor relapse were hematologic metastasis. Therefore, it is possible that the pathological T and N factor did not correlate with postoperative tumor relapse in our study because of the deviation in the sample collection. Further investigations based on the large-scale collection of samples without deviation is needed to confirm our results. Conclusions Our microarray analysis using the archived FFPE samples of surgically removed ESCC tumors from 19 training set patients identified 8 miRNAs that were differentially expressed between the patients with and without postoperative tumor relapse. The higher expression of miR-574-3p, which showed the most significant inverse association with postoperative tumor relapse, was significantly associated with better overall survival. Quantitative real-time PCR experiments using frozen samples of ESCC tumors from 12 independent validation set patients confirmed that the higher expression of miR-574-3p was significantly associated with non-tumor relapse and better overall survival. Although we only examined a small number of the cases, to the best of our knowledge, this is the first study to show the association between the expression of miR-574-3p and postoperative outcomes in ESCC patients. Further investigations based on the large-scale collection of samples are awaited to assess the clinical use of miR-574-3p as prognostic predictor in patients with ESCC. Molecular biological investigations to define the mechanisms by which miR-574-3p suppresses malignant feature of ESCC using cell lines also provide us with the basis of its prognostic significance and a molecular target to develop novel diagnostic and/or therapeutic strategies. Additional file Additional file 1: Table S1. The relationship between the expression of the 20 miRNAs in the tumors (T/N ratio) and clinicopathological features of the patients. (XLSX 20 kb) Abbreviations ESCCEsophageal squamous cell carcinoma miRNAsMicroRNAs FFPEFormalin-fixed paraffin-embedded ROCReceiver-operating characteristic AUCArea under the curve Acknowledgements Not applicable. Funding This work was supported by JSPS KAKENHI Grant Numbers JP15 K10088 and JP15 K10089. Availability of data and material The authors cannot share the raw data of the microarray analysis because of intellectual property. Authors’ contributions TO was involved in the conception and design, informed consent, surgery, experiments, and manuscript writing. HK, TM, SS, and TN were involved in the conception, experiments, and manuscript writing. YS was involved in the conception and design and manuscript writing. 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 Written informed consent was obtained from patients with approval by the Institutional Review Board in the Toyama University Hospital (No.20-75). ==== Refs References 1. Thallinger CM Raderer M Hejna M Esophageal cancer: a critical evaluation of systemic second-line therapy J Clin Oncol 2011 29 4709 4714 10.1200/JCO.2011.36.7599 22067408 2. Ando N Kato H Igaki H Shinoda M Ozawa S Shimizu H A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907) Ann Surg Oncol 2012 19 68 74 10.1245/s10434-011-2049-9 21879261 3. Hara H Tahara M Daiko H Kato K Igaki H Kadowaki S Phase II feasibility study of preoperative chemotherapy with docetaxel, cisplatin, and fluorouracil for esophageal squamous cell carcinoma Cancer Sci 2013 104 1455 1460 10.1111/cas.12274 23991649 4. Baek D Villén J Shin C Camargo FD Gygi SP Bartel DP The impact of microRNAs on protein output Nature 2008 455 464 471 10.1038/nature07242 5. Lin RJ Xiao DW Liao LD Chen T Xie ZF Huang WZ MiR-142-3p as a potential prognostic biomarker for esophageal squamous cell carcinoma J Surg Oncol 2012 105 175 182 10.1002/jso.22066 21882196 6. Sobin LH Gospodarowicz M Wittekind C TNM Classification of Malignant Tumours, 7th Edition, UICC International Union Against Cancer 2010 2010 Hoboken, NJ Wiley-Blackwell 7. Osawa S Shimada Y Sekine S Okumura T Nagata T Fukuoka J MicroRNA profiling of gastric cancer patients from formalin-fixed paraffin-embedded samples Oncol Lett 2011 2 613 619 22848236 8. Zhu Y Xia Y Niu H Chen Y MiR-16 induced the suppression of cell apoptosis while promote proliferation in esophageal squamous cell carcinoma Cell Physiol Biochem 2014 33 1340 1348 10.1159/000358701 24852767 9. Ansari MH, Irani S, Edalat H, Amin R, Mohammadi RA. Deregulation of miR-93 and miR-143 in human esophageal cancer. Tumour Biol. 2015;1 [Epub ahead of print]. 10. Yu H Duan B Jiang L Lin M Sheng H Huang J Serum miR-200c and clinical outcome of patients with advanced esophageal squamous cancer receiving platinum-based chemotherapy Am J Transl Res 2013 6 71 77 24349623 11. Komatsu S Ichikawa D Hirajima S Kawaguchi T Miyamae M Okajima W Plasma microRNA profiles: identification of miR-25 as a novel diagnostic and monitoring biomarker in oesophageal squamous cell carcinoma Br J Cancer 2014 111 1614 1624 10.1038/bjc.2014.451 25117812 12. Nie J Ge X Geng Y Cao H Zhu W Jiao Y miR-34a inhibits the migration and invasion of esophageal squamous cell carcinoma by targeting Yin Yang-1 Oncol Rep 2015 34 311 317 25954903 13. Xiang Z Dong X Sun Q Li X Yan B Clinical significance of up-regulated miR-181a in prognosis and progression of esophageal cancer Acta Biochim Biophys Sin 2014 46 1007 1010 10.1093/abbs/gmu083 25230784 14. Sharma P Saraya A Gupta P Sharma R Decreased levels of circulating and tissue miR-107 in human esophageal cancer Biomarkers 2013 18 322 330 10.3109/1354750X.2013.781677 23627613 15. Guo Y Chen Z Zhang L Zhou F Shi S Feng X Distinctive microRNA profiles relating to patient survival in esophageal squamous cell carcinoma Cancer Res 2008 68 26 33 10.1158/0008-5472.CAN-06-4418 18172293 16. Gu J Wang Y Wu X MicroRNA in the pathogenesis and prognosis of esophageal cancer Curr Pharm Des 2013 19 1292 1300 10.2174/1381612811319350009 23092349 17. Zhang C Wang C Chen X Yang C Li K Wang J Expression profile of microRNAs in serum: a fingerprint for esophageal squamous cell carcinoma Clin Chem 2010 56 1871 1879 10.1373/clinchem.2010.147553 20943850 18. Wu C Li M Hu C Duan H Prognostic role of microRNA polymorphisms in patients with advanced esophageal squamous cell carcinoma receiving platinum-based chemotherapy Cancer Chemother Pharmacol 2014 73 335 341 10.1007/s00280-013-2364-x 24288122 19. Hummel R Sie C Watson DI Wang T Ansar A Michael MZ MicroRNA signatures in chemotherapy resistant esophageal cancer cell lines World J Gastroenterol 2014 20 14904 14912 10.3748/wjg.v20.i40.14904 25356050 20. Zhan C Yan L Wang L Jiang W Zhang Y Xi J Landscape of expression profiles in esophageal carcinoma by The Cancer Genome Atlas data Dis Esophagus 2015 21. Hanniford D Zhong J Koetz L Gaziel-Sovran A Lackaye DJ Shang S A miRNA-based signature detected in primary melanoma tissue predicts development of brain metastasis Clin Cancer Res 2015 21 4903 4912 10.1158/1078-0432.CCR-14-2566 26089374 22. El-Awady RA Hersi F Al-Tunaiji H Saleh EM Abdel-Wahab AH Al Homssi A Epigenetics and miRNA as predictive markers and targets for lung cancer chemotherapy Cancer Biol Ther 2015 16 1056 1070 10.1080/15384047.2015.1046023 25962089 23. He D Wang J Zhang C Shan B Deng X Li B Down-regulation of miR-675-5p contributes to tumor progression and development by targeting pro-tumorigenic GPR55 in non-small cell lung cancer Mol Cancer 2015 14 73 10.1186/s12943-015-0342-0 25889562 24. Wang Y Yan S Liu X Zhang W Li Y Dong R miR-1236-3p represses the cell migration and invasion abilities by targeting ZEB1 in high-grade serous ovarian carcinoma Oncol Rep 2014 31 1905 1910 24573236 25. Qian J Li R Wang YY Shi Y Luan WK Tao T MiR-1224-5p acts as a tumor suppressor by targeting CREB1 in malignant gliomas Mol Cell Biochem 2015 403 33 41 10.1007/s11010-015-2334-1 25648114 26. Chen J Jiang CC Jin L Zhang XD Regulation of PD-L1: a novel role of pro-survival signalling in cancer Ann Oncol 2016 27 409 416 10.1093/annonc/mdv615 26681673 27. Tao J Wu D Xu B Qian W Li P Lu Q microRNA-133 inhibits cell proliferation, migration and invasion in prostate cancer cells by targeting the epidermal growth factor receptor Oncol Rep 2012 27 1967 1975 22407299 28. Namkung J Kwon W Choi Y Yi SG Han S Kang MJ Molecular subtypes of pancreatic cancer based on miRNA expression profiles have independent prognostic value J Gastroenterol Hepatol 2015 29. Okumura T Shimada Y Omura T Hirano K Nagata T Tsukada K MicroRNA profiles to predict postoperative prognosis in patients with small cell carcinoma of the esophagus Anticancer Res 2015 35 719 727 25667451 30. Kan T Sato F Ito T Matsumura N David S Cheng Y The miR-106b-25 polycistron, activated by genomic amplification, functions as an oncogene by suppressing p21 and Bim Gastroenterology 2009 136 1689 1700 10.1053/j.gastro.2009.02.002 19422085 31. Lerebours F Cizeron-Clairac G Susini A Vacher S Mouret-Fourme E Belichard C miRNA expression profiling of inflammatory breast cancer identifies a 5-miRNA signature predictive of breast tumor aggressiveness Int J Cancer 2013 133 1614 1623 10.1002/ijc.28171 23526361 32. Östling P Leivonen SK Aakula A Kohonen P Mäkelä R Hagman Z Systematic analysis of microRNAs targeting the androgen receptor in prostate cancer cells Cancer Res 2011 71 1956 1967 10.1158/0008-5472.CAN-10-2421 21343391 33. Harata K Ishiguro H Kuwabara Y Kimura M Mitsui A Ogawa R MicroRNA-34b has an oncogenic role in esophageal squamous cell carcinoma Oncol Lett 2010 1 685 689 22966364 34. Cao J Song Y Bi N Shen J Liu W Fan J DNA methylation-mediated repression of miR-886-3p predicts poor outcome of human small cell lung cancer Cancer Res 2013 73 3326 3335 10.1158/0008-5472.CAN-12-3055 23592755 35. Guzel E Karatas OF Semercioz A Ekici S Aykan S Yentur S Identification of microRNAs differentially expressed in prostatic secretions of patients with prostate cancer Int J Cancer 2015 136 875 879 10.1002/ijc.29054 24976077 36. Pignot G Cizeron-Clairac G Vacher S Susini A Tozlu S Vieillefond A microRNA expression profile in a large series of bladder tumors: identification of a 3-miRNA signature associated with aggressiveness of muscle-invasive bladder cancer Int J Cancer 2013 132 2479 2491 10.1002/ijc.27949 23169479 37. Su Y Ni Z Wang G Cui J Wei C Wang J Aberrant expression of microRNAs in gastric cancer and biological significance of miR-574-3p Int Immunopharmacol 2012 13 468 475 10.1016/j.intimp.2012.05.016 22683180 38. Della Vittoria Scarpati G Falcetta F Carlomagno C Ubezio P Marchini S De Stefano A A specific miRNA signature correlates with complete pathological response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer Int J Radiat Oncol Biol Phys 2012 83 1113 1119 10.1016/j.ijrobp.2011.09.030 22172905 39. Vimalraj S Miranda PJ Ramyakrishna B Selvamurugan N Regulation of breast cancer and bone metastasis by microRNAs Dis Markers 2013 35 369 387 10.1155/2013/451248 24191129 40. Sun Y Zhang T Wang C Jin X Jia C Yu S MiRNA-615-5p functions as a tumor suppressor in pancreatic ductal adenocarcinoma by targeting AKT2 PLoS One 2015 10 10.1371/journal.pone.0119783 25856297 41. Li W Chang J Wang S Liu X Peng J Huang D miRNA-99b-5p suppresses liver metastasis of colorectal cancer by down-regulating mTOR Oncotarget 2015 6 24448 24462 10.18632/oncotarget.4423 26259252 42. Yang Y Liu H Wang X Chen L Up-regulation of microRNA-664 inhibits cell growth and increases cisplatin sensitivity in cervical cancer Int J Clin Exp Med 2015 8 18123 18129 26770409 43. Chiyomaru T Yamamura S Fukuhara S Hidaka H Majid S Saini S Genistein up-regulates tumor suppressor microRNA-574-3p in prostate cancer PLoS One 2013 8 10.1371/journal.pone.0058929 23554959 44. Tatarano S Chiyomaru T Kawakami K Enokida H Yoshino H Hidaka H Novel oncogenic function of mesoderm development candidate 1 and its regulation by MiR-574-3p in bladder cancer cell lines Int J Oncol 2012 40 951 959 22179486 45. Deng F Zhou K Cui W Liu D Ma Y Clinicopathological significance of wnt/β-catenin signaling pathway in esophageal squamous cell carcinoma Int J Clin Exp Pathol 2015 8 3045 3053 26045816 46. Lin G Sun XJ Han QB Wang Z Xu YP Gu JL Epidermal growth factor receptor protein overexpression and gene amplification are associated with aggressive biological behaviors of esophageal squamous cell carcinoma Oncol Lett 2015 10 901 906 26622592

PMC005xxxxxx/PMC5002116.txt

==== Front Cardiovasc DiabetolCardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 43710.1186/s12933-016-0437-yOriginal InvestigationEffects of 6-month eicosapentaenoic acid treatment on postprandial hyperglycemia, hyperlipidemia, insulin secretion ability, and concomitant endothelial dysfunction among newly-diagnosed impaired glucose metabolism patients with coronary artery disease. An open label, single blinded, prospective randomized controlled trial Sawada Takahiro +81-79-293-3131taktak319@yahoo.co.jp 1Tsubata Hideo hideot1122@yahoo.co.jp 1Hashimoto Naoko nhashimoto@hbhc.jp 2Takabe Michinori mictak1@hotmail.com 2Miyata Taishi taishi820@hotmail.com 1Aoki Kosuke kooske2@yahoo.co.jp 1Yamashita Soichiro galmonbozia20@yahoo.co.jp 1Oishi Shogo soishi@hbhc.jp 1Osue Tsuyoshi t.ohsue@gmail.com 1Yokoi Kiminobu kyokoi@hbhc.jp 1Tsukishiro Yasue yasue-oka@nifty.com 1Onishi Tetsuari tonishi@med.kobe-u.ac.jp 1Shimane Akira ashimane@hbhc.jp 1Taniguchi Yasuyo ytaniguchi0@gmail.com 1Yasaka Yoshinori yoshiysk@yahoo.co.jp 1Ohara Takeshi oohara@hp.pref.hyogo.jp 2Kawai Hiroya hikawai@hbhc.jp 1Yokoyama Mitsuhiro Mitsuhiro_Yokoyama@pref.hyogo.lg.jp 11 Division of Cardiovascular Medicine, Department of Internal Medicine, Hyogo Prefectural Himeji Cardiovascular Center, 520 Saisho-Kou, Himeji, Hyogo 670-0981 Japan 2 Division of Diabetes and Endocrinology, Hyogo Prefectural Himeji Cardiovascular Center, 520, Saisho-Kou, Himeji, Hyogo 670-0981 Japan 26 8 2016 26 8 2016 2016 15 1 12115 7 2016 12 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 Recent experimental studies have revealed that n-3 fatty acids, such as eicosapentaenoic acid (EPA) regulate postprandial insulin secretion, and correct postprandial glucose and lipid abnormalities. However, the effects of 6-month EPA treatment on postprandial hyperglycemia and hyperlipidemia, insulin secretion, and concomitant endothelial dysfunction remain unknown in patients with impaired glucose metabolism (IGM) and coronary artery disease (CAD). Methods and results We randomized 107 newly diagnosed IGM patients with CAD to receive either 1800 mg/day of EPA (EPA group, n = 53) or no EPA (n = 54). Cookie meal testing (carbohydrates: 75 g, fat: 28.5 g) and endothelial function testing using fasting-state flow-mediated dilatation (FMD) were performed before and after 6 months of treatment. The primary outcome of this study was changes in postprandial glycemic and triglyceridemic control and secondary outcomes were improvement of insulin secretion and endothelial dysfunction. After 6 months, the EPA group exhibited significant improvements in EPA/arachidonic acid, fasting triglyceride (TG), and high-density lipoprotein cholesterol (HDL-C). The EPA group also exhibited significant decreases in the incremental TG peak, area under the curve (AUC) for postprandial TG, incremental glucose peak, AUC for postprandial glucose, and improvements in glycometabolism categorization. No significant changes were observed for hemoglobin A1c and fasting plasma glucose levels. The EPA group exhibited a significant increase in AUC-immune reactive insulin/AUC-plasma glucose ratio (which indicates postprandial insulin secretory ability) and significant improvements in FMD. Multiple regression analysis revealed that decreases in the TG/HDL-C ratio and incremental TG peak were independent predictors of FMD improvement in the EPA group. Conclusions EPA corrected postprandial hypertriglyceridemia, hyperglycemia and insulin secretion ability. This amelioration of several metabolic abnormalities was accompanied by recovery of concomitant endothelial dysfunction in newly diagnosed IGM patients with CAD. Clinical Trial Registration UMIN Registry number: UMIN000011265 (https://www.upload.umin.ac.jp/cgi-open-bin/ctr/ctr.cgi?function=brows&action=brows&type=summary&recptno=R000013200&language=E) Electronic supplementary material The online version of this article (doi:10.1186/s12933-016-0437-y) contains supplementary material, which is available to authorized users. Keywords Eicosapentaenoic acidImpaired glucose metabolismPostprandial hyperglycemiaPostprandial insulin secretionPostprandial hypertriglyceridemiaEndothelial dysfunctionissue-copyright-statement© The Author(s) 2016 ==== Body Background Type 2 diabetes mellitus (DM) is a prevalent disease and recognized as a major risk factor for coronary artery disease (CAD) [1, 2]. Epidemiological studies and meta-analyses indicate that dietary modifications that increase the relative abundance of dietary n-3 polyunsaturated fatty acids (PUFA), such as eicosapentaenoic acid (EPA) and docosahexaenoic acid, can reduce the risks of DM as well as CAD [3–5] Most animal experiments also document beneficial effects of n-3 PUFA on insulin sensitivity, secretion, and glucose metabolism under condition of obesity, insulin resistance and DM [6, 7]. These findings suggest that n-3 PUFA may help control and prevent DM possibly due to improvement of insulin secretion and sensitivity. However, data from clinical trials have been conflicting. Some studies indicated that n-3 PUFA improved insulin sensitivity in humans, whereas others found that n-3 PUFA had no insulin sensitizing effects [8]. Laria et al. reported that dietary EPA and docosahexaenoic acid did not alter peripheral insulin sensitivity, postprandial glucose disposal, or insulin secretion in insulin-resistant, non-diabetic overweight individual [9]. In contrast, a randomized clinical trial in overweight patients with DM revealed a beneficial effect of purified EPA on glucose homeostasis and insulin sensitivity [10]. Thus, the effect of EPA on the prevention or exacerbation of DM still remains unknown and has not yet been studied in clinical intervention trials in high-risk patients with impaired glucose metabolism (IGM). We hypothesized that long-term EPA treatment would predominantly improve postprandial state of insulin secretion and postprandial hyperglycemia and hypertriglyceridemia, and lead to prevent development of DM in patients with newly diagnosed IGM. In addition, these metabolic effects would be associated with improvement in concomitant endothelial dysfunction. Therefore, the present study was designed to examine the effects of EPA on postprandial plasma metabolic parameters and fasting endothelial function among newly diagnosed IGM patients with CAD. Methods Participants This study was open label, single blinded, randomized controlled study approved by the Ethics Committee of Hyogo Prefectural Himeji Cardiovascular Center and complied with the Declaration of Helsinki. Informed written consent was obtained from all eligible patients before randomization to either EPA treatment or non-EPA treatment groups. This study is registered in the UMIN Clinical Trials Registry under the identifier UMIN000011265 (https://www.upload.umin.ac.jp/cgi-open-bin/ctr/ctr.cgi?function=brows&action=brows&type=summary&recptno=R000013200&language=E). All potential participants who had documented chronic CAD, and with hemoglobin A1c levels <6.5 % (using the National Glycohemoglobin Standardization Program method), fasting plasma glucose (PG) levels ≥116 mg/dL or fasting plasma triglyceride (TG) levels ≥150 mg/dL by screening laboratory examination, were recruited from outpatients between July 2013 and December 2014. CAD was defined as stenosis of >50 % of the diameter of a coronary artery on angiography or computed tomography coronary angiography, or a history of myocardial infarction, percutaneous coronary intervention, or coronary artery bypass surgery. They underwent a 75-g oral glucose tolerance test (OGTT) and were diagnosed with DM or impaired glucose tolerance (IGT) according to Japan Diabetic Society criteria (DM: fasting PG ≥126 mg/dL and PG ≥200 mg/dL at 2 h after the OGTT; IGT: PG ≥140 mg/dL at 2 h after the OGTT). Patients were eligible to participate if they had newly diagnosed IGM (defined as IGT or DM during the OGTT). The exclusion criteria were presence of established DM, treatment history by anti-diabetic agent, systemic disease, including hepatic disease, renal disease (serum creatinine levels ≥2.5 mg/dL), collagen disease, infection, or malignancy and acute coronary syndrome. As shown in Fig. 1, 155 patients underwent OGTT screening, and 118 patients were diagnosed with IGM. At enrollment, all patients met at least once with a dietician for nutritional guidance and were encouraged to start and maintain a low-calorie diet and mild-to-moderate exercise levels. And then, 118 patients were randomly assigned to the EPA group (59 patients) or the non-EPA group (59 patients). Randomization was performed by means of random, permuted blocks of four in sealed envelope. We used EPADEL soft capsule (Mochida Pharmaceutical Co. Ltd., Tokyo, Japan) containing 900 mg of highly (≥98 %) purified EPA ethyl ester per capsule. In this drug product, EPA is actually purified from long-chain polyunsaturated fatty acids present in fish oil. We adopted the most widely used therapeutic dose of 900 mg of EPA ethyl ester capsule administered orally twice a day immediate after meals, as these dose of EPA has been shown to possess the beneficial effects in humans [11, 12].Fig. 1 Flow chart of patient recruitment and participation All patients were followed without hypoglycemic drugs and insulin treatment during the study period. And all participants were also requested to maintain any concomitant drug treatment throughout the study. At each visit, we questioned the participants regarding any adverse events. Anthropometric measurements (body weight and body mass index) and systolic and diastolic blood pressures were measured at randomization and after 6 months of treatment. Blood pressure was measured with the patient in the supine position after 15 min of rest. Blood biochemistry Blood samples were collected from all patients after an overnight fast and were used to determine PG, hemoglobin A1c, 1,5-anhydro-glucitol, immune reactive insulin (IRI), total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein (LDL) cholesterol, TG, remnant-like particle cholesterol, and C-reactive protein. All biochemical analyses were performed using a commercially available kit. Hemoglobin A1c levels were measured using high-performance liquid chromatography. IRI concentrations were measured using a chemiluminescent enzyme immunoassay and remnant-like particle cholesterol was measured using an assay kit (Japan Immunoresearch Laboratories Co., Ltd., Takasaki, Japan). Plasma total fatty acid concentrations were measured by a central laboratory (BML Inc. Saitama, Japan), which was previously described [12, 13]. Plasma fatty acid composition was determined by capillary gas chromatography. Briefly, plasma lipids were extracted by Folch’s procedure, and then fatty acids were methylated with boron trifluoride and methanol, and methylated fatty acids were analyzed using the SHIMAZU GC-17A gas chromatograph (Shimazu Corporation, Kyoto, Japan) and a BPX70 capillary column (0.25 mm ID*30 m; SGE international Ltd., Melbourne, Australia) [13]. Cookie meal test A cookie meal test (CMT) was performed for glucose and fat loading at randomization and 6-month follow-up. The cookie consisted of carbohydrates (75 g; 85 % flour starch, 15 % maltose), butter fat (28.5 g), and protein (8 g) with 592 kcal (SARAYA Corp., Osaka, Japan). The CMT is sufficient for providing information regarding glucose intolerance and postprandial hypertriglyceridemia. The same values of cut-off point is proposed to be used at 2h PG for the evaluation of DM and IGT as WHO criteria for OGTT with the exception of exocrine pancreatic dysfunction [14]. Therefore, we used the CMT to simultaneously evaluate postprandial hyperglycemia (using the same criteria as for the OGTT) and postprandial hypertriglyceridemia. All participants completed a 12-h overnight fast before the CMT. The cookie was ingested with water within a 20-min period, and blood samples for PG, IRI, and TG measurement were obtained at 0, 1 and 2 h after the participant had ingested half volume of the cookie [14]. These values were reported as PG-1h, PG-2h, IRI-1h, IRI-2h, TG-1h and TG-2h, respectively. As an index of postprandial hyperglycemia, we calculated the incremental glucose peak [(the maximal incremental increase in PG at any point after meal loading)—(fasting PG)]. We also calculated the area under the response curves for PG (AUC-PG) and IRI (AUC-IRI) using the trapezoid rule. And then, we calculated AUC-IRI/AUC-PG as an index of postprandial insulin secretion ability. The homeostasis model assessment ratio (HOMA-R = fasting IRI × fasting PG/405) was used as an index of insulin resistance. Incremental TG peak [(the maximal incremental increase in TG at any point after loading)—(fasting TG)] was calculated as an index of postprandial hypertriglyceridemia. The trapezoid rule was used to calculate the area under the TG-response curve (AUC-TG). Small dense LDL particles, which are more susceptible to oxidation, have been suggested to be more atherogenic than large buoyant LDL [15, 16]. In this study, the TG/HDL-C ratio was used to assess the presence of small dense LDL particles [17]. Measurement of endothelial function We selected percent change in flow-mediated dilatation (FMD) as the surrogate marker for endothelial function [18–20] Before undergoing FMD testing, the participants were instructed to fast for >12 h and to abstain from any medications, smoking, alcohol, caffeine, and antioxidant vitamins during that time. All participants rested for at least 15 min in a seated position in a quiet dark air-conditioned room (22–25 °C) before the FMD measurements, which was previously described [18, 19]. In brief, a longitudinal image of the right or left brachial artery was recorded at baseline using high-resolution ultrasonography and a 10-MHz linear array transducer probe (UNEX, Nagoya, Japan). A forearm-cuff was inflated for 5 min at 50 mmHg above the systolic blood pressure just prior to FMD measurement. After cuff deflation, the diastolic diameter of the brachial artery was semi-automatically and continuously recorded for 2 min using software-equipped instrument that could monitor arterial diameter. The %FMD was estimated as the percent change in the vessel diameter, which corresponded to the maximum dilatation that was reached during reactive hyperemia divided by the baseline value. Because the %FMD value is highly dependent on the baseline diameter of the vessel, we compared the baseline diameters and absolute changes in the brachial artery’s diameters for each group. We have also previously confirmed that there is excellent intra- and inter-observed agreement for the %FMD measurement [18, 19]. Primary and secondary endpoints The primary endpoint of this study was postprandial changes in glucose concentration (incremental glucose peak) and postprandial changes in TG (incremental TG peak), derived from CMT after 6 months treatment of EPA. The secondary endpoints was improvement in postprandial insulin secretion ability and concomitant endothelial dysfunction. Statistical analysis All statistical analyses were performed using MedCalc software (Version 9.3; Mariakerke, Belgium), and a two-tailed P < 0.05 was considered statistically significant. Sample size was determined by power analysis using preliminary data obtained in our laboratory with the following assumptions: Type I error of 0.05 (two-tailed), power of 90 %, difference in the incremental glucose peak and incremental TG peak between before and after EPA treatment of 20 and 20 mg/dL, respectively, and a standard deviation of 30 and 25, respectively. Therefore, a minimum of 31 patients would yield 90 % power to detect a significant difference. All variables were checked for normal distribution using Kolmogorov–Smirnov algorithm. Continuous variables with normal distribution were reported as mean ± standard deviation, otherwise as median and interquartile ranges. Inter-group comparisons of normal distributed data were performed using the unpaired Student’s t test, and Mann–Whitney test was used if the variables did not indicate normal distribution. Comparison of categorical variables was performed using χ2 test. Differences between the biochemical data with normal distribution before and after treatment were compared using a paired t test, otherwise using the Wilcoxon rank-sum test. Multivariable regression analyses were used to identify factors that were independently associated with the changes in incremental glucose peak and %FMD. Biologically plausible factors were included in the original model. Potential factors of interest with a univariate P ≤ 0.1 were entered into multivariate models. Results Study population Among the 118 eligible patients, five patients in the EPA group and six patients in the non-EPA group withdrew their consent before second CMT. Therefore, the final analyses included data from 54 patients in the EPA group and 53 patients in the non-EPA group (Fig. 1). During the study period, one patient in the EPA group and one patient in the non-EPA group underwent percutaneous coronary intervention due to progression of CAD. However, there were no major adverse events such as massive bleeding complication and gastrointestinal disorder requisite to treat that were related to the EPA treatment. Baseline characteristics, laboratory data, and anthropometric measurements As shown in Table 1, the patients’ baseline characteristics and concomitant drug use were similar for groups. Most baseline laboratory findings were similar between the two groups, although fasting PG levels in the EPA group were lower than those in the non-EPA group. At baseline, resting brachial artery diameters and levels of endothelial function impairment were similar between the two groups. After 6 months treatment, both groups exhibited significant reductions in body weight and body mass index although there were no significant inter-group differences in these changes (Table 2).Table 1 Comparison of baseline characteristics and concomitant use drugs between the two groups Non-EPA group (n = 53) EPA group (n = 54) P value Age (y.o) 68.9 ± 8.8 67.8 ± 9.1 0.49 Sex (male, n, %) 43 (81.1 %) 44 (81.5 %) 0.84 Coronary risk factor  Hypertension (n, %) 49 (92.5 %) 48 (88.9 %) 0.76  Dyslipidemia (n, %) 53 (100 %) 54 (100 %) 0.99  Smoking (n, %) 4 (7.5 %) 5 (9.3 %) 0.98 Concomitant use drug  Statin (n, %) 44 (83.0 %) 46 (85.2 %) 0.97  Calcium channel blocker (n, %) 30 (56.6 %) 27 (50.0 %) 0.62  ACEI/ARB (n, %) 35 (66.0 %) 40 (74.1 %) 0.49 Values are presented as number (%) or mean ± standard deviation, as indicated ACEI angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker Table 2 Comparison of anthropometric analysis between baseline and 6 months, and comparison of absolute change from baseline between the two groups Variable Non-EPA group (n = 53) EPA group (n = 54) P value Baseline 6-month Baseline 6-month Weight (kg) 67.8 ± 10.0 66.3 ± 9.9* 67.6 ± 11.2 66.4 ± 11.0* 0.97  Absolute Δ −1.1 (−3.0, 0.0) −1.0 (−2.8, 0.0) 0.49 Body mass index (kg/m2) 25.4 ± 2.4 24.8 ± 2.4* 25.3 ± 2.9 24.8 ± 2.8* 0.90  Absolute Δ −0.5 (−1.1, 0.0) −0.4 (−1.0, 0.0) 0.53 Systolic blood pressure (mmHg) 135.6 ± 15.0 133.0 ± 17.1 132.9 ± 19.5 129.7 ± 19.0 0.35  Absolute Δ −3.0 (−13.0, 8.0) −3.0 (−10.0, 4.0) 0.89 Diastolic blood pressure (mmHg) 78.6 ± 9.8 77.5 ± 10.9 77.2 ± 11.7 76.3 ± 8.2 0.51  Absolute Δ −2.0 (−7.0, 3.3) 0.0 (−7.0, 6.0) 0.63 P values represent comparison of each values between groups at 6 months except for absolute Δ As for absolute Δ of each values, P values represent comparison between the two groups Values are presented as mean ± standard deviation or medians and interquartile ranges, as indicated * P < 0.001 vs baseline Changes in fasting lipid and glucose profiles The fasting lipid and glucose profiles at baseline and 6 months are shown in Table 3. After 6 months of treatment, the median EPA/arachidonic acid (AA) ratio significantly increased in the EPA group (from 0.31 to 1.08, P < 0.0001), and significantly increased HDL-C levels also occurred in the EPA group. Significant reductions in the levels of LDL cholesterol, TG, TG/HDL-C ratio, and remnant-like particle cholesterol were observed in the EPA group, but not in the non-EPA group. The EPA group exhibited significantly increased 1,5-anhydro-glucitol levels, although no changes were observed in the non-EPA group. The changes of TG, TG/HDL-C ratio and 1,5-anhydro-glucitol levels from baseline were also significantly larger in the EPA group than the non-EPA group. Neither group exhibited significant changes in fasting levels of PG, IRI, hemoglobin A1c, and HOMA-R.Table 3 Comparison of biochemical data and flow-mediated dilatation data between baseline and 6 months, and comparison of absolute change from baseline between the two groups Variable Non-EPA group (n = 53) EPA group (n = 54) P value Baseline 6-month Baseline 6-month Fasting PG (mg/dL) 111.2 ± 10.4 109.9 ± 10.8 105.8 ± 11.1† 104.5 ± 9.1 0.001  Absolute Δ −1.0 (−6.0, 4.0) −1.0 (−5.0, 4.0) 0.69 Hemoglobin A1c (NGSP; %) 6.1 ± 0.4 6.1 ± 0.4 6.0 ± 0.3 5.9 ± 0.4 0.02  Absolute Δ 0.0 (−0.1, 0.13) 0.0 (−0.1, 0.10) 0.49 1,5-anhydro-glucitol (μg/mL) 16.5 ± 6.1 16.1 ± 6.4 18.5 ± 7.8 19.3 ± 8.0* 0.02  Absolute Δ −0.6 (−1.8, 1.2) 0.9 (−0.8, 1.9) 0.03 Fasting IRI (µU/mL) 6.4 ± 3.3 6.1 ± 2.9 6.6 ± 3.4 6.5 ± 3.0 0.49  Absolute Δ −0.1 (−1.4, 0.7) −0.2 (−2.0, 1.8) 0.63 HOMA-R 1.6 (1.0, 2.4) 1.6 (1.1, 2.0) 1.6 (1.0, 2.3) 1.6 (1.0, 2.2) 0.86  Absolute Δ 0.0 (−0.4, 0.2) 0.0 (−0.5, 0.5) 0.67 Total cholesterol (mg/dL) 171.7 ± 31.9 169.1 ± 27.7 171.8 ± 30.8 167.4 ± 29.4 0.76  Absolute Δ −1.0 (−12.5, 8.5) −3.5 (−13.0, 7.0) 0.45 LDL cholesterol (mg/dL) 96.7 ± 25.1 95.5 ± 23.3 97.9 ± 27.6 92.5 ± 26.8** 0.53  Absolute Δ 1.0 (−10.2, 8.0) −4.0 (−13.0, 3.0) 0.09 HDL-C (mg/dL) 48.2 ± 9.8 48.2 ± 11.6 46.7 ± 9.3 50.1 ± 13.2** 0.51  Absolute Δ 0.0 (−4.3, 3.0) 2.0 (−3.0, 8.0) 0.05 TG (mg/dL) 123.3 ± 59.2 120.8 ± 62.2 137.8 ± 67.6 103.0 ± 38.1*** 0.10  Absolute Δ −3.0 (−35.0, 19.0) −24.0 (−54.0, −3.0) 0.005 TG/HDL-C ratio 2.46 (1.43, 3.54) 2.18 (1.38, 3.54) 2.58 (1.65, 4.63) 2.02 (1.35, 2.65)*** 0.39  Absolute Δ −0.04 (−0.72, 0.33) −0.73 (−1.44, −0.12) 0.0004 Remnant-like particle-cholesterol (mg/dL) 9.1 ± 6.1 8.4 ± 6.1 10.4 ± 6.4 7.3 ± 3.9*** 0.26  Absolute Δ −1.0 (−3.1, 1.1) −1.7 (−4.8, −0.3) 0.07 EPA (μg/mL) 84.0 ± 50.7 81.1 ± 45.0 69.3 ± 40.7 196.4 ± 51.4*** <0.0001  Absolute Δ −1.0 (−23.5, 24.0) 128.0 (98.0, 164.0) <0.0001 AA (μg/mL) 205.6 ± 55.3 201.2 ± 48.1 206.3 ± 52.6 177.0 ± 39.7*** 0.006  Absolute Δ −2.0 (−34.8, 22.5) −30.0 (−52.0, −3.0) 0.001 Docosahexaenoic acid (μg/mL) 162.5 ± 64.6 152.2 ± 57.7 158.9 ± 70.0 134.2 ± 48.1** 0.08  Absolute Δ −18.0 (−44.3, 15.8) −21.5 (−55.0, 12.0) 0.27 EPA/AA ratio 0.35 (0.25, 0.52) 0.36 (0.27, 0.46) 0.31 (0.21, 0.49) 1.08 (0.88, 1.45)*** <0.0001  Absolute Δ −0.01 (−0.11, 0.10) 0.78 (0.52, 1.05) <0.0001 Docosahexaenoic acid/AA 0.75 (0.59, 1.02) 0.74 (0.63, 0.96) 0.78 (0.53, 0.97) 0.77 (0.56, 0.95) 0.92  Absolute Δ −0.01 (−0.19, 0.11) −0.01 (−0.15, 0.11) 0.71 C-reactive protein (mg/dL) 0.09 (0.04, 0.14) 0.06 (0.04, 0.12) 0.10 (0.04, 0.22) 0.06 (0.03, 0.11)*** 0.55  Absolute Δ −0.01 (−0.05, 0.02) −0.01 (−0.08, 0.00) 0.05 %FMD (%) 4.1 ± 1.8 4.0 ± 1.6 3.6 ± 1.7 5.2 ± 1.9*** 0.0005  Absolute Δ 0.1 (−0.6, 0.8) 1.6 (0.7, 2.5) <0.0001 Rest brachial artery diameter (mm) 4.27 ± 0.55 4.23 ± 0.58 4.23 ± 0.60 4.23 ± 0.58 0.72  Absolute change in brachial artery diameter (mm) 0.17 ± 0.07 0.17 ± 0.06 0.15 ± 0.07 0.22 ± 0.07*** 0.0005  Absolute Δ 0.01 (−0.03, 0.03) 0.06 (0.02, 0.08) <0.0001 P values represent comparison of each values between groups at 6 months except for absolute Δ As for absolute Δ of each values, P values represent comparison between the two groups Values are presented as mean ± standard deviation or medians and interquartile ranges, as indicated PG plasma glucose, IRI immune reactive insulin, HOMA-R homeostasis model assessment ratio, LDL low-density lipoprotein, HDL-C high-density lipoprotein cholesterol, TG triglyceride, EPA eicosapentaenoic acid, AA arachidonic acid, FMD flow-mediated dilatation † P < 0.05 vs baseline values of non-EPA group * P < 0.05 vs baseline, ** P < 0.01 vs baseline, *** P < 0.0001 vs baseline Changes in the CMT results The CMT data from baseline and after 6 months of treatment are shown in Figs. 2, 3 and Table 4. The baseline CMT revealed that, among patients from the non-EPA group, 37 patients (69.8 %) had IGT and 16 patients (30.2 %) had DM. Similarly, among patients from the EPA group, 38 patients (70.4 %) had IGT and 16 patients (29.6 %) had DM (Fig. 2). After 6 months, the EPA group included 16 patients (29.6 %) with normal glucose tolerance, 31 patients (57.4 %) with IGT and seven patients (13.0 %) with DM. The non-EPA group included three patients (5.7 %) with normal glucose tolerance, 35 patients (66.0 %) with IGT and 15 patients (28.3 %) with DM. With regard to changes in glucometabolic category, deteriorated patients were significantly larger in the non-EPA group, while improved patients were significantly larger in the EPA group (deteriorated glucose tolerance; seven patents in the non-EPA group vs two patients in the EPA group, improved glucose tolerance; 11 patients in the non-EPA group vs 25 patients in the EPA group, p = 0.01).Fig. 2 Categorical allocation according to glucose tolerance at baseline and after 6 months. Numbers on the lines represent the number of patients transferred between glucose categories. Deteriorated patients were significantly larger in the non-EPA group, while improved patients were significantly larger in the EPA group (deteriorated glucose tolerance; seven patents in the non-EPA group vs two patients in the EPA group, improved glucose tolerance; 11 patients in the non-EPA group vs 25 patients in the EPA group, p = 0.01). Comparison was performed using χ2 test Fig. 3 Plasma glucose (PG), immune reactive insulin (IRI), and triglyceride (TG) levels during cookie meal test at baseline and 6 months. Bars indicate SEM. * indicates P < 0.05 vs baseline. Comparisons of data between before and after treatment were performed using a paired t test Table 4 Comparison of cookie meal test data between baseline and 6 months, and comparison of absolute change from baseline between the two groups Variable Non-EPA group (n = 53) EPA group (n = 54) P value Baseline 6-month Baseline 6-month Glucose tolerance test  NGT (n, %) 0 (0.0 %) 3 (5.7 %) 0 (0.0 %) 16 (29.6 %)** 0.001  IGT (n, %) 37 (69.8 %) 35 (66.0 %) 38 (70.4 %) 31 (57.4 %) 0.36  DM (n, %) 16 (30.2 %) 15 (28.3 %) 16 (29.6 %) 7 (13.0 %)* 0.05 Fasting PG (mg/dL) 111.2 ± 10.4 109.9 ± 10.8 105.8 ± 11.1† 104.5 ± 9.1 0.009  Absolute Δ −1.0 (−6.0, 4.0) −1.0 (−5.0, 4.0) 0.69 PG-1h (mg/dL) 200.3 ± 32.9 197.0 ± 33.3 190.9 ± 34.4 172.9 ± 26.8** 0.0001  Absolute Δ −5.0 (−23.0, 14.0) −15.5 (−33.0, 1.0) 0.006 PG-2h (mg/dL) 185.1 ± 25.5 180.1 ± 30.8 179.7 ± 32.5 162.2 ± 32.9** 0.004  Absolute Δ −9.0 (−21.3, 19.3) −18.0 (−33.0, 0.0) 0.03 AUC-PG 348.5 ± 42.1 342.0 ± 46.9 333.2 ± 48.9 306.4 ± 41.4** 0.0001  Absolute Δ −7.0 (−13.3, 2.1) −22.3 (−45.0, 10.0) 0.002 Incremental glucose peak (mg/dL) 95.0 ± 27.7 92.1 ± 29.1 90.9 ± 30.3 75.4 ± 26.2** 0.002  Absolute Δ −6.0 (−19.3, 15.3) −16.5 (−33.0, −2.0) 0.006 Fasting IRI (µU/mL) 6.4 ± 3.3 6.1 ± 2.9 6.6 ± 3.4 6.5 ± 3.0 0.49  Absolute Δ −0.1 (−1.4, 0.7) −0.2 (−2.0, 1.8) 0.63 IRI-1h (µU/mL) 53.9 ± 29.7 50.7 ± 25.2 54.3 ± 32.2 57.3 ± 29.8 0.28  Absolute Δ 0.1 (−10.6, 10.4) 0.7 (−7.0, 13.1) 0.27 IRI-2h (µU/mL) 64.4 ± 30.2 58.6 ± 27.2 61.8 ± 35.0 60.1 ± 30.5 0.78  Absolute Δ −6.0 (−15.3, 5.2) 2.5 (−15.5, 10.9) 0.2 AUC-IRI 89.3 ± 42.1 83.0 ± 36.6 88.5 ± 48.3 89.7 ± 42.4 0.38  Absolute Δ −3.7 (−19.5, 8.6) 4.9 (−7.4, 11.4) 0.08 AUC-IRI/AUC-PG 0.23 (0.18, 0.35) 0.24 (0.17, 0.29) 0.24 (0.17, 0.33) 0.27 (0.19, 0.36)** 0.06  Absolute Δ −0.01 (−0.04, 0.03) 0.02 (−0.01, 0.06) 0.003 Fasting TG (mg/dL) 123.3 ± 59.2 120.8 ± 62.2 137.8 ± 67.6 103.0 ± 38.1** 0.10  Absolute Δ −3.0 (−35.0, 19.0) −24.0 (−54.0, −3.0) 0.005 TG-1h (mg/dL) 138.5 ± 61.8 136.2 ± 64.0 150.7 ± 64.7 121.1 ± 37.7** 0.17  Absolute Δ 0.0 (−29.5, 16.5) −20.5 (−57.0, 4.0) 0.01 TG-2h (mg/dL) 168.9 ± 74.8 171.6 ± 89.3 198.3 ± 78.7 150.8 ± 44.7** 0.14  Absolute Δ −2.0 (−33.8, 27.3) −42.0 (−80.0, 10.0) 0.0001 AUC-TG 285.0 ± 126.5 282.0 ± 137.3 316.8 ± 135.1 250.0 ± 76.3** 0.13  Absolute Δ −5.5 (−60.5, 37.3) −51.5 (−127.0, 5.0) 0.002 Incremental TG peak (mg/dL) 45.2 ± 31.2 51.9 ± 43.7 60.5 ± 29.8† 47.8 ± 24.7* 0.54  Absolute Δ 4.0 (−12.0, 21.3) −11.5 (−24.0, 3.0) 0.005 P values represent comparison of each values between groups at 6 months except for absolute Δ As for absolute Δ of each values, P values represent comparison between the two groups Values are presented as mean ± standard deviation or medians and interquartile ranges, as indicated DM diabetes mellitus, IGT impaired glucose tolerance, NGT normal glucose toleransce, PG plasma glucose, AUC area under the response curve, IRI immune reactive insulin, TG triglyceride † P < 0.05 vs baseline values of non-EPA group * P < 0.01 vs baseline, ** P < 0.0001 vs baseline Significant reductions in AUC-PG and incremental glucose peak were only observed for the EPA group (Table 4; Fig. 3). Despite significant improvements in AUC-PG, the EPA group exhibited a mild increase in AUC-IRI, while the non-EPA group exhibited decreased AUC-IRI accompanied by AUC-PG decrease (Fig. 3). Thus, the AUC-IRI/AUC-PG ratio, which indicated postprandial insulin secretion ability, significantly increased only in the EPA group. Compared to baseline data, the EPA group exhibited significant reductions in TG-1h, TG-2h, and fasting TG levels. Furthermore, incremental TG peak and AUC-TG decreased only in the EPA group. As for the comparison of absolute change from baseline between the EPA and non-EPA groups, the changes of PG-1h, PG-2h, AUC-PG, incremental glucose peak, fasting TG, TG-1h, TG-2h, AUC-TG, and incremental TG peak were significantly higher in the EPA group than the non-EPA group. Furthermore, the changes of AUC-IRI/AUC-PG ratio from baseline was significantly larger in the EPA group than the non-EPA group. The multiple regression analysis showed that EPA treatment and lower baseline PG levels were independent factor for predicting improvement of incremental glucose peak (Additional file 1: Table S1). In order to adjust the difference of baseline PG levels between the two groups, additional analysis was performed among patients with similar baseline PG levels of ≤110 mg/dL (Additional file 2: Table S2). The absolute changes of incremental glucose and TG peak in the EPA group remained larger than those in the non-EPA group, and also the change of AUC-IRI/AUC-PG ratio in the EPA group was significantly higher than that in the non-EPA group. Comparing inflammatory markers and endothelial function As shown in Table 3, the improvement of C-reactive protein levels was significantly higher in the EPA group than in the non-EPA group (P = 0.05). Although both groups exhibited impaired endothelial function at baseline, the EPA group exhibited a significant improvement in %FMD after 6 months of treatment. In contrast, endothelial dysfunction was unchanged in the non-EPA group. The results of the univariate and multivariate regression analyses for predicting improvements in %FMD are shown in Table 5. In the EPA group, improvements in the TG/HDL-C ratio and incremental TG peak were independent predictors of %FMD improvements.Table 5 Regression analysis for predicting % flow-mediated dilatation improvement Variable Non-EPA group (n = 53) EPA group (n = 54) Univariate Multivariate Univariate Multivariate t P value t P value t P value t P value Age 2.865 0.006 2.505 0.01 0.085 0.93 0.013 0.99 Sex −1.640 0.11 −1.038 0.30 −0.106 0.92 0.055 0.96 Changes in weight −0.123 0.90 −0.045 0.95 Changes in fasting PG 0.837 0.41 1.531 0.13 Changes in fasting IRI 0.584 0.56 1.093 0.28 Changes in Hemoglobin A1c 0.437 0.66 0.399 0.69 Changes in 1,5-anhydro-glucitol −1.463 0.15 1.542 0.13 Changes in HOMA-R 0.027 0.98 0.596 0.55 Changing in PG-1h −0.675 0.50 −1.305 0.20 Changing in PG-2h 1.776 0.08 −1.092 0.28 −0.923 0.36 Changing in AUC-PG 0.049 0.96 −0.945 0.35 Changing in incremental glucose peak −0.030 0.98 −2.227 0.03 −1.054 0.30 Changing in IRI-1h −0.896 0.37 0.008 0.99 Changing in IRI-2h 1.549 0.13 −1.156 0.25 Changing in AUC-IRI −0.154 0.88 −0.511 0.61 Changing in AUC-IRI/AUC-PG −0.276 0.78 −0.001 0.99 Changes in total cholesterol −0.680 0.50 −0.349 0.73 Changes in LDL cholesterol −0.531 0.60 0.280 0.78 Changes in HDL cholesterol −0.085 0.93 0.328 0.74 Changes in TG −1.925 0.05 0.274 0.79 −2.062 0.04 Changes in log TG/HDL ratio −1.234 0.22 −2.009 0.05 −2.001 0.05 Changing in TG2h 1.234 0.22 −3.842 0.0003 Changing in AUC-TG −2.140 0.03 −0.288 0.77 −2.727 0.009 Changing in incremental TG peak −1.413 0.16 −4.337 <0.0001 −3.560 0.0008 Changes in remnant-like particle cholesterol −2.282 0.03 −2.022 0.05 −0.369 0.71 Changes in EPA −0.269 0.79 −0.213 0.83 Changes in AA −0.174 0.86 −0.142 0.89 Changes in docosahexaenoic acid −0.752 0.46 −1.290 0.20 Changes in EPA/AA −0.469 0.64 0.109 0.91 Changes in C-reactive protein −0.560 0.58 0.328 0.74 PG plasma glucose, IRI immune reactive insulin, HOMA-R homeostasis model assessment ratio, AUC area under the response curve, LDL low-density lipoprotein, HDL high-density lipoprotein, TG triglyceride, EPA eicosapentaenoic acid, AA arachidonic acid Discussion Six months of EPA treatment improved not only fasting atherogenic lipid derangements and postprandial hypertriglyceridemia, but also postprandial hyperglycemia and insulin secretion ability. The simultaneous improvements in several metabolic abnormalities were associated with improvements in concomitant endothelial dysfunction among newly diagnosed IGM patients with CAD who received EPA. Effects of EPA on postprandial hyperglycemia Accumulating data from animal models confirm that n-3 PUFA have numerous beneficial effects on health and diseases, such as anti-arrhythmic, vasodilatory, anti-inflammatory, anti-thrombotic, and lipid lowering actions [21–23]. N-3 PUFA also have a number of metabolic effects including improvement in insulin secretion, sensitivity and anti-obesity action [7, 8]. The recent experiments have revealed that n-3 PUFA act as ligands of several G-protein- coupled receptors (GPCRs) including GPCR 40 on pancreatic β cells and GPCR 120 on enteroendocrine cells of gastrointestinal tract [24, 25]. These actions enforce the beneficial effects on insulin secretion and sensitivity, and glucose homeostasis. However, many previous clinical trials have reported inconsistent results regarding the association of n-3 PUFA intake with the incidence of DM and glycemic control [8–10]. In the present study, we found that EPA treatment resulted in significant improvements in postprandial insulin secretion and glucose homeostasis among patients with IGM and significantly ameliorated the development of DM. Although we did not assess glucagon-like peptide-1 concentrations, we speculate that EPA-activated GPCRs might be actively involved in postprandial insulin secretion. These results suggest that EPA might provide a greater protective effect against developing DM among pre-diabetic patients. The results from the Japan EPA Lipid Intervention Study (JELIS) trial [11] support this possibility, as long-term treatment with EPA resulted in a significant 33 % reduction in the incidence of new-onset DM among patients with IGM (U.S. Patent Application Publication No. US2015/0250754 A1). Although a few previous studies have reported that n-3 PUFA improved insulin resistance [10], we did not observe any significant improvement in HOMA-R for the EPA group. There are two possible explanations for this finding. First, our mean pre-treatment HOMA-R was 1.6, which was near the upper limit for normal HOMA-R results in the Japanese population, and might preclude the detection of any significant change. Second, EPA treatment did not alter fasting glycemic control at 6 months. Because HOMA-R was calculated using fasting PG and insulin levels, EPA treatment did not significantly improve HOMA-R. Many human studies have also failed to demonstrate that n-3 PUFA exert a protective effect on insulin sensitivity [8, 9]. Thus, future studies should be warranted in selected study populations. Most trials have involved the use of diets supplemented by intake of fish, fish oil or capsule containing fish oil extracts. These contains a number of other fatty acids and different components. Thus, an evaluation of the specific effects of each n-3 PUFA was not possible. We adopted highly purified EPA ethyl ester, which was approved by Japan’s Ministry of Health, Labour and Welfare for the treatment of peripheral artery disease and hyperlipidemia, to examine its effects on postprandial glucose and insulin metabolism [11, 12]. Effects of EPA on atherogenic dyslipidemia and postprandial hypertriglyceridemia Despite best evidence-based statin therapy, it is clear that there persists unacceptably high residual cardiovascular events [26]. The importance of atherogenic dyslipidemia, defined as elevated TG-rich lipoproteins and their remnants and low HDL-C, has been recognized as a key driver of residual cardiovascular risk in individuals with metabolic disease, even if LDL cholesterol are well controlled [26, 27]. Type 2 DM commonly represents elevated TG, low HDL-C, and the predominance of small dense LDL particles due to insulin resistance [28]. Thereby, patients with DM are thought to belong to the highest cardiovascular disease risk category despite being treated by statin [1, 2, 29]. Several methods have been developed for the measurement of atherogenic small dense LDL levels, including ultracentrifugation, gradient gel electrophoresis, nuclear magnetic resonance spectroscopy and heparin magnesium precipitation methods [30, 31]. However, there is no standard procedure for clinical practice. Maruyama et al. demonstrated that the TG/HDL-C ratio was associated with LDL particle size [17]. Previous reports revealed that a high TG/HDL-C ratio was associated with cardiovascular disease and all-cause death among patients with DM or metabolic syndrome [17, 32, 33] and the amelioration of this abnormality was associated with improvements in endothelial dysfunction [34]. Thus, high TG/HDL-C ratio was regarded as an important residual cardiovascular risk and a therapeutic target for decreasing cardiovascular events in DM patients. Therefore, we used the TG/HDL-C ratio as the marker of small dense LDL particle in this study. Long-term ingestion of fish oils (which include n-3 PUFA) by normal volunteers and patients with hypertriglyceridemia dramatically reduced fasting TG levels and postprandial hyperlipidemia that were produced by fat loading [35, 36]. In this context, n-3 PUFA have been suggested to predominantly suppress both hepatic and intestinal ApoB secretion and synthesis. EPA treatment also reduced TG levels with minimal changes in HDL-C levels for both healthy volunteers and patients with CAD [11, 21–23] and significantly decreased postprandial TG elevation and concomitantly improved postprandial endothelial dysfunction among healthy individuals [35]. The present study also revealed that 6 months of EPA treatment significantly reduced postprandial hypertriglyceridemia (incremental TG peak and AUC-TG) and remnant-like particle cholesterol as well as fasting TG levels and TG/HDL-C ratio, and that these improvements were associated with improvements in endothelial dysfunction among patients with newly diagnosed IGM. Therefore, these results suggest that EPA has beneficial effects on postprandial hypertriglyceridemia and other atherogenic dyslipidemia, such as remnant-like particle cholesterol and small dense LDL particles, and may significantly contribute to the residual risk reduction of cardiovascular events among patients with IGM. Indeed, JELIS trials have demonstrated that EPA treatment (1800 mg/day) significantly reduced the risk of major coronary events during an average follow-up of 4.6 years, and sub-analyses revealed that EPA treatment was especially beneficial among patients with IGM or patients with high TG and low HDL-C levels [11, 37, 38]. Effects of EPA on FMD Endothelial dysfunction contributes to the onset and progression of atherosclerosis [39]. Furthermore, endothelial dysfunction is associated with many cardiovascular risk factors, such as dyslipidemia, hypertension, DM, and obesity. Moreover, postprandial hyperglycemia and hypertriglyceridemia cause temporal and persistent endothelial dysfunction, and are thought to be associated with increased risk of CAD, especially among pre-diabetic and diabetic patients [18, 35]. The pathophysiology of FMD, based on reactive hyperemia, has extensively been studied and a previous large-scale population-based study demonstrated that brachial FMD predicted the incidence of cardiovascular events among adults [20]. Therefore, we selected percent change in FMD as the surrogate marker for the risk of future cardiovascular events. EPA or re-composition of plasma membrane fatty acids following EPA administration (high EPA/AA ratio) directly improves endothelial dysfunction and provides a vasodilatory effect among patients with CAD and hyperlipidemia [40, 41]. Moreover, EPA upregulates endothelial nitric oxide synthase and augments nitric oxide production through the activation and translocation of endothelial nitric oxide synthase in endothelial cells [42]. In the present study, we measured plasma EPA levels and the EPA/AA ratio to monitor the bioavailability of EPA. There is accumulating evidence that blood EPA levels and the EPA/AA ratio predict the risk of cardiovascular events among the general population and patients with CAD [43]. It appears that EPA possesses a variety of biological actions including anti-inflammatory and anti-thrombotic effects, which may contribute to the prevention of cardiovascular events [22, 23, 44, 45]. Indeed, the present study also revealed improvements in inflammatory marker by using EPA and then improved endothelial dysfunction. Furthermore, based on EPA’s beneficial effects on metabolic derangements and endothelial protection, it appears that n-3 PUFA may reduce the risk of developing DM and/or cardiovascular events among patients with IGM. Effects on EPA on fasting glucose control Most observational studies have reported an association between glucose control and cardiovascular disease [46]. Furthermore, several randomized controlled studies of patients with type 2 DM have demonstrated that intensive glucose control was associated with a significant reduction in the rate of major cardiovascular events [47]. Therefore, glycemic control plays an important role in preventing CAD among patients with diabetes. However, the present study found no changes in the EPA group’s fasting PG, IRI, ad hemoglobin A1c levels after 6 months of treatment. In this context, the STOP-NIDDM trial reported that acarbose (an α-glucosidase inhibitor) reduced the risk of cardiovascular events among patients with IGT, although these patients did not experience any changes in their fasting PG [48]. The authors speculated that a reduction in postprandial hyperglycemia and TG levels might reduce oxidative stress and endothelial dysfunction, which would subsequently be associated with the prevention of cardiovascular events. Study limitations Several limitations in the present study warrant consideration. First, this study was open-label, single-blinded and number of participants was relatively small. Second, baseline PG level of the EPA group was significantly lower than that of the non-EPA group and multiple regression analysis for predicting incremental glucose peak improvement revealed lower baseline PG levels and EPA treatment. Whether the difference of baseline PG level between the two groups might influence the results of the present study was tested by additional statistical analysis. The improvement of incremental glucose and TG peak in the EPA group remained larger than that in the non-EPA group and the change of AUC-IRI/AUC-PG ratio in the EPA group was significantly higher than that in non-EPA group even among patients with similar baseline PG levels ≤110 mg/dL. Therefore, these results supported the improvements of postprandial hyperglycemia, hypertriglyceridemia and insulin secretory ability might be due to an effect of EPA. Third, cohort studies have reported that n-3 PUFA consumption was inversely associated with the incidence of DM in Asia, whereas it was positively associated in North America and Europe [4]. Thus, our results may not generalize to other geographical or racial populations. Fourth, docosahexaenoic acid has an in vitro agonistic effect on GPCR 40 that is stronger than the effect of EPA [24], and both EPA and docosahexaenoic acid have numerous distinct biological effects. Therefore, future studies are needed to explore the effects of docosahexaenoic acid on the incidences of DM and glucose homeostasis. Conclusions In conclusion, EPA corrected not only postprandial hypertriglyceridemia but also postprandial hyperglycemia and insulin secretion ability. This amelioration of metabolic abnormalities was associated with improvements in concomitant endothelial dysfunction among newly diagnosed IGM patients with CAD. Though lifestyle changes remain a cornerstone in all strategy for prevention of DM, this study suggests that addition of EPA could accelerate the prevention of DM and CAD. Additional files 10.1186/s12933-016-0437-y Multiple regression analysis for predicting incremental glucose peak improvement. 10.1186/s12933-016-0437-y Comparison of cookie meal test data between baseline and 6 months, and comparison of absolute change from baseline among patients with baseline plasma glucose <110 mg/dL. Abbreviations DMdiabetes mellitus CADcoronary artery disease PUFApolyunsaturated fatty acids EPAeicosapentaenoic acids IGMimpaired glucose metabolism PGplasma glucose TGtriglyceride OGTToral glucose tolerance test IGTimpaired glucose tolerance IRIimmune reactive insulin HDL-Chigh-density lipoprotein cholesterol LDLlow-density lipoprotein CMTcookie meal test AUCarea under the response curve HOMA-Rhomeostasis model assessment ratio FMDflow-mediated dilatation AAarachidonic acid GPCRG-protein coupled receptor Authors’ contributions All authors were involved in reporting the results of this study and all approved the final version of the submitted manuscript. TS, HT, MT, NH, TO and MY contributed in the conception, design and planning of the study. All authors were involved in acquisition of data and critical revision of the manuscript for important intellectual content. TS and HT did the statistical analysis. Manuscript writing: TS and MY. TS is responsible for the overall content and serves as guarantor. All authors read and approved the final manuscript. Acknowledgements Not applicable. Competing interests The authors declare that they have no competing interests. Funding There is no grant support for the present study. ==== Refs References 1. Mark KH Moliterno DJ Granger CB Miller DP White HD Wilcox RG Influence of diabetes mellitus on clinical outcome in the thrombolytic era of acute myocardial infarction. GUSTO-I Investigators. Global utilization of streptokinase and tissue plasminogen actibator for occluded coronary arteries J Am Coll Cardiol 1997 30 171 179 10.1016/S0735-1097(97)00118-6 9207639 2. Malmberg K Yusuf S Gerstein HC Brown J Zhao F Hunt D Impact of diabetes on long-term prognosis in patients with unstable angina and non-Q wave myocardial infarction: results of the OASIS (Organization to Assess Strategies for Ischemic Syndrome) Registry Circulation 2000 102 1014 1017 10.1161/01.CIR.102.9.1014 10961966 3. Nanri A Mizoue T Noda M Takahashi Y Matsushita Y Poudel-Tandukar K Japan Public Health Center-based Prospective Study Group Fish intake and type 2 diabetes in Japanese men and women: the Japan Public Health Center-based Prospective Study Am J Clin Nutr 2011 94 884 891 10.3945/ajcn.111.012252 21775559 4. Wallin A Di Giuseppe D Orsini N Patel PS Forouhi NG Wolk A Fish consumption, dietary long-chain n-3 fatty acids, and risk of type 2 diabetes: systematic review and meta-analysis of prospective studies Diabetes Care 2012 35 918 929 10.2337/dc11-1631 22442397 5. Wilkinson P Leach C Ah-Sing EE Hussain N Miller GJ Millward DJ Influence of alpha-linolenic acid and fish-oil on markers of cardiovascular risk in subjects with an atherogenic lipoprotein phenotype Atherosclerosis 2005 181 115 124 10.1016/j.atherosclerosis.2004.12.029 15939062 6. Kato T Shimano H Yamamoto T Ishikawa M Kumadaki S Matsuzaka T Palmitate impairs and eicosapentaenoate restores insulin secretion through regulation of SREBP-1c in pancreatic islet Diabetes 2008 57 2382 2392 10.2337/db06-1806 18458149 7. Bhaswant M Poudyal H Brown L Mechanisms of enhanced insulin secretion and sensitivity with n-3 unsatulated fatty acids J Nutr Biochem 2015 26 571 584 10.1016/j.jnutbio.2015.02.001 25841249 8. Flachs P Rossmeisl M Kopecky J The effect of n-3 fatty acids on glucose homeostasis and insulin sensitivity Physiol Res 2014 63 Suppl 1 S93 S118 24564669 9. Lalia AZ Johnson ML Jensen MD Hames KC Port JD Lanza IR Effects of dietary n-3 fatty acids on hepatic and peripheral insulin sensitivity in insulin-resistant humans Diabetes Care 2015 38 1228 1237 10.2337/dc14-3101 25852206 10. Sarbolouki S Javanbakht MH Derakhshanian H Hosseinzadeh P Zareei M Hashemi SB Eicosapentaenoic acid improves insulin sensitivity and blood sugar in overweight type 2 diabetes mellitus patients: a double-blind randomised clinical trial Singapore Med J 2013 54 387 390 10.11622/smedj.2013139 23900468 11. Yokoyama M Origasa H Matsuzaki M Matsuzawa Y Saito Y Ishikawa Y Japan EPA lipid intervention study (JELIS) investigators Effects of eicosapentaenoic acid on major coronary events in hypercholesterolaemic patients (JELIS): a randomised open-label, blinded endpoint analysis Lancet 2007 369 1090 1098 10.1016/S0140-6736(07)60527-3 17398308 12. Yokoyama M Origasa H JELIS Investigators Effects of eicosapentaenoic acid on cardiovascular events in Japanese patients with hypercholesterolemia: rational, design, and baseline characteristics of the Japan EPA Lipid Intervention Study (JELIS) Am Heart J 2003 146 613 620 10.1016/S0002-8703(03)00367-3 14564313 13. Itakura H Yokoyama M Matsuzaki M Saito Y Origasa H Ishikawa Y JELIS Investigators Relationships between plasma fatty acid composition and coronary artery disease J Atheroscler Thromb 2011 18 99 107 10.5551/jat.5876 21099130 14. Harano Y Miyawaki T Nabiki J Shibachi M Adachi T Ikeda M Development of cookie test for the simultaneous determination of glucose intolerance, hyperinsulinemia, insulin resistance and postprandial dyslipidemia Endocr J 2006 53 173 180 10.1507/endocrj.53.173 16618974 15. Gardner CD Fortmann SP Krauss RM Association of small low-density particles with the incidence of coronary artery disease in men and women JAMA 1996 276 875 881 10.1001/jama.1996.03540110029028 8782636 16. El Harchaoui K van der Steeg WA Stroes ES Kuivenhoven JA Otvos JD Wareham NJ Value of low-density lipoprotein particle number and size as predictors of coronary artery disease in apparently healthy men and women: the EPIC-Norfolk Prospective Population Study J Am Coll Cardiol 2007 49 547 553 10.1016/j.jacc.2006.09.043 17276177 17. Maruyama C Imamura K Teramoto T Assessment of LDL particle size by triglyceride/HDL-cholesterol ratio in non-diabetic, healthy subjects without prominent hyperlipidemia J Atheroscler Thromb 2003 10 186 191 10.5551/jat.10.186 14564088 18. Emoto T Sawada T Hashimoto M Kageyama H Terashita D Mizoguchi T Effect of 3-month repeated administration of miglitol on vascular endothelial function in patients with diabetes mellitus and coronary artery disease Am J Cardiol 2012 109 42 46 10.1016/j.amjcard.2011.08.005 21944671 19. Sawada T Emoto T Motoji Y Hashimoto M Kageyama H Terashita D Possible association between non-invasive parameter of flow-mediated dilatation in brachial artery and whole coronary plaque vulnerability in patients with coronary artery disease Int J Cardiol 2013 166 613 620 10.1016/j.ijcard.2011.11.101 22172589 20. Yeboah J Folsom AR Burke GL Johnson C Polak JF Post W Predictive value of brachial flow-mediated dilation for incident cardiovascular events in a population-based study: the multi-ethnic study of atherosclerosis Circulation 2009 120 502 509 10.1161/CIRCULATIONAHA.109.864801 19635967 21. Harris WS Park Y Isley WL Cardiovascular disease and long-chain omega-3 fatty acids Curr Opin Lipidol 2003 14 9 14 10.1097/00041433-200302000-00003 12544655 22. Saravanan P Davidson NC Schmidt EB Calder PC Cardiovascular effects of marine omega-3 fatty acids Lancet 2010 376 540 550 10.1016/S0140-6736(10)60445-X 20638121 23. De Caterina R n-3 fatty acids in cardiovascular disease N Engl J Med 2011 364 2439 2450 10.1056/NEJMra1008153 21696310 24. Itoh Y Kawamata Y Harada M Kobayashi M Fujii R Fukusumi S Free fatty acids regulate insulin secretion from pancreatic beta cells through GPR40 Nature 2003 422 173 176 10.1038/nature01478 12629551 25. Oh DY Talukdar S Bae EJ Imamura T Morinaga H Fan W GPR120 is an omega-3 fatty acid receptor mediating potent anti-inflammatory and insulin-sensitizing effects Cell 2010 142 687 698 10.1016/j.cell.2010.07.041 20813258 26. Fruchart JC Davignon J Hermans MP Al-Rubeaan K Amarenco P Assmann G Residual risk reduction initiative (R3i). Residual macrovascular risk in 2013: what have we learned? Cardiovasc Diabetol 2013 2014 13 26 27. Chapman MJ Ginsberg HN Amarenco P Andreotti F Boren J Catapano AL European Atherosclerosis Society Consensus panel Triglyceride-rich lipoprotein and hig-density lipoprotein cholesterol in patients at high risk of cardiovascular disease: evidence and guidance for management Eur Heart J 2011 32 1345 1361 10.1093/eurheartj/ehr112 21531743 28. Wu L Parhofer KG Diabetic dyslipidemia Metabolism 2014 63 1469 1479 10.1016/j.metabol.2014.08.010 25242435 29. Nesto RW Beyond low-density lipoprotein: addressing the atherogenic lipid triad in type 2 diabetes mellitus and the metabolic syndrome Am J Cardiovasc Drugs 2005 5 379 387 10.2165/00129784-200505060-00005 16259526 30. Chung M Lichtenstein AH Ip S Lau J Balk EM Comparability of methods for LDL subfraction determination: a systematic review Atherosclerosis 2009 205 342 348 10.1016/j.atherosclerosis.2008.12.011 19162266 31. Blake GJ Otvos JD Rifai N Ridker PM Low-density lipoprotein particle concentration and size as determined by nuclear magnetic resonance spectrosopy as predictors of cardiovascular disease in women Circulation 2002 106 1930 1937 10.1161/01.CIR.0000033222.75187.B9 12370215 32. Vega GL Barlow CE Grundy SM Leonard D DeFina LF Triglyceride-to-high-density-lipoprotein-cholesterol ratio is an index of heart disease mortality and of incidence of type 2 diabetes mellitus in men J Investig Med 2014 62 345 349 10.2310/JIM.0000000000000044 24402298 33. Sarwar N Danesh J Eiriksdottir G Sigurdsson G Wareham N Bingham S Triglycerides and the risk of coronary heart disease: 10,158 incident cases among 262,525 participants in 29 Western prospective studies Circulation 2007 115 450 458 10.1161/CIRCULATIONAHA.106.637793 17190864 34. Sawada T Shiotani H Terashita D Nagasawa Y Kim SS Koide M Comparison of effects of alpha-glucosidase inhibitors and glinide drugs on endothelial dysfunction in diabetic patients with coronary artery disease Circ J 2014 78 248 255 10.1253/circj.CJ-13-0918 24225338 35. Miyoshi T Noda Y Ohno Y Sugiyama H Oe H Nakamura K Omega-3 fatty acids improve postprandial lipemia and associated endothelial dysfunction in healthy individuals—a randomized cross-over trial Biomed Pharmacother 2014 68 1071 1077 10.1016/j.biopha.2014.10.008 25458786 36. Tinker LF Parks EJ Behr SR Schneeman BO Davis PA (n-3) fatty acid supplementation in moderately hypertriglyceridemic adults changes postprandial lipid and apolipoprotein B responses to a standardized test meal J Nutr 1999 129 1126 1134 10356076 37. Oikawa S Yokoyama M Origasa H Matsuzaki M Matsuzawa Y Saito Y JELIS Investigators Japan Suppressive effect of EPA on the incidence of coronary events in hypercholesterolemia with impaired glucose metabolism: sub-analysis of the Japan EPA Lipid Intervention Study (JELIS) Atherosclerosis 2009 206 535 539 10.1016/j.atherosclerosis.2009.03.029 19447387 38. Saito Y Yokoyama M Origasa H Matsuzaki M Matsuzawa Y Ishikawa Y JELIS Investigators Japan Effects of EPA on coronary artery disease in hypercholesterolemic patients with multiple risk factors: sub-analysis of primary prevention cases from the Japan EPA Lipid Intervention Study (JELIS) Atherosclerosis 2008 200 135 140 10.1016/j.atherosclerosis.2008.06.003 18667204 39. Heitzer T Schlinzig T Krohn K Meinertz T Munzel T Endothelial dysfunction, oxidative stress, and risk of cardiovascular events in patients with coronary artery disease Circulation 2001 104 2673 2678 10.1161/hc4601.099485 11723017 40. Tagawa H Shimokawa H Tagawa T Kuroiwa-Matsumoto M Hirooka Y Takeshita A Long-term treatment with eicosapentaenoic acid augments both nitric oxide-mediated and non-nitric oxide-mediated endothelium-dependent forearm vasodilatation in patients with coronary artery disease J Cardiovasc Pharmacol 1999 33 633 640 10.1097/00005344-199904000-00017 10218735 41. Yamakawa K Shimabukuro M Higa N Asahi T Ohba K Arasaki O Eicosapentaenoic acid supplementation changes fatty acid composition and corrects endothelial dysfunction in hyperlipidemic patients Cardiol Res Pract 2012 2012 754181 23326753 42. Li Q Zhang Q Wang M Zhao S Ma J Luo N Eicosapentaenoic acid modifies lipid composition in caveolae and induces translocation of endothelial nitric oxide synthase Biochimie 2007 89 169 177 10.1016/j.biochi.2006.10.009 17125900 43. Ninomiya T Nagata M Hata J Hirakawa Y Ozawa M Yoshida D Association between ratio of serum eicosapentaenoic acid to arachidonic acid and risk of cardiovascular disease: the Hisayama Study Atherosclerosis 2013 231 261 267 10.1016/j.atherosclerosis.2013.09.023 24267237 44. Doi M Nosaka K Miyoshi T Iwamoto M Kajiya M Okawa K Early eicosapentaenoic acid treatment after percutaneous coronary intervention reduces acute inflammatory responses and ventricular arrythmias in patients with acute myocardial infartion: a randomized, controlled study Int J Cardiol 2014 176 577 582 10.1016/j.ijcard.2014.08.055 25305703 45. Micallef MA Munro IA Garg ML An inverce relationship between plasma n-3 fatty acids and C-reactive protein in healthy individuals Eur J Clin Nutr 2009 63 1154 1156 10.1038/ejcn.2009.20 19352379 46. Selvin E Marinopoulos S Berkenblit G Rami T Brancati FL Powe NR Meta-analysis: glycosylated hemoglobin and cardiovascular disease in diabetes mellitus Ann Intern Med 2004 141 421 431 10.7326/0003-4819-141-6-200409210-00007 15381515 47. Hayward RA Reaven PD Wiitala WL Bahn GD Reda DJ Ge L VADT Investigators Follow-up of glycemic control and cardiovascular outcomes in type 2 diabetes N Engl J Med 2015 372 2197 2206 10.1056/NEJMoa1414266 26039600 48. Chiasson JL Josse RG Gomis R Hanefeld M Karasik A Laakso M STOP-NIDDM Trial Group Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial JAMA 2003 290 486 494 10.1001/jama.290.4.486 12876091

PMC005xxxxxx/PMC5002117.txt

==== Front BMC Health Serv ResBMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 167110.1186/s12913-016-1671-4Research ArticleBudget impact of secondary hyperparathyroidism treatment in chronic kidney disease in an Ecuadorian social security hospital Manjarres Luis lmanjarresb@iess.gob.ec 1Sanchez Pilar pilitadoc@yahoo.com 1Cabezas María C. maria.cabezas@hrservicesec.com 24Fornasini Marco mfornasini@udla.edu.ec 3Freire Valeria fm216@hotmail.com 2Albert Adelin aalbert@ulg.ac.be 51 Nephrology Service, Carlos Andrade Marin Hospital, Quito, Ecuador 2 Medical School, Pontifical Catholic University of Ecuador, Quito, Ecuador 3 Translational Research Center, Universidad de las Américas (UDLA), Quito, Ecuador 4 Health & Research Services, Quito, Ecuador 5 Public Health Department, University of Liège, Liège, Belgium 26 8 2016 26 8 2016 2016 16 1 4432 4 2016 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 Chronic kidney disease (CKD) is a disorder with high morbidity and mortality worldwide whose complications generate multiple costs. In Ecuador, only a few healthcare institutions have implemented management protocols aimed to reduce costs and to improve the quality of life of patients. The aim of this study is to evaluate the short-term (1-year) and long-term (5-year) costs and savings in the management of secondary hyperparathyroidism (SHPT) of hemodialyzed CKD patients by comparing calcitriol and paricalcitol in a large social security hospital in Quito, Ecuador. Methods The estimation model assessed the resources used in the management of SHPT by comparing prospectively the cost savings within 1-year and 5-year time horizon with calcitriol and paricalcitol. Hospitalization, erythropoietin (EPO), treatment doses, intravenous iron consumption, and medical supplies were estimated according international references, based on the initial parathormone level (iPTH) of patients. The Ecuadorian National Reference costs (2014–2015) and institutional costs were used to calculate treatment costs. A statistical sensitivity analysis was also performed. Results The study was based on data from 354 patients of whom 147 (41.4 %) had a value of iPTH in the range 300–600 pg/ml, 45 (12.8 %) in the range 601–800 pg/ml, and 162 (45.7 %) over 800 pg/ml. The 1-year estimated costs per patient for calcitriol and paricalcitol, respectively, were: medication, 63.88 USD and 1,123.44 USD; EPO, 19,522.95 USD and 16,478 USD; intravenous iron 143.21 USD and 187.76 USD. Yearly hospitalization costs per patient were 11,647.99 USD with calcitriol and 8,019.41 USD with paricalcitol. Total yearly costs per patient amounted to 31,378.02 USD with calcitriol and 25,809.50 USD with paricalcitol. Total savings using paricalcitol were 5,568.52 USD per patient compared with calcitriol. The 5-year cumulative medication costs were 319 USD for calcitriol and 2,403 USD for paricalcitol; EPO with calcitriol was 97,615 USD and with paricalcitol 82,394 USD; intravenous iron with calcitriol was 716 USD and paricalcitol 939 USD. Hospitalization costs for patients with calcitriol and paricalcitol were 43,095 USD and 62,595 USD, respectively. Total savings using paricalcitol amounted 32,414 USD per patient compared with calcitriol. Conclusions Paricalcitol use generated more cost savings than calcitriol after 1 and 5 years. Keywords Renal kidney diseaseCalcitriolParicalcitolAbbvie Ecuadorissue-copyright-statement© The Author(s) 2016 ==== Body Background Chronic kidney disease (CKD) is a clinical condition, which affects approximately 10 % of the adult population with high morbidity and mortality worldwide [1, 2]. According to the Vivenkanad study in 2013, the overall prevalence of CKD ranged between 8 and 16 %. The same study reported that in Latin America, the prevalence increased by about 7 % during the last few years [3]. The increasing frequency of type II diabetes and hypertension have greatly contributed to increased prevalence of CKD [4, 5]. A common complication of CKD is secondary hyperparathyroidism (SHPT), which generates high treatment costs. SHPT increases the frequency of skeletal and cardiovascular disorders [6]. SHPT is also known to worsen renal-driven anemia because higher levels of circulating parathyroid hormone is associated with a lower response to recombinant human erythropoietin [7]. Furthermore, SHPT is associated with increased hospitalization rates, especially when parathormone (PTH) levels exceed 600 pg/ml [8]. Thus, SHPT generates high costs to healthcare systems [9]. An Argentinian study published in 2013 reported that among 1210 CKD hemodialyzed patients, 26.7 % had an initial parathyroid hormone (iPTH) > 300 pg/ml [3]. In Ecuador, the total number of SHTP cases is not known; however, several studies based on relatively small samples have exhibited differing prevalence levels. As an illustration, a study (88 patients) in Guayaquil, Ecuador reported 3.4 % of SPTH cases in dialysis patients in 2009 [10], while another study (2012), in Ambato, Ecuador, on 54 hemodialyzed patients with CKD reported 32 % of the population with values > 450 pg/ml of iPTH [11]. The cost of care for CKD is high, especially in end stage renal disease (ESRD). In England (2012), the cost of CKD in 2009–2010 was estimated at £1.44 to £1.45 billion [12, 13]. Patients with ESRD consumed important economic resources from healthcare systems, with an estimated annual cost of 41,341.05 USD [14, 15]. Moreover, costs of CKD treatment have been increasing steadily; for example in the United States, they grew by 57 % between 1999 and 2004 [16]. In Latin American countries, the costs range from 10,956 USD to 14,654 USD per year per patient [17]. Calcitriol has been considered to be the first choice drug for treating SHPT in ESRD [18]. It is effective and inexpensive, although its use is frequently limited for patients with either hypercalcemia or hypophosphatemia. Other drugs like paricalcitol have been developed to treat these conditions [18]. Unlike calcitriol, paricalcitol has minimal impact on serum calcium and phosphorus [9]. However, paricalcitol is much more expensive than calcitriol. In Colombia (2016), the price of one vial of paricalcitol (5 μg) is 79,400 Colombian pesos (26.46 USD), while 0.25 μg of calcitriol costs 5,018 Colombian pesos (1.67 USD), reflecting the higher cost of paricalcitol, according to the Colombian Ministry of Health [19]. In Ecuador (2016), the price of paricalcitol is greater than calcitriol according the price of one vial of paricalcitol is 31.25 USD, while 0.25 μg of calcitriol costs 0.21 USD according the Ecuadorian Social Security Institute [20]. The aim of this comparative study was to evaluate the budget impact associated with calcitriol and paricalcitol in the management of secondary hyperparathyroidism (SHPT) in hemodialyzed patients with chronic kidney disease (CKD) 1-year and 5-year time horizons. Methods The budget impact model was developed in Microsoft Excel to calculate the costs of management of hemodialyzed patients hospitalized with SHPT due to chronic kidney disease in a social security hospital in Quito, Ecuador. The study included data of patients from the nephrology department with a diagnosis of SHPT, who were under dialysis treatment between September 2013 and October 2014. Eligible patients were identified using code E21.1 of the International Classification of Diseases (ICD-10) for SHPT. The only inclusion criterion was that the patient had been under SHPT dialysis treatment with either paricalcitol or calcitriol. Data collection consisted of a retrospective chart review of the institutional database. Serum iPTH was recorded during the previous 12 months in patients with calcitriol and paricalcitol treatment. The study developed a model to estimate the annual cost of paricalcitol and calcitriol treatment. The analysis used international studies published to calculate the costs of management and treatment. We implemented an estimation to calculate the impact budget over 1 year and 5-year periods. Literature review A systematic search was conducted in PubMed and Medline database in order to identify all randomized clinical trials of SHPT treatment, particularly those that compared paricalcitol and calcitriol treatments with respect to efficacy and which were conducted between January 2004 and December 2013. Trial selection criteria were based on study variables such as hospitalization days, iron, and erythropoietin consumption. Paricalcitol and calcitriol doses were determined from international effectiveness trials and guidelines (see Fig. 1).Fig. 1 Flowchart for study selection The study of Naves Diaz was the reference to classify the initial PTH level (iPTH). The study classified patient iPTH level in three groups (group I: 300–600; group II: 601–800; group III: >800 pg/ml) in order to determine the amount of paricalcitol required for SHPT treatment [21]. Llach study provided the paricalcitol dose reference for group I. For groups II and III, doses were calculated based on a simple linear extrapolation [22]. The National Kidney Foundation Kidney Disease Outcomes Quality Initiative (NKF KDOQI) guidelines were the reference to determine the amount of calcitriol needed to treat SHPT. The guidelines recommend 0.25 μg of calcitriol per day [23]. The Capuano study was used as the reference to determine erythropoietin and intravenous iron yearly costs. This study showed a significant difference in the use of erythropoietin when using paricalcitol versus calcitriol [24]. As shown in Table 1, there are statistically significant differences in the weekly consumption of erythropoietin and intravenous iron in patients treated with paricalcitol and calcitriol. The calculations in the study used the annual consumption for each therapy (Table 1).Table 1 Annual consumption of erythropoietin and intravenous iron according Paricalcitol and Calcitriol therapies Parameter Paricalcitol Calcitriol P-value Erythropoietin Weekly Average Consumption 11,758 UI 13,930 UI <0.05 Intravenous Iron Weekly Average Consumption 59 mg 45 mg <0.05 Source: Capuano, A., 2009 Hospitalization were calculated based on the Dobrez study based on a sample of 11,443 patients, of whom 4,611 received treatment with paricalcitol and 6,832 with calcitriol [25]. The model used the averages of patients’ hospitalized rates and hospitalization days for each SHPT therapy (See Table 2).Table 2 Hospitalization rates in patients with Paricalcitol and Calcitriol Parameter Paricalcitol Calcitriol Hospitalized patients per year per each 100 patients in treatment 59.6 75.2 Average number of hospitalization days per year 17.2 19.8 Source: Dobrez, DG., 2004 Cost evaluation The cost estimation included the resources used in the management of SHPT, such as erythropoietin and intravenous iron. Other variables considered in the model were erythropoietin consumption, intravenous iron consumption, hospitalization days, and medical supplies costs that were calculated based on international references [21–25]. Treatments costs were derived from “The Ecuadorian National Reference Costs”, published in December 2014 (see Table 3). Inflation was estimated from International Monetary Fund data.Table 3 Unit cost according the National Reference Category USD Paricalcitol 1 μg 1.25 Calcitriol 1 μg 0.84 Hospitalization Day 116.76 Erythropoietin 1 UI 0.03 Iron 1 mg 0.06 The paricalcitol price was established using the databases from the institution. The study calculated the cost of 1 μg for use in the model. To calculate the annual cost of paricalcitol for each iPTH group, the model used the following formula: ParicalcitolannualcostperPTHgroup=ParicalcitolmicrogramsrequiredperyearperPTHgroup×Paricalcitolmicrogramsprice The average paricalcitol cost per year was used to estimate the cost per patient treatment using the following formula: Paricalcitolcostperpatient=ΣParicalcitolannualcostsperPTHgroup×%patientsperPTHgroup For calcitriol, the model considered the different presentations and the institutional price to estimate the drug cost per microgram. Annual costs of calcitriol treatment per patient in the model were calculated using the following formula: Calcitriolannualcostperpatient=Calciotrioldose0,25μg×μgCalcitriolPrice×365days Erythropoietin was calculated based on 1 International Unit (IU) of cost according the institutional cost of erythropoietin. The annual cost of erythropoietin was calculated by multiplying the mean unit cost by the weekly consumption of erythropoietin and then multiplying by 52 weeks. The cost of annual intravenous iron was calculated based on its institutional cost multiplied by weekly consumption and then multiplied by 52 weeks. Hospitalization cost per patient was based on the price of 1 day of hospitalization according the “Ecuadorian National Reference Costs” including estimate the annual inflation. Calculations were based on a simple linear extrapolation according the iPTH level and the literature review. Subjects with missing data were not included in the analysis. A one-way sensitivity analysis was conducted by varying the following parameters by ± 10 %, hospitalization costs per day, average erythropoietin costs, average intravenous iron costs, treatment iPTH, hospitalization rate reduction, annual hospitalization reduction, erythropoietin weekly savings costs of paricalcitol and calcitriol, and the number of hemodialysis per month; an inflation rate by ± 2.5 % was included. Results The study was based on data from 354 SHPT patients whose iPTH levels were registered. The average estimates presented in Table 4 were calculated according to the Naves-Diaz study, divided into three iPTH l groups achieving, respectively, 435 pg/ml in the first group, 668 pg/ml in the second group, and 1236 pg/ml in the third group.Table 4 Initial PTH levels (n = 354 patients) Basal PTH groupa No. of patients PTH mean level (pg/ml) PTH serum levels 300-600 pg/ml 118 435 601-800 pg/ml 67 668 >800 pg/ml 169 1, 236 aBasal PTH ranges were according Naves-Diaz Study [21] Secondary hyperthyroidism costs The projected cost per patient to treat SHPT using paricalcitol across time was 1,123.44 USD the first year and 2,403 USD the fifth year. The cost for calcitriol treatment was 63.88 USD the first year, and 319 USD the fifth year, as shown in Table 5.Table 5 Cumulative costs per patient short and long term in USD 1st Year 5th Year Paricalcitol Calcitriol Paricalcitol Calcitriol (A) (B) (A) (B) SHPT treatment 1,123.44 63.88 2,403 319 Erythropoietin 16,478 19,522.95 82,394 97,615 Intravenous iron 187.76 143.21 939 716 Hospitalization 8,019.41 11,647.99 43,095 62,595 Total 25,809.50 31,378.02 128,831 161,245 Hence, treatment with paricalcitol represented costs reductions of 5,568.52 USD the first year and 32,414.00 USD the fifth year (Fig. 2).Fig. 2 Cumulative savings per patient short and long term in US dollars Sensitivity analyses The results of the one-way sensitivity analyses are shown in Table 6. The present model yielded a decrease around 18 % of costs in all parameters. Sensitivity analysis shown that the average treatment of SHPT with paricalcitol was less expensive than that of calcitriol.Table 6 Cumulative Net Savings in 5 years 5-year savings Parameter Minimum savings (USD) Maximum savings (USD) Initial value (USD) Daily hospitalization costs (±10 %) 52,427 80,178 66,303 Erythropoietin costs (±10 %) 61,389 72,683 66,303 Intravenous iron costs (±10 %) 66,214 66,391 66,303 Annual inflation 2014–2016 (±2.5 %) 56,397 76,876 66,303 Treatment costs per iPTH group (I: 29 %, II: 17 %, III: 42 %) 66,303 89,994 66,303 Hemodialysis per month (±10 %) 79.66 52,983 66,303 Weekly savings erythropoietin administration (±10 %) 60,164 72,435 66,303 Hospitalization rate reduction (±10 %) 58,276 74,329 66,303 Annual hospitalization reduction (±10 %) 61,667 70,938 66,303 Discussion This study shows that paricalcitol was the best economic option to treat patients with SHPT, saving 2,679.31 USD in 1 year and 16,249 USD in 5 years per patient treated. Treatment costs with paricalcitol were 18,799 USD for the short term (1-year) and 91,744 USD for the long term (5-year). Despite its higher costs compared with calcitriol, paricalcitol showed higher effectiveness and benefits. Patients treated with paricalcitol presented fewer hospitalizations (59.6 %), generating important savings in direct costs in the management of patients with SHPT. According to the study of Nuijten et al. (2009) conducted in the United States, treatment with paricalcitol represented annual savings of 1,941 USD compared to calcitriol [26]. This results of that study are similar to those presented here, where paricalcitol savings per patient/year were 2,679.31 USD in the short term (1-year). The study published by Lorenzoni et al. (2014) demonstrated that paricalcitol used during the pre-dialysis stage and at early stage of SHPT in Italy represented an overall reduction in direct medical costs from 1′782,921 to 1′622,357 Euro for the healthcare system. The savings in a hypothetical cohort of 1,000 patients would amount to 1′971,500 Euro, which is less than the savings per 1,000 patients of 16′249,000 USD calculated in our model [27]. However, Lorenzoni study was implemented in pre-dialysis stage. Another study, conducted in Mexico by Sanchez-Casillas et al. (2013), reported a total long term (5-year) cost of 24,532.88 USD using paricalcitol therapy compared to 35,633.36 USD using calcitriol therapy, which reflects greater costs for both therapies compared to those calculated in our study (91,744 USD vs 107,933 USD) [28]. Nevertheless, the Mexican study also found cost savings of paricalcitol compared with calcitriol in SHPT. The work of Sprague et al. (2003) in the United States demonstrated that paricalcitol is more effective over PTH levels in SHPT patients, achieving a reduction of iPTH levels over 50 % after 18 weeks of treatment with less sustained hypercalcemia [29]. The control of serum calcium in SHPT patients is important for avoiding cardiovascular complications and parathyroid hyperplasia, thus leading to fewer hospitalizations, as described in the 2006 Cheng et al. review [30]. A study by Rosery et al. (2006) conducted in 6376 patients indicated that paricalcitol treatment resulted in a reduction of 84 % of hospitalizations in 1 year compared to calcitriol treatment, producing savings of 5,394 USD associated with paricalcitol treatment [31]. In sum, these studies demonstrated the cost savings of paricalcitol consistent with the present study. In our model, erythropoietin was the most expensive medication used in the management of SHPT patients. Capuano et al. described the erythropoietin international units recommended for patients treated with paricalcitol compared with calcitriol. They showed that there is evidence of a need of lower amount of erythropoietin with paricalcitol therapy [24]. Afsar et al. and Riccio et al. (2015) showed that paricalcitol increased hemoglobin levels, decreased urinary protein excretion, and generated lower resistance to erythropoietin treatment; in addition, it did not interfere with erythropoietin synthesis [32, 33]. The present study has some limitations, most importantly, that the study population was limited to the main social security general hospital in Quito, Ecuador. Another limitation is that the study was not able to compare the cost of paricalcitol with calcinanet, which is not available in the Ecuadorian market. In addition, our model did not take into account the CKD stage of disease only the iPTH. Finally, the model was developed using economic simulations based on international parameters, which might not be applicable in Ecuador. Conclusions In conclusion, the average annual cost per patient for paricalcitol therapy was 18,799, USD of which 80 % represented hospitalization and erythropoietin consumption. The use of paricalcitol, according to PTH initial level, represented substantial economic savings of 2,679 USD in the first year and 16,249 USD in the fifth year, when compared to calcitriol. The present study supports the use of paricalcitol as the first choice drug to treat SHPT in patients with chronic renal insufficiency. Abbreviations CKDChronic kidney disease EPOErythropoietin ESRDEnd stage renal disease ICD-10International classification of diseases iPTHInitial parathormone IUInternational Unit NKF KDOQINational kidney foundation kidney disease outcomes quality initiative (NKF KDOQI) PTHParathormone SHPTSecondary hyperparathyroidism (SHPT) We wish to thank Professor Will Watters and MS. Jennifer Mullin for their assistance with the critical review of the document. Funding The project was supported by Abbvie Ecuador laboratories and paper publication was sponsored by Pontifical Catholic University of Ecuador and Universidad de las Américas. Availability of data and materials Erythropoietin consumption, intravenous iron consumption, hospitalization days, and medical supplies costs were calculated based on international references included in the present manuscript. Treatments costs were derived from “The Ecuadorian National Reference Costs” and inflation was estimated from data of the International Monetary Fund. Authors’ contributions LM carried out and supervised the design, implementation, and acquisition of data in the study. PS was involved in revising the manuscript for important intellectual content. MC carried out the design, implementation, and acquisition of data used in the study. MF was involved in revising the manuscript for important intellectual content. VF was involved in the project implementation and acquisition of data. AA was involved in revising the manuscript critically for important intellectual content. All authors read and approved the final manuscript. Authors’ information Luis Manjarres is an MD with a specialization in nephrology and is chief of nephrology service at Carlos Andrade Marin Hospital in Quito, Ecuador. Pilar Sanchez is an MD with a specialization in nephrology. She is the chief of the dialysis service at Carlos Andrade Marin Hospital in Quito, Ecuador. Maria de Carmen Cabezas is MD and PhD. She is a researcher at the School of Medicine, Pontifical Catholic University of Ecuador in Quito, Ecuador. Marco Fornasini is MD and PhD. He is a full time professor and researcher at the Universidad de las Americas. Quito, Ecuador; his expertise is in the epidemiology of chronic diseases and infectious diseases. Valeria Freire is MD (c) at the School of Medicine of the Pontifical Catholic University of Ecuador. Quito, Ecuador. Adelin Albert, PhD, is professor emeritus of medical informatics and biostatistics, Department of Public Health, University of Liège, Liège, Belgium. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. The study was an economic simulation. Ethics approval and consent to participate Not applicable. The study was an economic simulation. ==== Refs References 1. Braun L Sood V Hogue S Lieberman B Copley-Merriman C High burden and unmet patient needs in chronic kidney disease Int J Nephrol Renovasc Dis 2012 5 151 163 23293534 2. Stengel B Combe C Jacquelinet C Briancon S Fouque D Laville M Frimat L Pascal C Herpe YE Deleuze JF The French Chronic Kidney Disease-Renal Epidemiology and Information Network (CKD-REIN) cohort study Nephrol Dial Transplant 2014 29 8 1500 1507 10.1093/ndt/gft388 24064325 3. Jha V Garcia-Garcia G Iseki K Li Z Naicker S Plattner B Saran R Wang AY Yang CW Chronic kidney disease: global dimension and perspectives Lancet 2013 382 9888 260 272 10.1016/S0140-6736(13)60687-X 23727169 4. Cusumano AM Gonzalez Bedat MC Chronic kidney disease in Latin America: time to improve screening and detection Clin J Am Soc Nephrol 2008 3 2 594 600 10.2215/CJN.03420807 18272831 5. Singh AK Farag YM Mittal BV Subramanian KK Reddy SR Acharya VN Almeida AF Channakeshavamurthy A Ballal HS PG Epidemiology and risk factors of chronic kidney disease in India–results from the SEEK (Screening and Early Evaluation of Kidney Disease) study BMC Nephrol 2013 14 114 10.1186/1471-2369-14-114 23714169 6. Thomas R Kanso A Sedor JR Chronic kidney disease and its complications Prim Care 2008 35 2 329 344 10.1016/j.pop.2008.01.008 18486718 7. Drueke TB Eckardt KU Role of secondary hyperparathyroidism in erythropoietin resistance of chronic renal failure patients Nephrol Dial Transplant 2002 17 Suppl 5 28 31 10.1093/ndt/17.suppl_5.28 12091604 8. Block GA Klassen PS Lazarus JM Ofsthun N Lowrie EG Chertow GM Mineral metabolism, mortality, and morbidity in maintenance hemodialysis J Am Soc Nephrol 2004 15 8 2208 2218 10.1097/01.ASN.0000133041.27682.A2 15284307 9. Joy MS Karagiannis PC Peyerl FW Outcomes of secondary hyperparathyroidism in chronic kidney disease and the direct costs of treatment J Manag Care Pharm 2007 13 5 397 411 17605511 10. Ramirez J PREVALENCIA DE HIPERPARATIROIDISMO SECUNDARIO EN PACIENTES SOMETIDOS A HEMODIÁLISIS” CENTRO DE DIALISIS “DR. CARLOS ELIZALDE MONTEVERDE” 2010 Guayaquil UNIVERSIDAD DE GUAYAQUIL 11. Tapia M EFECTOS COLATERALES POR LA DISMINUCIÓN DEL CALCIO EN PACIENTES CON INSUFICIENCIA RENAL CRÓNICA ATENDIDOS EN LA UNIDAD RENAL CONTIGO DE LA CIUDAD DE LATACUNGA EN EL PERIODO OCTUBRE 2011–FEBRERO 2012 2012 Ambato Universidad Técnica de Ambato 12. St Peter WL Khan SS Ebben JP Pereira BJ Collins AJ Chronic kidney disease: the distribution of health care dollars Kidney Int 2004 66 1 313 321 10.1111/j.1523-1755.2004.00733.x 15200439 13. Kerr M Bray B Medcalf J O’Donoghue DJ Matthews B Estimating the financial cost of chronic kidney disease to the NHS in England Nephrol Dial Transplant 2012 27 3 iii73 80 22815543 14. Zhang QL Rothenbacher D Prevalence of chronic kidney disease in population-based studies: systematic review BMC Public Health 2008 8 117 10.1186/1471-2458-8-117 18405348 15. Rodriguez-Iturbe B Bellorin-Font E End-stage renal disease prevention strategies in Latin America Kidney Int Suppl 2005 98 S30 36 10.1111/j.1523-1755.2005.09806.x 16108968 16. Foley RN Collins AJ End-stage renal disease in the United States: an update from the United States Renal Data System J Am Soc Nephrol 2007 18 10 2644 2648 10.1681/ASN.2007020220 17656472 17. Pecoits-Filho R Campos C Cerdas-Calderon M Fortes P Jarpa C Just P Luconi P Lugon JR Pacheco A Paniagua R Policies and health care financing issues for dialysis in Latin America: extracts from the roundtable discussion on the economics of dialysis and chronic kidney disease Perit Dial Int 2009 29 Suppl 2 S222 226 19270223 18. Bolasco P Treatment options of secondary hyperparathyroidism (SHPT) in patients with chronic kidney disease stages 3 and 4: an historic review Clin Cases Miner Bone Metab 2009 6 3 210 219 22461248 19. Ministerio de Salud y Protección Social Social SIdIdlP SISPRO Aplicativos Misionales 2016 Colombia Ministerio de Salud y Protección Social 20. Ecuador GNdlRd Sistema Oficial de Contratación Pública 2016 Ecuador Ecuador GNdlRd 21. Naves-Diaz M Passlick-Deetjen J Guinsburg A Marelli C Fernandez-Martin JL Rodriguez-Puyol D Cannata-Andia JB Calcium, phosphorus, PTH and death rates in a large sample of dialysis patients from Latin America. The CORES Study Nephrol Dial Transplant 2011 26 6 1938 1947 10.1093/ndt/gfq304 20513773 22. Llach F Yudd M Paricalcitol in dialysis patients with calcitriol-resistant secondary hyperparathyroidism Am J Kidney Dis 2001 38 5 Suppl 5 S45 50 10.1053/ajkd.2001.28114 11689387 23. National Kidney F K/DOQI clinical practice guidelines for bone metabolism and disease in chronic kidney disease Am J Kidney Dis 2003 42 4 Suppl 3 S1 201 14520607 24. Capuano A Serio V Pota A Memoli B Andreucci VE Beneficial effects of better control of secondary hyperparathyroidism with paricalcitol in chronic dialysis patients J Nephrol 2009 22 1 59 68 19229819 25. Dobrez DG Mathes A Amdahl M Marx SE Melnick JZ Sprague SM Paricalcitol-treated patients experience improved hospitalization outcomes compared with calcitriol-treated patients in real-world clinical settings Nephrol Dial Transplant 2004 19 5 1174 1181 10.1093/ndt/gfh123 15004264 26. Nuijten M Andress DL Marx SE Sterz R Chronic kidney disease Markov model comparing paricalcitol to calcitriol for secondary hyperparathyroidism: a US perspective Curr Med Res Opin 2009 25 5 1221 1234 10.1185/03007990902844097 19335321 27. Lorenzoni V Pierotti F Turchetti G A Budget Impact Analysis (Bia) Of The Use Of Paricalcitol For The Treatment Of Secondary Hyperparathyroidism (Shpt) In End Stage Renal Disease Patients Value Health 2014 17 7 A467 10.1016/j.jval.2014.08.1313 27201327 28. Sanchez-Casillas JL Ramirez-Lopez-De-N MG Long-Term Cost Comparison Between Paricalcitol and Calcitriol for the Treatment of Secondary Hyperparathyroidism in Chronic Kidney Disease in Mexico Value Health 2013 16 7 A688 10.1016/j.jval.2013.08.2048 29. Sprague SM Llach F Amdahl M Taccetta C Batlle D Paricalcitol versus calcitriol in the treatment of secondary hyperparathyroidism Kidney Int 2003 63 4 1483 1490 10.1046/j.1523-1755.2003.00878.x 12631365 30. Cheng S Coyne D Oral paricalcitol for the treatment of secondary hyperparathyroidism in chronic kidney disease Ther Clin Risk Manag 2006 2 3 297 301 10.2147/tcrm.2006.2.3.297 18360604 31. Rosery H Bergemann R Marx SE Boehnke A Melnick J Sterz R Williams L Health-economic comparison of paricalcitol, calcitriol and alfacalcidol for the treatment of secondary hyperparathyroidism during haemodialysis Clin Drug Investig 2006 26 11 629 638 10.2165/00044011-200626110-00002 17163297 32. Riccio E Sabbatini M Bruzzese D Capuano I Migliaccio S Andreucci M Pisani A Effect of paricalcitol vs calcitriol on hemoglobin levels in chronic kidney disease patients: a randomized trial PLoS One 2015 10 3 10.1371/journal.pone.0118174 25781618 33. Afsar B Agca E Turk S Comparison of erythropoietin resistance in hemodialysis patients using calcitriol, cinacalcet, or paricalcitol J Clin Pharmacol 2015 55 11 1280 1285 10.1002/jcph.556 26032009

PMC005xxxxxx/PMC5002118.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 68610.1186/s12974-016-0686-4ResearchSiponimod (BAF312) prevents synaptic neurodegeneration in experimental multiple sclerosis Gentile Antonietta 12Musella Alessandra 1Bullitta Silvia 1Fresegna Diego 12De Vito Francesca 12Fantozzi Roberta 3Piras Eleonora 4Gargano Francesca 4Borsellino Giovanna 4Battistini Luca 4Schubart Anna 5Mandolesi Georgia +39-06-50170-3211g.mandolesi@hsantalucia.it 1Centonze Diego 231 Laboratory of Neuroimmunology and Synaptic Transmission, IRCCS Fondazione Santa Lucia, Centro Europeo di Ricerca sul Cervello (CERC), 00143 Rome, Italy 2 Multiple Sclerosis Research Unit, Department of Systems Medicine, Tor Vergata University, 00133 Rome, Italy 3 Unit of Neurology and Neurorehabilitation, IRCCS Istituto Neurologico Mediterraneo (INM) Neuromed, 86077 Pozzilli, IS Italy 4 Neuroimmunology Unit, IRCCS Fondazione Santa Lucia-CERC, 00143 Rome, Italy 5 Novartis Institutes of Biomedical Research, Basel, Switzerland 26 8 2016 26 8 2016 2016 13 1 2071 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 Data from multiple sclerosis (MS) and the MS rodent model, experimental autoimmune encephalomyelitis (EAE), highlighted an inflammation-dependent synaptopathy at the basis of the neurodegenerative damage causing irreversible disability in these disorders. This synaptopathy is characterized by an imbalance between glutamatergic and GABAergic transmission and has been proposed to be a potential therapeutic target. Siponimod (BAF312), a selective sphingosine 1-phosphate1,5 receptor modulator, is currently under investigation in a clinical trial in secondary progressive MS patients. We investigated whether siponimod, in addition to its peripheral immune modulation, may exert direct neuroprotective effects in the central nervous system (CNS) of mice with chronic progressive EAE. Methods Minipumps allowing continuous intracerebroventricular (icv) infusion of siponimod for 4 weeks were implanted into C57BL/6 mice subjected to MOG35-55-induced EAE. Electrophysiology, immunohistochemistry, western blot, qPCR experiments, and peripheral lymphocyte counts were performed. In addition, the effect of siponimod on activated microglia was assessed in vitro to confirm the direct effect of the drug on CNS-resident immune cells. Results Siponimod administration (0.45 μg/day) induced a significant beneficial effect on EAE clinical scores with minimal effect on peripheral lymphocyte counts. Siponimod rescued defective GABAergic transmission in the striatum of EAE, without correcting the EAE-induced alterations of glutamatergic transmission. We observed a significant attenuation of astrogliosis and microgliosis together with reduced lymphocyte infiltration in the striatum of EAE mice treated with siponimod. Interestingly, siponimod reduced the release of IL-6 and RANTES from activated microglial cells in vitro, which might explain the reduced lymphocyte infiltration. Furthermore, the loss of parvalbumin-positive (PV+) GABAergic interneurons typical of EAE brains was rescued by siponimod treatment, providing a plausible explanation of the selective effects of this drug on inhibitory synaptic transmission. Conclusions Altogether, our results show that siponimod has neuroprotective effects in the CNS of EAE mice, which are likely independent of its peripheral immune effect, suggesting that this drug could be effective in limiting neurodegenerative pathological processes in MS. Keywords StriatumSynaptic transmissionGABANeurodegenerationExperimental autoimmune encephalomyelitisParvalbumin neuronhttp://dx.doi.org/10.13039/100008792Novartis Pharmaissue-copyright-statement© The Author(s) 2016 ==== Body Background Multiple sclerosis (MS) is an inflammatory neurodegenerative disease of the central nervous system (CNS), affecting young adults. Clinically, two main forms of MS can be distinguished, the relapsing-remitting (RRMS) and the progressive, where the latter is described as the gradual progression of clinical disability in a patient either with a preceding relapsing course (secondary progressive MS (SPMS)) or without a preceding relapsing course (primary progressive MS (PPMS)) [1]. Such phenotypic distinction seems to rely on predominance of distinct pathogenic mechanisms that appear different, though poorly understood, throughout RRMS, PPMS, and SPMS [2]. In general, although inflammation has a recognized pivotal role in triggering the cascade of events leading to both white and gray matter (GM) damage, anti-inflammatory or immunomodulatory drugs have been found successful only in RRMS [3], with the recent exception of ocrelizumab, which was found to reduce disability progression also in PPMS [4]. A forthcoming challenge is therefore to identify drugs able to interfere with the mechanisms leading to neurodegeneration in PPMS and in SPMS. GM pathology in MS is increasingly recognized to contribute to the progression of the disease. In fact, GM damage occurs early in the disease course, is independent of demyelination, and is associated with neurological and neuropsychological disability [5]. Several studies in experimental autoimmune encephalomyelitis (EAE), the murine model of MS, have underscored an inflammation-dependent synaptopathy affecting different brain structures (striatum, hippocampus, cerebellum) and occurring independently of demyelination [6–10]. In such brain areas of EAE mice, the excitatory glutamatergic transmission is potentiated, while the GABAergic inhibitory transmission is reduced: this imbalance in synaptic transmission leads to excitotoxicity and subsequently to neurodegeneration [11]. Of note, glutamate receptor antagonists and gamma-amino butyric acid (GABA) agonists exert beneficial effects in both EAE and MS [12–16]. This synaptopathy has been proposed as a valuable therapeutic target [11]. Several drugs are under investigation for their neuroprotective effects in progressive MS. In particular, an ongoing phase-III clinical trial is currently conducted with the sphingosine-1-phosphate receptor (S1P) modulator, siponimod (BAF312) [17]. Siponimod is a next-generation S1P modulator selective for S1P1 and S1P5 that showed good safety and tolerability in humans. This compound has recently successfully passed a phase-II clinical trial in RRMS [18]. Compared to fingolimod, an S1P modulator selective for S1P1, S1P3, S1P4, and S1P5, the half-life of siponimod is shorter, allowing recovery of peripheral lymphocyte count to a normal range within 1 week after treatment cessation. The risk of bradycardia is mitigated using a dose titration scheme during treatment initiation [19, 20]. Most of the effects of siponomid are attributed to S1P1 expressed on lymphocytes, thereby preventing lymphocyte trafficking into the brain, but both S1P1 and S1P5 are widely expressed in brain-resident cells, like neurons, microglia, astroglia, and oligodendrocytes [21]. The aim of the present study was therefore to assess the central effects of siponimod in mice with myelin oligodendrocyte glycoprotein peptide 35-55 (MOG35-55)-induced EAE, a model of progressive MS, to provide a substrate of the putative neuroprotective effects of this drug. Methods EAE model EAE was induced as described previously [7, 20]. Mice were injected subcutaneously at the flanks with 200 μg of MOG35-55 emulsion to induce EAE by active immunization. The emulsion was prepared under sterile conditions using MOG35-55 (85 % purity; Espikem) in 300 μl of complete Freund’s adjuvant (CFA; Difco) containing Mycobacterium tuberculosis (8 mg/ml, strain H37Ra; Difco) and emulsified with phosphate buffer solution (PBS). All animals were injected with 500 ng of pertussis toxin (Sigma) intravenously on the day of immunization and 2 days later. Control animals received the same treatment as EAE mice without the immunogen MOG peptide, including complete CFA and pertussis toxin (referred to as hereafter as “CFA”). Animals were scored daily for clinical symptoms of EAE according to the following scale: 0 = no clinical signs, 1 = flaccid tail, 2 = hindlimb weakness, 3 = hindlimb paresis, 4 = complete bilateral hindlimb paralysis, and 5 = death due to EAE; intermediate clinical signs were scored by adding 0.5 [15, 22]. For each animal, the onset day was recorded as the day post-immunization (dpi) when it showed the first clinical manifestations. Experiments were carried out in accordance with Internal Institutional Review Committee, the European Directive 2010/63/EU and the European Recommendations 526/2007, and the Italian D.Lgs26/2014. All the efforts were made to minimize the number of animals utilized and their suffering. Siponimod formulation for minipump and surgery Siponimod (Novartis Pharma AG) was dissolved in a solution containing 10 % Solutol/Kolliphor HS15 (BASF Pharma Solutions)—final pH range between 6 and 7—at a final concentration of 2 mg/ml. This preparation allowed stability of the drug for up to 6 weeks at 37 °C. One week before immunization, mice were implanted with subcutaneous osmotic minipumps allowing continuous intracerebroventricular (icv) infusion of either vehicle or siponimod for 4 weeks (three sets of immunizations) [8, 22]. Different siponimod dosages were tested: 4.5, 0.45, and 0.225 μg/day. T cell absolute count T cell absolute count was performed on blood samples kept from the mandibular vein of the mouse. For the phenotypic characterization of cell populations, the following antibodies were used: CD8-FITC (Miltenyi Biotec), CD25-APC (Pharmingen), CD3-PE-Vio770 (Miltenyi Biotec), CD4-APC-Vio770 (Milteny Biotec), CD45R (B220)-Violblu (Milteny Biotec), NK1.1-PE (Milteny Biotec). At predetermined optimal concentrations, 100 μl of blood was stained by incubation with the antibodies. Fifty microliters of CountBright Absolute Counting Beads (Molecular Probes) was added, and, following lysis of red blood cells, cells were acquired on a CyAn Cytometer (Beckman Coulter). By comparing the ratio of bead events to cell events, absolute numbers of cells in the sample were calculated. Some experiments were performed by acquiring the stained blood samples on the CytoFLEX cytometer (Coulter), equipped with a volumetric sample injection module, which enables volumetric sampling and provides absolute cell counts for all samples without the use of beads. Determination of siponimod in mouse blood by LC-MS For quantitative determination, 10 spiked samples from 0.5 up to 10,000 ng/ml were prepared in the same matrix. Proteins were removed by protein precipitation by adding an organic solvent mixture. The organic layer was evaporated to dryness, and the residue was re-dissolved in HPLC buffer B, containing 5 mM ammonium formate. Aliquots of 2 μl were directly injected on a Agilent Eclipse Plus, RRHD 2.1 × 50 mm reversed-phase column with 1.8-μm particles, and kept at 40 °C. For separation, a linear gradient from 50 to 100 % B within 1.7 min was used. Solvent A was 0.2 % formic acid in water and solvent B 0.2 % formic acid in acetonitrile. The flow was kept at 500 μl/min during the whole cycle. For detection, the column effluent was guided directly to the electrospray source of the Agilent 6490 triple quadrupole MS with parameters optimized for siponimod. Compound and a structure-related internal standard were detected as their [MH]+ ions with the multiple reaction monitoring (MRM) transition 517.3 → 159.0, 416.1 for siponimod. For data processing, the compound to internal standard ratio of the extracted ion chromatograms was used. The calculation was based on a second order fitted and 1/x weighted calibration curve (r2 = 0.9987). The accuracy of the 10 individually prepared calibration samples was better than 15 %. The recovery of the compounds was 94.3 % (RSD = 1.3 %). The precision of the control samples (n = 3), distributed over the whole series, was better than RSD = 1.8 %. The LOQ for siponimod was found to be 0.5 ng/ml, based on the lowest calibration sample. Electrophysiology Mice were killed by cervical dislocation, and corticostriatal coronal slices (200 μm) were prepared from fresh tissue blocks of the brain with the use of a vibratome [6, 15]. Single slices were then transferred to a recording chamber and submerged in a continuously flowing oxygenated artificial cerebrospinal fluid (ACSF) (34 °C, 2–3 ml/min) gassed with 95 % O2–5 % CO2. The composition of the control ACSF was (in mM) as follows: 126 NaCl, 2.5 KCl, 1.2 MgCl2, 1.2 NaH2PO4, 2.4 CaCl2, 11 Glucose, 25 NaHCO3. Whole-cell patch clamp recordings were made with borosilicate glass pipettes (1.8 mm o.d.; 4–8 MΩ), in voltage-clamp mode, at the holding potential (HP) of −80 mV. Spontaneous excitatory and inhibitory post-synaptic currents (EPSCs, IPSCs) were recorded from medium spiny neurons (MSNs) as in Centonze et al. [15] and Rossi et al. [6]. Siponimod (1 μM) was added to the bath solution for 1 h, before recording. Synaptic events were stored by using P-CLAMP 9.2 (Axon Instruments) and analyzed offline on a personal computer with Mini Analysis 5.1 (Synaptosoft, Leonia, NJ, USA) software. Offline analysis was performed on spontaneous synaptic events recorded during fixed time epochs (1–2 min, three to five samplings), sampled every 5 or 10 min. One to six cells per animal were recorded, and four animals were sacrificed for each experimental group. Western blot Animals were killed by cervical dislocation at 24 dpi. Striata were quickly removed, snap frozen in dry ice, and homogenized as in Mandolesi et al. [8] and Gentile et al. [22]. Soon after blocking with 5 % milk in Tris buffered solution (TBS), the membrane was incubated for 1 h with mouse anti-β-actin primary antibody (1:20,000; Sigma-Aldrich) followed by anti-mouse IgG HRP (1:10,000, GE Healthcare, formerly Amersham Biosciences) secondary antibody. Immunodetection was performed by ECL reagent (Amersham) and membrane was exposed to film (Amersham). Next, a mouse anti-glial fibrillary amino acid protein (GFAP, 1:2000, Immunological Science, over night) primary antibody was used in combination with anti-mouse IgG HRP (1:4000 for GFAP; GE Healthcare, formerly Amersham Biosciences) secondary antibody. Immunodetection was performed as for β-actin. Densitometric analysis of bands was performed by NIH ImageJ software (http://rsb.info.nih.gov/ij/). Western blot (WB) results are presented as data normalized to control CFA values. RNA extraction and quantitative real-time PCR Striata were dissected in RNAse-free conditions. Total RNA was extracted according to the standard miRNeasy Micro kit protocol (Qiagen). Next, 350 ng of total RNA was reverse-transcribed using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystem), and 10 ng of complementary DNA (cDNA) was amplified with SensiMix SYBR Hi-Rox Kit (Bioline; Meridian Life Science) in triplicate using the Applied Biosystem 7900HT Fast Real-Time PCR system. Messenger RNA (mRNA) relative quantification was performed using the comparative cycle threshold (2−ΔΔCt) method. β-actin was used as endogenous control. Primer sequences: Ionized calcium binding adaptor protein-1 (Iba-1, NM_019467): 5′-GACAGACTGCCAGCCTAAGACAA-3′ (sense), 5′-CATTCGCTTCAAGGACATAATATCG-3′ (antisense); β-ACTIN (NM_007393): 5′-CCTAGCACCATGAAGATCAAGATCA-3′ (sense), 5′-AAGCCATGCCAATGTTGTCTCT-3′ (antisense). For CD3+ mRNA detection, 17.5 ng of the same cDNA was amplified with SensiMix II Probe (Bioline; Meridian Life Science) by using TaqMan gene expression assays (ID Mm00599684_g1 and ID Mm00607939; Applied Biosystem). Immunofluorescence and confocal microscopy Mice from two different immunization experiments were deeply anesthetized and intracardially perfused with ice-cold 4 % paraformaldehyde. Brains were post-fixed for 2 h and equilibrated with 30 % sucrose at least one overnight. Immunofluorescence was performed as in Mandolesi et al. [8] and Gentile et al. [22]. The following primary antibodies were used: rabbit anti-Iba1 (1:750; Wako Chemicals, USA), rabbit anti-GFAP (1:500; DAKO), rat anti-CD3 (1:300; AbD Serotec). Secondary antibodies used are as follows: Alexa Fluor-488 (1:200; Life Technologies) and Cy3-conjugated donkey anti-rabbit or anti-rat (1:200; Jackson ImmunoResearch). 4′,6-diamidino-2-phenylindole (DAPI; 0.01 mg/ml) was used to visualize nuclei. All images were acquired using a LSM7 Zeiss confocal laser-scanner microscope (Zeiss, Göttingen, Germany) and processed by NHI ImageJ software [8, 22]. Immunohistochemistry A section every third (for a total of eight sections per animal) throughout the rostro-caudal extent of the striatum was selected for staining and examination. Freshly cut 30-μm serial sections, after blocking of nonspecific staining with 10 % normal donkey serum, were incubated overnight with a rabbit polyclonal antibody against parvalbumin (PV) (1:1000; Immunological Science), then with a biotinylated rabbit secondary antibody (1:500; Vector Laboratories, Burlingame, CA, USA) for 1 h, and finally with Extravidin (1:1000; SIGMA) for 1 h. Reaction was developed in a freshly prepared diaminobenzidine (DAB)/H2O2 (DAB tablet, SIGMA). Sections were mounted on glass slides and coverslipped under Eukitt. Stereology Quantitative observations were limited to the dorsal striatum of the left hemisphere. Using the Stereo Investigator System (MicroBrightField Europe e.K., Magdeburg, Germany) composed of a Zeiss Axioimager.M2 microscope and MicroBrightField’s Stereo Investigator software package, an optical fractionator, stereological design was applied to obtain unbiased estimates of total PV+ cells. Sampling grids and magnifications were adjusted to obtain a relatively constant number of cells sampled and a coefficient of error (CE Gunderson) of ≤0.1. A tri-dimensional optical dissector counting probe (x, y, z dimension of 200 × 200 × 25 μm, respectively) was applied. Counts were performed using a ×20 objective. Total number was estimated according to the following formula: N=∑Q×1/ssf×1/asf×1/tsf. where ΣQ represents the total number of neurons counted in all optically sampled fields of the dorsal striatum, ssf is the section sampling fraction, asf is the area sampling fraction, and tsf is the thickness sampling fraction. Cell culture supernatant Luminex assay The BV2 immortalized murine microglial cell line was pre-treated for 1 h with siponimod 0.1 μM in dimethyl sulfoxide (DMSO), before incubation with tumor necrosis factor (TNF; 200 U/ml; Milteny Biotec); control cells received equal volume of DMSO (n = 3 per condition). The 24-h conditioned media was assayed for IL-6 and RANTES by Luminex assay, according to the manufacturer instructions (R&D systems). The plate was read on a Luminex-200 instrument (Luminex Corp., Austin, TX). Concentrations were calculated by using a standard 5P-logistic weighted curve generated for each target and expressed as picograms per milliliter (pg/ml). Statistical analysis For each type of experiment, at least three mice per group were employed. Throughout the text, “n” refers to the number of animals, except for electrophysiology, where it means the number of cells. Data were presented as the mean ± S.E.M. The significance level was established at p < 0.05. Statistical analysis was performed using unpaired Student’s T test for comparisons between two groups and non-parametric Mann-Whitney test, where needed. Multiple comparisons were analyzed by one-way ANOVA for independent measures followed by Tukey’s HSD or Newman-Keuls. Results Intracerebroventricular injection of 0.45 μg/day siponimod ameliorates clinical score of EAE mice without affecting peripheral CD3+ cell count In order to assess whether siponimod has direct neuroprotective effects, the drug was delivered directly into the brain by means of continuous intracerebroventricular infusion, starting 1 week before the induction of EAE. During the acute phase of the disease, several biochemical and electrophysiological analyses were performed on both the peripheral and the CNS compartment (Fig. 1a, timeline of the experimental design).Fig. 1 Intracerebroventricular administration of 0.45 μg/day siponimod attenuates EAE motor deficits without inducing peripheral lymphocyte depletion. a Schematic representation of the experimental design used in the study. Minipump implantation and siponimod/vehicle release preceded the induction of EAE by 1 week. All examinations were performed on animals during the peak of the symptomatic phase of the disease 18–24 dpi, both in the peripheral blood samples and in the striatum of the animals. b CD3+ lymphocytes were counted in the peripheral blood of EAE mice receiving vehicle or different siponimod dosages: significant reduction was observed for the 4.5 μg/day dosage. Siponimod 4.5 μg/day vs vehicle ##p < 0.01 unpaired T test; siponimod 0.45 μg/day vs vehicle p > 0.05 unpaired T test; siponimod 0.225 μg/day vs vehicle p > 0.05 unpaired T test. c Siponimod concentrations were determined in peripheral blood samples of EAE mice receiving intracerebroventricular release of different siponimod dosages: the highest concentration (4.5 μg/day) used was significantly higher compared to both 0.450 μg/day and 0.225 μg/day, while there were no statistically differences between the other two dosages. Tukey’s post hoc ***p < 0.001. d Representative clinical course of EAE mice treated with three different siponimod dosages: EAE disease progression is strongly affected by 4.5 μg/day dosage during the symptomatic phase of the disease and significantly attenuated by 0.45 μg/day concentration in the dpi range of 18–24. Daily statistical significance was evaluated by non-parametric Mann-Whitney test First, we tested three dosages of the drug (4.50, 0.450, and 0.225 μg/day) for their possible impact on peripheral T lymphocytes: siponimod is lipophilic and therefore able to cross the blood brain barrier (BBB). At 18 dpi, during the symptomatic phase of the disease, we counted the total number of CD3+ lymphocytes in peripheral blood samples of EAE mice receiving different dosages of siponimod or vehicle: while the highest dose of siponimod (4.5 μg/day; n = 3) elicited a strong reduction of T lymphocytes in the blood of EAE mice compared to EAE-vehicle mice (n = 4, p < 0.001; unpaired T test), the intermediate (0.450 μg/day; n = 6) and the lowest (0.225 μg/day; n = 4) dosages of siponimod induced a slight and not significant drop in T lymphocyte counts with respect to EAE-vehicle mice (unpaired T test; p > 0.05) (Fig. 1b). This result was consistent with siponimod concentrations found in peripheral blood samples (Fig. 1c): intracerebroventricular release of 4.5 μg/day siponimod (n = 4) resulted in a high amount of the drug in the peripheral blood (169.9 ± 31.49 nM), significantly different from that measured in the blood of mice receiving 0.45 μg/day (n = 3) and 0.225 μg/day (n = 5; one-way ANOVA p < 0.01; Tukey’s post hoc comparisons: 4.5 μg/day vs 0.45 and 0.225 μg/day, p < 0.001). The concentrations of siponimod detected in mice receiving 0.45 or 0.225 μg/day were detectable, but low (13 ± 4.7 nM and 6.6 ± 1.5 nM for the 0.45 and 0.225 μg/day groups, respectively). These levels are not expected to have major effects on peripheral lymphocytes in the majority of animals as confirmed by minor effects observed on peripheral lymphocyte count. From a clinical point of view, such preventive and CNS-directed treatment reduced motor disability associated to EAE in a dose-dependent manner (Fig. 1d). The highest dose of siponimod used in this study fully inhibited EAE development in treated mice compared to EAE-vehicle (EAE-siponimod n = 5, EAE-vehicle n = 9; from day 15 to 24 p < 0.001, non-parametric Mann-Whitney). This result, together with that shown in Fig. 1b, c, suggests that siponimod effects on peripheral lymphocytes mediate its disease-modifying activity, when administered at this daily concentration. Interestingly, while the 0.450 μg/day (n = 6) and the 0.225 μg/day (n = 6) doses exerted comparable effects on peripheral lymphocytes, only the former significantly ameliorated the disease severity (p < 0.05, non-parametric Mann-Whitney) in the dpi range of 21–24 of the disease compared to EAE-vehicle (n = 9) (Fig. 1d). Bringing together these data, we decided to carry out the study using the 0.450 μg/day dose to better dissect the neuroprotective action of siponimod. Siponimod icv treatment reduces gray matter inflammation in EAE mice Activation of microglia and astrocytes are hallmarks of both MS and EAE [23]. Both cell populations have been demonstrated to sustain inflammation within the CNS during autoimmune attack through antigen presentation and/or cytokine/chemokine secretion [24–26]. Moreover, we previously demonstrated that microglia and astroglia together with infiltrating lymphocytes play a pivotal role in inducing the synaptic alterations observed in EAE brain [8–10, 15, 27]. We investigated by immunohistochemistry the effect of siponimod on brain inflammatory reaction during EAE. Coronal striatal slices were probed for the GFAP, a marker upregulated during astrocyte activation. As depicted in Fig. 2a, a substantial reduction of astrogliosis was observed in EAE-siponimod striatum (green fluorescence for GFAP staining, DAPI counterstaining in gray) compared to EAE-vehicle. Qualitative data of confocal microscopy images were confirmed by western blot analysis of protein lysates from EAE-siponimod striatum compared to EAE-vehicle striatum: GFAP levels, normalized to β-actin, were reduced by 50 % in the striatum of EAE mice treated with siponimod (Fig. 2b, b’; n = 5 for EAE-vehicle, n = 5 for EAE-siponimod; p < 0.05, unpaired T test).Fig. 2 CNS-directed delivery of siponimod reduces astrogliosis in the EAE striatum. a Confocal microscopy images depict the striatum of EAE-vehicle and siponimod mice stained for the astrocyte marker GFAP (green fluorescence, counterstained with DAPI, gray). Expression of GFAP is markedly reduced in EAE-siponimod striatum compared to EAE-vehicle striatum (scale bar: 100 μm). The images are representative of stained sections from three animals per groups from two immunizations. The panel in b shows western blot comparing EAE-siponimod and EAE-vehicle striatal extracts probed for anti-GFAP antibody. Densitometric analysis of the bands in b’ reveals reduced GFAP content, relative to β-actin, in EAE-siponimod lysates compared to EAE-vehicle. WB data are normalized to EAE-vehicle values. Unpaired T test was used for two-group analysis. Values are means ± SEM, *p < 0.05 Microgliosis was assessed by immunofluorescence and qPCR. As shown in Fig. 3a, the staining for the ionized calcium binding adaptor protein-1 (IBA1), marker of microglia/macrophage, is more pronounced in EAE-vehicle than in EAE-siponimod striatum (red fluorescence, counterstaining with DAPI in gray). Moreover, in the same slices, icv treatment with siponimod reduced the number of infiltrating lymphocytes compared to the icv-vehicle control group (Fig. 3a, green fluorescence). Quantitative analysis of IBA1 and CD3 transcripts by qPCR confirms reduced microglia activation and lymphocyte infiltration in the striatum of EAE mice receiving siponimod in comparison to EAE-vehicle (Fig. 3b; n = 4 for EAE-vehicle, n = 5 for EAE-siponimod; p < 0.05, unpaired T test).Fig. 3 Attenuated microgliosis and reduced lymphocyte infiltration characterizes the striatum of EAE mice treated with siponimod. a Double immunostaining of coronal striatal sections (in gray DAPI nuclei) showing expression of Iba1-positive microglia/macrophage cells (red) and of CD3 (green) in EAE-siponimod and EAE-vehicle mice. Only few lymphocytes could be detected in the striatum of mice treated with siponimod, and microgliosis is visibly reduced in the same animals. Scale bar: 100 μm. b The levels of IBA1 and CD3 mRNA were quantified by qRT-PCR in the striatum of EAE-siponimod versus EAE-vehicle mice using β-actin as internal control. The histogram shows that both IBA1 and CD3 mRNA were significantly reduced in the striatum of siponimod-treated mice. Statistical significance was evaluated by unpaired T test: *p < 0.05 In vitro application of siponimod reduces IL-6 and RANTES release by TNF-treated microglial cells Microglia have been clearly involved in neuronal damage and T lymphocyte recruitment in the brain, by releasing pro-inflammatory cytokines and chemokines, like IL-6 and RANTES, which have been implicated in EAE and MS pathogenesis [28–32]. We tested the hypotheses that siponimod might reduce microglia activation. To this aim, we used an in vitro model of microglia activation. BV2 microglial cells were pre-treated with 0.1 μM siponimod prior to 24-h stimulation with TNF. By Luminex assay (Fig. 4a), we found that siponimod prevented the release of IL-6 induced by TNF stimulation of BV2 cells (one-way ANOVA: control cells 12.06 ± 0.55 pg/ml vs TNF 15.61 ± 0.40 pg/ml, p < 0.05 Tukey’s post hoc; TNF + siponimod 12.01 ± 0.21 pg/ml vs TNF, p < 0.05 Tukey’s post hoc). Siponimod in the absence of TNF did not alter the basal secretion of IL-6 (control cells vs vehicle + siponimod 11.6 ± 1.11 pg/ml, p > 0.05 Tukey’s post hoc).Fig. 4 In vitro siponimod application to activated BV2 microglial cell reduces IL-6 and RANTES release. BV2 cells were pre-treated with siponimod prior to stimulation with TNF. The levels of secreted IL-6 (a) and RANTES (b) were measured by Luminex assay on collected media. Application of siponimod to non-activated cells does not modulate the secretion of both molecules. Data were expressed as picograms per milliliter (pg/ml) and were analyzed by one-way ANOVA, followed by Tukey’s post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001 TNF induced a strong release of RANTES in cell culture supernatant of BV2 cells (one-way ANOVA: control cells 596.2 ± 17.4 pg/ml vs TNF 1794 ± 33.65 pg/ml, p < 0.001 Tukey’s post hoc; Fig. 4b). Siponimod reduced the amount of RANTES released by TNF-stimulated cells (TNF vs siponimod + TNF 1621 ± 21.67 pg/ml, p < 0.01 Tukey’s post hoc; siponimod + TNF vs control p < 0.01 Tukey’s post hoc) (Fig. 4b). As for IL-6, siponimod did not alter the release of RANTES by BV2 microglial cells in the absence of TNF (control vs siponimod-TNF 698.4 ± 16.44 pg/ml, p > 0.05, Tukey’s post hoc). Siponimod treatment recovers GABAergic transmission alterations in EAE We next asked whether siponimod might improve the synaptic transmission alterations typical of EAE brains [11], by recording both the glutamatergic and the GABAergic currents from striatal MSNs. We previously demonstrated that both the presynaptic (frequency) and post-synaptic (kinetic) properties of the spontaneous glutamatergic currents (sEPSCs) are altered in the striatum of EAE mice during the acute phase of the disease [15]. Here, we first investigated the effect on sEPSCs after in vivo treatment with siponimod 0.45 μg/day on sEPSCs, by recording striatal slices in the dpi range of 20–24. Siponimod did not correct neither the kinetic parameters of rise time, decay time, and half width (Fig. 5a–a”) nor the frequencies (Fig. 5a’–a”) of EAE sEPSC (n = 8 for both groups; p > 0.05, unpaired T test for all comparisons).Fig. 5 Siponimod treatment corrects GABAergic defects in EAE striatum. a The glutamatergic transmission was recorded from MSNs of EAE-vehicle and EAE-siponimod mice. The analysis of the increased kinetic properties (rise time, decay time, and half width) of sEPSCs showed that there were no differences between the two groups (unpaired T test, p > 0.05). a’ The frequencies of the glutamatergic currents, increased in EAE striatum, were not corrected by siponimod (unpaired T test, *p > 0.05). a” Electrophysiological traces representative of glutamatergic currents recorded from MSNs of EAE-vehicle and EAE-siponimod mice. b Siponimod restores normal GABAergic frequencies in the striatum of acute phase EAE mice (unpaired T test,*p < 0.05). b’ Electrophysiological traces representative of GABAergic currents recorded from MSNs of EAE-vehicle and EAE-siponimod mice. c The frequencies of the GABAergic currents, reduced in EAE striatum, were completely rescued after 1 h of siponimod incubation in EAE slices (unpaired T test, **p < 0.01) Conversely, siponimod icv treatment significantly ameliorated GABAergic defects in the striatum of EAE mice, by increasing GABA frequencies, which are considerably reduced in the symptomatic phase of the disease [6]. Siponimod restored normal values of frequency of spontaneous inhibitory post-synaptic currents (sIPSCs) (Fig. 5b, b’; EAE-vehicle 0.61 ± 0.11 Hz, n = 10; EAE-siponimod: 1.52 ± 0.12 Hz, n = 8; p < 0.05, unpaired T test), suggesting that siponimod might have neuroprotective effects by preventing GABAergic transmission deficits. Furthermore, in order to better explain the protective role exerted by siponimod on GABA synapses, we incubated EAE slices taken during the symptomatic phase of the disease (20–25 dpi) with siponimod (1 μM) or vehicle (DMSO) for 1 h. The in vitro treatment with siponimod was able to rescue the GABAergic impairment induced by EAE (Fig. 5c). The sIPSC frequency was higher in EAE-siponimod slices (1.47 ± 0.61 Hz, n = 8) in comparison to EAE untreated slices (0.78 ± 0.11 Hz; n = 8; p < 0.01, unpaired T test), reaching values indistinguishable from control slices. These data support the idea that siponimod exerts its central protective role likely by reducing resident immune cell activation. Siponimod promotes PV+ interneuron survival GABAergic interneurons, together with axon collaterals from MSNs themselves, are the main source of GABA inputs to striatal MSNs [33, 34]. The loss of parvalbumin-positive (PV+) GABAergic interneurons together with the noxious effects of pro-inflammatory cytokines has been proposed as major determinant of the reduced GABAergic tone in the striatum of EAE mice [6]. Therefore, we assessed whether siponimod could prevent PV+ interneuron death in the striatum of EAE mice. To avoid possible interference of the damage caused by the cannula inserted in the right lateral ventricle, we performed unbiased stereological counts in the left dorsal striatum of EAE mice. In Fig. 6a, there are representative images of coronal slices stained with anti-parvalbumin antibody. This experimental approach confirmed the reduction of the total number of PV+ neurons in EAE-vehicle (1885 ± 117.7, n = 5) compared to control CFA-vehicle striatum (2664 ± 156.6, n = 3) and, more interestingly, showed a recovery induced by siponimod (2293 ± 115.3, n = 5; one-way ANOVA p < 0.01; Newman-Keuls post hoc comparisons: CFA-vehicle vs EAE-vehicle, p < 0.01; CFA-vehicle vs EAE-siponimod, p > 0.05; EAE-vehicle vs EAE-siponimod, p < 0.05; Fig. 6b).Fig. 6 Brain infusion of siponimod promotes the survival of PV+ GABAergic interneurons during the course of EAE. a Low magnification images (×5 objective) of coronal striatal sections from control CFA-vehicle, EAE-vehicle, and EAE-siponimod showing area selected for PV counting on serial sections (dorsal striatum) and DAB-immunolabeled interneurons (scale bar: 200 μm): at high magnification, the density of PV+ cells is visibly reduced in EAE-vehicle compared to that in CFA-vehicle with a partial recovery in EAE-siponimod slices. Yellow arrows indicate PV-labeled cells (scale bar: 20 μm). b Histogram represents the quantitative stereological counts of PV+ cells performed on striatal slices: the intracerebroventricular treatment with siponimod significantly increases the total number of PV+ cells in the dorsal left striatum of EAE mice. Data are expressed as mean ± S.E.M. Statistical comparisons showed in the graph were calculated with Newman-Keuls post hoc: CFA-vehicle vs EAE-vehicle **p < 0.01; CFA-vehicle vs EAE-siponimod p > 0.05; EAE-vehicle vs EAE-siponimod*p < 0.05 These data indicate that siponimod prevents GABAergic interneuron loss in the striatum of symptomatic EAE mice. Discussion Discovering drugs effective in progressive forms of MS is a hard challenge for both researchers and physicians working in the field of MS. Promoting neuroprotection is a promising therapeutic strategy in the light of the relevant contribution of neurodegeneration to MS pathogenesis and progression [2]. GM damage, including neuron and synaptic loss, is currently considered as the major contributor to both cognitive impairment and motor disability in MS patients [35, 36]. Recently, it has been highlighted that inflammation-driven synaptic abnormalities are pathological hallmarks of both EAE and MS brains: the sustained imbalance between glutamatergic and GABAergic transmission, occurring independently of demyelination and axonal loss, is supposed to induce excitotoxic neurodegeneration [11]. Preventing such inflammatory neurodegenerative mechanisms is a novel and poorly explored therapeutic strategy. In the present study, we investigated whether siponimod could be neuroprotective in EAE mice, by limiting striatal synaptopathy. Major findings of this study are that the GABAergic transmission alterations typical of the EAE striatum are rescued by siponimod treatment, likely due to reduced local inflammatory reaction and to increased survival of PV+ interneurons, which are the main GABAergic input to MSNs. In an attempt to discriminate the peripheral from the central effects of siponimod, we used a direct and preventive CNS administration of the drug, choosing the concentration of siponimod able to ameliorate the disease severity, but with minimal impact on blood lymphocyte count. Adjusting the administration dose was our primary goal to avoid that the observed beneficial effects could be secondary to its peripheral immunomodulatory action. In fact, by virtue of its lipophilic nature, siponimod easily crosses the BBB reaching peripheral tissues. Not surprisingly, the highest dosage (4.5 μg/day) tested in this study resulted in significant blood exposure of siponimod and a strong reduction of T cells in the blood, likely as a result of retention in lymph nodes, and completely prevented EAE development. The lower dosages (0.45 and 0.225 μg/day) had both a minimal impact on peripheral lymphocyte count with siponimod blood levels reaching only low nanomolar range. Since the 0.225 μg/day dose had no effect on the clinical score, we chose the 0.45 μg/day dosage due to its beneficial effect on disease severity, even if this was limited to the peak of the acute phase (20–24 dpi). However, since the clinical symptoms are mainly due to lesions in the spinal cord and since we expect a gradient of the compound between the site of administration and the site of the spinal cord lesions, a lack of strong clinical efficacy does not rule out more potent effects at the proximity of the administration site. Moreover, although we cannot exclude any peripheral effect of siponimod, we might assume that the observed recovery in clinical score could be linked to central effects of this drug. Accordingly, this treatment induced attenuation of microgliosis and astrogliosis with reduced number of infiltrating CD3+ lymphocytes in the striatum of EAE mice during the symptomatic phase of the disease. A possible explanation for the low number of infiltrating lymphocytes could be a reduced inflammatory reaction in the brain, caused by the preventive and central treatment with siponimod, likely promoting a different tissue distribution of T cells compared to EAE untreated mice. Indeed, CNS-resident cells, such as microglia, astroglia, oligodendrocytes, and neurons, express sphingosine receptors, although with some differences in the receptor subunit composition [21]. Most of the anti-inflammatory effects of S1PR agonists, like fingolimod and siponimod, in the CNS have been associated to their action on astroglia and microglia [37, 38]. Microglia activation occurs early in the EAE brain and spinal cord, before the appearance of motor deficits [25, 39, 40]. By secreting a number of pro-inflammatory cytokines and chemotactic factors, microglia are involved in the amplification of the inflammatory reaction and the recruitment of monocyte-blood-born cells and T cells into the brain. In our experimental paradigm of preventive and central treatment with siponimod, siponimod may prevent the chain of noxious events triggered by microglia in the EAE brain, reducing MOG-specific lymphocyte recruitment. To prove this hypothesis, we used an in vitro simplification of the in vivo experimental approach: we pre-treated microglial cells with siponimod prior to activation with TNF. We found that siponimod reduced microglial release of IL-6 and RANTES, some of the most relevant cytokines/chemokines involved in EAE/MS pathogenesis [28–32]. Indeed, although IL-6 levels in the CSF of MS patients have reported to be unaltered, likely indicating that IL-6 is not a specific signature of the disease [41], it can be hypothesized that this cytokine has a local role in the modulation of the inflammatory reaction, since its expression was detected in glial cells of MS lesions [42]. Conversely, it is worth noting that RANTES levels have been found increased in the CSF of MS patients [43] and, most notably, they correlated with inflammation and synaptic excitability [44]. We have recently demonstrated that inflammation causes an imbalance between the glutamatergic and the GABAergic transmission in the EAE striatum [6, 15]. Notably, such alterations are putative therapeutic targets, and some drugs currently used in MS treatment have already been found able to interfere with the chain of events leading to such dysregulation [45–47]. Of note, fingolimod, given systemically and starting the day of immunization, was shown to strongly ameliorate clinical deficits and to correct the glutamatergic transmission alterations of EAE striatum [45]. Here, we demonstrated that siponimod rescued GABAergic transmission defects, without improving the glutamatergic transmission in MSNs. Such differences between the results obtained with the two S1P modulators probably stems from the way of administration of the drug and the dose/drug exposure. The oral delivery of fingolimod likely prevented T lymphocytes from migrating into the brain, thus explaining the marked effect on disease progression and the excitatory currents. Notably, even if in a small number, infiltrating T cells, which have been highly involved in glutamatergic alteration in the striatum of EAE mice [15] may be able to affect excitatory transmission. Furthermore, an acute exposure (1 h) of siponimod to slices of EAE mice reproduced the same effect of the in vivo treatment on GABA transmission, reinforcing the idea that the compound may improve the inhibitory tone in the striatum of EAE mice by interacting with resident immune cells. In the present study, we observed that icv treatment with siponimod recovered the frequencies of the sIPSCs in EAE striatum. This specific impairment of the GABAergic neurotransmission has been linked to inflammation, although a selective cytokine or chemokine has not been recognized, and to the reduced number of GABAergic interneurons contacting MSNs in the EAE striatum [6]. The death of PV+ interneurons is a pathological hallmark of both MS and EAE brains [6, 48], and the mechanisms leading to such degeneration are still unknown. Inflammation is thought to play a role in this neuronal death, which has been observed in other brain inflammatory conditions [49] and in rodent chronic stress paradigms [50]. In the striatum of EAE mice, GABAergic defects occur throughout the disease course [6]. In our study, we found that siponimod promotes the survival of these neurons, thus explaining the recovery of GABAergic neurotransmission, in spite of the potentiated glutamatergic transmission. Further studies are needed to dissect the mechanisms of neuroprotection exerted by siponimod on PV+ interneurons: other mechanisms involving neuroprotective factors, like the brain-derived neurotrophic factor (BDNF), may be implied [51]. It is worth noting that reduced GABA concentrations in the hippocampus and sensory-motor cortex have recently been correlated with physical disability in progressive MS patients [52]. This finding strongly corroborates the role of the inflammatory synaptopathy in MS progression. Interestingly, in our study, the attenuation of the motor symptoms of EAE mice treated with siponimod was observed in the symptomatic phase of the disease in connection to the improved tone of the GABAergic transmission in the striatum, a subcortical brain area involved in motor control [53] and compromised in MS [54]. Conclusions In conclusion, the present investigation highlights a neuroprotective activity of siponimod with respect to a relevant feature of neurodegeneration affecting MS brains, the reduced tone of GABAergic transmission. Abbreviations BBBBlood brain barrier BDNFBrain-derived neurotrophic factor CFAComplete Freund’s adjuvant CNSCentral nervous system DABDiaminobenzidine DAPI4′,6-Diamidino-2-phenylindole DMSODimethyl sulfoxide dpiDay post-immunization EAEExperimental autoimmune encephalomyelitis GABAGamma-amino butyric acid GFAPGlial fibrillary amino acid protein GMGray matter IBA1Ionized calcium binding adaptor protein-1 icvIntracerebroventricular ILInterleukin MOGMyelin oligodendrocyte glycoprotein MRMMultiple reaction monitoring MSMultiple sclerosis MSNMedium spiny neuron PBSPhosphate buffer solution PPMSPrimary progressive MS PVParvalbumin qPCRQuantitative-polymerase chain reaction RANTESRegulated on activation, normal T cell expressed and secreted RRMSRelapsing-remitting MS S1PSphingosine-1-phosphate receptor sEPSCSpontaneous excitatory post-synaptic currents sIPSCsSpontaneous inhibitory post-synaptic currents SPMSSecondary progressive MS TBSTris buffered solution TNFTumor necrosis factor WBWestern blot Acknowledgements The authors thank Vladimiro Batocchi for helpful technical assistance. Funding The study was supported by a Novartis Pharma grant to D.C. Availability of data and materials Data supporting the conclusions of this article are presented in the manuscript. Authors’ contributions AG, AM, GB, LB, AS, DC, and GM conceived and designed the study. AG, AM, SB, DF, FDV, RF, EP, FG, and AS acquired and analyzed the data. AG, AM, AS, DC, and GM drafted the manuscript. All the authors critically revised the article for important intellectual content and approved the version to be published. AG and AM contributed equally to this work, as first authors. DC and GM contributed equally to this work, as senior authors. Competing interests DC: advisory board member of Almirall, Bayer Schering, Biogen, Genzyme, GW Pharmaceuticals, Merck Serono, Novartis, Teva; honoraria for speaking from Almirall, Bayer Schering, Biogen Idec, Genzyme, GW Pharmaceuticals, Merck Serono, Novartis, Sanofi, Teva; principal investigator in clinical trials for Bayer Schering, Biogen Idec, Merck Serono, Mitsubishi, Novartis, Roche, Sanofi, Teva; grants for clinical and preclinical research from Bayer, Biogen, Merck Serono, Novartis, Teva. LB: honoraria for speaking from Teva and Genzyme; research grants from Teva. AS: employee of Novartis Pharma, Basel, Switzerland. Consent for publication Not applicable. Ethics approval and consent to participate All animal experiments described in this study were conducted at IRCCS Fondazione Santa Lucia, according to the guidelines set by the Internal Institutional Review Committee, the European Directive 2010/63/EU and the European Recommendations 526/2007 and the Italian D.Lgs 26/2014. All efforts were made to minimize the number of animals used and their suffering. ==== Refs References 1. Lublin FD Reingold SC Cohen JA Cutter GR Sørensen PS Thompson AJ Defining the clinical course of multiple sclerosis: the 2013 revisions Neurology 2014 83 278 286 10.1212/WNL.0000000000000560 24871874 2. Calabrese M Magliozzi R Ciccarelli O Geurts JJ Reynolds R Martin R Exploring the origins of grey matter damage in multiple sclerosis Nat Rev Neurosci 2015 16 147 158 10.1038/nrn3900 25697158 3. Salvetti M Landsman D Schwarz-Lam P Comi G Thompson AJ Fox RJ Progressive MS: from pathophysiology to drug discovery Mult Scler 2015 21 1376 1384 10.1177/1352458515603802 26362902 4. Fyfe I In the news: ocrelizumab excites ECTRIMS Nat Rev Neurol 2015 11 667 5. Ciccarelli O Barkhof F Bodini B De Stefano N Golay X Nicolay K Pathogenesis of multiple sclerosis: insights from molecular and metabolic imaging Lancet Neurol 2014 13 807 822 10.1016/S1474-4422(14)70101-2 25008549 6. Rossi S Muzio L De Chiara V Grasselli G Musella A Musumeci G Impaired striatal GABA transmission in experimental autoimmune encephalomyelitis Brain Behav Immun 2011 25 947 956 10.1016/j.bbi.2010.10.004 20940040 7. Mandolesi G Grasselli G Musella A Gentile A Musumeci G Sepman H GABAergic signaling and connectivity on Purkinje cells are impaired in experimental autoimmune encephalomyelitis Neurobiol Dis 2012 46 414 424 10.1016/j.nbd.2012.02.005 22349452 8. Mandolesi G Musella A Gentile A Grasselli G Haji N Sepman H Interleukin-1β alters glutamate transmission at purkinje cell synapses in a mouse model of multiple sclerosis J Neurosci 2013 33 12105 12121 10.1523/JNEUROSCI.5369-12.2013 23864696 9. Nisticò R Mango D Mandolesi G Piccinin S Berretta N Pignatelli M Inflammation subverts hippocampal synaptic plasticity in experimental multiple sclerosis PLoS One 2013 8 e54666 10.1371/journal.pone.0054666 23355887 10. Mori F Nisticò R Mandolesi G Piccinin S Mango D Kusayanagi H Interleukin-1β promotes long-term potentiation in patients with multiple sclerosis Neuromolecular Med 2014 16 38 51 10.1007/s12017-013-8249-7 23892937 11. Mandolesi G Gentile A Musella A Fresegna D De Vito F Bullitta S Synaptopathy connects inflammation and neurodegeneration in multiple sclerosis Nat Rev Neurol 2015 11 711 724 10.1038/nrneurol.2015.222 26585978 12. Plaut GS Effectiveness of amantadine in reducing relapses in multiple sclerosis J R Soc Med 1987 80 91 93 3550084 13. Cutter NC Scott DD Johnson JC Whiteneck G Gabapentin effect on spasticity in multiple sclerosis: a placebo-controlled, randomized trial Arch Phys Med Rehabil 2000 81 164 169 10.1016/S0003-9993(00)90135-7 10668769 14. Pitt D Werner P Raine CS Glutamate excitotoxicity in a model of multiple sclerosis Nat Med 2000 6 67 70 10.1038/71555 10613826 15. Centonze D Muzio L Rossi S Cavasinni F De Chiara V Bergami A Inflammation triggers synaptic alteration and degeneration in experimental autoimmune encephalomyelitis J Neurosci 2009 29 3442 3452 10.1523/JNEUROSCI.5804-08.2009 19295150 16. Bhat R Axtell R Mitra A Miranda M Lock C Tsien RW Inhibitory role for GABA in autoimmune inflammation Proc Natl Acad Sci U S A 2010 107 2580 2585 10.1073/pnas.0915139107 20133656 17. Subei AM Cohen JA Sphingosine 1-phosphate receptor modulators in multiple sclerosis CNS Drugs 2015 29 565 575 10.1007/s40263-015-0261-z 26239599 18. Selmaj K Li DK Hartung HP Hemmer B Kappos L Freedman MS Siponimod for patients with relapsing-remitting multiple sclerosis (BOLD): an adaptive, dose-ranging, randomised, phase 2 study Lancet Neurol 2013 12 756 767 10.1016/S1474-4422(13)70102-9 23764350 19. Gergely P Nuesslein-Hildesheim B Guerini D Brinkmann V Traebert M Bruns C The selective sphingosine 1-phosphate receptor modulator BAF312 redirects lymphocyte distribution and hasspecies-specific effects on heart rate Br J Pharmacol 2012 167 1035 1047 10.1111/j.1476-5381.2012.02061.x 22646698 20. Shakeri-Nejad K Aslanis V Veldandi UK Mooney L Pezous N Brendani B Effects of therapeutic and supratherapeutic doses of siponimod (BAF312) on cardiac repolarization in healthy subjects Clin Ther 2015 37 2489 2505 10.1016/j.clinthera.2015.09.006 26519230 21. Groves A Kihara Y Chun J Fingolimod: direct CNS effects of sphingosine 1-phosphate (S1P) receptor modulation and implications in multiple sclerosis therapy J Neurol Sci 2013 328 9 18 10.1016/j.jns.2013.02.011 23518370 22. Gentile A Fresegna D Federici M Musella A Rizzo FR Sepman H Dopaminergic dysfunction is associated with IL-1β-dependent mood alterations in experimental autoimmune encephalomyelitis Neurobiol Dis 2015 74 347 358 10.1016/j.nbd.2014.11.022 25511803 23. Compston A Coles A Multiple sclerosis Lancet 2008 372 1502 1517 10.1016/S0140-6736(08)61620-7 18970977 24. Muzio L Martino G Furlan R Multifaceted aspects of inflammation in multiple sclerosis: the role of microglia J Neuroimmunol 2007 191 39 44 10.1016/j.jneuroim.2007.09.016 17936915 25. Ajami B Bennett JL Krieger C McNagny KM Rossi FM Infiltrating monocytes trigger EAE progression, but do not contribute to the resident microglia pool Nat Neurosci 2011 14 1142 1149 10.1038/nn.2887 21804537 26. Moreno M Bannerman P Ma J Guo F Miers L Soulika AM Conditional ablation of astroglial CCL2 suppresses CNS accumulation of M1 macrophages and preserves axons in mice with MOG peptide EAE J Neurosci 2014 34 8175 8185 10.1523/JNEUROSCI.1137-14.2014 24920622 27. Grasselli G Rossi S Musella A Gentile A Loizzo S Muzio L Abnormal NMDA receptor function exacerbates experimental autoimmune encephalomyelitis Br J Pharmacol 2013 168 502 517 10.1111/j.1476-5381.2012.02178.x 22924679 28. Eugster HP Frei K Kopf M Lassmann H Fontana A IL-6-deficient mice resist myelin oligodendrocyte glycoprotein-induced autoimmune encephalomyelitis Eur J Immunol 1998 28 2178 2187 10.1002/(SICI)1521-4141(199807)28:07<2178::AID-IMMU2178>3.0.CO;2-D 9692887 29. Sørensen TL Tani M Jensen J Pierce V Lucchinetti C Folcik VA Expression of specific chemokines and chemokine receptors in the central nervous system of multiple sclerosis patients J Clin Invest 1999 103 807 815 10.1172/JCI5150 10079101 30. Baranzini SE Elfstrom C Chang SY Butunoi C Murray R Higuchi R Transcriptional analysis of multiple sclerosis brain lesions reveals a complex pattern of cytokine expression J Immunol 2000 165 6576 6582 10.4049/jimmunol.165.11.6576 11086101 31. dos Santos AC Barsante MM Arantes RM Bernard CC Teixeira MM CCL2 and CCL5 mediate leukocyte adhesion in experimental autoimmune encephalomyelitis—an intravital microscopic study J Neuroimmunol 2005 162 122 129 10.1016/j.jneuroim.2005.01.020 15833367 32. Furlan R Rovaris M Martinelli Boneschi F Khademi M Bergami A Immunological patterns identifying disease course and evolution in multiple sclerosis patients J Neuroimmunol 2005 165 192 200 10.1016/j.jneuroim.2005.04.012 15949850 33. Kawaguchi Y Wilson CJ Augood SJ Emson PC Striatal interneurones: chemical, physiological and morphological characterization Trends Neurosci 1995 18 527 535 10.1016/0166-2236(95)98374-8 8638293 34. Gustafson N Gireesh-Dharmaraj E Czubayko U Blackwell KT Plenz D A comparative voltage and current-clamp analysis of feedback and feedforward synaptic transmission in the striatal microcircuit in vitro J Neurophysiol 2006 95 737 752 10.1152/jn.00802.2005 16236782 35. De Stefano N Matthews PM Filippi M Agosta F De Luca M Evidence of early cortical atrophy in MS: relevance to white matter changes and disability Neurology 2003 60 1157 1162 10.1212/01.WNL.0000055926.69643.03 12682324 36. Roosendaal SD Bendfeldt K Vrenken H Polman CH Borgwardt S Grey matter volume in a large cohort of MS patients: relation to MRI parameters and disability Mult Scler 2011 17 1098 1106 10.1177/1352458511404916 21586487 37. Jackson SJ Giovannoni G Baker D Fingolimod modulates microglial activation to augment markers of remyelination J Neuroinflammation 2011 8 76 10.1186/1742-2094-8-76 21729281 38. Colombo E Di Dario M Capitolo E Chaabane L Newcombe J Martino G Fingolimod may support neuroprotection via blockade of astrocyte nitric oxide Ann Neurol 2014 76 325 337 10.1002/ana.24217 25043204 39. Haji N Mandolesi G Gentile A Sacchetti L Fresegna D Rossi S TNF-α-mediated anxiety in a mouse model of multiple sclerosis Exp Neurol 2012 237 296 303 10.1016/j.expneurol.2012.07.010 22836148 40. Gentile A De Vito F Fresegna D Musella A Buttari F Bullitta S Exploring the role of microglia in mood disorders associated with experimental multiple sclerosis Front Cell Neurosci 2015 9 243 10.3389/fncel.2015.00243 26161070 41. Uzawa A Mori M Uchida T Masuda H Ohtani R Kuwabara S Increased levels of CSF CD59 in neuromyelitis optica and multiple sclerosis Clin Chim Acta 2016 453 131 3 10.1016/j.cca.2015.12.013 26686775 42. Maimone D Guazzi GC Annunziata P IL-6 detection in multiple sclerosis brain J Neurol Sci 1997 146 1 59 65 10.1016/S0022-510X(96)00283-3 9077497 43. Szczuciński A Losy J CCL5, CXCL10 and CXCL11 chemokines in patients with active and stable relapsing-remitting multiple sclerosis Neuroimmunomodulation 2011 18 1 67 72 10.1159/000317394 20720435 44. Mori F, Nisticò R, Nicoletti CG, Zagaglia S, Mandolesi G, Piccinin S et al. RANTES correlates with inflammatory activity and synaptic excitability in multiple sclerosis. Mult Scler. 2016. doi:10.1177/ 1352458515621796. 45. Rossi S Lo Giudice T De Chiara V Musella A Studer V Motta C Oral fingolimod rescues the functional deficits of synapses in experimental autoimmune encephalomyelitis Br J Pharmacol 2012 165 861 869 10.1111/j.1476-5381.2011.01579.x 21740406 46. Gentile A Rossi S Studer V Motta C De Chiara V Musella A Glatiramer acetate protects against inflammatory synaptopathy in experimental autoimmune encephalomyelitis J Neuroimmune Pharmacol 2013 8 651 663 10.1007/s11481-013-9436-x 23370991 47. Musella A Mandolesi G Gentile A Rossi S Studer V Motta C Cladribine interferes with IL-1β synaptic effects in experimental multiple sclerosis J Neuroimmunol 2013 264 8 13 10.1016/j.jneuroim.2013.08.009 24045165 48. Clements RJ McDonough J Freeman EJ Distribution of parvalbumin and calretinin immunoreactive interneurons in motorcortex from multiple sclerosis post-mortem tissue Exp Brain Res 2008 187 459 465 10.1007/s00221-008-1317-9 18297277 49. Ji MH Qiu LL Tang H Ju LS Sun XR Zhang H Sepsis-induced selective parvalbumin interneuron phenotype loss and cognitive impairments may be mediated by NADPH oxidase 2 activation in mice J Neuroinflammation 2015 12 182 10.1186/s12974-015-0401-x 26416717 50. Czeh B Simon M van der Hart MG Schmelting B Hesselink MB Fuchs E Chronic stress decreases the number of parvalbumin-immunoreactive interneurons in the hippocampus: prevention by treatment with a substance P receptor (NK1) antagonist Neuropsychopharmacology 2005 30 67 79 10.1038/sj.npp.1300581 15470372 51. Allen SJ Watson JJ Shoemark DK Barua NU Patel NK GDNF, NGF and BDNF as therapeutic options for neurodegeneration Pharmacol Ther 2013 138 155 175 10.1016/j.pharmthera.2013.01.004 23348013 52. Cawley N Solanky BS Muhlert N Tur C Edden RA Wheeler-Kingshott CA Reduced gamma-aminobutyric acid concentration is associated with physical disability in progressive multiple sclerosis Brain 2015 138 2584 2595 10.1093/brain/awv209 26304151 53. Haber SN The place of dopamine in the cortico-basal ganglia circuit Neuroscience 2014 282C 248 257 10.1016/j.neuroscience.2014.10.008 25445194 54. Tao G Datta S He R Nelson F Wolinsky JS Narayana PA Deep gray matter atrophy in multiple sclerosis: a tensor based morphometry J Neurol Sci 2009 282 39 46 10.1016/j.jns.2008.12.035 19168189

PMC005xxxxxx/PMC5002119.txt

==== Front BMC GenomicsBMC GenomicsBMC Genomics1471-2164BioMed Central London 297210.1186/s12864-016-2972-zResearch ArticleIdentification of protein-damaging mutations in 10 swine taste receptors and 191 appetite-reward genes Clop Alex alex.clop@cragenomica.es 1Sharaf Abdoallah 12Castelló Anna 1Ramos-Onsins Sebastián 1Cirera Susanna 3Mercadé Anna 1Derdak Sophia 45Beltran Sergi 45Huisman Abe 6Fredholm Merete 3van As Pieter 7Sánchez Armand armand.sanchez@uab.cat 181 Centre for Research in Agricultural Genomics (CRAG) CSIC-IRTA-UAB-UB, Campus UAB, 08193 Cerdanyola del Valles, Catalonia, Spain 2 Faculty of agriculture, Ain Shams University, Khalifa El-Maamon st, Abbasiya sq, 11566 Cairo, Egypt 3 Department of Veterinary Clinical and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Grønnegårdsvej 3, 1870 Frederiksberg, Denmark 4 CNAG-CRG, Centre for Genomic Regulation (CRG), Barcelona Institute of Science and Technology (BIST), Baldiri i Reixac 4, 08028 Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 Hypor, a Hendrix Genetics company, Spoorstraat 69, 5831 CK, Boxmeer, The Netherlands 7 Hendrix Genetics Research & Technology Centre, Hendrix Genetics B.V, Spoorstraat 69, 5831 CK, Boxmeer, The Netherlands 8 Departament de Ciència Animal i dels Aliments, Universitat Autònoma de Barcelona (UAB), 08193 Cerdanyola del Valles, Catalonia, Spain 26 8 2016 26 8 2016 2016 17 1 6856 8 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 Taste receptors (TASRs) are essential for the body’s recognition of chemical compounds. In the tongue, TASRs sense the sweet and umami and the toxin-related bitter taste thus promoting a particular eating behaviour. Moreover, their relevance in other organs is now becoming evident. In the intestine, they regulate nutrient absorption and gut motility. Upon ligand binding, TASRs activate the appetite-reward circuitry to signal the nervous system and keep body homeostasis. With the aim to identify genetic variation in the swine TASRs and in the genes from the appetite and the reward pathways, we have sequenced the exons of 201 TASRs and appetite-reward genes from 304 pigs belonging to ten breeds, wild boars and to two phenotypically extreme groups from a F2 resource with data on growth and fat deposition. Results We identified 2,766 coding variants 395 of which were predicted to have a strong impact on protein sequence and function. 334 variants were present in only one breed and at predicted alternative allele frequency (pAAF) ≥ 0.1. The Asian pigs and the wild boars showed the largest proportion of breed specific variants. We also compared the pAAF of the two F2 groups and found that variants in TAS2R39 and CD36 display significant differences suggesting that these genes could influence growth and fat deposition. We developed a 128-variant genotyping assay and confirmed 57 of these variants. Conclusions We have identified thousands of variants affecting TASRs as well as genes involved in the appetite and the reward mechanisms. Some of these genes have been already associated to taste preferences, appetite or behaviour in humans and mouse. We have also detected indications of a potential relationship of some of these genes with growth and fat deposition, which could have been caused by changes in taste preferences, appetite or reward and ultimately impact on food intake. A genotyping array with 57 variants in 31 of these genes is now available for genotyping and start elucidating the impact of genetic variation in these genes on pig biology and breeding. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-2972-z) contains supplementary material, which is available to authorized users. Keywords Taste receptorsAppetite-reward pathwaysCoding genetic variationGenotyping arrayFuture association studieshttp://dx.doi.org/10.13039/501100003329Ministerio de Economía y Competitividad (ES)AGL2013-44978-RClop Alex http://dx.doi.org/10.13039/501100003329Ministerio de Economía y Competitividad (ES)CSD2007-00036http://dx.doi.org/10.13039/501100003329Ministerio de Economía y Competitividad (ES)IPT-2012-0378-060000Clop Alex http://dx.doi.org/10.13039/501100003329Ministerio de Economía y Competitividad (ES)RYC-2011-07763Clop Alex http://dx.doi.org/10.13039/501100004837Ministerio de Ciencia e Innovación (ES)AGL2010-22358-C02-01http://dx.doi.org/10.13039/501100003329Ministerio de Economía y Competitividad (ES)PTQ-12-05391Derdak Sophia http://dx.doi.org/10.13039/501100001825Styrelsen for Forskning og Innovation (DK)UNIK - food, Fitness & PharmaFredholm Merete issue-copyright-statement© The Author(s) 2016 ==== Body Background There are five canonical tastes that are sensed in the taste buds of the tongue, which include salty, sour, sweet, umami and bitter. While salty and sour are detected by ion channels, the other three are sensed by a group of G-protein coupled receptors called taste receptors (TASRs). Sweet and umami are appetizing tastes that characterize energy-rich food sources, namely sugar and amino acid molecules, respectively, and are sensed by the TAS1R sub-type TAS1R1, TAS1R2 and TAS1R3 [1]. On the other hand, the – unpleasant - bitter taste indicating the presence of toxic molecules, is sensed by TAS2Rs, also known as bitter taste receptors [1], which include a variable list of highly polymorphic genes with many species-specific orthologs. The annotation of the pig genome contains ten TAS2Rs according to the Ensembl database (www.ensembl.org). In the recent years, it has become obvious that TASRs are expressed in many other tissues and have additional chemo-sensing functions. For example, they are present in the respiratory system where they regulate innate immunity and infection [2], and in sperm they have been linked to motility and acrosomal reaction [3]. In the gastro-intestinal tract, TASRs detect the molecules that are on transit and stimulate the appetite and reward (AR) circuitries to promote the appropriate feeding behaviour, thus keeping energy balance and body homeostasis [4, 5]. The AR mechanisms are highly interconnected and involve complex networks containing nutrients, neuropeptides, neurotransmitters, hormones and their related receptors and enzymes. These pathways engage the gastrointestinal tract, pancreas, liver, muscle, adipose tissue and brain. Appetite-related genes such as leptin (LEP), leptin receptor (LEPR), cholecystokinin (CCK), Ghrelin (GHRL), Agouti-related protein (AgRP), neuropeptide Y (NPY), proopiomelanocortin (POMC) and melanocortin 4 receptor (MC4R) encode for products that inhibit or excite the dopamine, epinephrine, norepinephrine, serotonin, and glutamate receptor pathways [6, 7] and modulate food intake and energy balance. In a nutshell, ghrelin and LEP are two hormones with opposite excitatory and inhibitory effects on the same neurons secreting appetite inducing NPY and AgRP or the feeding inhibitor POMC. These neuropeptides in turn, inactivate or excite MC4R, which is a hunger repressor. Ghrelin is secreted by the stomach when it is empty and LEP is released by adipocytes as a response to high energy stores [6]. CCK is a hormone secreted in the duodenum as a response to luminal fat and protein and is a strong inhibitor of food intake probably by decreasing gastric emptying and stimulating the vagus nerve [8]. LEP and ghrelin also inhibit and excite dopamine secretion, respectively [9]. Dopamine signalling in certain parts of the brain promotes appetite. Ghrelin secretion during fasting also promotes glutamate release. This neurotransmitter, via a large catalogue of receptors, will also excite NPY, AgRP and inhibit POMC neurons and boost appetite. The catabolic product of glutamate, gamma aminobutyric acid (GABA) also boosts appetite but using different neuronal mechanisms [10]. Moreover, glutamate is also able to excite dopamine-secreting neurons in appetite-relevant areas of the brain thereby indirectly promoting the feeling of hunger [9]. In contrast, GABA can inhibit the same neurons and indirectly promote satiety [9]. Serotonin, a neurotransmitter that is mostly present in the gastrointestinal tract but also at much lower levels in the central nervous system, modulates gastrointestinal motility, mood and appetite [11]. Serotonin inhibits appetite by stimulating its receptors HTR2C and HTR1B, which in turn, activate the well-known appetite-inhibitors POMC and MC4R and inhibit the appetite-promoter genes NPY and AGRP. Epinephrine/norepinephrine are two additional neurotransmitters that seem to be key in food intake and in keeping energy balance and have been shown to respond to starvation and low glucose levels in blood by activating the secretion of ghrelin [12]. In humans, the recent advent of whole genome [13] and exome [14] sequencing has shown that mutations severely impacting on protein sequence are more abundant than previously thought, although due to purifying selection, they tend to have very low allele frequencies. Thus, protein-damaging polymorphisms in TASRs and AR genes are likely to have an important impact on a broad range of traits including feed intake, immune function, behaviour or fertility both in livestock and humans. Consequently, understanding how these variants affect phenotypes in farm animals may both help improving the sustainability of the animal breeding sector as well as benefit bio-medical research. Scientific interest in this gene family in the pig is now emerging and it has recently been shown that TASRs are expressed in multiple porcine systems (immune, gastro intestinal, spermatogenic, etc. [15, 16]). The recent publication of the swine genome sequence and annotation [17] and the development of genome capture assays open unprecedented possibilities to identify deleterious genetic variation. For instance, 295 coding variants in swine TASRs have recently been identified after analyzing the low coverage whole genome sequences of 79 domestic and wild pigs from Europe and Asia [18]. Motivated by their functional relevance in the pig, we have sequenced ten porcine canonical TASRs and 191 AR genes in 304 pigs from multiple breeds in 16 DNA pools with the aim to identify coding polymorphisms and to provide a catalogue of potentially deleterious mutations likely to affect the function of these genes. Moreover, we have also detected indications that some of these variants might be associated with growth and fatness giving new insights in regard to the potential phenotypic relevance of polymorphisms within these genes. Results Sequencing statistics We initially selected 459 kb of target genomic DNA (gDNA) covering the exons from the 12 TASRs and 201 AR genes. After genome capturing, sequencing and read mapping, we successfully covered 372 kb of target sequence with a read depth at each nucleotide position (DP) above 1,000 in the 16 libraries as a whole. This corresponds to 81 % of the initial target size and 201 genes fully or partially sequenced to a DP > 1,000. The poorly sequenced genes include TAS2R3, ENSSSCG000000029894 (a swine ortholog of human TAS2R16), and 10 AR genes. The list of successfully sequenced genes is shown in Additional file 1. After genome capture, sequencing and read mapping, 162,848,637 reads mapped to the target gDNA regions. Variant identification in TASR and AR genes We successfully sequenced (DP ≥ 1,000) 14,598 bp (94.5 % of the initial selection) of TASR exons and identified 219 coding variants in TASR exons, 113 of which (52.1 %) do not have a dbSNP identifier and are thus considered novel (Additional file 2). Two TASR variants had the alternative allele fixed in the 16 pools and are thus likely to be either errors or private variants in the Duroc animal used to generate the reference sequence. We excluded them from our list of putative polymorphisms. Ten of the 217 remaining variants were classified by snpEff to have a high impact (H) on the coding sequence and consequently, on the function of the gene according to the gene annotation in the swine genome (Table 1). These include four single nucleotide variants (SNVs) and six short indels, which cause three stop-codon gains, one stop-codon loss, five frame-shift, and one novel splice-site donor. These variants affect four TASRs with a clear over-representation in TAS1R1 (Table 1). Three of these variants affecting TAS1R1 and TAS1R3 had a predicted minor allele frequency (pMAF) ≤ 0.01 (Tables 2 and 3). In addition, we identified 125 non-synonymous-coding variants and one codon-deletion, which are classified by snpEff as having a moderate (M) impact. Thirty-four of the non-synonymous changes were predicted to be deleterious (Mdel) by SIFT [19] (Table 1). The remaining M variants were either SIFT predicted as tolerated (Mtol) or did not yield any prediction. Hence, we have identified 44 variants (10 H and 34 Mdel) that are likely to have an important effect on swine TASR function. Remarkably, all TASRs showed H or Mdel variants (Additional file 2). Finally, 81 variants were predicted to be synonymous changes with no apparent impact (L) on the subjacent proteins (Table 1). On average, the variants with strong impact on protein sequence (H and Mdel) were predicted to be rarer in the species than those having a mild impact (Mtol and L) (Table 2) according to pMAF.Table 1 Number of variants across the TASR and AR gene groups per each impact class High impact Moderate impact Low impact Gene Splice Stop gained Stop lost Start loss Frame shift Mdel Mtol/SIFT unknown Silent Start gained % strong impact Total Total TASRs 1 3 1 0 5 34 92 80 1 20.3 % 217  TAS1R1 (umami) 1 1 1 0 3 4 9 12 0 32.2 % 31  TAS1R3 (sweet and umami) 0 1 0 0 0 1 0 11 1 14.3 % 14  TAS2Rs (bitter) 0 1 0 0 2 29 83 57 0 18.5 % 172 AR 37 16 0 4 17 277 615 1,559 24 14.5 % 2,549 Total segregating 38 19 1 4 22 311 707 1,639 25 2,766 Alternative allele fixed 4 0 0 0 4 0 9 9 1 27 Total 42 19 1 4 26 311 716 1,648 26 2,793 Splice: variants predicted to alter either donor or acceptor splice sites; % strong impact: percentage of H + Mdel variants Table 2 Variant distribution per effect and pAAF within each gene group Variant frequency class Strong impact Mild impact TASR Very rare (pMAF < 0.01) 28 (63.6 %) 51 (29.5 %) Rare (pMAF = [0.010–0.019]) 6 (13.6 %) 28 (16.2 %) Common (pMAF = [0.020–0.979]) 10 (22.7 %) 94 (54.3 %) Total number 44 173 pMAF min-max (average) 0.0016–0.1910 (0.028) 0.0016–0.4920 (0.091) AR Very rare (pMAF < 0.01) 184 (52.6 %) 560 (25.5 %) Rare (pMAF = [0.010-0.019]) 41 (11.7 %) 235 (10.7 %) Common (pMAF = [0.020–0.979]) 125 (35.7 %) 1,403 (63.8 %) Total number 350 2,198 pMAF min-max (average) 0.0017–0.4800 (0.029) 0.0016–0.4980 (0.0992) The percentages are for the total number of variants within the groups: TASR strong impact, TASR mild impact, AR strong impact and AR mild impact Table 3 List of rare H variants Variant ID Effect Gene pMAF chr12_15398873_C_T STOP_GAINED ACE 0.0017 chr12_62594159_A_G START_LOST ALDH3A2 0.0086 chr7_103136276_A_G START_LOST ALDH6A1 0.0017 chr7_103129739_C_G SPLICE_SITE_ACCEPTOR ALDH6A1 0.0034 chr9_110061864_G_A STOP_GAINED CD36 0.0017 chr7_28072405_A_C SPLICE_SITE_DONOR NOTCH4 0.0017 JH118674.1_43285_C_T SPLICE_SITE_DONOR GRIA1 ortholog (ENSSSCG00000024560) 0.0049 chr16_66618439_A_T START_LOST GABRG2 0.0091 chr16_66618438_C_A,T START_LOST GABRG2 0.0017 chr1_64002416_A_G SPLICE_SITE_DONOR GABRR1 0.0034 GL896494.1_7864_G_T STOP_GAINED GPR179 0.0052 chr13_203412188_T_A SPLICE_SITE_ACCEPTOR GRIK1 0.0058 chr1_77441762_G_A SPLICE_SITE_ACCEPTOR GRIK2 0.0034 chr6_49600987_C_T SPLICE_SITE_DONOR GRIN2D 0.0034 chr6_49600986_A_G SPLICE_SITE_DONOR GRIN2D 0.0034 chr13_37136736_G_T SPLICE_SITE_ACCEPTOR GRM2 0.0017 chr13_37136747_T_G SPLICE_SITE_DONOR GRM2 0.0034 chr7_34927184_A_C SPLICE_SITE_DONOR GRM4 0.0017 chr7_34893335_T_G SPLICE_SITE_ACCEPTOR GRM4 0.0017 chr18_3214922_AC_A FRAME_SHIFT HTR5A_human_ortholog (ENSSSCG00000030573) 0.0030 chr18_3151609_AG_A FRAME_SHIFT HTR5A_human_ortholog (ENSSSCG00000023549) 0.0020 chr3_55922763_C_T SPLICE_SITE_ACCEPTOR NPAS2 0.0020 chr1_14768986_T_G SPLICE_SITE_ACCEPTOR OPRM1 0.0017 chr5_85218678_C_A STOP_GAINED PAH ortholog (ENSSSCG00000000856) 0.0017 chr5_85218708_C_A STOP_GAINED PAH ortholog (ENSSSCG00000000856) 0.0017 chr5_85218747_G_T SPLICE_SITE_DONOR PAH ortholog (ENSSSCG00000000856) 0.0017 chr3_23446168_C_T STOP_GAINED SCNN1G 0.0017 chr16_85870363_G_C SPLICE_SITE_DONOR SLC6A3 0.0017 chr16_85870364_T_A SPLICE_SITE_DONOR SLC6A3 0.0017 chr6_58114941_G_A STOP_GAINED TAS1R3 0.0016 chr6_62357984_C_T STOP_GAINED TAS1R1 0.0065 chr6_62363203_G_C STOP_LOST TAS1R1 0.0065 The variant identifier (ID) contains information on chromosome _ position _ reference allele _ alternative allele Likewise, we sequenced 357,844 (80.1 % of the initial selection) bp of exonic sequence from AR genes at a DP > 1,000, and identified 2,570 variant positions in AR exons, most of which were bi-allelic SNVs. 1,092 (42.4 %) variants were not annotated in dbSNP (Additional file 2). Four of these positions were multi-allelic with each allele predicted to have a different effect on protein sequence. Of the 2,574 assigned variant effects, 25 had the alternative allele fixed in the 16 pools. From the remaining 2,549 variants, 350 were classified as deleterious (H and Mdel) and affected 123 genes (Table 1 and Additional file 2). As for TASRs, H and Mdel as a whole tended to have lower pMAF than Mtol and L (Table 2). We identified 74 H variants in AR genes, 29 of which had pMAFs ≤ 0.01 and are thus considered rare. These H rare variants affected 21 genes (Table 3). We also plotted the distribution of the AR gene variants predicted to have a strong impact on protein (H and Mdel) and those with a mild impact (Mtol and L) along the protein sequence divided in 10 consecutive position bins of equal amino acid length. We noted that the strong impact variants tended to be more abundant at the end of the protein (bin9 + bin10), and that these also tended to have, on average, larger pMAF. This trend was not observed in the set of mild impact variants (Fig. 1).Fig. 1 Distribution and pMAF of strong (H + Mdel) and mild (Mtol + L) impact variants along the protein body of AR genes. The bar plots for (a) the strong impact effect variants were made with 310 premature stop, frame-shift and Mdel. The barplot for the mild impact effect variants (b) were made with 2,174 Mtol, in-frame indels and synonymous variants. The dots indicate the average pMAF in each of the 10 position percentile bins. Percentile bins divide the protein body of each gene in ten portions of equal size Per breed variant distribution We compared the ten purebred pools and observed that, as expected, larger pools contained more variants (both in TASR and AR genes). Nonetheless, the two Duroc pools together, with 45 samples, displayed less genetic diversity (1,042 variants), than the Large White, Landrace and the Pietrain, which had a similar number of animals and were above 1,300 variants each (Table 4). The Asian pool harboured the largest number of variants (1,589 variants with a pool size = 22) while the Iberian pool was ranked as the most homogenous (746 and pool size = 13). Overall, we observed that the most ancient breeds, Mangalitza, Iberian, Majorcan Black were less variable in our set of genes (Table 4). Given that a higher number of samples seem to be correlated to a higher number of variants, we also compared the genetic variability using the Watterson estimate [20] at the neutrally evolving synonymous sites (the third nucleotide position within each codon), which corrects the number of mutated sites by the number of animals included in the analysis. This value (Bazna: 0.00118, Mangalitza: 0.00101, Majorcan Black: 0.00103, Iberian: 0.00095, Duroc: 0.00090, Pietrain: 0.00131, Landrace: 0.00130, Large White: 0.00130, wild boar: 0.00109 and Asian: 0.00182) showed that all the populations had similar variability with the exception of the Asian pool, which genetic variability doubled that of the other breeds.Table 4 Number of variants and number of unique variants within each breed and per impact class Duroc Pietrain Large white Landrace Bazna Mangalitza Iberian Majorcan black Wild boar Asian Variant effect nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique nr (%) nr unique Strong impact 133 (12.7) 1 144 (10.9) 1 140 (10.5) 1 124 (9.4) 0 99 (10.2) 2 87 (11.3) 4 92 (12.3) 2 90 (10.4) 1 123 (12.4) 6 163 (10.2) 26 High 44 (4.2) 1 46 (3.5) 0 49 (3.7) 0 33 (2.5) 0 28 (2.9) 0 29 (3.8) 1 30 (4.0) 1 30 (3.5) 0 35 (3.5) 0 48 (3.0) 9 Moderate_Deleterious 89 (8.5) 0 98 (7.4) 1 91 (6.8) 1 91 (6.9) 0 71 (7.3) 2 58 (7.5) 3 62 (8.3) 1 60 (6.9) 1 88 (8.9) 6 115 (7.2) 17 Mild impact 909 (87.3) 5 1177 (89.1) 5 1194 (89.5) 5 1195 (90.6) 3 868 (89.8) 11 685 (88.7) 8 654 (87.7) 9 778 (89.6) 21 870 (87.6) 32 1426 (89.7) 195  Moderate_Tolerated 279 (26.8) 4 342 (25.9) 1 332 (24.9) 2 342 (25.9) 1 257 (26.6) 2 192 (24.9) 2 187 (25.1) 5 223 (25.7) 9 252 (25.4) 15 389 (24.5) 50 Low 630 (60.5) 1 835 (63.2) 4 862 (64.6) 3 853 (64.7) 2 611 (63.2) 9 493 (63.8) 6 467 (62.6) 4 555 (63.9) 12 618 (62.2) 17 1037 (65.3) 145 Total 1042 6 (0.6 %)a 1321 6 (0.5 %)a 1334 6 (0.4 %)a 1319 3 (0.2 %)a 967 13 (1.3 %)a 772 12 (1.6 %)a 746 11 (1.5 %)a 868 22 (2.5 %)a 993 38 (3.8 %)a 1589 221 (13.9 %)a Pool size 45 41 39 40 15 12 13 17 22 22 This table was done using the 2,574 variants in TASR and AR genes including the multi-allelic with different effects and these with the alternative allele fixed in all the populations aPercentage of variants that are breed-specific We also compared the percentage of variants that were H or Mdel per breed. No obvious differences were observed (p-val = 0.07) and the Duroc and Landrace were at the top (12.7 %) and bottom (9.4 %) ends, respectively (Table 4). Altogether, 26 TASR variants were present in a single breed at pAAF ≥ 0.1 and might thus be breed-specific. Not surprisingly, the Asian pool, which involved 15 Chinese Meishan and 7 Vietnamese pigs, showed the highest proportion of pool-specific variants, with 12 being present only in this group (Additional file 3). The wild boar also displayed several allelic particularities, with nine unique variants. Noteworthy, seven of these wild boar variants, all with similar pAAF, mapped to TAS2R1 (Additional file 3). Overall, five (one H and four Mdel) breed-specific variants were predicted to have a strong impact on protein sequence (Additional file 3). The H variant is a stop gain in TAS1R1 that is present in 17 % of the Mangalitza genomes, respectively. We also identified two variants, a synonymous (L) and a non-synonymous tolerated (Mtol) SNVs that affect TAS2R1, that whilst being present at very high frequencies or even fixed in all breeds (pAAF ≥ 0.5), were absent in the Asian pool (Additional file 3). We detected 306 AR coding variants that were uniquely present or absent in one breed (Additional file 3). Of these, 35 variants involving 32 genes were predicted to be of functional importance (Additional file 3). As in TASRs, the breed specific H and Mdel variants in the AR genes were more abundant in the Asian and the wild boar with 20 and five unique features, respectively. It was noteworthy that two of these H and Mdel variants, both in the Asian pools, were close to fixation (pAAF ≥ 0.9). These variants map to the serotonin receptor, HTR3C, and to the cytochrome P450 gene, CYP2A6 (Additional file 3). We performed hierarchical clustering using an Unweighted Pair Group Method with Arithmetic Mean (UPGMA) based on the 2,523 (217 TASR and 2,306 AR) variants that were present in at least one of the ten breeds and with pAAF information available in all these populations. One-hundred variants were excluded from the purebred comparison since they were unique to the F2 animals. The resulting phylogenetic dendrogram is in line with what has been shown in other studies (Fig. 2). Briefly, the Western (European and USA) breeds cluster together and the Asian pool form a separate branch. Within the Western cluster, the Duroc is the only member of a distant branch whilst the European commercial breeds (Large White, Landrace and Pietrain) belong to another sub-group and the more ancient breeds Iberian, Majorcan Black, and Mangalitza cluster together with the wild boar.Fig. 2 Phylogenetic dendrogram with the 10 breeds. The numbers indicate the support for each node according to 1,000 bootstrap iterations pAAF and phenotype relationships in the F2 groups We also wanted to see whether we were able to detect an indication of an effect of the variants on production traits. This was investigated by comparing the pAAFs of two F2 pools from the same experimental population, each belonging to one of the tails of the phenotypic distribution, for average daily gain and retroperitoneal fat content. Out of the 217 TASR coding variants, 97 segregated in the F2 resource, and within these, eight displayed significantly different pAAFs in five TASRs (Table 5 and Additional file 4). Remarkably, four TAS2R4 variants showed significant differences between the pools. Likewise, we could compare 1,280 variants in AR genes segregating in this population and identified 56 significant differences (p-val ≤ 0.05) involving 25 genes (Table 5 and Additional file 4). After correcting for multiple testing (p-val ≤ 0.05/(97 + 1,280) = 0.00003), one M variant in TAS2R39 and three variants in CD36 remained significant (Table 5).Table 5 Genes with variants showing significantly different pAAF between F2_F and F2_L Gene Total number Number and type of variants with significant differences is pAAF Fisher Test value range (min-max) ADRB1 1 1 synonymous 0.0467 ALDH1B1 3 3 synonymous 0.0253 ALDH2 1 1 non-synoymous tolerated 0.0358 ALDH3B2 7 1 non-synoymous deleterious; 6 synonymous 0.0003–0.0137 ALDH6A1 1 1 synonymous 0.0253 CD36 3 1 frame-shift; 1 non-synoymous deleterious; 1 synonymous 1.5xE-05a - 2.58xE-05a DISC1 3 3 synonymous 0.0280–0.0357 TVPR1 human ortholog (ENSSSCG00000017863) 5 1 non-synoymous deleterious; 4 synonymous 0.0111–0.0315 ALDH8A1 ortholog (ENSSSCG00000023457) 1 1 synonymous 0.0387 FOS 2 1 non-synoymous tolerated; 1 synonymous 0.0284 GABRA3 1 1 splice-site donor/acceptor 0.0253 GABRA6 1 1 synonymous 0.0047 GPR179 9 2 non-synoymous deleterious; 4 non-synoymous tolerated; 3 synonymous 0.0137–0.0324 GPRC5B 1 1 non-synoymous deleterious 0.0253 GPRC5C 1 1 non-synoymous tolerated 0.0258 GRM1 3 3 synonymous 0.0047–0.0383 GRM8 1 1 synonymous 0.0178 HTR1B 1 1 synonymous 0.0253 HTR3C 1 1 synonymous 0.0357 LEPR 1 1 synonymous 0.0253 MCHR2 2 2 synonymous 0.0115–0.0253 MTNR1B 1 1 non-synoymous tolerated 0.0357 P2RX2 2 2 synonymous 0.0094–0.0324 P2X7 3 3 synonymous 0.0006–0.0324 SIM1 1 1 synonymous 0.0115 TAS2R16 ortholog (ENSSSCG00000016433) 1 1 synonymous 0.0315 TAS2R39 1 1 codon change 6.16xE-08a TAS2R4 4 2 non-synoymous tolerated; 2 synonymous 0.0002–0.0158 TAS2R41 1 1 non-synoymous tolerated 0.0207 TAS2R60 1 1 synonymous 0.0351 Total 64 aVariants with Fisher test values remaining significant after Bonferroni correction for the multiple testing for 1,377 variants (p-val ≤ 0.05/(1,377)) Selection and genotype-based validation of a set of protein-damaging variants With the double aim to validate the subset of variants with more likely damaging impact and, to evaluate by genetic association, their impact in pig breeding, we designed a TaqMan genotyping assay targeting 128 putative polymorphisms. We originally aimed at including all the H variants found in the study and to fill the remaining of the assay with several M variants to cover all TASRs and several AR genes. However, the assay had design requirements that not all the variants fulfilled. Hence, we could only include nine H variants in TASR, 69 H variants in AR genes, 33 M variants in TASRs and 17 M variants in AR genes. To validate the selected variants, we genotyped 237 (Additional file 5) of the 304 sequenced pigs. We could not genotype the remaining 67 pigs as all their DNA was used for the sequencing step. 9 % of the assays failed to either amplify or clearly differentiate the genotype clusters, and 46 % did not show the alternative allele in any sample. 57 variants identified in our variant calling pipeline were validated by the TaqMan assay. The 57 polymorphisms (3 TASR H, 12 AR H, 31 TASR M and 11 AR M) affect eight TASRs and 23 AR genes (Additional file 6). All genotyped variants showed to be in Hardy-Weinberg Equilibrium across breeds (data not shown). Although none of the four genotyped rare polymorphisms (all H) showed a homozygote state for the minor allele, one H SNV (chr16_45548628_G_A), unique to the Vietnamese pigs and with a MAF slightly above the rare variant threshold (MAF = 0.013), was found in the homozygous state in two animals. This variant was predicted to cause a stop codon gain in the very last amino acid of the canonical HTR1A protein. Another introduction of a stop codon (chr8_38038074_G_A) was found in the middle of the Gamma Aminobutyric Acid receptor GABRG1, and was present only in Meishan pigs showing a MAF = 0.022. We next compared the predicted and the matched genotype-observed frequency of the minor allele and detected a strong correlation (r2 = 0.93) between the two (Fig. 3).Fig. 3 Linear regression and correlation between pMAF and the genotype observed MAF in the 57 genotyped polymorphisms We also carried a genetic association analysis with the genotypes and phenotypic values for retroperitoneal fat and daily gain in the F2. 36 of the 57 variants segregated in the F2 animals. The array contains two of the variants (chr2_3566521_A_C and chr18_7019623_G_A) that showed significant pAAF differences between the F2 groups with the Fisher test. For daily gain, only the marker chr2_3566521_A_C was nominally significant (p-val = 0.025) (Additional file 7) but it did not reach significance after Bonferroni correction for multiple testing (p-val ≤ 0.05/36 = 0.0013). For retroperitoneal fat, two markers, chr9_46397500_A_G and chr18_7019387_A_T, were nominally significant (p-val = 0.028 and 0.044, respectively) but again, none reached statistical significance after Bonferroni correction (Additional file 7). Chr18_7019623_G_A, which was significant at the pAAF comparison between both F2 groups, did not reach significance for any of the two traits (p-val = 0.09 and 0.1 for daily gain and retroperitoneal fat, respectively). Discussion TASRs and the components of the AR circuitry are key genes in keeping body homeostasis as they recognize chemical molecules that could be both sources of energy or threatening toxins and promote an adequate response. With the aim to characterize the coding genetic variation affecting swine TASRs and the AR circuitry genes, we sequenced 16 gDNA pools corresponding to 304 pigs from ten breeds and European wild boar and from two pools of an experimental F2 population with records on growth and retroperitoneal fat content. We have mapped thousands of coding region genetic variants, hundreds of which are expected to have a strong impact on protein sequence, some of which are breed specific. By comparing the pAAFs of these variants in two F2 pools divergent for growth and fat deposition, we also identified many genotype - phenotype relationships. Our data provides detailed information of the genetic variation present in TASRs and AR genes. We also developed an assay to genotype 128 of the most functionally relevant variants which is available to perform association studies with relevant traits in pig populations. Technical considerations Differences in DNA extraction methods and inaccuracies with respect to quantification might have a negative impact on the even distribution of sequencing reads among the genomes within a pool thereby reducing the accuracy of the pAAF/pMAFs as a proxy of the real allelic frequency. Nonetheless, the comparison of predicted versus observed allele frequencies in the 57 genotyped variants showed that pAAF/pMAFs were very good predictors and supply additional confidence of the quality of our results (Fig. 3). To the best of our knowledge, this is the first published study from a targeted genome enrichment experiment in swine. As this approach reduces the sequencing throughput requirements, we were able to sequence the target sub-genome (the exome of 201 TASRs and AR genes) in the largest number of pigs (n = 304) reported to date, in a single experiment. By mapping nearly 163 million reads on the target exons, we reached an average depth of sequencing of 72× for each of the haploid genomes. This read depth allowed us to detect all variants present in the pools regardless of their frequency. Given that we sequenced 304 animals harbouring in total 608 chromosome sets (2 alleles each), we were able to detect rare variants with MAF ≥ 0.0016. No previous high throughput sequencing study in pigs reached this sensitivity to detect rare variants. We identified a large number of variant events, most mapping to exon flanking regions including promoters, introns, and upstream and downstream segments. Although some of these non-coding variants could be functional, their regulatory relevance is difficult to predict and was not the subject of our study, which focused mainly, on the detection of protein-damaging coding variation. Variant identification Overall, we have identified 2,793 coding variants but 27 of these are fixed in our sample set and are thus likely to be either errors in or private to the reference sequence. We annotated 217 variants segregating in the 10 canonical swine TASRs in our samples. Da Silva et al. [18] described 279 coding variants in their list of 21 taste and nutrient receptor genes after sequencing 79 pigs. However, Da Silva and co-authors included 8 genes that we did not study. These are 5 likely-paralogs of canonical TASRs and 3 genes that are not canonical TASRs but that are related to tasting fat (GPR41, and GPR84) and amino acids (GPRC6A). On the contrary, we included 2 canonical TASRs (TAS2R4 and TAS2R40) not studied by Da Silva et al. Moreover, 4 we classified as AR genes (GPR40, GPR120, CASR, GRM1 and GRM4) that are shared in both studies as they are not canonical TASRs or are receptors belonging to the glutamate pathway. 138 of the 279 variants, 71 in canonical TASR and 67 in the other shared genes, are common to both studies. The difference in the catalogue of identified variants is likely to be mostly due to the fact that Da Silva et al. studied a different set of animals, some belonging to breeds we did not target (Hampshire, creole, Brazilian and Tamworth). 395 of the 2,766 “segregating” variants are predicted to have a high impact (H and Mdel) on the protein sequence of 133 genes and are thus very likely to disrupt or strongly alter their function (Table 1 and Additional file 2). Some of these variants showed allelic frequency relationship with daily gain and retro-peritoneal fat deposition in the F2 resource with phenotypic records (Table 5 and Additional file 4). In keeping with previous findings in human [13], we also detected that H and M variants tend to be more abundant and have higher pAAFs in TASRs than in non-sensory genes (Tables 1 and 2) which indicates that TASRs are subject to balancing selection. TASRs are comparable to the major histocompatibility complex genes (HLA in humans and SLA in pigs) as both are expressed at the surface of cells to detect particular molecules that could be hazardous for the body. Whilst HLA and SLA detect antigens to promote an immune reaction, TASRs sense chemical compounds to stimulate appropriate responses (reward, acrosomal reaction, smooth muscle contraction or immune function among others) depending on the cell type that is involved. Thus, a healthy animal population needs to be highly polymorphic in these genes in order to maximize its adaptability and survival to variable environments facing multiple threads. As expected, H and Mdel variants had on average, lower pMAFs than Mtol or L variants (Table 2), as the former are more likely to be deleterious and consequently, subject to purifying selection. This is well exemplified by the recent exome sequencing experiments that have successfully identified rare deleterious variants causing rare and severe mendelian disorders in humans [14]. Indeed, H and Mdel variants are more abundant and frequent at the 3′-end of the protein, where they are less likely to have an impact of the function of the protein (Fig. 1). H rare variants Remarkably, we found 32 rare H variants which mapped to TAS1R1, TAS1R3 and 20 AR genes (Table 3). These mutations are predicted to fully disrupt the function of the affected gene and might thus have an impact on taste and AR perceptions. Eight of the 20 AR genes carrying rare H variants, have been directly associated to taste, food intake, body size, diabetes or triglyceride levels. Two of them, GRM4 and GABRR1, are glutamate receptors that have been associated to body size [21] and feeding behaviour [22], respectively. Natural knock-out pigs, (i.e., pigs homozygous or compound heterozygous for rare H variants) in these genes may have severe consequences in taste perception and feeding attitude. Reverse genetics experiments to study the phenotypic changes that occur in such natural knock-outs would be very informative to understand the importance of these mutations on feeding behaviour and on a broad range of other traits of relevance in animal breeding and bio-medical sciences. However, the identification of such homozygotes may require the screening of thousands of pigs. A convenient alternative would be to identify heterozygous animals and cross them to generate these natural knock-outs. Breed particularities We found several variants with breed particularities that could both explain in part the population history of these breeds and be the result of genetic adaptation to the particular environment or artificial selection. Overall, the pools of commercial breeds (Large White, Landrace, Pietrain, Duroc) contained more variants than the pools of the ancient/traditional breeds (Iberian, Majorcan Black, Mangalitza, Bazna). The sequencing of the commercial breeds involved a larger number of samples and consequently, more genetic variation could be captured. This difference is possibly due also to first, population bottlenecks and a low effective population size within ancient breeds and second, to the introgression of Asian germplasm in the European commercial breeds [23, 24] (Table 4). The Duroc pool is somehow in between European commercial and traditional breeds, with a relative low number of variants (Table 4). However, this was expected as the pig reference genome sequence was obtained from a Duroc animal and is therefore expected to share more similarities with our Duroc pool. On the other hand, the Asian breeds showed the largest number of variants (Table 4) which was also expected as our pool was made of two different swine populations (15 Meishan and 7 Vietnamese pigs) and also, these animals diverged from the European counterparts around one million years ago [25]. It is worth pointing out that as we merged the Meishan and the Vietnamese pigs, we cannot have specific information for these populations. Although the comparison of both Asian breeds would have been interesting, our aim was not to study the genetic variation within these breeds but to identify a large catalogue of variants by sequencing a set of divergent populations. For this reason, and to optimise the high throughput sequencing resources, we merged all Asian animals. Although we tried to minimise co-ancestry between samples, we do not have pedigree data and hence, we cannot exclude the possibility that the trend on the reduction in the genetic diversity observed in some breeds is due to close familiar relationships. This is particularly true for our Mangalitza samples, which come from a highly inbred population from Romania and might thus not represent the genetic diversity existing within the Mangalitza breed as a whole. To better determine genetic diversity within each breed, we compared the genetic variability in the synonymous sites, which are considered to be neutral in evolutionary terms, in our TASR and AR genes using the method described by Watterson. This method corrects the number of variants by the number of individuals that were analysed [20]. We further corrected these values by the expected number of silent sites in the sequenced cds. Genetic variability among western breeds is highly similar (0.00090–0.00131 variants per mutant site) being lowest in Duroc and highest in Pietrain. The genetic variability seems to be higher in the Asian pool (0.00182). We also compared the relative abundance of H + Mdel variants in each breed. In contrast to the recent results by Bianco and co-authors [26] who detected a higher ratio of deleterious variants in Western than in Asian breeds, we did not identify large differences across pools. This inconsistency could be explained by the fact that we interrogated a particular set of genes whilst Bianco et al. assessed all annotated genes in the pig genome. Also, both studies screened different animals from distinct breeds. Furthermore, the two studies used different sequencing strategies. Whilst we deep-sequenced 304 pigs in a pool-based strategy, Bianco et al. performed low depth whole-genome shotgun sequencing of 128 pigs. Our approach is better suited for the identification of rare variants and this difference could have altered the catalogue of variants identified in both studies. Therefore, the comparison of the two datasets needs to be taken with caution. Of note, we observed that 25 % of the 28 variants identified in TAS2R1 were wild boar specific (Additional file 3). This could indicate a particular haplotype that might have been lost in the European domestic breeds by artificial selection. Alternatively, as our wild boar gDNA pool was made with samples from three different European locations (Catalonia, Belgium, Romania), these variants, which have a relatively low pAAF [0.16–0.18], could well belong to one of these populations with no germplasm contributed to the analyzed domestic breeds. Noteworthy, some of these TAS2R1 variants are predicted to have a strong effect on the protein and could thus indicate adaptive selection to particular foods or environments (Additional file 3). Especially relevant are the breed-specific H variants with relatively high frequency in the affected breed (pAAF > 0.1) (Additional file 3). The stop codon that prematurely truncates TAS1R1 right in the middle of the protein in 17 % of the Mangalitza genomes might impair the ability to sense umami and might thus affect food preferences in a similar way as described in giant pandas, which lack a functional TAS1R1 potentially disrupting preferences for protein-rich sources [27]. Furthermore, two H variants in HTR3C, and CYP2A6 are unique and almost fixed in the Asian breeds (Additional file 3). As the pAAF between Asian and European pools are so dramatically different, we hypothesize that this might reflect an AR adaptation to very different environments. Variant and phenotype relationships In our study, we aimed not only at identifying damaging variants but also at checking whether we were able to detect potential relationships with production traits. This comparison was done with only 38 pigs, which would typically be a very small number of animals and as such, the power to detect genetic associations is low. Thus, it does not aim at finding statistically significant genetic associations but at detecting a trend that could indicate this association. Indeed, we identified several allele - phenotype relationships that could indicate real genetic associations (Table 5). We believe that comparing the pAAFs of the two F2 pools was a good strategy for the identification of allele - phenotype relationships, as these animals share a common genetic background and the only criteria used to make the pools, was their extreme and opposed phenotypic characteristics. Four of the 11 (36 %) TAS2R4 variants segregating in our F2 resource showed significant differences at the nominal level in pAAF between the obese and fast growing F2_F pool and the lean slow growing F2_L pool (Table 5). As this is a large percentage, we believe that this could be a real association and that a polymorphism, perhaps a regulatory variant not assessed here, is in part responsible of the phenotypic differences between the divergent groups of pigs. Five of the six genes with significant allelic differences at the nominal level (all with p-val ≤ 0.01) between the F2 pools have been directly linked to both taste and growth (umami taste and GRM1 [28], fat taste and CD36 [29], taste in general and P2RX2/P2RX7 [30] and weight gain and GABRA6 [31]) (Table 5) and could thus indicate a difference in the eating behaviour. This would explain the difference in growth and fat deposition between the F2 groups. Remarkably, three variants in CD36, a gene associated to fat taste, had significant pAAF differences between the high and the low fat deposition pigs after correction for multiple testing (Table 5). Two of the markers, chr2_3566521_A_C and chr18_7019623_G_A, in the ALDH3B1 ortholog ENSSSCG00000026349 and in TAS2R41, respectively, were also genotyped in the F2 animals with TaqMan probes using the OpenArray technology. Only chr2_3566521_A_C was nominally significant for daily gain (p-val = 0.025), but did not reach the significance threshold after multiple testing (Additional file 7). chr18_7019623_G_A was not significant, but a marker 236 bp upstream from this SNP, chr18_7019387_A_T, was nominally significant (p-val = 0.044) for retroperitoneal fat content (Additional file 7). We seek to screen larger populations with phenotypic data to determine whether these associations could become significant. Selection and genotype-based validation of SNPs We also developed a genotyping array to validate the variants with highest damaging potential regardless of their frequency. We believe that a proportion of the 59 non-polymorphic positions might be real low-frequency polymorphisms that were present only in animals that we did not genotype due to the lack of available gDNA. However, we cannot rule out the possibility that some of these variants could simply be false positives caused by the erroneous mapping of some reads to highly similar regions. For example, none of the 17 frame-shift variants we included could be confirmed as a polymorphism but the high false positive rate among indels in high throughput sequencing experiments is well documented. Most of the H variants failed to proof polymorphic but they had very low pMAFs whilst most of the M variants, which on average showed higher pMAFs, were confirmed (Additional file 6). None of the rare H polymorphisms displayed the homozygous state for the minor allele in the genotyped samples. However, we identified two H variants that approached the rare MAF threshold, which showed the three genotypic classes in our animals (Additional file 6). The Vietnamese rare variant in HTR1A is a stop codon that was homozygous in two Vietnamese pigs. Nonetheless, we cannot consider these two animals as natural knockouts as this mutation is located at the very last amino acid of the canonical HTR1A protein, and it is not expected to have an impact on HTR1A function. Therefore, the low frequency of the minor allele might not reflect the existence of purifying selection forces. The Meishan H variant in GABRG1 is also an introduction of a stop codon, but this is located in the middle of the gene and is thus likely to disrupt its function. This gene increases neuronal activity and is associated to eating disorders and anxiety [32] and even with alcohol dependence in humans. Thus, natural GABRG1 knock-out pigs carrying this stop codon may show both signs of anxiety and a particular eating behaviour. We have developed the first assay for massive genotyping of variants with high functional potential in swine TASR and AR genes, and we now seek to perform association analysis on pig populations with phenotypic records for eating attitude, feed intake, growth, obesity, but also on semen quality and fertility, infection and immunity and behaviour abnormalities such as stress-related stereotypies and tail biting. This approach will help us understanding the impact of genetic variation in TASR and AR genes on traits of interest for the pig breeding industry. This genotyping assay will also allow us to identify natural knock-outs that could then be used in reverse genetics studies. Nonetheless, we acknowledge that this array tags a very small proportion (eight TASRs and 23 reward) of the genes involved in taste, appetite and reward. Ideally, the list of variants and genes should be extended to achieve a comprehensive analysis of the impact of these gene pathways in pig breeding. Conclusions We detected 2,766 variants predicted to have a potential (high, moderate or low) impact in the protein sequence of 201 TASR and AR genes. Of these, 395 were predicted to strongly impact on the protein sequence of the 10 TASRs and 123 AR genes and consequently, in their function. The importance of these genes in many traits contributing to body homeostasis has been well documented in human and animal models but remains unexplored in livestock. We have found significant relationships between the pAAF of some variants and growth and fat deposition. This, albeit not more than a mere indication, strongly encourages further studying the effect of these genes on traits of interest in body homeostasis and animal breeding. For this reason, we have developed a genotyping array with a subset of these variants and have validated 57 by genotyping the initially sequenced animals. This array is now ready to be used in genetic association studies for relevant traits including taste preferences, food intake, fertility or behaviour. Although this array is not comprehensive and does not contain all the variants that we identified, it contains a careful selection of the most likely deleterious variants and involves eight TASRs and 23 AR genes. Methods Selection of target regions The genomic regions of TASR exons were selected using the genome annotations accessible via Ensembl’s Biomart (www.ensembl.org/biomart; version 72, June 2013). We selected two TAS1Rs and the 10 annotated TAS2R genes (Additional file 1). TAS1R2 could not be included due to its unknown genomic location at the time of selection. In total, 166 genes from the AR circuitries were selected from a study aimed at understanding the genetic signatures in giant panda that confer the highly selective diet based on bamboo only [33]. In their study, these authors retrieved the 166 genes from 4 review articles on appetite and food intake behaviour as described in their material and methods. In addition, 35 AR genes were identified by searching NCBI’s PubMed database using the keywords: “appetite”, “food intake”, “dopamine”, “serotonin”, “glutamate receptor”, “epinephrine”, “norepinephrine”, “reward”. We used Ensembl’s Biomart to identify and select the genomic coordinates of the exons from the swine AR orthologous genes (Additional file 1). Samples We used gDNA from 266 pigs belonging to eleven breeds or populations including Large White (United Kingdom), Landrace (Denmark), Pietrain (Belgium), Duroc (United States), Iberian (mainland Spain and Portugal), Majorcan Black (Balearic islands), Bazna (Romania), Mangalitza (Hungary), Meishan (China), Vietnamese pot-bellied and to the European wild boar. The Large White, Landrace, Pietrain and Duroc have been subjected to strong selection pressures in the last decades and at present are the most commonly used in intensive production systems. On the contrary, Iberian (mainland Spain and Portugal), Majorcan Black (Balearic islands), Mangalitza (Hungary, although our animals come from a closed herd in Romania) and Bazna (a Romanian breed obtained in the 19th century by crossing Mangalitza and Berkshire) are local breeds with much lower selection intensity and with very localised geographic locations and relatively small productions. The Large White was strongly introgressed with pigs from Asian origin back in the 18th and 19th century. The Landrace was developed in the late 19th century by crossing native Danish breeds with Large White pigs. All samples were collected from farms in Catalonia, Canada and France. Samples were selected based on DNA availability in our DNA archive. Although we do not have pedigree information, we tried to minimise co-ancestry by selecting samples from different farms when possible with the exception of the Mangalitza, which samples come from a close and highly inbred population and are thus likely to have strong familiar relationships. No phenotypic data is available from these samples. Furthermore, DNA from 38 F2 pigs from an experimental intercross created with the aim of studying obesity-related traits was used. This experimental intercross is described in detail by Kogelman et al. [34]. Briefly, the F2 population was created by inseminating seven Large White and seven Duroc sows with Göttingen minipig semen from 14 males. Both Large White and Duroc breeds have been selected for leanness and growth traits during many years, while Göttingen minipigs are prone to obesity. 563 F2 pigs were created. The animals were housed at a regular pig farm, and slaughtered at a commercial slaughterhouse under veterinary supervision. Tissue and blood samples were collected at slaughter. Extensive phenotypic collection was performed from birth to slaughter (242 ± 48 days) including obesity, obesity-related, and metabolic phenotypes; and measurements of fat compartments at slaughter [34]. The retroperitoneal fat compartment was removed from the carcass by blunt dissection and weighted on a bench scale. To calculate daily gain, body weight was measured individually at birth and at 7 month of age (220 ± 45 days). The 38 F2 pigs used in this study were selected to represent two divergent groups with respect to average daily gain (g/day) and retroperitoneal fat content (kg). The two divergent groups, i.e. the fast growing, obese (F2_F) and the slow growing, lean (F2_L), showed highly significant differences (T-test) for both traits (Additional file 8). gDNAs from purebred animals were extracted using different tissues (blood and a variety of solid tissues) and protocols including standard phenol - chloroform - Isoamyl alcohool organic extraction and the Charge Switch gDNA Micro Tissue kit (Invitrogen). DNA from the F2 pigs was extracted from EDTA stabilized blood using a salting out procedure. Samples were pooled in 16 tubes on a per-breed basis using semi-equal amounts of gDNA as measured by Nanodrop. The sample size of the pools ranged between 12 and 24 (Additional file 5). Likewise, the sample size of each breed ranged between 12 and 45 (Additional file 5). The F2 animals belong to an experimental population in Denmark and they were subject to animal care, maintenance and experimental work according to the ‘Animal Maintenance Act’ (Act 432 dated 09/06/2004) and the approval from the Danish Animal Experimentation Board (J.nr. 2007/561-1434). Specialized professionals at each institution obtained all the other blood and tissue samples following standard routine monitoring procedures and relevant guidelines. No animal experiment has been performed in the scope of this research. Capture of genomic regions, library prep and high throughput sequencing The 16 gDNAs pools were subjected to genomic capture and library preparation following Agilent’s SureSelect protocol for Illumina paired-end sequencing. Briefly, three μg of porcine gDNA were sheared on a Covaris™ E220 instrument. The fragment size (150–300 bp) and the quantity were confirmed with the Agilent 2100 Bioanalyzer 1000 chip. Fragmented DNA was end-repaired, adenylated and ligated to Agilent specific paired-end adapters. The DNA with adapter-modified ends was PCR amplified (six cycles, Herculase II fusion DNA polymerase). PCR product size and quantity were determined on the Agilent 2100 Bioanalyzer DNA 1000 assay and hybridized to the genomic capture baits for 24 h at 65 °C (Applied Biosystems 2720 Thermal Cycler). The hybridization mix was washed in the presence of magnetic beads (Dynabeads MyOne Streptavidin T1, Life Technologies) and the eluate was PCR amplified (16 cycles) in order to add the indexed tags using 6 bp SureSelectXT indexes for Illumina. The final library size and concentration was determined on an Agilent 2100 Bioanalyzer 1000 assay. Each library was sequenced on an Illumina HiSeq 2000 instrument in a fraction of a sequencing lane following the manufacturer’s protocol, with paired end run of 2x101bp. Images analysis, base calling and quality scoring of the run were processed using the manufacturer’s software Real Time Analysis (RTA 1.13.48) and followed by generation of FASTQ sequence files by Illumina’s proprietary CASAVA software. Read mapping and variant calling Reads were hard trimmed from the end of the read up to the first base with a quality of at least 10. Reads with at least 40 nt of length were mapped to Sus scrofa reference version 3 (http://hgdownload.cse.ucsc.edu/goldenPath/susScr3/bigZips/susScr3.fa.gz). As the list of TASRs was clearly shorter than the catalogue of AR genes and to minimize type I error (false negative variants), we applied a slightly different read mapping strategy for the two gene sets: 1) For TASR variant detection, reads were mapped first with the GEM toolkit [35] allowing up to four mismatches, and unmapped reads were then aligned to the swine genome using the more permissive BFAST read aligner [36]. We further manually curated the TASR variant list by removing these variants that clustered in high variant density regions as these indicate the presence of wrongly mapped reads - most of them probably aligned by Bfast - and thus false variant calls; 2) For the detection of variants in the AR genes, reads were mapped using only GEM as the manual curation would have been too labour intensive and prone to errors. Alignment (.bam) files containing only properly paired, uniquely mapped reads without duplicates were submitted to variant calling. Each pool was processed separately. The ploidy of the pool was calculated as two times the number of individuals in the pool and input as an optional argument in GATK 3.1 UnifiedGenotyper [37]. For variant calling, read numbers were down-sampled to 1,000 reads per position. Single pool variant calls were merged into a multi-sample.vcf file using GATK CombineVariants [37]. To confirm that variant positions not called in certain pools had a homozygous reference genotype, a second round of single pool variant calling was performed restricted to the list of variant positions in the merged.vcf file. Subsequently, results were merged again. Functional annotations were added using snpEff [38] with the Sscrofa10.2.69 database, and variants were classified according to their predicted impact as High (H), Moderate (M), Low (L) and Modifier (Additional file 9). Porcine dbSNP version 138 and porcine SIFT scores and deleteriousness prediction were annotated using snpSift [39] and the Ensembl Variant Effect Predictor (VEP) online tool (http://www.ensembl.org/Tools/VEP). M variants were further classified as deleterious (Mdel) or tolerated (Mtol) according to SIFT predictions. Genes of interest and the original target region of the capture experiment were annotated using vcftools [40]. Base counts at variant positions in the merged.vcf file were annotated using GATK Variant Annotator [37]. We used the proportion of reads carrying each allele as an estimator of the allelic frequency for both the alternative allele (pAAF) and for the minor allele (pMAF). Identification of breed-specific variants We searched for allelic variants that were uniquely present or uniquely absent in only one breed. We chose these variants that were either breed-specific but with a pAAF in the specific breed ≥ 0.1 and the variants that having a pAAF ≥ 0.5 in all breeds, were absent in only one breed. These breed specific features were assessed on the 2,793 variants identified in TASR and AR genes. Phylogenetic tree We calculated the pair-wise Pearson correlation of the pAAFs from the eight European breeds, the wild boars and the Asian animals (Meishan and Vietnamese together) with the 2,523 variants with the alternative allele present in at least one breed and with allelic information in the ten breeds and the wild boar. We then constructed a phylogenetic tree using an UPGMA method. These calculations were done using an in-house developed R script (Do.upgma.pops.bootstrap; https://bioinformatics.cragenomica.es/numgenomics//people/sebas/software/software.html). Assessment of allele frequency – phenotype relationships pAAFs at each variant position between the two F2 pools F2_F and F2_L were compared and the significance of these differences were determined using the Fisher exact test in an R environment. We excluded multi-allelic variants and variants with no read count in at least one pool from the analysis. After this filtering, we were able to compare 1,377 variants. OpenArray design, genotyping and variant validation In order to validate the most likely deleterious variants and to develop a genotyping assay to be used in future association studies in pig populations with relevant phenotypes, we developed a genotyping array containing 128 of such potential polymorphisms. We chose the TaqMan® OpenArray® Real-Time technology for Genotyping (Life Technologies) and designed the baits using the online Custom Assay Design tool. This design requires the absence of polymorphisms in the 2 bp window centred at the target variant and has also some constrains on the probe’s melting temperature. As a consequence, we could not design assays for all the variants using this approach. We first selected all the H variants in TASRs and AR genes and supplemented the assay with M variants in both gene groups. The genotyping was performed in a QuantStudioTM 12 K Flex Real-Time PCR System (Life Technologies). This platform is a high performance, high-throughput technology based on real-time PCR, which enables to run up to 12,000 data points, including SNV, small insertions and deletions (indels), simultaneously. 250 ng of gDNA and master mix were loaded to the OpenArray plates using the AccuFillTM robotic system (Life Technologies), filled with an immersion fluid and sealed. OpenArray plates were genotyped according to the manufacturer’s recommendations. Genotype analysis was performed using both Taqman Genotyper version 1.3 and Symphoni Suite software (Life Technologies). Genetic association of OpenArray genotypes with daily gain and retroperitoneal fat in the F2 We used a univariate mixed model to determine genetic associations of the 36 polymorphisms genotyped and segregating in the F2 animals with the phenotypic records for daily gain weight and retroperitoneal fat content. The analysis was calculated with the software GEMMA [41]. Abbreviations AR, appetite and reward; DP, read depth at the position; gDNA, genomic DNA; H, high impact variants; L, variant with no apparent impact; M, moderate impact variant; Mdel, moderate deleterious variant; Mtol, moderate tolerated variant; pAAF, predicted alternative allele frequency; pMAF, predicted minor allele frequency; SNV, single nucleotide variation; TASR, taste receptor; UPGMA, unweighted pair group method with arithmetic mean Additional files Additional file 1: List of successfully sequenced TASR and AR genes. Successfully sequenced genes include these genes fully or partially sequenced at a DP > 1,000 considering all the libraries together. (XLSX 25 kb) Additional file 2: List of variants identified in TASR and AR genes. This table contains all the variants, including the multi-allelic with more than one effect on protein sequence and these which alternative allele was fixed in all the animals. (XLSX 302 kb) Additional file 3: List of breed-specific variants. This table contains these variants that were present in a single breed at pAAF ≥ 0.1 or present in all breeds at pAAF ≥ 0.5 but absent in one single population. (XLSX 154 kb) Additional file 4: List of variants displaying significant differences in pAAF between F2_F and F2_L. (XLSX 145 kb) Additional file 5: Breed and sample size information for the 16 gDNA pools. Large White, Landrace, Duroc and Pietrain were sequenced in two pools according to their commercial origin. In the pool size column, the number between brackets indicates the sample size of each pool. The cells in the column ‘number of genotyped pigs’ indicate the number of pigs that were successfully genotyped for all or some of the variants included in the OpenArray. (XLSX 15 kb) Additional file 6: List of variants confirmed to be real polymorphisms by genotyping. MAF = observed minor allele frequency; * H variants which MAF is close to that of rare variants and display the three genotypic classes in the genotyped animals. (XLSX 12 kb) Additional file 7: Results on the genetic association of 36 polymorphisms with daily gain and retroperitoneal fat content in the 38 F2 pigs. (XLSX 13 kb) Additional file 8: Phenotype mean and standard deviation for F2_F and F2_L and the T-test p-value for both daily gain and retroperitoneal fat. (XLSX 16 kb) Additional file 9: Variant impact and effect classification according to SNPeff. This table describes the types of genetic variants according to their effect on protein sequence. This table has been modified from the SNPeff manual (http://snpeff.sourceforge.net/SnpEff_manual.html#input). (XLSX 11 kb) Acknowledgements We are grateful to Marcel Amills, Joan Tibau and Valentin Balteanu for contributing samples. We are thankful to Claudia Cabrera and Giuseppe Saldi for the R script to calculate the Fisher exact test. Funding This work was funded by grants from the Spanish Ministry of Science and Innovation (project AGL2010-22358-C02-01) and the Spanish Ministry of Economy and Competitiveness (AGL2013-44978-R; CSD2007-00036, IPT-2012-0378-060000). We would also like to thank the Danish Ministry of Science, Technology and Innovation for the grant to the “UNIK Project for Food Fitness and Pharma for Health” that funded the development of the F2 resource. Alex Clop acknowledges the Ramon y Cajal Fellowship program from the Spanish Ministry of Economy and Competitiveness (RYC-2011-07763). Sophia Derdak is supported by the Parc Científic de Barcelona through the Torres Quevedo subprogram (MICINN) under grant agreement PTQ-12-05391. Availability of data and material The fastq files for each pool have been submitted to SRA (study accession SRP078796), linked to NCBI’s BioProject PRJNA328570; http://www.ncbi.nlm.nih.gov/bioproject/PRJNA328570). Novel variants have been submitted to dbSNP (http://www.ncbi.nlm.nih.gov/SNP) and accession application is in progress. These variants are also listed in Additional file 2. The phylogenetic data has been uploaded to Treebase and it can be accessed through the url http://purl.org/phylo/treebase/phylows/study/TB2:S19540. Authors’ contributions A.Cl.and A.Sa. designed the experiment. A.Sh and A.Cl. searched for TASR and AR exonic regions. A.Ca. performed gDNA extractions. S.C. and M.F. generated the F2 animals and obtained the phenotypic data. A.M. did the genotyping using the OpenArray technology. S.D. and S.B. contributed the sequencing steps from wetlab library prep to all the bioinformatic analysis until variant calling. S.R. developed the R script to carry UPGM analysis and helped calculating genetic variability with the Watterson’s method. A.Cl. also contributed analysis. A.H. and P.vA. contributed samples. A.Cl. wrote the main text. All authors 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. Ethics approval and consent to participate The F2 animals belong to an experimental population in Denmark and they were subject to animal care, maintenance and experimental work according to the ‘Animal Maintenance Act’ (Act 432 dated 09/06/2004) and the approval from the Danish Animal Experimentation Board (J.nr. 2007/561-1434). The samples obtained from pigs other than the F2 generation were privately owned for non-research purposes. The owners provided consent for the use of these samples for research. For these samples, specialized professionals at each institution obtained all the blood and tissue samples following standard routine monitoring procedures and relevant guidelines. No animal experiment has been performed in the scope of this research. ==== Refs References 1. Bachmanov AA Beauchamp GK Taste receptor genes Annu Rev Nutr 2007 27 389 414 10.1146/annurev.nutr.26.061505.111329 17444812 2. Lee RJ Kofonow JM Rosen PL Siebert AP Chen B Doghramji L Xiong G Adappa ND Palmer JN Kennedy DW Kreindler JL Margolskee RF Cohen NA Bitter and sweet taste receptors regulate human upper respiratory innate immunity J Clin Invest 2014 124 1393 1405 10.1172/JCI72094 24531552 3. Meyer D Voigt A Widmayer P Borth H Huebner S Breit A Marschall S de Angelis MH Boehm U Meyerhof W Gudermann T Boekhoff I Expression of tas1 taste receptors in mammalian spermatozoa: Functional role of tas1r1 in regulating basal ca 2+ and camp concentrations in spermatozoa PLoS One 2012 7 e32354 10.1371/journal.pone.0032354 22427794 4. Depoortere I Taste receptors of the gut: emerging roles in health and disease Gut 2014 63 179 190 10.1136/gutjnl-2013-305112 24131638 5. Loper HB La Sala M Dotson C Steinle N Taste perception, associated hormonal modulation, and nutrient intake Nutr Rev 2015 73 83 91 10.1093/nutrit/nuu009 26024495 6. Fulton S Appetite and reward Front Neuroendocrinol 2010 31 85 103 10.1016/j.yfrne.2009.10.003 19822167 7. Rask-Andersen M Olszewski PK Levine AS Schiöth HB Molecular mechanisms underlying anorexia nervosa: Focus on human gene association studies and systems controlling food intake Brain Res Rev 2010 62 147 164 10.1016/j.brainresrev.2009.10.007 19931559 8. De Graaf C Blom WAM Smeets PAM Stafleu A Hendriks HFJ Biomarkers of satiation and satiety Am J Clin Nutr 2004 79 946 961 15159223 9. Palmiter RD Is dopamine a physiologically relevant mediator of feeding behavior? Trends Neurosci 2007 30 375 381 10.1016/j.tins.2007.06.004 17604133 10. Delgado TC Glutamate and GABA in appetite regulation Front Endocrinol (Lausanne) 2013 4 AUG 1 8 23355833 11. Mohammad-Zadeh LF, Moses L, Gwaltney-Brant S. Serotonin: a review. J Vet Pharmacol Therap. 2008;31:187–99. 12. Goldstein JL Zhao T Li RL Surviving Starvation: Essential Role of the Ghrelin-Growth Hormone Axis Surviving Starvation: Essential Role of the Ghrelin-Growth Hormone Axis Cold Spring Harb Symp Quant Biol 2011 76 121 127 10.1101/sqb.2011.76.010447 21785007 13. Gudbjartsson DF Helgason H Gudjonsson SA Zink F Oddson A Gylfason A Besenbacher S Magnusson G Halldorsson BV Hjartarson E Sigurdsson GT Stacey SN Frigge ML Holm H Saemundsdottir J Helgadottir HT Johannsdottir H Sigfusson G Thorgeirsson G Sverrisson JT Gretarsdottir S Walters GB Rafnar T Thjodleifsson B Bjornsson ES Olafsson S Thorarinsdottir H Steingrimsdottir T Gudmundsdottir TS Theodors A Large-scale whole-genome sequencing of the Icelandic population Nat Genet 2015 47 435 444 10.1038/ng.3247 25807286 14. Ng SB Buckingham KJ Lee C Bigham AW Tabor HK Dent KM Huff CD Shannon PT Jabs EW Nickerson DA Shendure J Bamshad MJ Exome sequencing identifies the cause of a mendelian disorder Nat Genet 2010 42 30 35 10.1038/ng.499 19915526 15. Kiuchi S Yamada T Kiyokawa N Saito T Fujimoto J Yasue H Genomic structure of swine taste receptor family 1 member 3, TAS1R3, and its expression in tissues Cytogenet Genome Res 2006 115 51 61 10.1159/000094801 16974084 16. Zhang J Yin Y Shu XG Li T Li F Tan B Wu Z Wu G Oral administration of MSG increases expression of glutamate receptors and transporters in the gastrointestinal tract of young piglets Amino Acids 2013 45 1169 1177 10.1007/s00726-013-1573-2 23943043 17. Groenen MA Archibald AL Uenishi H Tuggle CK Takeuchi Y Rothschild MF Rogel-Gaillard C Park C Milan D Megens HJ Li S Larkin DM Kim H Frantz LA Caccamo M Ahn H Aken BL Anselmo A Anthon C Auvil L Badaoui B Beattie CW Bendixen C Berman D Blecha F Blomberg J Bolund L Bosse M Botti S Bujie Z Analyses of pig genomes provide insight into porcine demography and evolution Nature 2012 491 393 398 10.1038/nature11622 23151582 18. da Silva EC de Jager N Burgos-Paz W Reverter A Perez-Enciso M Roura E Characterization of the porcine nutrient and taste receptor gene repertoire in domestic and wild populations across the globe BMC Genomics 2014 15 1057 10.1186/1471-2164-15-1057 25471201 19. Kumar P Henikoff S Ng PC Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm Nat Protoc 2009 4 1073 1081 10.1038/nprot.2009.86 19561590 20. Watterson GA On the Number of Segregating Sites in Genetical Models without Recombination Theor Popul Biol 1975 7 256 276 10.1016/0040-5809(75)90020-9 1145509 21. Wang L Zhang L Yan H Liu X Li N Liang J Pu L Zhang Y Shi H Zhao K Wang L Genome-wide association studies identify the loci for 5 exterior traits in a large white x Minzhu pig population PLoS One 2014 9 e103766 10.1371/journal.pone.0103766 25090094 22. Do DN Strathe AB Ostersen T Jensen J Mark T Kadarmideen HN Genome-wide association study reveals genetic architecture of eating behavior in pigs and its implications for humans obesity by comparative mapping PLoS One 2013 8 e71509 10.1371/journal.pone.0071509 23977060 23. Clop A Amills M Noguera JL Fernández A Capote J Ramón MM Kelly L Kijas JM Andersson L Sànchez A Estimating the frequency of Asian cytochrome B haplotypes in standard European and local Spanish pig breeds Genet Sel Evol 2004 36 97 104 10.1186/1297-9686-36-1-97 14713412 24. Fang M Andersson L Mitochondrial diversity in European and Chinese pigs is consistent with population expansions that occurred prior to domestication Proc Biol Sci 2006 273 1803 1810 10.1098/rspb.2006.3514 16790414 25. Frantz LA Schraiber JG Madsen O Megens HJ Bosse M Paudel Y Semiadi G Meijaard E Li N Crooijmans RP Archibald AL Slatkin M Schook LB Larson G Groenen MA Genome sequencing reveals fine scale diversification and reticulation history during speciation in Sus Genome Biol 2013 14 R107 10.1186/gb-2013-14-9-r107 24070215 26. Bianco E Nevado B Ramos-Onsins SE Pérez-Enciso M A Deep Catalog of Autosomal Single Nucleotide Variation in the Pig PLoS One 2015 10 e0118867 10.1371/journal.pone.0118867 25789620 27. Zhao H Yang JR Xu H Zhang J Pseudogenization of the umami taste receptor gene Tas1r1 in the giant panda coincided with its dietary switch to bamboo Mol Biol Evol 2010 27 2669 2673 10.1093/molbev/msq153 20573776 28. Shigemura N Shirosaki S Ohkuri T Sanematsu K Islam AA Ogiwara Y Kawai M Yoshida R Ninomiya Y Variation in umami perception and in candidate genes for the umami receptor in mice and humans Am J Clin Nutr 2009 90 764 769 10.3945/ajcn.2009.27462M 29. Ozdener MH Subramaniam S Sundaresan S Sery O Hashimoto T Asakawa Y Besnard P Abumrad NA Khan NA CD36- and GPR120-mediated Ca2+ signaling in human taste bud cells mediates differential responses to fatty acids and is altered in obese mice Gastroenterology 2014 146 995 1005 10.1053/j.gastro.2014.01.006 24412488 30. Vandenbeuch A Larson ED Anderson CB Smith SA Ford AP Finger TE Kinnamon SC Postsynaptic P2X3-containing receptors in gustatory nerve fibres mediate responses to all taste qualities in mice J Physiol 2015 5 1113 1125 10.1113/jphysiol.2014.281014 25524179 31. Han DH Bolo N Daniels MA Lyoo IK Min KJ Kim CH Renshaw PF Craving for alcohol and food during treatment for alcohol dependence: Modulation by T allele of 1519T > C GABAAα6 Alcohol Clin Exp Res 2008 32 1593 1599 10.1111/j.1530-0277.2008.00734.x 18616664 32. Bloss CS Berrettini W Bergen AW Magistretti P Duvvuri V Strober M Brandt H Crawford S Crow S Fichter MM Halmi KA Johnson C Kaplan AS Keel P Klump KL Mitchell J Treasure J Woodside DB Marzola E Schork NJ Kaye WH Genetic Association of Recovery from Eating Disorders: The Role of GABA Receptor SNPs Neuropsychopharmacology 2011 36 2222 2232 10.1038/npp.2011.108 21750581 33. Jin K Xue C Wu X Qian J Zhu Y Yang Z Yonezawa T Crabbe MJC Cao Y Hasegawa M Zhong Y Zheng Y Why does the giant panda eat bamboo? a comparative analysis of appetite-reward-related genes among mammals PLoS One 2011 6 e22602 10.1371/journal.pone.0022602 21818345 34. Kogelman LJ Kadarmideen HN Mark T Karlskov-Mortensen P Bruun CS Cirera S Jacobsen MJ Jørgensen CB Fredholm M An F2 pig resource population as a model for genetic studies of obesity and obesity-related diseases in humans: Design and genetic parameters Front Genet 2013 4 29 10.3389/fgene.2013.00029 23515185 35. Marco-Sola S Sammeth M Guigó R Ribeca P The GEM mapper: fast, accurate and versatile alignment by filtration Nat Methods 2012 9 1185 1189 10.1038/nmeth.2221 23103880 36. Homer N Merriman B Nelson SF BFAST: An alignment tool for large scale genome resequencing PLoS One 2009 4 e7767 10.1371/journal.pone.0007767 19907642 37. DePristo MA Banks E Poplin R Garimella KV Maguire JR Hartl C Philippakis AA del Angel G Rivas MA Hanna M McKenna A Fennell TJ Kernytsky AM Sivachenko AY Cibulskis K Gabriel SB Altshuler D Daly MJ A framework for variation discovery and genotyping using next-generation DNA sequencing data Nat Genet 2011 43 491 498 10.1038/ng.806 21478889 38. Cingolani P Platts A Wang LL Coon M Nguyen T Wang L Land SJ Ruden DM Lu X A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3 Fly (Austin) 2012 6 1 13 10.4161/fly.19695 39. Cingolani P Patel VM Coon M Nguyen T Land SJ Ruden DM Lu X Using Drosophila melanogaster as a model for genotoxic chemical mutational studies with a new program, SnpSift Front Genet 2012 3 35 10.3389/fgene.2012.00035 22435069 40. Danecek P Auton A Abecasis G Albers CA Banks E DePristo MA Handsaker RE Lunter G Marth GT Sherry ST McVean G Durbin R The variant call format and VCFtools Bioinformatics 2011 27 2156 2158 10.1093/bioinformatics/btr330 21653522 41. Zhou X Stephens M Genome-wide efficient mixed-model analysis for association studies Nat Genet 2012 44 821 824 10.1038/ng.2310 22706312

PMC005xxxxxx/PMC5002120.txt

==== Front BMC EcolBMC EcolBMC Ecology1472-6785BioMed Central London 9110.1186/s12898-016-0091-yResearch ArticlePopulation density of the western burrowing owl (Athene cunicularia hypugaea) in Mexican prairie dog (Cynomys mexicanus) colonies in northeastern Mexico Ruiz Ayma Gabriel gabriel.ruizaym@uanl.edu.mx 1Olalla Kerstupp Alina alina.olallakrs@uanl.edu.mx 1Macías Duarte Alberto alberto.macias@ues.mx 2Guzmán Velasco Antonio antonio.guzman@uanl.mx 1González Rojas José I. jose.gonzalezr@uanl.mx 11 Facultad de Ciencias Biologicas, Universidad Autonoma de Nuevo Leon, Ave. Universidad s/n. Cd. Universitaria, 66455 San Nicolas de los Garza, Nuevo Leon Mexico 2 Ley Federal del Trabajo S/N, Universidad Estatal de Sonora, Col. Apolo, 83100 Hermosillo, Sonora Mexico 26 8 2016 26 8 2016 2016 16 1 381 11 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 The western burrowing owl (Athene cunicularia hypugaea) occurs throughout western North America in various habitats such as desert, short-grass prairie and shrub-steppe, among others, where the main threat for this species is habitat loss. Range-wide declines have prompted a need for reliable estimates of its populations in Mexico, where the size of resident and migratory populations remain unknown. Results Our objective was to estimate the abundance and density of breeding western burrowing owl populations in Mexican prairie dog (Cynomys mexicanus) colonies in two sites located within the Chihuahuan Desert ecoregion in the states of Nuevo Leon and San Luis Potosi, Mexico. Line transect surveys were conducted from February to April of 2010 and 2011. Fifty 60 ha transects were analyzed using distance sampling to estimate owl and Mexican prairie dog populations. We estimated a population of 2026 owls (95 % CI 1756–2336) in 2010 and 2015 owls (95 % CI 1573–2317) in 2011 across 50 Mexican prairie dog colonies (20,529 ha). Conclusions The results represent the first systematic attempt to provide reliable evidence related to the size of the adult owl populations, within the largest and best preserved Mexican prairie dog colonies in Mexico. Keywords Chihuahuan DesertDistance samplingGrasslandMexican prairie dogMexicoPopulationWestern burrowing owlissue-copyright-statement© The Author(s) 2016 ==== Body Background Rigorous estimates of regional population size are critical for the development and assessment of avian conservation strategies, particularly for species undergoing shifts in their distribution and range. The western burrowing owl (Athene cunicularia hypugaea) (Fig. 1), a species with special conservation status throughout much of its range, has experienced range-wide shifts from southern Canada to northern Mexico [1]. Western burrowing owls belong to a grassland bird guild that is threatened by habitat loss [2]. The species uses open habitats such as grasslands, deserts and areas of disturbance [3]. These owls also prefer areas with discontinuous vegetation and low growth shrubs, allowing them to increase visibility for hunting, vigilance against predators and caring of burrows [4, 5].Fig. 1 Western burrowing owl in the colony of Mexican prairie dog, in Chihuahuan Desert Published data from owl populations vary within the range of distribution in North America. For example, in the 1990's population estimates of this species in Canada and the United States of America (USA) ranged from as low as 2000–20,000 to as high as 20,000–200,000 individuals [6]. In Canada, the populations have declined abruptly and even disappeared from British Columbia and Manitoba [7]. Previous reports indicate a wide variation of population trends ranging from stable in some areas in the USA and Canada, to reduced, extirpated or increasing in others [2, 7–16]. Local density estimates a range of 13–31 owls/km2 in Canada (Manitoba, Alberta, Saskatchewan and British Columbia) and the USA (Arizona, California, Colorado, Idaho, Iowa, Kansas, Minnesota, Montana, Nebraska, New Mexico, North and South Dakota, Oklahoma, Oregon, Texas, Utah and Washington) during the western burrowing owl breeding season indicating variation in the density estimates [1, 7, 9, 14, 17–26]. In Mexico, the federal government classifies the western burrowing owl under the category of special protection [27]. Habits of the western burrowing owl such as summer diet, prey selection, movement of juveniles, selection of nesting sites and threats remain poorly known. Densities estimated during the breeding season in 2002 in Mexico include 14.1 owls/km2 near Mexicali [28], 3.2 pairs/km2 in Yaqui-Mayo Valley, Sonora, 4.5 pairs/km2 in Valle del Fuerte, Sinaloa, and 4.7 pairs/km2 in Valle de Culiacan, Sinaloa [29]. Winter season density estimates in central Mexico include 11 owls/km2 in Guanajuato [30] and 5.2 owls/km2 in Nuevo Leon [31]. The western burrowing owl has been strongly associated with two species of prairie dogs in Mexico, the Mexican prairie dog (Cynomys mexicanus) and black-tailed prairie dog (C. ludovicianus) (Fig. 2) [9, 32–35]. Both of these species are federally listed in Mexico as endangered and threatened, respectively [27]. The black-tailed prairie dog is distributed from Saskatchewan in Canada to southern Montana and Nebraska in the United States to northern Chihuahua and Sonora in Mexico where the colonies are fragmented and isolated. The habitat occupied by the species of prairie dog is herbs, grasses and shrubs. Currently, the regions supporting black-tailed prairie dog colonies cover 18,500 ha [36]. The Mexican prairie dog is endemic of central and northern of Mexico within the states of Coahuila, Nuevo Leon, Zacatecas and San Luis Potosi, in colonies covering approximately 25,000 ha. These two species of dogs have lost more than 80 % of their original range [37]. Mexican prairie dog colonies provide burrows and foraging opportunities for breeding burrowing owls, which apparently keep the prairie dog population stable, despite disturbance and loss of habitat in prairie dog colonies caused by expanding agricultural and cattle grazing activities [38], the use of pesticides, collisions with vehicles, diseases, predators, and urbanization [1, 2, 7, 11, 32, 39–44].Fig. 2 Mexican prairie dogs sampling sites, located in the Chihuahuan Desert within the states Coahuila, NL and SLP, Mexico Based on the problems and the lack of knowledge mentioned above, in this study we estimate the abundance and density of western burrowing owls in colonies of Mexican prairie dog in northeastern Mexico. Density/abundances of western burrowing owls and their association with Mexican prairie dog colonies provide relevant conservation information to ensure the long-term persistence of both species. In addition, this study can be integrated across North America to establish baseline range-wide population estimate(s) to improve our understanding of the recent range-wide shifts in owl populations. Methods Study area Our study sites were located in Nuevo Leon (NL) and San Luis Potosi (SLP) within the Chihuahuan Desert ecoregion [45] (Fig. 2) that is part of the physiographic region known as the Mexican Plateau within the Mexican states of Coahuila, Zacatecas, NL and SLP. The semi-arid climate features temperatures ranging from 6 to 25 °C with an annual average of 16 °C [46]. Average precipitation totals 427 mm [47]. Previously, studies in NL have been conducted in the areas known as Llano de la Soledad (23°53′N, 100°42′W) and Compromiso (23°53′N, 100°42′W). These areas maintain the largest Mexican prairie dog populations, including those at Martha (25°0′N, 100°40′W), Concha (25°1′N, 100°35′W), and Hediondilla (24°57′N, 100°42′W). Western burrowing owls of SLP were studied in Llano del Manantial (24°7′N, 100°55′W) and Gallo (24°12′N, 100°54′W) in the municipality of Vanegas. The Llano de la Soledad has been provided with several conservation designations by the NL government such as State Natural Protected Area [48], and Important Site for Bird Conservation [49]. This site hosts several vulnerable, endemic and migratory species [50, 51]. The dominant vegetation in Mexican prairie dog colonies is characterized by halophytic grassland and consists largely of Muhlenbergia villiflora, Muhlenbergia repens, Pleuraphis mutica, Sporobolus airoides, Frankenia gypsophila and Dalea gypsophila. Other coexisting plant communities include microphyllus vascular plants and rosette shrubs [31, 52–56]. From 50 colonies of prairie dogs existing in NL and SLP, nine were selected for sampling. These colonies were selected based on the following characteristics: spatial continuity of the community and a lack of fragmentation, conservation status of the site, vegetation type that was homogeneous enough to contain at least one complete transect. The sampled colonies covered about 55 % of the area available for all colonies of Mexican prairie dogs in the southern Chihuahuan Desert. Sampling was conducted between February and April in both 2010 and 2011. The transect line method was used [57]. Fifty transects (each 2 km long × 0.3 km wide and ≥0.5 km apart from each other) were traveled using the remote sampling method by the observer as described below to estimate the density of adult owls [58]. The number (n) of transect routes for each area was: Soledad (n = 28) and Compromiso (15), Marta (2) and Concha (2) in NL; Manantial (2) and Gallo (1) in SLP. We walked each transect at a constant rate using a global positioning system (GPS) to ensure a straight survey line. Owls were detected visually or with binoculars. Then, the perpendicular distance from the transect line route was measured using a laser rangefinder (15–815 m, Leica Rangemaster 900, Optics Planet, Inc. Northbrook, IL, USA). To meet the assumptions of distance sampling, only adult owls were recorded on the ground outside the burrows or without movement [58, 59]. If the bird under observation moved because of the presence of the observer, registering the perpendicular distance was performed at the original site without the observer leaving their sighting transect travel line. Those adult owls flying with an unknown initial location were not documented. To reduce bias and avoid an overestimation of population density, only adult owls were recorded. Considering the extreme desert climate, personal observations made during previous years and different criteria of previous authors related to the activities of owls, the field observations were conducted from 0600 to 1200 h [20, 23, 60, 61]. Data analysis We used program DISTANCE ver. 6.0 to obtain western burrowing owl density estimates from distance sampling [62]. DISTANCE calculates density and abundance using modeling detection probability as a function of the perpendicular distance to the transect in a series of monotonic models. Several standard detection functions (uniform, half-normal, or hazard-rate) with cosine series adjustment were evaluated using the Akaike information criterion (AIC). We used the AIC to select the model with the most parsimonious detection function in DISTANCE [58, 59, 63]. We pooled all data to estimate a single detection function (probability of detection, g (x), at a given distance (x) from the transect) because we did not anticipate effects of environmental features on detection, such as age (adult) and factor STATE (levels: NL and SLP). We considered serial adjustments of one to three parameters. We did not truncate the data because the frequencies of long distances observations were better maintained in this manner [4]. The estimator of density (D^) is given by the expression: D^=n^f^(0)2L where f^(0) is the probability density function of detection distances from the line evaluated at zero distance, calculated in DISTANCE as the average number of individuals per detection [62]. The standard error of density SE (D^), assuming a Poisson distribution of counts, can be approximated using the delta method as follows [58]: SE(D^)=D^1n+Var(f^(0))(f^(0))2, where Var(f^(0))=(SE(f^(0))2 also is a direct output of DISTANCE. The component cluster size was omitted from the above formulas because virtually all detections were individual records. Estimates of density and their standard errors were used to test statistical differences in density between states and years using a Wald test [64]. Values are presented as mean ± SE. Overall estimates of western burrowing owl density (and their SE) at the nine sampled colonies (9620 ha) were obtained by pooling detection distance data by year. These estimates were then multiplied by the total area of the 50 colonies of the Mexican prairie dog described for the southern part of the Chihuahuan Desert to provide yearly estimates of owl population size through the range of Mexican prairie dog: 38 in NL (19,802 ha) and 12 in SLP (727 ha) [37]. On average, 55 % of the surface reported for the Mexican prairie dog colony complex was sampled [37]. Results Density and population size Colonies were stable during the years 2010–2011 and were not destroyed or fragmented (agriculture, livestock) during this time. During the 2010 and 2011 sampling periods, 235 detections of at least one owl were recorded. The estimates of western burrowing owl density in the 50 prairie dog colony complex were 9.8 ± 1.0 ind/km2 (CV 0.107) in 2010 and 9.8 ± 1.0 ind/km2 (CV 0.108) in 2011. The owl density estimate for NL was 8.8 ± 1.0 ind/km2 (CV 0.114) in 2010 and 7.3 ± 0.9 ind/km2 (CV 0.123) in 2011. For SLP, the owl population density was 26.7 ± 6.2 ind/km2 (CV 0.236) in 2010 and 47 ± 8.4 ind/km2 (CV 0.180) in 2011 (Table 1). No significant differences were found among western burrowing owl densities (Wald test, p = 0.431) and the paired states of NL (p = 0.967) and SLP (p = 0.635).Table 1 Western burrowing owl population density between 2010 and 2011 in Mexican prairie dog colonies in NL and SLP, Mexico Modela Db Nc Estimated density (owl/ha) CVf No. colonyg Area (ha) Average 95 % ICd 95 % ICe Global  2010 0.1 119 949 788 1039 0.107 9 9620* 2026 1765 2326 50 20,529**  2011 0.09 116 944 783 1035 0.108 9 9620* 2015 1753 2317 50 20,529** NL  2010 0.08 100 809 698 937 0.114 6 9170* 1747 1508 2024 50 20,529**  2011 0.074 82 678 578 794 0.124 6 9170* 1464 1248 1716 50 20,529** SLP  2010 0.26 19 118 37 87 0.236 3 450* 190 60 141 12 727**  2011 0.47 34 211 48 167 0.180 3 450* 341 170 397 12 727** *Total area of sampled colonies **Total area of colonies in SLP and NL aModel base done AIC criteria: half-normal + cosine bWestern burrowing owl density (owl/ha) cTotal number of detections in both years dUpper confidence intervals eLower confidence intervals fVariation coefficient for the estimated density gNumber of Mexican prairie dog colonies Applying the overall yearly estimates of western burrowing owl density to the entire area of the 50-colony complex of prairie dogs in NL and SLP resulted in a population size of 2026 (CV 0.173) in 2010 and 2015 (CV 0.213) in 2011. For colonies in NL, an average population size of 1747 (CV 0.178) was obtained in 2010 and 1464 (CV 0.218) for 2011, while in SLP, population estimates were between 190 (CV 0.312) and 341 (CV 0.322) for each year. Discussion To date, many density estimates have been made for the western burrowing owl in Canada and the USA, with quite variable results [1, 6, 7, 9, 11–15, 20–23, 25, 26]. The resulting variation in the population sizes can be attributed to the size of sample area, methodology, analytical precision, timing, observer skill, and so on; these have contributed to an inexact picture of the density of the western burrowing owl populations [6, 15]. Therefore, a comparison of our results with those of the USA and Canada could be difficult. During the last 30 years, the North American Breeding Bird Survey has estimated a negative trend for the western burrowing owl population for Canada and the USA. Similarly, the United States Geological Survey (2014) has reported the same negative trend in the Chihuahuan Desert region [16]. Even though Mexico has not established systematic surveys that allow the establishment of a population trend, some studies (the present one included) can form the basis to achieve this goal of documenting population trends in the future. In NL and SLP, the average density of breeding pairs (9.8 ind/km2) in 2010 and 2011 is greater than that reported by Macias-Duarte in Sonora (6.4 ind/km2) and similar to the Sinaloa average (9.2 ind/km2) [29]. However, in Baja California, Itubarria-Rojas reported an average of 14.1 ind/km2, which is a value higher than that determined by the present study [28]. This difference could be caused by the habitat quality among sites as reported in NL and SLP where burrow competition is related to the abundance of prairie dogs per colony or Baja California where the owls use irrigation canals to create burrows. Our overall estimates of population size for western burrowing owls reveal the relative importance of Mexican prairie dog colonies to owl population viability. No previous data related to population size estimates in owls is available for the study area. However, the precision of these estimates must be taken with caution because of the variability between sites. However, we believe the extrapolation is correct because we sampled over 55 % of the current area with the active prairie dog colonies in both states. The range of the western burrowing owl in northeastern (NL, SLP, and Coahuila) Mexico includes viable colonies of Mexican prairie dogs. These areas provide an optimal habitat for the prairie dogs to feed on grasses and this contributes to a low height of herbaceous plants and allows the owls greater visual access to the foraging area. This species uses prairie dog colonies as a place for nesting, protection against climatic factors (extreme temperatures, flooding by rain, and strong winds). The owls also respond to alarm calls by prairie dogs, alerting them to the presence of predators. The western burrowing owl colonies in Mexico have declined from 88 colonies to 53, equivalent to a loss of 37 % in 10 years (1992–2003) [37, 38]. Many of the problems in northeastern Mexico that involve the western burrowing owl are directly related to loss of habitat from agriculture, but some direct mortality has been caused by collisions with vehicles. However, another possible cause of morbidity and mortality could be the direct or incidental (by bioaccumulation) exposure to pesticides used in neighboring areas. Conclusions These results represent the first systematic effort to address the conservation status of the western burrowing owl populations in Mexican prairie dog colonies located in northeastern Mexico. This geographic area is considered to contain the largest preserved Mexican prairie dog colonies in the country and deserves attention from the scientific and conservation communities. Furthermore, these results contribute new information to our understanding of the population dynamics of this kind of species across North America, and highlight the urgent need to preserve grasslands, particularly those in the southern part of the Chihuahuan Desert, which harbor many bird species cataloged as threatened or endangered. Abbreviations Regions USA: United States of America; NL: Nuevo Leon; SLP: San Luis Potosi. Units km: kilometers; ind/km2: individual per square kilometer; ha: hectare; n: number of line transects; m: meters; mm: millimeter; °C: celsius; hr: hours. Statistical gx: probability of detection; x: given distance; ± SE: standard error; CV: coefficient variation; p: probability; AIC: akaike information criterion; IC: confidence intervals. Orientation N: north; W: western; GPS: global positioning system. Authors’ contributions GRA conceived of and designed the study, collected the data and performed data analysis. AOK, AMD, AGV and JIGR contributed to study design and provided advice for data collection and analysis. All authors participated in drafting the manuscript. All authors read and approved the final manuscript. Acknowledgements We are grateful to Pronatura Noreste Asociacion Civil for Cynomys ludovicianus photo (Fig. 2) and their facilities for hospitality throughout the sampling period. Competing interests The authors declare that they have no competing interests. Availability of data and materials Data are available in the Dryad database (http://dx.doi.org/10.5061/dryad.pm362). Description of supplementary files can be found at [65]. Ethics All protocols were performed according to the ethical guidelines adopted by the ethic committee of the Facultad de Ciencias Biologicas of the Universidad Autonoma de Nuevo Leon, as well by the current environmental Mexican laws. Only field observations were made. There was no animal handling. However in order to comply with the Mexican regulations we have a permit (SGPA/DGVS/01588/10), granted by the Secretaria del Medio Ambiente y Recursos Naturales/Subsecretaria de Gestion para la Proteccion Ambiental/Direccion General de Vida Silvestre. Funding This research was funded with resources from the Universidad Autonoma de Nuevo Leon through the support program for Scientific and Technological Research (PAICyT). ==== Refs References 1. Macias-Duarte A Conway CC Distributional changes in the western burrowing owl (Athene cunicularia hypugaea ) in North America from 1967 to 2008 J Raptor Res 2015 49 75 83 10.3356/JRR-14-00004.1 2. ACA. Commission for Environmental Cooperation. In: North America conservation action plan (Athene cunicularia hypugaea). Commission for Environmental Cooperation. Printed in Canada; 2005. p. 1–55. http://www.cec.org Accessed 15 Apr 2009. 3. Clark RJ A review of the taxonomy and distribution of burrowing owl (Speotyto cunicularia ) J Raptor Res 1997 9 14 23 4. Coulombe HN Behavior and population ecology of the burrowing owl, Speotyto cunicularia , in the Imperial Valley of California Condor 1971 73 162 176 10.2307/1365837 5. Howell GR, Webb S. A Guide to the birds of Mexico and Central America. Oxford University Press; 1995. pp. 364. 6. James PC Espie RHM Current status of the burrowing owl in North America: an agency survey J Raptor Res 1997 9 3 5 7. COSEWIC. Assessment and update status report on burrowing owl Athene cunicularia in Canada. Committee on the status of endangered wildlife in Canada. Ottawa; 2006. pp. 31. 8. Desante DF, Ruhlen ED, Adamany SL, Butron KM, Amin S. A census of burrowing owls in central California in 1991. In: Lincer J, Steenhof K, editors. The burrowing owl, its biology and management including the proceedings of the first international burrowing owl symposium; 1997. 9. Desmond MJ Savidge JA Factors influencing burrowing owl (Speotyto cunicularia ) nest densities and numbers in western Nebraska Am Midl Nat 1996 136 143 148 10.2307/2426639 10. Clayton KM Schmutz JK Is the decline of burrowing owls Speotyto cunicularia in prairie Canada linked to changes in great plains ecosystems? Bird Conserv Int 1999 9 2 163 185 10.1017/S0959270900002288 11. Arrowood PC Finley CA Thompson C Analyses of burrowing owl populations in New Mexico J Raptor Res 2001 35 4 362 370 12. Korfanta NM Ayers LW Anderson SH McDonald DB A preliminary assessment of burrowing owl status in Wyoming J Raptor Res 2001 35 337 343 13. Sheffield SR Howery M Current status, distribution, and conservation of the burrowing owl in Oklahoma J Raptor Res 2001 35 351 356 14. Murphy RK Hasselbland DW Grondahl CD Sidle JG Martin RE Feed DW Status of the burrowing owl in North Dakota J Raptor Res 2001 35 322 330 15. Klute DS, Green TM, Howe WH, Jones ST, Shaffer JL, Sheffield SR, Zimmerman TS. Status assessment and conservation plan for the western burrowing owl in the United States. US Department of Interior, Fish & Wildlife Service, Biological Technical Publication FWS/BTP-R6001-2003, Washington, D.C.; 2003. 16. Sauer JR, Hines J E, Fallon JE, Pardieck KL, Ziolkowski DJ. Link the North American breeding bird survey, results and analysis 1966–2013. Version 01.30.2015 USGS Patuxent Wildlife Research Center, Laurel. USA; 2014. 17. Butts KO Lewis JC The importance of prairie dog towns to burrowing owls in Oklahoma Proc Okla Acad Sci 1982 62 46 52 18. Enriquez-Rocha P Rangel-Salazar DW Presence and distribution of Mexican owls: a review J Raptor Res 1993 27 154 160 19. Trulio L Burrowing owl demography and habitat use at two urban sites in Santa Clara County, California J Raptor Res 1997 9 84 89 20. Conway CJ Simon JC Comparison of detection probability associated burrowing owl survey methods J Wildl Manag 2003 67 3 501 511 10.2307/3802708 21. Desante DF Ruhlen ED Rosenberg DK Density and abundance of burrowing owls in the agricultural matrix in the Imperial Valley Stud Avian Biol 2004 27 116 119 22. Manning JA. Burrowing owl population size in the Imperial Valley, California: survey and sampling methodologies for estimation. Final report to the Imperial irrigation district, Imperial, California, USA; 2009. http://www.iid.com/Modules/ShowDocument.aspx?documentid=8172. Accessed 15 Apr 2009. 23. Tipton HC Doherty PF Dreitz VJ Abundance and density of mountain plover (Charadrius montanus ) and burrowing owl (Athene cunicularia ) in eastern Colorado Auk 2009 126 493 499 10.1525/auk.2009.07172 24. Berardelli D Desmond JM Murray L Reproductive success of burrowing owls in urban and grassland habitats in southern New Mexico Wilson J Ornithol 2010 122 1 51 59 10.1676/09-034.1 25. Crowe D, Longshore K. Population status and reproduction ecology of the western burrowing owl (Athene cunicularia hypugaea) in Clark Contry, Nevada. Report final 2005.USGS-582-P. United States Geological Survey; 2010. pp. 31. 26. Wilkerson RL Sigel RB Distribution an abundance of western burrowing owls (Athene cunicularia hypugaea ) in southeastern California Southwest Nat 2011 56 3 378 384 10.1894/F11-KF-16.1 27. NOM-059-SEMARNAT-2010. Protección ambiental-Especies nativas de Mexico de flora y fauna silvestres-categorias de riesgo y especificaciones para su inclusion, exclusion o cambio-lista de especies en riesgo. DIARIO OFICIAL DE LA FEDERACION; 2010. http://www.profepa.gob.mx/innovaportal/file/435/1/NOM_059_SEMARNAT_2010.pdf. Accessed 30 Dec 2010. 28. Itubarria-Rojas H. Estimacion de abundacia y afinidad de habitat del tecolote llanero (Athene cunicularia) en el Valle de Mexicali California y Sonora, Mexico. Universidad Autonoma de Guadalajara. Facultad de Ciencias Quimicas y Biologicas. Tesis de licenciatura. Guadalajara, Jalisco, Mexico; 2002. pp. 36. 29. Macias-Duarte A. Change in migratory behavior as possible explanation for burrowing owl population declines in northern latitudes. The University of Arizona. School of Natural Resources and the Environmental. PhD Thesis. USA; 2011. pp. 145. 30. Valdez-Gomez HE Holroyd GL The burrowing owl, habits and distribution center in western Mexico Boletin de la Sociedad de Ciencias Naturales de Jalisco 2000 1 57 63 31. Cruz-Nieto MA. Ecologia invernal de la lechuza llanera (Athene cunicularia), en los pastizales ocupados por los perritos llanero Mexicano (Cynomys mexicanus), Nuevo Leon, Mexico. Universidad Autonoma de Nuevo Leon. Facultad de Ciencias Biologicas. Laboratorio de Ornitologia. San Nicolas de los Garza. Tesis Doctorado en Ciencias; 2006. pp. 118. 32. Desmond MJ Savidge JA Eskridge KM Correlations between burrowing owl and black-tailed prairie dog declines: a 7-year analysis J Wild Manag 2000 64 4 1067 1075 10.2307/3803217 33. Griebel RL Savidge JA Factors related to body condition of nestling burrowing owls in Buffalo Gap National Grassland, South Dakota Wilson J Ornithol 2003 115 477 480 34. McNicolle JL. Burrowing owl (Athene cunicularia) nest site selection in relation to prairie dog colony characteristics and surrounding land-use practices in Janos, Chihuahua, Mexico. Las Cruces, New Mexico, New Mexico State University. Ms Thesis. USA; 2005. pp. 54. 35. Ruiz-Ayma G. Exito reproductive, entrega de presas y dieta del tecolote (Athene cunicularia hypugaea) en el complejo de la colonias de perrito de la pradera Mexicano (Cynomys mexicanus) en Galeana, Nuevo Leon, Mexico. Universidad Autonoma de Nuevo Leon. Facultad de Ciencias Biologicas. Tesis Maestria en Ciencias. Mexico; 2009. pp. 85. 36. Ceballos GOG. Los mamiferos silvestres de Mexico. Ed. Fondo de Cultura Economica de España; 2009. pp. 986. 37. Carrera MMA. Situacion actual, estrategias de conservacion y bases para recuperacion del perrito llanero mexicano (Cynomys mexicanus). Universidad Autonoma de Mexico. Tesis Maestria en Ciencias. Mexico; 2008. pp. 72. 38. Scott-Morales L Estrada E Chavez-Ramirez M Cotera M Continued decline in geographic distribution of Mexican prairie dog (Cynomys mexicanus ) J Mammal 2004 85 6 1095 10.1644/BER-107.1 39. Green GA Anthony RG Nesting success and habitat relationships of burrowing owl in the Columbia Basin Oregon Condor 1989 91 347 354 10.2307/1368313 40. Haug EA, Millsap BA, Martell MS. Burrowing owl (Speotyto cunicularia). In: Poole A, Gill F, editors, The birds of North America, No. 61. Academy of Natural Sciences, Philadelphia, and American Ornithologists’ Union, Washington, DC; 1993. pp. 20. 41. Sheffield SR. Current status, distribution, and conservation of the burrowing owl (Speotyto cunicularia) in Midwestern North America. 1997. In: Duncan JR, Johnson DH, Nicholls TH, editors. Biology and conservation of owls of the Northern Hemisphere, USDA Forest Service, General Technical Report NC-190. North Central Forest Experiment Station, St. Paul, Minnesota;1997. pp. 399–407. 42. Wellicome TI. Effects of food on reproduction in burrowing owl (Athene cunicularia) during three stages of the breeding season. University of Alberta. Edmonton. PhD Thesis. Canada; 2000. pp. 113. 43. Holroyd GR Rodriguez RE Sheffield S Conservation of the burrowing owl in western North America, challenges and recommendations J Raptor Res 2001 35 3 399 407 44. McDonald D, Korfanta M, Lantz SJ. The burrowing owl (Athene cunicularia): a technical conservation assessment Wyoming, USDA Forest Service, Rocky Mountain Region; 2004. 45. CONABIO. Comision Nacional para el Conocimiento y Uso de la Biodiversidad. Mexico; 2008. http://www.conabio.gob.mx/informacion/metadata/gis/ecort08gw.xml?_xsl=/db/metadata/xsl/fgdc_html.xsl&_indent=no. Accessed 26 Sept 2009. 46. CONAGUA. Comisión Nacional del Agua. Consulta base de datos. Distrito Federal, Mexico. http://www.smn.cna.gob.mx/es/emas. Accessed 15 Sept 2009. 47. Instituto Nacional De Estadística Geografia e Informatica. Conjunto de datos vectoriales de la carta de uso del suelo y vegetacion, escala 1:250,000, Serie III. INEGI. Mexico; 2005. 48. Periodico Oficial. Monterrey, N. L., Gobierno Constitucional del Estado Libre y Soberano de Nuevo Leon, Mexico. Tomo CXXXIX; 2002. 49. WHSRN. Designación de sitio en categoria de importancia internacional para la conservacion de aves playeras de la red hemisferica de reservas para aves playeras. 2005. http://www.whsrn.org/site-profile/llano-de-la-soledad. Accessed 03 Dec 2014. 50. Macias-Duarte A, Panjabi AO, Pool D, Youngberg E, Levandoski G. Wintering grassland bird density in Chihuahuan Desert grassland priority conservation areas, 2007–2011. Rocky Mountain Bird Observatory, Brighton, CO, RMBO Technical Report INEOTROP- MXPLAT-10-2; 2011. pp. 164. 51. Del Coro-Arizmendi, Marquez VL. Areas de importancia para la conservacion de las aves, CONABIO & Fondo Mexicano para la Conservacion de la Naturaleza; 2000. http://www.conabioweb.conabio.gob.mx/aicas/doctos/NE-36.html. Accessed 03 Dec 2010. 52. Johnston MC Past and present grassland of southern Texas and northeastern Mexico Ecology 1963 44 456 466 10.2307/1932524 53. Rojas MP. Generalidades sobre la vegetacion del estado de Nuevo Leon y datos acerca de su flora. Facultad de Ciencias. Mexico D.F, Universidad Nacional Autonoma de Mexico; 1965. pp. 124. 54. Rivera RE. Caracterizacion y productividad invernal de tres areas de pastizal habitat para la lechuza llanera (Athene cunicularia) en el Municipio de Galeana, Nuevo Leon; Mexico. Universidad Autonoma de Nuevo Leon. Facultad de Ciencias Biologicas. Tesis licenciatura. México; 2006. pp. 72. 55. Garcia RAM. Habitat reproductivo del gorrion de Worthen (Spizella wortheni) en cuatro localidades del noreste de Mexico. Universidad Autonoma de Nuevo Leon. Facultad de Ciencias Biologicas. Tesis licenciatura. México; 2008. pp. 71. 56. Martinez RLM. Caracterizacion de los sitios de anidación del gorrion de Worthen (Spizella wortheni) en los estados de Nuevo Leon y Coahuila de Zaragoza, Mexico. Universidad Autonoma de Nuevo Leon. Facultad de Ciencias Biologicas. Tesis licenciatura. México; 2009. pp. 47. 57. Ralph CJ, Geupel GR, PYLE P, Martin TE, Desante DF, Milá B. Manual de metodos de campo para el monitoreo de aves terrestres. General Technical Report PSW–GTR–159, USDA Forest Service, Albany; 1996. 58. Buckland ST Anderson DR Burnham KP Laake JL Borchers DL Thomas L Introduction to distance sampling 2001 3 Oxford Oxford University Press 466 59. Buckland ST Anderson DR Burnham KP Laake JL Borchers DL Thomas L Introduction to distance sampling 2004 4 Oxford Oxford University Press 435 60. Manning JA Factors affecting detection probability of burrowing owls in southwest agroecosystem environments J Wild Manag 2011 75 1558 1567 10.1002/jwmg.192 61. Manning JA Kaler RSA Effects of survey methods on burrowing owl behaviors J Wild Manag 2011 75 525 530 10.1002/jwmg.86 62. Thomas L Buckland ST Rexstad EA Laake JL Strindberg S Hedley SL Bishop JRB Marques TA Burnham KP Distance software: design and analysis of distance sampling surveys for estimating population size J Appl Ecol 2010 47 5 14 10.1111/j.1365-2664.2009.01737.x 20383262 63. Burnham KP Anderson DR Model selection and multimodel inference: a practical information-theoretic approach 2002 2 New York Springer 448 64. MCulloch CE, Searle SR, Neuhaus JM. Generalized, linear and mixed models. 2nd ed. Wiley: New York; 2008. 65. Ruizayma G, Olallakerstupp A, Maciasduarte A, Antonio G, Gonzalezrojas JI. Data from: population density of the burrowing owl (Athene cunicularia hypugaea) in Mexican prairie dog (Cynomys mexicanus) colonies at northeastern México. BMC Ecol. http://dx.doi.org/10.5061/dryad.pm362.

PMC005xxxxxx/PMC5002121.txt

==== Front Zoological LettZoological LettZoological Letters2056-306XBioMed Central London 5310.1186/s40851-016-0053-4Research ArticleSubstrate vibrations mediate behavioral responses via femoral chordotonal organs in a cerambycid beetle Takanashi Takuma takanasi@affrc.go.jp 1Fukaya Midori viridisetviridis@gmail.com 23Nakamuta Kiyoshi nakamuta@faculty.chiba-u.jp 16Skals Niels niels.skals@gmail.com 14Nishino Hiroshi nishino@es.hokudai.ac.jp 51 Department of Forest Entomology, Forestry and Forest Products Research Institute, Tsukuba, Japan 2 Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo, Japan 3 College of Bioresource Sciences, Nihon University, Fujisawa, Japan 4 Department of Geosciences and Natural Resource Management, University of Copenhagen, Frederiksberg C, Denmark 5 Research Institute for Electronic Science, Hokkaido University, Sapporo, Japan 6 Present address: Department of Horticulture, Chiba University, Matsudo, Japan 26 8 2016 26 8 2016 2016 2 1 1831 5 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 Vibrational senses are vital for plant-dwelling animals because vibrations transmitted through plants allow them to detect approaching predators or conspecifics. Little is known, however, about how coleopteran insects detect vibrations. Results We investigated vibrational responses of the Japanese pine sawyer beetle, Monochamus alternatus, and its putative sense organs. This beetle showed startle responses, stridulation, freezing, and walking in response to vibrations below 1 kHz, indicating that they are able to detect low-frequency vibrations. For the first time in a coleopteran species, we have identified the sense organ involved in the freezing behavior. The femoral chordotonal organ (FCO), located in the mid-femur, contained 60–70 sensory neurons and was distally attached to the proximal tibia via a cuticular apodeme. Beetles with operated FCOs did not freeze in response to low-frequency vibrations during walking, whereas intact beetles did. These results indicate that the FCO is responsible for detecting low-frequency vibrations and mediating the behavioral responses. We discuss the behavioral significance of vibrational responses and physiological functions of FCOs in M. alternatus. Conclusions Our findings revealed that substrate vibrations mediate behavioral responses via femoral chordotonal organs in M. alternatus. Electronic supplementary material The online version of this article (doi:10.1186/s40851-016-0053-4) contains supplementary material, which is available to authorized users. Keywords BehaviorVibrationSense organColeopterahttp://dx.doi.org/10.13039/501100001700Ministry of Education, Culture, Sports, Science, and Technology2458007515K07327Takanashi Takuma Fukaya Midori Council for Science, Technology and Innovation, Cross-ministerial Strategic Innovation Promotion Programthe FFPRI Encouragement Model in Support of Researchers with Family Responsibilitiesissue-copyright-statement© The Author(s) 2016 ==== Body Background Many animals are sensitive to substrate-borne vibrations. Vibration detection is an important sense that is used for intra- and interspecific interactions in diverse animal taxa [1, 2]. Specifically, vibrations transmitted through plants propagate well, allowing plant-dwelling animals to detect approaching conspecifics or predators without relying on other signals [3, 4]. Insects exhibit a range of behaviors in response to vibrations [1, 2]. A ‘startle response’ is a fast jerky movement with short latency elicited by vibrations; it is considered to be a preparatory behavior that enables locomotion to follow in a smooth behavioral sequence [5–7]. Vibrations may also elicit abrupt cessation of ongoing movements, such as freezing behavior or thanatosis (long-lasting freezing) [6–12]. The functional significance of vibration detection can be classified into: i) predator–prey interactions, including prey localization and antipredatory behavior, and ii) social interactions, including sexual signals, aggressive signals, and heterospecific signals [1, 2, 4, 13–17]. Although a number of studies have shown that coleopteran insects detect vibrations, and that they exhibit behavioral responses [11–16], the sense organs mediating such responses are largely unknown. An electrophysiological study showed that unidentified sensory neurons originating from the tibia and tarsus responded to low-frequency vibrations in Scarabaeidae, Carabidae, and Silphidae [18]. Studies of orthopteran insects showed that the primary organs sensitive to vibrations are internal mechanoreceptors, called ‘chordotonal organs’, located in the legs and other body appendages; these organs are also known to participate in the motor control of joints [8, 19–23]. The Japanese pine sawyer beetle, Monochamus alternatus (Coleoptera: Cerambycidae), is the vector of the pine wood nematode, which kills pine trees [24]. In this study, we investigated the behavioral responses of M. alternatus to vibrations and identified a chordotonal organ in the leg that detects vibrations transmitted through the tree. Methods Insects Dead pine trees, Pinus thunbergii and Pinus densiflora, infested with larvae of Monochamus alternatus were collected at the Forestry and Forest Products Research Institute and its Chiyoda Experimental Station in Ibaraki Prefecture, Japan in February to March during the 2006–2013 period, and were kept within a screen-caged house in natural conditions. In June through July, adults emerging from the dead pine logs were collected and kept separately in plastic cups (ca. 10 cm diam., 6 cm high) at 25 °C and 50–60 % relative humidity with a 16:8-h L:D cycle. A few twigs (ca. 10 cm long) of P. thunbergii and P. densiflora were provided as diet and replaced every 3–7 days. Male and female mature adults (>2 weeks old) were used. All behavioral experiments were performed at room temperature (23–26 °C) during the light-on period. Vibration stimuli and behavioral responses Pulsed sine waves of 100 ms duration ranging from 25 Hz to 10 kHz were shaped by commercial software (0105; NF Corp., Yokohama, Japan). The duration included 5 ms rise and 5 ms fall times, irrespective of frequency. The vibration stimuli, continuously looped back by a function generator (WF1945; NF Corp.) at intervals of 900 ms, were applied to a beetle via a vibration exciter (type 4809; Brüel & Kjær, Nærum, Denmark) connected to a type 2718 power amplifier. Frequencies (Hz) and amplitudes (zero-to-peak accelerations, m/s2) of vibrations were measured by attaching a piezoelectric charge accelerometer (type 4371 or 4393; Brüel & Kjær) to the center of the steel plate on which the beetle was placed. The signals were amplified by a type 2692 conditioning amplifier and displayed on an oscilloscope (DS-8822P; Iwatsu Test Instruments Corp., Tokyo, Japan). Behavioral responses of M. alternatus to vibration stimuli were observed under various conditions (Experiments 1–3), as follows. Experiment 1 An intact beetle was allowed to move freely on a flat steel plate (6 × 6 cm, 1 mm thick) attached to a vibration exciter placed on a desktop vibration isolator (UM-0405; Nippon Boushin Industry Co., Ltd, Numazu, Japan). Movements of the antennae and/or legs (startle response) and warning sound production using the prothorax and mesoscutum (stridulation) [16, 25] from a stationary position were categorized as behavioral responses to vibration stimuli in the inactive state. To determine the behavioral thresholds of these distinct responses, five male and five female beetles were used (Fig. 1a). The amplitude of the stimulus was gradually raised in 10-dB steps using the variable gain control of the power amplifier until a response appeared. The vibration-amplitude threshold was defined as the slightest acceleration necessary to evoke the response at a given frequency. When the response to vibration at a particular amplitude was clear, the response was counted. When the response to vibration was unclear, the same amplitude was applied again, with an inter-stimulus interval >1 min to avoid habituation. Six to nine different frequencies from 25 Hz to 10 kHz were presented to individuals in random order. After determining the threshold for a given frequency, subsequent thresholds for different stimulus frequencies were determined at intervals of >1 min. Immediately after each behavioral test, the acceleration on the center of the steel plate was measured as described above. No distinct differences in the thresholds were observed between the sexes; thus, the thresholds among different frequencies were analyzed in pooled individuals (n = 10) using the Kruskal–Wallis test, with multiple comparisons.Fig. 1 Setups for behavioral experiments. a An individual Monochamus alternatus beetle was allowed to move freely from a standstill on a steel plate attached to the vibration exciter (Experiment 1). b A naturally dried pine rod attached to the vibration exciter was tilted at ca. 70° from the horizontal. Freezing responses during walking or initiation of walking from a standstill were observed (Experiment 2). c The pine rod was attached vertically to the vibration exciter. Freezing responses were observed (Experiment 3) Experiment 2 Instead of the steel plate of Experiment 1, a rod of naturally dried pine trunk (3 cm diam., 30 cm long) attached to the vibration exciter with a screw (5 mm diam., 13 mm long) was hung from the ceiling of a soundproof box (90 × 90 × 70 cm) with a thick rope (Fig. 1b). The pine rod was tilted at ca. 70 ° from the horizontal. A beetle was gently transferred to the edge of the rod. Continuous waves of 100 Hz were applied to the beetle via the rod after it was allowed to rest or walk. Freezing responses during walking (i.e., cessation of walking), or initiation of walking from a stationary position were observed in intact beetles. Prior to the behavioral tests, vibration amplitudes as accelerations on the surface of the rod at 15 cm from the exciter were determined to be 0.03 m/s2. The numbers of freezing or walking behaviors in the presence of vibrations were compared with those in the absence of vibrations (control) using Fisher’s exact probability test. Experiment 3 A naturally dried pine rod was attached vertically to the vibration exciter, which was placed on the desktop vibration isolator (Fig. 1c). Freezing responses during walking were observed in beetles with operated femoral chordotonal organs (FCOs), in sham-operated beetles (with femoral integument damaged with microscissors), and in intact beetles. The conditions used in this experiment were more suitable for observing the response than those in Experiment 2, because it allowed the beetles to walk up and down the rod. The FCO- and sham-operated beetles were allowed to recover for at least 1 day. As described in Experiment 1, the amplitude of stimuli at a set frequency was gradually raised from 0.01 to 7 m/s2 until a freezing response appeared. From an intact beetle, the threshold of the freezing response was determined as described in Experiment 1. After each behavioral test on intact beetles, the acceleration on the surface of the rod at 15 cm from the exciter was measured. Differences in the response at the same frequency were determined by Fisher’s exact probability test and Ryan’s multiple comparison test. The thresholds among different frequencies were tested by the Kruskal-Wallis test. Although the exciter generated airborne sounds at frequencies above 500 Hz, the beetles did not exhibit any behavioral responses to sounds with similar frequencies and amplitudes broadcast from a speaker. Chordotonal organs The FCOs and other chordotonal organs were stained by backfills from the main leg nerve (n = 12). The beetle was briefly anesthetized with carbon dioxide and then fixed ventral-side-up on a beeswax plate using insect pins. To stain peripheral nerves in the leg, the main leg nerve was cut at the terminus of the thoracic ganglion and its peripheral cut end placed into the tip of a tapered glass capillary tube filled with a 1 % micro-Ruby solution (MW = 3000; Invitrogen, Carlsbad, CA, USA). After fixation in 4 % paraformaldehyde solution for 6 h, specimens were dehydrated through an ethanol series, cleared in methyl salicylate, and viewed under a confocal microscope (LSM510; Zeiss, Jena, Germany). The stained chordotonal organs are shown in a false color (green). Optical sections (1.2 μm thick) were reconstructed two-dimensionally using commercial software (Amira ver. 3.1; FEI Visualization Sciences Group, Burlington, VT, USA) linked to the LSM510. For FCO surgery, cell clusters (scoloparia) attached to the apodemes of all six femora of anesthetized beetles were carefully removed with microscissors immediately after opening a flap of the overlaying cuticle. The flaps were replaced to minimize damage to the surrounding muscles and tracheae. FCO-operated beetles were capable of walking although they exhibited deficits in the righting response (after turning them onto their backs) [26], which was slower than in intact beetles (6.8 s and 1.2 s, respectively; means of three measurements on two operated and two intact beetles) (Additional file 1: Video S1). Results Monochamus alternatus exhibited startle responses (twitch movements) and/or stridulation when subjected to vibrations at different frequencies (Fig. 2). Vibrations frequently evoked startle responses, and occasionally stridulation with or without a startle response. The threshold progressively increased with frequency from 25 Hz to 1 kHz (Fig. 2), although the thresholds from 3.5 to 23.5 m/s2 were not significantly different. Higher amplitudes >100 m/s2 were needed to evoke the responses between 2.5 kHz and 10 kHz. Thus, the beetles were most sensitive to low frequencies (<1 kHz).Fig. 2 Behavioral thresholds to vibrations in Monochamus alternatus. Startle response (twitch movements) and stridulation from a standstill were categorized as behavioral responses to pulsed sine waves with different frequencies and amplitudes. Thresholds (mean ± SEM) with the same letters are not significantly different by the Kruskal-Wallis test with multiple comparisons (p > 0.05). For setup, see Fig. 1a Freezing or walking in response to vibrations was observed significantly more frequently than in the absence of vibrations in the controls (Fig. 3). The behavioral choice was state-dependent. When low-amplitude vibrations at 100 Hz were presented, 53 % of the beetles froze during walking, while 47 % did not freeze. Similarly, 60 % of the beetles initiated walking from a standstill in response to the vibrations, while 40 % did not. In contrast, in the absence of vibrations (controls), 7 % of the beetles froze and 0 % walked. The others showed no response. Startle responses to vibrations were rarely observed prior to walking and were never observed prior to freezing.Fig. 3 Proportions of Monochamus alternatus responding to vibrations at 100 Hz. The behavioral responses, freezing and walking, were significantly different from the controls in the absence of vibrations by Fisher’s exact probability test (**, p < 0.001). The numbers in brackets represent the total number of individuals. For setup, see Fig. 1b Examination of backfills of the leg nerves at different proximo-distal levels revealed that the subgenual organ, a chordotonal organ specialized for vibration detection in orthopteran and heteropteran species [20, 27], is absent from all tibiae in M. alternatus (Fig. 4a–b), in agreement with a previous report on coleopteran species [18]. However, we identified the FCO as the largest chordotonal organ in each leg (Fig. 4a–g). The FCO shares general morphological characteristics with FCOs of other insects [20] but consists of a single scoloparium (cell cluster embedded in connective tissue) (Fig. 4h, i). The scoloparium of the FCOs in all legs was located in mid-femur (Fig. 4c–e), but the FCO scoloparium of the metathoracic leg was located more distally, resulting in a shorter cuticular apodeme than those of the pro- and mesothoracic legs. The FCOs contained about 60–70 sensory neurons (Fig. 4h, i) with no observable differences among the legs or between sexes. The FCO scoloparium was firmly attached to the anterior cuticle, and was distally connected to the proximal tibial joint via a ligament and the apodeme (Fig. 4c). A pair of neurons extended dendrites into a single scolopale cap (Fig. 4j), as previously reported for chordotonal sensilla of insect FCOs [20]. Compared with FCOs, the tibial and tarsal chordotonal organs were small and contained approximately 15 and six sensory neurons, respectively (Fig. 4k, l).Fig. 4 Leg chordotonal organs (COs) of Monochamus alternatus. a Photograph of female adult in resting posture, and diagram indicating the locations of COs in a femur (c), a tibia (b, f), and a tarsus (g). b Retrograde nerve staining of the prothoracic tibia showing absence of subgenual organ (b in a). c–e The FCO scoloparia (indicated by red arrows) in the prothoracic (c), mesothoracic (d), and metathoracic (e) legs. The cuticular apodeme is indicated by dotted lines. The FCO scoloparium in the metathoracic legs was located more distally compared with those in the pro- and mesothoracic legs. (f) The tibial CO in the distal region of the mesothoracic leg (f in a). g The tarsal CO in tarsal segment III of the metathoracic leg (g in a). h The prothoracic FCO scoloparium viewed posteriorly. i Three-dimensional reconstruction of the scoloparium h, viewed anteriorly. j High-power confocal stacks of the FCO in the mesothoracic leg. Pairs of cell bodies (arrowheads) inserted into common scolopale caps (arrows) are shown with different colors. k Magnified view of the tibial CO (f). l Magnified view of the tarsal CO (g) We compared freezing responses in FCO-operated, sham-operated, and intact beetles during walking (Fig. 5). When 100-Hz and 1-kHz vibrations were presented to intact beetles, a majority (95 and 65 %, respectively) showed freezing responses at various stimulus amplitudes. Similarly, 75 % of sham-operated beetles responded to 100 Hz. In contrast, only 9.5 and 0 % of FCO-operated beetles responded to 100 Hz and 1 kHz vibrations, respectively. Freezing responses to 100 Hz and 1 kHz differed significantly between the FCO-operated beetles and the other groups, but at 20 Hz the differences were not significant. The behavioral thresholds of intact beetles (mean ± SEM) were 0.41 ± 0.08 m/s2 at 20 Hz, 0.33 ± 0.06 m/s2 at 100 Hz, and 0.27 ± 0.06 m/s2 at l kHz (n = 8–18). There were no significant differences in thresholds among 20 Hz, 100 Hz, and l kHz.Fig. 5 Proportions showing freezing responses to vibrations during walking in femoral-chordotonal-organ (FCO)-operated, sham-operated, and intact beetles. Significant differences in responses at the same frequency were detected with Fisher’s exact probability test (**, p < 0.001) and Ryan’s multiple comparison test for proportions at 100 Hz (different letters, p < 0.05). The numbers in brackets represent the total number of individuals. n.e.: not examined. For setup, see Fig. 1c Discussion Monochamus alternatus showed startle responses and stridulation from a standstill, when subjected to a broad range of vibrations below 1 kHz. In addition, the beetles froze or walked in response to vibrations at 100 Hz. Beetles with operated FCOs did not show freezing behavior, suggesting that the FCOs detect low-frequency vibrations and mediate this behavior. This finding is in accordance with a report that the cricket Gryllus bimaculatus with all FCOs operated tended not to exhibit long-lasting freezing behavior [8]. Freezing behavior mediated by excitation of sensory neurons in FCOs seems widespread across insects. Monochamus alternatus showed different thresholds for the behavioral responses. Low-frequency vibrations of amplitudes 3.5–23.5 m/s2 induced the startle response and stridulation, whereas lower amplitudes of 0.3–0.4 m/s2 induced the freezing response. Similar differences in the stimuli needed to trigger the startle and freezing responses have been reported in P. fortunei [7]. Although the threshold was unclear, walking from a standstill was evoked by continuous vibrations with a low amplitude at 0.03 m/s2 in M. alternatus. Repeated exposures to vibrations above the threshold allow M. alternatus to walk, after initially showing the startle response. What is the behavioral significance of the vibrational responses in M. alternatus? Some of the responses may be associated with anti-predator behavior. Approaching predators (e.g., birds [28]) cause low-frequency vibrations through a tree, which may elicit the startle and freezing responses mediated by the FCO of M. alternatus. For example, the cerambycid beetle Hylotrupes bajulus [16] and other insects [5, 9, 11] exhibit these responses, presumably as a defense against predators. Freezing and motionless insects are capable of hiding without emitting vibrational and/or other cues to predators [10, 11]. In addition, M. alternatus stridulates in response to vibrations. In cerambycids and other beetles, stridulation is regarded as a defensive, disturbance, or warning signal to potential predators [25, 29]. Startle and freezing responses may also serve for conspecific recognition in M. alternatus. Detection of approaching conspecifics by their vibrations could allow insects to prepare for subsequent behaviors, e.g., escaping or mating [30]. P. fortunei are able to detect vibrations from conspecifics landing and walking on the host leaf [7]. In addition to vibrations, contact sex pheromones and visual cues play important roles in conspecific recognition in cerambycids [31, 32]. Hence, vibrations may play an important role in both inter- and intraspecific interactions in concert with other sensory cues. We identified for the first time the femoral chordotonal organ of a coleopteran species as a sensory organ detecting vibrations. In the absence of any specialized vibration detectors such as subgenual organs, the FCO, the largest leg chordotonal organ in M. alternatus, is suggested to play a pivotal role in the detection of low-frequency vibrations below 1 kHz. The FCO of M. alternatus possesses only a single scoloparium, which morphologically resembles the distal scoloparium in a locust [19, 20]. Considering that the distal scoloparia are sensitive to tibial movements and mediate reflexes in the leg muscles of a locust [19] and a stick insect [22], the FCOs of M. alternatus are likely to be bifunctional sensory organs that detect: i) small, fast small movements, e.g., accelerations through the tibia; and ii) large, slow movements, e.g., displacements of tibia. In fact, M. alternatus with operated FCOs took more time to right themselves (after turning them onto their backs) (Additional file 1: Video S1), an action that requires coordination of leg muscles. Possibly, pairs of neurons within a sensillum of the M. alternatus FCO have different physiological properties, as reported in the paired neurons of the antennal chordotonal organ in a cockroach [33]. Furthermore, the shorter apodeme of the metathoracic FCOs compared with the pro- and mesothoracic FCOs might be related to physiological properties (e.g., vibration detection, proprioception) in M. alternatus. Further studies are needed to determine the relationships between function and structure in M. alternatus FCOs. Conclusions Our findings revealed that a cerambycid beetle showed behavioral responses, such as startle and freezing, when subjected to vibrations. For the first time, the internal mechanoreceptors, ‘chordotonal organs’, responsible detecting vibrations in a coleopteran species was identified. Micro-ablation of the femoral chordotonal organs in all legs completely abolished vibration-mediated freezing behavior. Freezing behavior may be associated with defense against predators. Additional file Additional file 1: Video S1. Righting in FCO-operated and intact beetles. Beetles righted themselves after they were caused to fall upside down onto a sheet of paper. The movie was recorded with a Sony DCR-DVD201 Handycam at 30 frames per second. (MOV 3446 kb) Abbreviations COChordotonal organ FCOFemoral chordotonal organ MWMolecular weight SEMStandard error of mean We thank M. Jinkawa and Y. Suzuki (Forestry and Forest Products Research Institute) for the loan of the instruments. This work was supported in part by the Ministry of Education, Culture, Sports, Science and Technology KAKENHI Grant Nos. 80332477 (TT, HN), 24580075 (MF, TT), and 15 K07327 (MF), Council for Science, Technology and Innovation, Cross-ministerial Strategic Innovation Promotion Program (TT), and the FFPRI Encouragement Model in Support of Researchers with Family Responsibilities (TT). Funding This work was partly supported by Ministry of Education, Culture, Sports, Science and Technology KAKENHI Grant Nos. 80332477 (TT, HN), 24580075 (MF, TT), and 15 K07327 (MF), Council for Science, Technology and Innovation, Cross-ministerial Strategic Innovation Promotion Program (TT), and the FFPRI Encouragement Model in Support of Researchers with Family Responsibilities (TT). The funders had no role in the design of the study and collection, analysis, or interpretation of data, or in the writing of the manuscript. Availability of data and materials The datasets supporting the conclusions of this article are included within the article and its additional file. Authors’ contributions TT, NS, and HN designed the study. All authors wrote the manuscript. TT and MF carried out the behavioral experiments; HN carried out anatomical experiments; TT and KN collected and reared insects. 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 Laboratory-maintained insects were used in all experiments. Ethical approval and consent to participate were not required for this work. ==== Refs References 1. Hill PSM Vibrational Communication in Animals 2008 Cambridge Harvard University Press 2. Cocroft RB Rodríguez RL The behavioral ecology of insect vibrational communication Bioscience 2005 55 323 34 10.1641/0006-3568(2005)055[0323:TBEOIV]2.0.CO;2 3. Michelsen A Fink F Gogala M Traue D Plants as transmission channels for insect vibrational songs Behav Ecol Sociobiol 1982 11 269 81 10.1007/BF00299304 4. McVean A Field LH Communication by substratum vibration in the New Zealand tree weta, Hemideina femorata (Stenopelmatidae: Orthoptera) J Zool 1996 239 101 22 10.1111/j.1469-7998.1996.tb05440.x 5. Friedel T The vibrational startle response of the desert locust Schistocerca gregaria J Exp Biol 1999 202 2151 9 10409486 6. Bullock TH Eaton RC Comparative neuroethology of startle, rapid escape and giant fiber-mediated responses Neural mechanisms of startle behaviour 1984 New York Plenum Press 1 13 7. Tsubaki R Hosoda N Kitajima H Takanashi T Substrate-borne vibrations induce behavioral responses of a leaf-dwelling cerambycid Paraglenea fortune Zool Sci 2014 31 789 94 10.2108/zs140029 25483790 8. Nishino H Sakai M Behaviorally significant immobile state so called of thanatosis in the cricket Gryllus bimaculatus DeGeer: its characterization, sensory mechanism and function J Comp Physiol A 1996 179 613 24 10.1007/BF00216126 9. Rohrseitz K Kilpinen O Vibration transmission characteristics of the legs of freely standing honeybees Zoology 1997 100 80 4 10. Miyatake T Katayama K Takeda Y Nakashima A Sugita A Mizumoto M Is death-feigning adaptive? Heritable variation in fitness difference of death-feigning behaviour Proc R Soc Lond B 2004 271 2293 6 10.1098/rspb.2004.2858 11. Kojima W Ishikawa Y Takanashi T Deceptive vibratory communication: pupae of a beetle exploit the freeze response of larvae to protect themselves Biol Lett 2012 8 717 20 10.1098/rsbl.2012.0386 22675138 12. Acheampong S Mitchell BK Quiescence in the Colorado potato beetle, Leptinotarsa decemlineta Entomol Exp Appl 1997 82 83 9 10.1046/j.1570-7458.1997.00116.x 13. Hanrahan SA Kirchener WH Acoustic orientation and communication in Desert tenebrionid beetles in sand dunes Ethology 1994 97 26 32 10.1111/j.1439-0310.1994.tb01026.x 14. Goulson D Birch MC Wyatt TD Mate location in the deathwatch beetle, Xestobium rufovillosum De Geer (Anobiidae): orientation to substrate vibrations Anim Behav 1993 47 899 907 10.1006/anbe.1994.1122 15. Kojima W Takanashi T Ishikawa Y Vibratory communication in the soil: pupal signals deter larval intrusion in a group-living beetle Trypoxylus dichotoma Behav Ecol Sociobiol 2012 66 171 9 10.1007/s00265-011-1264-5 16. Breidbach O Studies on the stridulation of Hylotrupes bajulus (L.) (Cerambycidae, Coleoptera): communication through support vibration- morphology and mechanics of the signal Behav Processes 1986 12 169 86 10.1016/0376-6357(86)90055-0 24897351 17. Travassos MA Pierce NE Acoustics, context and function of vibrational signalling in a lycaenid butterfly-ant mutualism Anim Behav 2000 60 13 26 10.1006/anbe.1999.1364 10924199 18. Schneider W Über den Erschütterungssinn von Käfern und Fliegen Z vergl Physiol 1950 32 287 302 10.1007/BF00340690 19. Field LH Pflüger HJ The femoral chordotonal organ: a bifunctional orthopteran (Locusta migratoria ) sense organ? Comp Biochem Physiol 1989 93A 729 43 10.1016/0300-9629(89)90494-5 20. Field LH Matheson T Chordotonal organs in insects Adv Insect Physiol 1998 27 1 228 10.1016/S0065-2806(08)60013-2 21. Field LH Burrows M Reflex effects of the femoral chordotonal organ upon leg motor neurones of the locust J Exp Biol 1982 101 265 85 22. Stein W Sauer E Physiology of vibration-sensitive afferents in the femoral chordotonal organ of the stick insect J Comp Physiol A 1999 184 253 63 10.1007/s003590050323 23. Nishino H Sakai M Field LH Two antagonistic functions of neural groups of the femoral chordotonal organ underlie thanatosis in the cricket Gryllus bimaculatus DeGeer J Comp Physiol A 1999 185 143 55 10.1007/s003590050373 24. Kobayashi F Yamane A Ikeda T The Japanese pine sawyer beetle as the vector of pine wilt disease Ann Rev Entomol 1984 29 115 35 10.1146/annurev.en.29.010184.000555 25. Alexander RD Moore TE Woodruff RE The evolutionary differentiation of stridulatory signals in beetles (Insecta: Coleoptera) Anim Behav 1963 11 111 2 10.1016/0003-3472(63)90018-6 26. Frantsevich L Righting kinematics in beetles (Insecta: Coleoptera) Arthropod Struct Dev 2004 33 221 35 10.1016/j.asd.2004.05.007 18089036 27. Nishino H, Mukai H, Takanashi T. Chordotonal organs in hemipteran insects: unique peripheral structures but conserved central organization revealed by comparative neuroanatomy. Cell Tiss Res. (in press). 28. Inoue M Predation of the Japanese pine sawyer (Monochamus alternatus ) by wild birds Bull Tottori Pre For Exp Stn 1987 30 47 71 29. Masters WM Insect disturbance stridulation: characterization of airborne and vibrational components of the sound J Comp Physiol A 1980 135 259 68 10.1007/BF00657254 30. Fauziah BA Hidaka T Tabata K The reproductive behaviour of Monochamus alternatus Hope Appl Entomol Zool 1987 22 272 85 31. Allison JD Borden JH Seybold SJ A review of the chemical ecology of the Cerambycidae (Coleoptera) Chemoecology 2004 14 123 50 10.1007/s00049-004-0277-1 32. Fukaya M Akino T Yasuda T Yasui H Wakamura S Visual and olfactory cues for mate orientation behaviour in male white-spotted longicorn beetle, Anoplophora malasiaca Entomol Exp Appl 2004 111 111 5 10.1111/j.0013-8703.2004.00147.x 33. Ikeda S Toh Y Okamura J Okada J Intracellular responses of antennal chordotonal sensilla of the American cockroach Zool Sci 2004 21 375 83 10.2108/zsj.21.375 15118224

PMC005xxxxxx/PMC5002122.txt

==== Front BMC Health Serv ResBMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 168210.1186/s12913-016-1682-1ErratumErratum to: Development of a nurse home visitation intervention for intimate partner violence Jack Susan M. jacksm@mcmaster.ca 189Ford-Gilboe Marilyn 2Wathen C. Nadine 39Davidov Danielle M. 49McNaughton Diane B. 5Coben Jeffrey H. 69Olds David L. 7MacMillan Harriet L. 89for NFP IPV Research Team 1 School of Nursing, McMaster University, 1280 Main Street West, Hamilton, ON Canada 2 Arthur Labatt Family School of Nursing, The University of Western Ontario, London, ON Canada 3 Faculty of Information and Media Studies, The University of Western Ontario, London, ON Canada 4 Department of Emergency Medicine, West Virginia University, Morgantown, WV USA 5 College of Nursing, Rush University, Chicago, IL USA 6 Departments of Emergency Medicine and Community Medicine, West Virginia University, Morgantown, WV USA 7 Prevention Research Center for Family and Child Health, University of Colorado Denver, Denver, CO USA 8 Offord Centre for Child Studies, McMaster University, Hamilton, ON Canada 9 West Virginia University Injury Control Research Center, Morgantown, WV USA 26 8 2016 26 8 2016 2016 16 1 445© 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 For full transparency, the authors wish to update the competing interest statement of this article [1] with the following information: The Prevention Research Center for Family and Child Health (PRC), directed by Dr. David Olds at the University of Colorado School of Medicine, has a contract with the Nurse-Family Partnership© to conduct research to improve the NFP program and its implementation. Dr. David Olds was employed by this center at the time the current study was conducted. The contract with the NFP contributes to the salaries and operations of the PRC. Dr. Olds is the founder of the Nurse-Family Partnership. The online version of the original article can be found under doi:10.1186/1472-6963-12-50. ==== Refs Reference 1. Jack Development of a nurse home visitation intervention for intimate partner violence BMC Health Services Research 2012 12 50 10.1186/1472-6963-12-50 22375908

PMC005xxxxxx/PMC5002123.txt

==== Front J Cardiothorac SurgJ Cardiothorac SurgJournal of Cardiothoracic Surgery1749-8090BioMed Central London 53410.1186/s13019-016-0534-8Case ReportOpen-heart surgery using a centrifugal pump: a case of hereditary spherocytosis http://orcid.org/0000-0002-5931-2020Matsuzaki Yuichi +81-3-3353-8111bokumatsuzaki@hotmail.co.jp Tomioka Hideyuki stomioka@hij.twmu.ac.jp Saso Masaki masaki1021@gmail.com Azuma Takashi azumaro2008@gmail.com Saito Satoshi ssaitos@hij.twmu.ac.jp Aomi Shigeyuki saomi@hij.twmu.ac.jp Yamazaki Kenji syamazak@hij.twmu.ac.jp The Heart Institute Japan, Department of Cardiovascular Surgery, Tokyo Women’s Medical University, 8-1 Kawada-cho, Shinjyuku-ku, Tokyo, 162-8666 Japan 26 8 2016 26 8 2016 2016 11 1 1387 3 2016 23 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 Hereditary spherocytosis is a genetic, frequently familial hemolytic blood disease characterized by varying degrees of hemolytic anemia, splenomegaly, and jaundice. There are few reports on adult open-heart surgery for patients with hereditary spherocytosis. Case presentation We report a rare case of an adult open-heart surgery associated with hereditary spherocytosis. A 63-year-old man was admitted for congestive heart failure due to bicuspid aortic valve, aortic valve regurgitation, and sinus of subaortic aneurysm. The family history, the microscopic findings of the blood smear, and the characteristic osmotic fragility confirmed the diagnosis of hereditary spherocytosis. Furthermore, splenectomy had not been undertaken preoperatively. The patient underwent a successful operation by means of a centrifugal pump. Haptoglobin was used during the cardiopulmonary bypass, and a biological valve was selected to prevent hemolysis. No significant hemolysis occurred intraoperatively or postoperatively. Conclusion There are no previous reports of patients with hereditary spherocytosis, and bicuspid aortic valve. We have successfully performed an adult open-heart surgery using a centrifugal pump in an adult patient suffering from hereditary spherocytosis and bicuspid aortic valve. Keywords Hereditary spherocytosisAdult cardiacBicuspid aortic valveCardiopulmonary bypassHaptoglobinCentrifugal pumpissue-copyright-statement© The Author(s) 2016 ==== Body Background Hereditary spherocytosis (HS) is a genetic, frequently familial hemolytic blood disease characterized by varying degrees of hemolytic anemia, splenomegaly, and jaundice. The disease is associated with various defects in any of the number of the proteins responsible for maintaining the shape and flexibility of the red blood cell (RBC), resulting in osmotically fragile and characteristically spherical RBCs. Cardiopulmonary bypass (CPB) can exacerbate hemolysis and subsequent renal dysfunction. There are few reports on open-heart surgery for adult patients with HS. This study reports the case of a 63-year-old man with HS who underwent aortic valve replacement, valval aneurysm patch closure, and ascending aorta replacement. Case presentation The patient was a 63-year-old man with HS, idiopathic aortic bileaflet, and subaortic aneurysm (SAA) caused by infective endocarditis. The SAA was incidentally discovered during a clinical evaluation for aortic regurgitation (AR); this was due to idiopathic aortic bileaflet and infective endocarditis that were diagnosed 10 years ago. Because of the patient’s HS history and severe anemia, a cardiologist conducted the medical follow-up. The patient’s LV systolic function worsened gradually and he had frequent episodes of dyspnea. Moreover, transapical echocardiography showed a decrease in the AR grading. On admission, the hemoglobin level was 8.7 mg/dL, hematocrit was 23.8 %, platelet count was 18.2 × 104/μL, blood urea nitrogen was 15.1 mg/dL, and creatinine was 0.93 mg/dL. The physical examination revealed an anemic and icteric conjunctiva, and his skin appeared jaundiced. The laboratory findings of the patient’s blood smear revealed a type 1 + polychromasia. The transthoracic echocardiogram showed a bicuspid aortic valve (adhesions on the left and right coronary cusps), a 10-mm SAA prolapsing at the atrio-ventricular continuity, an ascending aorta measuring 38 mm, and a grade IV central AR (Fig. 1). The left ventricular ejection fraction was preserved. The computer tomography angiography revealed a right coronary sinus aneurysm in the sagittal and axial plane.Fig. 1 Subaortic aneurysm showed by TTE A surgical intervention was indicated because of severe AR due to the bicuspid aortic valve, moderate dilation of ascending aorta, and SAA. The procedure was performed through a median sternotomy. A CPB was instituted by ascending aortic cannulation and bicaval drainage. Left atrial venting was carried out from the right upper pulmonary vein. Cardioplegia was performed by selective antegrade and retrograde perfusion with blood Frem’s solution and cold blood. Before initiating the CPB, we transfused four units of RBCs (hemoglobin of 12.1 mg/dL). Furthermore, human haptoglobin was also transfused. At the time of the CPB, the centrifuge pump was selected rather than the roller pump, and perfusion index was 2.5 L/m2/min. First, the epicardial membrane was used for the repair of the SAA (Figs. 2 and 3). Second, aortic valve replacement was performed by using the biological valve (Magna EASE 21 mm, Carpentier-Edwards, Japan). Third, an ascending aorta replacement was performed under moderate hypothermia and cross-clamping condition. The aortic cross-clamp time was 182 min, and CPB time was 234 min. The urine remained clear throughout and after CPB, indicating that no significant hemolysis had occurred during the procedure. The patient was extubated on day 1, and all the vasoactive infusions were discontinued on day 3 postoperatively.Fig. 2 Subaortic aneurysm Fig. 3 Repair of the Subaortic aneurysm (SAA) Despite the 1.5 L RBC transfusion during the operation, the immediate postoperative blood smear did not show any spherocytes and other abnormal RBC morphologies. Postoperative bleeding was minimal, and further RBC transfusions were performed on day 4 and 14 postoperatively. The echocardiography showed no residual AR, and the patient was discharged from the hospital on day 18 without any hemolysis. At the 6-month follow-up, he was doing well. Recent echocardiography showed no residual AR, and good left ventricular function. Hematogenic investigation showed the hemoglobin level to be maintained at 10.6 g/dL without any transfusion. The total bilirubin and lactate dehydrogenase levels were 3.9 mg/dL, and 303 IU/L, respectively (Table 1).Table 1 Changes in the laboratory data Before operation Ope Day 3 Day 10 Day 18 6 months RBC (106/μL)) 243 339 265 276 362 295 Hb (g/dl) 8.7 10.5 7.9 8.2 11 10.6 Hct (%) 23.8 29.8 24.7 25.1 32.9 28.7 T.B. (mg/dL) 3.3 9.8 6.6 1.6 2 3.9 LDH (IU/L) 141 265 380 262 302 303 Haptoglobin 24 <10 <10 <10 <10 22 Discussion Hereditary spherocytosis, an autosomal dominant or recessive trait most commonly (though not exclusively) found in the Northern European and Japanese families, affects one in 2000 individuals [1]. There are few publications about the HS management during an open-heart surgery and there were no reports of patients with HS, bicuspid aortic valve, and SAA. A variety of approaches have been proposed for HS patients to avoid hemolysis during surgery. These include cardiac surgery without CPB, preemptive splenectomy [2], administration of haptoglobin to reduce plasma-free hemoglobin [3, 4], use of poloxamer 188 (a non-ionic antihemolytic detergent that protects RBC membrane during CPB) [5], or simply proceeding with surgery employing CPB without any special measures [5–7]. No significant hemolysis or renal failure has been reported in these instances and no trial has compared these various approaches. It must be noted that in Japan, poloxamer 188 is not permitted for use on humans; therefore, we did not consider this when analyzing our case. Although splenectomy is the only treatment for HS, we did not proceed because this patient had the high operative risk and he did not present with splenomegaly. Our patient had multiple risks contraindicating cardiac surgery; he had a bicuspid aortic valve and an enlarged ascending aorta, which necessitated a valve change. In addition, he had SAA and his HB level was low (8.7 mg/dL); thus, a straightforward procedure was not possible. For all these reasons and for the clinical safety of the patient, we had to consider other alternatives. To reduce the risk of hemolysis during operation, we had four measures as follows. First, RBC (HB of 10 mg/dL) transfusion was performed preoperatively. Second, aortic valve replacement was performed by using a biological valve, probably less hemolysis than mechanical valve. Third, we administered haptoglobin. Because haptoglobin can link to the free hemoglobin and change the complex form, this link can help the hepatic metabolism and prevent renal damage. In this case, although haptoglobin was transfused during and after the operation, haptoglobin level remained low. This could be explain by the fact that the speed of hemolysis for an HS patient is high, implying that the patient requires substantial amount of haptoglobin. Finally, we used a centrifugal pump because the risk of hemolysis is less with a centrifugal pump rather than with a roller pump [8]. The hematology-related approach, the improvement in surgical techniques, and the CPB technology will certainly reduce the complications associated with surgical intervention in patients with HS. Conclusion The outcome in this case was good as evidenced by the no residual AR, lack of hemolysis, and the preserved splenic function. We can conclude that we have successfully performed an adult open-heart surgery by using a centrifugal pump in a patient suffering from HS and bicuspid aortic valve. Abbreviations ARAortic regurgitation CPBCardiopulmonary bypass HSHereditary spherocytosis RBCRed blood cell SAASubaortic aneurysm Funding This study did not avail of any source of funding. Author’s contributions All authors read and approved the final manuscript. Competin interest The authors declare that they have no competing interests. Ethics approval and consent to participate Written informed consent was obtained from the patient for the publication of this report and of any accompanying images. Furthermore, the research committee of our institute approved this study. ==== Refs References 1. Perrotta S Gallagher PG Mohandas N Hereditary spherocytosis Lancet 2008 372 1411 26 10.1016/S0140-6736(08)61588-3 18940465 2. Gayyed NL Bouboulis N Holden MP Open heart operation in patients suffering from hereditary spherocytosis Ann Torac Surg 1993 55 1497 500 10.1016/0003-4975(93)91097-7 3. Kaminishi Y Atsumi N Terada Y Nakamura K Gomi S Mitsui T Anatomic correction of double-outlet right ventricle associated with hereditary spherocytosis-a case report Nihon Kyobu Geka Gakkai 1996 44 2164 71 4. Kawahira Y Kishimoto H Iio M Ikawa S Ueda H Kayatani F Open heart operation in a young child with spherocytosis Ann Thorac Surg 1994 58 1166 8 10.1016/0003-4975(94)90481-2 7944773 5. Aoyagi S Kawano H Tomoeda H Hiratsuka R Kawara T Open heart operation in a patient with hereditary spherocytosis: a case report Ann Thorac Cardiovascular Surg 2001 7 375 7 6. Moyes DG Rogers MA Coleman AJ Cardiopulmonary bypass in hereditary spherocytosis: a case report Thorax 1971 26 131 2 10.1136/thx.26.1.131 5101263 7. Moyes DG Holloway AM Hutton WS Correction of Fallot’s tetralogy in a patient suffering from hereditary spherocytosis S Afr Med J 1974 48 1535 6 4419248 8. Onitsuka T Nakamura K Kuwabara M Yonezawa T Shibata K Koga Y Mitral and aortic valve replacement with tricuspid annuloplasty in a patient suffering from hereditary spherocytosis Gen Thorac Cardiovasc Surg 1991 39 8 1184 7

PMC005xxxxxx/PMC5002124.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 338110.1186/s12889-016-3381-6Research ArticleActive children through incentive vouchers – evaluation (ACTIVE): a mixed-method feasibility study http://orcid.org/0000-0003-1117-6127Christian Danielle 01792606716d.l.christian@swansea.ac.uk 1Todd Charlotte c.e.todd@swansea.ac.uk 1Hill Rebecca r.a.hill@swansea.ac.uk 1Rance Jaynie j.y.rance@swansea.ac.uk 2Mackintosh Kelly k.mackintosh@swansea.ac.uk 3Stratton Gareth g.stratton@swansea.ac.uk 3Brophy Sinead s.brophy@swansea.ac.uk 11 Swansea University Medical School, Swansea University, Swansea, SA2 8PP UK 2 College of Health and Human Science, Swansea University, Swansea, SA2 8PP UK 3 College of Engineering, Bay Campus, Swansea University, Swansea, SA1 8EN UK 26 8 2016 26 8 2016 2016 16 1 8902 4 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.Background Adolescents face many barriers to physical activity, demonstrated by the decline in physical activity levels in teenage populations. This study aimed to assess the feasibility of overcoming such barriers via the implementation of an activity-promoting voucher scheme to teenagers in deprived areas. Methods All Year 9 pupils (n = 115; 13.3 ± 0.48 years; 51 % boys) from one secondary school in Wales (UK) participated. Participants received £25 of activity vouchers every month for six months for physical activity or sporting equipment. Focus groups (n = 7), with 43 pupils, and qualitative interviews with teachers (n = 2) were conducted to assess feasibility, in addition to a process evaluation utilising the RE-AIM framework. Quantitative outcomes at baseline, five months (during intervention) and twelve months (follow-up) included: physical activity (accelerometer), aerobic fitness (12 min Cooper run) and self-reported activity (PAQ-A). Motivation to exercise (BREQ-2) was measured three months post-baseline and at follow-up. Results Qualitative findings showed that vouchers encouraged friends to socialise through activity, provided opportunities to access local activities that pupils normally could not afford, and engaged both those interested and disinterested in physical education. Improvements in weekend moderate-to-vigorous physical activity and reductions in sedentary behaviour were observed in both sexes. Boys’ fitness significantly improved during the voucher scheme. ‘Non-active’ pupils (those not meeting recommended guidelines of 60 mins∙day−1) and those with higher motivation to exercise had higher voucher use. Conclusions Adolescents, teachers and activity providers supported the voucher scheme and felt the vouchers enabled deprived adolescents to access more physical activity opportunities. Voucher usage was associated with improved attitudes to physical activity, increased socialisation with friends and improved fitness and physical activity; presenting interesting avenues for further exploration in a larger intervention trial. Electronic supplementary material The online version of this article (doi:10.1186/s12889-016-3381-6) contains supplementary material, which is available to authorized users. Keywords AdolescentTeenagerPhysical activityAccelerometerVoucherDeprivedMixed methodshttp://dx.doi.org/10.13039/501100000374British Medical AssociationJoan Dawkins Grantissue-copyright-statement© The Author(s) 2016 ==== Body Background Physical inactivity in youth has been linked to a number of health issues in later life [1] and is considered a key contributor to increases in obesity [2]. This is concerning since a notable decline in physical activity (PA) is observed during adolescence [3] and lower PA levels, fitness and higher cardiometabolic risk scores are reported in children in deprived areas [4–7]. Therefore, approaches to address physical inactivity in this target group are urgently required. The main barriers to PA for adolescents are reported to be cost, accessibility, lack of parental support and lack of local facilities [8–10]. The Households Below Average Income data (HBAI) for the United Kingdom (UK) showed that the most disadvantaged families were unable to purchase leisure equipment for their children, such as sports kits or bicycles, due to prohibitive costs, and nearly a quarter of disadvantaged families reported a lack of outdoor space or facilities for their children to play safely [11]. Participating in organised activities more than once a week is associated with improved fitness [12]. However, participation is often lower for children in deprived areas and children with inactive parents [13]. Additionally, there is a difference between what young people would like to do and what is available [14], meaning they may not have the opportunity to participate in desired activities. For example, research has reported that girls often see competitive sports as a barrier to participation [15]. Therefore, if the majority of activities available are competitive, there is likely to be a lower uptake and lower physical activity among girls. This study examines whether placing decision making with young people, through activity vouchers, can increase empowerment and engagement to shape activity provision. Previous research has reported positive effects on PA levels following provision of activity vouchers to adults in socially deprived communities [16, 17]. However, the extent to which behaviour change can be sustained following such schemes remains uncertain [16, 18]. Furthermore, vouchers were used as an adjunct to motivational interviewing or provision of PA information in these studies. Provision of PA vouchers as a stand-alone intervention has not been evaluated among adolescents. Recent research among primary school children reported a positive increase in the percentage of time spent using active transport following the incentive of entry into a cash or voucher prize draw for utilising active transport every weekday [19]. Increases in fitness centre use have also been found among college students following monetary incentives [20]. However, these two previous studies [19, 20] focused on financial reward for participation rather than direct provision of activity enabling vouchers. This study aimed to conduct a phase 1 feasibility study of the ACTIVE (Active Children through Incentive Vouchers - Evaluation) scheme among adolescents in a deprived community, in order to assess acceptability of the study design and explore the effects of physical activity, fitness and motivation to exercise. Methods Recruitment One secondary school in a deprived catchment area in South Wales (UK) was approached to take part in the scheme. The school was classified as deprived as it: i) reported 54 % eligibility for free school meals (FSM) [21]; ii) was eligible for ‘Community Regeneration Initiatives’, ‘Communities First Funding’ and ‘Objective One’ priority areas; and iii) is in one of the more socially deprived areas in Wales [22]. All Year 9 pupils (13.3 ± 0.48 years) were eligible to receive activity vouchers (n = 115, 59 boys; 49 eligible for FSM). Intervention The intervention involved provision of £25 of vouchers (five vouchers in increments of £5) per month for six months. Vouchers could be used to: i) enrol in existing activities; ii) fund coaches or new activities directly in communities or at their school, such as Zumba and Boxercise and iii) purchase new sporting equipment for themselves or their school. Recognised providers (i.e., leisure centres, clubs, and dance providers) were recruited during development stages, and their logos were printed on the vouchers to enable easy identification of where they could be used. Vouchers were numbered and contained the participant’s name with a section to complete the type and duration of activity and the signature of the participant and provider. Vouchers were treated like a cash transaction, but no change was provided in order to prevent non-PA-related purchases. At the end of the month, all vouchers were collected from each provider with accompanying invoices for monetary reimbursement. A trained facilitator regularly attended the school to highlight activities available in the area, provide advice on how best to access activities, and discuss methods of overcoming individual barriers to exercise. The facilitator liaised between local sporting providers and pupils to ensure vouchers were redeemable when accessing facilities and identify new coaches or facilities. Outcomes Qualitative outcomes Prior to the intervention, one focus group was conducted with Year 9 pupils (n = 10, 5 boys) to examine how best to introduce the scheme, establish preferred PA providers, and identify potential barriers to participation (Fig. 1). Pupils were purposively allocated to focus groups depending on sex and deprivation, but were then selected randomly to participate from these groups by selecting every 10th pupil from the separate lists derived from the purposive sampling. Deprivation was classified at an individual level through FSM eligibility [21].Fig. 1 ACTIVE intervention timeline Two experienced researchers were present during focus groups (DC & SB or DC & RH), which took place in an empty classroom and lasted one hour on average. Each session was digitally-recorded following consent from pupils and parents, for later transcription, and followed a semi-structured topic guide (Fig. 2). No a priori hypothesis was determined, and themes and codes emerged through data analysis following transcription. All subsequent focus groups followed the same methodology with five to six pupils present at any one time [23].Fig. 2 Example of topic guide for focus groups and interviews A second focus group was conducted mid-intervention with five adolescents who were not using their vouchers, in order to identify barriers to use and suggest changes to improve uptake. Additionally, four focus groups (two girl groups (n = 11) and two boy groups (n = 12)) and two interviews with school teachers (Head of Year 9 and Head of P.E.) were conducted post-intervention. These interviews followed a similar line of questioning to that of the pupils (Fig. 2) but mainly concentrated on the teachers’ views of the scheme and whether they had perceived any changes in their pupils’ behaviour as a result of ACTIVE. A final focus group (n = 5) was conducted during the six month follow-up to examine longer-term changes in PA. Quantitative outcomes Quantitative outcomes included objectively measured PA, self-reported PA and aerobic fitness. More details can be found in Additional file 1. Data analysis Thematic analysis was used to analyse qualitative data. Focus group and interview transcripts were coded independently by two members of the research team and discrepancies were resolved through discussion. The themes and findings were presented back to participants during follow-up as suggested in the Consolidated Criteria for reporting qualitative research (COREQ) checklist [24]. Results The RE-AIM framework was used to evaluate the reach, effectiveness, adoption, implementation and maintenance of the project [25]. Reach All 115 pupils in Year 9 were eligible to participate and were provided with ACTIVE vouchers. At least one voucher was used by 81 % (48/59) of boys and 77 % (43/56) of girls. At least one voucher was used by 76 % (37/49) of deprived participants and 83 % (53/64) of non-deprived. Mid-intervention reasons for non-participation were lack of time, lack of opportunity to use vouchers, and apprehension over where vouchers were accepted. Efficacy/effectiveness Voucher usage During ACTIVE, 1464/3450 (42.6 %) of vouchers were used. Boys used more vouchers than girls (807 boys: 657 girls). Girls were more likely to use their vouchers to buy equipment than boys, 61 % of vouchers compared to 42 %, respectively [Difference 18.9 % (95 % CI: 13.8 %–23.8 %)]. Popular uses included purchasing scooter equipment (22.2 % boys, 38.5 % girls), football sessions (16.7 % boys), skateboarding and skateboard equipment (15.4 % boys), rollerblades (11.3 % girls), and attending the local waterpark (21.9 % boys, 13.7 % girls) (Table 1). Vouchers were predominantly used to attend pre-existing activities in groups or to buy equipment, as opposed to funding new coaches. Pupils did suggest activities or new activity providers to the facilitator but did not tend to organise new coaches or activities themselves. There was a preference for voucher expenditure at weekends or during half-terms, averaging 11 vouchers per weekend day and eight vouchers per half-term day. Normal school days showed the least voucher usage with an average of only five vouchers being spent per weekday.Table 1 Frequency data of voucher usage stratified by gender Activity/Equipment Total Girls Boys Assault Course 31 31 0 Badminton 5 5 0 Basketball 2 0 2 Basketball Equipment 10 2 8 Boxercise 1 0 1 Boxing 2 0 2 Boxing Equipment 4 0 4 Climbing 9 7 2 Cycling accessories 29 9 20 Football 135 0 135 Football Equipment 2 2 0 Gym (One off session) 37 28 9 Gym (Membership) 18 13 5 Laserzone 135 66 69 Rollerblades 88 74 14 Rugby Equipment 31 12 19 Scooter Equipment 432 253 179 Scootering 24 16 8 Skateboard Equipment 107 34 73 Skateboarding 55 4 51 Surfing Equipment 10 0 10 Swim Equipment 2 2 0 Swimming 23 9 14 Tennis Equipment 1 0 1 Waterpark 267 90 177 Missing data 4 0 4 Total 1464 657 807 Evidence of impact on PA and fitness A brief overview of the evidence of impact will be presented here, but further details can be found in Additional file 1. A marginal increase in moderate-to-vigorous PA (MVPA) was observed during the scheme. However, when stratified by day of the week, weekend MVPA showed a significant increase during the intervention in both sexes. A significant decrease in sedentary behaviour (SB) was reported overall during the intervention, though this was not sustained twelve months post-baseline. Furthermore, the intervention was associated with improvements in fitness post-intervention; significant for boys. Adoption Of the 24 activity providers participating, 16 were utilised (Table 1). Providers not used were those initiated by researchers at baseline and were predominantly structured activities including; dance classes, swimming clubs, gymnastics classes, and football coaching. Implementation The ACTIVE scheme was implemented as intended with pupils receiving vouchers through school teachers. Positive feedback was received from teachers, who reported the scheme was feasible to run through school and was not too intrusive on school time. Maintenance Quantitative data showed promise for sustainability both in terms of engagement and impacts on health behaviours (Additional file 1), though these need to be assessed on a larger scale before accepting conclusiveness. Qualitative responses regarding the scheme are discussed below with pupil and teacher quotes presented in Tables 2 and 3, respectively.Table 2 Qualitative data from pupils regarding positive aspects of the voucher scheme and recommendations for improvement Positive aspects of the voucher scheme Removal of financial barrier “…like when you go like with your friends and you don’t have the vouchers you like have got to worry about what you’re gonna pay, who’s got money, who hasn’t, but like this just everyone’s got it and you can just go…” (girl) Changed attitudes towards physical activity “…Well I didn’t really like sports until I started using the vouchers. I do a bit more sport than I used to…” (girl) “…The girls, they do gym and all that like when you goes out, you just sees them all the time…” (boy). “…I think it has worked because some people I think have got a gym membership after they went to the gym and when they stopped, when the vouchers stopped, I think they carried on going which they wouldn’t have before we had the vouchers…” (girl) Increased choice / flexibility “…With the vouchers we can do it anytime… you get more of a choice with the vouchers…” (boy) “…That’s why nobody likes P.E. because it forces them to do things they don’t want to do…”(girl) “…I like this input, yeah, like get a say in what we’d like…”(girl) “…that’s why I chose equipment because when it (the vouchers) finishes you can stop doing things but if you’re broke that much you can still use it…” (girl) "It’s cheaper to like pay for activities, like to go swimming and that than for like scooters and that’s more expensive so the voucher would be more better for equipment…"(girl) Increased awareness of local facilities “…I don’t know, the stuff that was on the back of it, it looked, fair play, it was like you want to go and do it so basically that’s what I went in to do in places and I did stuff. Because they’re on the back you can do it so why not try it out…”(girl) Increased social interaction “…It builds your friendship more as well from places…because like it’s doing something fun and you also doing it with your friends, like mixing with them…”(boy) "It’s made everyone active because the people who sometimes aren’t active might be friends with someone who is active and they have more opportunities to go and do different stuff (laughter) that they haven’t done before with the vouchers. And then they get like the unfit people and drag them along because they’re unfit…"(girl) “…you can do exercise in places like I enjoy going to LaserZone, I find it really fun, there for like half an hour and I’m sweating pints so…” (boy) “…because we’ve got the money, we’re not on the streets just making trouble, we’re out doing something…” (boy) Increased competence / body image “…Personally I think it’s helped me because I’ve become more fit and better stamina and things like that but other people in the year… when I was doing it, I had a six pack…” (boy) “…Yeah, it makes like unfit people going to be like fit because like they can go swimming and stuff..” (boy) Recommendations for improvement Provision of an electronic card “…Yeah the paper falls apart when its wet…”(boy) “…because then sometimes they say it’s like only £3 and you’ve got a fiver then you’ve lost £2. So if you go twice you’ve, say you’ve spent £6 and then you’ve lost £4 what you could have spent on another one, on another sports…” (boy) “…Reckon you should like make a little card or something, we don’t have to carry paper around with us everywhere…and say how much we’ve got left on our cards and then use it…” (girl) Vouchers redeemable for public transport “…I think if we could improve the vouchers they should have like bus passing and that…I think it might get more people active because it’s easier to get places if it’s free to get there, it’s free to do it…”(boy) Increased range of opportunities “If they had treadmills in school, I’d be happy…Actually treadmills at school, that would keep everyone active…I don’t know if you look like(name of sports centre)…all the equipment they’ve got, just bring some things into school like that and like they could impel the people like to be more active…”(boy) “....I was going to buy some weights but I don’t know where I could buy them…”(boy) “…I’d like to see like, I’ve been like looking online for like female boxing and stuff like that, they don’t really do that…” (girl) “…It’s only like one thing in (name of area) and that’s the(name of activity centre) (laughs) isn’t it basically…”(girl) “…Getting all of us just on a trip....Yeah, and go up to (next town) for the day, like spend your vouchers…”(girl) “…Everyone’s going to go paintballing, you’d have loads going for paintballing…” (boy) Use during school holidays “…I would like one of them vouchers just for the summer holidays because you know like when parents are buying the new like school bags and everything and it’s going to be expensive to get everything…and then you’ve got the vouchers you don’t have to ask for money…” (girl) “…When you get home you just want to relax, it’s like I’d use them more weekends probably…I’d use them more in the holidays probably…” (girl) Positive aspects of the voucher scheme (pupils) Removal of financial barrier Participants commented that vouchers enabled them to fund activities or buy equipment not normally accessible to them due to parental monetary constraints (Table 2). Pupils mentioned that prior to the scheme, if one or two adolescents could not afford the activity, the whole group did not go. Vouchers enabled participation by providing everyone the same opportunities regardless of economic background. It was felt that this sense of inclusiveness reduced the stigma linked to deprivation experienced by those unable to participate due to lack of financial resources. Changed attitudes towards PA Participants noted that their attitudes towards PA changed for the better, as for many the only experiences of activity had been through school P.E. They suggested seeing people from their year group in local activity centres reinforced the acceptability of participating in activities outside school; activity centres became a venue for socialising, and activity was seen as an opportunity to have fun with friends. Pupils remarked that the vouchers stimulated an interest in PA, which continued after the scheme ended, and believed this long term stimulation of interest would not have happened without the scheme. Increased choice/flexibility The large number of companies supporting the scheme provided flexibility so that one week a pupil could, for example, buy a skateboard, next week join a gym, and a month later, attend the waterpark. The option to spend vouchers on equipment, as well as activity, was praised as this gave financially-disadvantaged families the opportunity to buy longer-lasting, more expensive sporting equipment such as scooters, which were previously unaffordable. Additionally, participants liked that vouchers were not restricted to specific times or days, which would otherwise reduce accessibility. Increased awareness of local facilities Participants reported an increased awareness of local services and opportunities to try existing PA provisions in their area. Increased social interaction Some participants commented that the vouchers allowed them to spend more time doing activities with family and friends, which increased their enjoyment. In fact, several noted that those less inclined to take part in activity were now participating as they would go to be sociable and have fun with friends. One boy mentioned that he welcomed the opportunity to do something productive in the evenings, minimising his involvement in anti-social behaviour. Increased competence/body image The scheme also led to reported increases in competence and improved body image (particularly among boys), possibly due to improved self-efficacy. Recommendations for improvement of voucher scheme (pupils) Most participants were positive regarding the scheme and many were keen for the vouchers to return in the next academic year. However, there were recommendations for improvements (Table 2). Provision of an electronic card Some participants reported the paper vouchers ripping or disintegrating when wet, so the predominant suggestion was provision of an electronic card, also ensuring the exact activity cost could be deducted, instead of payments being made in increments of £5. Participants suggested they would be less likely to lose an electronic card. Vouchers redeemable for public transport It was suggested that vouchers should be redeemable for transport or that pupils should have access to free bus passes through the local council to access facilities outside of the community. Pupils advocated that this would address transport barriers, increasing accessibility and reducing dependence on parents/guardians. Increased range of opportunities Girls felt there were more activities on offer for boys and would like to see more choice of girls’ activities. The choice was restricted somewhat by facilities on offer within the area. Participants were encouraged to set up new activities, however, girls preferred to use pre-existing facilities. Participants wanted access to bigger sports shops, which were unable to participate due to company restrictions allowing participation in nationwide schemes only. Smaller shops were available, but tended to be further from participants’ neighbourhoods. More participants may have used their vouchers for equipment had bigger, more accessible shops been available. The ability to use vouchers outside of the immediate area was recommended in order to further increase choice and opportunities for new activities. Use during school holidays With the scheme ending in July, recurring views suggested the scheme would be greatly missed during school holidays, as they would have more time for activity engagement. Furthermore, parents would have less money in July for them to spend on activities due to spending large amounts on new equipment for the impending school year. Teacher’s opinions Increased opportunities Teachers spoke highly of the scheme stating it provided pupils with opportunities they may not otherwise have had and encouraged adolescents to spend their leisure time more positively (Table 3).Table 3 Qualitative data from teachers regarding positive aspects of the voucher scheme and recommendations for improvement Positive aspects of the voucher scheme Increased range of opportunities "…I think that our kids have benefitted from it hugely because they have very limited extra income to do anything with and are very much led by their parents a lot of the time and a lot of the parents have quite a negative attitude towards exercise anyway and it’s broadened their horizons because it’s allowed them to go and try things that maybe they wouldn’t have had access to before because of financial constraints…" (T2) "…obviously it’s a good scheme because it provides the children with greater opportunities to do things outside of school and anything to encourage them to use their time positively has got to be good…" (T1) Positive restriction to physical activity or sport-related equipment "…it’s good that the vouchers can only be spent on the sport because even if the family’s more affluent there’s nothing to say they’d pass money onto the children who would then use it to buy activities in some way, they would possibly spend it on PlayStation games or mobile phones and stuff…" (T1) Engaged non-sporty individuals "…it's engaged a lot more of the girls, a lot of the girls that were really disaffected with PE before because of the more traditional sports that we often offer, have bothered to come down, to see us to talk to us about the options, chasing vouchers etcetera and that’s led to us having more engaging conversations which in turn has allowed them to feel a bit more relaxed down here and not so under pressure to perform, perform, perform all the time and it’s just more about the taking part” (T2) "… bridges have been built with certain girls who just disliked coming to PE in the past and now those bridges have been built and relationships have improved I think yeah, well hopefully, as far I’m concerned… it’s allowed them, a lot of them to take part more with a smile on their face and enjoy it more than they have done in the past…" (T2) "…Oh (name of child) went out and bought footballs and basketballs and she’s not a sporty person, to say the least. So yeah, I’d say there are some who’ve done things I wouldn’t have thought they would…" (T1) Recommendations for improvement Increase accessibility “…In a more affluent area, you’ve got two parents probably, they’ve got a car each and it’ll just happen and also from the cultural side of it, people from those areas probably themselves are taken places by their parents to do whatever activity they did and they’ll see it as an expectation for them to take their children, which you may not get up here…” (T1) “…Or you make the activity local to the children so most of our lot come from (deprived area) so you put stuff on at the Centre or the youth club up there so they can get to it…or putting coaches in because I mean every community’s got a bit of open land… Yeah, community, yeah, smaller community things…” (T1) Positive restriction to PA or sport-related equipment The restriction of vouchers for PA was seen as an advantage as this prevented money being spent on other things, such as ‘Playstation’ games. Distributing vouchers directly to participants enabled them to be more involved as there was no need to negotiate money through parents, thereby removing a potential barrier. Engaged non-sporty individuals The Head of P.E. reported that the scheme enabled staff to build relationships with pupils who were usually disengaged with school-based PA. Pupils actively approached teaching staff to discuss the vouchers and where they could be used, leading to more pupils attending P.E. classes and willingly participating. Recommendations Teachers’ recommendations centred on increasing accessibility of existing activities, such as improved public transport to distant facilities or introduction of activities directly into communities, reducing the need for dependence on parents for transportation. Discussion This study demonstrated that the ACTIVE scheme is feasible and strongly supported by pupils and teachers. Interestingly, teachers commented on increased engagement in P.E. from those usually disinterested. This may be due to some children avoiding sport that is competitive [26]; the vouchers counteracted this by empowering participants to choose activities they wanted to participate in, leading to an increased sense of involvement. This was significant as attitudes towards PA influence participation in adolescents [27, 28]. Indeed, whilst removal of cost was a large incentive to participate in PA, qualitative analysis showed that the scheme appeared to act through changing pre-conceived attitudes towards PA. Therefore, PA became a fun opportunity to socialise with friends, especially for girls, and the scheme allowed relationships with family and friends to develop through co-participation in activity. This is consistent with previous research reporting participation with friends promoted increased activity in young people [29, 30] and motivation to be physically active [31]. In our study, spontaneous, unplanned activities were favoured, such as ‘Laserzone’ or skateboarding; consistent with the behaviour of children from lower socioeconomic status [32]. This supports previous research suggesting adolescents see structured, timetabled activities as a barrier to PA [15]. Analysis of voucher usage revealed that boys were more likely to use vouchers for activity, and increased PA levels would most likely be the main contributor to increased fitness. Girls preferred to purchase equipment, such as roller blades and scooters. However, focus groups indicated that some lacked the ability to rollerblade, so were unlikely to increase their PA enough to significantly improve fitness. Competence and skill level have previously been reported as barriers to PA among teenage girls [8, 15]. Thus, future work should explore how to improve fundamental movement skills in order to enable greater use of sporting equipment, and further increase PA. Qualitative responses demonstrated that improving activity levels is a complex, multi-faceted issue; the scheme addressed a number of issues but additional barriers to PA such as transport, time and variety remained. Recommendations included provision of an electronic card, expansion to neighbouring counties, and administration of vouchers in school holidays. Another major barrier to PA confirmed in previous work was transport to and from facilities [13]. In this study, both pupils and teachers recommended reducing public transport costs and increasing opportunities for free bus passes. Improving access to facilities through better transport would greatly enhance the success of future similar interventions. These recommendations provide useful insights into barriers experienced by adolescents, which are invaluable for ensuring their needs are met when developing future studies. Early observations presented favourable changes in PA, SB and fitness in both boys and girls. However, these changes appeared to be more pronounced at the weekend. This is consistent with the qualitative reports of increased time for activity and increased voucher usage during non-school days. This is promising given the literature reports less activity during weekends [33]. Conversely, weekday data demonstrated increased time spent in SB. No explanation for this was presented qualitatively but it could be a compensatory result of undertaking more activity during the weekend [34]. However, a recent review has reported existing research regarding the presence of compensation as inconclusive [35]. An additional point to note is that although the school itself was classified as deprived (54 % FSM eligibility), not all pupils who participated were explicitly classified as deprived; a common limitation for school-based projects. Given the small sample size, it was encouraging to see such positive differences in behaviour. However, the next step is to examine effect size and uptake in a larger range of schools, including comparators, to examine who benefits most from vouchers and whether the collective distribution of vouchers amongst adolescents plays a role in engagement. Cost-effectiveness, to explore the cost-benefit for changes in physical activity, and long-term sustainability also warrant further exploration. Additionally, this study only examined PA, SB and fitness. However, when children choose activities such as weight lifting or yoga, it is likely that different aspects of fitness (not measured here) would improve. Therefore, further research is required to examine which other outcomes are affected when teenagers do activities of their choice rather than traditional sports. Conclusions This study demonstrates that ACTIVE is a feasible approach to increasing PA and fitness amongst adolescents from a low socioeconomic background. This was achieved predominantly through changing attitudes towards PA and reducing cost barriers, allowing those from deprived backgrounds more opportunities to access activities. The feasibility of this approach on a larger scale, alongside a greater understanding of effectiveness, needs exploration. Abbreviations ACTIVE, active children through incentive vouchers—evaluation; BREQ-2, behavioural regulation in exercise questionnaire; COREQ, consolidated criteria for reporting qualitative research; CRT, cooper run test; FSM, free school meals; HBAI, households below average income data; LPA, light physical activity; MVPA, moderate-to-vigorous physical activity; P.E., physical education; PA, physical activity; PAQ-A, physical activity questionnaire for adolescents; RE-AIM framework, used to evaluate reach, effectiveness, adoption, implementation and maintenance of an intervention; SB, sedentary behaviour; UK, United Kingdom Additional file Additional file 1: Evidence of effect. The data within Additional file 1 includes more details regarding the quantitative outcomes from the ACTIVE study such as physical activity, sedentary behaviour, fitness and motivation to exercise. (PDF 239 kb) Acknowledgements All authors would like to thank the staff at the participating school for their co-operation during the study and also the Year 9 pupils for their views and opinions as well as participation. Dedication: This work was designed with Professor Non Thomas. Non died April 11th 2012 and is greatly missed by us all. Funding The work was funded by the British Medical Association in conjunction with support from The Centre for the Development and Evaluation of Complex Interventions for Public Health Improvement (DECIPHer), a UKCRC Public Health Research Centre of Excellence. Availability of data and materials The dataset supporting the conclusions of this article is available in The Secure Anonymised Information Linkage Databank (SAIL) repository [http://www.saildatabank.com]. Authors’ contributions The study was designed by DC and SB. Data collection was performed by DC, SB and RH. Data analysis was performed by DC, SB, CT and KM. GS, JR and SB provided supervision throughout. The first draft of the paper was written by DC and all authors provided critical input and revisions for all further drafts. 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 Ethical approval was granted by the Swansea University Research Ethics Committee. Parents and participants provided consent and assent, respectively, for specific research-focused outcome measures to be undertaken, such as self- and accelerometer-assessed PA. ==== Refs References 1. Fernandes RA Zanesco A Early physical activity promotes lower prevalence of chronic diseases in adulthood Hypertens Res 2010 33 9 926 931 10.1038/hr.2010.106 20574424 2. World Health Organisation Population based prevention strategies for childhood obesity: Report of the WHO forum and technical meeting, Geneva 15–17 December 2009 2010 3. Brodersen NH Steptoe A Boniface DR Wardle J Trends in physical activity and sedentary behaviour in adolescence:ethnic and socioeconomic differences Br J Sports Med 2007 41 3 140 144 10.1136/bjsm.2006.031138 17178773 4. Brooks F van der Sluijs W Klemera E Morgan A Magnusson J Nic Gabhainn S Roberts C Smith R Currie C Young people’s health in Great Britain and Ireland: findings for the health behaviour in School-Aged Children Survey 2006 2009 5. Brophy S Rees A Knox G Baker JS Thomas NE Child fitness and father’s BMI are important factors in childhood obesity: A school based cross-sectional study PLoS One 2012 7 5 e36597 10.1371/journal.pone.0036597 22693553 6. Houston E Baker J Buchan D Stratton G Fairclough S Foweather L Gobbi R Graves LF Hopkins N Boddy LM Cardiorespiratory fitness predicts clustered cardiometabolic risk in 10–11.9-year-olds Eur J Pediatr 2013 172 7 913 918 10.1007/s00431-013-1973-z 23440481 7. Boddy LM Murphy MH Cunningham C Breslin G Foweather L Gobbi R Graves LEF Hopkins ND Auth MKH Stratton G Physical activity, cardiorespiratory fitness, and clustered cardiometabolic risk in 10- to 12-year-old school children: The REACH Y6 study Am J Hum Biol 2014 26 4 446 451 10.1002/ajhb.22537 24599609 8. Slater A Tiggemann M Uncool to do sport: A focus group study of adolescent girls’ reasons for withdrawing from physical activity Psychol Sport Exercise 2010 11 6 619 626 10.1016/j.psychsport.2010.07.006 9. Brophy S Crowley A Mistry R Hill R Choudhury S Thomas NE Rapport F Recommendations to improve physical activity among teenagers-A qualitative study with ethnic minority and European teenagers BMC Public Health 2011 11 1 412 10.1186/1471-2458-11-412 21627781 10. Charlton R Gravenor MB Rees A Knox G Hill R Rahman MA Jones K Christian D Baker JS Stratton G Factors associated with low fitness in adolescents-A mixed methods study BMC Public Health 2014 14 1 764 10.1186/1471-2458-14-764 25074589 11. Department for Work and Pensions Households Below Average Income: An analysis of the income distribution 1994/95 – 2008/09 2010 London Department for Work and Pensions 12. Drenowatz C Steiner RP Brandstetter S Klenk J Wabitsch M Steinacker JM Organised sports, overweight, and physical fitness in primary school children in Germany J Obesity 2013 2013 7 10.1155/2013/935245 13. Zahner L Muehlbauer T Schmid M Meyer U Puder JJ Kriemler S Association of sports club participation with fitness and fatness in children Med Sci Sports Exerc 2009 41 2 344 350 10.1249/MSS.0b013e318186d843 19127191 14. Corder K Atkin AJ Ekelund U van Sluijs EMF What do adolescents want in order to become more active? BMC Public Health 2013 13 1 718 10.1186/1471-2458-13-718 23914878 15. Knowles A Niven A Fawkner S A qualitative examination of factors related to the decrease in physical activity behaviour in adolescent girls during the transition from primary to secondary school J Phys Activity Health 2011 8 8 1084 1091 10.1123/jpah.8.8.1084 16. Harland J White M Drinkwater C Chinn D Farr L Howel D The Newcastle exercise project: a randomised controlled trial of methods to promote physical activity in primary care BMJ 1999 319 7213 823 832 10.1136/bmj.319.7213.828 17. Lowther M, Mutrie N, Scott EM. Promoting physical activity in a socially and economically deprived community: a 12 month randomized control trial of fitness assessment and exercise consultation. J Sports Sci. 2002;20(7):577–88. 18. Moller AC, GeneMcFadden H, Hedeker D, Spring B. Financial motivation undermines maintenance in an intensive diet and activity intervention. Journal of Obesity. 2012;2012(740519):8. doi:10.1155/2012/740519 19. Cuffe HE Harbaugh WT Lindo JM Musto G Waddell GR Evidence on the efficacy of school-based incentives for healthy living Econ Educ Rev 2012 31 6 1028 1036 10.1016/j.econedurev.2012.07.001 20. Pope L Harvey-Berino J Burn and earn: a randomized controlled trial incentivizing exercise during fall semester for college first year students Prev Med 2013 56 3 197 201 10.1016/j.ypmed.2012.12.020 23295170 21. Shuttleworth I The Relationship between Social Deprivation as Measured by Individual Free School Meal Eligibility and Educational Attainment at GCSE in Northern Ireland: A Preliminary Investigation British Educational Res J 1995 21 4 487 504 10.1080/0141192950210404 22. Welsh Government Welsh Index of Multiple Deprivation Summary Report Cardiff: Statistical Directorate, Statistical Publication Unit 2011 23. Morgan M Gibbs S Maxwell K Britten N Hearing children’s voices: methodological issues in conducting focus groups with children aged 7–11 years Qualitative Res 2002 2 1 5 20 10.1177/1468794102002001636 24. Tong A Sainsbury P Craig J Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups International J Qual Health Care 2007 19 6 349 357 10.1093/intqhc/mzm042 25. Glasgow RE Vogt TM Boles SM Evaluating the Public Health Impact of Health Promotion Interventions: The RE-AIM Framework Am J Public Health 1999 89 9 1322 1327 10.2105/AJPH.89.9.1322 10474547 26. Trudeau F, Shephard RJ. Physical education, school physical activity, school sports and academic performance. International Journal of Behavioral Nutrition and Physical Activity. 2008;5:10. doi:10.1155/2012/740519 27. De la Haye K Robins G Mohr P Wilson C How physical activity shapes, and is shaped by adolescent friendships Soc Sci Med 2011 73 5 719 728 10.1016/j.socscimed.2011.06.023 21802807 28. Mulhall P Reis J Begum S Early adolescent participation in physical activity: correlates with individual and family characteristics J. Phys. Act. Health 2011 8 2 244 252 10.1123/jpah.8.2.244 21415452 29. Jago R Brockman R Fox KR Cartwright K Page AS Thompson JL Friendship groups and physical activity: qualitative findings on how physical activity is initiated and maintained among 10–11 year old children Int J Behav Nutr Phys Act 2009 6 1 4 10.1186/1479-5868-6-4 19138411 30. Jago R Macdonald-Wallis K Thompson JL Page AS Brockman R Fox KR Better with a Buddy: Influence of Best Friends on Children’s Physical Activity Med Sci Sports Exerc 2011 43 2 259 265 10.1249/MSS.0b013e3181edefaa 20581714 31. Salvy SJ, Roemmich JN, Bowker JC, Romero ND, Stadler PJ, Epstein LH. Effect of Peers and Friends on Youth Physical Activity and Motivation to be Physically Active. Journal of Pediatric Psychology. 2009;34(2):217-225. doi:10.1093/jpepsy/jsn071. 32. Brockman R Jago R Fox K Thompson J Cartwright K Page A “Get off the sofa and go and play”: Family and socioeconomic influences on the physical activity of 10–11 year old children BMC Public Health 2009 9 1 253 10.1186/1471-2458-9-253 19622143 33. Fairclough SJ Ridgers ND Welk G Correlates of children’s moderate and vigorous physical activity duing weekdays and weekends J. Phys. Act. Health 2012 9 129 137 10.1123/jpah.9.1.129 22232499 34. Rowland TW The biological basis of physical activity Med Sci Sports Exerc 1998 30 3 392 399 10.1097/00005768-199803000-00009 9526885 35. Gomersall SR Rowlands AV English C Maher C Olds TS The ActivityStat hypothesis: the concept, the evidence and the methodologies Sports Med 2013 43 2 135 149 10.1007/s40279-012-0008-7 23329607

PMC005xxxxxx/PMC5002125.txt

==== Front BMC GastroenterolBMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 52210.1186/s12876-016-0522-4Research ArticleEffect of mobile phone reminder messages on adherence of stent removal or exchange in patients with benign pancreaticobiliary diseases: a prospectively randomized, controlled study Gu Yong 13259800503@126.com 12Wang Limei 13310983711@163.com 3Zhao Lina zhaolina@fmmu.edu.cn 4Liu Zhiguo liuzhiguo@fmmu.edu.cn 3Luo Hui fmmulh@163.com 3Tao Qin 505456739@qq.com 3Zhang Rongchun Zrc.700502@163.com 3He Shuixiang 83169397@qq.com 1Wang Xiangping windxp2013@163.com 3Huang Rui hedakangxiaoyu@163.com 3Zhang Linhui kangxiaoyu@hotmail.com 3Pan Yanglin panyanglin@gmail.com 3Guo Xuegang xuegangguo@gmail.com 31 Department of Gastroenterology, the first affiliated hospital of Xi’an Jiao Tong university, Xi’an, China 2 Digestive System Department, Shaanxi Provincial Crops Hospital of Chinese People’s Armed Police Force, Xi’an, China 3 Xijing Hospital of Digestive Diseases, Fourth Military Medical University, Xi’an, Shannxi China 4 Department of Radiotherapy, Xijing Hospital, Xian, China 26 8 2016 26 8 2016 2016 16 1 10518 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 Plastic and covered metal stents need to be removed or exchanged within appropriate time in case of undesirable complications. However, it is not uncommon that patients do not follow the recommendation for further stent management after Endoscopic Retrograde Cholangiopancreatography (ERCP). The effect of short message service (SMS) intervention monthly on the stent removal/exchange adherence in patients after ERCP is unknown at this time. Methods A prospective, randomized controlled study was conducted. After receiving regular instructions, patients were randomly assigned to receive SMS reminding monthly (SMS group) for stent removal/exchange or not (control group). The primary outcome was stent removal/exchange adherence within appropriate time (4 months for plastic stent or 7 months for covered stent). Multivariate analysis was performed to assess factors associated with stent removal/exchange adherence within appropriate time. Intention-to-treat analysis was used. Results A total of 48 patients were randomized, 23 to the SMS group and 25 to the control. Adherence to stent removal/exchange was reported in 78.2 % (18/23) of patients receiving the SMS intervention compared with 40 % (10/25) in the control group (RR 1.98, 95 % CI 1.16–3.31; p = 0 · 010). Among patients with plastic stent insertion, the median interval time from stent implantation to stent removal/exchange were 90 days in the SMS group and 136 days in the control respectively (HR 0.36, 95 % CI 0.16–0.84, p = 0.018). No difference was found between the two groups regarding late-stage stent-related complications. The rate of recurrent abdominal pain tended to be lower in SMS group without significant difference (8.7 vs 28 %, p = 0.144). Multivariate logistic regression analyses revealed that SMS reminding was the only factor associated with adherence of stent removal/exchange (OR 6.73, 95 % CI 1.64–27.54, p = 0.008). Conclusion This first effectiveness trial demonstrated that SMS reminding monthly could significantly increase the patient adherence to stent removal/exchange after ERCP. Trial registration The study was respectively registered on July 10 in 2016 at ClinicalTrials.gov (NCT02831127). Keywords Short message serviceAdherenceStent exchangeERCPBiliary stricturehttp://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China8117228881372388Pan Yanglin issue-copyright-statement© The Author(s) 2016 ==== Body Background Endoscopic implantation of plastic or covered metal stents is widely used in a variety of benign pancreaticobiliary diseases, including duct stricture, large or difficult stones, bile or pancreatic duct leak, etc. [1–4]. There are some complications after stent insertion, such as stent occlusion, proximal or distal migration, secondary duct injury and even the failure of stent removal [5–8]. For plastic stents, occlusion is the main disadvantage, limiting their patency to around 3 months. For fully covered metal stents, stent migration, occlusion and even the failure of stent removal may happen after long-term implantation [8, 9]. The longer the stents areplaced, more likely the complications may happen. Although the optimal time of stent placement has not been well established, it has been recommended that plastic stent should be removed/exchanged within 3–4 months and covered metal stent be removed within 6 months [10]. However, it is not uncommon that patients with stent implantation do not follow the recommendation of further stent management [11]. With the stents left in biliary or pancreatic duct for a long-term period, stone formation, acute duct inflammation and even chronic pancreatitis and secondary sclerosing cholangitis can happen. Occasionally, breakage of the stent can be also found [12]. Some patients in this situation may need emergent endoscopic management or even surgery. In addition, endoscopic management may thus be technically challenging, and the treatment cost can be increased. Many methods have been used to improve the adherence of patients in medical service [13–15]. With the advance of mobile technology and popular use of mobile phones, it is believed that the patient-centered outcome (e.g. suppressed viral loads due to antivirus treatment) can be improved by mobile telecommunication with the timely support of a patient by a health professional [13]. Here we hypothesize that mobile technology, reminding the patients the necessity of stent management in time by short message service (SMS), may increase the patient adherence. The purpose of this prospectively randomized, controlled study is to evaluate the effect of SMS intervention monthly on the stent removal/exchange adherence in patients with benign pancreaticobiliary diseases after ERCP. Methods Patients This is a prospective, randomized, controlled study with consecutive patients with benign pancreaticobiliary diseases undergoing endoscopic stent insertion at Endoscopy Center of Xijing Hospital of Digestive Diseases in China. The study protocol and informed consent form were approved by Institutional Review Board of Xijing Hospital (protocol number: 20160707–1). The study was respectively registered on July 10 in 2016 at ClinicalTrials.gov (NCT02831127). The informed consent was obtained from all patients. Patients more than 18 years old with plastic or covered stent implantation for the drainage of bile or pancreatic juice were eligible for participation in the study. Patients should be able to communicate via SMS by mobile phones of themselves or relatives living together. Exclusion criteria included: 1. primary or secondary sclerosing cholangitis (PSC), 2. malignant or suspected malignant stricture of biliary or pancreatic duct, 3.implantation of pancreatic duct (PD) stent for prevention of post-ERCP pancreatitis, 4.expected survival time less than 6 months, 5. plan of surgery within 6 months, 6. pregnant or lactating women, 7. patients who could not give informed consent. Written informed consent was obtained from all the patients. Patients were randomized (1:1) to either the SMS intervention (SMS group) or standard care (control group) after stent insertion by opening an opaque and sealed envelope. The envelopes were randomized by using computer-generated random numbers generated by one of the investigators (HR) who kept the randomization key under lock until the inclusion of the last patient. At least two telephone numbers of all patients or their relatives living together were recorded in case of failed connection later. In the beginning of the enrollment, all patients were instructed not to tell doctors, nurses and investigators whether they received SMS reminding or not. The investigator (ZLN) performing data analysis was blinded to the allocation until the final analysis was finished. Endoscopic treatment The diagnosis of all the patients was primarily based on symptoms, surgical history, chemical test and imaging modalities (contrast-enhanced CT or ultrasound). All patients underwent MRCP for determination of etiology and the site of stricture. Only the patients with benign stricture of CBD or PD were considered eligible for this study. During ERCP, tissue samples were obtained with brush and/or forceps to confirm the benign nature of the stricture when clinically indicated. Single or multiple plastic stents (8.5Fr, Advanix, Boston Scientific, Natick, MA) or a fully covered self-expandable metal stent (FCSEMS) (Wallflex, Boston Scientific, Natick, MA) was inserted across the site of obstruction. The length of the stent varied depending on the anatomic location of the stricture. No covered metal stent was placed in PD. The number and type of the stents was determined based on the characteristics of stricture or diseases, which was determined at the discretion of the attending endoscopists. Intervention After stent implantation, all patients received oral and written instructions about further management. If single or multiple plastic stents were inserted, patients were informed to come back to the hospital at 3 months for stent removal/exchange; if FCSEMS was inserted, they were informed to come back to the hospital at 6 months after ERCP. Patients in SMS group received additional reminding by SMS messages from an investigator (TQ) blinded to further clinical data collection. Each month after stent implantation, the investigator sent a text message by SMS to inform patients the necessity of regular stent removal/exchange and the disadvantage of delayed management, and to remind them the appropriate date to come back to the hospital for stent management. Patients were requested to respond by SMS and were encouraged to contact the investigator if they had any questions about stent management. Patients in control group were not contacted after ERCP. At the end of the study, all the patients who did not come back to the hospital were called and informed again to return for further stent management. Follow-up was at least 6 months for all patients. Outcome measurement The primary outcome was stent removal/exchange adherence within appropriate time (4 months for plastic stent or 7 months for covered stent). Secondary outcomes were stent-related complications, including cholangitis, stent migration and abdominal pain. Statistical analysis At the beginning of the study, a sample size calculation was performed. Based on our previous experience, only 1/3 of patients in common practice will readmit for stent removal/exchange within appropriate time. The adherence in SMS group was estimated to be 80 %. To detect the difference with a significance level (α) of .05 and a power of 80 % with a 2-tailed test, we calculated that at least 42 patients were needed. However, about 10 % of patients might be lost during follow up. Thus, we estimated that totally 48 patients would be enough for the detection of a significant difference in the primary outcome. Intention-to-treat (ITT) analysis was used to assess primary outcome from all evaluable patients. Relative risk (RR) was reported for adherence, with an RR more than 1 suggesting better outcome for SMS intervention group. Since only a small group of patients would be included, categorical variables, such as adherence rate of stent exchange/removal and complication rates, were analyzed using Fisher’s exact test. Continuous variables were expressed as means with standard deviations and analyzed with student’s t-test. Cumulative proportion of patients readmitting with plastic stent implanted during follow up was determined by the Kaplan-Meier method, and the difference was assessed using the log-rank test. To assess factors associated with stent removal/exchange adherence, multivariate logistic analysis was performed using variables with p values of <0.1 in the univariate logistic analysis. Forward stepwise method was used in the multivariate model. Analyses were performed with SPSS software version 19.0 for Windows (SPSS Inc, IBM Company). A p value <0.05 was considered statistically significant. Results From Feb in 2012 to Oct in 2013, 193 consecutive patients were enrolled. After screening, 145 patients were excluded, including 22 with inadequate phone access, 78 with definite or suspected malignancy, 3 with PSC, 25 with prophylactic PD stent implanted and 17 with declined participation. Finally, 48 patients were randomly assigned to the SMS group (n = 23) or to control group (n = 25). After randomization, all the patients in SMS group responded by SMS or phone call. However, 1 subject with distal stricture of CBD in SMS group underwent unplanned surgery because of pancreatic cancer. The subject flow is detailed in Fig. 1. All baseline characteristics but alkaline phosphatase (336.5 ± 324.2 U/L in SMS group vs. 125.8 ± 76.2 U/L in control, p = 0.003) between the two groups were well balanced (Table 1).Fig. 1 Flowchart of the study Table 1 Baseline of the characteristics of patients SMS group (n = 23) Control (n = 25) P value Age 54.4 ± 15.0 52.2 ± 19.5 0.672 Male (%) 14 (60.9 %) 11 (44 %) 0.265 Smoking 7 3 0.162 Drinking 4 6 0.727 Education 1.000  Elementary school or less 5 6  High school or higher 18 19 Payment 0.610  By insurance 22 22  By self 1 3 Previous Stenting 0.511  Yes 7 5  No 16 20 Previous surgery 0.818  Cholecystectomy 8 7  Liver transplantation 0 1  Other 3 2 Main symptom 0.467  Jaundice 7 4  Fever 1 2  Abdominal pain 13 16 Chemical test before ERCP  White blood cell (×109) 5.8 ± 1.3 5.6 ± 2.8 0.714  Total bilirubin (mg/dL) 60.6 ± 93.3 30.0 ± 65.9 0.193  Alkaline phosphatase (U/L) 336.5 ± 324.2 125.8 ± 76.2 0.003 Stricture site 0.849  Proximal CBD 6 7  Distal CBD 11 10  PD 6 8 Reason for stenting 0.501  Biliary benign stricture 10 12  Pancreatic benign stricture 10 8  Other 3 5 Stent type 0.719  Plastic stent (average number of stents) 18 (1.39) 21 (1.43)  FCSEMS 5 4 ERCP complication 1.000  Pancreatitis 1 1  Biliary infection 2 1 CBD common bile duct, PD pancreatic duct, FCSEMS fully covered self-expanded metal stent In ITT analysis, adherence to stent removal/exchange was reported in 78.2 % (18/23) of patients receiving the SMS intervention compared with 40 % (10/25) in the control group (relative risk [RR] 1.98, 95 % CI 1.16–3.31; p = 0 · 010) (Table 2). Among patients undergoing insertion of plastic stent (n = 39), adherence to stent removal/exchange was 77.8 % in SMS group and 33.3 % in control (p = 0.010). The cumulative proportions of patients coming back to the hospital during follow up are shown in Fig. 2. The mean interval time between stent implantation and stent removal/exchange was 90 days in SMS group and 136 days in the control group respectively (hazard ratio [HR] 0.36, 95 % CI 0.16–0.84, p = 0.018).As shown in Table 2, no difference was found regarding FCSEMS removal adherence between the two groups (80 vs. 75 %, p = 1.000). There were also no differences between the two groups with regard to stent-related complications, such as cholangitis (9 vs 8 %, p = 1.000), stent migration (13 vs. 8 %, p = 0.653) and recurrent abdominal pain (9 vs. 28 %, p = 0.144). However, the rate of recurrent abdominal pain tended to be lower in SMS group (8.7 vs 28 %, p = 0.144).Table 2 Outcomes of SMS reminding compared with standard care SMS group (n = 23) Control (n = 25) P value Stent removal/exchange adherence, n (%) 18/23 (78 %) 10/25 (40 %) 0.010  Plastic stent (<4 month) 14/18 (78 %) 7/21 (33 %) 0.010  FCSEMS (<7 month) 4/5 (80 %) 3/4 (75 %) 1.000 Stent-related complications, n (%)  Cholangitis 2 (9 %) 2 (8 %) 1.000  Stent migration 3 (13 %) 2 (8 %) 0.653  Recurrent pain 2 (9 %) 7 (28 %) 0.144 FCSEMS, fully covered self-expanded metal stent Fig. 2 Kaplan-Meier survival analysis of proportions of patients with plastic stent implanted undergoing stent removal/exchange later in SMS group (n = 17) and control (n = 21). p = 0.018 by log-rank test Multivariate logistic regression analyses were performed to identify any significant factors for stent removal or exchange adherence. The factors analyzed were age, gender, history of surgery, education level, pre-ERCP total bilirubin level, location of stenosis, stent type, stent number, reasons for stenting, post-ERCP complications and SMS reminding or not. As shown in Table 3, only SMS reminding were significantly associated with adherence of stent removal/exchange (OR 6.73, 95 % CI 1.64–27.54, p = 0.008).Table 3 Multivariate logistic regression analysis of the association between patient characteristics and stent removal/exchange adherence Variable Adherence p value OR 95 % CI SMS reminding No 1 Yes 6.73 1.64–27.54 0.008 Surgery history No 1 Yes 3.20 0.74–13.80 0.119 Stent type Plastic 1 Metal 2.34 0.33–16.71 0.398 Stent number Single 1 Multiple 2.10 0.51–8.73 0.306 Among patients coming back to the hospital finally (19 in SMS group vs 18 in the control, p = 0.297), 11 in the SMS group and 9 in the control group underwent plastic stent exchange (p = 0.746). The remaining patients in both groups needed no further management after stent removal and clearance of biliary or pancreatic duct. Discussion Plastic stents and covered metal stent are commonly used for the drainage and relief of benign stricture of biliary and pancreatic ducts [1–4]. It is suggested that these stents should be removed or exchanged within 3–6 months to prevent late complications [7, 8]. Although patients are usually instructed the details of further stent management, some of them may be not compliant with the recommendation. The reasons may include: 1, the unawareness of the necessity of regular stent removal/exchange; 2, the unawareness of the possible complications of delayed stent management; 3, forgetting the appropriate date to come back to the hospital for stent management; 4, financial consideration. Here we found that SMS reminding monthly could significantly increase the patient adherence to stent removal/exchange. This is, to our knowledge, the first effectiveness trial assessing the ability of a mobile health technology intervention to influence the stent removal/exchange adherence. Patients’ forgetfulness is considered one of the main reasons for missed appointments. There are many modes of communicating reminders for appointments to patients, such as face-to-face communication, postal messages, phone calls and SMS [16]. The later represent a convenient, less time-consuming and inexpensive delivery medium for improving the adherence of healthcare appointments. Studies that compare the outcomes of SMS reminding versus other methods for the patients with removable stents is of interest. With the better adherence to stent removal or exchange, it could be expected that the stent-related complications due to long-term placement of plastic or cover metal stents might be reduced [7, 8, 17, 18]. However, the late-stage complications between the two groups in this study were not significantly different, although the rate of recurrent abdominal pain tended to be lower after SMS reminding. The reason may be due to small numbers of patients enrolled in each subgroup. The power of the study may be insufficient to detect the differences of stent-related complications and identify more predictive factors related to stent removal/exchange adherence. There are some other limitations of this study. Firstly, the follow up time of this study is relatively short. It has been recommended that multiple plastic stents should be placed and exchanged for at least one year for long-term stricture of biliary stricture [19]. With better adherence to plastic stent exchange, it will be interesting to further evaluate the long-term resolution rate of biliary stricture after SMS reminding. Secondly, although patients with plastic stents in SMS group had better adherence to stent removal or exchange, no difference was found regarding the adherence to covered metal stent management. It is necessary to enroll more patients with covered metal stent to investigate whether they will be also benefit from SMS reminding. Thirdly, although number of patients undergoing placement of covered metal stent was similar between the two group, whether patients received metal stent were determined at the discretion of the attending endoscopists. The possible bias of patient selection may have impacts on the adherence in metal group. Last but not the least, the present study was performed in one tertiary center in a less developed area in China. The adherence rate without interference seems to be quite low (40 %). The beneficial effect of SMS on adherence of stent removal/exchange needs to be further investigated in other settings, especially in centers with higher adherence of stent removal/exchange. Conclusions In conclusion, our study demonstrated that SMS reminding could improve the patient adherence to stent removal/exchange within appropriate time for the first time. SMS reminding could shorten the mean interval time between stent implantation and stent removal/exchange. Patients with stent implantation might be benefit from SMS reminding strategy. Abbreviation CIConfident interval CTComputed tomography ERCPEndoscopic retrograde cholangiopancreatography FCSEMSFully covered self-expandable metal stent. HRhazard ratio ITTIntention-to-treat MRCPMagnetic resonance cholangiopancreatography OROdd ratio PDPancreatic duct PSCPrimary sclerosing cholangitis RRRelative risk SMSShort message service Acknowledgement We are grateful to the doctors and nurses working in Xijing Hospital of digestive diseases for the help in conducting of this study. Funding This work was supported in part by the National Natural Science Foundation of China to Yanglin Pan (No. 81172288 and 81372388). Availability of data materials All the data supporting the conclusions of this study had been presented in this paper. The raw individual datasets will not be shared in order to protect patient confidentiality. Authors’ contributions Study concept and design: PYL, GY; acquisition of data: WLM,; analysis and interpretation of data: ZLN, LH,; drafting of the manuscript: GY PYL, HSX; critical revision of the manuscript for important intellectual content: ZRC, GXG, LZG; statistical analysis: HR, WXP, ZLH; administrative and material support: GXG. 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 The study protocol and informed consent form were approved by Institutional Review Board of Xijing Hospital (protocol number: 20160707–1). The informed consent was obtained from all patients. ==== Refs References 1. Tuvignon N Liguory C Ponchon T Long-term follow-up after biliary stent placement for postcholecystectomy bile duct strictures: a multicenter study Endoscopy 2011 43 208 16 10.1055/s-0030-1256106 21365514 2. Tabibian JH Asham EH Han S Endoscopic treatment of postorthotopic liver transplantation anastomotic biliary strictures with maximal stent therapy Gastrointest Endosc 2010 71 505 12 10.1016/j.gie.2009.10.023 20189508 3. Mahajan A Ho H Sauer B Temporary placement of fully covered self-expandable metal stents in benign biliary strictures: midterm evaluation Gastrointest Endosc 2009 70 303 9 10.1016/j.gie.2008.11.029 19523620 4. Jong EA Moelker A Leertouwer T Percutaneous transhepatic biliary drainage in patients with postsurgical bile leakage and nondilated intrahepatic bile ducts Dig Surg 2013 30 444 50 10.1159/000356711 24434644 5. ASGE Technology Assessment Committee Pfau PR Pleskow DK Pancreatic and biliary stents Gastrointest Endosc 2013 77 319 27 10.1016/j.gie.2012.09.026 23410693 6. Khashab MA Kim K Hutfless S Predictors of early stent occlusion among plastic biliary stents Dig Dis Sci 2012 57 2446 50 10.1007/s10620-012-2178-4 22573343 7. Slattery E Kale V Anwar W Role of long-term biliary stenting in choledocholithiasis Dig Endosc 2013 25 440 3 10.1111/j.1443-1661.2012.01399.x 23808949 8. Balmadrid B Kozarek R Prevention and management of adverse events of endoscopic retrograde cholangiopancreatography Gastrointest Endosc Clin N Am 2013 23 385 403 10.1016/j.giec.2012.12.007 23540966 9. Baron TH Covered self-expandable metal stents for benign biliary tract diseases Curr Opin Gastroenterol 2011 27 262 7 10.1097/MOG.0b013e3283438a26 21248636 10. Kasher JA Corasanti JG Tarnasky PR Amulticenter analysis of safety and outcome of removal of a fully coveredself-expandable metal stent during ERCP Gastrointest Endosc 2011 73 1292 7 10.1016/j.gie.2011.01.043 21470603 11. Lawrence C Romagnuolo J Payne KM Low symptomatic premature stent occlusion of multiple plastic stents for benign biliary strictures: comparing standard and prolonged stent change intervals Gastrointest Endosc 2010 72 558 63 10.1016/j.gie.2010.05.029 20638060 12. Haapamäki C Kylänpää L Udd M Randomized multicenter study of multiple plastic stents vs covered self-expandable metallic stent in the treatment of biliary stricture in chronic pancreatitis Endoscopy 2015 47 605 10 10.1055/s-0034-1391331 25590182 13. Lester RT Ritvo P Mills EJ Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): a randomised trial Lancet 2010 376 1838 45 10.1016/S0140-6736(10)61997-6 21071074 14. Margolis KL Asche SE Bergdall AR Effect of home blood pressure telemonitoring and pharmacist management on blood pressure control: a cluster randomized clinical trial JAMA 2013 310 46 56 10.1001/jama.2013.6549 23821088 15. Liu X Luo H Zhang L Telephone-based re-education on the day before colonoscopy improves the quality of bowel preparation and the polyp detection rate: a prospective, colonoscopist-blinded, randomised, controlled study Gut 2014 63 125 30 10.1136/gutjnl-2012-304292 23503044 16. Gurol-Urganci I de Jongh T Vodopivec-Jamsek V Mobile phone messaging reminders for attendance at healthcare appointments Cochrane Database Syst Rev 2013 12 CD007458 24310741 17. Romagnuolo J Cotton PB Recording ERCP fluoroscopy metrics using a multinational quality network: establishing benchmarks and examining time-related improvements Am J Gastroenterol 2013 108 1224 30 10.1038/ajg.2012.388 23912400 18. Kachaamy TA Faigel DO Improving ERCP quality and decreasing risk to patients and providers Expert Rev Gastroenterol Hepatol 2013 7 531 40 10.1586/17474124.2013.824703 23985002 19. Coté GA Slivka A Tarnasky P Effect of covered metallic stents compared with plastic stents on benign biliary stricture resolution: a randomized clinical trial JAMA 2016 315 1250 7 10.1001/jama.2016.2619 27002446

PMC005xxxxxx/PMC5002126.txt

==== Front BMC Health Serv ResBMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 168110.1186/s12913-016-1681-2ErratumErratum to: Implications of early and guideline adherent physical therapy for low back pain on utilization and costs Childs John D. childsjd@gmail.com 1Fritz Julie M. 2Wu Samuel S. 3Flynn Timothy W. 4Wainner Robert S. 4Robertson Eric K. 5Kim Forest S. 6George Steven Z. 71 Army Medical Department Center and School, US Army-Baylor University Doctoral Program in Physical Therapy, 3151 Scott Rd., Rm. 2307, JBSA Fort Sam Houston, San Antonio, TX 78234 USA 2 Department of Physical Therapy, University of Utah, 520 Wakara Way, Salt Lake City, UT 84108 USA 3 Department of Health Services Research, Management and Policy, College of Public Health and Health Professions, University of Florida, 1329 SW 16th St., Rm. 5231, Gainesville, FL 32610-0177 USA 4 EIM School of Physical Therapy, South College, 3904 Lonas Dr, Knoxville, TN 37909 USA 5 Doctor of Physical Therapy Program, University of Texas at El Paso, 500 W. University Avenue, El Paso, TX 79968 USA 6 US Army Medical Department Center and School, US Army-Baylor MHA/MBA Program, 3599 Winfield Scott Rd., Bldg. 2841, JBSA Fort Sam Houston, San Antonio, TX 78234-6135 USA 7 Department of Physical Therapy, Director, Brooks-PHHP Research Collaboration, University of Florida, P.O. Box 100154, Gainesville, FL 32610-0154 USA 26 8 2016 26 8 2016 2016 16 1 444© 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 There is a typographical error on p. 7 of the article [1], where it states, “…24.0 % (n = 17,175) were categorized as receiving early physical therapy that was also adherent to the recommendation for active treatment, 19.2 % (n = 13,742) received delayed physical therapy that was adherent, 23,993 (33.5 %) received delayed and adherent care, and 16,649 (23.3 %) received physical therapy that was delayed and non-adherent.” As can be seen, reference to the “delayed and adherent” category is mentioned twice. Based on Table 4, the second reference to “delayed and adherent” care should have read “early and non-adherent” care. The corrected text is as follows:“…24.0 % (n = 17,175) were categorized as receiving early physical therapy that was also adherent to the recommendation for active treatment, 19.2 % (n = 13,742) received delayed physical therapy that was adherent, 23,993 (33.5 %) received early and non-adherent care, and 16,649 (23.3 %) received physical therapy that was delayed and non-adherent.” This typographical correction makes the text in the manuscript now consistent with the data in Table 4, which is correct as originally published. The online version of the original article can be found under doi:10.1186/s12913-015-0830-3. ==== Refs Reference 1. Childs et al. Implications of early and guideline adherent physical therapy for low back pain on utilization and costs. BMC Health Serv Res. 2015; 15:150. doi: 10.1186/s12913-015-0830-3

PMC005xxxxxx/PMC5002127.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 127210.1186/s12906-016-1272-0Research ArticleAn investigation of the use of acupuncture in stroke patients in Taiwan: a national cohort study Weng Shu-Wen aaaa6712@gmail.com 12Chen Ta-Liang tlc@tmu.edu.tw 345Yeh Chun-Chieh b8202034@gmail.com 67Liao Chien-Chang jacky48863027@yahoo.com.tw 1345Lane Hsin-Long imcold0404@hotmail.com 8Lin Jaung-Geng jglin@mail.cmu.edu.tw 19Shih Chun-Chuan 886-7-615-1100hwathai@seed.net.tw 8101 Graduate Institute of Chinese Medicine, College of Chinese Medicine, China Medical University, Taichung, 404 Taiwan 2 Department of Chinese Medicine, Taichung Hospital, Ministry of Health and Welfare, Taichung 403, Taiwan 3 Department of Anesthesiology, Taipei Medical University Hospital, Taipei, 110 Taiwan 4 Department of Anesthesiology, School of Medicine, College of Medicine, Taipei Medical University, Taipei 110, Taiwan 5 Health Policy Research Center, Taipei Medical University Hospital, Taipei, 110 Taiwan 6 Department of Surgery, China Medical University Hospital, Taichung, 110 Taiwan 7 Department of Surgery, University of Illinois, Chicago, Illinois USA 8 School of Chinese Medicine for Post-Baccalaureate, College of Medicine, I-Shou University, Kaohsiung City, 824 Taiwan 9 Department of Healthcare Administration, Asia University, Taichung, 413 Taiwan 10 Ph.D. Program for the Clinical Drug Discovery from Botanical Herbs, College of Pharmacy, Taipei Medical University, Taipei 110, Taiwan 26 8 2016 26 8 2016 2016 16 1 32123 1 2016 9 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 Acupuncture is considered a complementary and alternative medicine in many countries. The purpose of this study was to report the pattern of acupuncture use and associated factors in patients with stroke. Methods We used claims data from Taiwan’s National Health Insurance Research Database and identified 285001 new-onset stroke patients in 2000–2008 from 23 million people allover Taiwan. The use of acupuncture treatment after stroke within one year was identified. We compared sociodemographics, coexisting medical conditions, and stroke characteristics between stroke patients who did and did not receive acupuncture treatment. Results The use of acupuncture in stroke patients increased from 2000 to 2008. Female gender, younger age, white-collar employee status, higher income, and residence in areas with more traditional Chinese medicine (TCM) physicians were factors associated with acupuncture use in stroke patients. Ischemic stroke (odds ratio [OR] 1.21, 95 % confidence interval [CI] 1.15–1.28), having no renal dialysis (OR 2.76, 95 % CI 2.45–3.13), receiving rehabilitation (OR 3.20, 95 % CI 3.13–3.27) and longer hospitalization (OR 1.23, 95 % CI 1.19–1.27) were also associated with acupuncture use. Stroke patients using rehabilitation services were more likely to have more acupuncture visits and a higher expenditure on acupuncture compared with stroke patients who did not receive rehabilitation services. Conclusions The application of acupuncture in stroke patients is well accepted and increasing in Taiwan. The use of acupuncture in stroke patients is associated with sociodemographic factors and clinical characteristics. Keywords AcupunctureComplementary and alternative medicineStrokeTraditional Chinese medicineUsehttp://dx.doi.org/10.13039/501100004663Ministry of Science and Technology, TaiwanMOST103-2320-B-214-010-MY2MOST104-2314-B-038-027-MY2Liao Chien-Chang Shih Chun-Chuan Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of ExcellenceMOHW105-TDU-B-212-133019issue-copyright-statement© The Author(s) 2016 ==== Body Background With the increasing use of complementary and alternative medicine (CAM) worldwide [1], the 1-year prevalence of CAM use in the United States and United Kingdom were found to be as high as 33.2 % and 26.3 %, respectively [2, 3]. The estimated out-of-pocket cost for CAM was $33.9 billion in the USA in 2007 [4]. Acupuncture is considered a subtype of traditional Chinese medicine (TCM) [5], which has been used for at least 2000 years in China, and it has gained attention in the United States since 1971. Currently, it is widely used in many countries [1, 5]. A cross-sectional survey showed that at least 1.5 % of adults had used acupuncture in the past 12 months in the United States in 2007 [6]. Stroke remains the leading cause of adult disabilities worldwide, with an estimated direct medical cost of $20.6 billion in the United States in 2010 [7]. The 1-year cost of stroke ranges from $7,342 to $146,149 per patient in several countries [8]. Pneumonia, urinary tract infection, pain, dysphagia, depression, and recurrent stroke are common complications after stroke [9–11]. Acupuncture has been accepted in stroke rehabilitation in many countries, and the treatment is relatively safe and effective in improving post-stroke chronic symptoms, such as disability, shoulder pain, and dysphagia [5, 12–15]. In the United States, acupuncture was used more frequently in stroke compared with non-stroke patients [16]. A previous study also showed the high use of TCM among stroke patients in Taiwan [17]. However, limited information is available on the pattern of use of acupuncture in stroke patients. Using the National Health Insurance Research Database, we conducted a nationwide, population-based cohort study to evaluate the pattern of acupuncture use in stroke patients. Another purpose of this study was to report the factors associated with acupuncture use among adult stroke patients. Methods Source of data Since 1996, all medical claims of insured beneficiaries have been documented in the National Health Insurance Research Database, which was established by Taiwan’s National Health Research Institute. Information available for this study included gender, birth date, disease codes, health care rendered, medicines prescribed, diagnoses at admission and discharge, and medical institutions and physicians providing services. This study employed the All Stroke Database, which consisted of all prevalent and incident stroke patients across Taiwan between 2000 and 2008 [17–19]. Ethical statement Insurance reimbursement claims used in this study were obtained from Taiwan’s National Health Insurance Research Database, which is available for academic access. This study was conducted in accordance with the Helsinki Declaration. To protect personal privacy, the electronic database was decoded with patient identifications scrambled for further public access for research. Although the National Health Research Institute regulations do not require informed consent due to the use of decoded and scrambled patient identification, this study was approved by Taiwan’s National Health Research Institute (NHIRD-100-122) and the Institutional Review Board of E-DA Hospital, Kaohsiung, Taiwan (2014012) [17–19]. Study design and population We identified 285001 newly diagnosed, hospitalized stroke patients aged ≥20 years between 2000 and 2008 as our eligible study subjects from 23 million people allover Taiwan. Those with a previous stroke according to a physician’s diagnosis were excluded until 1996. To confirm that all stroke patients in our study were incident cases, only new-onset stroke cases were included. The outcome of this study was the prevalence of acupuncture use in people with a new diagnosis of stroke in the first year. This study compared sociodemographic factors, coexisting medical conditions, and stroke characteristics between stroke patients who used and did not use acupuncture. Criteria and definition We defined stroke according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM 430–438). Coexisting medical conditions included diabetes mellitus (ICD-9-CM 250), hypertension (ICD-9-CM 401–405), hyperlipidemia (ICD-9-CM 272.0–272.4), and myocardial infarction (ICD-9-CM 410 and 412). According to the administration codes (D8, D9) from reimbursement claims, regular renal dialysis (including hemodialysis and/or peritoneal dialysis) was also considered a coexisting medical condition among stroke patients in this study. We classified the frequency of acupuncture visits into quartiles. Stroke patients in the highest quartile of acupuncture visits were defined as high acupuncture users. Medical expenditures on acupuncture were also classified into quartiles. Stroke patients who were in the highest quartile of acupuncture expenditure were considered high acupuncture expenditure patients. As Taiwan has 359 townships and city districts, we calculated the population density (persons/km2) of each of these administrative units. Based on the population density, these units were stratified into tertiles to designate areas of low, moderate, and high urbanization. We calculated the density of traditional Chinese physicians (traditional Chinese physicians/10,000 persons) based on the number of traditional Chinese physicians per 10,000 residents in each administrative unit. The first, second, and third tertiles were considered areas with low, moderate, and high physician densities, respectively. Based on the Ministry of Health and Welfare criteria, low income status was defined as qualification for waived medical co-payments. Statistical analysis To observe the trend of acupuncture use, the annual prevalence of stroke patients using acupuncture treatment was calculated from 2000 to 2008. We used chi-square tests to compare the difference in sociodemographics, coexisting medical conditions, and characteristics of hospitalization between stroke patients who did and did not use acupuncture. Univariate and multivariate logistic regression analyses were performed to calculate crude and adjusted odds ratios (ORs) and 95 % confidence intervals (CIs) that measured the relationships between acupuncture use and associated characteristics in stroke patients. These characteristics included sex, age, income status, occupation, urbanization, density of traditional Chinese physicians in the area, disease history, use of rehabilitation, type of stroke, and in-hospital characteristics. All analyses were performed using Statistical Analysis Software version 9.1 (SAS Institute Inc., Cary, North Carolina, USA). A two-sided probability value of <0.05 was considered statistically significant. Results The prevalence of acupuncture use among stroke patients increased from 12 % in 2000 to 17 % in 2008 (p < 0.0001) (Fig. 1). A higher incidence of acupuncture use was found in men, younger patients, higher income people, white-collar employees, residents living in highly urbanized areas, and areas with more TCM physicians.Fig. 1 Prevalence of Acupuncture use among stroke patients in Taiwan from 2000–2008 (*Cochran-Armitage Trend Test) The multivariate logistic regression analysis yielded the ORs of factors associated with acupuncture use in stroke patients (Tables 1 and 2), including female gender (OR 1.04, 95 % CI 1.01–1.06), age 30–39 years (OR 4.05, 95 % CI 3.77–4.36), very high income status (OR 1.55, 95 % CI 1.47–1.65), white-collar employee status (OR 1.16, 95 % CI 1.12–1.20), residence in highly urbanized areas (OR 1.44, 95 % CI 1.37–1.52), residence in areas with more TCM physicians (OR 1.43, 95 % CI 1.39–1.48), and use of other types of rehabilitation (OR 3.20, 95 % CI 3.13–3.27). Acupuncture users also experienced greater incidences of hypertension (OR 1.18, 95 % CI 1.15–1.21) and hyperlipidemia (OR 1.30, 95 % CI 1.26–1.35) but lower incidences of myocardial infarction (OR 1.18, 95 % CI 1.09–1.28) and renal dialysis (OR 2.76, 95 % CI 2.45–3.13). Ischemic stroke (OR 1.21, 95 % CI 1.15–1.28) and longer hospitalization (OR 1.23, 95 % CI 1.19–1.27) were also associated with acupuncture use.Table 1 Comparison of sociodemographic characteristics between stroke patients with and without acupuncture treatment in 2000–2008 Acupuncture use p-value OR (95 % CI)* No (N = 242213) Yes (N = 42788) Sex n (%) n (%) <0.0001  Women 101,397 (85.5) 17,225 (14.5) 1.04 (1.01–1.06)  Men 140,816 (84.6) 25,563 (15.4) 1.00 (reference) Age, years <0.0001  20–29 1819 (83.1) 369 (16.9) 3.44 (3.04–3.89)  30–39 5680 (78.9) 1517 (21.1) 4.05 (3.77–4.36)  40–49 20,319 (79.3) 5298 (20.7) 3.79 (3.60–4.00)  50–59 39,190 (79.6) 10,054 (20.4) 3.56 (3.39–3.73)  60–69 58,684 (82.6) 12,374 (17.4) 3.05 (2.91–3.18)  70–79 74,891 (87.8) 10,424 (12.2) 2.04 (1.95–2.13)   ≥ 80 41,630 (93.8) 2752 (6.2) 1.00 (reference)  Mean ± SD 67.4 ± 13.3 62.5 ± 12.5 <0.0001 Income <0.0001  Very low 13,611 (86.1) 2203 (13.9) 1.00 (reference)  Low 71,645 (85.4) 12,283 (14.6) 1.23 (1.16–1.30)  Moderate 42,567 (84.9) 7565 (15.1) 1.31 (1.24–1.38)  High 88,550 (87.1) 13,152 (12.9) 1.30 (1.23–1.38)  Very high 25,840 (77.3) 7585 (22.7) 1.55 (1.47–1.65) Occupation <0.0001  White collar 73,946 (82.0) 16,188 (18.0) 1.16 (1.12–1.20)  Blue collar 115,906 (86.5) 18,067 (13.5) 1.00 (reference)  Other 52,361 (86.0) 8533 (14.0) 1.06 (1.02–1.11) Urbanization <0.0001  Low 17,323 (90.3) 1858 (9.7) 1.00 (reference)  Moderate 93,336 (87.4) 13,509 (12.6) 1.21 (1.15–1.28)  High 131,554 (82.8) 27,421 (17.3) 1.44 (1.37–1.52) Density of TCM <0.0001  Low 64,074 (88.8) 8117 (11.2) 1.00 (reference)  Moderate 121,661 (84.6) 22,199 (15.4) 1.17 (1.14–1.21)  High 56,478 (81.9) 12,472 (18.1) 1.43 (1.39–1.48)  Mean ± SD 1.6 ± 1.2 1.8 ± 1.3 <0.0001 *Logistic regression model included sociodemographics and medical conditions; Hosmer-Lemeshow goodness of fit, p-value = 0.0006; c-statistic = 0.71; CI, confidence interval; OR, odds ratio; TCM, traditional Chinese medicine Table 2 Medical conditions of stroke patients with and without acupuncture treatment in 2000–2008 Acupuncture No (N = 285329) Yes (N = 47832) p-value OR (95 % CI)* Rehabilitation <0.0001  No 166,526 (90.7) 17,087 (9.2) 1.00 (reference)  Yes 75,687 (74.7) 25,701 (25.3) 3.20 (3.13–3.27) Diabetes <0.0001  No 172,698 (85.2) 29,983 (14.8) 1.00 (reference)  Yes 69,515 (84.4) 12,805 (15.6) 1.00 (0.98–1.03) Hypertension <0.0001  No 93,695 (86.0) 15,290 (14.0) 1.00 (reference)  Yes 148,518 (84.4) 27,498 (15.6) 1.18 (1.15–1.21) Hyperlipidemia <0.0001  No 219,658 (85.4) 37,503 (14.6) 1.00 (reference)  Yes 22,555 (81.0) 5285 (19.0) 1.30 (1.26–1.35) MI <0.0001  No 236,962 (84.9) 42,037 (15.1) 1.18 (1.09–1.28)  Yes 5251 (87.5) 751 (12.5) 1.00 (reference) Dialysis <0.0001  No 238,140 (84.9) 42,500 (15.1) 2.76 (2.45–3.13)  Yes 4073 (93.4) 288 (6.6) 1.00 (reference) Type of Stroke <0.0001  Hemorrhage 53,950 (83.4) 10,714 (16.6) 1.02 (0.96–1.08)  Ischemia 173,573 (85.2) 30,155 (14.8) 1.21 (1.15–1.28)  Other 14,690 (88.5) 1919 (11.6) 1.00 (reference) LOS, days <0.0001  1–5 84,534 (87.4) 12,172 (12.6) 1.00 (reference)  6–9 63,969 (85.9) 10,499 (14.1) 1.06 (1.03–1.09)  10–14 34,261 (83.8) 6605 (16.2) 1.16 (1.12–1.20)  15–19 16,537 (82.7) 3465 (17.5) 1.22 (1.17–1.28)   ≥20 43,092 (81.1) 10,047 (18.9) 1.23 (1.19–1.27)  Mean ± SD 12.8 ± 17.5 14.9 ± 17.0 <0.0001 *Logistic regression model included sociodemographics and medical conditions; Hosmer-Lemeshow goodness of fit, p-value = 0.001; c-statistic = 0.71; CI, confidence interval; LOS, length of stay; MI, myocardial infarction; OR, odds ratio The average number of acupuncture visits in stroke patients was higher in males than in females (6.5 ± 7.8 vs. 6.4 ± 7.7, p < 0.001) (Tables 3 and 4). Stroke patients who had a low income, were white-collar employees, lived in highly urbanized areas and areas with more traditional Chinese physicians, used rehabilitation services, suffered from hemorrhagic stroke or ischemia, or had a longer hospitalization made more acupuncture visits. Patients who had more acupuncture treatment visits also had a higher expenditure on acupuncture.Table 3 Post-stroke visits and expenditure on acupuncture treatment within one year in stroke patients in 2000–2008 by sociodemographics Medical visits Medical expenditure n Mean ± SD p-value Mean ± SD p-value Sex 0.0155 0.0264  Women 17,225 6.4 ± 7.7 212 ± 258  Men 25,563 6.5 ± 7.8 219 ± 271 Age 0.0342 <0.0001  20–29 369 6.4 ± 7.0 219 ± 272  30–39 1517 6.8 ± 7.8 240 ± 302  40–49 5298 6.7 ± 8.2 226 ± 279  50–59 10,054 6.5 ± 7.7 217 ± 267  60–69 12,374 6.4 ± 7.6 210 ± 254  70–79 10,424 6.4 ± 7.7 216 ± 263   ≥80 2752 6.2 ± 7.9 208 ± 267 Income <0.0001 <0.0001  Very low 12,172 5.0 ± 6.5 168 ± 220  Low 10,499 5.7 ± 7.1 187 ± 234  Moderate 6605 6.7 ± 7.9 217 ± 262  High 3465 7.6 ± 8.4 246 ± 276  Very high 10,047 8.5 ± 8.9 295 ± 321 Occupation <0.0001 <0.0001  White collar 16,188 6.9 ± 8.0 229 ± 276  Blue collar 18,067 6.1 ± 7.4 203 ± 253  Other 8533 6.6 ± 7.9 219 ± 269 Urbanization <0.0001 <0.0001  Low 1858 5.2 ± 6.8 173 ± 229  Moderate 13,509 6.0 ± 7.4 202 ± 259  High 27,421 6.8 ± 8.0 226 ± 270 Density TCM <0.0001 <0.0001  Low 8117 5.9 ± 7.2 198 ± 250  Moderate 22,199 6.5 ± 7.8 215 ± 263  High 12,472 6.7 ± 7.9 230 ± 279 TCM, traditional Chinese medicine Table 4 Post-stroke visits and expenditure on acupuncture treatment within one year in stroke patients in 2000–2008 by medical condition Medical visits Medical expenditure n Mean ± SD p-value Mean ± SD p-value Rehabilitation <0.0001 <0.0001  No 17,087 4.8 ± 6.3 157 ± 206  Yes 25,701 7.6 ± 8.4 256 ± 292 Diabetes mellitus 0.02 0.01  No 29,983 6.5 ± 7.8 218 ± 268  Yes 12,805 6.3 ± 7.7 211 ± 259 Hypertension 0.77 0.93  No 15,290 6.5 ± 7.8 216 ± 269  Yes 27,498 6.5 ± 7.7 216 ± 263 Hyperlipidemia 0.35 0.98  No 37,503 6.5 ± 7.8 216 ± 266  Yes 5285 6.4 ± 7.6 216 ± 259 MI 0.14 0.15  No 42,037 6.5 ± 7.8 216 ± 266  Yes 751 6.1 ± 7.0 204 ± 240 Renal dialysis 0.0002 <0.0001  No 42,500 6.5 ± 7.8 217 ± 266  Yes 288 5.1 ± 6.1 161 ± 220 Length of stay, days <0.0001 <0.0001  1–5 12,172 5.0 ± 6.5 168 ± 220  6–9 10,499 5.7 ± 7.1 187 ± 234  10–14 6605 6.7 ± 7.9 217 ± 262  15–19 3465 7.6 ± 8.4 246 ± 276   ≥20 10,047 8.5 ± 8.9 295 ± 321 Type of Stroke <0.0001 <0.0001  Hemorrhage 10,714 7.4 ± 8.5 253 ± 296  Ischemia 30,155 6.3 ± 7.5 207 ± 256  Other 1919 4.7 ± 6.3 154 ± 200 MI, myocardial infarction Discussion Our study found that the prevalence of acupuncture use among stroke patients significantly increased from 2000 to 2008 in Taiwan, and the sociodemographic characteristics were highly correlated with recent acupuncture use. In contrast to the previous report that was based on a cross-sectional sample [17], this study included all of Taiwan’s stroke patients and evaluated the patterns of acupuncture use. The incidence of acupuncture use in the general population has been reported in the United States as 4.1 % in 2002 and 6.8 % in 2007 [6, 20]. The increasing use of acupuncture was investigated in the western countries [6, 20]. However, Chinese herbal medicine is not common in western countries. In Taiwan, Chinese herbal medicine and acupuncture were covered in the traditional Chinese medicine which is commonly used. Frequent use of acupuncture treatment was found in people with chronic diseases, such as osteoarthritis [21], cancer [22], and stroke [16]. The use of CAM is a trend, and it is not surprising that our study found that the use of acupuncture in stroke patients increased from 12 % in 2000 to 17 % in 2008. Demographic factors, such as age and sex, are associated with the patient’s choice of acupuncture [20, 23]. Compared with men, women were more likely to use acupuncture in this study. Several surveys also showed similar findings, namely, women had a higher use of TCM than men [17, 24]. Younger stroke patients were more likely to use acupuncture than older patients in this study. The association between young or middle age and acupuncture use was investigated in previous studies [17, 20, 23, 24]. The finding of better functional outcomes in younger stroke patients is not unexpected, as they had more home support and motivation [25]. A previous study suggested that young people seek more effective ways to improve their well-being and health and to relieve disease symptoms [26]. It is reasonable that younger stroke patients had a higher tendency to use acupuncture and to have a higher acupuncture expenditure than older patients in this study. The increasing use of TCM is somewhat related to the growth of the number of TCM physicians in Taiwan [17]. A previous investigation reported that the density of TCM increased from 1.39 physicians per 10,000 residents in 1996 to 1.78 physicians per 10,000 residents in 2001 [27]. Our results also showed that the prevalence of acupuncture use in stroke patients increased with the increasing density of TCM physicians. High urbanization was associated with TCM use in a previous survey [17, 24]. Because acupuncture is a subtype of TCM, it is not surprising that we found in the present study that people who lived in an urbanized area were more likely to use acupuncture. Economic growth is a determinant of physician supply and utilization of medical services [28]. In Taiwan, TCM has become an increasingly popular form of medicine, particularly after the implementation of the National Health Insurance in the medical care system since 1995. In this study, the frequency of acupuncture use and related insurance-paid expenditure were higher in stroke patients with a low income than in those without a low income. Co-payment is considered an important factor in the use of medical services [29]. According to the Ministry of Health and Wealth [17–19], patients with low-income status who do not need to pay a co-payment when receiving medical services may have more medical visits for acupuncture treatment than stroke patients with higher incomes. Among the stroke patients, diabetes, hypertension, hyperlipidemia, myocardial infarction, and kidney insuffucuency were common coexisting medical conditions that were also considered comorbidities in this study [30]. Acupuncture treatment can lower blood pressure [31, 32]. Moreover, evidence-based studies have shown acupuncture’s beneficial effects in addressing physical illness in stroke patients [12–14, 33]. We found that patients with comorbidities of hypertension or hyperlipidemia were more likely to undergo acupuncture treatment. People with more chronic diseases were likely to use TCM, which was confirmed in previous surveys [17, 24]. In this study, we found that stroke patients who used acupuncture were more likely to simultaneously undergo conventional medical rehabilitation. Previous reports showed a high use of CAM in stroke survivors in many countries [16, 34]. Medical pluralism, such as adopting more than one medical system or the use of both conventional medicine and CAM for health and illness, is also common in Taiwan [35, 36]. Stroke patients with longer hospitalizations were more likely to undergo acupuncture treatment. Stroke patients who had a longer length of hospital stay may have more neurological impairment [37], so they may require more or longer rehabilitation for stroke. Our study showed that the frequency of acupuncture visits and related expenditure were higher in stroke patients after a longer length of hospital stay. The incidence of ischemia is higher than that of other stroke subtypes [38]; however, subarachnoid hemorrhage resulted in a higher re-admission rate or mortality than ischemic stroke, and patients with a subarachnoid hemorrhage used more medical resources following stroke admission [39]. Conventional therapy with acupuncture treatment for acute ischemic stroke for four weeks has been shown to improve self-care ability and quality of life compared with sham acupuncture [40]. In this study, patients with ischemic stroke were more likely to use acupuncture than those with a subarachnoid hemorrhage. The frequency of acupuncture visits and related expenditures were higher in ischemic stroke patients. This study had some limitations. First, we used retrospective medical claims data from health insurance claims data that lacked detailed patient information on clinical risk scores (e.g., National Institute of Health Stroke Scale score, Barthel index, and Rivermead index) and lifestyle, physical, psychiatric, and biochemical measures. We were unable to determine whether these factors were causally related to acupuncture use. Second, our study used ICD-9-CM codes claimed by physicians for the diagnosis of stroke without clarifying the severity of disease. Third, information on folk therapy was not available in the National Insurance Research Database. In addition, in-come measurement and out-come measurements of stroke patients are unavailable in this study, which are valuable indicators for efficacy of acupuncture treatment. It is also one of our study limitations. Finally, this study was based on cross-sectional analyses of acupuncture use in stroke patients. Understanding the benefit of acupuncture in stroke patients requires further cohort studies. Conclusions In conclusion, the application of acupuncture in stroke patients is well accepted and increasing in Taiwan. The use of acupuncture in stroke patients is associated with sociodemographic factors and clinical characteristics. Abbreviations CAMComplementary and alternative medicine CIConfidence interval ICD-9-CMInternational classification of diseases, 9th revision, clinical modification OROdds ratio TCMTraditional Chinese medicine Dr. Chien-Chang Liao has equal contribution with the first author; Prof. Jaung-Geng Lin has equal contribution with the corresponding author. This study is based in part on data obtained from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health and managed by the National Health Research Institutes. The interpretation and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health, or National Health Research Institutes. Funding This study was supported in part by Shuang Ho Hospital, Taipei Medical University (104TMU-SHH-23), Taiwan’s Ministry of Science and Technology (MOST105-2629-B-038-001, MOST104-2314-B-038-027-MY2, MOST103-2320-B-214-010-MY2), Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (MOHW105-TDU-B-212-133019). Availability of data and material All data supporting the study is presented in the manuscript or available upon request from the corresponding author of this manuscript (Chun-Chuan Shih) at Email: hwathai@seed.net.tw. Authors’ contributions All authors contributed substantially toward the design of the study, the analysis and interpretation of the data, drafting and revising the manuscript. All authors approved the final version. Competing interests The authors declare that they have no competing interests. Consent for publication All authors have read and agreed to all the contents for publication. Ethics approval and consent to participate This study was approved by Taiwan’s National Health Research Institute (NHIRD-100-122) and the Institutional Review Board of E-DA Hospital, Kaohsiung, Taiwan (2014012). ==== Refs References 1. Harris PE Cooper KL Relton C Thomas KJ Prevalence of complementary and alternative medicine (CAM) use by the general population: a systematic review and update Int J Clin Pract 2012 66 924 39 10.1111/j.1742-1241.2012.02945.x 22994327 2. Barnes PM Bloom B Nahin RL Complementary and alternative medicine use among adults and children: United States, 2007 Natl Health Stat Report 2008 12 1 23 19361005 3. Hunt KJ Coelho HF Wider B Perry R Hung SK Terry R Complementary and alternative medicine use in England: results from a national survey Int J Clin Pract 2010 64 1496 502 10.1111/j.1742-1241.2010.02484.x 20698902 4. Nahin RL Barnes PM Stussman BJ Bloom B Costs of complementary and alternative medicine (CAM) and frequency of visits to CAM practitioners: United States, 2007 Natl Health Stat Report 2009 18 1 14 19771719 5. Chon TY Lee MC Acupuncture Mayo Clin Proc 2013 88 1141 6 10.1016/j.mayocp.2013.06.009 24079683 6. Upchurch DM Rainisch BW A sociobehavioral wellness model of acupuncture use in the United States, 2007 J Altern Complement Med 2014 20 32 9 10.1089/acm.2012.0120 23414108 7. Go AS Mozaffarian D Roger VL Benjamin EJ Berry JD Blaha MJ Heart disease and stroke statistics--2014 update: a report from the American Heart Association Circulation 2014 129 e28 292 10.1161/01.cir.0000441139.02102.80 24352519 8. Luengo-Fernandez R Gray AM Rothwell PM Costs of stroke using patient-level data: a critical review of the literature Stroke 2009 40 e18 23 10.1161/STROKEAHA.108.529776 19109540 9. Kumar S Selim MH Caplan LR Medical complications after stroke Lancet Neurol 2010 9 105 18 10.1016/S1474-4422(09)70266-2 20083041 10. Tong X Kuklina EV Gillespie C George MG Medical complications among hospitalizations for ischemic stroke in the United States from 1998 to 2007 Stroke 2010 41 980 6 10.1161/STROKEAHA.110.578674 20203317 11. Ingeman A Andersen G Hundborg HH Svendsen ML Johnsen SP In-hospital medical complications, length of stay, and mortality among stroke unit patients Stroke 2011 42 3214 8 10.1161/STROKEAHA.110.610881 21868737 12. Sze FK Wong E Or KK Lau J Woo J Does acupuncture improve motor recovery after stroke? A meta-analysis of randomized controlled trials Stroke 2002 33 2604 19 10.1161/01.STR.0000035908.74261.C9 12411650 13. Wu P Mills E Moher D Seely D Acupuncture in poststroke rehabilitation: a systematic review and meta-analysis of randomized trials Stroke 2010 41 e171 9 10.1161/STROKEAHA.109.573576 20167912 14. Lee JA Park SW Hwang PW Lim SM Kook S Choi KI Acupuncture for shoulder pain after stroke: a systematic review J Altern Complement Med 2012 18 818 23 10.1089/acm.2011.0457 22924414 15. Zhao XF Du Y Liu PG Wang S Acupuncture for stroke: evidence of effectiveness, safety, and cost from systematic reviews Top Stroke Rehabil 2012 19 226 33 10.1310/tsr1903-226 22668677 16. Shah SH Engelhardt R Ovbiagele B Patterns of complementary and alternative medicine use among United States stroke survivors J Neurol Sci 2008 271 180 5 10.1016/j.jns.2008.04.014 18485369 17. Liao CC Lin JG Tsai CC Lane HL Su TC Wang HH An investigation of the use of traditional chinese medicine in stroke patients in Taiwan Evid Based Complement Alternat Med 2012 2012 387164 23304199 18. Liao CC Su TC Sung FC Chou WH Chen TL Does hepatitis C virus infection increase risk for stroke? A population-based cohort study PLoS One 2012 7 e31527 10.1371/journal.pone.0031527 22363662 19. Liao CC Chang PY Yeh CC Hu CJ Wu CH Chen TL Outcomes after surgery in patients with previous stroke Br J Surg 2014 101 1616 22 10.1002/bjs.9639 25224956 20. Burke A Upchurch DM Dye C Chyu L Acupuncture use in the United States: findings from the National Health Interview Survey J Altern Complement Med 2006 12 639 48 10.1089/acm.2006.12.639 16970534 21. Jong MC van de Vijver L Busch M Fritsma J Seldenrijk R Integration of complementary and alternative medicine in primary care: what do patients want? Patient Educ Couns 2012 89 417 22 10.1016/j.pec.2012.08.013 23031611 22. Gansler T Kaw C Crammer C Smith T A population-based study of prevalence of complementary methods use by cancer survivors: a report from the American Cancer Society’s studies of cancer survivors Cancer 2008 113 1048 57 10.1002/cncr.23659 18680170 23. Zhang Y Lao L Chen H Ceballos R Acupuncture use among American adults: what acupuncture practitioners can learn from National Health Interview Survey 2007? Evid Based Complement Alternat Med 2012 2012 710750 22474517 24. Shih CC Liao CC Su YC Tsai CC Lin JG Gender differences in traditional Chinese medicine use among adults in Taiwan PLoS One 2012 7 e32540 10.1371/journal.pone.0032540 22539941 25. Knoflach M Matosevic B Rücker M Furtner M Mair A Wille G Functional recovery after ischemic stroke--a matter of age: data from the Austrian Stroke Unit Registry Neurology 2012 78 279 85 10.1212/WNL.0b013e31824367ab 22238419 26. Chang LC Huang N Chou YJ Lee CH Kao FY Huang YT Utilization patterns of Chinese medicine and Western medicine under the National Health Insurance Program in Taiwan, a population-based study from 1997 to 2003 BMC Health Serv Res 2008 8 170 10.1186/1472-6963-8-170 18691428 27. Yang CH Huang YT Hsueh YS Redistributive effects of the National Health Insurance on physicians in Taiwan: a natural experiment time series study Int J Equity Health 2013 12 13 10.1186/1475-9276-12-13 23374629 28. Cooper RA Getzen TE Laud P Economic expansion is a major determinant of physician supply and utilization Health Serv Res 2003 38 675 96 10.1111/1475-6773.00139 12785567 29. Kiil A Houlberg K How does copayment for health care services affect demand, health and redistribution? A systematic review of the empirical evidence from 1990 to 2011 Eur J Health Econ 2014 15 813 28 10.1007/s10198-013-0526-8 23989938 30. Lackland DT Elkind MS D’Agostino R Sr Dhamoon MS Goff DC Jr Higashida RT Inclusion of stroke in cardiovascular risk prediction instruments: a statement for healthcare professionals from the American Heart Association/American Stroke Association Stroke 2012 43 1998 2027 10.1161/STR.0b013e31825bcdac 22627990 31. Flachskampf FA Gallasch J Gefeller O Gan J Mao J Pfahlberg AB Randomized trial of acupuncture to lower blood pressure Circulation 2007 115 3121 9 10.1161/CIRCULATIONAHA.106.661140 17548730 32. Wang J Xiong X Liu W Acupuncture for essential hypertension Int J Cardiol 2013 169 317 26 10.1016/j.ijcard.2013.09.001 24060112 33. Geeganage C Beavan J Ellender S Bath PM Interventions for dysphagia and nutritional support in acute and subacute stroke Cochrane Database Syst Rev 2012 10 CD000323 23076886 34. Shin YI Yang CY Joo MC Lee SG Kim JH Lee MS Patterns of using complementary and alternative medicine by stroke patients at two university hospitals in Korea Evid Based Complement Alternat Med 2008 5 231 5 10.1093/ecam/nem025 18604256 35. Shih SF Lew-Ting CY Chang HY Kuo KN Insurance covered and non-covered complementary and alternative medicine utilisation among adults in Taiwan Soc Sci Med 2008 67 1183 9 10.1016/j.socscimed.2008.06.011 18649983 36. Shih CC Su YC Liao CC Lin JG Patterns of medical pluralism among adults: results from the 2001 National Health Interview Survey in Taiwan BMC Health Serv Res 2010 10 191 10.1186/1472-6963-10-191 20604913 37. Luengo-Fernandez R Gray AM Rothwell PM Population-based study of determinants of initial secondary care costs of acute stroke in the United Kingdom Stroke 2006 37 2579 87 10.1161/01.STR.0000240508.28625.2c 16946157 38. Tsai CF Thomas B Sudlow CL Epidemiology of stroke and its subtypes in Chinese vs white populations: a systematic review Neurology 2013 81 264 72 10.1212/WNL.0b013e31829bfde3 23858408 39. Chang KC Lee HC Huang YC Hung JW Chiu HE Chen JJ Cost-effectiveness analysis of stroke management under a universal health insurance system J Neurol Sci 2012 323 205 15 10.1016/j.jns.2012.09.018 23046751 40. Shen PF Kong L Ni LW Guo HL Yang S Zhang LL Acupuncture intervention in ischemic stroke: a randomized controlled prospective study Am J Chin Med 2012 40 685 93 10.1142/S0192415X12500516 22809024

PMC005xxxxxx/PMC5002128.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 356710.1186/s12889-016-3567-yResearch ArticleEffectiveness of psychotherapeutic consultation in the workplace: a controlled observational trial http://orcid.org/0000-0002-8295-2632Rothermund Eva eva.rothermund@uni-ulm.de 1Gündel Harald harald.guendel@uniklinik-ulm.de 1Rottler Edit edit.rottler@uniklinik-ulm.de 1Hölzer Michael hoelzer_michael@sonnenbergklinik.de 2Mayer Dorothea dorothea.mayer@daimler.com 3Rieger Monika monika.rieger@med.uni-tuebingen.de 4Kilian Reinhold reinhold.kilian@bkh-guenzburg.de 51 Department of Psychosomatic Medicine and Psychotherapy, University Hospital Ulm, Albert-Einstein-Allee 23, 89081 Ulm, Germany 2 ZfP Suedwuerttemberg, Sonnenbergklinik, 70597 Stuttgart, Germany 3 Health and Safety Sindelfingen, Daimler AG, 71059 Sindelfingen, Germany 4 Institute for Occupational and Social Medicine and Health Services Research, University Clinic Tuebingen, Competence Centre Health Services Research, Medical Faculty Tuebingen, 72074 Tuebingen, Germany 5 Department of Psychiatry II, University Hospital Ulm, BKH Guenzburg, 89312 Guenzburg, Germany 26 8 2016 26 8 2016 2016 16 1 8913 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 This study compares the effectiveness of psychotherapeutic consultation in the workplace (PSIW) with psychotherapeutic outpatient care (PSOC) in Germany. Methods Work ability (WAI), quality of life (SF-12), clinical symptoms (PHQ) and work-related stress (MBI, IS) were assessed in 367 patients seeking mental health care via two routes (PSIW n = 174; PSOC n = 193) before consultation and 12 weeks later. Changes in outcome variables were assessed using covariance analysis with repeated measures (ANCOVA) with sociodemographic variables (propensity score method), therapy dose, setting and symptom severity as covariates. Results The PSIW and PSOC groups included 122 and 66 men respectively. There were 102 first-time users of mental healthcare in the PSIW group and 83 in the PSOC group. There were group differences in outcome variables at baseline (p < 0.05); PSIW patients were less impaired overall. There were no group difference in sociodemographic variables, number of sessions within the offer or symptom severity. There was no main effect of group on outcome variables and no group*time interaction. Work-related stress indicators did not change during the intervention, but work ability improved in both groups (F = 10.149, p = 0.002; baseline M = 27.2, SD = 8.85); follow-up M = 28.6, SD = 9.02), as did perceived mental health (SF-12 MCS), depression (PHQ-9) and anxiety (PHQ-7). Effect sizes were between η2 = 0.028 and η2 = 0.040. Conclusions Psychotherapeutic consultation is similarly effective in improving patients’ functional and clinical status whether delivered in the workplace or in an outpatient clinic. Offering mental health services in the workplace makes it easier to reach patients at an earlier stage in their illness and thus enables provision of early and effective mental health care. Trial registration DRKS00003184, retrospectively registered 13 January 2012. Keywords Common mental disorderWorkplace based interventionPsychotherapyMental healthOccupational psychiatryHealth services researchyoung scientists programme run by the German network Health Services Research Baden-Wuerttemberg of the Ministry for Science, Research and Arts in collaboration with the Ministry for Work and Social Affairs, Family, Women, and Senior Citizens, Baden-Wuerttn.a.Rothermund Eva issue-copyright-statement© The Author(s) 2016 ==== Body Background Common mental disorders (CMD) will be one of the main contributors to the global economic burden of non-communicable diseases by 2030 [1]. Recently it has been anticipated that the direct and indirect costs of CMD amount to 3.5 % of the gross domestic product of industrial countries on average [2]. CMD and even sub-threshold mental health conditions have adverse effects on everyday life, work capacity and even mortality [3–6]. Because of their prevalence and costs CMD have an enormous impact on public health [7]. The increase in the prescription of psychotropic drugs, such as antidepressants, demonstrates that in practice this is the preferred approach to treatment, although it has been shown that a combination of psychotherapy and medication is the most effective treatment for most CMD [8, 9]. German guidelines recommend pharmacotherapy or psychotherapy as treatments for mild to moderate depressive disorders and explicitly recommend a combination of both for severe depression [10]. The treatment gap for CMD is estimated at 55 % globally while the improvement of treating CMD effectively has been considered to be very encouraging [11, 12]. Long waiting periods for initial consultations and psychotherapy are obstacles to appropriate treatment in healthcare systems across Europe, including in Belgium, Denmark, the Netherlands and Germany [13–17]. The interface between primary and secondary care may represent another barrier to effective treatment. In most countries general practitioners (GPs) mediate access to secondary services, i.e. specialist care. Unfortunately GPs often struggle to identify mental health problems [15, 18]. Fear of stigmatisation is another obstacle to timely treatment [19]. Thus there is a need to strengthen psychotherapeutic approaches by making them more accessible. 20 % of the working age population experience CMD at any one time [20]. The workplace has been identified as a key social context for early identification and treatment of mental health problems [2, 21, 22]. Focusing on what improves return to work after being sick listed due to depression [23] or on the effectiveness of work-directed interventions within the return to work process [24] restricted to randomised controlled trials showed positive but weak effects on work-related outcomes such as returning to work faster. Putting these findings together and including controlled trials Pomaki et al. [25] emerged important factors of how workplace-based interventions improve work disability outcomes: facilitation of access to clinical treatment, and availability of workplace-based psychotherapeutic interventions [25]. Chronic mental health problems, including subclinical symptoms, and impaired work ability are strong predictors of long sickness absence [26, 27]; this underlines the importance of early intervention. Collaborative care offers (CCO) such as “psychotherapeutic consultation in the workplace” have been established as part of programmes to promote mental health in the workplace and are well-accepted by patients [28, 29]. Although the workplace environment is considered by many experts as an appropriate setting for a psychosocial intervention, there are currently no studies on the effectiveness of such interventions. A standard form of mental health treatment, short-term psychotherapeutic outpatient care (PSOC), has been adapted for delivery in the workplace under the label “psychotherapeutic consultation in the workplace” (PSIW) [30–32]. In this article we present the results of an observational controlled trial on the effectiveness of the PSIW program in comparison to the PSOC as a measure of routine care. Methods Study design Due to reasons of the involved companies, staff policy and legal restrictions it was not possible to conduct a randomised controlled trial. Therefore, a prospective, controlled, observational trial comparing 174 employees from three companies who received PSIW with 193 outpatients from two clinics who received PSOC was conducted. Outcome indicators were assessed prior to the initial consultation (t1) and 12 weeks later (t2) [33]. Study context In the German healthcare system patients with CMD are treated by physicians specialising in psychiatry or psychosomatic medicine or by psychological psychotherapists. Treatment is usually delivered through private practices, the outpatient clinics of psychosomatic hospitals and psychosomatic departments or psychosomatic outpatient clinics at general hospitals [34]. In the German healthcare system PSOC is covered by statutory health insurance as well as by the private health insurance, and nearly 100 % of the population is covered by health insurance. Thus comprehensive care should be available to all those who need it [35]. In spite of this the treatment gap for CMD in Germany is comparable to that in other European countries [36–38]. Intervention Psychotherapeutic consultation in the workplace (PSIW) PSIW is provided by a mental health expert (medical or psychological psychotherapist) who, although not employed by the host company, may be paid by the company (as an external contractor) or by the company’s health insurance funds. The service is available to all staff free of charge. Staff members are usually informed about the service by the company physician. In one company employees must be referred to PSIW by the company physician but in others self-referral is possible. If self-referral is available information about PSIW (nature of the service; location; how to make an appointment) is communicated via paper and online. In the first session the user’s clinical status and needs are being assed. This assessment is used to determine severity of the mental health problem and whether workplace consultation is a suitable treatment option or whether additional or more intensive mental health care (e.g. outpatient psychotherapy; short-term (8–10 sessions) psychotherapy; psychopharmacological treatment; inpatient treatment) is needed. Each session lasts 50–60 minutes and a maximum of four sessions can be offered under the PSIW programme. After assessment the PSIW user will be informed about any further therapeutic steps that are indicated, motivating the service user to obtain additional help and providing specific advice to help individuals overcome barriers to service use. This involves providing the PSIW user with information about CMD and psychotherapeutic approaches to treatment. The strengths and resources of the patient are stressed and further treatments are recommended. If appropriate information about self-help books, counselling centres, e.g. family or drug and addiction counselling, and other services such as workshops on relaxation techniques is provided. Recommendations are made according to the clinical guidelines of the Association of the Scientific Medical Societies in Germany (AWMF) as they exist e.g. for depression [10]. PSIW takes place only in companies that agree to offer the service and provide structural support. Company support includes appointment making and providing a suitable location for consultations. Subject to patient consent the next steps in treatment are determined in a case conference involving the occupational health physicians and the psychotherapeutic consultant. Psychotherapeutic outpatient care (PSOC) Initial PSOC treatment is limited to two sessions and the core elements are assessment of clinical symptoms and service needs, provision of information about CMD and treatment methods and recommendations for further treatment. Referral to PSOC is predominantly via GPs. Self-referral is encouraged and information for patients drafted via the clinic websites. Eligibility criteria and recruitment process To be eligible for the study participants had to be at least 18 years old, capable of understanding and writing German and currently employed. Participants in the PSIW group had to be employed by one of the participating companies. Participants in the PSOC group were recruited consecutively from two outpatient clinics, the University Clinic of Psychosomatic Medicine and Psychotherapy, Ulm, and Sonnenbergklinik, Division of Psychosomatic Medicine of the ZfP Suedwuerttemberg, Stuttgart, from June 2012 to January 2013. Participants in the PSIW group were recruited consecutively from November 2011 to June 2013. Assessment Assessment of clinical status was made by means of self-administrated questionnaires (Table 1). Questionnaires for baseline assessment were handed over at the first meeting. Follow-up questionnaires were sent out by mail. Data were double entered manually and checked twice. Work ability Ability to work was the primary outcome indicator and was assessed using the work ability index (WAI), a self-report instrument used to assess current and future work ability and work demand management based on behavioural measures [39]. Use of mental health services and symptom duration Lifetime history of psychotherapeutic treatment and number of contacts with mental health services over the previous 12 months were recorded. Duration (in months) of the symptoms which had prompted the consultation was recorded by the therapist. Mental health and somatic symptoms Depression was assessed with the 9-item Patient Health Questionnaire - depression scale (PHQ-9) [40], anxiety with the 7-item Patient Health Questionnaire - generalised anxiety disorder scale (PHQ-7) [41] and somatic symptom severity with the 15-item Patient Health Questionnaire - somatisation (PHQ-15) [42]. Interpretation of PHQ scores was based on the diagnostic criteria of the DSM-IV and ICD-10, using the recommended cut-off of 10 or above to distinguish between clinical and non-clinical levels of symptoms [43]. Health-related quality of life Health-related quality of life was assessed using the SF-12, a validated, short version of the SF-36, to measure the functional health status of patients [44]. Weighted summation provided summary scores for perceived mental health (MCS; mental health component score) and perceived physical health (PCS; physical health component score). Work-related stress We assessed work-related stress using the German version (MBI-GS-D) of the Maslach Burnout Inventory (MBI), which is used to assess burnout syndrome as a manifestation of mental exhaustion [45]. The three components of burnout are: emotional exhaustion (EE), depersonalisation (DP) and reduced personal accomplishment (PA). Respondents indicate the frequency with which they experience each item. Sum scores were used instead of means for each subscale. Irritation Irritation is defined as subjectively perceived emotional and cognitive strain in occupational contexts [46, 47]. We assessed irritation with the irritation scale (IS). Statistical analyses Sample size determination We calculated that a sample of 220 participants would be needed to detect a medium (effect size f = 0.25) difference in WAI with a power of 0.95 at a significance level of p < 0.05 using repeated measures ANOVA. Bias control The propensity adjustment method was used for bias control [48]. Propensity scores were estimated as the conditional probability of belonging to the PSIW group using a logistic regression model including age, gender, family status (marital status, living with children or not, living with other adults or not), job status (school-leaving qualification, full or part-time employment, occupation: uneducated, skilled or white collar worker, managerial staff), duration of psychopathological symptoms, history of psychiatric or psychotherapeutic treatment (outpatient treatment: lifetime, previous 12-month, inpatient treatment: lifetime, previous 12-month, other physician’s care: previous 12-month, psychosocial treatment within company: previous 12-month, other like pastoral advice service etc.), and baseline clinical differences as independent variables. Number of sessions and presence of treatment between the first consultation and follow-up assessment were included in the model as covariates. Latent profile analysis (LPA) [49] including the WAI, PHQ-9 and the SF-12 was used to provide an overall indication of clinical differences at baseline. Details of the LPA procedure and results are published elsewhere (Rothermund et al. submitted). On the basis of the LPA participants were classified into four classes (1 = severely mentally ill, 2 = moderately affected with low quality of life, 3 = moderately affected with low work ability, 4 = less affected but at risk) which captured global differences in baseline clinical measures. Class membership was included in the logistic regression model as a categorical independent variable with class four as the reference category. Effectiveness analysis The effectiveness of the intervention was estimated with respect to primary and secondary outcomes on an intention to treat (ITT) basis. Missing endpoint data were imputed as last observation carried forward (LOCF). Change in outcome indicators was analysed by covariance analysis with repeated measures (ANCOVA) with propensity scores as covariates. Statistical analyses were performed using SPSS version 21. Results A total of 367 participants were included in the analyses (PSIW n = 174; PSOC n = 193) (see flow diagram, Fig. 1 and Table 2 for sociodemographic data). Due to missing propensity scores a total of 17 cases (4.6 %) was excluded from the analyses. Follow-up data were obtained of 60 % of participants. Reasons for dropout were not given. Non-respondents were younger (38, SD 11.3) than respondents (45, SD 10.7). No significant differences were found for gender, educational level, symptom severity or any of the psychometric measures. The PSIW group was slightly older (45 years, SD = 10.1) than the comparison group (40 years, SD = 12.1) and PSIW participants tended to have been experiencing symptoms for a shorter period of time (38 months, SD = 65.4) than PSOC participants (51 months, SD = 72.9). The majority of the PSIW group were men (n = 122, 70 %) and just over a third of the PSOC group was men (n = 66, 34 %). The PSIW group contained 102 first-time users of mental health services (65 %); the corresponding figure for the PSOC group was 83 (45 %). 12-month usage of mental health services (number of participants who had at least one contact with mental health services) was lower in the PSIW group (n = 65) than the PSOC group (n = 119).Fig. 1 Flow diagram showing participants numbers at different stages Table 1 Measures and instruments used Measure Instrument Range Category boundaries/interpretation Source WAI work ability index 7–49 49–44 very good work ability 43–37 good work ability 36–28 moderate work ability 27–7 very low work ability [64] SF-12-MCS SF-12-mental component score 0–100 to compare: normative German sample 1994: 51.2, psychosomatic inpatients 27 [44, 65] SF-12-pcs SF-12-physical component score 0–100 to compare: normative German sample 1994: 46.3, psychosomatic inpatients 40 PHQ-9-depr patient health questionnaire depression 0–27 0 to 4 minimal symptom burden 5 to 9 mild symptom burden 10 to 14 moderate symptom burden >15 severe symptom burden [43] PHQ-15-som patient health questionnaire somatoform symptom severity 0–30 PHQ-7-anx patient health questionnaire anxiety 0–21 ISGI irritation scale global index 8–56 no irritation 8–16 low irritation 17–26 moderate irritation 27–38 strong irritation 39–56 [46] MBI-EE burnout-emotional exhaustion 5–30 <17.5 unremarkable >17.5 alerting >25 critical [45] MBI-DP burnout- depersonalisation 5–30 low scores are healthier MBI-PA burnout- personal accomplishment 6–36 high scores are healthier Table 2 Sample description Characteristic total sample PSIW PSOC P value n = 367 n = 174 n = 193 age (years) mean (SD) 42.94 (11.47) 45.20 (10.12) 40.05 (12.07) *** b symptom duration (months) mean (SD) 44.94 (69.68) 38.02 (65.41) 51.06 (72.87) n.s. b male gender n (%) 188 (51.2) 122 (70.1) 66 (34.2) *** a in a stable relationship n (%) 239 (68.5) 126 (75.9) 113 (61.7) * a school-leaving qualification n (%) 364 (100) 171 (100) 193 (100) still at school 4 (1.1) 1 (0.6) 3 (1.6) n.s. a low (9 years) 106 (29.1) 57 (33.3) 49 (25.4) medium (10 years) 133 (36.6) 55 (32.2) 78 (40.4) high (<10 years) 121 (33.2) 58 (33.9) 63 (32.6) first-time user (no previous in- or out-patient treatment) n (%) 185 (54.1) 102 (65.4) 83 (44.6) ***a mental health care system user in the previous 12 months n (%) 184 (50.1) 65 (37.4) 119 (61.7) ***a SD standard deviation, n number, n.s. not significant, a chi-square test, b t-test * p < 0.05, ** p < 0.01, *** p < 0.001 Table 3 Effects of the PSIW and PSOC interventions on work ability and clinical and work-related parameters at baseline and follow-up Variable Group Baseline mean (SD) Follow-up mean (SD) Group effect, F-value, p, η2 Time effect F-value, p, η2 Group*Time effect, F-value, p, η2 WAI PSIW 29.5 (8.02) 30.8 (8.32) (1:317) = 1.832, 0.177, 0.006 (1:317) = 10.149, 0.002, 0.031 (1:317) = 1.602, 0.206, 0.005 PSOC 25.3 (9.07) 26.8 (9.21) all 27.2 (8.85) 28.6 (9.02) aΔ PSIW vs. PSOC p = 0.000 SF-12 MCS PSIW 33.3 (11.13) 37.5 (11.77) (1:315) = 0.155, 0.694, 0.000 (1:315) = 9.093, 0.003, 0.028 (1:315) = 6.950, 0.009, 0.022 PSOC 29.8 (10.24) 32.7 (11.07) all 31.4 (10.78) 34.9 (11.62) aΔ PSIW vs. PSOC p = 0.002 SF-12 PCS PSIW 45.8 (11.13) 45.4 (10.96) (1:315) = 0.321, 0.572, 0.001 (1:315) = 0.858, 0.355, 0.003 (1:315) = 0.679, 0.411, 0.002 PSOC 43.8 (11.31) 44.5 (10.90) all 44.7 (11.36) 45.0 (10.92) aΔ PSIW vs. PSOC p = 0.146 PHQ-9 PSIW 11.2 (5.71) 9.6 (6.05) (1:334) = 0.600, 0.439, 0.002 (1:334) = 13.946, 0.000, 0.040 (1:334) = 1.821, 0.178, 0.005 PSOC 14.4 (6.24) 12.4 (6.44) all 12.9 (6.20) 11.1 (6.40) aΔ PSIW vs. PSOC p = 0.000 PHQ-15 PSIW 10.5 (5.41) 9.7 (5.44) (1:319) = 0.527, 0.468, 0.002 (1:319) = 3.444, 0.064, 0.011 (1:319) = 1.891, 0.170, 0.006 PSOC 11.4 (5.44) 10.9 (5.56) all 11.0 (5.44) 10.3 (5.53) aΔ PSIW vs. PSOC p = 0.196 PHQ-7 PSIW 9.5 (5.36) 8.2 (5.31) (1:341) = 0.500, 0.480, 0.001 (1:341) = 12.259, 0.001, 0.035 (1:341) = 0.858, 0.355, 0.003 PSOC 11.3 (5.53) 9.5 (5.43) all 10.4 (5.50) 8.9 (5.41) aΔ PSIW vs. PSOC p = 0.005 IS PSIW 30.7 (11.74) 29.7 (12.23) (1:317) = 1.453, 0.229, 0.005 (1:317) = 2.243, 0.135, 0.007 (1:317) = 1.719, 0.147, 0.007 PSOC 35.2 (12.12) 34.2 (12.33) all 33.0 (12.13) 32.1 (12.47) aΔ PSIW vs. PSOC p = 0.001 MBI-EE PSIW 3.9 (1.27) 3.8 (1.27) (1:325) = 0.012, 0.912, 0.000 (1:325) = 4.199, 0.041, 0.013 (1:325) = 0.758, 0.385, 0.002 PSOC 4.5 (1.24) 4.3 (1.18) all 4.2 (1.28) 4.1 (1.25) aΔ PSIW vs. PSOC p = 0.000 MBI-DP PSIW 3.2 (1.11) 3.2 (1.10) (1:316) = 0.150, 0.699, 0.000 (1:316) = 0.524, 0.470, 0.002 (1:316) = 0.008, 0.927, 0.000 PSOC 3.5 (1.22) 3.5 (1.18) all 3.3 (1.17) 3.4 (1.15) aΔ PSIW vs. PSOC p = 0.015 MBI-PA PSIW 4.5 (0.82) 4.5 (0.78) (1:324) = 0.760, 0.384, 0.002 (1:324) = 0.437, 0.509, 0.001 (1:324) = 0.001, 0.980, 0.000 PSOC 4.4 (0.91) 4.3 (0.92) all 4.4 (0.86) 4.4 (0.86) aΔ PSIW vs. PSOC p = 0.115 Legend: a Δ PSIW vs. PSOC (p): Results of a t-test comparing PSIW and PSOC groups at baseline, η 2 partial square eta (effect size), p level of significance At two companies most referrals to PSIW were made by occupational physicians or social workers; at the third company employees were usually self-referred. For a more detailed description of procedures of PSIW see Rothermund et al. 2014 [31]. Baseline data (Table 3) At baseline the PSIW group had higher work ability (M = 29.5, SD = 8.02) than the PSOC group (M = 25.3, SD = 9.07; p < 0.001). The PSIW group also had better perceived mental health (SF-12 MCS: M = 33.3, SD = 11.13) than the PSOC group (M = 29.8, SD = 10.24, p = 0.002) at baseline. Depression scores at baseline were also lower in the PSIW group (M = 11.3, SD = 5.72) than the PSOC group (M = 14.5, SD = 6.28, p < 0.001). Anxiety scores showed a similar pattern (PSIW group: M = 9.6, SD = 5.36; PSOC group: M = 11.2, SD = 5.71; p < 0.001). There were no group differences in physical health (SF-12 PCS) and somatic symptom severity (PHQ-15) at baseline. Work-related stress was also lower in the PSIW group at baseline. IS scores were lower in the PSIW group (M = 30.7, SD = 11.74) than the PSOC group (M = 35.2, SD = 12.12, p = 0.001). The PSIW group also had lower scores on two aspects of burnout. EE was lower in the PSIW group (M = 3.9, SD = 1.27) than in the PSOC group (M = 4.5, SD = 1.24; p < 0.001) and the PSIW group showed less DP (M = 3.2, SD = 1.11) than the PSOC group (M = 3.5, SD = 1.22; p = 0.015). There was no baseline group difference in PA. Overall, patients that were recruited in the workplace context (PSIW group) appeared to be healthier than patients that were using established services. Effects of PSIW and PSOC on work ability and clinical and work-related parameters (Table 3) Both groups showed a similar improvement in work ability over time (baseline M = 27.2, SD = 8.85; follow-up M = 28.6, SD = 9.02; F = 10.149, p = 0.002). At follow-up the mean work ability of the PSIW group (M = 30.8, SD = 8.32) was higher than that of the PSOC group (M = 26.8, SD = 9.21). However there was no group effect and no group*time interaction for WAI score. None of the 3 covariates (propensity score; number of sessions; presence of further treatment) had an effect on WAI. Both groups showed similar improvements in perceived mental health (SF-12 MCS; baseline M = 31.4, SD = 10.78; follow-up M = 34.9, SD = 11.62; F = 9.093, p = 0.003), depression (PHQ-9; baseline M = 12.9, SD = 6.20; follow-up M = 11.1, SD = 6.40; F = 13.946, p < 0.001) and anxiety (PHQ-7; baseline M = 10.4, SD = 5.50; follow-up M = 8.9, SD = 5.41; F = 12.259, p = 0.001). The effect sizes for the changes were small (η2 between 0.028 and 0.040). There was no effect of group and no group*time interaction on these outcome parameters. There were no changes in measures of physical health (perceived physical health measured as SF-12 PCS score; somatoform symptoms, measured as PHQ-15 score) or work-related stress (all aspects of burnout: EE, DP and PA; IS score). Discussion In this trial we assessed work ability, quality of life, clinical symptoms and work-related stress in 367 patients who received mental health services via two programmes, PSIW and PSOC. Patients were assessed before their first consultation and 12 weeks later and in the analyses we controlled for variance in sociodemographic factors, therapy dose, setting and symptom severity. Our results indicate that psychotherapeutic consultation is similarly effective in improving patients’ functional and clinical status whether delivered in the workplace or in a psychotherapeutic outpatient clinic. Our data clearly suggest that workplace-based services reach people with mental health problems at an earlier state of illness than the traditional health care system. Thus, psychotherapeutic care in an occupational healthcare setting can be considered an early and effective form of mental health care. The work ability of individuals improved after the initial intervention independent of the setting in which patients sought help; this was a small effect (baseline WAI: M = 27.2 points, SD = 8.02; follow-up WAI: M = 28.6, SD = 9.02; η2 = 0.031; p = 0.002). The same pattern of improvement was observed for depression (baseline PHQ-9: M = 12.9, SD = 6.20; follow-up PHQ-9: M = 11.1, SD = 6.40; F = 13.946, p < 0.001) and anxiety (baseline PHQ-7: M = 10.4, SD = 5.50; follow-up PHQ-7: M = 8.9, SD = 5.41; F = 12.259, p = 0.001). Perceived mental health (SF-12 MCS), also known as health related quality of life, also improved (baseline M = 31.4, SD = 10.78; follow-up M = 34.9, SD = 11.62; F = 9.093, p = 0.003). Effect sizes for all improvements were small, between η2 = 0.028 and η2 = 0.040. The small effects are plausible as the interventions were very short: one or two sessions, primarily focusing on diagnostic assessment and facilitation of access to other clinical services. A recent study showed that better work ability at baseline was associated with better outcomes, e.g. early return to work [27]. This study focused on short time effects, as the follow-up assessment occurred 12 weeks after the initial consultation. This limitation may have obscured better outcomes in the group treated in the workplace due to the earlier detection of illness. Nevertheless the parallel improvement of clinical parameters and work ability is meaningful, as we know that recovery from depression does not automatically mean that a person is once more capable of dealing with the demands of everyday life [50–52]. One might have expected the PSIW group to show less spontaneous improvement than the PSOC group on the grounds that they were healthier and would therefore show a smaller response to the initial very short, primarily diagnostic intervention offered. Alternatively one might expect the PSIW group to show greater improvement as prognosis for mental health conditions is usually better in cases where intervention is provided early in the course of disease [53–55]. The findings of this study are in line with findings on return to work process following CMD-related absence. Vlasveld et al. 2012 [56] investigated employees absent from work with major depressive disorder and found that collaborative care produced similar effects to care as usual when outcome was measured as a continuous variable with PHQ-9; however the response rate was higher in the collaborative care group. This probably reflects an aspect of the collaborative care which was not captured in the analyses. Van der Feltz-Cornelis et al. 2010 [57] showed that return to work was faster in an intervention group than a group who received care as usual. The intervention consisted of treatment by occupational physicians who were advised by psychiatric consultants. Both groups showed comparable improvements in symptom severity and quality of life. During the 12-week period between first consultation and follow-up assessment there was no change in work-related stress (irritation, burnout) in either group, although there were improvements in overall mental well-being, depression and anxiety. There were also no changes on the scales measuring aspects of physical health, namely somatoform symptoms (PHQ-15) and perceived physical health (SF-12 PCS). The lack of effects on work-related stress and physical health may have been because the PSIW programme did not target work-related health or physical health but focused on diagnostic assessment and motivation for further psychotherapeutic treatment. Limitations and strengths Our observation that the group accessing psychotherapeutic services in the workplace did not improve more than those accessing services via conventional outpatient clinics may have been influenced by the short length of the observation period of only 12 weeks. We know from research on maintaining gains from psychotherapy that psychotherapeutic interventions may have long term effects and hence follow-up for at least at 12 month after intervention is recommended [58–60]. Our study also suffers from the familiar limitations associated with lack of a randomised group assignment. Thus, bias control was addressed using the propensity score method [61] with only rare (max. 6 %) reduction of sample size. The controlled observational design allowed us to manage the problem of artificial deterioration of the sample. Randomised controlled trials may suffer deterioration of the sample, due to the better control of intervening variables, which often leads to homogenous samples that differ from samples under routine conditions [60, 62, 63]. In our study the intervention was studied under routine conditions and thus the study has good external validity and the findings should be generalizable. A loss to follow-up of 40 % is quite high. The only detectable difference between non-respondents and respondents was age (38, SD 11.3 vs 45, SD 10.7 years). Thus we conclude the results might not be generalizable to younger workers. Another limitation is that we did not collect data on therapeutic recommendations and whether they were followed. Such data would have helped us to interpret the pattern of changes between the baseline and follow-up assessment and will therefore be collected and analysed in our next investigation. Finally, the sample was drawn from a single region of Germany rather than from across Europe and so the findings should only be generalised to other countries with different healthcare systems with caution. Conclusions We conclude that the PSIW intervention is effective over a 3 month-follow-up period, covering diagnostic assessment and the very early psychotherapeutic sessions. The effects of the intervention were most marked with respect to social functioning (work ability) and mental health in patients with sub-syndromal illness and CMD. To make the intervention more effective in this setting than it is in its current form it would need to be tailored even more to the specific needs of the target group. Exhaustion is one of the major topics mentioned as reason for utilising PSIW (qualitative data in preparation for publication). Ongoing interpersonal conflicts, problems regulating demands and efforts, or problems in prioritising are likely drivers for exhaustion on the individual side. On the organisational side working conditions may contribute to this problem. A short time psychotherapeutic intervention (up to 10 sessions) should be offered to detect relevant areas of conflict and to consequently analyse and address specific individual and/or organisational reasons in more detail. Together with consecutive motivation for further clinical treatment if necessary this approach should make a short-time intervention even more powerful. Showing that the intervention is effective in the new setting workplace while reaching different users especially due to gender and course of disease drafts the need for in depth information about: which ingredients, barriers, drivers and synergies can be identified in that kind of health care offer in the workplace? We are currently analysing qualitative data which address these questions. Abbreviations ANCOVACovariance analysis with repeated measures AWMFAssociation of the scientific medical societies in Germany CCOCollaborative care offers CMDCommon mental disorder DPDepersonalisation EEEmotional exhaustion GPGeneral practitioner ISIrritation scale ITTIntention to treat LOCFLast observation carried forward LPALatent profile analysis MMean MBIMaslach burnout inventory NNumber n.s.Not significant PADepersonalisation PHQ-7/-9/-15Patient health questionnaire 7 items generalised anxiety/9 items depression/15 items somatisation PSIWPsychotherapeutic consultation in the workplace PSOCPsychotherapeutic outpatient care SDStandard deviation SF-12 mcs/pcsShort form-12 Items mental health component score/physical health component score WAWork ability WAIWork ability index Acknowledgements We thank all participating patients, Anita Gutbrod, Petra Beschoner and cooperating occupational consultants, Daniel Mauss, Marc Krueger, Bernhard Reiter, Jutta-Mueller Nuebling and Andreas Woerner for patient recruitment, Gabi Kreutzer-Stierle and Firuzan Sari-Kundt for data management. Funding The study was supported by the young scientists programme run by the German network “Health Services Research Baden-Wuerttemberg” of the Ministry for Science, Research and Arts in collaboration with the Ministry for Work and Social Affairs, Family, Women, and Senior Citizens, Baden-Wuerttemberg and by the company Airbus. The work of the Institute of Occupational and Social Medicine and Health Services Research Tuebingen is supported by an unrestricted grant from the employers’ association for the metal and electric industry in Baden-Wuerttemberg (Suedwestmetall). Availability of data and materials The datasets generated and analysed during the current study are available from the corresponding author on reasonable request after consultation with the cooperating partners of the study. Authors’ contributions EVR, RK, MH, DM, MAR, and HG conceived and designed the study. EVR, EDR, HG, and RK were responsible for the analysis and interpretation of the data. EVR, EDR, MH and DM worked on the acquisition of the data. EVR and HG obtained funding and are the study guarantors. EVR, HG, and RK drafted the manuscript. MAR, EDR, DM, and MH critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript. Competing interests EVR is senior physician and HG and MH director of a clinic that offers the service of PSIW to interested organisations and companies. Consent for publication not applicable. Ethics approval and consent to participate The Medical Ethical Board of the University Medical Centre of the University of Ulm gave approval for the study (No224/11). All participants gave written consent to participate. ==== Refs References 1. Bloom DE, Cafiero E, Jané-Llopis E, Abrahams-Gessel S, Bloom L, Fathima S, Feigl AB, Gaziano T, Mowafi M, Pandya A, Prettner K, Rosenberg L, Seligman B, Stein AZ, Weinstein C: The Global Economic Burden of Noncommunicable Diseases. Geneva: World Economic Forum; 2012. 2. OECD Fit mind, fit job: from evidence to practice in mental health and work, mental health and work 2015 Paris OECD Publishing 3. Bertilsson M Vaez M Waern M Ahlborg G Hensing G A prospective study on self-assessed mental well-being and work capacity as determinants of all-cause sickness absence J Occup Rehabil 2015 25 52 64 10.1007/s10926-014-9518-5 24898192 4. Cuijpers P Vogelzangs N Twisk J Kleiboer A Li J Penninx BW Differential mortality rates in major and subthreshold depression: Meta-analysis of studies that measured both Br J Psychiatry 2013 202 22 7 10.1192/bjp.bp.112.112169 23284149 5. Russ TC Stamatakis E Hamer M Starr JM Kivimaki M Batty GD Association between psychological distress and mortality: individual participant pooled analysis of 10 prospective cohort studies Br Med J 2012 345 e4933 10.1136/bmj.e4933 22849956 6. Jain G Roy A Harikrishnan V Yu S Dabbous O Lawrence C Patient-reported depression severity measured by the PHQ-9 and impact on work productivity: results from a survey of full-time employees in the United States J Occup Environ Med 2013 55 252 8 10.1097/JOM.0b013e31828349c9 23439268 7. Vos T Barber RM Bell B Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013 Lancet 2015 386 743 800 10.1016/S0140-6736(15)60692-4 26063472 8. Cuijpers P Sijbrandij M Koole SL Andersson G Beekman AT Reynolds CF Adding psychotherapy to antidepressant medication in depression and anxiety disorders: a meta-analysis World Psychiatry 2014 13 56 67 10.1002/wps.20089 24497254 9. Huhn M Tardy M Spineli LM Kissling W Förstl H Pitschel-Walz G Leucht C Samara M Dold M Davis JM Leucht S Efficacy of pharmacotherapy and psychotherapy for adult psychiatric disorders a systematic overview of meta-analyses JAMA Psychiatry 2014 71 706 15 10.1001/jamapsychiatry.2014.112 24789675 10. DGPPN, BÄK, KBV, AWMF, AkdÄ, BPtK, BApK, DAGSHG, DEGAM, DGPM, DGPs, DGRW. Unipolare Depression. S3-Leitlinie/Nationale Versorgungsleitlinie Unipolare Depression - Langfassung. last updat. DGPPN BÄK KBV AWMF AkdÄ BPtK BApK DAGSHG DEGAM DGPM DGPs DGRW; 2009. Available from: www.depression.versorgungsleitlinien.de; [cited: 23.08.2016]; doi:10.6101/AZQ/000239. 11. OECD Making mental health count: the social and economic costs of neglecting mental health care OECD health policy studies 2014 Paris OECD Publishing 12. Chisholm D Sweeny K Sheehan P Rasmussen B Smit F Cuijpers P Saxena S Scaling-up treatment of depression and anxiety: a global return on investment analysis Lancet Psychiatry 2016 0366 1 10 13. OECD. Mental Health and Work: Denmark. Paris: OECD Publishing; 2013. 14. OECD. Mental Health and Work: Belgium. Paris: OECD Publishing; 2013. 15. OECD. Mental Health and Work: Netherlands. Paris: OECD Publishing; 2014. 16. Schulz H Barghaan D Harfst T Koch U Psychotherapeutische Versorgung [Psychotherapeutic healthcare] Gesundheitsberichterstattung des Bundes 2008 41 1 47 17. Kruse J, Herzog W. Zur ambulanten psychosomatischen/psychotherapeutischen Versorgung in der kassenärztlichen Versorgung in Deutschland - Formen der Versorgung und ihre Effizienz - Zwischenbericht 2012 [About psychosomatic and psychotherapeutic outpatient care in Germany - e. Berlin: Kassenärztliche Bundesvereinigung (KBV); 2012. 18. Munk-Jørgensen P Allgulander C Dahl AA Foldager L Holm M Rasmussen I Virta A Huuhtanen M-T Wittchen H-U Prevalence of generalized anxiety disorder in general practice in Denmark, Finland, Norway, and Sweden Psychiatr Serv 2006 57 1738 44 10.1176/ps.2006.57.12.1738 17158488 19. Clement S Schauman O Graham T Maggioni F Evans-Lacko S Bezborodovs N Morgan C Rüsch N Brown JSL Thornicroft G What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies Psychol Med 2015 45 11 27 10.1017/S0033291714000129 24569086 20. OECD Sick on the job?: myths and realities about mental health and work 2012 Paris OECD Publishing 21. Gilbody S Bower P Rick J Better care for depression in the workplace: Integrating occupational and mental health services Br J Psychiatry 2012 200 442 3 10.1192/bjp.bp.111.103598 22661675 22. Leka S Cox T The European Framework for Psychosocial Risk Management : PRIMA-EF 2008 Nottingham Institute of Work, Health and Organisation 23. Nieuwenhuijsen K Faber B Verbeek JH Neumeyer-Gromen A Hees HL Verhoeven AC van der Feltz-Cornelis CM Bültmann U Interventions to improve return to work in depressed people Cochrane Database Syst Rev 2014 12 CD006237 25470301 24. van Vilsteren M van Oostrom SH de Vet HC Franche R-L Boot CR Anema JR Workplace interventions to prevent work disability in workers on sick leave Cochrane Database Syst Rev 2015 10 CD006955 26436959 25. Pomaki G Franche R-L Murray E Khushrushahi N Lampinen TM Workplace-based work disability prevention interventions for workers with common mental health conditions: a review of the literature J Occup Rehabil 2012 22 182 95 10.1007/s10926-011-9338-9 22038297 26. Brouwers EPM Terluin B Tiemens BG Verhaak PFM Predicting return to work in employees sick-listed due to minor mental disorders J Occup Rehabil 2009 19 323 32 10.1007/s10926-009-9198-8 19760489 27. Ekberg K Wåhlin C Persson J Bernfort L Öberg B Early and late return to work after sick leave: predictors in a cohort of sick-listed individuals with common mental disorders J Occup Rehabil 2015 25 1 11 10.1007/s10926-015-9570-9 25604212 28. Brohan E Evans-Lacko S Henderson C Murray J Slade M Thornicroft G Disclosure of a mental health problem in the employment context: qualitative study of beliefs and experiences Epidemiol Psychiatr Sci 2014 23 289 300 10.1017/S2045796013000310 23919992 29. Burman-Roy S Butterworth M Madan I Henderson M Harvey SB Which patients are seen by an occupational psychiatry service? Occup Med (Lond) 2013 63 507 9 10.1093/occmed/kqt103 23990507 30. Mayer D Schmidt H Hoelzer M “Psychosomatische Sprechstunde “und “Psychosomatische Grundversorgung in der Arbeitsmedizin”. [“Psychosomatic consultation” and “Basic Psychosomatic care in occupational medicine”.] Arbeitsmed Sozialmed Umweltmed 2010 45 593 7 31. Rothermund E Gündel H Kilian R Hölzer M Reiter B Mauss D Rieger MA Müller-Nübling J Wörner A von Wietersheim J Beschoner P Behandlung psychosomatischer Beschwerden im Arbeitskontext - Konzept und erste Daten. [Treatment of psychosomatic disorders occurring in the vocational context – concept and first findings] Z Psychosom Med Psychother 2014 60 177 89 24877574 32. Preiser C, Rothermund E, Wittich A, Gündel H, Rieger MA. Psychosomatic consultation in the workplace - opportunities and limitations of the services offered. Results of a qualitative study. Int. Arch. Occup. Environ. Health. 2016;89:599-608. 33. Rothermund E Kilian R Hoelzer M Mayer D Mauss D Krueger M Rieger MA Guendel H “Psychosomatic consultation in the workplace” -- a new model of care at the interface of company-supported mental health care and consultation-liaison psychosomatics: design of a mixed methods implementation study BMC Public Health 2012 12 780 10.1186/1471-2458-12-780 22974257 34. Diefenbacher A Psychiatry, psychosomatic medicine and the general hospital in Germany World Psychiatry 2003 2 95 7 16946905 35. Kruse J Larisch A Hofmann M Herzog W Ambulante psychosomatische und psychotherapeutische Versorgung in Deutschland - Versorgungsprofile abgebildet durch Daten der Kassenärztlichen Bundesvereinigung (KBV) [Psychosomatic and psychotherapeutic outpatient care in Germany – Capacity profile based Z Psychosom Med Psychother 2013 59 254 72 24085478 36. Alonso J, Angermeyer MC, Bernert S, Bruffaerts R, Brugha TS, Bryson H, de Girolamo G, Graaf R, Demyttenaere K, Gasquet I, Haro JM, Katz SJ, Kessler RC, Kovess V, Lepine JP, Ormel J, Polidori G, Russo LJ, Vilagut G, Almansa J, Arbabzadeh-Bouchez S, Autonell J, Bernal M, Buist-Bouwman MA, Codony M, Domingo-Salvany A, Ferrer M, Joo SS, Martinez-Alonso M, Matschinger H, Mazzi F, Morgan Z, Morosini P, Palacin C, Romera B, Taub N, Vollebergh WAM. Prevalence of mental disorders in Europe: results from the European Study of the Epidemiology of Mental Disorders (ESEMeD) project. Acta Psychiatr. Scand. Suppl. 2004;109:21–27. 37. Kessler RC Üstün B The WHO world mental health surveys. Global perspectives of mental health surveys. first 2008 New York Cambridge University Press 38. Mack S Jacobi F Gerschler A Strehle J Höfler M Busch MA Maske UE Hapke U Seiffert I Gaebel W Zielasek J Maier W Wittchen H-U Self-reported utilization of mental health services in the adult German population – evidence for unmet needs ? Results of the DEGS1-Mental Health Module (DEGS1-MH) Int J Methods Psychiatr Res 2014 23 289 303 10.1002/mpr.1438 24687693 39. Tuomi K Ilmarinen J Jahkola A Katajarinne L Tulkki A Work ability index 1998 2 Helsinki Finnish Institute of Occupational Health 40. Kroenke K Spitzer RL Williams JBW The PHQ-9: validity of a brief depression severity measure J Gen Intern Med 2001 16 606 13 10.1046/j.1525-1497.2001.016009606.x 11556941 41. Spitzer RL Kroenke K Williams JBW Löwe B A brief measure for assessing generalized anxiety disorder: The gad-7 Arch Intern Med 2006 166 1092 7 10.1001/archinte.166.10.1092 16717171 42. Kroenke K Spitzer RL Williams JBW The PHQ-15: validity of a new measure for evaluating the severity of somatic symptoms Psychosom Med 2002 64 258 66 10.1097/00006842-200203000-00008 11914441 43. Kroenke K Spitzer RL Williams JBW Löwe B The patient health questionnaire somatic, anxiety, and depressive symptom scales: a systematic review Gen Hosp Psychiatry 2010 32 345 59 10.1016/j.genhosppsych.2010.03.006 20633738 44. Morfeld M Bullinger M Kirchberger I SF-36 Fragebogen zum Gesundheitszustand [Short form-36 health survey questionnaire] 2011 2 Goettingen Hogrefe 45. Maslach L Jackson S Leiter M Maslach Burnout inventory general version 1996 3 Palo Alto Consulting Psychologists Press 46. Mohr G Rigotti T Müller A Irritation - ein Instrument zur Erfassung psychischer Beanspruchung im Arbeitskontext. Skalen- und Itemparameter aus 15 Studien [Irritation - an instrument assessing mental strain in working contexts. Scale and item parameters from 15 studies] Zeitschrift für Arbeits- und Organ 2005 49 44 8 10.1026/0932-4089.49.1.44 47. Mohr G Müller A Rigotti T Normwerte der Skala Irritation: Zwei Dimensionen psychischer Beanspruchung [Standardisation data of the Irritation Scale. Two dimensions of mental strain] Diagnostica 2005 51 12 20 10.1026/0012-1924.51.1.12 48. Austin PC An introduction to propensity score methods for reducing the effects of confounding in observational studies Multivariate Behav Res 2011 46 399 424 10.1080/00273171.2011.568786 21818162 49. Lanza ST Rhoades BL Latent class analysis: an alternative perspective on subgroup analysis in prevention and treatment Prev Sci 2013 14 157 68 10.1007/s11121-011-0201-1 21318625 50. Buist-Bouwman MA Ormel J De Graaf R Vollebergh W Functioning after a major depressive episode: complete or incomplete recovery? J Affect Disord 2004 82 363 71 15555687 51. Pachoud B Plagnol A Leplege A Outcome, recovery and return to work in severe mental illnesses Disabil Rehabil 2010 32 1043 50 10.3109/09638281003775618 20450409 52. Lagerveld SE Blonk RWB Brenninkmeijer V Wijngaards-de Meij L Schaufeli WB Work-focused treatment of common mental disorders and return to work: a comparative outcome study J Occup Health Psychol 2012 17 220 34 10.1037/a0027049 22308965 53. Dell’Osso B Glick ID Baldwin DS Altamura AC Can long-term outcomes be improved by shortening the duration of untreated illness in psychiatric disorders? A conceptual framework Psychopathology 2013 46 14 21 10.1159/000338608 22890286 54. Cosci F Fava GA Staging of mental disorders: systematic review Psychother Psychosom 2013 82 20 34 10.1159/000342243 23147126 55. Beck A Crain L Solberg L Unützer J Glasgow R Maciosek M Whitebird RR Does severity of depression predict magnitude of productivity loss? Am J Manag Care 2014 20 e294 301 25295792 56. Vlasveld MC van der Feltz-Cornelis CM Adèr HJ Anema JR Hoedeman R van Mechelen W Beekman ATF Collaborative care for major depressive disorder in an occupational healthcare setting Br J Psychiatry 2012 200 510 1 10.1192/bjp.bp.111.095687 22116977 57. van der Feltz-Cornelis CM Hoedeman R de Jong FJ Meeuwissen JA Drewes HW van der Laan NC Adèr HJ Faster return to work after psychiatric consultation for sicklisted employees with common mental disorders compared to care as usual. A randomized clinical trial Neuropsychiatr Dis Treat 2010 6 375 85 10.2147/NDT.S11832 20856601 58. Blonk RW Brenninkmeijer V Lagerveld SE Houtman IL Return to work: a comparison of two cognitive behavioural interventions in cases of work-related psychological complaints among the self-employed Work Stress 2006 20 129 44 10.1080/02678370600856615 59. Imamura K Kawakami N Furukawa TA Matsuyama Y Shimazu A Umanodan R Kawakami S Kasai K Does Internet-based cognitive behavioral therapy (iCBT) prevent major depressive episode for workers? A 12-month follow-up of a randomized controlled trial Psychol Med 2015 45 1 11 10.1017/S0033291714003006 25065902 60. Lambert M Lambert M The efficacy and effectiveness of psychotherapy Bergin Garfield’s Handb. Psychother. Behav. Chang 2013 6 New York Wiley 169 218 61. Freemantle N Marston L Walters K Wood J Reynolds MR Petersen I Making inferences on treatment effects from real world data: propensity scores, confounding by indication, and other perils for the unwary in observational research BMJ 2013 347 f6409 10.1136/bmj.f6409 24217206 62. Tschuschke V Freyberger HJ Zur aktuellen Situation der Psychotherapiewissenschaft und ihrer Auswirkungen – eine kritische Analyse der Lage.[On the present situation in psychotherapy and its implications – A critical analysis of the facts] Z Psychosom Med Psychother 2015 61 122 38 26175169 63. Cuijpers P Cristea IA How to prove that your therapy is effective, even when it is not: a guideline Epidemiol Psychiatr Sci 2015 28 1 8 10.1017/S2045796015000864 26411384 64. Hasselhorn HM Freude G Der Work Ability Index – ein Leitfaden [The Work Ability Index – a practical guide] 2007 Dortmund/Berlin/Dresden Bundesanstalt für Arbeitsschutz und Arbeitsmedizin 65. Cieza A Ewert T Üstün B Chatterji S Kostanjsek N Stucki G Development of ICF Core Sets for patients with chronic conditions J Rehabil Med 2004 36 9 11 10.1080/16501960410015353 15370742

PMC005xxxxxx/PMC5002129.txt

==== Front Malar JMalar. JMalaria Journal1475-2875BioMed Central London 148910.1186/s12936-016-1489-xResearchReductions in malaria in pregnancy and adverse birth outcomes following indoor residual spraying of insecticide in Uganda Muhindo Mary K. marymkakuru@gmail.com 1Kakuru Abel akakuru@idrc-uganda.org 1Natureeba Paul pnature@gmail.com 1Awori Patricia patriciaawori@yahoo.com 1Olwoch Peter polwoch@idrc-uganda.org 1Ategeka John ategekajohnnie@yahoo.com 1Nayebare Patience pnayebare@ymail.com 1Clark Tamara D. Tamara.Clark@ucsf.edu 2Muehlenbachs Atis vkd6@cdc.gov 3Roh Michelle Michelle.Roh@ucsf.edu 2Mpeka Betty Betty_Mpeka@ugandairs.com 4Greenhouse Bryan bryan.greenhouse@ucsf.edu 2Havlir Diane V. Diane.Havlir@ucsf.edu 2Kamya Moses R. mkamya@infocom.co.ug 5Dorsey Grant grant.dorsey@ucsf.edu 2http://orcid.org/0000-0001-6305-758XJagannathan Prasanna pras.jagannathan@ucsf.edu 21 Infectious Diseases Research Collaboration, Kampala, Uganda 2 Department of Medicine, University of California, San Francisco, USA 3 Centers for Disease Control and Prevention, Atlanta, Georgia 4 Uganda Indoor Residual Spraying Phase II Project, Abt Associates, Inc, Kampala, Uganda 5 Department of Medicine, Makerere University College of Health Sciences, Kampala, Uganda 26 8 2016 26 8 2016 2016 15 1 4374 8 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 Indoor residual spraying of insecticide (IRS) is a key intervention for reducing the burden of malaria in Africa. However, data on the impact of IRS on malaria in pregnancy and birth outcomes is limited. Methods An observational study was conducted within a trial of intermittent preventive therapy during pregnancy in Tororo, Uganda. Women were enrolled at 12–20 weeks of gestation between June and October 2014, provided with insecticide-treated bed nets, and followed through delivery. From December 2014 to February 2015, carbamate-containing IRS was implemented in Tororo district for the first time. Exact spray dates were collected for each household. The exposure of interest was the proportion of time during a woman’s pregnancy under protection of IRS, with three categories of protection defined: no IRS protection, >0–20 % IRS protection, and 20–43 % IRS protection. Outcomes assessed included malaria incidence and parasite prevalence during pregnancy, placental malaria, low birth weight (LBW), pre-term delivery, and fetal/neonatal deaths. Results Of 289 women followed, 134 had no IRS protection during pregnancy, 90 had >0–20 % IRS protection, and 65 had >20–43 % protection. During pregnancy, malaria incidence (0.49 vs 0.10 episodes ppy, P = 0.02) and parasite prevalence (20.0 vs 8.9 %, P < 0.001) were both significantly lower after IRS. At the time of delivery, the prevalence of placental parasitaemia was significantly higher in women with no IRS protection (16.8 %) compared to women with 0–20 % (1.1 %, P = 0.001) or >20–43 % IRS protection (1.6 %, P = 0.006). Compared to women with no IRS protection, those with >20–43 % IRS protection had a lower risk of LBW (20.9 vs 3.1 %, P = 0.002), pre-term birth (17.2 vs 1.5 %, P = 0.006), and fetal/neonatal deaths (7.5 vs 0 %, P = 0.03). Conclusion In this setting, IRS was temporally associated with lower malaria parasite prevalence during pregnancy and at delivery, and improved birth outcomes. IRS may represent an important tool for combating malaria in pregnancy and for improving birth outcomes in malaria-endemic settings. Trial Registration Current Controlled Trials Identifier NCT02163447 Electronic supplementary material The online version of this article (doi:10.1186/s12936-016-1489-x) contains supplementary material, which is available to authorized users. Keywords Malaria in pregnancyPlacental malariaPlasmodium falciparumIndoor residual sprayingVector-borne diseasehttp://dx.doi.org/10.13039/100000071National Institute of Child Health and Human DevelopmentPO1 HD059454Havlir Diane V. http://dx.doi.org/10.13039/100000060National Institute of Allergy and Infectious DiseasesU19 AI089674K23AI100949Dorsey Grant Jagannathan Prasanna issue-copyright-statement© The Author(s) 2016 ==== Body Background In sub-Saharan Africa, over 30 million pregnancies occur annually in areas where malaria is endemic, and each year malaria in pregnancy is estimated to cause nearly one million low birth weight (LBW) deliveries and up to 100,000 infant deaths [1–3]. Given this high burden of disease, the World Health Organization (WHO) recommends the implementation of malaria preventive measures in all African countries where Plasmodium falciparum remains endemic, including the use of long-lasting, insecticide-treated nets (LLINs) and intermittent preventive treatment during pregnancy (IPTp) with sulfadoxine-pyrimethamine (SP) [4]. Despite these measures, rates of placental malaria and poor birth outcomes remain persistently high in many parts of Africa. In a recent cross-sectional study performed in Tororo, a highly endemic district in Uganda where nearly 70 % of pregnant women reported using a LLIN, the prevalence of placental malaria was 62 % in women who had received ≥two doses of IPTp-SP, and nearly 10 % of children were born with LBW [5]. This and other studies have raised concern for waning efficacy of LLINs given the emergence of resistance to pyrethroid insecticides [6–8], as well as waning efficacy of SP given widespread prevalence of drug-resistant parasites [9]. New interventions to prevent malaria during pregnancy and improve birth outcomes are urgently needed. Indoor residual spraying (IRS) has been shown to be very effective in reducing vector densities, parasite prevalence and malaria morbidity in sub-Saharan Africa [10–14]. IRS can provide significant added protection together with LLINs compared to LLINs alone, particularly in areas with significant pyrethroid resistance [15–18], although evidence has been mixed [19, 20]. In 2006, Uganda began using IRS in selected districts, initially in the epidemic-prone areas of southwestern Uganda but later in the highly endemic areas of northern Uganda, with significant reductions in malaria morbidity and slide positivity rates [10, 14]. However, data quantifying the impact of IRS on malaria in pregnancy and birth outcomes are lacking. Recently, a clinical trial was conducted comparing IPTp with dihydroartemisinin-piperaquine (DP) with IPTp-SP among HIV-uninfected pregnant women. This study was conducted in Tororo, Uganda, where, in separate studies, the East African International Centres of Excellence in Malaria Research (ICEMR) have been conducting cohort studies and entomology surveys since 2011 [21]. IPTp-DP was well tolerated, and significantly reduced parasite prevalence during pregnancy and the risk of placental malaria [22]. Midway through this trial, the President’s Malaria Initiative (PMI), through its implementing partner Abt Associates, initiated IRS in Tororo district for the first time [23]. Measures of malaria during pregnancy were compared before and after the implementation of IRS, and associations between IRS exposure during pregnancy and adverse birth outcomes were evaluated, among participants enrolled in this trial. Methods Study site and participants Tororo is a rural district in southeastern Uganda with an entomologic inoculation rate estimated at 310 infectious bites per person year in 2012 [21]. Following a universal LLIN distribution campaign in November 2013 [23], 95 % of household in the region reported owning at least one LLIN [24]. Between June and October 2014, 300 women were enrolled into a double-blinded, placebo-controlled trial of three-dose SP vs three-dose DP vs monthly DP for IPTp. Details of the parent study have been described elsewhere [22]. Briefly, participants were HIV-uninfected pregnant women at least 16 years of age of all gravidities with an estimated gestational age of 12–20 weeks confirmed by ultrasound. The sub-study described in this report includes all women followed through delivery (n = 289). From December 2014 to February 2015, Abt, in cooperation with the Ugandan Ministry of Health, initiated IRS in Tororo district using bendiocarb wettable powder, a carbamate insecticide [23]. Homes were sprayed once and following spraying, houses were marked with the date of spraying. Home visitors obtained the exact date of spraying from each participant’s household. If a household was not sprayed, the reasons for a lack of spraying were obtained and if the surrounding households in the village were sprayed, the date of spraying was obtained from surrounding homes. Study procedures and follow-up At enrolment, women underwent a standardized examination and received a LLIN. Study participants were randomized in a 1:1:1 ratio to three-dose SP vs three-dose DP vs monthly DP for IPTp, as previously described [22]. Women received all of their medical care at a study clinic open daily. Routine visits were conducted every 4 weeks, including collection of dried blood spots (DBS). Women were encouraged to come to the clinic any time they were ill. Those who presented with a documented fever (tympanic temperature >38.0 °C) or history of fever in the previous 24 h had blood collected for a thick blood smear. If the smear was positive, the patient was diagnosed with malaria and treated with artemether-lumefantrine. Women were encouraged to deliver at the hospital adjacent to the study clinic. Women delivering at home were visited by study staff at the time of delivery or as soon as possible afterwards. At delivery a standardized assessment was completed, including evaluation for birth weight and collection of specimens, including placental tissue and DBS of placental blood. Laboratory methods DBS were tested for the presence of malaria parasites using loop-mediated isothermal amplification (LAMP) [25]. Placental tissues were processed for histological evidence of placental malaria as previously described [22]. Histopathology slides were read in duplicate using a standardized case record form by two independent readers and any discrepant results resolved by a third reader. Blood smears were collected in febrile women during pregnancy, stained with 2 % Giemsa and read by laboratory technologists. A blood smear was considered negative when the examination of 100 high power fields did not reveal asexual parasites. All slides were read by a second microscopist and a third reviewer settled any discrepant readings. Entomological surveys Entomological surveys were conducted monthly beginning in June 2011 from 100 households enrolled in a separate longitudinal cohort study in Nagongera sub-county, Tororo district, as part of the ICEMR programme in Uganda [21]. Each month, miniature CDC light traps (Model 512; John W Hock Company, Gainesville, FL, USA) were positioned with the light 1 m above the floor at the foot end of the bed where a cohort study participant slept. Traps were set at 19.00 and collected at 07.00 h the following morning by field workers. Primary exposure variable A woman was considered ‘protected under IRS’ if her house was directly sprayed, or if the surrounding village was sprayed, 14 days after spraying to account for the average incubation period for P. falciparum. The duration of pregnancy under protection of IRS was calculated as the date of delivery—(date of IRS + 14 days). The total duration of pregnancy was calculated as the date of delivery—estimated date of conception (ultrasound-confirmed). The proportion of time during a woman’s pregnancy that was under protection of IRS was then calculated as the duration of pregnancy under IRS/total duration of pregnancy. Three categories of protection were defined: no protection, >0–20 % of pregnancy protected, and >20–43 % of pregnancy protected. Outcomes Outcomes assessed during pregnancy included parasite prevalence by LAMP and the incidence of malaria, calculated as the number of episodes per person years (ppy) at risk. Outcomes assessed at birth included the prevalence of parasitaemia at delivery by LAMP and evidence of placental malaria (parasites or pigment) by histopathology. Placental histopathology was also classified according to whether moderate-high grade pigment deposition was present (defined as pigment detected in >5 % of high power 40× fields) [26]. Birth outcomes assessed included birth weight, LBW (<2500 g), pre-term delivery (<37 weeks), and fetal/neonatal deaths, including spontaneous abortion, stillbirth and neonatal death within 4 weeks of delivery. For women giving birth to twins, delivery outcomes were based on whether the outcome was present in either child/placenta. Statistical methods Data were double entered into an Access database. Data analysis was done using Stata version 14 (Stata Corp, College Station, TX, USA). Baseline characteristics between groups were compared using the χ2 test. Repeated prevalence measures and the daily risk of malaria during pregnancy were compared using generalized estimating equations with a log-binomial family, after adjustment for gravidity, age, IPTp arm, and gestational age when study drugs were started, and presented as adjusted risk ratios (aRR). For results stratified by IPTp arm, women randomized to either IPTp-DP arm were considered together. Dichotomous outcomes at delivery were compared using multivariate logistic regression. Continuous outcomes at delivery were compared using multivariate linear regression, with adjustment as above. As a secondary sensitivity analysis, propensity scores and inverse probability weighting were used to assess the average causal effect of IRS exposure with outcomes at delivery. All P-values were two-sided and values <0.05 considered statistically significant. Results Characteristics of study site and participants Between June and October 2014, 300 women were enrolled prior to IRS, and 289 women were followed through delivery from October 2014 to May 2015 (Fig. 1). From December 2014 to February 2015, 123,924 of 145,574 households (85.1 %) in Tororo district received IRS. Among women in the study cohort, 95 % of households were sprayed between 1 December, 2014 and 31 January, 2015, with the remaining sprayed in February 2015. Monthly measurements of female Anopheles mosquitoes collected per household per night in Tororo were lower from February to May 2015, after IRS, compared with February–May 2014 (5.4 vs 33.7 female Anopheles mosquitoes per house per night, P < 0.001) (Fig. 1).Fig. 1 Temporal trends with enrolment, delivery and entomological measures. Shown are the number of women enrolled (blue bars) and number of women delivered (green bars) each month between 2014 and 2015. The red line (2013–2014) and yellow line (2014–2015) represent the number of female anopheline mosquitoes collected per household per night (right y axis) in Tororo district from 100 randomly selected households enrolled in the International Centers of Excellence in Malaria Research cohort [21] Of 289 women followed through delivery, 155 (53.6 %) were exposed to IRS before the birth of their child; 137 of these women’s households were directly sprayed, and 18 households (11.6 %) were indirectly protected through spraying of the surrounding village. Of the 134 women not exposed to IRS during pregnancy, six lived outside of the IRS spraying area in neighbouring Busia district, and 128 were exposed to IRS after the birth of their child. The median duration of pregnancy under protection of IRS was 46 days (IQR 20–68 days), and women were stratified into three groups based on the proportion of pregnancy under the protection of IRS: 0 (n = 134), >0–20 % (n = 90), and >20–43 % (n = 65). At enrolment, the prevalence of malaria parasites was >55 % and similar between the three groups (Table 1). Women with no IRS protection during pregnancy were slightly younger, more likely primigravid, and more likely to have study drugs initiated at 20 weeks gestation vs 16 weeks gestation, than women with IRS protection during pregnancy. IPTp treatment assignments, maternal weight gain, and household wealth were similar between groups (Table 1).Table 1 Characteristics of women stratified by duration of pregnancy under the protection of IRS Characteristic Proportion of pregnancy under the protection of IRS None (n = 134) >0–20 % (n = 90) >20–43 % (n = 65) Malaria parasites detected at enrolment, n (%) 76 (56.7 %) 52 (57.8 %) 39 (60.0 %) Age at enrolment in years, mean (SD)a 21.1 (3.9) 22.9 (4.0) 22.6 (4.0) Gravidity, n (%)a  1 64 (47.8 %) 23 (25.6 %) 16 (24.6 %)  2 37 (27.6 %) 25 (27.8 %) 25 (38.5 %)  >3 33 (24.6 %) 42 (46.7 %) 24 (36.9 %) Study drugs started at 16 vs 20 weeks GA, n (%)a 61 (45.5 %) 85 (94.4 %) 57 (87.7 %) Assigned IPTp arm, n (%)  3-dose SP 45 (33.6 %) 34 (37.8 %) 23 (35.4 %)  3-dose DP 41 (30.6 %) 26 (28.9 %) 22 (33.9 %)  Monthly DP 48 (35.8 %) 30 (33.3 %) 20 (30.8 %) Maternal weight gain during pregnancy, mean kg/week (SD) 0.25 (0.15) 0.25 (0.12) 0.24 (0.15) Household wealth index, n (%)  Lowest tertile 45 (33.6 %) 34 (37.8 %) 20 (30.8 %)  Middle tertile 45 (33.6 %) 23 (25.6 %) 27 (41.5 %)  Highest tertile 44 (32.8 %) 33 (36.7 %) 18 (27.7 %) aP < 0.001 Malaria incidence and parasite prevalence during pregnancy, before and after IRS Parasite prevalence at enrolment and before study drug initiation was stable and high prior to IRS (see Additional file 1). After study drug initiation, the incidence of malaria during pregnancy was significantly higher before IRS (0.49 episodes ppy) compared to after IRS (0.10 episodes ppy, aRR 0.20, P = 0.02) after adjustment for gravidity, age, IPTp arm, and gestational age when study drugs were started. Similarly, the prevalence of malaria parasites by LAMP during pregnancy was significantly higher before IRS (20.0 %) compared to after IRS (8.9 %, aRR 0.40, P < 0.001). Declines were observed in women randomized to both IPTp-SP and IPTp-DP, although there was evidence of interaction between IPTp and IRS (P = 0.07, Fig. 2). Among women randomized to IPTp-SP, the prevalence of parasitaemia was 44.0 % before IRS, before declining to 22.0 % after IRS (aRR 0.49, P = 0.003). Among women randomized to IPTp-DP, parasite prevalence was 8.7 % before IRS vs 1.1 % after IRS (aRR 0.13, P = 0.003). Women living in households directly sprayed vs those indirectly sprayed were both protected by IRS. Among women living in households directly sprayed, parasite prevalence was 20.2 % before IRS vs 8.5 % after IRS (aRR 0.42, 95 % CI 0.26–0.70, P = 0.001). Among women living in households indirectly sprayed, parasite prevalence was 34.2 % before IRS vs 5.0 % after IRS (aRR 0.12, 95 % CI 0.03–0.59, P = 0.009), consistent with community-wide benefits of IRS on parasitaemia [11].Fig. 2 Parasite prevalence during pregnancy by calendar month before, during and after IRS. Shown is the predicted probability of having a positive (+) LAMP result during pregnancy, stratified by IPTp arm, after enrolment and initiation of study drugs, over the period of time of the study. Parasite prevalence point estimates and standard errors obtained using generalized estimating equations after adjustment for gravidity, age, and gestational age when study drugs were started. Grey-shaded bar shows the time period of IRS in Tororo district. SP (black): sulfadoxine-pyrimethamine; DP (blue): dihydroartemisinin-piperaquine Association between IRS and placental malaria At delivery, the prevalence of malaria parasites in placental blood by LAMP was significantly higher in women with no IRS protection (16.8 %) compared to women with >0–20 % (1.1 %, P = 0.001) or >20–43 % IRS protection during pregnancy (1.6 %, P = 0.006) (Table 2). The prevalence of any placental malaria by histopathology, as well as the prevalence of moderate to high-grade pigment deposition, was higher in women with no IRS protection compared with women with IRS protection (Table 2), but the differences were not statistically significant. All 105 placentas that were positive for placental malaria by histopathology had pigment in fibrin that was indicative of past infection, and seven (11 %) had parasites indicative of concomitant active infection. All seven with active infection were not exposed to IRS prior to delivery.Table 2 Associations between proportion of pregnancy under the protection of IRS and outcomes measured at birth Outcome Proportion of pregnancy under the protection of IRS None >0–20 % >20–43 % Placental blood positive for malaria parasites by LAMPa  Risk 22/131 (16.8 %) 1/88 (1.1 %) 1/61 (1.6 %)  aORb (95 % CI) Reference group 0.03 (0–0.25) 0.05 (0.01–0.41)  P value 0.001 0.006 Any evidence of placental malaria by histopathologya  Risk 63/132 (47.7 %) 25/88 (28.4 %) 17/62 (27.4 %)  aORb (95 % CI) Reference group 0.77 (0.35–1.69) 0.63 (0.27–1.48)  P value 0.51 0.29 Moderate-high grade pigment deposition by histopathologya  Risk 37/132 (28.0 %) 11/88 (12.5 %) 9/62 (14.5 %)  aORb (95 % CI) Reference group 0.52 (0.20–1.36) 0.54 (0.20–1.47)  P value 0.18 0.23  Risk 28/134 (20.9 %) 9/90 (10.0 %) 2/65 (3.1 %) LBW (<2500 g)  aORb (95 % CI) Reference group 0.29 (0.12–0.75) 0.08 (0.02–0.39)  P value 0.01 0.002  Risk 23/134 (17.2 %) 3/90 (3.3 %) 1/65 (1.5 %) Pre-term delivery (<37 weeks)  aORb (95 % CI) Reference group 0.13 (0.03–0.53) 0.05 (0.01–0.43)  P value 0.005 0.006 Fetal/neonatal deaths  Risk 10/134 (7.5 %) 1/90 (1.1 %) 0/65 (0 %)  aORb (95 % CI) Reference group 0.10 (0–0.86) 0 (n/a)  P value 0.03 0.03 LBW low birth weight; aOR adjusted odds ratio aIncludes all subjects with evaluable outcomes of interest bOdds ratio adjusted for gravidity, household wealth, presence of parasites at enrollment, gestational age study drugs started, and assigned IPTp treatment arm Association between IRS and birth outcomes The prevalence of LBW was significantly higher in infants born to mothers with no IRS protection (20.9 %) compared to those with >0–20 % (10.0 %, P = 0.01) or >20–43 % IRS protection (3.1 %, P = 0.002) (Table 2). Compared to infants born to women with no IRS protection, mean birth weight was 196 g higher among infants born to women with 0–20 % IRS protection (95 % CI 51–340 g, P = 0.008) and 257 g higher among infants born to women with >20–43 % IRS protection (95 % CI 105–409 g, P = 0.001) (Fig. 3). When restricting the analysis to women without histologic evidence of placental malaria, the prevalence of LBW was similarly higher in infants born to mothers with no IRS protection (18.8 %) compared to those with >0–20 % (7.9 %, P = 0.07) or >20–43 % IRS protection (2.2 %, P = 0.03).Fig. 3 Impact of IRS on birth weight. Shown are birth weights (clear circles) from infants born to mothers with no IRS exposure (n = 134), >0–20 % exposure (n = 90), and >20–43 % exposure (n = 65). Also shown are model adjusted means and 95 % CI (lines and error bars), and P values comparing birth weights between groups, adjusted for gravidity, gestation age study drugs started, wealth category, LAMP at enrolment, and treatment arm The prevalence of pre-term birth was significantly higher in infants born to women with no IRS protection (17.2 %) compared to those with >0–20 % (3.3 %, P = 0.01) or >20–43 % IRS protection (1.5 %, P = 0.006) (Table 2). Of 11 fetal/neonatal deaths (three spontaneous abortions, three still births, and five neonatal deaths), ten (7.5 % of pregnancies) occurred among those with no IRS protection, and none occurred among those with 20–43 % IRS protection (P = 0.03). Given the imbalances noted between IRS protection strata, a sensitivity analysis was performed to evaluate associations between IRS and outcomes at delivery using propensity scores with inverse probability weighting, with similar results (see Additional file 2). In all analyses, there was no significant interaction observed between IRS and assigned IPTp group. Discussion While conducting a clinical trial comparing novel IPTp strategies in a high-transmission setting in Uganda with near-universal LLIN coverage, the unexpected opportunity arose to evaluate the additional impact of IRS on measures of malaria during pregnancy and adverse birth outcomes. In this setting, the period following protection with IRS was associated with a significantly lower incidence of symptomatic malaria and prevalence of parasitaemia during pregnancy. Reductions in parasite prevalence were observed among women randomized to either IPTp-SP or IPTp-DP, suggesting additional benefits of IRS when given along with LLINs and IPTp. Importantly, the period following protection with IRS was also associated with a significantly lower prevalence of malaria parasites in placental blood at the time of delivery, and a significantly lower risk of LBW, pre-term delivery, and fetal/neonatal deaths. WHO recommends IRS as a central part of malaria control policy in Africa, but coverage rates have been low (<10 %) and have declined in recent years, possibly due to increased costs from spraying with non-pyrethroid insecticides [4]. Although IRS has been shown to be very effective in reducing vector densities and malaria morbidity, particularly among children [10–18], no study has assessed its impact on malaria morbidity during pregnancy and birth outcomes. In eastern Uganda, rates of pyrethroid resistance among Anopheles mosquitoes and the prevalence of molecular markers of SP resistance among P. falciparum parasites are very high [6, 27]. This may explain the high parasite prevalence prior to IRS among pregnant women given LLINs and randomized to IPTp-SP, as well as the significant reduction in parasite prevalence both during pregnancy and at delivery following implementation of carbamate-containing IRS. Although significant differences in placental histopathology with increasing IRS exposure were not observed, it is possible that histopathologic changes may reflect early malaria exposure during pregnancy [28], as all women became pregnant prior to the implementation of IRS, and nearly 60 % of women had evidence of asymptomatic parasitaemia at enrolment [22]. Notably, IRS was associated with significant reductions in the risk of LBW and pre-term delivery, and no fetal or neonatal deaths occurred among women with >20 % IRS protection during pregnancy. LBW and pre-term delivery are multifactorial disorders, and relatively few interventions have been shown to improve these outcomes, especially in resource-limited settings [29–31]. It is possible that IRS may have led to improved birth outcomes by preventing placental infection by malaria parasites [2, 32]. However, malaria in pregnancy is thought to contribute to 20 % of LBW in sub-Saharan Africa [2], significantly less than the reductions observed in this study, and improvements in birth outcomes were observed even in women without histological evidence of placental malaria. Furthermore, not all interventions that have been shown to prevent malaria infection during pregnancy have been shown to improve birth outcomes [33]. In addition to targeting Anopheles mosquitoes and preventing malaria, IRS may be targeting other vectors such as, potentially, Aedes and culicine mosquitoes, argasid ticks and fleas that are capable of transmitting a variety of vector-borne pathogens [34] that could be associated with poor pregnancy outcomes. However, data on additional pathogens that may have been prevented by IRS were unavailable in this study. There are several limitations to this study. As this was a pre-post observational study nested within a trial of IPTp, definitive causal inferences regarding the effect of IRS on pregnancy and birth outcomes cannot be drawn. Unmeasured confounders, including temporal and seasonal changes, may have affected these results, although parasite prevalence prior to IRS was stable and high, lessening this possibility. There were also significant differences between the IRS exposure groups, including a higher proportion of first pregnancies among women in the group not exposed to IRS during pregnancy. Multivariate adjustment and a secondary sensitivity analysis using propensity scores and inverse probability weighting were performed to adjust for these differences, although it is possible that residual confounding remains. Furthermore, it is possible that given the timing of IRS within the context of this trial, a cohort effect may have led to these results, specifically with regard to birth outcomes. However, women were enrolled over a period of 6 months, and delivered over a period of 6 months, reducing the possibility that this form of bias was responsible for these results. Finally, due to sample size and study design, several aspects which would further help to inform policy were incompletely addressed. These include a determination of effect duration, as well as the relative contributions to protection from direct household versus community spraying. Conclusions The provision of IRS, in combination with LLINs and IPTp, was temporally associated with a reduction in the incidence of malaria during pregnancy, prevalence of malaria parasites during pregnancy and in placental blood, as well as improved birth outcomes. IRS may represent an important tool for combating malaria in pregnancy and for improving birth outcomes in malaria-endemic settings. Additional and larger controlled evaluations, including monitoring for malaria in pregnancy metrics in settings where IRS is currently being implemented, are warranted to further assess this promising strategy, and to inform policy. Additional files 10.1186/s12936-016-1489-x Parasite prevalence during pregnancy by calendar month at enrolment and visits prior to study drug initiation. Shown is the predicted probability of having a positive (+) LAMP result during pregnancy at enrolment and before initiation of study drugs. Parasite prevalence point estimates and standard errors obtained using generalized estimating equations after adjustment for gravidity and age. 10.1186/s12936-016-1489-x Sensitivity analysis: estimation of effect of any IRS vs no IRS protection on outcomes measured at birth. Authors’ contributions MKM, AK, GD, and PJ conceived and designed the study. MKM, AK, PNat, PO, JA, PNay, AM, and BM participated in data collection. MKM, AK, MR, BG, GD, and PJ participated in the data analysis. All authors participated in the writing of the manuscript, read and approved the final manuscript. All authors read and approved the final manuscript. Acknowledgements We are grateful to all the parents and guardians for kindly giving their consent and the study participants for their cooperation. We thank all the members of the study team in Uganda. Competing interests None of the authors declare any potential conflicts of interest, including financial interests and relationships and affiliations relevant to the subject of the manuscript. The authors declare that they have no competing interests. Availability of data and material All data generated or analysed during this study are included in this published article and its supplementary information files. Ethics approval and consent to participate Informed consent was obtained from all study participants. Ethical approval was obtained from the Uganda National Council of Science and Technology, the Makerere University School of Medicine Research and Ethics Committee, the Makerere University School of Biomedical Sciences Research and Ethics Committee, and the University of California, San Francisco, Committee on Human Research. Funding Research reported in this publication was supported by the National Institutes of Health under award numbers PO1HD059454-03, U19AI089674, K23AI100949. Additional support was received from the Centers for Diseases Control (AM) and the President’s Malaria Initiative (BM). The content is solely the responsibility of the authors and does not necessarily represent the official position of the NIH, Centers for Diseases Control and Prevention, or the President’s Malaria Initiative. ==== Refs References 1. Dellicour S Tatem AJ Guerra CA Snow RW ter Kuile FO Quantifying the number of pregnancies at risk of malaria in 2007: a demographic study PLoS Med 2010 7 e1000221 10.1371/journal.pmed.1000221 20126256 2. Desai M ter Kuile FO Nosten F McGready R Asamoa K Brabin B Epidemiology and burden of malaria in pregnancy Lancet Infect Dis 2007 7 93 104 10.1016/S1473-3099(07)70021-X 17251080 3. Walker PG ter Kuile FO Garske T Menendez C Ghani AC Estimated risk of placental infection and low birthweight attributable to Plasmodium falciparum malaria in Africa in 2010: a modelling study Lancet Glob Health 2014 2 e460 e467 10.1016/S2214-109X(14)70256-6 25103519 4. WHO. World Malaria Report Geneva 2015 Switzerland World Health Organization 2015 5. Arinaitwe E Ades V Walakira A Ninsiima B Mugagga O Patil TS Intermittent preventive therapy with sulfadoxine-pyrimethamine for malaria in pregnancy: a cross-sectional study from Tororo, Uganda PLoS One 2013 8 e73073 10.1371/journal.pone.0073073 24023811 6. Mawejje HD Wilding CS Rippon EJ Hughes A Weetman D Donnelly MJ Insecticide resistance monitoring of field-collected Anopheles gambiae s.l. populations from Jinja, eastern Uganda, identifies high levels of pyrethroid resistance Med Vet Entomol 2013 27 276 283 10.1111/j.1365-2915.2012.01055.x 23046446 7. Ramphul U Boase T Bass C Okedi LM Donnelly MJ Müller P Insecticide resistance and its association with target-site mutations in natural populations of Anopheles gambiae from eastern Uganda Trans R Soc Trop Med Hyg 2009 103 1121 1126 10.1016/j.trstmh.2009.02.014 19303125 8. Verhaeghen K Van Bortel W Trung HD Sochantha T Keokenchanh K Coosemans M Knockdown resistance in Anopheles vagus , An. sinensis , An. paraliae and An. peditaeniatus populations of the Mekong region Parasit Vectors 2010 3 59 10.1186/1756-3305-3-59 20646327 9. ter Kuile FO van Eijk AM Filler SJ Effect of sulfadoxine-pyrimethamine resistance on the efficacy of intermittent preventive therapy for malaria control during pregnancy: a systematic review JAMA 2007 297 2603 2616 10.1001/jama.297.23.2603 17579229 10. Bukirwa H Yau V Kigozi R Filler S Quick L Lugemwa M Assessing the impact of indoor residual spraying on malaria morbidity using a sentinel site surveillance system in Western Uganda Am J Trop Med Hyg 2009 81 611 614 10.4269/ajtmh.2009.09-0126 19815875 11. Zhou G Githeko AK Minakawa N Yan G Community-wide benefits of targeted indoor residual spray for malaria control in the western Kenya highland Malar J 2010 9 67 10.1186/1475-2875-9-67 20199674 12. Pluess B Tanser FC Lengeler C Sharp BL Indoor residual spraying for preventing malaria Cochrane Database Syst Rev 2010 4 CD006657 20393950 13. Skarbinski J Mwandama D Wolkon A Luka M Jafali J Smith A Impact of indoor residual spraying with lambda-cyhalothrin on malaria parasitemia and anemia prevalence among children less than 5 years of age in an area of intense, year-round transmission in Malawi Am J Trop Med Hyg 2012 86 997 1004 10.4269/ajtmh.2012.11-0621 22665608 14. Kigozi R Baxi SM Gasasira A Sserwanga A Kakeeto S Nasr S Indoor residual spraying of insecticide and malaria morbidity in a high transmission intensity area of Uganda PLoS One 2012 7 e42857 10.1371/journal.pone.0042857 22880123 15. Hamel MJ Otieno P Bayoh N Kariuki S Were V Marwanga D The combination of indoor residual spraying and insecticide-treated nets provides added protection against malaria compared with insecticide-treated nets alone Am J Trop Med Hyg 2011 85 1080 1086 10.4269/ajtmh.2011.10-0684 22144448 16. West PA Protopopoff N Wright A Kivaju Z Tigererwa R Mosha FW Indoor residual spraying in combination with insecticide-treated nets compared to insecticide-treated nets alone for protection against malaria: a cluster randomised trial in Tanzania PLoS Med 2014 11 4 e1001630 10.1371/journal.pmed.1001630 24736370 17. Gimnig JE Otieno P Were V Marwanga D Abong’o D Wiegand R The effect of indoor residual spraying on the prevalence of malaria parasite infection, clinical malaria and anemia in an area of perennial transmission and moderate coverage of insecticide treated nets in Western Kenya PLoS One 2016 11 e0145282 10.1371/journal.pone.0145282 26731524 18. Hamainza B Sikaala CH Moonga HB Chanda J Chinula D Mwenda M Incremental impact upon malaria transmission of supplementing pyrethroid-impregnated long-lasting insecticidal nets with indoor residual spraying using pyrethroids or the organophosphate, pirimiphos methyl Malar J 2016 15 100 10.1186/s12936-016-1143-7 26893012 19. Pinder M Jawara M Jarju LB Salami K Jeffries D Adiamoh M Efficacy of indoor residual spraying with dichlorodiphenyltrichloroethane against malaria in Gambian communities with high usage of long-lasting insecticidal mosquito nets: a cluster-randomised controlled trial Lancet 2015 385 1436 1446 10.1016/S0140-6736(14)61007-2 25498847 20. Lines J Kleinschmidt I Is malaria control better with both treated nets and spraying? Lancet 2015 385 1375 1377 10.1016/S0140-6736(14)61306-4 25498846 21. Kamya MR Arinaitwe E Wanzira H Katureebe A Barusya C Kigozi SP Malaria transmission, infection, and disease at three sites with varied transmission intensity in Uganda: implications for malaria control Am J Trop Med Hyg 2015 92 903 912 10.4269/ajtmh.14-0312 25778501 22. Kakuru A Jagannathan P Muhindo MK Natureeba P Awori P Nakalembe M Dihydroartemisinin–piperaquine for the prevention of malaria in pregnancy N Engl J Med 2016 374 928 939 10.1056/NEJMoa1509150 26962728 23. President’s Malaria Initiative, Uganda. Malaria operational plan FY 2014. 2014. http://www.pmi.gov/docs/default-source/default-document-library/malaria-operational-plans/fy14/uganda_mop_fy14.pdf?sfvrsn=14. Accessed 12 Jan 2015. 24. Uganda Ministry of Health. Malaria Indicator Survey (MIS) 2014–15. 2015. https://www.dhsprogram.com/pubs/pdf/MIS21/MIS21.pdf. Accessed 3 Jan 2016. 25. Hopkins H Gonzalez IJ Polley SD Angutoko P Ategeka J Asiimwe C Highly sensitive detection of malaria parasitemia in a malaria-endemic setting: performance of a new loop-mediated isothermal amplification kit in a remote clinic in Uganda J Infect Dis 2013 208 645 652 10.1093/infdis/jit184 23633405 26. Muehlenbachs A Fried M McGready R Harrington WE Mutabingwa TK Nosten F A novel histological grading scheme for placental malaria applied in areas of high and low malaria transmission J Infect Dis 2010 202 1608 1616 10.1086/656723 20929353 27. Mbonye AK Birungi J Yanow SK Shokoples S Malamba S Alifrangis M Prevalence of Plasmodium falciparum resistance markers to sulfadoxine-pyrimethamine among pregnant women receiving intermittent preventive treatment for malaria in Uganda Antimicrob Agents Chemother 2015 59 5475 5482 10.1128/AAC.00507-15 26124154 28. Muehlenbachs A Nabasumba C McGready R Turyakira E Tumwebaze B Dhorda M Artemether-lumefantrine to treat malaria in pregnancy is associated with reduced placental haemozoin deposition compared to quinine in a randomized controlled trial Malar J 2012 11 150 10.1186/1475-2875-11-150 22554092 29. Piso B Zechmeister-Koss I Winkler R Antenatal interventions to reduce preterm birth: an overview of cochrane systematic reviews BMC Res Notes 2014 7 265 10.1186/1756-0500-7-265 24758148 30. Iams JD Romero R Culhane JF Goldenberg RL Primary, secondary, and tertiary interventions to reduce the morbidity and mortality of preterm birth Lancet 2008 371 164 175 10.1016/S0140-6736(08)60108-7 18191687 31. WHO WHO recommendations on interventions to improve preterm birth outcomes 2015 Geneva World Health Organization 32. Cottrell G Moussiliou A Luty AJ Cot M Fievet N Massougbodji A Submicroscopic Plasmodium falciparum infections are associated with maternal anemia, premature births, and low birth weight Clin Infect Dis 2015 60 1481 1488 25694651 33. Desai M Gutman J L’Lanziva A Otieno K Juma E Kariuki S Intermittent screening and treatment or intermittent preventive treatment with dihydroartemisinin-piperaquine versus intermittent preventive treatment with sulfadoxine-pyrimethamine for the control of malaria during pregnancy in western Kenya: an open-label, three-group, randomised controlled superiority trial Lancet 2015 386 2507 2519 10.1016/S0140-6736(15)00310-4 26429700 34. Borchert JN Eisen RJ Atiku LA Delorey MJ Mpanga JT Babi N Efficacy of indoor residual spraying using lambda-cyhalothrin for controlling nontarget vector fleas (Siphonaptera) on commensal rats in a plague endemic region of northwestern Uganda J Med Entomol 2012 49 1027 1034 10.1603/ME11230 23025183

PMC005xxxxxx/PMC5002130.txt

==== Front Reprod Biol EndocrinolReprod. Biol. EndocrinolReproductive Biology and Endocrinology : RB&E1477-7827BioMed Central London 18610.1186/s12958-016-0186-5Research17β-estradiol modifies human spermatozoa mitochondrial function in vitro http://orcid.org/0000-0002-9802-374XKotwicka Malgorzata mkotwic@ump.edu.pl 1Skibinska Izabela iskibinska@ump.edu.pl 1Jendraszak Magdalena mahwaxx@yahoo.com 1Jedrzejczak Piotr piotrjedrzejczak@gmail.com 21 Department of Cell Biology, Faculty of Health Sciences, Poznan University of Medical Sciences, Rokietnicka 5D, 60-806 Poznan, Poland 2 Division of Infertility and Reproductive Endocrinology, Faculty of Medicine I, Poznan University of Medical Sciences, Polna 33, 60-535 Poznan, Poland 26 8 2016 26 8 2016 2016 14 1 5015 3 2016 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 It is assumed that spermatozoa are target cells for estrogens however, the mechanism of their action is not fully understood. The aim of this study was to investigate the influence of 17β-estradiol (E2) on the human spermatozoa mitochondrial function. Methods The effects on spermatozoa of E2 at final concentrations of 10−10, 10−8 and 10−6 M were studied regarding the following phenomena: (1) kinetics of intracellular free calcium ions changes (using Fluo-3), (2) mitochondrial membrane potential ΔΨm (using JC-1 fluorochrome), (3) production of superoxide anion in mitochondria (using MitoSOX RED dye), (4) spermatozoa vitality (propidium iodide staining) and (5) phosphatidylserine membrane translocation (staining with annexin V marked with fluorescein). Results E2 initiated rapid (within a few seconds) dose dependent increase of intracellular free calcium ions concentration. E2 was changing the mitochondrial membrane potential: 10−8 M initiated significant increase of percentage of high ΔΨm spermatozoa while the 10−6 M induced significant decrease of high ΔΨm cells. In spermatozoa stimulated with E2 10−6 M a significant increase of mitochondrial superoxide anion level was observed. 2 h incubation of spermatozoa with E2 did not alter cells vitality nor stimulated phosphatidylserine membrane translocation, for all three doses. Conclusions 17β-estradiol affected the human spermatozoa mitochondrial function. E2 in low concentration improved while in high concentration might deteriorate mitochondrial function. Keywords EstrogenHuman spermatozoaMitochondrial membrane potentialMitochondrial superoxide anionNational Research Centre (PL)N N 401 077037Kotwicka Malgorzata issue-copyright-statement© The Author(s) 2016 ==== Body Background It has been shown that estrogens play important role in the regulation of male reproductive system [1–3]. Previous studies revealed that human spermatozoa can be considered target cells for estrogens. Impact of estrogens comprises sperm capacitation, acrosome reaction, motility and fertilizing ability [4–7]. 17β-estradiol (E2), the most powerful estrogen, affects target cells via estrogen receptors (ESRs): ESR1 and ESR1. Spermatozoa are transcriptionally inactive cells. E2 nongenomic signal transduction using intracellular secondary messengers is the only pathway possible [8, 9]. Nongenomic E2 effects are mediated via either membrane bound receptors or interaction with other membrane proteins and/or lipids [10]. Classic nuclear ESR1 and ESR2 receptors were described within cell membrane [11, 12]. However, considering absence of transmembrane domains in both ESRs, post-translational ESRs modification or ESRs binding with other membrane proteins are suggested. Kim et al. suspected that 45 and 66 kDa splicing ESRs variants, lacking the A/B domain, are capable to act as integral transmembrane proteins [13]. G protein coupled estrogen receptor (GPER) representing alternative pathway for rapid nongenomic answer was detected in human, boar and stallion spermatozoa [14, 15]. Both subtypes of classical ESRs are expressed in ejaculated spermatozoa however, the reports concerning their localization are not unanimous [14, 16, 17]. In our own unpublished study we observed strong expression of ESR1 and ESR2 in the midpiece region of human spermatozoa, supposing their presence in spermatozoal mitochondria. Our findings support the results obtained by Solakidi et al. with MitoTracker Red CMXRos mitochondrial marker usage [17]. Also, the study by Guido et al., exploiting the colloidal gold, indicated the presence of ESRs in spermatozoal mitochondria obtained from fertile men. The mitochondrial ESR2 expression was significantly stronger than that of the ESR1. Moreover, in patients with variocele, significant decrease of ESRs expression in the midpiece regions and tail or sporadically no ESRs expression were observed [7]. Assuming the ESRs are present in sperm mitochondria, their functions should be affected by estrogens. Estrogens can have an impact on mitochondrial function, however the mechanism is not fully understood [18–20]. It is known that due to high lipid content mitochondria are reservoirs of cell estrogens [21]. It was indicated that besides passive estrogen diffusion, the mechanism of rapid estrogen transport to mitochondria is present within the cell, probably via receptor mediated endocytosis [22]. ESRs presence was revealed in somatic cells mitochondria. In most mitochondria the ESR2 seems to be the dominant receptor type even if both types were detected. Mitochondrial ESR2 mass analysis revealed various receptor isoforms [23]. The mechanism controlling ESR transport to mitochondria is poorly understood. It is suggested that in mitochondria, similarly as in the nucleus, the ESRs play the role of transcription factors [24]. It is postulated that estrogens act on mitochondria, not only by affecting the mitochondrial DNA. It was observed that they stimulate local, mitochondrial increase of free calcium ions (Ca2+) concentration. It was suggested that estrogens inhibit sodium dependent efflux of free calcium ions from mitochondria [25]. As a consequence of the increase of mitochondrial free calcium ions concentration, the increase of synthesis of reactive oxygen species (ROS) such as superoxide anion (O2*−), hydrogen peroxide (H2O2) and hydroxyl radical (OH−) occurs. This indicates that estrogen induced increase of mitochondrial Ca2+ concentration stimulates the activity of mitochondrial nitric oxide synthase (mtNOS), leading to inhibition of cytochrome c oxidase activity [26]. The aim of present study was to investigate the influence of 17β-estradiol on human spermatozoa mitochondrial function, based on the analysis of mitochondrial membrane potential changes and detection of mitochondrial superoxide anion. Methods Semen of 10 normozoospermic men (according to WHO 2010 criteria) was analyzed. Material was obtained after sexual abstinence of 3–5 days. Spermatozoa with high motility were isolated by the swim-up technique. Ham’s F-10 medium was used as sperm cell extender. Isolated cells were incubated with E2 in final concentrations of 10−10, 10−8 and 10−6 M. Spermatozoa stimulated by Ham’s F-10 medium were used as controls. Spermatozoa mitochondrial membrane potential (ΔΨm) was noted at 5, 10, 15, 20, 25, 30 and 120 min after exposure to E2. Changes of Ca2+ level were examined throughout 400 s after exposure to E2, every 10 s. Sperm vitality, phosphatidylserine membrane translocation and mitochondrial superoxide anion level were examined at 2 h after exposure to E2. Changes in intracellular free calcium ions level Fluo-3 (Molecular Probes; Ex/Em = 488/526 nm) was used to study changes in free calcium ions level in human sperm. Spermatozoa (1 × 106 cells/mL) were incubated with 4 μM Fluo-3 for 45 min at 37 °C according to the manufacturer’s protocol. For confocal microscopy, spermatozoa were immobilized in 1 % (w/v) agarose and then treated with E2. Microscopic images were used for gating single sperm cells in which fluorescence changes were recorded. Forty images were collected (every 10 s) and used to study the kinetics of intracellular free calcium ions changes. Spermatozoa were observed using LSM 510 confocal microscope (Zeiss, Jena, Germany) equipped with a Plan Apochromat 63×/1.4 Oil DIC objective. Sperm cells stimulated with Ham’s F-10 medium were used as a control for fluorescence intensity changes. Detection of mitochondrial membrane potential To evaluate spermatozoa mitochondrial membrane potential (ΔΨm) the 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolocarbocyanine iodide (JC-1; Molecular Probes) was used. This is a lipophilic cationic compound that has the unique ability to label spermatozoa depending on either low or high mitochondrial potential. In the case of spermatozoa with high mitochondrial membrane potential (ΔΨm > 80–100 mV), the JC-1 forms aggregates emitting red to orange fluorescence (wavelength of 590 nm). In spermatozoa with low mitochondrial potential (ΔΨm < 80–100 mV), the JC-1 forms monomers emitting green fluorescence (wavelength of 525 to 530 nm). In both cases the excitation wavelength was 488 nm. JC-1 was diluted in DMSO (dimethyl sulfoxide) and added to cell suspension at a final concentration of 1 μM. The cells were incubated in darkness for 30 min in temperature of 37 °C. Afterwards the cell suspension was washed twice (5 min × 2400 rpm) with the use of Ham’s F-10 medium. JC-1 fluorescence emissions in spermatozoa treated with valinomycin (100 nM) was used as a control that prevents JC-1 aggregation. Valinomycin permeabilizes the mitochondrial membrane for K+ ions, and thus, dissipates the mitochondrial electrochemical potential. The results were expressed as the percentage of cells exhibiting high mitochondrial membrane potential. Detection of mitochondrial superoxide anion In order to estimate the amount of superoxide anion produced in spermatozoal mitochondria, we used MitoSOX Red fluorochrome (Molecular Probes). The analysis was performed according to the method described by Koppers et al. [27] with the use of confocal microscope and flow cytometer. MitoSOX Red stock solution (5 mM diluted in DMSO) was added to cell suspension (20 × 106 cells per mL) to give the final concentration of 2 μM. Cells were incubated for 15 min in darkness at 37 °C and afterwards washed twice with F-10 medium (5 min at 600 × g). Microscopic observation was made under LSM 510 confocal microscope (Carl Zeiss GmbH, Germany). MitoSOX Red fluorescence was measured using flow cytometer FACSCalibur (Becton–Dickinson, USA). The results were expressed as the percentage of MitoSOX positive cells and as mean fluorescence intensity of MitoSOX positive cells. Sperm vitality and phosphatidylserine membrane translocation To determine phosphatidylserine membrane translocation (PST) from the inner to the outer layer of the plasma membrane, the annexin-V labeled with fluorescein (AnV-FLUOS) (Roche Molecular Diagnostics, Darmstadt, Germany) was used. Simultaneously, to distinguish between viable and dead spermatozoa the propidium iodide (PI) staining was used, in the final concentration of 0.125 μg/L (Sigma-Aldrich, St. Louis, MO). Double staining was conducted according to manufacturer’s recommendations. Flow cytometry The fluorescence signals of labeled spermatozoa were analyzed by flow cytometer FACSCalibur. 10 000 cells were examined for each experiment. The fluorescence of An-V-FLUOS and PI was excitated by argon laser (488 nm) and emission of An-V-FLUOS was measured in the FL1 channel (515–545 nm), while the red fluorescence of PI was detected in the FL3 channel (650 nm). The fluorescence of MitoSOX Red was analyzed in the FL2 channel (561–603 nm). The emission of JC-1 monomers and aggregates was measured in the FL1 channel (515–545 nm) and FL2 channel (561–603 nm), respectively. All data was collected and analyzed using CellQuest Pro software (v.5.2.1) (Becton–Dickinson). Statistical analysis The analysis was performed using Statistica 10 software (StatSoft Inc., Tulsa, OK, USA). Nonparametric Kruskal-Wallis test with Dunn’s post hoc test were applied. Data were presented as mean ± SD and considered statistically significant at P < 0.05. Results Intracellular free calcium ions level changes In most viable spermatozoa, the highest concentration of intracellular free calcium ions was observed within the midpiece and distal part of head (Fig. 1a). E2 caused a rapid, transient increase of intracellular free calcium level. The reaction was observed at 10 s after stimulation and lasted a few minutes. The kinetics of the reaction were dose depended (Fig. 1b).Fig. 1 Kinetics of intracellular free calcium ions concentration changes in human spermatozoa after E2 stimulation. a Spermatozoa stained with Fluo-3 reveal high concentration of free calcium ions in midpiece. b Representative reaction after stimulation with 17β-estradiol in 10−10 M, 10−8 M or 10−6 M final concentration. The individual lines in the graphs correspond to the reactions of individual sperm cells. Arrows indicate the moment of E2 administration. Ch1-fluorescence channel for Fluo-3 Mitochondrial membrane potential The percentage of cells with high membrane potential revealed from 60 to 92 %. No significant changes of membrane potential were observed in controls during 2 h incubation. Stimulation of human sperm cells with E2 caused significant changes of ΔΨm. The 2 h incubation of sperm cells with E2 at a final concentration of 10−10 M did not result in statistically significant changes of ΔΨm (P > 0.05), while E2 at a final concentrations of 10−8 M and 10−6 M caused a significant increase (P = 0.004) or a significant decrease (P = 0.04) of the percentage of sperm cells with high ΔΨm, respectively (Fig. 2 and Fig. 3). E2 induced ΔΨm changes at final concentration of 10−8 M were observed at 5 min while the changes at final concentration of 10−6 M were observed at 1 min after stimulation.Fig. 2 Analysis of spermatozoa mitochondrial membrane potential after exposure to E2. Kinetics of the changes of percentage of spermatozoa with high mitochondrial membrane potential measured at 1, 5, 10, 15, 20, 25, 30 and 120 min after exposure to E2 with the use of JC-1 fluorochrome. Data obtained from 10 separate analyses expressed as means ± standard deviation Fig. 3 Changes of percentage of spermatozoa with high and low mitochondrial membrane potential after E2 stimulation. R2 and R3 gates represent cells with high (JC-1 aggregates emitting red to orange fluorescence) and low (JC-1 monomers emitting green fluorescence) mitochondrial membrane potential after E2 stimulation, respectively. Representative reactions: a-c increase of percentage of high ΔΨ spermatozoa after 10−8 M E2 stimulation; e–g-decrease of percentage of high ΔΨ spermatozoa after 10−6 M E2 stimulation. d, h-controls after 120 min stimulation Detection of mitochondrial superoxide anion The swim-up isolated fraction contained two cells subpopulations: MitoSOX−and MitoSOX+ (Fig. 4). The percentage of MitoSOX positive cells ranged from 6.7 to 42 %. The 2 h incubation of sperm cells with E2 did not reveal significant change of the percentage of MitoSOX positive cells, for any of the concentrations used (P > 0.05) (Fig. 5a). However, a significant increase of MitoSOX fluorescence intensity was observed in cells stimulated with E2 at concentration of 10−6 M (P < 0.01) (Fig. 5b and Fig. 6).Fig. 4 Detection of mitochondrial superoxide anion level with MitoSOX Red fluorochrome. a Nonlabelled cells (controls). b–c Swim-up isolated fractions of two semen samples Fig. 5 Changes of mitochondrial superoxide anion level (using MitoSOX Red dye) after 17β-estradiol stimulation. a Percentage of MitoSOX-positive sperm treated with E2. b Significant dose-dependent increase in intensity of MitoSOX fluorescence after E2 stimulation. Data obtained from 10 separate analyses are expressed as means ± standard deviation. Different superscript letters above each bar denote significant difference, P < 0.05 Fig. 6 Representative examples of increase of MitoSOX fluorescence intensity after E2 stimulation. FL-2-channel of MitoSOX Red fluorescence Effects of 17β-estradiol on sperm vitality Flow cytometry analyses identified four fractions of spermatozoa: (1) An-V−/PI− viable sperm without PST, (2) An-V+/PI− viable sperm with PST, (3) An-V−/PI+ dead sperm without PST and (4) An-V+/PI+ dead sperm with PST. The sperm cells percentage of each fraction was 81.2 ± 6.1 %, 0.4 ± 0.3 %, 15.9 ± 5.5 % and 2.4 ± 1.4 %, respectively. It did not change after 2 h incubation with E2, for all concentrations used. Discussion In this study the effect of E2 on mitochondrial membrane potential and mitochondrial superoxide level was examined. At the same time the effect of E2 on sperm vitality as well as changes of intracellular free calcium ions concentration were investigated. E2 caused significant changes of mitochondrial membrane potential. When stimulating cells with E2 10−10 M, changes were not relevant. Concentration of 10−8 M caused distinct increase of percentage of high ΔΨm spermatozoa however, concentration of 10−6 M induced significant decrease of percentage of high ΔΨm spermatozoa. ΔΨm is an indicator of mitochondrion energetic potential. High ΔΨm denotes a functionally undisturbed mitochondrion. Positive correlation between decreased effectiveness of mitochondria and both poor sperm motility and poor capacity to fertilize the egg was described [28]. It is postulated that changes of ΔΨm can be used as an efficient tool estimating the fertilizing potential of spermatozoa devoted to in vitro fertilization [28, 29]. ΔΨm changes observed in the present study suggest that low estrogen concentration favorably influenced mitochondrial function, which, in turn, can be associated with positive impact on the vitality and motility of sperm cells. Several studies reported estrogen effects in mature spermatozoa [4–7]. Ded et al. [5] postulated significant differences in the response to tested estrogens at different capacitation time and among individual animals. It cannot be excluded that the observed calcium ions increase and the ΔΨm increase after E2 10−10 M stimulation are both the elements of the ongoing capacitation process. Observations regarding increase of sperm mitochondrial membrane potential under capacitation conditions have been previously made [30]. A sharp decline of ΔΨm after stimulation with higher E2 concentration appears to be an interesting finding, especially when taking into consideration the common exposure of contemporary men to xenoestrogens and their probable synergistic action with endogenous estrogens. Studies conducted on somatic cells indicate that mitochondrial estrogen’s effect largely depends on hormone dose and cell type. It was shown that estrogens influence the transportation of calcium ions to mitochondria [31, 32]. It is suggested that estrogen induced increase of free calcium ions concentration, caused by estrogens, could activate protein phosphatase which leads to dephosphorylation of cytochrome c oxidase. This, in turn, would influence the increase of membrane mitochondrial potential and production of reactive oxygen species [33]. It is proposed that estrogens can modulate ΔΨm by influencing the phosphorylation of protein complexes of the respiratory chain [33]. It can be assumed that 17β-estradiol induced changes of ΔΨm in spermatozoa could be a consequence of the influx of free calcium ions into the mitochondria. It is important to notice that the spermatozoa, in contrast to the somatic cells, present small capacities of calcium ions buffering within intracellular organelle. Calcium ions storage occurs mainly within mitochondria, as well as in acrosome and posterior part of the nucleus however, a long lasting increased calcium ions concentration may trigger pathological pathways leading to apoptosis. Excessive increase of mitochondrial concentration of free calcium ions is connected to the significant decrease in membrane mitochondrial potential. This, in turn, activates megachannels and in consequence causes release of cytochrome c to cytoplasm [25]. It is that estradiol induces apoptosis of rat’s spermatogenic cells by lowering the hyperpolarization of mitochondrial membrane [34]. Our study demonstrated, that sperm cells stimulated with high E2 concentrations presented decline of ΔΨm accompanied by increased mitochondrial superoxide anion level. Changes of mitochondrial ΔΨm are closely related to the issue of the oxidative stress and its influence on cell functions, the situation, when the antioxidant defensive systems fail and when the level of enzymatic and nonenzymatic molecules presenting antioxidant properties is low. It is suggested that oxidative stress is an etiological factor in various disorders, such as cardiovascular disease, neoplasms, diabetes, nervous system degenerative diseases, male and female infertility [27, 35]. To be noticed, spermatozoa are the first cells in which potential susceptibility to oxidative damages was proven. Sperm cells are very sensitive to oxidative stress mainly because of their structure. Among other, they contain polyunsaturated fatty acids (in particular docosahexaenoic acid), including six double bonds per molecule, which makes spermatozoa specific aggregates of electrons, susceptible to oxidation and other structural modifications. These modifications affect mostly cell membrane and are connected with its disturbed liquidity. This, in turn, leads to decreased motility, disturbances in acrosome reaction or in spermatozoon-oocyte fusion [35, 36]. Moreover, spermatozoa are exposed to oxidative stress because of their specific size, localization of intracellular antioxidative enzymes and limited ability of DNA repair [27, 35]. In context of the studies reporting relation between the PST and the ΔΨm decrease [37, 38] one could expect that the observed ΔΨm decline and the simultaneous increase of mitochondrial superoxide anion level in sperm stimulated with high E2 concentrations shall be accompanied with presence of sperm apoptosis markers. However, we did not observed significant increase of PST positive cells or of alive cells percentage. We speculate on the impact of the swim-up because this technique isolates spermatozoa with propoer motility and morphology as well as eliminates the spermatozoa with PST [39]. Nonapoptotic human spermatozoa with intact plasma membrane reveal the highest fertilizing potential [40]. We speculate that for this fraction the ΔΨm decrease and oxidative stress after 120 min E2 incubation did not resulted in PST. On the other hand, Grunewald et al. indicates possibility of interaction of the capacitation and apoptosis signaling systems that enables the capacitation process by prevention of apoptosis [41]. E2 involvement in capacitation could be the reason why apoptosis markers were not observed. Both the expression of ESR in the mitochondrial location and the observation that E2 regulates the movement of spermatozoa [7] suggest that estrogens may be involved in the metabolism of mitochondria. Mitochondria take part in fundamental cell processes such as (1) cellular respiration, (2) oxidative phosphorylation, (3) apoptosis, (4) synthesis of lipids, heme, amino acids, nucleotides, steroid hormones and (5) ions homeostasis. It is suggested that mitochondria are reservoirs for estrogens. Furthermore, in mitochondria of various somatic cells both the ESR1 and ESR2 have been localized. According to up to date literature it can be assumed that estrogens take part in the regulation of mitochondrial functions. In somatic cells estrogens cause increase of mitochondrial mRNA level of mtDNA encoded proteins. That indicates that estrogens can influence the level of gene expression [18, 42]. In regular conditions mitochondria produce small amounts of ROS that can be easily neutralized by cell antioxidants. Due to low concentration the mitochondrial ROS can play role of signal molecules. Previously, estrogens were considered only as antioxidants protecting for example from neurodegenerative diseases. Some studies indicate that 17β-estradiol protects mitochondria from oxidative stress in somatic cells [43, 44]. The presented results indicate that in case of sperm cells the effect of estrogen is dependent on the hormone concentration. It is said that the presence of estrogens may lead to increase of the level of reactive oxygen species in mitochondria by induction of nitric oxide synthase or by inhibiting cytochrome c oxidase. Also the increase of mitochondrial level of Ca2+ promotes ROS formation. Bennetts et al. observed that some metabolites of estradiol and xenoestrogens cause oxidative stress in human spermatozoa. An intense reaction was observed after stimulation with, for example, diethylstilbestrol or catechol estrogen derivatives. In case of genistein the effect was observed only in case of high doses of substrate used. No significant changes were observed under the influence of either 17β-estradiol, nonyphenol or bisphenol A [33]. Intracellular increase of free calcium ions level observed after E2 stimulation, suggests it is an important second messenger in nongenomic estrogen action. Conclusion In conclusion, 17β-estradiol affects the human spermatozoa mitochondrial function. Results obtained in this study, in which spermatozoa were stimulated with high 17β-estradiol concentration suggest, that excessive exposition of spermatozoa on this type of may negatively affect the biology of sperm cells. This can be assumed because of observed decrease of mitochondrial membrane potential and increase of the concentration of mitochondrial superoxide anion. Consequently, it can be one of mechanisms by which the long lasting exposition of men to xenoestrogens decreases fertility potential of male gametes. Abbreviations AnV-FLUOSAnnexin-V labelled with fluorescein DMSODimethyl sulfoxide E217β-estradiol ESR1Estrogen receptor type 1 ESR2Estrogen receptor type 2 ESRsEstrogen receptors Ex/EmExcitation/emission Fluo-3Fluorescence indicator of intracellular calcium GPERG protein coupled estrogen receptor JC-15,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimi-dazolylcarbocyanine iodide MitoSOX™Red mitochondrial superoxide indicator mtNOSMitochondrial nitric oxide synthase PIPropidium iodide PSTPhosphatidylserine membrane translocation ROSReactive oxygen species SDStandard deviation WHOWorld Health Organization ΔΨmMitochondrial membrane potential Acknowledgements Not applicable. Funding The National Research Centre grant NN 401 077037 funds were received in support of this work. Availability of data and materials The information of all chemicals used in the study is available in the PubChem Substance Database. Authors’ contributions KM conceived the study, participated in its design and coordination, participated in the results analysis, performed the statistical analysis, and drafted the manuscript. SI carried out the molecular studies, participated in the results analysis and drafted the manuscript. JM carried out the molecular studies. JP participated in study design and helped to draft the manuscript. All authors read and approved the final manuscript. Competing interest The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study protocol was approved by the Institutional Review Board of the Poznan University of Medical Sciences (No 119/09). All the involved patients provided written informed consent. ==== Refs References 1. Carreau S Bouraima-Lelong H Delalande C Estrogens: new players in spermatogenesis Reprod Biol Society for Biology of Reproduction 2011 11 174 193 2. Vincenzo R Bruno M Cesare C Chiara D Lucia Z Daniele S Estrogens and male reproduction 2013 South Dartmouth Endotext. MDText.com, Inc 3. Akingbemi BT Estrogen regulation of testicular function Reprod Biol Endocrinol 2005 3 51 10.1186/1477-7827-3-51 16188042 4. Adeoya-Osiguwa S Markoulaki S Pocock V Milligan SR Fraser LR 17beta-Estradiol and environmental estrogens significantly affect mammalian sperm function Hum Reprod 2003 18 100 107 10.1093/humrep/deg037 12525448 5. Ded L Dostalova P Dorosh A Dvorakova-Hortova K Peknicova J Effect of estrogens on boar sperm capacitation in vitro Reprod Biol Endocrinol 2010 8 87 10.1186/1477-7827-8-87 20626847 6. Filannino A Stout TA Gadella BM Sostaric E Pizzi F Colenbrander B Dose–response effects of estrogenic mycotoxins (zearalenone, alpha-and beta-zearalenol) on motility, hyperactivation and the acrosome reaction of stallion sperm Reprod Biol Endocrinol BioMed Central Ltd 2011 9 134 10.1186/1477-7827-9-134 7. Guido C Perrotta I Panza S Middea E Avena P Santoro M Human sperm physiology: Estrogen receptor alpha (ERα) and estrogen receptor beta (ERβ) influence sperm metabolism and may be involved in the pathophysiology of varicocele-associated male infertility J Cell Physiol 2011 226 3403 3412 10.1002/jcp.22703 21344398 8. Losel RM Falkenstein E Feuring M Schultz A Tillmann H-C Rossol-Haseroth K Nongenomic steroid action: controversies, questions, and answers Physiol Rev 2003 83 965 1016 10.1152/physrev.00003.2003 12843413 9. Wehling M Lösel R Non-genomic steroid hormone effects: Membrane or intracellular receptors? J Steroid Biochem Mol Biol 2006 102 180 183 10.1016/j.jsbmb.2006.09.016 17113980 10. Levin ER Plasma membrane estrogen receptors Trends Endocrinol Metab 2009 20 477 482 10.1016/j.tem.2009.06.009 19783454 11. Boonyaratanakornkit V Scaffolding proteins mediating membrane-initiated extra-nuclear actions of estrogen receptor Steroids Elsevier Inc 2011 76 877 884 12. Levin ER Minireview: Extranuclear steroid receptors: roles in modulation of cell functions Mol Endocrinol 2011 25 377 384 10.1210/me.2010-0284 20861220 13. Kim HP Lee JY Jeong JK Bae SW Lee HK Jo I Nongenomic stimulation of nitric oxide release by estrogen is mediated by estrogen receptor alpha localized in caveolae Biochem Biophys Res Commun 1999 263 257 262 10.1006/bbrc.1999.1348 10486286 14. Rago V Giordano F Brunelli E Zito D Aquila S Carpino A Identification of G protein-coupled estrogen receptor in human and pig spermatozoa J Anat 2014 224 732 736 10.1111/joa.12183 24697543 15. Arkoun B Gautier C Delalande C Barrier-Battut I Guenon I Goux D Stallion spermatozoa: Putative target of estrogens; presence of the estrogen receptors ESR1, ESR2 and identification of the estrogen-membrane receptor GPER Gen Comp Endocrinol Elsevier Inc 2014 200 35 43 10.1016/j.ygcen.2014.02.016 16. Aquila S Sisci D Gentile M Middea E Catalano S Carpino A Estrogen receptor (ER)α and ERβ are both expressed in human ejaculated spermatozoa: Evidence of their direct interaction with phosphatidylinositol-3-OH kinase/Akt pathway J Clin Endocrinol Metab 2004 89 1443 1451 10.1210/jc.2003-031681 15001646 17. Solakidi S Psarra A-MG Nikolaropoulos S Sekeris CE Estrogen receptors alpha and beta (ERalpha and ERbeta) and androgen receptor (AR) in human sperm: localization of ERbeta and AR in mitochondria of the midpiece Hum Reprod 2005 20 3481 3487 10.1093/humrep/dei267 16123086 18. Gavrilova-Jordan LP Price TM Actions of steroids in mitochondria Semin Reprod Med 2007 25 154 164 10.1055/s-2007-973428 17447205 19. Tavares RS Martins FC Oliveira PJ Ramalho-Santos J Peixoto FP Parabens in male infertility-Is there a mitochondrial connection? Reprod Toxicol 2009 27 1 7 10.1016/j.reprotox.2008.10.002 19007877 20. Rajender S Rahul P Mahdi AA Mitochondria, spermatogenesis and male infertility Mitochondrion 2010 10 419 428 10.1016/j.mito.2010.05.015 20595008 21. Felty Q Roy D Estrogen, mitochondria, and growth of cancer and non-cancer cells J Carcinog 2005 4 1 10.1186/1477-3163-4-1 15651993 22. Chen JQ Eshete M Alworth WL Yager JD Binding of MCF-7 cell mitochondrial proteins and recombinant human estrogen receptors alpha and beta to human mitochondrial DNA estrogen response elements J Cell Biochem 2004 93 358 373 10.1002/jcb.20178 15368362 23. Psarra AMG Sekeris CE Nuclear receptors and other nuclear transcription factors in mitochondria: Regulatory molecules in a new environment Biochim Biophys Acta-Mol Cell Res 2008 1783 1 11 10.1016/j.bbamcr.2007.10.021 24. Chen JQ Delannoy M Cooke C Yager JD Mitochondrial localization of ERalpha and ERbeta in human MCF7 cells Am J Physiol Endocrinol Metab 2004 286 E1011 E1022 10.1152/ajpendo.00508.2003 14736707 25. Breitbart H Rubinstein S Gruberger M Calcium efflux mechanism in sperm mitochondria. Biochim. Biophys. Acta-Mol Cell Res 1996 1312 79 84 26. Tatoyan A Giulivi C Purification and characterization of a nitric-oxide synthase from rat liver mitochondria J Biol Chem 1998 273 11044 11048 10.1074/jbc.273.18.11044 9556587 27. Koppers AJ De Iuliis GN Finnie JM McLaughlin EA Aitken RJ Significance of Mitochondrial Reactive Oxygen Species in the Generation of Oxidative Stress in Spermatozoa J Clin Endocrinol Metab 2008 93 3199 3207 10.1210/jc.2007-2616 18492763 28. Paoli D Gallo M Rizzo F Baldi E Francavilla S Lenzi A Mitochondrial membrane potential profile and its correlation with increasing sperm motility Fertil Steril 2011 95 2315 2319 10.1016/j.fertnstert.2011.03.059 21507394 29. Marchetti P Ballot C Jouy N Thomas P Marchetti C Influence of mitochondrial membrane potential of spermatozoa on in vitro fertilisation outcome Andrologia 2012 44 136 141 10.1111/j.1439-0272.2010.01117.x 21714802 30. Paventi G Lessard C Bailey JL Passarella S In boar sperm capacitation l-lactate and succinate, but not pyruvate and citrate, contribute to the mitochondrial membrane potential increase as monitored via safranine O fluorescence Biochem Biophys Res Commun 2015 462 257 262 10.1016/j.bbrc.2015.04.128 25956060 31. Horvat A Nikezić G Petrović S Kanazir DT Nikezie G Petrovic S Binding of estradiol to synaptosomal mitochondria: physiological significance Cell Mol Life Sci 2001 58 636 644 10.1007/PL00000886 11361096 32. Petrović S Demajo M Horvat A Estradiol affects calcium transport across mitochondrial membrane in different brain regions Ann N Y Acad Sci 2005 1048 341 343 10.1196/annals.1342.032 16154947 33. Bennetts LE De Iuliis GN Nixon B Kime M Zelski K McVicar CM Impact of estrogenic compounds on DNA integrity in human spermatozoa: Evidence for cross-linking and redox cycling activities Mutat Res-Fundam Mol Mech Mutagen 2008 641 1 11 10.1016/j.mrfmmm.2008.02.002 34. Mishra DP Shaha C Estrogen-induced spermatogenic cell apoptosis occurs via the mitochondrial pathway: role of superoxide and nitric oxide J Biol Chem 2005 280 6181 6196 10.1074/jbc.M405970200 15545282 35. Gharagozloo P Aitken RJ The role of sperm oxidative stress in male infertility and the significance of oral antioxidant therapy Hum Reprod 2011 26 1628 1640 10.1093/humrep/der132 21546386 36. Kim SH Yu DH Kim YJ Apoptosis-like change, ROS, and DNA status in cryopreserved canine sperm recovered by glass wool filtration and Percoll gradient centrifugation techniques Anim Reprod Sci 2010 119 106 114 10.1016/j.anireprosci.2009.11.002 19942382 37. Aitken RJ Gibb Z Baker MA Drevet J Gharagozloo P Causes and consequences of oxidative stress in spermatozoa Reprod Fertil Dev 2015 28 1 10 10.1071/RD15325 27062870 38. Kotwicka M Filipiak K Jedrzejczak P Warchol JB Caspase-3 activation and phosphatidylserine membrane translocation in human spermatozoa: is there a relationship? Reprod Biomed Online 2008 16 657 663 10.1016/S1472-6483(10)60479-8 18492369 39. Kotwicka M Jendraszak M Skibinska I Jedrzejczak P Pawelczyk L Decreased motility of human spermatozoa presenting phosphatidylserine membrane translocation-cells selection with the swim-up technique Hum Cell 2013 26 28 34 10.1007/s13577-011-0024-1 21725868 40. Grunewald S Baumann T Paasch U Glander HJ Capacitation and acrosome reaction in nonapoptotic human spermatozoa Ann N Y Acad Sci 2006 1090 138 146 10.1196/annals.1378.015 17384256 41. Grunewald S Kriegel C Baumann T Glander HJ Paasch U Interactions between apoptotic signal transduction and capacitation in human spermatozoa Hum Reprod 2009 24 2071 2078 10.1093/humrep/dep178 19443455 42. Psarra AMG Solakidi S Sekeris CE The mitochondrion as a primary site of action of steroid and thyroid hormones: Presence and action of steroid and thyroid hormone receptors in mitochondria of animal cells Mol Cell Endocrinol 2006 246 21 33 10.1016/j.mce.2005.11.025 16388892 43. Lu A Frink M Choudhry M Schwacha MG Hubbard WJ Rue LW Mitochondria play an important role in 17beta-estradiol attenuation of H (2) O (2)-induced rat endothelial cell apoptosis Am J Physiol Endocrinol Metab 2007 292 E585 E593 10.1152/ajpendo.00413.2006 17018771 44. Wang J Green PS Simpkins JW Estradiol protects against ATP depletion, mitochondrial membrane potential decline and the generation of reactive oxygen species induced by 3-nitroproprionic acid in SK-N-SH human neuroblastoma cells J Neurochem 2001 77 804 811 10.1046/j.1471-4159.2001.00271.x 11331409

PMC005xxxxxx/PMC5002131.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 66610.1186/s12974-016-0666-8ResearchHigh-fat diet feeding differentially affects the development of inflammation in the central nervous system Guillemot-Legris Owein owein.guillemot@uclouvain.be 1Masquelier Julien julien.masquelier@uclouvain.be 1Everard Amandine amandine.everard@uclouvain.be 2Cani Patrice D. patrice.cani@uclouvain.be 2Alhouayek Mireille mireille.alhouayek@uclouvain.be 1Muccioli Giulio G. +32 2 764 7231giulio.muccioli@uclouvain.be 11 Bioanalysis and Pharmacology of Bioactive Lipids Research Group, Louvain Drug Research Institute (LDRI), Université catholique de Louvain (UCL), Av. E.Mounier, 72 (B1.72.01), 1200 Brussels, Belgium 2 Metabolism and Nutrition Research Group, WELBIO - Walloon Excellence in Life Sciences and BIOtechnology, Louvain Drug Research Institute, Université catholique de Louvain, Brussels, Belgium 26 8 2016 26 8 2016 2016 13 1 20620 6 2016 18 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 Obesity and its associated disorders are becoming a major health issue in many countries. The resulting low-grade inflammation not only affects the periphery but also the central nervous system. We set out to study, in a time-dependent manner, the effects of a high-fat diet on different regions of the central nervous system with regard to the inflammatory tone. Methods We used a diet-induced obesity model and compared at several time-points (1, 2, 4, 6, 8, and 16 weeks) a group of mice fed a high-fat diet with its respective control group fed a standard diet. We also performed a large-scale analysis of lipids in the central nervous system using HPLC-MS, and we then tested the lipids of interest on a primary co-culture of astrocytes and microglial cells. Results We measured an increase in the inflammatory tone in the cerebellum at the different time-points. However, at week 16, we evidenced that the inflammatory tone displayed significant differences in two different regions of the central nervous system, specifically an increase in the cerebellum and no modification in the cortex for high-fat diet mice when compared with chow-fed mice. Our results clearly suggest region-dependent as well as time-dependent adaptations of the central nervous system to the high-fat diet. The differences in inflammatory tone between the two regions considered seem to involve astrocytes but not microglial cells. Furthermore, a large-scale lipid screening coupled to ex vivo testing enabled us to identify three classes of lipids—phosphatidylinositols, phosphatidylethanolamines, and lysophosphatidylcholines—as well as palmitoylethanolamide, as potentially responsible for the difference in inflammatory tone. Conclusions This study demonstrates that the inflammatory tone induced by a high-fat diet does not similarly affect distinct regions of the central nervous system. Moreover, the lipids identified and tested ex vivo showed interesting anti-inflammatory properties and could be further studied to better characterize their activity and their role in controlling inflammation in the central nervous system. Electronic supplementary material The online version of this article (doi:10.1186/s12974-016-0666-8) contains supplementary material, which is available to authorized users. Keywords ObesityBioactive lipidAstrocyteMicrogliaHPLC-MSCortexCerebellumPEAhttp://dx.doi.org/10.13039/501100002661Fonds De La Recherche Scientifique - FNRSCC 1.5.034.10 - FRFC 2.4555.08 - J.0160.13Muccioli Giulio G. http://dx.doi.org/10.13039/501100000781European Research Council336452-ENIGMOCani Patrice D. FRFS-WELBIOWELBIO-CR-2012S-02RCani Patrice D. issue-copyright-statement© The Author(s) 2016 ==== Body Background Obesity and related disorders are becoming worldwide health issues [1–3]. Obesity is regarded as an inflammatory condition because of the associated low-grade inflammation [4–6] affecting the periphery and increasing the incidence of many pathologies such as cardiovascular diseases [7], asthma [8], or even cancer [9]. One of the proposed mechanisms leading to peripheral inflammation implicates the gut microbiota. More specifically, a high-fat diet (HFD) will change the balance between different populations of bacteria within the gut [10, 11]. This will lead to a disruption of the intestinal epithelium integrity that in turn will result in increased passage of endotoxins (such as lipopolysaccharides (LPS)) into the bloodstream that will then fuel the peripheral inflammatory tone [4, 12, 13]. The demonstration that disrupting LPS signaling (i.e., TLR4−/− mice or CD14−/− mice) protects from diet-induced obesity and metabolic disorders strongly supports the important role played by LPS in the pathophysiology of these disorders [12–15]. This increased peripheral inflammatory tone will also affect the central nervous system (CNS) and will increase the incidence of CNS pathologies such as cognitive impairments [16], Alzheimer’s disease [17], stroke [18], or dementia [19]. The impact of a HFD on the CNS was well characterized with regard to a specific region, the hypothalamus [20–22]. The hypothalamus has attracted the attention of many researchers because of its central role in food intake as well as in monitoring the availability of nutrients [23–25]. HFD feeding is associated with a disruption of the homeostasis in the hypothalamus, and more specifically with the activation of glial cells and increased inflammatory tone [24, 26, 27]. This leads to both leptin and insulin resistance thus worsening obesity [22]. However, much less is known about the repercussions of a HFD on the other regions of the CNS in terms of inflammation. Obesity and inflammation are closely related to lipids and their metabolism. Indeed, HFD feeding will lead to an increase in the intake of saturated fatty acids [28] and to the disruption of cholesterol homeostasis (an increase in LDL to HDL cholesterol ratio) [29, 30], both associated with deleterious effects. The adipose tissue will have to cope with an increased flow of free fatty acids that will trigger a low-grade inflammation through immunomodulatory changes of both specific T cell subtypes and macrophage polarization [31, 32]. The perception of lipids has dramatically changed from being mere energy substrate molecules to bioactive molecules involved in many physiological processes notably through the emergence of the lipidomic approach [33]. Lipids are recognized as central mediators involved in the onset, development, and resolution of inflammatory processes [34, 35]. Obesity alters the endogenous levels of several bioactive lipid families such as ceramides, phosphatidylcholines, and endocannabinoids [36–38]. In turn, some bioactive lipids exert either pro- or anti-inflammatory effects during obesity. For instance, ceramides will exert pro-inflammatory effects in the liver and will progressively lead to insulin resistance by tampering with the insulin signaling [6, 39, 40]. Conversely, n−3 polyunsaturated fatty acids show beneficial effects by counteracting HFD-induced adipose tissue inflammation [41]. Still, the potential involvement of other lipids needs to be addressed to better characterize the inflammatory tone deriving from obesity. In this study, we set out to characterize, at multiple time-points and in different CNS regions, the inflammatory tone induced by a HFD. We found that, depending on the CNS region, a HFD differentially affects the inflammatory tone. We, therefore, investigated whether changes in CNS lipid content could explain the differences in the inflammatory tone between CNS regions. Methods Animals and diets Nine-week-old male C57BL/6J mice (Charles River) were housed in a controlled environment (12-h day light cycle, lights off at 6 pm, controlled temperature and humidity). Upon arrival, they were randomly split into 12 groups of eight mice each (four mice/cage) and acclimated for 1 week. Then, six of these groups were given free access to a standard diet (AIN 93-M, Research Diets, New Brunswick, USA) and the remaining six groups were given free access to a HFD (D12492, Research Diets, New Brunswick, USA). For details in the composition of both diets, refer to Additional file 1: Table S1. For this experiment, we euthanized at each selected time-point (i.e., after 1, 2, 4, 6, 8, and 16 weeks) one group under standard diet and one group under a HFD. Mice were anesthetized using isoflurane after a 6-h fasting period and sacrificed by cervical dislocation. The cortex, cerebellum, and brainstem were carefully and rapidly recovered and snap-frozen in liquid nitrogen. The different adipose tissue depots (subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), epididymal adipose tissue (EAT), and brown adipose tissue (BAT)) were harvested and weighed. All the tissues collected were stored at −80 °C until further analysis. We performed this study in accordance with the European recommendation 2007/526/CE (which was transformed into the Belgian Law of May 29, 2013), regarding the protection of laboratory animals. The local ethics committee approved the protocol of the study (study agreement 2010/UCL/MD/022; lab agreement LA1230314). Cholesterol quantification Plasma total cholesterol was quantified, following manufacturer’s instructions, in the vena cava using the Cholesterol FS10 kit (DiaSys Diagnostic and Systems, Holzheim, Germany), which is based on an enzymatic reaction coupled with a spectrophotometric detection of the end-product. RNA preparation and RT-qPCR analysis Total RNA from tissues was extracted using TriPure reagent (Roche, Basel, Switzerland) according to the manufacturer’s instructions. cDNA was synthesized using an RT kit (Promega, GoScript™ Reverse Transcription System) from 1 μg of total RNA. qPCR was performed with a StepOnePlus instrument and software (Applied Biosystems, Foster City, CA, USA). PCR reactions were run using a SYBR Green mix (Promega, GoTaq® qPCR Master Mix). We measured each sample in duplicate during the same run. The following conditions were used for amplification: an initial holding stage of 10 min at 95 °C, then 45 cycles consisting of denaturation at 95 °C for 3 s, annealing at 60 °C for 26 s, and extension at 72 °C for 10 s. Products were analyzed by performing a melting curve at the end of the PCR reaction. Data are normalized to the 60S ribosomal protein L19 (RPL19) messenger RNA (mRNA) expression [42]. The sequences of the primers used are listed in Table 1.Table 1 Primer sequences Gene Forward primer (5′–3′) Reverse primer (5′–3′) CD11b GAACATCCCATGACCTTCCA GCTGGGGGACAGTAGAAACA CD11c ACGTCAGTACAAGGAGATGTTGGA ATCCTATTGCAGAATGCTTCTTTACC CD68 CTTCCCACAGGCAGCACAG AATGATGAGAGGCAGCAAGAGG Claudin 1 TTCGCAAAGCACCGGGCAGATACA GCCACTAATGTCGCCAGACCTGAAA Claudin 5 GTTAAGGCACGGGTAGCACT GTACTTCTGTGACACCGGCA COX-2 TGACCCCCAAGGCTCAAATAT TGAACCCAGGTCCTCGCTTA F4/80 TGACAACCAGACGGCTTGTG CAGGCGAGGAAAAGATAG GFAP TTCGCACTCAATACGAGGCA CTCCAGATCGCAGGTCAAG IL-1β TCGCTCAGGGTCACAAGAAA CATCAGAGGCAAGGAGGAAAAC IL-6 ACAAGTCGGAGGCTTAATTACACAT TTGCCATTGCACAACTCTTTTC iNOS AGGTACTCAGCGTGCTCCAC GCACCGAAGATATCTTCATG LBP AGTCCTGGGAATCTGTCCTTG ACTTGTGCCTTGTCTGGATG MCP-1 GCAGTTAACGCCCCACTCA TCCAGCCTACTCATTGGGATCA Occludin ATGTCCGGCCGATGCTCTC TTTGGCTGCTCTTGGGTCTGTAT RPL19 TGACCTGGATGAGAAGGATGAG CTGTGATACATATGGCGGTCAATC Serpina3n GGACATTGATGGTGCTGGTGAAT CTCCTCTTGCCCGCGTAGAA TNFα CCACCACGCTCTTCTGTCT TCCAGCTGCTCCTCCACTT ZO-1 TTTTTGACAGGGGGAGTGG TGCTGCAGAGGTCAAAGTTCAAG Lipid quantification Tissues (cerebellum or cortex) were homogenized in water (2.5 mL), and then the lipids were extracted following acidification, in the presence of internal standards, by adding 10 mL of chloroform (CHCl3) and 5 mL of methanol (MeOH). Following vigorous mixing and sonication, the samples were centrifuged and the organic layer was recovered and dried under a stream of N2. The resulting lipid extracts were purified by solid-phase extraction using silica and eluted with a mix of CHCl3 and MeOH. The resulting lipid fractions were analyzed by HPLC-MS using an LTQ-Orbitrap mass spectrometer (ThermoFisher Scientific) coupled to an Accela HPLC system (ThermoFisher Scientific). Analyte separation was achieved using a C-18 Phenomenex pre-column and a Kinetex LC-18 column (5 μm, 4.6 × 150 mm) (Phenomenex). For lysophosphosphatidylcholines, phosphatidylcholines, and sphingomyelins, mobile phases A and B were composed of MeOH-H2O 85:15 (v/v) and MeOH, respectively, with 5 mM of CH3COONH4. The gradient (0.25 mL/min) was designed as follows: transition from 100 % A to 100 % B over 15 min, followed by 100 % B linearly over 15 min followed by a subsequent re-equilibration at 100 % A. Analytes were ionized using an ESI source operated in positive mode. For the other lysophosphospholipids and for the phospholipids and sulfatides, mobile phases A and B were composed of MeOH-H2O-NH4OH 50:50:0.1 (v/v/v) and MeOH-NH4OH 100:0.1 (v/v), respectively. The gradient (0.4 mL/min) was designed as follows: transition from 100 % A to 100 % B over 30 min, followed by 100 % B linearly over 15 min, and followed by a subsequent re-equilibration at 100 % A. Analytes were ionized using an ESI source operated in negative mode. For N-acylethanolamines and ceramides, mobile phases A and B were composed of MeOH-H2O-acetic acid 75:25:0.1 (v/v/v) and MeOH-acetic acid 100:0.1 (v/v), respectively. The gradient (0.4 mL/min) was designed as follows: transition from 100 % A to 100 % B over 15 min, followed by 100 % B linearly over 45 min, followed by a subsequent re-equilibration at 100 % A. Analytes were ionized using an APCI source operated in positive mode. The signals of the lipids were normalized using the signal obtained for the corresponding internal standard. We used d4-PEA, 17:1-lysophosphatidylinositol, 17:0-lysophosphatidylcholine, 17:0-sulfatide, 17:0-ceramide, 17:0-sphingomyelin, and 17:0/17:0-PC. Data are presented as fold increase compared with levels found in control mice. Immunohistology During the sacrifice, sections of the cortex and cerebellum were transferred to a solution of 4 % PFA in PBS and stored at 4 °C for 24 h. Cryopreservation was performed by incubation in a solution of 20 % sucrose in PBS for a further 24 h at 4 °C. Finally, tissues were embedded in Tissue-Tek (Sakura Finetek, Zoeterwoude, The Netherlands) and kept at −80 °C. Sections were cut (30 μm) using a cryostat and then used for the detection of microglial cells (Iba-1) and astrocytes (GFAP). The sections were incubated in blocking solution containing 5 % normal donkey serum and 1 % Triton X-100 (Sigma-Aldrich, Seelze, Germany) in PBS for 60 min. The primary antibodies, rabbit anti-Iba1 (Wako Laboratory Chemicals, Japan) (1:1000 in PBS/triton 1 %) and direct rat anti-GFAP (1:250 in PBS/triton 1 %), were applied for 12 h at 4 °C. Tissues were then rinsed three times with PBS. The secondary antibody anti-rabbit Alexa 488 (Thermo Fisher Scientific) (1:100) was applied for 1 h at room temperature. Tissues were washed with PBS and nuclei were stained using Hoescht. Slides were mounted using Dako Fluorescence Mounting Medium. Stained slides were digitized using a Mirax Midi scanner (Carl Zeiss Micro-Imaging). Image acquisition was executed with Mirax Scan software (Zeiss). The obtained images were analyzed (by a researcher blinded to the treatment) using ImageJ software (http://imagej.nih.gov/ij/) and/or CellProfiler software (http://www.cellprofiler.org/). Primary glial cell culture and treatment C57BL/6J mice pups (post-natal day 2–3) were euthanized, the brain recovered, and their cerebral cortices dissected. Tissues were then mechanically dissociated by several sequences of pipetting and sedimenting, then centrifuged and resuspended in DMEM-F12 media (containing 10 % FBS and 100 units/mL of penicillin and 100 μg/mL of streptomycin). Cells were seeded in poly-lysine pre-coated flasks (two pups per flask) and incubated for 2 weeks with two media changes at day 5 and 10. After 14 days of culture, cells were trypsinized and secondary cultures were seeded overnight in poly-lysine pre-coated 24-well plates (150,000 cells/well). Cells were then incubated with fresh culture medium containing the compounds of interest (10 μM), and LPS (10 ng/mL, from E. coli 055:B5) was added 1 h later. After 8 h, the media was removed and Tripure© was added to the cells for mRNA analysis (see above). For all experiments, a control condition was performed where cells were only incubated with vehicle (DMSO, 0.2 %) in the absence of LPS. Inflammatory plasma cytokine quantification Plasma cytokines IL-1β, IL-10, and tumor necrosis factor α (TNFα) were quantified using a Bio-Plex Multiplex kit (Bio-Rad, Nazareth, Belgium) and measured by using Luminex technology (Bio-Plex 200; Bio-Rad) following the manufacturer’s instructions. Statistical analysis All data are presented as mean ± s.e.m. Statistical analysis was performed using GraphPad Prism version 5.0 for Windows (San Diego, CA). We used two-tailed Student’s t test for unpaired values to compare two groups, and when relevant, we used the Mann-Whitney test for the peripheral inflammation assessment. We used one-way ANOVA with Bonferroni’s post test or Kruskal-Wallis test with Dunn’s post test between HFD group and its respective CTL group (*P < 0.05; **P < 0.01; and ***P < 0.001) and between CNS regions (#P < 0.05; ##P < 0.01; and ###P < 0.001) for the comparison of inflammatory markers, immunohistological analysis, and the lipid levels for the two CNS areas studied. Finally, we used the one-way ANOVA test with Dunnett’s post test for the ex vivo experiments. For all statistical tests, statistical significance was taken when P < 0.05. Results and discussion Characterization of the obese phenotype We monitored the weight of mice for the different groups throughout the study. Body weight already increased after 4 days of HFD feeding and persistently increased over time. Conversely, the mice of the control groups did not significantly gain weight (Additional file 2: Figure SI 1). We also weighed the different white adipose tissue depots at each selected time-point. There is a sustained increase of the subcutaneous (SAT), epididymal (EAT), and visceral (VAT) adipose tissues over time (Additional file 2: Figure SI 2). We also measured the cholesterolemia of mice at the earliest (1 week) and latest (16 weeks) time-points. We found a clear increase for mice fed a HFD compared with their respective controls (Additional file 2: Figure SI 3a-b). Taken together, these data validate the obesogenic properties of the diet used. Peripheral inflammation induced by HFD feeding Because obesity is accompanied by a low-grade peripheral inflammation, we next sought to study the impact of the HFD on the peripheral inflammatory tone. Thus, we measured the mRNA expression of different inflammatory markers in the SAT at weeks 1 and 16 (Fig. 1a, b).. In this tissue, as might be expected for an obesogenic diet, after 16 weeks of HFD, we measured a significant increase in the expression of F4/80 (a macrophage marker), CD11c (a M1 polarization marker), LPS-binding protein (LBP), and interleukin-6 (IL-6) (Fig. 1b). These observations confirm the establishment and progression of an inflammatory tone induced by the HFD in the SAT. Because the liver is another organ that can be affected by inflammation in obese conditions, we also measured the expression of CD68, CD11c, and IL-6 in the liver of mice after 1 and 16 weeks of HFD (Fig. 1c, d). We found a significant increase in the expression of CD11c and IL-6 at 16 weeks (Fig. 1c, d). Finally, we measured the concentration of two pro-inflammatory cytokines (IL-1β and TNFα) and one anti-inflammatory cytokine (IL-10) in the plasma of mice at week 16 and found a significant increase of TNFα for mice fed a HFD compared with chow-fed mice (Additional file 2: Figure SI 3c). Taken together, these data validate the pro-inflammatory effects, in the periphery, of the obesogenic diet used.Fig. 1 Peripheral inflammatory tone induced by a HFD at weeks 1 and 16. mRNA relative expression of inflammatory and macrophage markers at a 1 week and b 16 weeks in the subcutaneous adipose tissue (SAT) and mRNA relative expression of inflammatory and macrophage markers at c 1 week and d 16 weeks in the liver. Results are expressed relative to the control diet group (CTL) set at 100 %. Data are mean ± s.e.m; Student’s t test or Mann-Whitney test between HFD group and its CTL group (*P < 0.05 and **P < 0.01) Central inflammation induced by HFD feeding Obesity is a well-established contributing factor increasing the incidence of peripheral pathologies. It is also well demonstrated that obesity induces inflammation in the hypothalamus [21, 22, 27]. However, much less is known about the effects of obesity on other CNS areas. We thought that two regions, the cerebellum and the cortex, were of particular interest because obesity induces morphological changes in these two areas. Indeed, obese patients display differences in gray matter density in these two specific regions when compared with lean subjects [43, 44]. Early onset obesity is also associated with several cerebellar abnormalities such as neuronal injuries, smaller volume, and compromised development [45, 46]. As for the cortex, it is an area responsible for the cognitive control of food intake [44, 47, 48]. Those specific changes could be either a cause or a consequence of obesity and further maintain dysregulations in food-oriented behaviors. In obesity settings, one major inflammatory pathway affected is the one involving the nuclear factor kappa B (NF-kB). Indeed, this transcription factor is considered as pivotal in the inflammatory tone deriving from obesity, and its involvement in the etiology of metabolic disorders has also been established [49, 50]. Therefore, we set out to study the expression of downstream genes of the NF-kB pathway comprising two cytokines, IL-1β and TNFα as well as the chemokine MCP-1 (known as monocyte chemoattractant protein-1) and the inducible enzyme cyclooxygenase-2 (COX-2). To characterize the changes in inflammatory tone in the cerebellum and cortex during the development of obesity, we measured the mRNA expression of these four inflammatory markers at the six selected time-points of our HFD study (Fig. 2). We found a marked increase of these inflammatory markers in the cerebellum as early as 1 week after HFD feeding. This early inflammation was not present in the cortex. Furthermore, after 16 weeks of HFD feeding, the expression of IL-1β, TNFα, and COX-2 was strikingly different between the two CNS regions considered. Indeed, we found a marked inflammation in the cerebellum of HFD-fed mice when compared with the cortex of the same mice where the inflammatory tone was similar to that of the CTLs.Fig. 2 Central inflammatory tone induced by a HFD at different time-points. mRNA expression of inflammatory markers at the different time-points in the cerebellum (CBL) and in the cortex (CTX) at a 1 week, b 2 weeks, c 4 weeks, d 6 weeks, e 8 weeks, and f 16 weeks. Data are mean ± s.e.m. The standard diet groups were set at 1. The white columns represent mice fed a standard diet and the black columns represent the mice fed a HFD. One-way ANOVA with Bonferroni’s post test or Kruskal-Wallis test with Dunn’s post test between HFD group and its respective CTL group (*P < 0.05; **P < 0.01; and ***P < 0.001) and between CNS regions (#P < 0.05; ##P < 0.01; and ###P < 0.001) Because the consequences of the HFD feeding on inflammation were clearly different in the cortex and cerebellum, we sought to determine whether this decreased inflammatory tone was specific to the cortex. Thus, we measured the inflammatory tone in the brainstem, at weeks 1 and 16, and we found an increased expression of IL-1β, TNFα, and MCP-1 at both time-points (Additional file 2: Figure SI 4), a profile similar to the one observed in the cerebellum. Thus, these results clearly suggest region-dependent adaptations of the CNS to the HFD. To explain these observations, we first assessed the expression of zonula occludens (ZO)-1, claudin 1 and 5, and occludin, four tight junction-forming proteins of the blood-brain barrier [51, 52]. Indeed, the integrity of the blood-brain barrier has been shown to be altered during obesity [53]. We found no variation in the expression of ZO-1 and claudin 5 in the cortex or cerebellum at week 16. However, claudin 1 and occludin expression were significantly increased in the cerebellum, whereas no variation was measured in the cortex at week 16 (Additional file 2: Figure SI 5). Of note, the mRNA expression of these proteins was shown to display consistent variations with proteins detected through immunohistochemistry [54]. We next sought to characterize further the differences between the cerebellum and the cortex after 16 weeks of HFD feeding by studying the activation state of the microglial cells and astrocytes, two major players in inflammatory processes of the CNS [55, 56]. These cells are known to be involved in the inflammatory processes induced by HFD feeding in the hypothalamus [26, 57]. To this end, we performed an immunohistological analysis of these two cell types in the cerebellum and the cortex of the mice after 16 weeks of HFD feeding. Iba-1 (ionized calcium-binding adapter molecule 1) immunostaining showed no differences in microglial cell activation between the control and the HFD groups, neither in the cerebellum nor in the cortex (Fig. 3a, b). These data are supported by the fact that the mRNA expression of CD11b and CD11c were not affected by the HFD (Fig. 3c, d). On the other hand, immunostaining for astrocytes performed using anti-GFAP (glial fibrillary acidic protein) antibodies showed a larger area occupied by astrocytes in the cerebellum upon HFD feeding, indicating an activated state, an outcome not present in the cortex (Fig. 3e, f). This was further supported by the enhanced GFAP mRNA expression induced by the HFD in the cerebellum compared with CTL mice as well as compared with the cortex of the same animal (Fig. 3g). Finally, to confirm the astrocytes’ state of activation in the cerebellum and the cortex, we studied the expression of an astrogliosis marker Serpina3n (the serine peptidase inhibitor clade A member 3N, also known as alpha-1 antitrypsin) [58, 59]. We found a significant increase of this marker in the cerebellum of HFD mice compared with CTL mice and compared with the cortex of the same mice at week 16 (Fig. 3h). Interestingly, we found no variation of these two astrocyte markers in the cortex between CTL and HFD-fed mice. Our observations seem to point towards the involvement of astrocytes and astrogliosis in the differential inflammatory tone measured in these two CNS regions.Fig. 3 Activation state of glial cells in the cortex (CTX) and the cerebellum (CBL) at week 16. a Representative photomicrographs of the cerebellum and cortex immunostained for microglial cells with Iba-1. b Quantification of the area occupied by microglial cells in the cortex and the cerebellum of CTL (white column) and HFD (black column) mice at week 16. Microglial cells were identified using immunohistology in both CNS structures with Iba-1. mRNA expression of c CD11b and d CD11c, both myeloid lineage markers in the cortex and cerebellum of HFD mice at 16 weeks. e Representative photomicrographs of the cortex and cerebellum immunostained for astrocytes with GFAP. f Quantification of the area occupied by astrocytes in the cerebellum and the cortex of CTL (white column) and HFD (black column) mice at week 16. Astrocytes were identified using immunohistology in both CNS structures with GFAP. mRNA expression of g GFAP and h Serpina3n (an astrogliosis marker) in the cortex and the cerebellum of CTL and HFD mice at week 16. Data are mean ± s.e.m. The standard diet groups were set at 100% for area quantification and at 1 for mRNA expression. The white columns represent mice fed a standard diet and the black columns represent the mice fed a HFD. One-way ANOVA with Bonferroni’s post test or Kruskal-Wallis test with Dunn’s post test between HFD group and its respective CTL group (*P < 0.05) and between CNS regions (#P < 0.05 and ##P < 0.01). Scale bar 50 μm HFD feeding alters lipid levels in the cerebellum and the cortex at week 16 As mentioned in the introduction, lipids are involved in the control of inflammation. Thus, we decided to perform a broad analysis of the lipids present in the cortex and cerebellum focusing on week 16 because mice displayed a distinct inflammatory tone at this specific time-point. We decided to focus our investigations on ceramides, dihydroceramides, sphingomyelins, sulfatides, and N-acylethanolamines as well as phospholipids and lysophospholipids because they are known to be involved in inflammation [60–64]. Globally, the ceramide and dihydroceramide species measured here displayed no variation in the cortex or in the cerebellum. Regarding the sphingomyelin species, their levels were increased in the cortex and showed no variation in the cerebellum. Among the phospholipids and lysophospholipids studied, we found that the levels of phosphatidylinositols, phosphatidylethanolamines, and lysophosphatidylcholines were increased in the cortex upon HFD feeding, while the HFD had much less effect on their levels in the cortex. Of note, we also found that the levels of the anti-inflammatory palmitoylethanolamide were increased in the cortex, but not in the cerebellum, upon HFD feeding (Fig. 4 and Additional file 1: Table S2-4).Fig. 4 Large-scale lipid screening in the cerebellum (CBL) and the cortex (CTX) at week 16. Phospholipids, lysophospholipids, ceramides, sphingomyelins, N-acylethanolamines, dihydroceramides, and sulfatides were measured in the cortex and cerebellum of CTL and HFD mice at week 16 by HPLC-MS. The color red represents an increase in lipid levels in the HFD mice compared with the CTL mice. The color green represents a decrease in lipid levels in the HFD mice compared with the CTL mice. Numerical data are reported in Additional file 1: Table S2-4. ND not detected. For phospholipids, the sum of R1 and R2 acyl chains are indicated, whereas for the other lipids the R acyl chain is mentioned Effects of the identified lipids on primary co-cultured astrocytes and microglia Our HPLC-MS analysis allowed us to identify several lipid classes differently affected by the HFD in the cortex and cerebellum. To determine which of these lipids could potentially be involved in the reduced inflammatory tone found in the cortex, we selected the lipids increased under a HFD in the cortex but decreased (or not affected) in the cerebellum. The lipids that fulfill these criteria are sphingomyelins, phosphatidylinositols, phosphatidylethanolamines, lysophosphatidylcholines, and palmitoylethanolamide. To determine whether these lipids could be in part responsible for the reduced inflammatory tone in the cortex, we tested their effect on microglia–astrocyte-mixed cultures activated by LPS. Incubation of the primary microglia–astrocyte-mixed cultures with LPS induced a strong increase in the expression of inflammatory markers (Fig. 5). While sphingomyelins had no effect on the inflammatory markers, we found that the phosphatidylethanolamines were able to decrease IL-1β expression while they increased IL-6 and MCP-1 expression. Lysophosphatidylcholines were only able to decrease MCP-1 expression while phosphatidylinositols were able to reduce the LPS-induced increase in mRNA expression of IL-1β, IL-6, and MCP-1 (Fig. 5a–c). Interestingly, palmitoylethanolamide was also able to reduce the LPS-induced expression of these three inflammatory markers in these cells. Although palmitoylethanolamide is a known anti-inflammatory and neuroprotective bioactive lipid, this is one of the first reports of the effect of phosphatidylinositols, phosphatidylethanolamines, and lysophosphatidylcholines in such settings. The use of a microglia–astrocyte-mixed culture allowed us to circumvent the bias due to metabolism and blood-brain barrier crossing of the tested lipids. However, additional studies are needed to further support the in vivo role of these lipids in the HFD-induced effects in the cortex.Fig. 5 Ex vivo testing of identified lipids. mRNA expression of a IL-1β, b MCP-1, c IL-6, and d iNOS in primary co-culture of astrocytes and microglia incubated with sphingomyelins (SM), phosphatidylinositols (PI), phosphatidylethanolamines (PE), lysophosphatidylcholines (LPC), or palmitoylethanolamide (PEA) prior to LPS activation. Results are expressed relative to the CTL group set at 100 %. The dotted line represents the levels obtained for the vehicle-treated LPS-activated group. Results are expressed as mean ± s.e.m. n = 2 in triplicate. One-way ANOVA with Dunnett’s post test. *P < 0.05; **P < 0.01; and ***P < 0.001 vs vehicle-treated LPS-activated group Conclusions In this study, we showed that the peripheral low-grade inflammation induced by a HFD does not affect the different regions of the CNS in the same way and that this inflammatory tone is also time-dependent. In this particular setting, we were able to identify the potential involvement of glial cells and, more precisely, astrocytes. Interestingly, these CNS inflammatory cells are also involved in the control of the blood-brain barrier permeability. In the cerebellum, we found activated astrocytes and increased expression of claudin 1 and occludin. This increased expression of tight junction proteins could be a potential mechanism aiming at restoring the blood-brain barrier integrity in order to reduce the inflammatory insult evidenced in this CNS area. These changes were absent in the cortex where the inflammatory tone was similar to the one of chow-fed mice. These findings further support that the low-grade inflammatory tone resulting from a HFD differentially affects the two specific regions of the CNS studied. We further characterized the specific micro-environment in these two CNS areas by measuring levels of bioactive lipids. Upon testing on primary co-culture of microglia and astrocytes, we identified phosphatidylinositols, lysophosphatidylcholines, and PEA as potential anti-inflammatory compounds as they were increased in the cortex of HFD mice at week 16 and were also able to decrease the expression of pro-inflammatory markers ex vivo. This study demonstrates that not all CNS regions are equal when facing obesity-driven inflammatory insults. Finally, this work paves the way for further research revolving around the effects of the identified lipids as anti-inflammatory compounds in obesity and other inflammatory settings influencing the CNS. Additional files Additional file 1: Table S1. Composition of the diets used in the diet-induced obesity model. Table S2. High-fat diet-induced changes in phospholipid and lysophospholipid cortical levels. Table S3. High-fat diet-induced changes in phospholipid and lysophospholipid cerebellar levels. Table S4. High-fat diet-induced changes in lipid levels. (PDF 354 kb) Additional file 2: Figure SI 1. Weight of the mice on standard (empty dots) and high-fat (full dots) diets in the different groups after (a) 1 week, (b) 2 weeks, (c) 4 weeks, (d) 6 weeks, (e) 8 weeks, and (f) 16 weeks. Data are mean ± s.e.m.; two-way ANOVA with post hoc Bonferroni test between HFD group and its CTL group *P < 0.05; **P < 0.01; and ***P < 0.001. Figure SI 2. Weight of the (a) subcutaneous adipose tissue, (b) epididymal adipose tissue, and (c) visceral adipose tissue at the selected time-points. Data are mean ± s.e.m.; Student’s t test or Mann-Whitney test between each HFD group (Black columns) and its respective CTL group (white columns) *P < 0.05; **P < 0.01; and ***P < 0.001. Figure SI 3. Total plasma cholesterol measured at (a) week 1 and (b) week 16. (c) Inflammatory markers measured in the plasma at week 16. Data are mean ± s.e.m.; Student’s t test between the HFD group and its CTL group *P < 0.05; ***P < 0.001. Figure SI 4. mRNA relative expression of inflammatory markers in the brainstem at (a) 1 week and (b) 16 weeks. The expression level in the control group is set at 100. Data are mean ± s.e.m.; Student’s t test between the HFD group and its CTL group *P < 0.05; **P < 0.01; and ***P < 0.001. Figure SI 5. mRNA relative expression of junction proteins of the blood-brain barrier at week 16 in the cortex and the cerebellum. The expression levels in the control groups were set at 100. Data are mean ± s.e.m. The white columns represent mice fed a standard diet and the black columns represent the mice fed a HFD. One-way ANOVA with Bonferroni’s post test and between HFD group and its respective CTL group *P < 0.05 and ***P < 0.001 and between CNS regions #P < 0.05 and ###P < 0.001. (PDF 67 kb) Abbreviations CBLCerebellum COX-2Cyclooxygenase-2 CTXCortex EATEpididymal adipose tissue GFAPGlial fibrillary acidic protein HFDHigh-fat diet Iba-1Ionized calcium-binding adapter molecule 1 IL-1βInterleukin-1β iNOSInducible nitric oxide synthase LBPLPS-binding protein LPCLysophosphatidylcholine LPSLipopolysaccharides MCP-1Monocyte chemoattractant protein 1 (or CCL2) PEPhosphatidylethanolamine PEAPalmitoylethanolamide PIPhosphatidylinositol SATSubcutaneous adipose tissue Serpina3nSerine peptidase inhibitor, clade A, member 3N (or alpha 1-antitrypsin) SMSphingomyelin TNFαTumor necrosis factor α VATVisceral adipose tissue ZO-1Zonula occludens-1 Acknowledgements The MASSMET platform is acknowledged for the use of the mass spectrometers. Funding OGL is a research fellow of the “Fonds pour la recherche dans l’industrie et l’agriculture” (FRIA, Belgium). AM and AE are postdoctoral researcher and PDC research associate from the FRS-FNRS, Belgium. GGM is the recipient of subsidies from the Fonds Spéciaux de Recherches (FSR, Université catholique de Louvain) and from the FRS-FNRS, Belgium (grants CC 1.5.034.10, FRFC 2.4555.08, J.0160.13). This work was supported by the Fonds de la Recherche Scientifique – FNRS and FRFS-WELBIO under Grant no. WELBIO-CR-2012S-02R. PDC is a recipient of ERC Starting Grant 2013 (European Research Council, Starting grant 336452-ENIGMO). Availability of data and materials Data supporting the conclusions of this article are presented in the paper and its supplementary files. Authors’ contributions OGL, MA, and PDC performed the animal procedures. MA and GGM performed the ex vivo experiments. OGL performed the molecular biology, plasma cholesterol assay, and immunohistology procedures. JM performed the large-scale lipid analysis and quantification. OGL and AE performed the multiplex assay on plasma samples. OGL, MA, and GGM wrote the manuscript. GGM, MA, and OGL designed the study. GGM supervised the study. 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 The local (Health Sector, Université catholique de Louvain) ethics committee on animal experimentations approved the protocol of the study (study agreement 2010/UCL/MD/022; lab agreement LA1230314). ==== Refs References 1. Ogden CL Yanovski SZ Carroll MD Flegal KM The epidemiology of obesity Gastroenterology 2007 132 2087 2102 10.1053/j.gastro.2007.03.052 17498505 2. Wang YC McPherson K Marsh T Gortmaker SL Brown M Health and economic burden of the projected obesity trends in the USA and the UK Lancet 2011 378 815 825 10.1016/S0140-6736(11)60814-3 21872750 3. Guh DP Zhang W Bansback N Amarsi Z Birmingham CL Anis AH The incidence of co-morbidities related to obesity and overweight: a systematic review and meta-analysis BMC Public Health 2009 9 88 10.1186/1471-2458-9-88 19320986 4. Cani PD Osto M Geurts L Everard A Involvement of gut microbiota in the development of low-grade inflammation and type 2 diabetes associated with obesity Gut Microbes 2012 3 279 288 10.4161/gmic.19625 22572877 5. Delzenne NM Neyrinck AM Backhed F Cani PD Targeting gut microbiota in obesity: effects of prebiotics and probiotics Nat Rev Endocrinol 2011 7 639 646 10.1038/nrendo.2011.126 21826100 6. Lumeng CN Saltiel AR Inflammatory links between obesity and metabolic disease J Clin Invest 2011 121 2111 2117 10.1172/JCI57132 21633179 7. Bastien M Poirier P Lemieux I Despres JP Overview of epidemiology and contribution of obesity to cardiovascular disease Prog Cardiovasc Dis 2014 56 369 381 10.1016/j.pcad.2013.10.016 24438728 8. Sideleva O Dixon AE The many faces of asthma in obesity J Cell Biochem 2014 115 421 426 10.1002/jcb.24678 24115053 9. Park J Morley TS Kim M Clegg DJ Scherer PE Obesity and cancer—mechanisms underlying tumour progression and recurrence Nat Rev Endocrinol 2014 10 455 465 10.1038/nrendo.2014.94 24935119 10. Cani PD and Everard A. Talking microbes: When gut bacteria interact with diet and host organs. Mol. Nutr. Food Res. 2016;60(1):58–66 11. Cani PD Amar J Iglesias MA Poggi M Knauf C Bastelica D Neyrinck AM Fava F Tuohy KM Chabo C Waget A Delmee E Cousin B Sulpice T Chamontin B Ferrieres J Tanti JF Gibson GR Casteilla L Delzenne NM Alessi MC Burcelin R Metabolic endotoxemia initiates obesity and insulin resistance Diabetes 2007 56 1761 1772 10.2337/db06-1491 17456850 12. Cani PD Bibiloni R Knauf C Waget A Neyrinck AM Delzenne NM Burcelin R Changes in gut microbiota control metabolic endotoxemia-induced inflammation in high-fat diet-induced obesity and diabetes in mice Diabetes 2008 57 1470 1481 10.2337/db07-1403 18305141 13. Kim KA Gu W Lee IA Joh EH Kim DH High fat diet-induced gut microbiota exacerbates inflammation and obesity in mice via the TLR4 signaling pathway PLoS One 2012 7 e47713 10.1371/journal.pone.0047713 23091640 14. Tsukumo DM Carvalho-Filho MA Carvalheira JB Prada PO Hirabara SM Schenka AA Araujo EP Vassallo J Curi R Velloso LA Saad MJ Loss-of-function mutation in Toll-like receptor 4 prevents diet-induced obesity and insulin resistance Diabetes 2007 56 1986 1998 10.2337/db06-1595 17519423 15. Davis JE Gabler NK Walker-Daniels J Spurlock ME Tlr-4 deficiency selectively protects against obesity induced by diets high in saturated fat Obesity (Silver Spring) 2008 16 1248 1255 10.1038/oby.2008.210 18421279 16. Misiak B Leszek J Kiejna A Metabolic syndrome, mild cognitive impairment and Alzheimer’s disease—the emerging role of systemic low-grade inflammation and adiposity Brain Res Bull 2012 89 144 149 10.1016/j.brainresbull.2012.08.003 22921944 17. Lee EB Mattson MP The neuropathology of obesity: insights from human disease Acta Neuropathol 2014 127 3 28 10.1007/s00401-013-1190-x 24096619 18. Kernan WN Inzucchi SE Sawan C Macko RF Furie KL Obesity: a stubbornly obvious target for stroke prevention Stroke 2013 44 278 286 10.1161/STROKEAHA.111.639922 23111440 19. Beydoun MA Beydoun HA Wang Y Obesity and central obesity as risk factors for incident dementia and its subtypes: a systematic review and meta-analysis Obes Rev 2008 9 204 218 10.1111/j.1467-789X.2008.00473.x 18331422 20. Williams LM Hypothalamic dysfunction in obesity Proc Nutr Soc 2012 71 521 533 10.1017/S002966511200078X 22954151 21. Thaler JP Guyenet SJ Dorfman MD Wisse BE Schwartz MW Hypothalamic inflammation: marker or mechanism of obesity pathogenesis? Diabetes 2013 62 2629 2634 10.2337/db12-1605 23881189 22. Thaler JP Schwartz MW Minireview: Inflammation and obesity pathogenesis: the hypothalamus heats up Endocrinology 2010 151 4109 4115 10.1210/en.2010-0336 20573720 23. Schwartz MW Woods SC Porte D Jr Seeley RJ Baskin DG Central nervous system control of food intake Nature 2000 404 661 671 10766253 24. Morton GJ Meek TH Schwartz MW Neurobiology of food intake in health and disease Nat Rev Neurosci 2014 15 367 378 10.1038/nrn3745 24840801 25. Lam TK Neuronal regulation of homeostasis by nutrient sensing Nat Med 2010 16 392 395 10.1038/nm0410-392 20376051 26. Buckman LB Thompson MM Moreno HN Ellacott KL Regional astrogliosis in the mouse hypothalamus in response to obesity J Comp Neurol 2013 521 1322 1333 10.1002/cne.23233 23047490 27. Thaler JP Yi CX Schur EA Guyenet SJ Hwang BH Dietrich MO Zhao X Sarruf DA Izgur V Maravilla KR Nguyen HT Fischer JD Matsen ME Wisse BE Morton GJ Horvath TL Baskin DG Tschop MH Schwartz MW Obesity is associated with hypothalamic injury in rodents and humans J Clin Invest 2012 122 153 162 10.1172/JCI59660 22201683 28. Kennedy A Martinez K Chuang CC LaPoint K McIntosh M Saturated fatty acid-mediated inflammation and insulin resistance in adipose tissue: mechanisms of action and implications J Nutr 2009 139 1 4 10.3945/jn.108.098269 19056664 29. Klop B Elte JW Cabezas MC Dyslipidemia in obesity: mechanisms and potential targets Nutrients 2013 5 1218 1240 10.3390/nu5041218 23584084 30. Tchernof A Despres JP Pathophysiology of human visceral obesity: an update Physiol Rev 2013 93 359 404 10.1152/physrev.00033.2011 23303913 31. Lee BC Lee J Cellular and molecular players in adipose tissue inflammation in the development of obesity-induced insulin resistance Biochim Biophys Acta 2014 1842 446 462 10.1016/j.bbadis.2013.05.017 23707515 32. Wellen KE Hotamisligil GS Obesity-induced inflammatory changes in adipose tissue J Clin Invest 2003 112 1785 1788 10.1172/JCI20514 14679172 33. Brown HA Murphy RC Working towards an exegesis for lipids in biology Nat Chem Biol 2009 5 602 606 10.1038/nchembio0909-602 19690530 34. Shimizu T Lipid mediators in health and disease: enzymes and receptors as therapeutic targets for the regulation of immunity and inflammation Annu Rev Pharmacol Toxicol 2009 49 123 150 10.1146/annurev.pharmtox.011008.145616 18834304 35. Serhan CN Pro-resolving lipid mediators are leads for resolution physiology Nature 2014 510 92 101 10.1038/nature13479 24899309 36. Nam M Choi MS Jung S Jung Y Choi JY Ryu DH Hwang GS Lipidomic profiling of liver tissue from obesity-prone and obesity-resistant mice fed a high fat diet Sci Rep 2015 5 16984 10.1038/srep16984 26592433 37. Muccioli GG Naslain D Backhed F Reigstad CS Lambert DM Delzenne NM Cani PD The endocannabinoid system links gut microbiota to adipogenesis Mol Syst Biol 2010 6 392 10.1038/msb.2010.46 20664638 38. Mutemberezi V, Masquelier J, Guillemot-Legris O,Muccioli GG. Development and validation of an HPLC-MS method for the simultaneous quantification of key oxysterols, endocannabinoids, and ceramides: variations in metabolic syndrome. Anal Bioanal. Chem. 2016;408(3):733–45 39. Schenk S Saberi M Olefsky JM Insulin sensitivity: modulation by nutrients and inflammation J Clin Invest 2008 118 2992 3002 10.1172/JCI34260 18769626 40. Chavez JA Summers SA A ceramide-centric view of insulin resistance Cell Metab 2012 15 585 594 10.1016/j.cmet.2012.04.002 22560211 41. Todoric J Loffler M Huber J Bilban M Reimers M Kadl A Zeyda M Waldhausl W Stulnig TM Adipose tissue inflammation induced by high-fat diet in obese diabetic mice is prevented by n-3 polyunsaturated fatty acids Diabetologia 2006 49 2109 2119 10.1007/s00125-006-0300-x 16783472 42. Alhouayek M Masquelier J Cani PD Lambert DM Muccioli GG Implication of the anti-inflammatory bioactive lipid prostaglandin D2-glycerol ester in the control of macrophage activation and inflammation by ABHD6 Proc Natl Acad Sci U S A 2013 110 17558 17563 10.1073/pnas.1314017110 24101490 43. Pannacciulli N Del PA Chen K Le DS Reiman EM Tataranni PA Brain abnormalities in human obesity: a voxel-based morphometric study Neuroimage 2006 31 1419 1425 10.1016/j.neuroimage.2006.01.047 16545583 44. Alonso-Alonso M Pascual-Leone A The right brain hypothesis for obesity JAMA 2007 297 1819 1822 10.1001/jama.297.16.1819 17456824 45. Miller JL Couch J Schwenk K Long M Towler S Theriaque DW He G Liu Y Driscoll DJ Leonard CM Early childhood obesity is associated with compromised cerebellar development Dev Neuropsychol 2009 34 272 283 10.1080/87565640802530961 19437203 46. Mueller K Sacher J Arelin K Holiga S Kratzsch J Villringer A Schroeter ML Overweight and obesity are associated with neuronal injury in the human cerebellum and hippocampus in young adults: a combined MRI, serum marker and gene expression study Transl Psychiatry 2012 2 e200 10.1038/tp.2012.121 23212584 47. Volkow ND Wang GJ Baler RD Reward, dopamine and the control of food intake: implications for obesity Trends Cogn Sci 2011 15 37 46 10.1016/j.tics.2010.11.001 21109477 48. Morton GJ Cummings DE Baskin DG Barsh GS Schwartz MW Central nervous system control of food intake and body weight Nature 2006 443 289 295 10.1038/nature05026 16988703 49. Baker RG Hayden MS Ghosh S NF-kappaB, inflammation, and metabolic disease Cell Metab 2011 13 11 22 10.1016/j.cmet.2010.12.008 21195345 50. Tornatore L Thotakura AK Bennett J Moretti M Franzoso G The nuclear factor kappa B signaling pathway: integrating metabolism with inflammation Trends Cell Biol 2012 22 557 566 10.1016/j.tcb.2012.08.001 22995730 51. Hawkins BT Davis TP The blood-brain barrier/neurovascular unit in health and disease Pharmacol Rev 2005 57 173 185 10.1124/pr.57.2.4 15914466 52. Ballabh P Braun A Nedergaard M The blood-brain barrier: an overview: structure, regulation, and clinical implications Neurobiol Dis 2004 16 1 13 10.1016/j.nbd.2003.12.016 15207256 53. Mauro C De RV Marelli-Berg F Solito E Metabolic syndrome and the immunological affair with the blood-brain barrier Front Immunol 2014 5 677 25601869 54. Jiao H Wang Z Liu Y Wang P Xue Y Specific role of tight junction proteins claudin-5, occludin, and ZO-1 of the blood-brain barrier in a focal cerebral ischemic insult J Mol Neurosci 2011 44 130 139 10.1007/s12031-011-9496-4 21318404 55. Tian L Ma L Kaarela T Li Z Neuroimmune crosstalk in the central nervous system and its significance for neurological diseases J Neuroinflammation 2012 9 155 10.1186/1742-2094-9-155 22747919 56. Ransohoff RM Brown MA Innate immunity in the central nervous system J Clin Invest 2012 122 1164 1171 10.1172/JCI58644 22466658 57. Yi CX Al-Massadi O Donelan E Lehti M Weber J Ress C Trivedi C Muller TD Woods SC Hofmann SM Exercise protects against high-fat diet-induced hypothalamic inflammation Physiol Behav 2012 106 485 490 10.1016/j.physbeh.2012.03.021 22483785 58. Zamanian JL Xu L Foo LC Nouri N Zhou L Giffard RG Barres BA Genomic analysis of reactive astrogliosis J Neurosci 2012 32 6391 6410 10.1523/JNEUROSCI.6221-11.2012 22553043 59. Abraham CR Reactive astrocytes and alpha1-antichymotrypsin in Alzheimer’s disease Neurobiol Aging 2001 22 931 936 10.1016/S0197-4580(01)00302-5 11755001 60. Kang SC Kim BR Lee SY Park TS Sphingolipid metabolism and obesity-induced inflammation Front Endocrinol (Lausanne) 2013 4 67 23761785 61. Maceyka M Spiegel S Sphingolipid metabolites in inflammatory disease Nature 2014 510 58 67 10.1038/nature13475 24899305 62. Alhouayek M Muccioli GG Harnessing the anti-inflammatory potential of palmitoylethanolamide Drug Discov Today 2014 19 1632 1639 10.1016/j.drudis.2014.06.007 24952959 63. Jeon SB Yoon HJ Park SH Kim IH Park EJ Sulfatide, a major lipid component of myelin sheath, activates inflammatory responses as an endogenous stimulator in brain-resident immune cells J Immunol 2008 181 8077 8087 10.4049/jimmunol.181.11.8077 19018000 64. Rodriguez-Cuenca S Barbarroja N Vidal-Puig A Dihydroceramide desaturase 1, the gatekeeper of ceramide induced lipotoxicity Biochim Biophys Acta 2015 1851 40 50 10.1016/j.bbalip.2014.09.021 25283058

PMC005xxxxxx/PMC5002132.txt

==== Front Stem Cell Res TherStem Cell Res TherStem Cell Research & Therapy1757-6512BioMed Central London 38810.1186/s13287-016-0388-yResearchChemotaxis-driven disease-site targeting of therapeutic adult stem cells in dystrophic epidermolysis bullosa Alexeev Vitali Vitali.Alexeev@jefferson.edu Donahue Adele Adele.Donahue@jefferson.edu Uitto Jouni Jouni.Uitto@jefferson.edu http://orcid.org/0000-0001-9813-7184Igoucheva Olga 215-503-5434Olga.Igoucheva@jefferson.edu Department of Dermatology and Cutaneous Biology, Sidney Kimmel Medical College, Thomas Jefferson University, 233 South 10th Street, BLSB, Rm. 430, Philadelphia, PA 19107 USA 27 8 2016 27 8 2016 2016 7 1 12427 6 2016 5 8 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 Dystrophic epidermolysis bullosa (DEB), a rare genodermatosis, is characterized by the formation of intra-epidermal blistering and the development of chronic nonhealing skin wounds. Recently, attempts have been made to develop cell-based therapies for this currently intractable disorder. The molecular mechanisms that govern directional migration of the adult stem cells, allowing their efficient and controlled homing to the skin affected with DEB, are poorly understood. The key mechanism that regulates recruitment of leukocytes and progenitor stem cells to distal anatomical tissues affected with disease is chemotaxis, which depends on the signaling molecules, chemokines, and acts primarily as part of the host defense and repair mechanism. Methods Comprehensive proteomic screening of chemokines in the blister fluids of DEB-affected mice was conducted to define the inflammatory and immune activities, thus providing potential to examine local biological mechanisms and define the protein signature within lesional skin as a potential marker of disease activity. Also, the therapeutic relevance of identified chemotactic pathways was investigated in vivo, providing a basis for future clinical investigations. Results Assessment of blister fluid-derived chemokines showed a persistent presence of several chemotactic molecules, including CXCL1 + 2 and CXCL5. The majority of blister-originated chemotactic signals were associated with preferential recruitment of CD45+CXCR2+ and CD11b+CXCR2+ leukocytes. Systemic transplantation of an enriched CXCR2 population of mouse adipose-derived stem cells (mADSC) into DEB-affected mice demonstrated effective recruitment of cells to the blistering skin under the influence of blister-derived ligands and deposition of therapeutic type VII collagen. Conclusions Collectively, these studies demonstrate that recruitment of mADSC into DEB skin is tightly controlled by disease-site chemotactic activities and suggest a potential mechanism for effective application of therapeutic stem cells for DEB. Keywords Adipose-derived stem cellsAdult stem cellsBlistering skinChemokine receptorsChemokinesChemotaxisEpidermolysis bullosaMigrationProteome analysisTransplantationType VII collagenNational Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health R01AR064286Igoucheva Olga issue-copyright-statement© The Author(s) 2016 ==== Body Background Dystrophic epidermolysis bullosa (DEB) is the mechanobullous disorder characterized by complete lack or dysfunction of type VII collagen, resulting in separation of the dermal–epidermal junction (DEJ) below the lamina densa following minor or insignificant trauma to the skin [1]. During the past two decades, various therapeutic approaches have been proposed and tested in preclinical and clinical settings [2]; however, effective therapies allowing correction of this debilitating disorder remain to be developed. In recent years, adult stem cell-based therapy was suggested as a potential approach to attenuate DEB progression. We and others have demonstrated that adult stem cells (ASC) of different origins (bone marrow, adipose tissue, blood vessels, and umbilical cord) produce a repertoire of extracellular matrix proteins, including type VII collagen, and a plethora of signaling molecules, such as growth factors and cytokines [3–8]. Also, we showed that intracutaneous transplantation of genetically normal ASC from bone marrow into DEB-affected mouse skin restores the DEJ via direct donation of the therapeutic type VII collagen [8]. These studies provided proof-of-concept data suggesting that adult stem cell-derived connective tissue proteins can participate directly in restoration of defective extracellular matrix and demonstrating the feasibility of systemic cellular therapy for DEB. In fact, data collected from ongoing clinical studies aimed at the application of whole bone marrow and bone marrow-derived mesenchymal stem cells (MSC) for DEB treatment showed that systemic allotransplantation accelerates wound healing, enhances angiogenesis, inhibits inflammation, and alleviates symptoms associated with the disease [9, 10]. However, despite encouraging results, these trials revealed that the majority of curative effects resulted from the secretion of various paracrine factors elicited by transplanted cells rather than from donation of the therapeutic protein to the affected cutaneous tissue [11]. Successful application of ASC for the correction of genodermatoses, including DEB, and treatment of chronic wounds which often accompany these disorders, mainly depends on the efficacy of stem cell homing to the damaged skin and is considered a major issue in stem cell-based therapy. The paucity and inefficient migration of ASC to the damaged skin have been documented in multiple wound healing studies [12, 13]. The causatives of the inefficient homing of ASC are still not defined but current data suggest that these deficiencies may arise, in part, due to poor extravasation of ASC from circulation into the cutaneous tissue. The molecular mechanisms regulating trafficking of the ASC to the skin, especially blistering skin, are poorly understood. Under physiological conditions, the key mechanism that regulates cell migration is chemotaxis, which depends on signaling molecules, chemokines. Being secreted from cells in damaged peripheral tissue, chemokines recruit chemokine receptor-expressing cells, such as leukocytes, and possibly stem cells as part of host defense and repair mechanisms. The crucial role of chemokines in directional migration of ASC to various tissues, such as the heart, lung, bone marrow, and intestine, has been demonstrated previously [14, 15]. Thus, knowledge of chemokines expressed by the targeted tissue and cognate chemokine receptors expressed on the surface of the targeted stem cells is a prerequisite for successful targeting of therapeutic ASC. The primary challenge is therefore to integrate our current knowledge of chemotaxis into a rational design of the guidance system to recruit ASC to the skin affected by certain genodermatoses, such as DEB. Because ASC are naturally involved in the regeneration/repair of multiple tissues, expression of chemokine receptors in ASC from various sources has been studied widely. Collectively, these data indicate that ASC express a limited repertoire of functional chemokine receptors [6, 16, 17]. Moreover, responsiveness of ASC to only a few chemokines was tested experimentally [16, 18–22]. To the contrary, from our knowledge of chemotaxis in repair processes, very limited information is currently available regarding chemotactic signals in the skin affected by various forms of hereditary blistering disorders. Skin-derived chemokines have been mostly characterized in association with recruitment of leukocytes during acute and chronic inflammation [23]. Elevated levels of CXCL12 and CCL21 were reported in blister fluids following skin burns, lichen planus, and cutaneous lupus erythematosus, where inflammatory and immune responses were supported by the recruitment of CXCR4+ and CCR7+ leukocytes, respectively [24–26]. However, wide expression of CXCL12 in multiple organs does not provide desirable specificity of ASC migration leading to significant “dilution” of the therapeutic effect. Importantly, the CXCL12–CXCR4 chemotactic axis was also shown to be responsible for the entrapment of systemically administered ASC in lungs. For the same token, it was suggested that an endogenously high level of CCL21 in secondary lymphoid organs is responsible for the observed nonspecific distribution of systemically transplanted ASC [27]. Collectively, these observations strongly suggest the need for alternative chemotactic pathways to target ASC specifically to the skin. Based on current knowledge of T-cell trafficking [23], it is plausible that systemically transplanted ASC can be recruited from circulation to the skin by the activation of specific receptors on the surface of transplanted stem cells, followed by transcutaneous migration mediated by stem cell receptors under the influence of the DEB skin-derived chemokines. In this study, we conducted proteomic screening of chemotactic molecules in the blister fluids of DEB-affected mice, the model for severe autosomal recessive DEB, to define the inflammatory and immune activities during the active disease state. Pair-wise comparison of identified chemokines in blister fluids and receptors on the surface of mouse adipose-derived stem cells (mADSC) allowed us to define potential chemotactic axes for targeting of the cells into the blistering skin. We investigated the therapeutic relevance of identified chemotactic pathways in vivo, providing a basis for future clinical investigations. Identified molecules provided insight into the mechanisms that govern directional migration and intracutaneous trafficking of systemically infused stem cells, thus permitting broad and effective application of the therapeutic cells for DEB treatment. Methods Mouse strains Wild-type C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Transgenic type VII collagen-deficient (Col7a1–/–) mice were generated in the Department of Dermatology and Cutaneous Biology, Thomas Jefferson University [28]. Transgenic type VII collagen-deficient hypomorphic (Col7a1f lNeo/f lNeo) mice were obtained from Freiburg Institute for Advanced Studies, Germany [29]. Both mouse models recapitulate the clinical, genetic, and ultrastructural features of human DEB. Targeted inactivation of the Col7a1 gene generated a severely affected collagen VII knockout mouse (Col7a1–/–), which is born with extensive cutaneous blistering and rapid demise during the first 2 weeks of life due to complications arising from blistering in the skin, oral mucosa, and esophagus and imbalanced digestion by the gastrointestinal track. The second model, a collagen VII hypomorphic mouse (Col7a1f lNeo/f lNeo) which had 10 % of the normal collagen VII levels in the skin, developed all of the symptoms of severe RDEB but the presence of one-tenth of normal full-length type VII collagen confers a milder phenotype and better prognosis than that in Col7a1 knockout mice. Blister fluid collection from DEB-affected mice Col7a1–/– and Col7a1f lNeo/f lNeo mice are born with a blistering phenotype. Hemorrhagic blisters are readily developed on paws and other parts on the body (e.g., abdomen, armpit, neck). The blister fluids were collected by needle piercing with an attached syringe, cleared by centrifugation, and stored at –70 °C until testing. Chemokine antibody arrays Proteome Profiler™ Mouse Chemokine Antibody Array (R&D Systems, Minneapolis, MN, USA) was employed to assay blister fluid samples derived from Col7a1–/– and Col7a1f lNeo/f lNeo mice, respectively. Twenty microliters of blister fluid was used to probe the chemokine antibody arrays according to the manufacturer’s instructions. Chemokine antibody array membranes were developed by standard enhanced chemiluminescence techniques as advised by the manufacturer. Acquisition of signals on mouse chemokine arrays was quantitatively determined using ScanAlize version 2.50 (Stanford University) and GEArray Expression Analysis Suite 2.0 software (SABiosciences, Frederick, MD, USA), which reads the images and matches them to the corresponding protein on the array. The net level of each protein was calculated by the mean of the individual spot intensity minus the mean of the background intensity. To provide normalization, the average level ratio of two principal genes was determined and introduced as a correction factor. Relative spot intensities are presented as mean ± SD. Microsoft Excel (Microsoft, Redmond, WA, USA) was utilized for statistical analysis. Isolation of mADSC and tissue culture conditions mADSC were isolated from subcutaneous fat of wild-type C57 BL/6 J mice. Following collection, specimens were washed in PBS + 1 % Pen/Strep (Gibco, Grand Island, NY, USA) twice, minced into small pieces, and digested in collagenase solution (0.1 μg collagenase I (Sigma, St. Louis, MO, USA) in 1 ml PBS and bovine serum albumin (BSA)). To obtain a single cell suspension, the digested tissue was applied to a 30 μm mesh separation filter (Miltenyi Biotec, Auburn, CA, USA). PBS + 1 % BSA solution was added to the mesh to quench the enzyme and flush any remaining cells through the filter. The suspension was centrifuged and the pellet was resuspended in 1 ml of DMEM/F12 and Glutamax + 10 % FBS (Gibco). Cells were plated in DMEM/F12 and Glutamax + 10 % FBS (Invitrogen, Grand Island, NY, USA) and grown to confluence. The adherent cells (passage 0) underwent negative selection using magnetic beads (MACS; Miltenyi Biotec) to remove contaminating endothelial CD31+ and mononuclear CD45+ cells. Briefly, cells were released by trypsin and centrifuged at 300 × g. The pellet was suspended in 2 mM EDTA and 0.5 % BSA (MACS buffer), mixed with CD31+ and CD45+ tagged microbeads, and incubated at 4 °C for 15 minutes. After centrifugation, the cells were suspended in MACS buffer and passed through an LS Separation Column mounted in the QuadroMACS separator magnetic unit. The resultant CD31–CD45– mouse cell population defined as mADSC was grown in standard DMEM/F12 and Glutamax + 10 % FBS media unless stated otherwise. Fluorescence-activated cell sorting analysis of mADSC mADSC were grown until confluent, trypsinized, and pelleted by centrifugation at 200 × g for 5 minutes. For fluorescence-activated cell sorting (FACS) analysis, ~1.0 × 105 cells were resuspended in 100 μl FACS buffer containing 1 % fetal bovine serum in PBS. For FACS analysis of surface receptors, each sample was incubated for 30 minutes at 4 °C with FITC-conjugated, Alexa488-conjugated, PerCP/Cy5.5-conjugated, PE-conjugated, or Alexa Fluor-647-conjugated antibodies against the surface markers CCR2, CCR3, CCR4, CCR5, CCR6, CCR7, CCR9, CCR10, CXCR1, CXCR2, CXCR3, CXCR4, CXCR6, CXCR5, CXCR6, and CXCR7 (eBioscience, San Diego, CA, USA) according to the manufacturer’s instructions. After incubation, the labeled cells were diluted with 2 ml of FACS buffer, pelleted and resuspended in 300 μl of FACS buffer. Generally, ~5 × 104 cells were analyzed per sample using the Guava flow cytometer (BD Biosciences, San Jose, CA, USA). Results were analyzed using GuavaSoft 2.7 software (BD Biosciences). Generation and characterization of CXCR2-overexpressing mADSC Full-length mouse Cxcr2 receptor with 3′ UTR was amplified from total mouse RNA via reverse transcription reaction using the Superscript II RT Kit (Invitrogen, Carlsbad, CA, USA) followed by PCR using PFU II high fidelity polymerase (Agilent Technologies, Santa Clara, CA, USA). Resultant cDNA was inserted into pEF2-TOPO vector. Integrity of the promoter and cDNA was verified by direct DNA sequencing. Minimally cultured mADSC (passages 1–2) were nucleofected with resultant plasmid (pEF1-mCxcr2) using Lonza nucleofection reaction (T-27 program, nucleofection kit V; Lonza, Cologne, Germany). Further, a pool of CXCR2-expressing cells was selected with Blasticidin (0.5 mg/ml; Invitrogen) for 10 days. Expression of CXCR2 in selected cells was confirmed by FACS and indirect immunofluorescence analyses. Surface expression of CXCR2 was determined by FACS using PE-conjugated antibodies as already described. For indirect immunofluorescence, CXCR2 immunocomplexes were detected with Alexa-Fluor488-conjugated secondary antibodies (Invitrogen). Nuclei were counterstained with 4′,6-diamidino-2-phenyl indol (DAPI; Sigma). Immunofluorescent images were obtained on a Nikon TS100F fluorescent microscope (Nikon, Melville, NY, USA). Labeling CXCR2+ mADSC and transplantation into Col7a1–/– mice All animal procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication no. 86-23) and approved by the Institutional Animal Care and Use Committee of the Thomas Jefferson University. For all transplantation studies, mADSC (passages 2–4) isolated from wild-type C57BL/6 J were labeled with a red lipophilic tracer DiOC18 (Molecular Probes, Grand Island, NY, USA) as we described previously [6]. For systemic transplantation, the 1–2-day-old Col7a1–/– neonates (n = 5/time point) received 0.6 × 106–0.8 × 106 cells in 10–15 μl PBS per mouse intraperitoneally. For direct viewing, transplanted cells were detected using an IVIS live-imaging system (IVIS Lumina XR; Caliper LifeSciences, Alameda, CA, USA). To avoid detection of autofluorescence, the exposure time of all imaged animals was minimized to a maximum of 1 second. Histological and immunofluorescent analyses Collected skin tissue was embedded into OCT compound (VWR, Radnor, PA, USA), frozen, and cryosectioned at a thickness of 7 μm. For histological analysis, sections were stained with H&E using a standard protocol. For direct viewing of transplanted mADSC, sections were fixed with 10 % formalin, stained with DAPI (Sigma) and evaluated using fluorescence microscopy. For indirect immunofluorescent analysis, cross-sections were stained with the following primary antibodies: goat anti-mouse CD45 polyclonal (BD Bioscience), rat anti-mouse CD11b monoclonal (BD Bioscience), rabbit anti-CXCR2 polyclonal (Bioss, Woburn, MA, USA), mouse anti-FLuc monoclonal (Abcam, Cambridge, MA, USA), rabbit anti-type VII collagen polyclonal (Millipore, Billerica, MA, USA), and goat anti-type IV collagen polyclonal (Millipore). Immunocomplexes were detected with AlexaFluor488-labeled and AlexaFluor594-labeled secondary antibodies (Invitrogen), respectively. Nuclei were counterstained with DAPI. Immunofluorescent images were obtained on a Nikon TS100F fluorescent microscope (Nikon). For analysis of the type VII collagen deposition, stained sections were analyzed using AutoQuant imaging software (AutoQuant Imaging Inc., Troy, NY, USA).The average value of the basement membrane zone (BMZ)-accolated fluorescence was calculated based on the ratio of the average value of the total fluorescence of wild-type skin (designated as 100 %) and the relative percentage of fluorescence in the skin of transplanted mice. Differences in fluorescence intensities were assessed using a Student’s t test, with p < 0.01 considered significant in all tests. Statistical analysis All data are expressed as mean ± SD. Statistical significance was assessed using two-tailed type 2 Student’s t test, where differences with p < 0.05 were considered significant. Results Proteomic screens of chemotactic molecules in DEB-affected mouse skin The chemokine expression profiles in blister fluids of Col7a1–/– knockout mice and Col7a1f lNeo/f lNeo hypomorphic mice were studied using Proteome Profiler™ Mouse Chemokine Antibody Array allowing the detection and quantification of 35 distinct chemotactic molecules. The chemokine arrays were probed using blister samples collected from blistering skin of mice. Proteomic screens of blister fluids collected from Col7a1–/– (n = 5) and Col7a1f lNeo/f lNeo (n = 5) mice showed consistently high levels of several chemotactic molecules, including CCL6, chemerin, CCL8, CCL9/CCL10, CCL12, CXCL1, CXCLl2, CXCL5, and CDF (Fig. 1a).Fig. 1 Analysis of chemokines in DEB-associated blister fluids and chemokine receptors on the surface of mADSC. a Chemokine profile in blister fluids of Col7a1 –/– (DEB) and Col7a1 f lNeo/f lNeo (Hypo DEB) mice conducted using Proteome Profiler™ Mouse Chemokine Antibody Array. Blister fluids were collected from blister-affected skin (paw, armpit, neck, and abdomen) of newborn mice. Spot intensities were normalized based on background levels and positive control signals. Data shown as mean ± SD from three independent arrays (p > 0.05). b Surface expression of chemokine receptors in primary mADSC on early passage (passages 2–3) detected by flow cytometry. Light gray curves, control antibody signal; dark shaded curves, specific antibody signal. Histograms are representative examples of chemokine receptor expression profiles obtained from five independent mADSC preparations. c Quantitation of flow cytometry results for expression of the chemokine receptors in primary mADSC. Data shown as mean ± SD of cell percentage. DEB dystrophic epidermolysis bullosa The analysis of biological function of identified chemokines suggested potential inflammatory infiltrate in blistering skin. For example, CCL6 has only been identified in rodents and is a potent chemoattractant of macrophages as well as B cells, CD4+ lymphocytes, and eosinophils. In mice, CCL6 is expressed in cells from neutrophil and macrophage lineages, and can be induced under conditions suitable for myeloid cell differentiation. The cell surface receptor for CCL6 is believed to be the chemokine receptor CCR1. Chemerin, also known as retinoic acid receptor responder protein 2 (RARRES2), is a chemoattractant protein that acts as a ligand for the G protein-coupled receptor CMKLR1, also known as ChemR23. Chemerin was found to stimulate chemotaxis of dendritic cells and macrophages to the site of inflammation. CCL8, also known as monocyte chemoattractant protein 2 (MCP-2), is chemotactic for and activates many different immune cells, including mast cells, eosinophils, and basophils as well as monocytes, T cells, and NK cells that are involved in the inflammatory response. CCL8 elicits its effects by binding to several different cell surface receptors, including CCR1, CCR2B, and CCR5. CCL9/CCL10, also called macrophage inflammatory protein-1 gamma (MIP-1γ), macrophage inflammatory protein-related protein-2 (MRP-2), or CCF18, in rodents attracts dendritic cells that possess the cell surface molecule CD11b and the chemokine receptor CCR1. CCL9 is constitutively expressed in macrophages and myeloid cells. CCL12, also known as monocyte chemotactic protein 5 (MCP-5) or MCP-1-related chemokine, specifically attracts eosinophils, monocytes, and lymphocytes. Its expression can be hugely induced in macrophages. CCL12 is a ligand for CCR2. CXCL1/2/3, also known as growth-regulated oncogene (GRO), can bind with high affinity to the IL-8 receptor type B and is a very potent neutrophil attractant and activator. CXCL5, also known as epithelial-derived neutrophil-activating peptide 78 (ENA-78), is produced following stimulation of cells with the inflammatory cytokines interleukin-1 or tumor necrosis factor alpha. CXCL5 is well known to have chemotactic and activating functions on neutrophils, mainly during acute inflammatory responses. This chemokine stimulates the chemotaxis of neutrophils possessing angiogenic properties and has been implicated in connective tissue remodeling. The cell surface receptor for CXCL5 is the chemokine receptor CXCR2. Collectively, all identified molecules are prominent chemotactic factors that activate and attract a variety of inflammatory cells to the sites of inflammation produced by tissue injury or disease. Also, the identified chemokine profiles provided several potential lead-chemotactic gradients, including CCL6–CCR1, CCL8–CCR1/CCR5, CCL9/10–CCR1, CCL12–CCR2, CXCL1/2–CXCR2, and CXCL5–CXCR2, for efficient recruitment of therapeutic stem cells into the skin affected with blisters. Expression profile of chemokine receptors in mADSC The presence of active chemokine receptors on mADSC was examined in order to determine potential migratory ability upon in-vivo transplantation into mice. A series of mADSC preparations were generated from subcutaneous fat of C57BL6J mice (n = 5), representing actively dividing cultures at early passage (passage 2). Cell surface expression for receptors on the primary mADSC was characterized by FACS analysis using fluorescently-labeled receptor-specific antibodies against a panel of available mouse chemokine receptors. The analysis showed that the mADSC express the limited repertoire of functional receptors, consistent with the published profiles from mouse adipose tissue (Fig. 1b). The overall level of receptor expression showed consistent reproducibility in all tested mADSC cultures. The expression of chemokine receptors appeared heterogeneous. A small percentage of cells was positive for CCR5 (4.5 ± 1.1 %), CCR6 (4.1 ± 3.8 %), CCR8 (0.9 ± 2.8 %), CXCR1 (3.0 ± 2.4 %), CXCR2 (2.5 ± 3.2 %), CXCR3 (4.2 ± 3.1 %), CXCR6 (0.3 ± 3.5 %), CXCR7 (2.2 ± 1.1 %), and CX3CR1 (2.3 ± 3.1 %). A higher percentage of cells was positive for CCR2 (7.8 ± 4.1 %), CCR3 (14.0 ± 4.3 %), CCR4 (6.5 ± 2.1 %), CCR7 (10.9 ± 5.2 %), CCR9 (8.2 ± 2.9 %), CCR10 (6.1 ± 4.1 %), CXCR4 (7.0 ± 1.7 %), and CXCR5 (8.7 ± 3.5 %) (Fig. 1c). Primary mADSC culture thus showed a heterogeneous phenotype, in which functional chemokine receptors are not expressed uniformly and only limited yet uncharacterized subsets of ADSC express specific receptors at relatively low levels, suggesting that only a small fraction of cells is available for effective recruitment of transplanted cells into the skin. Chemotactic recruitment of mADSC to DEB-affected blistering and nonblistering skin Analysis of the blistering and nonblistering skin in 1–2-day-old Col7a1–/– and wild-type neonatal mice demonstrated a significant infiltration of the DEB-affected skin with leukocytes (Fig. 2a). Infiltrates were detected at both ventral skin and the skin of the extremities, the most common sites of blister formation in DEB mice. However, it was particularly noticeable in the blistering skin where leukocyte infiltrate extended into the blister cavity (Fig. 2a). Immunophenotyping analysis showed that about 30 % of CD45+ leukocytes in the wild-type skin were represented by the CD11b+ cells. The percentage of CD11b+ cells in the DEB-affected nonblistering skin was slightly elevated whereas in blistering skin it was two times higher (Fig. 2a, b).Fig. 2 Analysis of wild-type and DEB-affected skin infiltration with leukocytes. a H&E staining and indirect immunofluorescent analysis of skin infiltration with leukocytes. Detected antigens are shown in respective colors above the panels (CD45, green; CXCR2, green; CD11b, red; colocalization of CD45 or CXCR2 and CD11b, yellow). Nuclei were counterstained with DAPI (blue). bc blister cavity. Scale bar = 100 μm. b Quantitation of the skin-infiltrating cells. Data presented as an average number or percentage of cells (as indicated on the Y axis) per microscopic field ± SD. *p <0.05. **p > 0.05. Data collected from at least 15 independent microscopic fields. CD-positive cell types are shown above the columns. H&E hematoxylin and eosin, WT wild-type skin, NB nonblistering, DEB dystrophic epidermolysis bullosa, Bl blistering DEB-affected skin Considering elevated levels of CXCR2 ligands, CXCL1, CXCL2, and CXCL5, in blister fluids, skin-infiltrating CXCR2+ cells were further quantitated. As expected, in normal mouse skin about 24 % of CD45+ leukocytes expressed CXCR2. In DEB-affected nonblistering skin, this percentage was higher, approaching 45 %. In the blistering DEB skin, more than 60 % of all leukocytes were CXCR2+. The prevalence of CD11b+ cells in cutaneous infiltrates and the high percentage of CXCR2+CD11b+ cells in the blistering skin suggest that higher levels of the CXCR2 ligands, CXCL1, CXCL2, and CXCL5, in blister fluids may direct intracutaneous migration of the CXCR2+ leukocytes. These findings also suggest that CXCR2 could mediate homing of the systemically infused therapeutic type VII collagen-producing mADSC to the blistering and nonblistering DEB-affected skin. Cxcr2-mediated homing of the mADSC to the DEB-affected skin FASC analysis showed the presence of a small percentage of CXCR2-expressing mADSC (2.5 ± 3.2 %) in the total pool of minimally cultured cells, suggesting that homing of systemically administered mADSC to the skin could be enhanced via the use of cells uniformly expressing CXCR2 receptor. Thus, we generated a stable cell line of mADSC expressing firefly luciferase (FLuc+-mADSC) and a derivative line coexpressing FLuc and mouse Cxcr2 genes (FLuc+CXCR2+-mADSC), respectively (Fig. 3a). One cohort of wild-type newborn mice (n = 3 per cohort) and two cohorts of Col7a1–/– newborn animals (n = 3 per cohort) received an intraperitoneal (IP) injection of FLuc+CXCR2+-mADSC and parental FLuc+-mADSC (5 × 105 cells per injection), respectively. Twenty-four hours after injection, FLuc signals were assessed using the AVIS live-imaging system. As expected, FLuc activity was found associated with the intraperitoneal cavity and spleen in all recipient mice (Fig. 3b). Importantly, FLuc activity signals were readily detectable in the blister sites located on the paws and the ventral skin of mice treated with FLuc+CXCR2+-mADSC but not in animals receiving FLuc+-mADSC transplant. An additional 24 hours resulted in the accumulation of the FLuc+CXCR2+-mADSC to the blister sites (Fig. 3b). H&E staining revealed an increased number of the skin-infiltrating cells at the blistering skin of mice receiving FLuc+CXCR2+-mADSC. Infiltrate was mostly noticeable at both ventral-associated and paw-associated blistering sites where penetration of infiltrating cells was readily seen in the blister cavity (Fig. 3c). Indirect immunofluorescent detection of the FLuc confirmed the presence of the transplanted FLuc+CXCR2++-mADSC in the blistering skin, whereas unselected FLuc+-mADSC were rarely detected at these sites (Fig. 3d). Double-immunostaining showed that 40–60 % of CXCR2+ cells express FLuc in ventral and paw-blistering skin (Fig. 3c, d), suggesting that more than 50 % of skin-recruited cells are represented by the transplanted FLuc+CXCR2+-mADSC. Further analysis of FLuc+CXCR2+-mADSC transplant showed the persistent presence of type VII collagen and FLuc double-positive cells in the upper dermis as well as within the blister cavity. These double-positive cells represented about 40–50 % of all transplanted FLuc+ cells (Fig. 3d, e). Significantly, short stretches of type VII collagen protein were also detected at the basal layer of the epidermis (Fig. 3d). Contrary, type VII collagen protein was undetectable in the skin of mice receiving unselected FLuc+-mADSC transplant. Collectively, these studies demonstrated that transplantation of mADSC uniformly expressing CXCR2+ receptor, unlike unselected counterparts, results in efficient recruitment of stem cells from circulation directly to the blistering skin, where cells exert their therapeutic function by production of collagen type VII protein in the proximity of the cutaneous BMZ.Fig. 3 Systemic transplantation of genetically engineered CXCR2+-mADSC into DEB-affected newborn mice. a Characterization of CXCR2 expression in primary mADSC and cells engineered to overexpress FLuc (FLuc+-mADSC) and FLuc and Cxcr2 genes (FLuc+CXCR2+-mADSC), respectively, by FACS and indirect immunofluorescent detection. Analyses were performed using gene-specific antibodies. (Fluc, red; CXCR2, green; colocalization of FLuc and CXCR2, yellow; nuclei were counterstained with DAPI). b In-vivo imaging of systemically administered FLuc+CXCR2+-mADSC and native FLuc+-mADSC into newborn Col7a1 –/– and wild-type mice, respectively, 24 and 48 hours after transplantation (as indicated on the left). White and orange arrows point to the blister location. Type of transplanted cells indicated below the image of a representative mouse. c Representative micrographs of H&E-stained sections of the skin collected from blistering sites (ventral skin and paws) of FLuc+-mADSC and FLuc+CXCR2+-mADSC transplanted animals, respectively. Black arrowheads point to infiltrating cells. bc blister cavity. d Indirect immunofluorescent analysis of mADSC recruitment to blistering sites (indicated to the left of the panels). Type of mADSC used for transplantation indicated above the panels. Detected antigens, FLuc, CXCR2, and Col7, are shown below the panels in corresponding colors. Scale bar = 100 μm. bc blister cavity, white arrows point to the stretches of the BMZ-associated type VI collagen. e Quantitation of cell recruitment to blistering sites. Y axis represents the percentage of positive cells per microscopic field. The FLuc+, CXCR2+, and Col7+ cells are shown below the columns. Data shown as mean ± SD. *p <0.05. **p > 0.05. DEB dystrophic epidermolysis bullosa, mADSC mouse adipose-derived stem cells Enrichment of the Cxcr2+-mADSC population for skin homing The obtained data demonstrate that genetically engineered CXCR2+-expressing mADSC can be effectively targeted to the blistering skin due to engagement and activation of CXCR2 by the DEB skin-producing ligands, CXCL1, CXCL2, and CXCL5. However, to make this approach clinically relevant, more generalized stem cell selection strategy to enrich the cells expressing lead receptors is required. To obtain a CXCR2+-enriched population, antibody-mediated positive selection of mADSC on paramagnetic microbeads with consequent removal of the beads via cleavage of the antibody-bead linker using CEllection biotin binder kit was tested. However, this approach did not produce a desirable number of viable mADSC (data not shown). Previously, we observed that tissue culture conditions can significantly affect secretion of chemokines and chemokine receptors in bone marrow-derived MSC. In fact, assessment of chemokine receptors on the surface of freshly isolated and minimally cultured cells demonstrated expression of multiple receptors on rather small populations of the cells. Further analysis of the cell surface expression of CXCR2+ and other potential epithelia-targeting receptors on mADSC cultured at different densities and in the presence/absence of selected growth factors/cytokines was thus tested. FACS-based assessment showed that when cell proliferation was repressed using a contact inhibition approach by plating high-density culture exceeding confluency three times and continued cultured for additional 2 days, a substantial induction of several chemokine receptors was detected (Fig. 4a). Exposure of dense cultures to granulocyte macrophage colony-stimulating factor (GM-CSF; 100 ng/ml) did not alter expression of chemokine receptors (data not shown). However, culturing of cells with epidermal growth factor (EGF; 20 ng/ml) for 48 hours boosted cell surface expression of CCR4 and CXCR4, whereas treatment with basic fibroblast growth factor (bFGF; 25 ng/ml) induced CCR3, CCR4, CCR6, and CXCR2 (Fig. 4b). Interleukin-12 (IL-12; 25 ng/ml) preferentially induced CCR3, CCR4, CCR9, and CXCR2. Interestingly, transforming growth factor beta (TGF-β; 10 ng/ml) affected cell surface expression of all examined receptors with predominant induction of CCR3, CCR6, and CXCR2. Averagely, exposure of cells to TGF-β led to the enrichment of the CXCR2+-mADSC population ranging from 40 to 55 %. Considering the abundance of CXCR2 ligands in blister fluids, expression of CCR3 ligands (CCL11 and CCL24) in the skin, and expression of CCR4 ligands (CCL17 and CCL22) in dermal microvasculature, treatment of cells with TGF-β was selected as a stimulatory factor for CXCR2+-mADSC enrichment.Fig. 4 Characterization of cell surface chemokine receptors on mADSC after treatment with selected growth factors and cytokines. a FACS-based quantitation of the chemokine receptor-positive populations of mADSC under different culture conditions. Selected chemokine receptors identified on the surface of the cells indicated above the profiles, treatment conditions indicated on the panel: 1× confluent culture, 3× high-density culture. Treatments that increase populations of chemokine receptor-positive cells are indicated to the right of the profiles. b Percentage of receptor-positive cells in confluent cultures. Treatment conditions are shown to the right of the graphs. Data shown as mean ± SD from three independent experiments. *p < 0.05 compared with control (1×) cells Transplantation of the CXCR2-enriched mADSC into DEB-affected mice To see whether the selected CXCR2+-mADSC provide an ADSC population with a functional receptor capable of directing extravasation at the blistering skin, unselected and CXCR2+-enriched mADSC were delivered via IP injection (1 × 106 cells/injection/animal) into newborn 1-day-old Col7a1–/– mice (n = 3), respectively. For transplantation of CXCR2+-mADSC, cells on early passage (passages 2–3) were plated in subconfluent density and cultured in the presence of TGF-β for 48 hours. To visualize the recruitment of cells to the blistering skin along natural CXCL1 + 2 and CXCL5 chemotactic axes, cells were labeled with Vibrant DiI red fluorescent tracker. As expected, during the first 24 hours transplanted cells were detected mostly at the abdomen, the site of injection, in both transplanted groups (Fig. 5a). However, DiI signals were readily detectable at blistering sites on the chest and neck of the mice receiving CXCR2+-mADSC. Col7a1–/– recipients receiving unselected mADSC transplants demised shortly after transplantation on days 3 and 4. Analysis of the blistering skin of mice receiving unselected mADSC did not show any appreciable recruitment of mADSC to the blistering skin. In contrast, all recipients transplanted with CXCR2+-mADSC survived for at least 1 week without obvious signs of deterioration, although animals started to show a runted phenotype. At day 7, migration of CXCR2+-mADSC from the intraperitoneal cavity to the blistering paws and neck was observed as judged by accumulation of DiI fluorescent signals. Direct immunofluorescent evaluation of the blistering skin sections confirmed the presence of DiI+CXCR2+ cells detectable in the mid and upper dermis and in close proximity to the blister cavity (Fig. 5b). Indirect immunofluorescent analysis revealed type VII collagen-positive transplanted cells just below the epidermis with few cells detectable at the basal layer. Double-immunostaining with antibodies specific for type IV collagen showed a clear colocalization of both proteins at the BMZ (Fig. 5c), suggesting functionality of the donated protein and partial restoration of the DEJ. Quantitation of the collagen VII-associated fluorescence by measuring integrated density at the DEJ showed that transplantation of the CXCR2+-mADSC led to an increase of the type VII collagen corresponding to 45 % of the wild-type level with the mean gray value corresponding to 65 % of the wild-type skin. Collectively, the in-vivo studies demonstrated that an enriched population of CXCR2+-mADSC can efficiently migrate toward the blistering skin characterized by the elevated levels of CXCR2 ligands and donate functional therapeutic type protein into the BMZ of DEB-affected mice.Fig. 5 Transplantation of CXCR2-enriched mADSC into DEB-affected mice. a In-vivo live imaging of neonatal DEB mice and 2 and 7 days after IP transplantation of the DiI-labeled native and CXCR2+-mADSC (red), respectively. b Direct fluorescent detection of the DiI-mADSC in the blistering skin of the recipient mice at 7 days after transplantation. Blister cavity (bc) outlined by dotted line. c Indirect immunofluorescent detection of type VII and type IV collagens in the skin of mADSC-treated and control mice (indicated above the panels). Detected antigens (in corresponding color) are shown to the left of the panels. Upper row, monochrome images of the sections stained with anti-type VII collagen antibodies. White arrowheads point to the BMZ. Blue, DAPI nuclear staining. Scale bar = 100 μm. DEB dystrophic epidermolysis bullosa, mADSC mouse adipose-derived stem cells Discussion Several clinical studies were conducted to evaluate the utility of the ASC in alleviating DEB-associated symptoms [9, 10, 30, 31]. Although some degree of success was reported, there is a growing body of evidence that the observed therapeutic outcome was mainly associated with the transplanted cell-derived paracrine effects rather than a direct contribution of the cells to the restoration of the BMZ integrity [10]. One of the potential pitfalls in achieving desirable therapeutic effects could be associated with the poor migration and homing of the systemically transplanted therapeutic cells to the DEB skin affected with blisters. Understanding the mechanisms regulating migration of systemically transplanted ASC is crucial to the success of any clinical strategy. Here we report a detailed study to address the chemotactic responsiveness of mADSC to natural chemotactic gradients in mice affected with severe DEB. To identify key players responsible for disease-targeting of systemically infused cells to the blistering skin, chemotactic signatures of the DEB-associated blister fluids were investigated by employing proteome profile screens. The content of chemotactic molecules in blister fluids showed consistently high levels of several chemokines, including CCL6, CCL8, CCL9/10, CCL12, CXCL1, CXCL2, and CXCL5. All identified molecules are prominent chemotactic factors that activate and attract various inflammatory cells, including lymphocytes, macrophages, neutrophils, and granulocytes. These data directly suggest that the presence of elevated levels of identified chemokines can provide favorable chemotactic axes (CCL8–CCR1, CCL12–CCR2, CXCL1 + 2–CXCR2, and CXCL5–CXCR2) for efficient recruitment of the therapeutic stem cells from the systemic compartment to the DEB-affected cutaneous tissue. From a clinical perspective, the CCL6–CCR1 and CCL9/10–CCR1 axes can be excluded because the ligands have only been described in rodents. Because several of the upregulated chemokines (CXCL1, CXCL2, and CXCL5) interact with CXCR2 chemokine receptor, natural recruitment of the CXCR2-positive leukocytes to the blistering and nonblistering skin of the DEB-affected mice was further assessed. Analysis demonstrated that up to 40 % and 70 % of all leukocytes in nonblistering and blistering skin, respectively, express CXCR2 on the cell surface, suggesting that CXCR2 ligands play an important role in recruiting CXCR2+ leukocytes to the blistering sites. Systemic transplantation of the mADSC engineered to uniformly express CXCR2 confirmed this notion and demonstrated preferential recruitment of the CXCR2+-mADSC cells to blistering skin when compared with a native cell population. This observation is in good agreement with prior studies when chemotactic molecules were used to target stem cells to specific organs [32, 33]. Recently, we showed that MSC overexpressing CCR4+ and CCR10+ could be recruited to epithelial tissue toward an ectopic gradient of endothelial cell-derived CCL17 and epithelia-derived CCL27 axes and take part in regeneration/repair of the skin [7]. Our findings supported the idea that CCR4 engagement is essential for binding to the endothelium and extravasation, whereas activation of CCR10 is crucial for the directional migration of the cells within the collagenous matrix. Also, these data suggested distinct yet synergistic CCR4-dependent and CCR10-dependent mechanisms of stem cell recruitment to the skin. There has been much work to investigate the role of the CXCL12–CXCR4 axis to target MSC to diverse tissues, including ischemic heart, brain, skeletal muscle, kidney, liver, and skin [34]. To date, this is the most studied axis in MSC homing to wounds [24, 35]. Despite this interest, the majority of these studies investigated cell homing to sites of burn injury or radiation. In a limited number of investigations in cutaneous wounds, it was reported that skin grafts can recruit bone marrow-derived MSC through stromal derived factor-1α (SDF-1α, also known as CXCL12)/CXCR4 interaction to enhance tissue regeneration. Data suggested that during a skin graft hypoxic damage to the skin results in release of a soluble SDF-1α, which in turn recruits CXCR4+ MSC from circulation to the graft. However, considering wide expression of CXCL12 in multiple organs and uninjured tissues, reduced targeting specificity can significantly dilute the therapeutic effect. Furthermore, HMGB1 released from the detached or blistered DEB epithelia was suggested to be a factor in mobilizing Lin–/PDGFRα+ bone marrow cells into circulation and recruitment of cells to the epidermis and dermis of grafted Col7-deficient skin on the back of GFP-BMT mice and acceleration of skin regeneration [36]. Although an interesting experimental concept, recruitment of PDGFRα+ cells by HMGB1 does not provide any homing specificity and strictly depends on the apoptotic and necrotic status of the tissue. Also, there are no in-vivo data supporting systemic recruitment of PDGFRα+ cells into the skin of DEB mice. Based on our comprehensive whole genome oligo microarray profiles, the primary ADSC naturally express PDGFRα+ [6]. Moreover, analysis of blister fluids from patients affected with DEB showed an extremely low level of PDGFRα+ ligand, PDGF-BB (data not yet published). However, as shown in this study and our prior work [7], transplantation of primary heterogeneous stem cells failed to provide any appreciable recruitment of cells into the blistering and nonblistering skin, suggesting that PDGFRα+ alone is not sufficient to drive effective migration of cells from circulation into the skin. Taken together, high levels of CXCL1, CXCL2, and CXCL5 ligands observed in DEB-associated blister fluids can provide stronger and more specific chemotactic attraction of systemically administered cells expressing cognate CXCR2 receptor on its surface to the DEB-affected cutaneous tissue. Generation of a stable mADSC cell line with engineered CXCR allowed achieving effective homing of systemically infused cells to blistering skin; however, genetic manipulation with cells may greatly complicate clinical application. To develop a strategy aimed at isolation and enrichment of the skin-homing ADSC, two approaches were considered: antibody-based separation of CXCR2+ cells on magnetic beads; and alteration of cell culture protocols. The former approach provided a good yield of freshly isolated cells after selection on paramagnetic microbeads but failed to produce sufficient number of viable cells following enzymatic cleavage of cells from the beads. The latter strategy, however, demonstrated that transient changes in cell culture condition may substantially enrich a population of cells expressing desired chemokine receptor. In fact, dense cultures produced induction of several chemokine receptors, including CCR3, CCR4, CCR6, CCR9, CXCR2, and CXCR4. Moreover, exposure of cells to TGF-β significantly increased population of cells expressing CXCR2. TGF-β, alternatively, resulted in induction of other potential skin-homing receptors, including CCR4 and CCR6, whereas incubation with bFGF altered CCR3 and CXCR2 expression. The presented data thus demonstrate that expression of lead receptors on mADSC could be altered by temporary changes in tissue culture conditions and suggest that the stem cells with disease-organ-specific homing capabilities could be generated, avoiding time-consuming, labor-retaining, and expensive selection protocols. At present, very limited information is available regarding the effect of various growth factors and cytokines on ADSC-specific cell surface receptor composition. The majority of studies are focused on optimization of stem cell proliferation, survival, and differentiation. In fact, TGF-β, PDGF, and FGF were shown to play an important role in proliferation of ADSC [37]. Currently, the focus of stem cell biology is primarily shifted on the trophic effects elicited by stem cells. Growing data suggest that therapeutic benefit may not be restricted to the repair function of cells alone, but also due to their transient paracrine actions. In fact, current research suggests that stem cells can secrete potent combinations of trophic factors, such as cytokines, chemokines, and growth factors, which modulate the molecular composition of the environment to evoke responses from resident cells [38]. Although important, the data presented here also show great promise and suggest the potential of chemotaxis-driven therapy to target exogenous cells to distal anatomical sites such as the skin, emphasizing the need for better cell selection strategies. The functionality of the CXCR2+-mADSC population enriched by exposing cells to TGF-β was further evaluated in a preclinical mouse model of severe DEB. The cells were systemically transplanted into neonatal mice via IP injection. Although intravenous (IV) injection would be more suitable for the systemic transplantation, an extremely high incidence of pulmonary embolism in newborn DEB mice did not allow pursuing the IV route. Nevertheless, IP injection led to a rapid recruitment and accumulation of the CXCR2+-mADSC in the blistering skin. Moreover, migrated cells were capable of secreting type VII collagen, with partial reconstitution of the damaged BMZ. Importantly, an estimate of the fluorescence intensity indicated that the type VII collagen level corresponded to about 50 % of the wild-type level. Prior studies demonstrated that 10 % of the BMZ-associated type VII collagen is enough to stabilize the skin of Col7a1f lNeo/f lNeo hypomorphic mice [29]. Moreover, injection of 20 × 106 fibroblasts resulted in an increase of type VII collagen to approximately 30 % of the wild-type level and provided resistance of the skin against shearing forces [29]. Our group also demonstrated that intradermal transplantation of only 0.5 × 106 congenic MSC led to an increase of type VII collagen to about 15 % of the wild-type level as well as restoration of the damaged DEJ and protection of the skin from mechanical damage [8]. The data presented here demonstrated that greater deposition levels of therapeutic protein could be achieved after systemic transplantation of a rather small number (0.5 × 106cells/injection) of CXCR2+-enriched mADSC, emphasizing the great potential of this approach for clinical application. Also, these data are in a good agreement with recent findings showing that preconditioning of MSC with TGF-β, TNF-α, and CXCL12 increases secretion of type VII collagen [39], suggesting that pretreatment of cells with a combination of factors stimulating both the lead receptor and the therapeutic protein may provide better outcome of transplantation therapy. Conclusions In summary, the current study provided experimental proof that the CXCR2+-mADSC population possesses blistering skin-homing capabilities allowing targeting of the therapeutic cells directly to DEB-affected skin along a natural chemotactic axis and deposition of type VII collagen into the BMZ with ultimate restoration of the integrity at the DEJ. Currently ongoing in-vitro and in-vivo studies will further refine the procedures for the enrichment of human ADSC with skin-homing capabilities and the elevated secretion of type VII collagen and other proteins necessary for proper assembly of the functional collagen fibers for future allogeneic transplantation to rescue the DEB phenotype in the patients. Further mechanistic studies to identify critical factors involved in chemotactic-based directional migration of the stem cells will hopefully play an important role in designing rational approaches toward increasing the efficiency of targeting the disease site. Abbreviations ASCAdult stem cells BMZBasement membrane zone BSABovine serum albumin CCLChemokine (C–C motif) ligand CCRChemokine (C–C motif) receptor CXCLChemokine (C–X–C motif) ligand CXCRChemokine (C–X–C motif) receptor DEBDystrophic epidermolysis bullosa DEJDermal–epidermal junction DMEM/F12Dulbecco’s modified Eagle medium EDTAEthylenediaminetetraacetic acid FACSFluorescence-activated cell sorting FBSFetal bovine serum H&EHematoxylin and eosin mADSCMouse adipose-derived stem cells MSCMesenchymal stem cells PBSPhosphate-buffered saline RT-PCRReverse-transcriptase polymerase chain reaction Acknowledgements Not applicable. Funding Research reported in this publication was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under Award Number R01AR064286 (to OI). 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 datasets supporting the conclusions of this article are included within the article. Authors’ contributions VA participated in the design of experiments, carried out the molecular analysis of blister fluids, cell transplantation into animals, interpretation and analysis of in-vitro and in-vivo data, and helped to draft the manuscript. AD participated in all experiments involving animals, including colony maintenance, genotyping, collection of biopsies, histological and immunofluorescence analyses, and helped to draft the manuscript. JU conceived of the study, participated in its design, and helped to draft the manuscript. OI was involved in all aspects of the study, including experimental design, characterization of cells, proteome analysis study, transplantation studies, analysis and interpretation of data, and manuscript writing. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethical approval and consent to participate Not applicable. ==== Refs References 1. Fine JD Eady RA Bauer EA Bauer JW Bruckner-Tuderman L Heagerty A Hintner H Hovnanian A Jonkman MF Leigh I McGrath JA Mellerio JE Murrell DF Shimizu H Uitto J Vahlquist A Woodley D Zambruno G The classification of inherited epidermolysis bullosa (EB): Report of the Third International Consensus Meeting on Diagnosis and Classification of EB J Am Acad Dermatol. 2008 58 931 50 10.1016/j.jaad.2008.02.004 18374450 2. Uitto J Bruckner-Tuderman L Christiano AM McGrath JA Has C South AP Kopelan B Robinson EC Progress toward treatment and cure of epidermolysis bullosa: summary of the DEBRA international research symposium EB 2015 J Invest Dermatol. 2016 136 352 8 10.1016/j.jid.2015.10.050 26802230 3. Wagner W Wein F Seckinger A Frankhauser M Wirkner U Krause U Blake J Schwager C Eckstein V Ansorge W Ho AD Comparative characteristics of mesenchymal stem cells from human bone marrow, adipose tissue, and umbilical cord blood Exp Hematol. 2005 33 1402 16 10.1016/j.exphem.2005.07.003 16263424 4. Jeong JA Hong SH Gang EJ Ahn C Hwang SH Yang IH Han H Kim H Differential gene expression profiling of human umbilical cord blood-derived mesenchymal stem cells by DNA microarray Stem Cells. 2005 23 584 93 10.1634/stemcells.2004-0304 15790779 5. Silva WA Jr Covas DT Panepucci RA Proto-Siqueira R Siufi JL Zanette DL Santos AR Zago MA The profile of gene expression of human marrow mesenchymal stem cells Stem Cells. 2003 21 661 9 10.1634/stemcells.21-6-661 14595126 6. Alexeev V Arita M Donahue A Bonaldo P Chu ML Igoucheva O Human adipose-derived stem cell transplantation as a potential therapy for collagen VI-related congenital muscular dystrophy Stem Cell Res Ther. 2014 5 21 10.1186/scrt411 24522088 7. Alexeev V Donahue A Uitto J Igoucheva O Analysis of chemotactic molecules in bone marrow-derived mesenchymal stem cells and the skin: Ccl27-Ccr10 axis as a basis for targeting to cutaneous tissues Cytotherapy 2013 15 171 84. e1 10.1016/j.jcyt.2012.11.006 23321329 8. Alexeev V Uitto J Igoucheva O Gene expression signatures of mouse bone marrow-derived mesenchymal stem cells in the cutaneous environment and therapeutic implications for blistering skin disorder Cytotherapy. 2011 13 30 45 10.3109/14653249.2010.518609 20854215 9. Wagner JE Ishida-Yamamoto A McGrath JA Hordinsky M Keene DR Woodley DT Chen M Riddle MJ Osborn MJ Lund T Dolan M Blazar BR Tolar J Bone marrow transplantation for recessive dystrophic epidermolysis bullosa N Engl J Med. 2010 363 629 39 10.1056/NEJMoa0910501 20818854 10. Petrof G Lwin SM Martinez-Queipo M Abdul-Wahab A Tso S Mellerio JE Slaper-Cortenbach I Boelens JJ Tolar J Veys P Ofuya M Peacock JL Martinez AE McGrath JA Potential of systemic allogeneic mesenchymal stromal cell therapy for children with recessive dystrophic epidermolysis bullosa J Invest Dermatol. 2015 135 2319 21 10.1038/jid.2015.158 25905587 11. Nagy N Almaani N Tanaka A Lai-Cheong JE Techanukul T Mellerio JE McGrath JA HB-EGF induces COL7A1 expression in keratinocytes and fibroblasts: possible mechanism underlying allogeneic fibroblast therapy in recessive dystrophic epidermolysis Bullosa J Invest Dermatol. 2011 131 1771 4 10.1038/jid.2011.85 21471992 12. Ojeh N Pastar I Tomic-Canic M Stojadinovic O Stem cells in skin regeneration, wound healing, and their clinical applications Int J Mol Sci. 2015 16 25476 501 10.3390/ijms161025476 26512657 13. Shingyochi Y Orbay H Mizuno H Adipose-derived stem cells for wound repair and regeneration Expert Opin Biol Ther. 2015 15 1285 92 10.1517/14712598.2015.1053867 26037027 14. Ratajczak MZ Kim C The use of chemokine receptor agonists in stem cell mobilization Expert Opin Biol Ther. 2012 12 287 97 10.1517/14712598.2012.657174 22263752 15. Melve GK Ersvssr E Kittang AO Bruserud O The chemokine system in allogeneic stem-cell transplantation: a possible therapeutic target? Expert Rev Hematol. 2011 4 563 76 10.1586/ehm.11.54 21939423 16. Honczarenko M Le Y Swierkowski M Ghiran I Glodek AM Silberstein LE Human bone marrow stromal cells express a distinct set of biologically functional chemokine receptors Stem Cells. 2006 24 1030 41 10.1634/stemcells.2005-0319 16253981 17. Chamberlain G Fox J Ashton B Middleton J Concise review: mesenchymal stem cells: their phenotype, differentiation capacity, immunological features, and potential for homing Stem Cells. 2007 25 2739 49 10.1634/stemcells.2007-0197 17656645 18. Binger T Stich S Andreas K Kaps C Sezer O Notter M Sittinger M Ringe J Migration potential and gene expression profile of human mesenchymal stem cells induced by CCL25 Exp Cell Res. 2009 315 1468 79 10.1016/j.yexcr.2008.12.022 19168060 19. Kalwitz G Endres M Neumann K Skriner K Ringe J Sezer O Sittinger M Haupl T Kaps C Gene expression profile of adult human bone marrow-derived mesenchymal stem cells stimulated by the chemokine CXCL7 Int J Biochem Cell Biol. 2009 41 649 58 10.1016/j.biocel.2008.07.011 18707017 20. Stich S Haag M Haupl T Sezer O Notter M Kaps C Sittinger M Ringe J Gene expression profiling of human mesenchymal stem cells chemotactically induced with CXCL12 Cell Tissue Res. 2009 336 225 36 10.1007/s00441-009-0768-z 19296133 21. Ringe J Strassburg S Neumann K Endres M Notter M Burmester GR Kaps C Sittinger M Towards in situ tissue repair: human mesenchymal stem cells express chemokine receptors CXCR1, CXCR2 and CCR2, and migrate upon stimulation with CXCL8 but not CCL2 J Cell Biochem. 2007 101 135 46 10.1002/jcb.21172 17295203 22. Novak L Igoucheva O Cho S Alexeev V Characterization of the CCL21-mediated melanoma-specific immune responses and in situ melanoma eradication Mol Cancer Ther. 2007 6 1755 64 10.1158/1535-7163.MCT-06-0709 17575105 23. Tubo NJ McLachlan JB Campbell JJ Chemokine receptor requirements for epidermal T-cell trafficking Am J Pathol. 2011 178 2496 503 10.1016/j.ajpath.2011.02.031 21641376 24. Avniel S Arik Z Maly A Sagie A Basst HB Yahana MD Weiss ID Pal B Wald O Ad-El D Fujii N Arenzana-Seisdedos F Jung S Galun E Gur E Peled A Involvement of the CXCL12/CXCR4 pathway in the recovery of skin following burns J Invest Dermatol. 2006 126 468 76 10.1038/sj.jid.5700069 16385346 25. Ichimura M Hiratsuka K Ogura N Utsunomiya T Sakamaki H Kondoh T Abiko Y Otake S Yamamoto M Expression profile of chemokines and chemokine receptors in epithelial cell layers of oral lichen planus J Oral Pathol Med. 2006 35 167 74 10.1111/j.1600-0714.2006.00402.x 16454813 26. Meller S Winterberg F Gilliet M Muller A Lauceviciute I Rieker J Neumann NJ Kubitza R Gombert M Bunemann E Wiesner U Franken-Kunkel P Kanzler H Dieu-Nosjean MC Amara A Ruzicka T Lehmann P Zlotnik A Homey B Ultraviolet radiation-induced injury, chemokines, and leukocyte recruitment: an amplification cycle triggering cutaneous lupus erythematosus Arthritis Rheum. 2005 52 1504 16 10.1002/art.21034 15880822 27. Sasaki M Abe R Fujita Y Ando S Inokuma D Shimizu H Mesenchymal stem cells are recruited into wounded skin and contribute to wound repair by transdifferentiation into multiple skin cell type J Immunol. 2008 180 2581 7 10.4049/jimmunol.180.4.2581 18250469 28. Heinonen S Mannikko M Klement JF Whitaker-Menezes D Murphy GF Uitto J Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa J Cell Sci. 1999 112 3641 8 10523500 29. Fritsch A Loeckermann S Kern JS Braun A Bosl MR Bley TA Schumann H von Elverfeldt D Paul D Erlacher M Berens von Rautenfeld D Hausser I Fassler R Bruckner-Tuderman L A hypomorphic mouse model of dystrophic epidermolysis bullosa reveals mechanisms of disease and response to fibroblast therapy J Clin Invest. 2008 118 1669 79 10.1172/JCI34292 18382769 30. Conget P Rodriguez F Kramer S Allers C Simon V Palisson F Gonzalez S Yubero MJ Replenishment of type VII collagen and re-epithelialization of chronically ulcerated skin after intradermal administration of allogeneic mesenchymal stromal cells in two patients with recessive dystrophic epidermolysis bullosa Cytotherapy. 2010 12 429 31 10.3109/14653241003587637 20230217 31. El-Darouti M Fawzy M Amin I Hay RA Hegazy R Gabr H El Maadawi Z Treatment of dystrophic epidermolysis bullosa with bone marrow non-hematopoeitic stem cells: a randomized controlled trial Dermatol Ther. 2016 29 96 100 10.1111/dth.12305 26439431 32. Hirota K Yoshitomi H Hashimoto M Maeda S Teradaira S Sugimoto N Yamaguchi T Nomura T Ito H Nakamura T Sakaguchi N Sakaguchi S Preferential recruitment of CCR6-expressing Th17 cells to inflamed joints via CCL20 in rheumatoid arthritis and its animal model J Exp Med. 2007 204 2803 12 10.1084/jem.20071397 18025126 33. Apostolaki M Manoloukos M Roulis M Wurbel MA Muller W Papadakis KA Kontoyiannis DL Malissen B Kollias G Role of beta7 integrin and the chemokine/chemokine receptor pair CCL25/CCR9 in modeled TNF-dependent Crohn’s disease Gastroenterology. 2008 134 2025 35 10.1053/j.gastro.2008.02.085 18439426 34. Lau TT Wang DA Stromal cell-derived factor-1 (SDF-1): homing factor for engineered regenerative medicine Expert Opin Biol Ther. 2011 11 189 97 10.1517/14712598.2011.546338 21219236 35. Toksoy A Muller V Gillitzer R Goebeler M Biphasic expression of stromal cell-derived factor-1 during human wound healing Br J Dermatol. 2007 157 1148 54 10.1111/j.1365-2133.2007.08240.x 17941943 36. Iinuma S Aikawa E Tamai K Fujita R Kikuchi Y Chino T Kikuta J McGrath JA Uitto J Ishii M Iizuka H Kaneda Y Transplanted bone marrow-derived circulating PDGFRalpha + cells restore type VII collagen in recessive dystrophic epidermolysis bullosa mouse skin graft J Immunol. 2015 194 1996 2003 10.4049/jimmunol.1400914 25601922 37. Josh F Tobita M Tanaka R Orbay H Ogata K Suzuki K Hyakusoku H Mizuno H Concentration of PDGF-AB, BB and TGF-beta1 as valuable human serum parameters in adipose-derived stem cell proliferation J Nippon Med Sch. 2013 80 140 7 10.1272/jnms.80.140 23657067 38. Baraniak PR McDevitt TC Stem cell paracrine actions and tissue regeneration Regen Med. 2010 5 121 43 10.2217/rme.09.74 20017699 39. Perdoni C McGrath JA Tolar J Preconditioning of mesenchymal stem cells for improved transplantation efficacy in recessive dystrophic epidermolysis bullosa Stem Cell Res Ther. 2014 5 121 10.1186/scrt511 25376815

PMC005xxxxxx/PMC5002133.txt

==== Front BMC CancerBMC CancerBMC Cancer1471-2407BioMed Central London 270610.1186/s12885-016-2706-2Research ArticleSurvival, recurrence and toxicity of HNSCC in comparison of a radiotherapy combination with cisplatin versus cetuximab: a meta-analysis Huang Jingwen 673515950@qq.com 1Zhang Jing zhangjing-med@foxmail.com 3Shi Changle 475515456@qq.com 3Liu Lei liuleihx@gmail.com 2Wei Yuquan yuquawei@vip.sina.com 21 Department of Medical Oncology, Cancer Center, West China Hospital, Sichuan University, Chengdu, China 2 State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, West China Medical School Sichuan University, Chengdu, China 3 West China Medical School, West China Hospital, Sichuan University, Chengdu, China 26 8 2016 26 8 2016 2016 16 1 68914 12 2015 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 Cisplatin-based treatment has been considered the standard treatment regimen of HNSCC. Cetuximab is an emerging target therapy that has potential therapeutic benefits over cisplatin. Nevertheless, curative effects of cisplatin-based chemoradiotherapy (CRT) versus cetuximab-based bioradiotherapy (BRT) are still controversial. Methods Potentially eligible studies were retrieved using PubMed, Embase and Medline. Basic characteristics of patients and statistical data were collected. A meta-analysis model was established to compare CRT and BRT. Results Thirty-one eligible studies and 4212 patients were found. The pooled HRs with 95 % confidence intervals (CIs) for OS and PFS were 0.32 [0.09, 0.55] and 0.51 [0.22, 0.80], respectively, and both were in favor of cisplatin. However, 3-year survival and recurrence analysis of the subgroups showed no differences between the two groups (p > 0.05). In subgroup analysis, oropharyngeal primary tumors exhibited improved results by cetuximab with a pooled HR of 1.56 [1.14, 2.13] for PFS. Additionally, the HPV+ status was a significant factor in positive outcomes with cetuximab with a pooled HR of 1.12 [0.46, 2.17] for OS. Conclusion Long-term use of BRT showed no significant difference compared with CRT, and both arms showed different aspects of toxicity. In subgroup analysis, taking the effects of treatment and adverse events into consideration, cetuximab plus radiation may show superior responses regarding OS and PFS in patients who have HPV+ or primary oropharyngeal HNSCC, respectively, but physicians should administer them with caution. Keywords HNSCCOropharynxHPVCisplatinCetuximabRadiotherapyPrognosisRecurrenceAdverse eventNational Natural Science Foundation of China, Beijing, China81101991Research Award Fund for New Young Teachers in Higher Education Institutions, China20120181120024issue-copyright-statement© The Author(s) 2016 ==== Body Background Squamous cell carcinoma of the head and neck (HNSCC) consists of cancers arising from the oral cavity, pharynx and larynx and comprises approximately 5 % of all cancers worldwide. The global incidence is increasing by half a million and causing more than 350,000 deaths every year [1, 2]. A limited number of patients with locally advanced disease are suitable for potentially curative surgery or definitive radiotherapy. Patients who are not candidates for surgery or definitive radiotherapy may receive chemotherapy plus radiation or systemic chemotherapy alone [3]. Cisplatin-based chemoradiotherapy is now considered to be the established standard, first-line chemotherapy to treat patients with locally advanced HNSCC [69]. Many large randomized studies and meta-analyses have demonstrated that cisplatin-based concurrent chemoradiotherapy regimens provide significantly higher response rates than radiotherapy alone [4, 5]. Epidermal growth factor receptor (EGFR) seems to be critical to cancer cell growth and proliferation, and the function of EGFR in these two settings appears to be different [6, 7]. Head and neck cancer cells exhibit this difference compared to normal cells without exception [8]. In addition, EGFR expression was markedly increased or over-expressed in HNSCC compared to normal tissue, which has been shown to be an independent prognostic factor for poor survival [9]. Thus, EGFR inhibitors have become a burgeoning strategy in anti-tumor treatment. To date, several monoclonal antibodies targeting EGFR have been successfully used in clinical practice with significant effects. Improved loco-regional control and prolonged survival time have already been achieved in lung and gastro-intestinal cancers [10–12]. Cetuximab, an EGFR-targeting monoclonal antibody, is the first targeted therapy to show therapeutic benefit in head and neck cancer [13] and received FDA approval for use in treating HNSCC in 2006 [14, 15]. The Bonner trial showed impressively increased survival outcomes and loco-regional control rates when comparing cetuximab plus radiation versus radiation alone [16]. The Merlano trial exhibited a promising treatment response from adding cetuximab to standard chemotherapy, with limited toxicity [17]. Clinical trials have shown that the addition of cetuximab to traditional treatment regimens (e.g., cisplatin plus radiation) could improve survival outcomes [18, 19]. However, this combination may lead to increased treatment-related toxicity and increased cost, and the administration of multiple drugs may worsen quality of life. Hence, we conducted a meta-analysis with the aims of gathering outcomes from clinical trials and obtaining a larger sample size to compare the curative effects between the administration of cisplatin-based chemoradiotherapy (CRT) or cetuximab-based bioradiotherapy (BRT) with regards to survival results, loco-regional control or distant metastasis (failure), and treatment-related adverse effects in patients with HNSCC. Methods Search strategy PubMed, Embase and Medline were searched on Mar 13, 2016. The following keywords were used to retrieve articles and abstracts: head and neck squamous cell carcinoma (HNSCC), cancers of larynx, cancers of oral tongue, cancers of oropharynx, cancers of laryngopharynx, cetuximab, cisplatin and radiotherapy. Study selection and inclusion/exclusion criteria Titles and abstracts were reviewed in all of the searched studies, and full texts were reviewed in potentially eligible studies according to our inclusion criteria. To avoid duplicated data, when more than one trial was completed with crossed data in a single center, only the largest most updated trials were included. In our meta-analysis, we used the following inclusion criteria: (1) studies containing patients with locally advanced HNSCC, including the following: cancers of the larynx, cancers of the oral tongue, cancers of the oropharynx, or cancers of the laryngopharynx; (2) studies comparing the administration of cisplatin-based chemotherapy versus cetuximab-based biotherapy; and (3) studies with available data regarding survival outcomes of patients included in the clinical trials. On the other hand, studies were excluded based on the following criteria: (1) articles that consisted of in vitro studies or were review articles; (2) studies with duplicated data, meaning that one analysis that had several articles reporting updated outcomes; and (3) studies containing metastatic and/or recurrent disease. Data extraction The following two investigators reviewed all of the articles independently: Huang JW and Shi CL. Any discrepancy was discussed until reaching a consensus. The data were independently extracted from eligible studies by two investigators (Huang JW and Shi CL), and then, the obtained data were integrated. The primary data consisted of HRs with a 95 % confidence interval (CI) or event/total patient numbers regarding survival outcomes, including OS and/or PFS and the recurrence rates, such as loco-regional and/or distant recurrence of disease in patients from cetuximab cohorts and cisplatin cohorts. The additional data obtained from the studies included the first author, publication year, patient source (region), median age, percentage of each sex, TNM stage at diagnosis, treatment regimens, tumor site (%), survival outcomes, recurrence rates, type of study, toxicity N (G3 ~ 4) in CRT vs. BRT groups, and attitude of the original studies. The statistical data for acquiring logHR and SE were also obtained, including HR with a 95 % CI, Kaplan–Meier survival curves with p values, and response rates of the over-expression cohort compared to the normal/lower expression cohort [20]. Statistical methods logHR and SE were required in our analysis. Some of the original papers provided logHR and SE directly, whereas other studies did not. As mentioned above, we utilized other data to calculate these values using methods developed by Parmar et al. (1998) [21], Williamson et al. (2002) [22], and Tierney et al. (2007) [23]. The logHRs and SEs were calculated with the methods described earlier when 1) there was a HR with 95 % CI or 2) there was a p value for the log-rank test with the Kaplan–Meier survival curve. Hazard ratio (HR) was used as the measure index to describe the survival outcomes and disease control rates between the BRT arm and CRT arm (we considered the cisplatin regimen as the standard regimen). As a result of the analysis of survival in patients, a significant outcome was defined by a p value < 0.05, while a p value > 0.05 indicated no significant difference between the two comparison arms. Pooled HRs > 1 combined with p < 0.05 indicate a narrow difference between the two groups, and the cetuximab arm showed higher event incidences. In contrast, pooled HRs < 1 indicated a lower incidence of events in the cetuximab cohort. Furthermore, pooled HRs > 2 or <0.5 denote a significant result. We use the term “positive” to indicate a better outcome related to cetuximab treatment and “negative” to indicate an absence of correlation between the two comparison arms or better outcomes in the cisplatin arm. In terms of heterogeneity, values of p < 0.10 or I2 > 50 % represent heterogeneity existing in the pooled HRs (Higgins et al., 2003) [24]. When homogeneity was minimal (p ≥ 0.10, I2 ≤ 50 %), a fixed-effects model was applied for secondary analysis; otherwise, a random-effects model was used. All of the earlier calculations and publication bias were measured using the Begg’s funnel plot, which was performed by STATA 11.0 (STATA Corporation, College Station, TX). This calculation for the current meta-analysis was performed using REVIEW MANAGER (version 5.0 for Windows; the Cochrane collaboration, Oxford, UK). The sensitive analysis, which aims to test for the heterogeneity of all of the included studies and to determine if heterogeneity arose from any single study, was performed by STATA 11.0 (STATA Corporation, College Station, TX). In the analytic figure, an absence of heterogeneity is indicated by the containment of the studies within the constricted interval (defined between lower CI limit and Upper CI limit), while the existence of a single study far outside the confidence interval indicates that the heterogeneity is due to that individual study. Results Eligible studies We initially obtained 794 studies from PubMed, Embase and Medline. After reviewing these abstracts, 73 potentially relevant studies were identified as candidates for a full-text review. We excluded 42 studies for the following reasons: twenty-one were clinical trials focused on CRT vs. CRT plus cetuximab, four were reviews, three were posters without follow-up statistics on the studies, and seven were in vitro studies (Fig. 1).Fig. 1 Selection of Studies Finally, we enrolled 31 eligible articles containing survival outcomes [25–50]. These eligible studies were published from 2008 to 2016 and included a total of 4212 patients, ranging from 24 to 421 patients per study (median, 126). The basic clinical characteristics of patients and other useful information are shown in Table 1.Table 1 Basic characteristics of included studies Author Type of study Region Age Stage Female (%) Therapy regimens HPV statue (+) Survival Cisplatin Cetuximab Cisplatin Cetuximab Cisplatin Cetuximab Vermorken JB RS America CRT: 57.8 BRT: 57 Stage III/IV NR RT-CIS VS. RT-CET NR 2-yrOS 1-yr PFS Caudell JJ RS America CRT: 55 BRT: 54 Stage III/V BRT: 20.7 % CRT: 22.3 % CCRT VS. Concurrent RT+CET NR 1-yr OS 2-yr OS L.D. Koutcher RS America NR RT-CIS VS. RT-CET 16 (42 %) 8 (35 %) 30-mon PFS 30-mon OS Jensen AD RS Germany CRT: 38 BRT: 38 Stage III/IV NR RT-CIS VS. RT-CET NR 2-yr OS 2-yr PFS 2-yr L-PFS 2-yr D-PFS Koutcher L RS America CRT: 56 BRT: 66 Stage III/IV CRT:17 (13.6 %) BRT:11 (22.5 %) RT-CIS VS. RT-CET NR 2-yr FFS 2-yr OS 2-yr LRC Beijer, Y.J. RS Netherland Primary: 56 Primary: 64 Stage II-IV CRT Primary: 37 CET Primary: 43 RT-CIS VS. RT-CET NR 1-yr OS 1-yr DFS 2-year OS 2-yr DFS LRR Adjuvant: 59 Adjuvant: 56 CRT Adjuvant: 36 CET Adjuvant: 36 Ley J RS America CRT: 55 BRT: 62 Stage III/IV CRT: 16.7 BRT: 34.5 RT-CIS VS. RT-CET NR 3-yr DSS 3-yr LRR Ye AY RS Canada CRT: 57 BRT: 62 Stage III/IV CRT: 17 BRT: 14 RT-CIS VS. RT-CET NR 3-yr OS 3-yr DFS 3-yr LRC Pajares B RS Spain p16 Negative: 59 p16 positive: 57 Stage III/IV p16 Negative:7 p16 Positive: 6 RT-CIS VS. RT-CET 10 (18 %) 8 (15 %) 2-yr OS 2-yr DFS 2-yr LRR Lefebvre JL Phase II RCT France CRT: 57.5 BRT: 57.8 Stage II-IV CRT: 13.3 BRT: 1.7 RT-CIS VS. RT-CET NR 18-mon OS 18-mon LRR 36-mon OS M. Ghi RCT Italy 60 Stage III/IV 80.5 CCRT VS. Cet+RT NR 3-yr OS 3-yr PFS N. Riaz RCT America NR NR NR CCRT VS. Cet+RT 24 (56%) 11 (75%) NR Hu MH RCT Taiwan CRT: 55 BRT: 78 Stage III/IV CRT: 3.4 BRT: 3.7 CCRT VS. Cet+RT LRR 3-yr RFS 3-yr OS DM Levy A RCT Germany CRT: 58 BRT: 60 Stage III/IV CRT: 20 BRT: 23 CCRT VS. BRT NR 2-yr OS 2-yr LRC 2-yr DM Tang C RCT America CRT: 58 BRT: 73 Stage I-IV CRT: 10 BRT: 1 CCRT VS. Concurrent Cet+RT NR 2-yr LRC 2-yr EFS 2-yr OS Fayette J RS France 56 Stage III/IV 10 CCRT VS. Concurrent Cet+RT NR 5-yr OS 5-yr DFS Huang J RS Japan CRT: 55 BRT: 77 Stage III/IV IMRT/ cisplatin: 13 IMRT/ cetuxima:19 IMRT/CIS VS. IMRT/CET NR LRC DM OS CSS Shapiro LQ RS America NR stage II-IV CRT:13.1 BRT: 22.4 IMRT/CIS VS. IMRT/CET NR 4-yr OS 4-yr LRF M.R. Kanakamedala RS America 53 NR NR RT-CIS VS. RT-CET NR LRC 3-yr OS 2-yr PFS Riaz N RS America NR NR NR RT-CIS VS. RT-CET NR NR Riaz N RS America NR Stage III/IV CRT: 21 BRT: 22 RT-CIS VS. RT-CET 31 (86 %) 17 (74 %) NR Peddi P RS America CRT: 55 BRT: 61 Stage III/IV CRT: 26.7 BRT: 29.7 CCRT VS. Concurrent RT-CET NR 2-yr OS 2-yr PFS S.L.Galper RS America CRT: 58 BRT: 71 NR NR RT-CIS VS. RT-CET NR NR D.Borchiellini RS France CRT: 56 BRT: 57 NR CRT 16 % BRT 8 % RT-CIS VS. RT-CET NR NR Lorraine Walsh RS Ireland CRT: 57.5 BRT: 63 Stage III/IV CRT : 9 % BRT : 11.8 % RT-CIS VS. RT-CET NR NR Stefano Maria Magrini RCT America CRT: 67.5 BRT: 61 Stage III/IV CRT: 31 % BRT:26 % RT-CIS VS. RT-CET NR 2-yr OS Tobin J. Strom RS America CRT: 58 BRT: 62 Stage III/IV CRT: 16.2 BRT: 5.3 RT-CIS VS. RT-CET 43.4 % 41.2 % 2-yrOS Nadeem Riaz NR America CRT: 118 < 71 7 >71 BRT: 38 < 71 11>71 NR CRT : 21 % BRT: 22 % RT-CIS VS. RT-CET 86 74 3-yr LRC 3-yr OS 3-yr PFS CRT cisplatin-based chemoradiotherapy, BRT cetuximab-based bioradiotherapy, RT-CIS radiation plus cisplatin, RT-CET radiation plus cetuximab, CCRT concurrent chemoradiotherapy, yr year, mon month, HPV Human papillomavirus, RS retrospective study, RCT randomized controlled study, OS overall survival, PFS progression free survival, L-PFS local progression free survival, D-PFS distant progression free survival, FFS failure-free survival, LRR locoregional recurrence, DFS disease free survival, DSS disease specific survival, LRC locoregional control, RFS relapse-free survival, DM distant metastasis, EFS event-free survival, CCS cause-specific survival, CAD/CVD coronary artery disease/cardiovascular disease, COPD chronic obstructive pulmonary disease, PNS peripheral nervous system, NR not reference Comparison between cisplatin-based and cetuximab regarding overall survival Twenty-three settings of accommodated data showed patients’ overall survival (OS). In these trials, patients were scheduled to receive cisplatin-based chemotherapy plus radiation or cetuximab single agent plus radiation. The pooled HRs to compare OS between the two groups showed better outcomes with cisplatin-based therapy and the mathematic value is 0.32 [0.09, 0.55], p = 0.006 (Table 2; Fig. 2).Table 2 Pooled HRs (95 % Cl) comparing survival outcomes and recurrence between BRT & CRT Comparison Survival outcome Study N. Model HR (95 % Cl) P value Heterogeneity (p ,I2) Conclusion BRT vs. CRT OS 23 Random 0.32 [0.09, 0.55] 0.006 P < 0.00001; I² = 84.6 % Positive BRT vs. CRT OS for 2-yr 11 Random 0.44 [0.13, 0.76] 0.006 P < 0.0001; I² = 76.9 % Positive BRT vs. CRT OS for 3-yr 12 Random 0.21 [-0.14, 0.55] 0.241 P < 0.00001; I² = 88.8 % Negative BRT vs. CRT PFS 21 Random 0.51 [0.22, 0.80] 0.001 P < 0.00001; I² = 90.1 % Positive BRT vs. CRT PFS for 2-yr 10 Random 0.56 [0.20, 0.92] 0.002 P < 0.00001; I² = 88.2 % Positive BRT vs. CRT PFS for 3-yr 11 Random 0.45 [-0.05, 0.95] 0.076 P < 0.00001; I² = 91.8 % Negative BRT vs. CRT Locoregional control 19 Random 0.49 [0.14, 0.85] 0.007 P < 0.00001; I² = 91 % Positive BRT vs. CRT Locoregional control for 2-yr 9 Random 0.63 [0.09, 1.17] 0.023 P < 0.00001; I² = 83 % Positive BRT vs. CRT Locoregional control for 3-yr 10 Random 0.06 [-0.40, 0.52] 0.808 P < 0.00001; I² = 93.3 % Negative BRT vs. CRT Distant control 5 Random 0.25 [0-0.06, 0.56] 0.118 P < 0.00001; I² = 88.3 % Negative BRT vs. CRT OS for oropharynx 7 Random 0.13 [-0.03, 0.89] 0.743 P < 0.00001; I² = 84.8 % Negative BRT vs. CRT PFS for oropharynx 3 Random 1.56 [1.14, 2.13] 0.006 P < 0.00001; I² = 96 % Positive BRT vs. CRT Locoregional control for oropharynx 6 Random 1.75 [0.6, 5.26] 0.31 P < 0.00001; I² = 89.1 % Negative BRT vs. CRT OS for HPV+ 5 Fixed 1.12 [0.46, 2.17] 0.015 P = 0.22; I² = 38 % Positive BRT vs. CRT PFS for HPV+ 5 Random 0.80 [0.38, 1.67] 0.55 P < 0.00001; I² = 92 % Negative BRT vs. CRT Locoregional control for HPV+ 5 Random 1.17 [0.69, 2.00] 0.56 P = 0.01; I² = 71.1 % Negative CRT cisplatin-based chemoradiotherapy, BRT cetuximab-based bioradiotherapy, N number, OS overall survival, PFS progression-free survival, CI confidence interval, HR hazard ratio, yr year Fig. 2 Meta-analysis estimated OS comparing cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy. (a) subgroup of estimation of 2-yr OS; (b) subgroup of estimation of 3-yr OS. OS, overall survival Subgroup analysis As survival outcomes were largely influenced by time of observation, we categorized OS outcomes by year of estimation: 2-years, 3-years, or 5-years and beyond. The pooled HR for 2-year estimation was 0.44 [0.13, 0.76], p = 0.006, which supports better survival achieved with cisplatin-based therapy, while the 3-year or 5-year and beyond time assessments showed no significant difference between the two groups, with pooled HRs of 0.21 [-0.14, 0.55], p = 0.241and 0.95 [0.51, 1.74], p = 0.86, respectively (Table 2; Fig. 2). Human papillomavirus (HPV) infection state might contribute to pathogenesis of HNSCC, and it has previously been demonstrated that HPV positive (HPV+) cases showed better prognosis and prolonged survival in the cetuximab single agent group. The pooled HR is 1.12 [0.46, 2.17], p = 0.015 (Table 2; Fig. 3).Fig. 3 Meta-analysis compared cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy in estimating patients in HPV+ subgroup regarding to OS, PFS, and LRC. OS, overall survival; PFS, progression-free survival; LRC, locoregional control The oropharynx was shown to be distinct in prognosis and therapy response compared with HNSCC in other locations. On this account, we analyzed this group separately, and the results showed that patients with primary tumors in the oropharynx exhibited similar values of OS with a pooled HR of 0.13 [-0.03, 0.89], p = 0.743 (Table 2; Fig. 4).Fig. 4 Meta-analysis compared cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy in estimating patients with oropharyngeal primary tumor regarding to OS, PFS and LRC. OS, overall survival; PFS, progression-free survival; LRC, locoregional control Comparison between cisplatin-based and cetuximab therapies regarding progression-free survival Twenty-one studies published data including progression free survival (PFS). The PFS results displayed a similar tendency as the OS and the mathematic value for the pooled HR was 0.51 [0.22, 0.80], p = 0.001 (Table 2; Fig. 5).Fig. 5 Meta-analysis estimated PFS comparing cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy. (a) subgroup of estimation of 2-yr PFS; (b) subgroup of estimation of 3-yr PFS. PFS, progression-free survival Subgroup analysis As assessed in the OS data, we categorized PFS outcomes by time intervals of estimation: 2-years, or 3-years and beyond. The pooled HRs were 0.56 [0.20, 0.92], p = 0.002, and 0.45 [-0.05, 0.95], p = 0.076 for the 2-year and 3-years and beyond time assessments, respectively, which indicate that better survival was achieved with cisplatin-based therapy (Table 2; Fig. 5). For the HPV+ group, cetuximab-based therapy again showed outcomes superior to those of cisplatin-based therapy, and the pooled HR was 0.80 [0.38, 1.67], p = 0.55 (Table 2; Fig. 3). We also analyzed PFS separately in patients with oropharynx tumors, and those patients who received cetuximab-based regimens showed prolonged PFS compared with administration of cisplatin-based therapy; the pooled HR was 1.56 [1.14, 2.13], p = 0.006 (Table 2; Fig. 4). Comparison between cisplatin-based and cetuximab therapies regarding loco-regional containment Nineteen studies reported loco-regional control or loco-regional failure in patients with HNSCC. The pooled HR to compare OS between the two groups showed better outcomes with cisplatin-based therapy, and the mathematic value was 0.49 [0.14, 0.85], p = 0.007 (Table 2; Fig. 6).Fig. 6 Meta-analysis estimated LRC comparing cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy. (a) subgroup of estimation of 2-yr LRC; (b) subgroup of estimation of 3-yr LRC. LRC, locoregional control Subgroup analysis Loco-regional control, like other recurrence rates and survival outcomes, directly correlated to the estimated time interval, and thus, we categorized the loco-regional control rates by the year of estimation: 2-years, 3-years, or 5-years and beyond. The pooled HR for the 2-year estimation was 0.63 [0.09, 1.17], p = 0.023, which supports better survival achieved with cisplatin-based therapy, while the 3-years or 5-years and beyond time assessments showed no significant difference between the two groups, and the pooled HRs were 0.34 [-0.12, 0.79], p = 0.15 and 2.67 [0.47, 8.73], p = 0.27, respectively (Table 2; Fig. 6). Patients with HPV+ infection states showed a non-significantly better prognosis and prolonged survival in the cetuximab single agent group, and the pooled HR was 0.06 [-0.40, 0.52], p = 0.808 (Table 2; Fig. 3). Analysis of patients with primary tumors in the oropharynx showed no significant difference between the cisplatin and cetuximab groups with a pooled HR of -0.05 [-1.34, 0.35], p = 0.248 (Table 2; Fig. 4). Comparison between cisplatin-based and cetuximab therapies regarding distant metastasis Five studies reported incidences of distant metastases. The pooled HR was 0.25 [0-0.06, 0.56], p = 0.118, indicating no significant difference between cisplatin and cetuximab administration (Table 2; Fig. 7).Fig. 7 Meta-analysis estimated DM comparing cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy. DM, distant metastasis Assessment of adverse events Twenty-one types of acute toxicity or late toxicity in patients treated with cisplatin plus radiotherapy or cetuximab plus radiotherapy were assessed. The pooled HRs of all toxicities, including acute and late toxicities, showed no difference for patients who received cisplatin-based or cetuximab-based therapy, and the mathematic value was -0.34 [-0.72, 0.04], p = 0.079 (Fig. 8).Fig. 8 Meta-analysis compared cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy in estimating toxicities including acute and late toxicities Subgroup analysis We estimated individual toxicities separately, which is shown in Fig. 9. We found that incidence of toxicities, such as leukopenia (p = 0.00), acute kidney injury (p = 0.002), and neutropenia (p = 0.002), were significantly higher in the cisplatin plus radiotherapy regimen, while some dermatitis-related toxicities, such as acneiform rash (p = 0.002), displayed a higher incidence in the cetuximab plus radiotherapy regimen. Other toxicities showed no statistical significance between the two groups (Table 3; Fig. 9).Fig. 9 Meta-analysis compared cisplatin-based chemoradiotherapy versus cetuximab-based bioradiotherapy in estimating patients with oropharyngeal primary tumor and with single toxicity separately Table 3 Pooled HRs 95 % Cl for toxicity between CRT & BRT Comparison Adverse event/ Toxicity Study N. Model HR 95 % [Cl] P value Heterogeneity (p ,I2) Conclusion CRT vs. BRT Mucositis 7 Fixed 0.05 [-0.09, 0.19] p=0.493 P = 0.45; I² = 36.9 % Negative CRT vs. BRT Dysphagia 5 Fixed -0.07 [-0.35, 0.21], p=0.63 P = 0.89; I² = 0 % Negative CRT vs. BRT Xerostomia 2 Fixed 0.51 [0.09, 2.95], p=0.46 P = 0.17; I² = 46 % Negative CRT vs. BRT Laryngeal edema 2 Fixed 0.91 [0.71, 1.18] p=0.49 P = 0.89; I² = 0 % Negative CRT vs. BRT Acute kidney injury 5 Fixed -1.30 [-2.11, -0.49] p=0.002 P = 0.32; I² = 0 % Positive CRT vs. BRT Nausea or vomiting 4 Random -1.30 [-2.66, 0.06], p=0.061 P = 0.03; I² = 57.2 % Negative CRT vs. BRT Radiation dermatitis 4 Random 0.31 [-0.45, 1.08] p=0.419 P = 0.001; I² = 87.6 % Negative CRT vs. BRT Acneiform rash 5 Random 3.49 [1.23, 5.74] P=0.002 P = 0.87; I² = 81 % Positive CRT vs. BRT Neutropenia 3 Fixed -0.88 [-1.42, -0.33] p=0.002 P < 0.00001; I² = 0.0 % Positive CRT vs. BRT Ototoxicity 3 Fixed 0.16 [0.04, 0.69] p=0.10 P = 0.60; I² = 0 % Negative CRT vs. BRT Infectious 2 Fixed 3.31 [0.55, 19.87] p=0.19 P = 0.59; I² = 0 % Negative CRT vs. BRT Neuropathy 2 Fixed 0.80 [0.46, 1.41] p=0.44 P = 0.37; I² = 0 % Negative CRT vs. BRT Pain 2 Fixed 0.92 [0.80, 1.06] p=0.24 P = 0.74; I² = 0 % Negative CRT vs. BRT Leukopenia 4 Fixed -0.76 [-1.16, -0.36] P=0.001 P = 0.19; I² = 44.2 % Positive CRT vs. BRT Late toxicity 4 Fixed 1.11 [0.83, 1.47], p=0.48 P = 0.53; I² = 0 % Negative CRT vs. BRT Total toxicity 21 Random -0.34 [-0.72, 0.04] P=0.079 P < 0.00001; I² = 91.7 % Negative CRT cisplatin-based chemoradiotherapy, BRT cetuximab-based bioradiotherapy, N number, CI confidence interval, HR hazard ratio, CRT chemoradiothrapy, BRT bioradiothrapy Results from sensitive tests As shown in Fig. 10, all of the scattered points were restricted within the interval of the lower CI and upper CI limitations, which indicated that the heterogeneity was acceptable and constrained (Fig. 10).Fig. 10 Sensitive analysis evaluated heterogeneity of OS cohort, PFS cohort and related subgroups. (a) Evaluation in OS group: total, 2-yr and 3-yr. (b) Evaluation in PFS group: total, 2-yr, and 3-yr. OS, overall survival; PFS, progression-free survival Assessment of publication bias On the basis of Begg’s funnel plot, the p value was greater than 0.10, which indicates that the publication bias was acceptable in the analysis. According to Begg’s funnel plot analysis, the publication bias arising in the OS cohort (p = 0.758), the PFS cohort (p = 0.90), the loco-regional control cohort (p = 0.83) or the distant metastasis cohort (p = 0.854) was acceptable (Fig. 11).Fig. 11 Estimated Begg’s funnel plots of publication bias regarding OS, PFS, LRC, and DM cohort respectively. OS, overall survival; PFS, progression-free survival; LRC, locoregional control; DM, distant metastasis Discussion In this systemic review, we conducted a meta-analysis to compare the effect of cisplatin-based chemotherapy plus radiotherapy versus cetuximab plus radiotherapy in controlling the overall survival, progression-free survival, loco-regional recurrence and distant metastasis of locally advanced HNSCC. Meanwhile, different time periods of estimation, primary tumor sites in the oropharynx and HPV infection status were also taken into consideration. Our study demonstrated that in all settings of the estimated OS time duration, the outcomes were found to be better with cisplatin treatment; however, specifically observing the longer follow-up time intervals, patients between the two groups shared similar overall survival rates, as there was no statistical significance between the two groups with a follow-up time duration equal to or longer than 3 years. Progression-free survival and loco-regional control rates displayed similar tendencies as the OS rates. In subgroup analysis, tumors with a primary site in the oropharynx and tumors with HPV+ infection status showed non significantly better PFS and OS, respectively, with cetuximab single agent treatment plus radiotherapy, while no remarkable difference was observed between the remaining survival outcomes and loco-regional control in the two subgroups, indicating that equivalent effects of the two treatment regimens were achieved in these categories. Concurrent cisplatin-based therapy has been regarded as the standard treatment regimen for patients with HNSCC [51]; however, cisplatin has been reported to cause immediate treatment-related adverse events and delayed toxicity. Cetuximab, an emerging monoclonal antibody therapeutic, targeting epidermal growth factor receptor (EGFR), seemed promising to provide patients with an effective alternative treatment [52]. Whether cetuximab could replace cisplatin in definitive chemoradiotherapy for HNSCC remains controversial because cetuximab has a reasonably good toxicity profile [53] but the tumor control effect and survival benefit present inconsistent results. Therefore, to achieve better quality of life and avoid these aggressive treatment regimens, concurrent cetuximab plus radiation versus cisplatin plus radiation therapies have been compared. A recent meta-analysis including 15 studies comparing CCRT and concurrent cetuximab with radiotherapy, with various estimation time intervals, suggested that cisplatin usage improved OS and PFS [54], consistent with our results. Nevertheless, it seems that these drug responses and effects will benefit patients in certain circumstances. In our analysis, we showed that patients from selected subgroups of HNSCC might benefit from concurrent cetuximab plus radiotherapy. In our analysis, we found that both the oropharyngeal primary tumor and HPV+ subgroups showed differences regarding survival outcomes, possibly supporting the utility of cetuximab to a large extent. We focused on HPV+ patients, as the biological behavior of these tumors showed particularity. HPV is now considered to be an independent and important risk factor in HNSCC [55, 56]. Recently, Dayyani et al. published a meta-analysis, which showed that HPV infection has a critical impact on survival and response to therapy, and they also demonstrated that HPV+ status was not rare (HPV+ 22 %, with 86.7 % exhibiting HPV16+ genotype) [57]. However, some negative outcomes also exist, and no significant difference was shown between cisplatin and cetuximab with radiation in LAHNC [58]. One major obstacle in this work was the lack of information regarding the HPV/p16 status; thus, we suggest that patients should undergo HPV testing for this unique and separate biologic entity. In our analysis, patients in the HPV+ group achieved better OS due to the highly selective and biologic characteristics, which made the HPV+ group more suitable for the concurrent BRT treatment regimen than the whole HNSCC group. We also observed unique responses in patients who had primary lesions in the oropharynx. One comprehensive study estimated chemotherapy effects via tumor sites, and the results showed increased benefits only for oropharyngeal and laryngeal tumors [59, 60]. There are well-established patient risk factors associated with HPV infection in oropharyngeal cancer, and a higher incidence of HPV infection was found in cancers of the oropharynx [61, 62]. In our analysis, better PFS was observed in the oropharyngeal group rather than all cases of HNSCC, which could support the administration of cetuximab as a single agent plus radiation in this specific subgroup. Adverse effects are important additional parameters to be taken into consideration when comparing treatment regimens. In our analysis, we found that there were no significant differences between the two groups for all toxicity data. The cisplatin regimen resulted in adverse events, including high-grade neutropenia, leukopenia and acute kidney injury, while adverse events due to BRT included grade 3-4 acne-like rash and oral mucositis. We found that the incidence of adverse events was elevated in advanced cases. One recent study published by Lawrence D. Koutcher et al, showed serious grade radiation dermatitis with spontaneous bleeding in patients undergoing the BRT regimen [63], which could further decrease quality of life [64] and have a negative impact on cosmetic outcomes [65]. As the total incidence of adverse events did not show significance between cisplatin and cetuximab and the two regimens cause different adverse events in different aspects, doctors need to take toxicity into consideration and choose regimens according to each patient’s condition. To further confirm the quantity of evidence of the analyzed the data, heterogeneity and sensitive analysis were examined, and no obvious heterogeneity was detected; as shown in every group estimation, I2 was <50 %, with no exception. In addition, the further sensitive analysis assessed heterogeneity in detail and revealed only limited heterogeneity. In addition, as this is a meta-analysis, some limitations still exist. Primarily, only published data from prospective or retrospective studies were included in our meta-analysis, without individual data. Therefore, we could only use these integrated data, which may lead to patient selection bias as patient selection and reporting processes could not be controlled by us. In addition, we combined both retrospective and randomized trials in our meta-analysis, which could also contribute to the bias of this meta-analysis as the inclusion criteria of these two types of studies may not be the same and result in mixed data bias. Additionally, in pooled-data calculation processes, we chose multivariate data, if they were available. Otherwise, our calculated data consisted of univariate data without adjusting for some other influencing factors, such as age, sex, and histologic grade. This would represent a source of bias because multivariate studies examine the prognostic value independently, while univariate studies consider single factor. Conclusion In spite of all of the limitations and biases of our meta-analysis, we conclude that long-term use of cetuximab plus radiation showed no significant difference compared with cisplatin plus radiation for all of the survival and toxicity data examined. In subgroup analysis, cetuximab plus radiation may show superior responses regarding OS and PFS in patients who have HPV+ or primary oropharyngeal HNSCC, respectively, but physicians should administer them with caution. This analysis is a combination of current data. Previously, it was thought that cetuximab could cause fewer side effects and may be preferable to cisplatin, as they showed similar survival outcomes. However, we showed that the two regimens caused toxicity without significant differences, while cisplatin treatments exhibited better survival outcomes. Thus, with all of the limitations, we recommend further RCTs to determine the utility of cetuximab in HNSCC, especially in the oropharyngeal and/or HPV+ specific subgroups. The authors thank the study participants in each of the individual studies for their involvement. Funding This study was supported by the National Natural Science Foundation of China, Beijing, China (Grant No. 81101991) and Research Award Fund for New Young Teachers in Higher Education Institutions, China (Grant No. 20120181120024). The funder had no role in the study design, data collection and analysis, decision to publish, or preparation of the article. Availability of data and materials All relevant data are within the paper and its Supporting Information files. Authors’ contribution This study was designed by LL and WYQ, and the manuscript was written by HJW. Data were collected by HJW and SCL. Statistical process was done by SCL and ZJ. Figures and tables were made by HJW and SCL. All authors have read and approved the manuscript, and ensure that this is the case. 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. Parkin DM Bray F Ferlay J Pisani P Global cancer statistics, 2002 CA Cancer J Clin 2005 55 2 74 108 10.3322/canjclin.55.2.74 15761078 2. Jemal A Bray F Center MM Ferlay J Ward E Forman D Global cancer statistics CA Cancer J Clin 2011 61 2 69 90 10.3322/caac.20107 21296855 3. Bar-Ad V Palmer J Yang H Cognetti D Curry J Current management of locally advanced head and neck cancer: the combination of chemotherapy with locoregional treatments Semin Oncol 2014 41 6 798 806 10.1053/j.seminoncol.2014.09.018 25499638 4. Pignon JP le Maitre A Maillard E MACH-NC Collaborative Group: meta-analysis of chemotherapy in head and neck cancer (MACH-NC): An update on 93 randomized trials and 17,346 patients Radiother Oncol 2009 92 4 14 10.1016/j.radonc.2009.04.014 19446902 5. Cooper JS PAJAK TF Forestiere AA Postoperative irradiation with or without concomitant chemotherapy for high-risk squamous cell carcinoma of the head and neck N Engl J Med. 2004 350 1937 1944 10.1056/NEJMoa032646 15128893 6. Rubin GJ Chakraborty A Melhem MF Zeng Q Tweardy DJ Inhibition of epidermal growth factor receptor gene expression and function decreases proliferation of head and neck squamous carcinoma but not normal mucosal epithelial cells Oncogene 1997 15 4 409 16 10.1038/sj.onc.1201188 9242377 7. Kalyankrishna S Grandis JR Epidermal growth factor receptor biology in head and neck cancer J Clin Oncol 2006 24 17 2666 72 10.1200/JCO.2005.04.8306 16763281 8. Ozanne B Richards CS Hendler F Burns D Gusterson B Over-expression of the EGF receptor is a hallmark of squamous cell carcinomas J Pathol 1986 149 1 9 14 10.1002/path.1711490104 2425067 9. Ma X Huang J Wu X Epidermal growth factor receptor could play a prognostic role to predict the outcome of nasopharyngeal carcinoma: a meta-analysis Cancer Biomark 2014 14 4 267 77 24934370 10. Keedy VL Temin S Somerfield MR American society of clinical oncology provisional clinical opinion: epidermal growth factor receptor (EGFR) Mutation testing for patients with advanced non-small-cell lung cancer considering first-line EGFR tyrosine kinase inhibitor therapy J Clin Oncol 2011 29 15 2121 7 10.1200/JCO.2010.31.8923 21482992 11. Sequist LV von PJ Garmey EG Randomized phase II study of erlotinib plus tivantinib versus erlotinib plus placebo in previously treated non-small-cell lung cancer J Clin Oncol 2011 29 24 3307 15 10.1200/JCO.2010.34.0570 21768463 12. Shah MA Jhawer M Ilson DH Phase II study of modified docetaxel, cisplatin, and fluorouracil with bevacizumab in patients with metastatic gastroesophageal adenocarcinoma J Clin Oncol 2011 29 7 868 74 10.1200/JCO.2010.32.0770 21189380 13. Mehra R, Cohen RB, Burtness BA.The role of cetuximab for the treatment of squamous cell carcinoma of the head and neck. Clin Adv Hematol Oncol. 2008;6(10):742–50 14. Kono SA Haigentz M Jr Yom SS Saba N EGFR Monoclonal antibodies in the treatment of squamous cell carcinoma of the head and neck: a view beyond cetuximab Chemother Res Pract. 2012 2012 901320 23056941 15. Cohen MH Chen H Shord S Approval summary: Cetuximab in combination with cisplatin or carboplatin and 5-fluorouracil for the first-line treatment of patients with recurrent locoregional or metastatic squamous cell head and neck cancer Oncologist 2013 18 4 460 6 10.1634/theoncologist.2012-0458 23576486 16. Bonner JA Harari PM Giralt J Radiotherapy plus cetuximab for locoregionally advanced head and neck cancer: 5-year survival data from a phase 3 randomised trial, and relation between cetuximab-induced rash and survival Lancet Oncol 2010 11 1 21 8 10.1016/S1470-2045(09)70311-0 19897418 17. Merlano M Russi E Benasso M Corvò R Colantonio I Cisplatin-based chemoradiation plus cetuximab in locally advanced head and neck cancer: a phase II clinical study Ann Oncol 2007 22 3 712 7 10.1093/annonc/mdq412 20810547 18. Ang KK Zhang Q Rosenthal DI Randomized phase III trial of concurrent accelerated radiation plus cisplatin with or without cetuximab for stage III to IV head and neck carcinoma: RTOG 0522 J Clin Oncol 2014 32 27 2940 50 10.1200/JCO.2013.53.5633 25154822 19. Burtness B Goldwasser MA Flood W Mattar B Forastiere AA Phase III randomized trial of cisplatin plus placebo compared with cisplatin plus cetuximab in metastatic/recurrent head and neck cancer: an Eastern cooperative oncology group study J Clin Oncol 2005 23 34 8646 54 10.1200/JCO.2005.02.4646 16314626 20. JPT Higgins SG. Cochrane handbook for systematic reviews of interventions. The Cochrane Library. 2005 21. Parmar MK Torri V Stewart L Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints Stat Med 1998 17 24 2815 34 10.1002/(SICI)1097-0258(19981230)17:24<2815::AID-SIM110>3.0.CO;2-8 9921604 22. Williamson PR Smith CT Hutton JL Marson AG Aggregate data meta-analysis with time-to-event outcomes Stat Med 2002 21 22 3337 51 10.1002/sim.1303 12407676 23. Tierney JF Stewart LA Ghersi D Burdett S Sydes MR Practical methods for incorporating summary time-to-event data into meta-analysis Trials. 2007 8 16 10.1186/1745-6215-8-16 17555582 24. Higgins JP Thompson SG Deeks JJ Altman DG Measuring inconsistency in meta-analyses BMJ 2003 327 7414 557 60 10.1136/bmj.327.7414.557 12958120 25. Walsh L Toxicity of cetuximab versus cisplatin concurrent with radiotherapy in locally advanced head and neck squamous cell cancer (LAHNSCC) Radiother Oncol 2011 98 1 38 41 10.1016/j.radonc.2010.11.009 21159400 26. Caudell JJ Sawrie SM Spencer SA Locoregionally advanced head and neck cancer treated with primary radiotherapy: a comparison of the addition of cetuximab or chemotherapy and the impact of protocol treatment Int J Radiat Oncol Biol Phys 2008 71 3 676 81 10.1016/j.ijrobp.2007.10.040 18355979 27. L. D. Koutcher ES, D. Carlson NK, M. Fury SW, Z. Zhang QM, D. Pfister. Cisplatin (CDDP) and Radiation (RT) versus Cetuximab (C) and RT in the Context ofHuman Papillomavirus (HPV) and P16 in the Treatment of Locally Advanced Head and Neck Cancer (LAHNC). 2012. Annual ASTRO Meeting. 28. Jensen AD Krauss J Weichert W Disease control and functional outcome in three modern combined organ preserving regimens for locally advanced squamous cell carcinoma of the head and neck (SCCHN) Radiat Oncol. 2011 6 122 10.1186/1748-717X-6-122 21942981 29. Koutcher L Sherman E Fury M Concurrent cisplatin and radiation versus cetuximab and radiation for locally advanced head-and-neck cancer Int J Radiat Oncol Biol Phys 2011 81 4 915 22 10.1016/j.ijrobp.2010.07.008 20947269 30. Beijer YJ Koopman M Terhaard CH Braunius WW van Es RJ de Graeff A Outcome and toxicity of radiotherapy combined with chemotherapy or cetuximab for head and neck cancer: our experience in one hundred and twenty-five patients Clin Otolaryngol 2013 38 1 69 74 10.1111/coa.12002 22891945 31. Ley J Mehan P Wildes TM Cisplatin versus cetuximab given concurrently with definitive radiation therapy for locally advanced head and neck squamous cell carcinoma Oncology 2013 85 5 290 6 10.1159/000355194 24217231 32. Ye AY Hay JH Laskin JJ Wu JS Ho CC Toxicity and outcomes in combined modality treatment of head and neck squamous cell carcinoma: cisplatin versus cetuximab J Cancer Res Ther 2013 9 4 607 12 10.4103/0973-1482.126455 24518704 33. Pajares B Trigo JM Toledo MD Differential outcome of concurrent radiotherapy plus epidermal growth factor receptor inhibitors versus radiotherapy plus cisplatin in patients with human papillomavirus-related head and neck cancer BMC Cancer. 2013 13 26 10.1186/1471-2407-13-26 23331666 34. Lefebvre JL Pointreau Y Rolland F Induction chemotherapy followed by either chemoradiotherapy or bioradiotherapy for larynx preservation: the TREMPLIN randomized phase II study J Clin Oncol 2013 31 7 853 9 10.1200/JCO.2012.42.3988 23341517 35. M. Ghi AP, R. Orecchia SP, F. Bertoni NM. A Phase 2-3 Study Comparing Concomitant Chemoradiation Therapy (CRT) Versus Cetuximab/RT (CET/RT) With or Without Induction Docetaxel/Cisplatin/5-Fluorouracil (TPF) in Locally-Advanced Head and Neck Squamous Cell Carcinoma (LASCCHN) e Efficacy Results of the GSTTC Italian Study (NCT01086826). Int J Radiat Oncol Biol Phys. 2013 [eAbstract] . 36. N. Riaz AB, D. Adkins SR, J. Huang PC, J. Ley DK, W. Thorstad NL. Multi-institution Analysis of Concurrent Chemoradiation Therapy With Cisplatin (CDDP) Versus Cetuximab (C225) in Locally-Advanced Squamous Cell Carcinoma of the Head and Neck (LA-HNSCC): Can HPV Help Decide Which Agent. Int J Radiat Oncol Biol Phys 2008 [eAbstract]. 37. Hu MH Wang LW Lu HJ Cisplatin-based chemotherapy versus cetuximab in concurrent chemoradiotherapy for locally advanced head and neck cancer treatment Biomed Res Int. 2014 2014 904341 25110705 38. Levy A Blanchard P Bellefqih S Concurrent use of cisplatin or cetuximab with definitive radiotherapy for locally advanced head and neck squamous cell carcinomas Strahlenther Onkol 2014 190 9 823 31 10.1007/s00066-014-0626-0 24638267 39. Tang C Chan C Jiang W Concurrent cetuximab versus platinum-based chemoradiation for the definitive treatment of locoregionally advanced head and neck cancer Head Neck 2015 37 3 386 92 10.1002/hed.23609 24431011 40. Fayette J Bonnin N Ferlay C Neoadjuvant TPF in locally advanced head and neck cancer can be followed by radiotherapy combined with cisplatin or cetuximab: a study of 157 patients Anticancer Drugs 2013 24 6 623 9 23542750 41. Huang J Baschnagel AM Chen P A matched-pair comparison of intensity-modulated radiation therapy with cetuximab versus intensity-modulated radiation therapy with platinum-based chemotherapy for locally advanced head neck cancer Int J Clin Oncol 2014 19 2 240 6 10.1007/s10147-013-0540-y 23479120 42. Shapiro LQ Sherman EJ Riaz N Efficacy of concurrent cetuximab vs. 5-fluorouracil/carboplatin or high-dose cisplatin with intensity-modulated radiation therapy (IMRT) for locally-advanced head and neck cancer (LAHNSCC) Oral Oncol 2014 50 10 947 55 10.1016/j.oraloncology.2014.07.001 25132089 43. Kanakamedala MR SPG, Abraham RS AAA, Vijayakumar S aRDH. Comparison of Cisplatin Versus Cetuximab With Intensity Modulated Radiation Therapy in Locally Advanced Squamous Cell Carcinomas of Head and Neck: Toxicity, Patterns of Failure, and Survival Analysis. Radiat Oncol. 2014;90(15). Annual ASTRO Meeting. 44. N. Riaz AB, D. Adkins SR, J. Huang PC, J. Ley DK, W. Thorstad NL. Multi-institution Analysis of Concurrent Chemoradiation Therapy With Cisplatin (CDDP) Versus Cetuximab (C225) in Locally Advanced Squamous Cell Carcinoma of the Head and Neck (LAHNSCC): Can HPV Help Decide Which Agent. Int J Radiat Oncol Supplement 2014 [eAbstract]. 45. Peddi P Shi R Nair B Ampil F Mills GM Jafri SH Cisplatin, cetuximab, and radiation in locally advanced head and neck squamous cell cancer: a retrospective review Clin Med Insights Oncol. 2015 9 1 7 10.4137/CMO.S18682 25628515 46. S. L. Galper HD, Decker MGRaRH. Cetuximab versus cisplatin concurrent with IMRT in locally advanced head and neck cancer (LAHNC). J Clin Oncol ASCO Annual Meeting. 2009. 27(e17030). 47. Delphine Borchiellini CD, Jocelyn Gal BL, Marc Alfonsi KB, Christian Righini BG, Beltran M. Induction TPF followed by chemoradiation or radiotherapy-cetuximab in nonresectable advanced head and neck squamous cell carcinoma: A retrospective study in 164 patients . Int J Radiat Oncol. 2012;30(e16044). 48. Strom TJ Comparison of every 3 week cisplatin or weekly cetuximab with concurrent radiotherapy for locally advanced head and neck cancer Oral Oncol 2015 51 7 704 8 10.1016/j.oraloncology.2015.04.012 25936651 49. Riaz N Concurrent chemoradiotherapy with cisplatin versus cetuximab for squamous cell carcinoma of the head and neck Am J Clin Oncol 2016 39 1 27 31 10.1097/COC.0000000000000006 24401670 50. Magrini SM Cetuximab and radiotherapy versus cisplatin and radiotherapy for locally advanced head and neck cancer: a randomized phase II trial J Clin Oncol 2016 34 5 427 35 10.1200/JCO.2015.63.1671 26644536 51. Network NCC NCCN head and neck cancers guidelines (Version 1, 2016) 2016 America NCCN 52. Robert F Ezekiel MP Spencer SA Meredith RF Phase I study of anti-epidermal growth factor receptor antibody cetuximab in combination with radiationtherapy in patients with advanced head and neck cancer J Clin Oncol 2001 19 13 3234 43 11432891 53. Bonner JA Harari PM Giralt J Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck N Engl J Med 2006 354 6 567 78 10.1056/NEJMoa053422 16467544 54. Petrelli F Coinu A Riboldi V Concomitant platinum-based chemotherapy or cetuximab with radiotherapy for locally advanced head and neck cancer: a systematic review and meta-analysis of published studies Oral Oncol 2014 50 11 1041 8 10.1016/j.oraloncology.2014.08.005 25176576 55. Fakhry C Westra WH Li S Improved survival of patients with human papillomavirus-positive head and neck squamous cell carcinoma in a prospective clinical trial J Natl Cancer Inst 2008 100 4 261 9 10.1093/jnci/djn011 18270337 56. Gillison ML, Harris J, Westra W, et al. Survival outcomes bytumor human papillomavirus (HPV) in stage III-IV oropharyngeal cancer (OPC) in RTOG 0129 [Abstract]. Proc Am SocClinOncol 2009;27(Suppl.):15 s . 57. Dayyani F Etzel CJ Liu M Ho CH Lippman SM Tsao AS Meta-analysis of the impact of human papillomavirus (HPV) on cancer risk and overall survival in head and neck squamous cell carcinomas (HNSCC) Head Neck Oncol. 2010 2 15 10.1186/1758-3284-2-15 20587061 58. Riaz N, Sherman E, Koutcher L, et al. Concurrent chemoradiotherapy with cisplatin versus cetuximab for squamous cell carcinoma of the head and neck. Am J Clin Oncol. 2014. 59. Blanchard P Baujat B Holostenco V Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): a comprehensive analysis by tumour site Radiother Oncol 2011 100 1 33 40 10.1016/j.radonc.2011.05.036 21684027 60. Lill C, Kornek G, Bachtiary B, Selzer E, et al. Survival of patients with HPV-positive oropharyngeal cancer after radiochemotherapy is significantly enhanced.Wien Klin Wochenschr. 2011;123(7–8):215–21. 61. Kreimer AR Clifford GM Boyle P Franceschi S Human papillomavirus types in head and neck squamous cell carcinomas worldwide: a systematic review Cancer Epidemiol Biomarkers Prev 2005 14 2 467 75 10.1158/1055-9965.EPI-04-0551 15734974 62. Termine N Panzarella V Falaschini S HPV in oral squamous cell carcinoma vs head and neck squamous cell carcinoma biopsies: a meta-analysis (1988-2007) Ann Oncol 2008 19 10 1681 90 10.1093/annonc/mdn372 18558666 63. Koutcher LD Wolden S Lee N Severe radiation dermatitis in patients with locally advanced head and neck cancer treated with concurrent radiation and cetuximab Am J Clin Oncol 2009 32 5 472 6 10.1097/COC.0b013e318193125c 19487913 64. Elliott EA Phase III Trial of an emulsion containing trolamine for the prevention of radiation dermatitis in patients with advanced squamous cell carcinoma of the head and neck: results of Radiation Therapy Oncology Group Trial 99-13 J Clin Oncol 2006 24 13 2092 7 10.1200/JCO.2005.04.9148 16648511 65. Dorr W Hendry JH Consequential late effects in normal tissues Radiother Oncol 2001 61 3 223 31 10.1016/S0167-8140(01)00429-7 11730991

PMC005xxxxxx/PMC5002134.txt

==== Front Virol JVirol. JVirology Journal1743-422XBioMed Central London 60310.1186/s12985-016-0603-6Short ReportCirculating serotypes of dengue virus and their incursion into non-endemic areas of Pakistan; a serious threat Ali Amjad +92-4235293135-36amjadaliswat@gmail.com 12Ahmad Habib drhahmad@gmail.com 2Idrees Muhammad idreeskhan@cemb.edu.pk 31Zahir Fazli fazlizahirmian@yahoo.com 4Ali Ijaz ijaz.ali@comsats.edu.pk 51 Centre for Applied Molecular Biology, University of the Punjab, Lahore, Lahore-53700 Pakistan 2 Department of Genetics, Hazara University Mansehra, Mansehra, Khyber Pukhtunkhwa Pakistan 3 Vice Chancellor, Hazara University Mansehra, Mansehra, Pakistan 4 IBGE, The University of Agriculture, Peshawar, Khyber Pukhtunkhwa Pakistan 5 Department of Biosciences, COMSATS Institute of Information Technology, Islamabad, Pakistan 26 8 2016 26 8 2016 2016 13 1 1448 5 2016 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 Dengue virus is circulating in Pakistan since 1994, which causes major and minor outbreaks in many areas of the country. The incidence of dengue in Pakistan in past years mainly restricted to parts of Sindh and Punjab provinces. As such, a severe dengue outbreak appeared in Pakistan in 2011, particularly in Punjab province with Lahore as the most hit city (290 deaths). In 2013, for the first time in the history of Pakistan, dengue outbreak erupted in Swat District, Khyber Pakhtunkhwa, which claimed more than 57 lives. Hence this study was conducted to document circulating serotypes of dengue virus in Pakistan in 2011 and 2013 dengue outbreaks in two different territories/areas of the country. Methods In total, 1340 blood samples from people having dengue (ELISA positive) and/or dengue like symptoms from various cities/areas of Punjab and Swat, Khyber Pakhtunkhwa (KP) were collected and analyzed by reverse transcription polymerase chain reaction (RT-PCR) using serotype specific primers. Results The results indicated that all the four dengue virus serotypes were circulating in Punjab Province with highest frequency of DENV-2 (41.64 %) and DENV-3 (41.05 %). Similarly, DENV-2 (41.66 %) and DENV-3 (35.0 %) were dominant serotypes detected in KP-based people lived in Punjab. On the other hand only DENV-2 (40.0 %) and DENV-3 (60.0 %) were detected in Swat District. Furthermore an important observation noted in this study was mixed infection of DENV-2 and DENV-3 in Punjab in 2011 (3.81 %) and in people from KP infected in Punjab (8.33 %) which may account for the high mortality and morbidity rates as compared to previous outbreaks. Over all male population was mostly infected as compared to females and people in the age group between 15 to 45 was the highest infected group. Conclusions The findings of this study indicate that all four serotypes of dengue virus are circulating in Punjab whereas serotypes 2 and 3 introduced for the first time into Swat, KP in 2013; about 600 km away from Lahore, Punjab. Overall dengue virus serotypes 2 and 3 were the major outbreak-causing serotypes in Pakistan in 2011 and 2013. Dengue outbreak in Swat may be the continuation of previous dengue outbreaks in Punjab but it needs further research and investigation. Keywords Circulating serotypes of dengue virusDengue infectionEpidemiology of dengueSwatDengue PakistanKhyber Pakhtunkhwaissue-copyright-statement© The Author(s) 2016 ==== Body Background Dengue fever is an important mosquito-borne viral disease, caused by one of the four closely related but antigenically distinct dengue virus serotypes (DENV-1 to 4). According to World Health Organization (WHO), the annual incidence of dengue infection is 100 million, of which approximately 500,000 patients develop dengue hemorrhagic fever (DHF) that may lead to dengue shock syndrome (DSS) with a mortality rate of more than 2.5 % [1–3]. DENV has a positive sense RNA genome (genus Flavivirus; family Flaviviridae) of approximately 11 Kb in size. Infection with one serotype does not provide immunity against other serotypes, so a person living in an endemic area could have up to four DENV infections during his/her life span [4, 5]. All four serotypes of dengue are able to produce DHF or DSS in humans after infection is established [6]. In the tropical and subtropical climates of the world, the incidence of dengue is continuously on the rise due to suitable climatic conditions for dengue vector growth [7]. Furthermore, antibodies raised against the membrane protein prM are known not to neutralize the virus but they can cause the antibody-dependent enhancement (ADE) [8–10]. Due to ADE, and in order to protect individuals from DENV infections, it becomes critical to survey the dengue serotypes in any particular area. In August 1994, the first confirmed outbreak of DENV was reported in Karachi, Pakistan. Since then, moderate to severe dengue outbreaks were observed in many areas of Pakistan. Dengue infection did not contain to southern Pakistan (Karachi) but transmitted to northern parts (Punjab) of the country with the passage of time that caused great losses in terms of mortalities and morbidities [11–15]. In 2011, a devastating dengue outbreak reported in Lahore, the capital city of the Punjab, Pakistan. According to Health Departments in Pakistan, more than 23,000 people were hospitalized, of which at least 365 people died due to DHF and DSS [13, 14]. DENV-2 and DENV-3 and a single case of DENV-4 were previously reported the 2011 dengue outbreak in Punjab [16]. Unprecedentedly, just 1 year after the major outbreak in Lahore, a sever dengue outbreak occurred in 2013 in Swat District, Khyber Pukhtunkhwa (KP), which is about 600 km away from Lahore and has relatively temperate climatic conditions (average annual rainfall exceeds 1000 mm and mean annual temperature of about 18 °C) in contrast to Lahore, Punjab (average annual rainfall recorded is 430–767 mm and mean annual maximum temperature of about 28–31 °C) [17, 18]. The outbreak in Swat claimed 57 lives and thousands of people were hospitalized. Therefore, this study was conducted to determine the exact distribution of DENV serotypes in the 2011 and 2013 outbreaks and to shed light on the overall situation. Methods Collection of samples from patients during the years 2011 and 2013 During the 2011 dengue outbreak in Punjab, blood samples (n = 600) were collected from Enzyme-linked immunosorbent assay (ELISA) positive hospitalized patients (461 patients from five districts in Punjab province and 139 patients from six districts of KP-based people, lived in Punjab) (Table 1). ELISA test for the initial diagnosis of DENV had conducted by the respective health centers and then we collected blood samples from those patients (ELIZA positive) for further analysis such as RT-PCR. Similarly, during the 2013 outbreak of DENV in Swat District, blood samples (n = 740) were randomly collected from the hospitalized patients at Saidu Medical College Hospital in Swat, experiencing dengue like symptoms (Table 1). Blood samples were shifted to the Institute of Biotechnology and Genetic Engineering (IBGE), The University of Agriculture, Peshawar, Pakistan. Blood samples collected in Swat, KP were analyzed initially by SD Dengue Duo strips (Standard Diagnostics, Korea) tests for IgM, IgG and DENV NS1 glycoproteins. Statistical analysis was done using SPSS version 20.Table 1 Collection of samples during 2011 and 2013 dengue outbreaks and distribution of DEVN serotypes in affected people Province/Year District Samples PCR + ve DENV1 DENV2 DENV3 DENV4 Mixed M/F Age group [Years] <15 15–45 >45 Punjab/2011 Lahore 322 282 11 121 115 25 10 154/128 57 164 61 Sargodha 58 28 3 9 8 5 3 17/11 5 16 7 Sheikhupora 41 16 0 7 8 1 0 9/7 3 9 4 Kasur 29 13 0 4 8 1 0 9/4 2 7 4 RYK 11 2 0 1 1 0 0 2/0 0 2 0 Total 461 341 14 (4.10 %) 142 (41.64 %) 140 (41.05) 32 (9.38 %) 13 (3.81 %) 191/150 67 (19.64 %) 198 (58.06 %) 76 (22.28 %) KP/2011 Peshawar 63 26 3 10 8 2 3 21/5 4 15 7 Mardan 25 11 1 5 3 1 1 8/3 1 8 2 Nowshehra 19 8 1 4 2 0 1 6/2 1 4 3 Kohat 6 2 0 0 2 0 0 2/0 0 2 0 Charsadda 10 5 1 2 2 0 0 4/1 1 3 1 Swat 16 8 0 4 4 0 0 7/1 1 6 1 Total 139 60 6 (10.0 %) 25 (41.66 %) 21 (35.0 %) 3 (5.0 %) 5 (8.33 %) 48/12 8 (13.33 %) 38 (63.33 %) 14 (23.33 %) KP/2013 Swat 740 200 0 80 (40.0 %) 120 (60.0 %) 0 0 159/41 42 (21.0 %) 89 (44.50 %) 69 (34.50 %) KP Khyber Pakhtunkhwa, RYK Rahim Yar Khan, M/F Male/Female, DENV Dengue Virus Information about the patients and/or samples collected from various geographical areas of Pakistan in both these two dengue outbreaks is provided in Fig 1.Fig. 1 Provinces and Districts hit by dengue virus in 2011 and 2013 in Pakistan. a Map of Khyber Pakhtunkhwa, DENV hit areas are encircled. b Map of Punjab, DENV hit districts/areas are encircled RNA extraction and serotype-specific PCR RNA from all serologically positive blood samples was extracted using QIAGEN QIAamp viral RNA mini kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions. The extracted RNA was used for the identification of the DENV serotypes (1–4) using type-specific reverse transcription-polymerase chain reaction (RT-PCR) as described previously [1, 19]. Results Results of this study indicated that all four serotypes of DENV (1–4) are circulating in Punjab province and DENV-2 and DENV-3 in Swat, KP. DENV specific bands (511 bp) of the first round RT-PCR assay along with positive and negative controls are given in Fig 2.Fig. 2 Gel photograph of the first round RT-PCR product. M: 100 bp DNA ladder (GeneRuler); 1: Positive control (provided by Molecular genetics lab in IBGE, The University of Agriculture Peshawar, Pakistan); 2: Negative control; 3–4: 511 bp specific bands of the first round RT-PCR reaction, representing DENV presence in serum samples of patients The 2011 outbreak in Punjab province In Punjab province, 74 % samples from patients belonged to Punjab (341 out of 461) were positive for DENV-RNA as indicated by RT-PCR analysis and confirmed to have active dengue infection that comprised 56 % (191 out of 341) males and 44 % (150 out of 341) female patients (Table 1). Majority of the actively infected people were from Lahore (83 %) followed by other districts like Sargodha (9 %), Sheikhupura (5 %), Kasur (4 %) and Rahim Yar Khan (0.5 %) (Table 1). Importantly, mixed infections with DENV-2 and DENV-3 in same serum samples were detected in this study; like sample no. 2x that revealed both DENV-2 and DENV-3 (Figs. 3 and 4). Mixed infection comprised of DENV-2 and DENV-3 in patients belonged to Punjab was 3.81 % (13 out of 341). The most dominant DENV serotypes in Punjab province were DENV-2 (41.64 %) and DENV-3 (41.05 %). The incidence of DENV-4 and DENV-1 recorded was 9 and 4 % respectively.Fig. 3 Gel photograph of sample no. 2x, representing DENV-2. M: 100 bp DNA ladder (GeneRuler); 1–2: DENV-2 specific bands of 119 bp size (Sample no. 2×) Fig. 4 Gel photograph of samples tested for the presence of DENV-3. M: 100 bp DNA ladder (GeneRuler); 1: Sample no. 2z; 2: Sample no. 2y; 3: Sample no. 2w; 4: Sample no. 2v; 5: Sample no. 2u; 6: Sample no. 2 t; 7: Sample no. 2x, showing DENV-3 specific band of 290 bp Sample no. 2x along with other samples was tested with specific primers for the detection of DENV-3. Out of the seven samples tested, DENV-3 was detected in sample no. 2x that confirmed the mixed infection (Fig. 4). Similarly active dengue infection rate as revealed by RT-PCR assay in KP-based patients, lived in Punjab was 43 % (60 out of 139). Male population was mostly infected (80 %) as compared to female population (20 %). Majority of KP-based people who infected in Punjab were from Peshawar (43 %), Mardan (18 %), Nowshehra (13 %), Swat (13 %), Charsadda (8 %) and Kohat (3 %) (Table 1). Mixed infection (8.33 %) of DENV-2 and DENV-3 was also recorded in some patients (Table 1). Individuals between the ages of 15 to 45 years were mostly infected in the 2011 dengue outbreak. The 2013 outbreak in Swat, KP province Analysis of the 740 blood samples collected from patients in Swat District during the 2013 dengue outbreak showed that 618 patients were positive for dengue antibodies. However, active dengue infection was detected in only 200 patients out of 618 as revealed by RT-PCR assay. Only DENV-2 (40.0 %) and DENV-3 (60.0 %) were detected in patients and no evidence of DENV-1 and DENV-2 was recorded in Swat (Table 1). Figure 5 indicates the overall summary of the situation of DENV serotypes, circulating in Pakistan. Frequency distribution of serotypes clearly indicates that DENV-2 and DENV-3 are the dominant serotypes in Pakistan with some mixed infection (Fig. 5).Fig. 5 District wise frequency distribution of DENV serotypes in Punjab and KP in 2011 and in Swat, KP in 2013. DENV-2 and DENV-3 are the major outbreak causing serotypes in both the outbreaks as indicated. Blue, green, gray, violet and yellow colors indicate DENV-1, DENV-2, DENV-3, DENV-4 and mixed infection, respectively Discussion Dengue has been one of the major causes of hospitalization among people since the 1990s, particularly in South East Asia [20]. Increased population movement and spread of competent mosquito vectors are the factors contributing to the spread and establishment of the disease in temperate areas of the world [21]. Countries bordering Pakistan like India is facing dengue since 1996 and all four serotypes have been documented in Delhi [22]. Similarly, DENV-2, DENV-3 and DENV-4 have been detected in a study conducted in Bangladesh [23]. Prevalence of dengue infection has also reported from Iran [24]. In Pakistan, Dengue virus infection initially started in Karachi in 1994 and gradually expanded to other areas of the country with the passage of time [13, 14]. The geographical location of Pakistan with hot and humid summers particularly in Punjab, is an ideal place for the breeding of the Aedes mosquitoes [13] which are responsible for spreading the virus from one location to another. Environmental fragmentation which resulted from severe flood in 2010 further encouraged the breeding of the Aedes mosquitoes in Pakistan [25, 26]. Consequently, DENV infection reached epidemic level in 2011 in Pakistan. In this study, patients (n = 1340) from 12 different districts of Pakistan were analyzed for prevalence, distribution, and frequency of dengue virus. The results indicated that all four DENV serotypes circulate in Punjab Province and only DENV-2 and DENV-3 for the first time introduced into Swat District, KP (Table 1). This study showed that during the 2011 dengue outbreak in Punjab, all four serotypes of DENV were prevalent in both genders and capable of active infection in population but DENV-2 and DENV-3 were the most dominant serotypes. Consequently, this type of studies presented here become crucial to determine the possibility of antibody-dependent enhancement (ADE), which usually leads to DSS [2, 3]. Mixed infection in 2011 dengue outbreak highly suggests secondary infection and hence more infectivity and deaths (Table 1). In 2013 dengue outbreak in Swat, DENV-2 (40 %) and DENV-3 (60 %) were the dominant serotypes isolated from the infected people while no evidence of DENV-1 or DENV-4 was found (Table 1). Another aspect which may account for the presence of higher incidence of DENV-3 in Swat than DENV-2 (unlike in Punjab) may be the role of Ae. Albopictus (Fig. 5). This species is generally related to temperate type of environment in the ruler areas [27] like Swat, but it needs further investigation and we would address it in our coming project. In Swat, extensive deforestation and a new culture of cultivating orchards; which is currently the major means of income of the majority of people, has brought substantial changes in the agriculture landscaping. Severe flooding in 2010 and the recent DENV outbreak is an indication of the climate change in this scenic valley because dengue incidence is associated with change in climate [28]. Lake of fresh water supply system is another problem afflicting the population now-a-days and that’s why water pools have been dug inside orchards where people keep water for several days which provides enough time and favorable conditions for the breeding of Aedes mosquitoes. Results of this study indicate that the active and juvenile group (15–45 years) was the most affected. Number of the elderly infected people in Swat was considerably higher than those infected during the 2011 outbreak in Punjab. One possible reason may be the healthy environment of Swat District; allows people to remain active and contribute to the working class even in their late sixties and early seventies. More exposure and more activity during the day time provide more chances for the vector bites and hence higher infection rate. The gender wise distribution of DENV infection indicates no significant difference between the male and female populations (191 & 150 respectively) infected in Punjab during the 2011 outbreak. However, the gender wise difference (48 males & 12 females) in Swat District was quite evident during 2013 dengue outbreak (Table 1). The possible reason for this may be different social set ups at two distant locations with different cultural characteristics. Women are mostly restricted to remain inside homes in Swat District while in contrast there is a huge class of working women in Punjab which increases the chances of contracting the infection. A considerable number of patients were found to have mixed infection (Table 1) which indicates that the individuals with mixed infection had acquired infection in highly endemic pockets of the cities due to increased breading activity of the mosquitoes during the same season. The incidence of concurrent or mixed infection in Punjab in 2011 can be attributed to the long history of DENV in the region. In 2006, the first confirmed DENV outbreak erupted in Punjab and since then it is causing sporadic infection and/or major outbreaks. In 2006, 3000 confirmed cases were registered in Pakistan, out of which 52 people died [29]. In 2008, dengue infection continued in Pakistan with 1450 people infected in Punjab, of which 20 died. In that outbreak, ten patients out of 17 in Lahore were positive for DENV-4, five for DENV-2 and two for DENV-3 when analyzed through real time PCR [30]. In 2010, dengue outbreak occurred in Pakistan and more than 4000 cases were reported from Punjab that caused three deaths [13]. DENV outbreak in Pakistan in 2011 was the most severe in the history; affected more than 22,562 individuals in total and claimed 363 lives across Pakistan according to the reports of health departments. According to independent surveys, the number of infected people was more than 35,000 and number of deaths it caused was more than 420. Punjab once again was the most hit Province by DENV that infected more than 21,300 individuals, out of which 337 people died. Though the entire Punjab was affected but Lahore was the most hit city with 17,493 people infected, of which 290 people died. As compared to previous dengue outbreaks in Punjab, mortality and morbidity rate was highest in 2011. The reasons for such high rate of infectivity and deaths as compared to previous infections in Pakistan, may be the long history of dengue in the region, prevalence of all four serotypes (Table 1), mixed infection (Figs. 3 and 4) that usually leads to DHF and DSS [8–10] and climatic suitability for vector species. This is the first study in Pakistan that documents concurrent infection in 2011 dengue outbreak in Punjab (Table 1). Concurrent infection was not observed in 2013 dengue outbreak in Swat. The possible reasons behind this may be that this was the first incidence of dengue with no previous record of dengue in Swat. Secondly, the infection was in progress in Swat when we conducted this study and ultimately it is clear from the above history that for concurrent infection to appears, co-circulation of multiple DENV serotypes for fairly long time in an endemic area may be required. The incidence of concurrent infections has already been reported in some parts of the world like in Puerto Rico in 1982 [31], in India with overall 66.7 % concurrent infection of DENV-2 and DENV-3 [32] and in Bangladesh, where two patients were co-infected with DENV-2 and DENV-3 while one patients was co-infected with DENV-3 and DENV-4 [23]. We suggest that effective mosquito eradication strategies should be adopted as future outbreaks may cause high mortality due to more chances of secondary infections. Moreover, dengue occurrence in areas like Swat is lethal because Swat is famous for tourism and if dengue persists it would greatly destroy not only tourism in Swat but in the adjacent areas too. Further, thorough genetic characterization of the DENV serotypes should be carried out in order to facilitate futuristic vaccine and/or drug development strategies. Conclusions Dengue virus is continuously affecting the people of Pakistan. Dengue virus is spreading to remote, rural and previously non-endemic areas of Pakistan with DENV-2 and DENV-3 as the major spreading viruses. Rural areas of Pakistan are less developed and lack basic health infrastructure, education and jobs opportunities. Therefore concerted efforts should be employed to check the spread of dengue into theses less developed areas in order to prevent further damage and to protect lives of the common and poor people in particular. Furthermore, research should be initiated to uncover the factors behind this expansion of dengue virus into previously non-endemic areas so that preventive measures may be taken. Abbreviations DENVDengue virus DHFDengue hemorrhagic fever DSSDengue shock syndrome ELISAEnzyme-linked immunosorbent assay IBGEInstitute of Biotechnology and Genetic Engineering KPKhyber Pakhtunkhwa RT-PCRReverse transcriptase polymerase chain reaction The authors highly acknowledge the support of Higher Education Commission of Pakistan (HEC), Pakistan Medical and Research Council (PMRC), Relief International (RI) and the Directorate of Science and Technology (DoST) Khyber Pukhtunkhwa, Pakistan. Funding This project was supported by the Higher Education Commission of Pakistan (HEC) by providing funds to conduct research on dengue in Pakistan. Availability of data and materials All relevant information is provided in this current manuscript. Authors’ contributions AA, HA and IA designed the study. AA carried out the research work. MI helped in drafting the manuscript. FZ helped in data collection. All authors read and approved the final manuscript. Competing interest The authors declare that they have no competing interests. Ethics approval and consent to participate Consent of the individuals to participate in the study was taken in each case and approval of the Institutional ethics committee was accrued before undertaking this study. All the procedures were carried out in accordance with the declaration of Helsinki [33]. ==== Refs References 1. Ali A Nasim Z Rehman R Farzana AS Ali S Zahir F Iqbal A Ali I Khan AW Dengue virus serotype 2 and 3 causing high morbidity and mortality in Swat, Pakistan Biohelikon: Immun Dis 2013 1 1 3 2. World Health Organization: Dengue hemorrhagic fever: diagnosis, treatment and control. 2nd edition. Geneva: World Health Organization; 1997. http://www.who.int/csr/resources/publications/dengue/Denguepublication/en/. Accessed 2 Dec 2015. 3. Simmons CP, Farrar JJ, Nguyen V, Wills B. Dengue. N Engl J Med. 2012. doi:10.1056/NEJMra1110265. 4. Gubbler DJ Dengue and dengue hemorrhagic fever Clin Microbiol 1998 11 480 496 5. Guzman MG Halstead SB Artsob H Buchy P Farrar J Gubler DJ Dengue: a continuing global threat Nat Rev Microbiol 2010 8 S7 S16 10.1038/nrmicro2460 21079655 6. Kyle JL, Harris E. Global spread and persistence of dengue. Annu Rev Microbiol. 2008. doi:10.1146/annurev.micro.62.081307.163005. 7. World Health Organization. Dengue: guidelines for diagnosis, treatment, prevention and control. WHO Press; 2009. http://www.who.int/rpc/guidelines/9789241547871/en/. Accessed 2 Jan 2016. 8. Dejnirattisai W, Jumnainsong A, Onsirisakul N, Fitton P, Vasanawathana S, Limpitikul W, et al. Cross-Reacting Antibodies Enhance Dengue Virus Infection in Humans. Science. 2010. doi:10.1126/science.1185181. 9. Kliks SC Nisalak A Brandt WE Wahl L Burke DS Antibody-dependent enhancement of dengue virus growth in human monocytes as a risk factor for dengue hemorrhagic fever Am J Trop Med Hyg 1989 40 444 451 2712199 10. Morens DM Halstead SB Measurement of antibody-dependent infection enhancement of four dengue virus serotypes by monoclonal and polyclonal antibodies J Gen Virol 1990 71 2909 2914 10.1099/0022-1317-71-12-2909 2273390 11. Chan YC Salahuddin NI Khan J Tan HC Seah CL Li J Dengue hemorrhagic fever outbreak in Karachi, Pakistan Trans R Soc Trop Med Hyg 1995 89 619 620 10.1016/0035-9203(95)90412-3 8594672 12. Koo C, Nasir A, Hapuarachchi HC, Lee KS, Hasan Z, Ng LC, Khan E. Evolution and heterogeneity of multiple serotypes of Dengue virus in Pakistan, 2006–2011. Virol J. 2013. doi:10.1186/1743-422X-10-275. 13. Rasheed SB, Butlin RK, Boots M. A review of dengue as an emerging disease in Pakistan. Public Health. 2013. doi:10.1016/j.puhe.2012.09.006. 14. Shakoora MT Ayubb S Ayub Z Dengue fever: Pakistan’s worst nightmare WHO South-East Asia J Public Health 2012 9 3 229 231 15. Jamil B Hasan R Zafar A Bewley K Chamberlain J Mioulet V Dengue virus serotype 3, Karachi, Pakistan Emerg Infect Dis 2007 13 182 183 10.3201/eid1301.060376 17370547 16. Fatima ZA, Afzal SA, Idrees M, Rafique S, Akram M, Khubaib B, et al. Change in demographic pattern of dengue virus infection: evidence from 2011 dengue outbreak in Punjab, Pakistan. Public Health. 2013. doi:10.1016/j.puhe.2013.03.003. 17. National Rural Support Program. Evaluation & Research Report NRSP-MER/2011-IV. http://www.nrsp.org.pk/publications.html. Accessed 30 Jul 2016 18. Khattak MS Sajjad A Assessment of temperature and rainfall trends in Punjab province of Pakistan for the period 1961–2014 J Himal Earth Sci 2015 48 42 61 19. Lanciotti RS Calisher CH Gubler DJ Chang G-J Vorndam AV Rapid detection and typing of Dengue viruses from clinical samples by using reverse trasnciptasepolymerase chain reaction J Clin Microbiol 1992 30 545 551 1372617 20. World Health Organization. Dengue hemorrhagic fever: diagnosis, treatment and control. Geneva: World Health Organization; 1986. 21. Fontenille D Failloux AB Romi R Takken W Knols BGJ Should we expect Chikungunya and Dengue in Southern Europe? Emerging Pests and Vector-Borne Diseases in Europe 2007 Wageningen Academic 169 184 22. Gupta E, Dar L, Kapoor G, Broor S. The changing epidemiology of dengue in Delhi, India. Virol J. 2006. doi:10.1186/1743-422X-3-92. 23. Aziz MM Hasan KN Hasanat MA Siddiqui MA Salimullah M Chowdhury AK Predominance of the DEN-3 genotype during the recent dengue outbreak in Bangladesh Southeast Asian J Trop Med Public Health 2002 33 42 48 12118459 24. Aghaie A, John A, Sadegh C, Matthias N, Soudabeh B, Hashem KM. Frequency of dengue virus infection in blood donors in Sistan and Baluchestan province in Iran. Transfus Apher Sci. 2014. doi:10.1016/j.transci.2013.07.034. 25. Raheel U Faheem M Riaz MN Kanwal N Javed F Qadri I Dengue fever in the Indian subcontinent: an overview J Infect Dev Ctries 2011 5 239 247 21537064 26. Warraich H, Zaidi AK, Patel K. Floods in Pakistan: a public health crises. Bulletin of the World Health Organization; 2011. doi:10.2471/BLT.10.083386. 27. Joshi V Mouriya DT Sharma RC Persistence of Dengue-3 virus through transovarial transmission passage in successive generations of aedes aegypti mosquitoes Am J Trop Med Hyg 2002 67 2 158 161 12389940 28. Sharmin S, Viennet E, Glass K, Harley D. The emergence of dengue in Bangladesh: epidemiology, challenges and future disease risk. Trans R Soc Trop Med Hyg. 2015. doi:10.1093/trstmh/trv067. 29. Tang JW, Khanani MR, Zubairi AM, Lam WY, Lai F, Hashmi K, et al. A wide spectrum of dengue IgM and PCR positivity post-onset of illness found in a large dengue 3 outbreak in Pakistan. J Med Virol. 2008. doi:10.1002/jmv.21290. 30. Humayoun, MA, Waseem T, Jawa AA, Hashimi MS, Akram J. Multiple dengue serotypes and high frequency of dengue hemorrhagic fever at two tertiary care hospitals in Lahore during the 2008 dengue virus outbreak in Punjab, Pakistan. Int J Infect Dis. 2010. doi:10.1016/j.ijid.2009.10.008. 31. Araújo FM Nogueira RM de Araújo JM Ramalho IL Roriz ML de Melo ME Coelho IC Concurrent infection with dengue virus type-2 and DENV-3 in a patient from Ceara, Brazil Mem Inst Oswaldo Cruz 2006 101 925 928 10.1590/S0074-02762006000800017 17293990 32. Vinodkumar CS, Kalapannavar NK, Basavarajappa KG, Sanjay D, Gowli C, Nadig NG, Prasad BS. Episode of coexisting infections with multiple dengue virus serotypes in central Karnataka, India. J Infect Public Health. 2013. doi:10.1016/j.jiph.2013.01.004. 33. Christie B Doctors revise Declaration of Helsinki BMJ 2000 321 913 10.1136/bmj.321.7266.913 11030663

PMC005xxxxxx/PMC5002135.txt

==== Front BMC Med ImagingBMC Med ImagingBMC Medical Imaging1471-2342BioMed Central London 15610.1186/s12880-016-0156-6Technical AdvanceSparse reconstruction of compressive sensing MRI using cross-domain stochastically fully connected conditional random fields Li Edward 1Khalvati Farzad 2Shafiee Mohammad Javad 1Haider Masoom A. 2Wong Alexander a28wong@uwaterloo.ca 11 Department of Systems Design Engineering, University of Waterloo, Ontario, Waterloo Canada 2 Department of Medical Imaging, University of Toronto and Sunnybrook Research Institute, Toronto, Ontario Canada 26 8 2016 26 8 2016 2016 16 1 515 4 2016 15 8 2016 © The Author(s) 2016 Open Access This 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 Magnetic Resonance Imaging (MRI) is a crucial medical imaging technology for the screening and diagnosis of frequently occurring cancers. However, image quality may suffer from long acquisition times for MRIs due to patient motion, which also leads to patient discomfort. Reducing MRI acquisition times can reduce patient discomfort leading to reduced motion artifacts from the acquisition process. Compressive sensing strategies applied to MRI have been demonstrated to be effective in decreasing acquisition times significantly by sparsely sampling the k-space during the acquisition process. However, such a strategy requires advanced reconstruction algorithms to produce high quality and reliable images from compressive sensing MRI. Methods This paper proposes a new reconstruction approach based on cross-domain stochastically fully connected conditional random fields (CD-SFCRF) for compressive sensing MRI. The CD-SFCRF introduces constraints in both k-space and spatial domains within a stochastically fully connected graphical model to produce improved MRI reconstruction. Results Experimental results using T2-weighted (T2w) imaging and diffusion-weighted imaging (DWI) of the prostate show strong performance in preserving fine details and tissue structures in the reconstructed images when compared to other tested methods even at low sampling rates. Conclusions The ability to better utilize a limited amount of information to reconstruct T2w and DWI images in a short amount of time while preserving the important details in the images demonstrates the potential of the proposed CD-SFCRF framework as a viable reconstruction algorithm for compressive sensing MRI. Keywords Compressive sensingConditional random fieldsMagnetic resonance imaginghttp://dx.doi.org/10.13039/501100001804Canada Research Chairs229272issue-copyright-statement© The Author(s) 2016 ==== Body Introduction Magnetic Resonance Imaging (MRI) is a medical imaging technology that is currently used for diagnostic imaging of a wide range of diseases. In particular, since MRI does not use ionizing radiation, it has become a crucial imaging modality for screening frequently occurring cancers such as prostate cancer in men, breast cancer in women, as well as lung and colorectal cancer for both men and women. In 2015, 196,900 new cases of cancer (excluding non-melanoma skin cancers) were expected, with 51 % of these belonging to the four aforementioned types of cancer in Canada [1]. As such, cancer screening methods with accurate and reliable information such as MRI is highly desired. Of particular interest for cancer screening is multi-parametric MRI (MP-MRI) since more information can be acquired through different modalities. MP-MRI contains different techniques such as diffusion weighted imaging (DWI), correlated diffusion imaging (CDI) [2–4], dynamic contrast enhancement (DCE), T2-weighted (T2w) imaging, and T1-weighted (T1w) imaging [5]. Although this approach provides a more complete information, acquisition times are significantly longer which causes more patient discomfort and motion artifacts that decrease image quality. As a result, new methods to improve MRI acquisition times are highly desired to facilitate for reliable MP-MRI data acquisition. Compressive sensing has been demonstrated to be an effective strategy for reducing MRI acquisition times by acquiring significantly fewer samples in k-space. A complete signal can then be fully reconstructed through sparse, yet sufficient number of samples [6–8]. In MRI, compressive sampling strategies have been demonstrated to be highly effective at reducing acquisition time while maintaining image quality as different types of tissue structure have been shown to be sparse in certain domains [9]. Furthermore, different techniques have been proposed to improve the imaging process [10] as well as the reconstruction process [11–23] in compressive sensing. Due to the limited amount of data available through compressive sensing, advanced reconstruction algorithms are required to produce high quality and reliable images. Different methods have been proposed for sparse reconstruction of compressive sensing MRI [11–23]. As a notable example, Block [14] proposed an iterative image reconstruction technique using a modified total variation (TV) constraint [20, 21] for sparse reconstruction of compressive sensing brain MRI. Trzasko [15] introduced a homotopic l 0 minimization method for the sparse reconstruction of compressive sensing spinal MRI. Wong [12] extended upon this idea and proposed a regional sparsified domain for the sparse reconstruction of breast MRI. A similar technique was also demonstrated by Qu using combined sparsifying transforms and smoothed l0 norm minimization [13], where they showed that the use of combined transforms can improve image quality comprised of the reconstructed images from compressive sensing MRI when compared to methods using a single sparsifying transform. However, the downside of the l0 norm minimization is the fact that its performance significantly depends on the tuning parameters where these tuning parameters can greatly affect the convergence rate of the algorithm. Other ln optimization techniques such as the standard l2 (least squares) minimization can have high error rates as reported in [24]. An area that is little explored but can reap significant potential benefits is the application of random field modeling for improved sparse reconstruction of compressive sensing MRI. Random field modeling such as Markov random fields (MRF) [25, 26] and conditional random fields (CRF) [27] have long been shown to be powerful tools for incorporating spatial context within a probabilistic graphical modeling framework, which can have significant benefits for reconstructing images from sparse measurements. Despite powerful modeling capabilities and potential benefit to sparse reconstruction, one of the biggest hurdles in leveraging random field models for compressive sensing MRI is the fact that all MRI measurements are made in k-space, whereas the images are reconstructed in spatial domain. As the majority of random field models are typically modeled in a single domain, such models cannot be used directly for the purpose of sparse reconstruction of compressive sensing MRI. This is further complicated by the fact that the MRI measurements in k-space are sparse and incomplete, which make it difficult to leverage existing random field models for this problem. Therefore, a probabilistic graphical modeling framework that can consolidate the fact that partial measurements are made in a domain different than the desired states of the reconstruction images is needed to truly leverage the power of random field modeling for sparse reconstruction of compressed sensing MRI. This paper proposes a cross-domain Stochastically fully connected conditional random field (CD-SFCRF) approach for the reconstruction of compressive sensing MRI at below Nyquist sampling rates [28]. Inspired by the stochastic cliques method presented in [29], the proposed cross domain model is a significant extension upon this method by consolidating two different domains of spatial domain and k-space specifically for MRI compressive sensing. CD-SFCRF framework introduces constraints in both k-space and spatial domains within a stochastically fully connected graphical model to produce improved MRI reconstruction. The proposed CD-SFCRF framework has the ability to utilize spatial and data driven consistencies in the spatial domain along with data driven consistencies in the k-space domain pertaining to sparse measurements while maintaining edge features and structural details in the reconstructed images. Phantom MRI data as well as prostate MRI data captured using T2w and DWI imaging modalities, which also yields apparent diffusion coefficient (ADC) map images, were used to illustrate the efficacy of the proposed CD-SFCRF framework for sparse reconstruction of compressive sensing MRI. To the best of the authors’ knowledge, this is the first time that constraints in both k-space and spatial domains are used in conjunction within a stochastically fully connected graphical model for the sparse reconstruction of compressive sensing MRI, which is the main contribution of this paper. The paper is formatted as follows. The methodology behind the proposed CD-SFCRF framework is described in Section “Methodology”. The experimental setup is described in Section “Experimental setup”. Results and discussions are presented and discussed in Sections “Results” and “Discussion”, respectively. Finally, the conclusion is presented in Section “Conclusions”. Methodology In MRI, measurements are made in the k-space [30], with the lower frequency coefficients in the k-space containing coarse-grained contrast information while higher frequency coefficients contain fine-grained image detail information. The MRI measurements from the k-space are transformed into the spatial domain to form the reconstructed MRI image. Most compressive sensing strategies [6, 15] sparsely sample the k-space to reduce image acquisition time significantly. Therefore, to fully utilize available information in the reconstruction process, data-driven constraints in the k-space domain and data and spatial driven constraints in the spatial domain would be highly beneficial in improving image reconstruction quality from compressive sensing MRI. Motivated by this, the proposed cross-domain stochastically fully connected conditional random field (CD-SFCRF) introduced here for the purpose of sparse reconstruction of compressive sensing MRI, extends upon the seminal work on stochastically fully connected conditional random fields (SFCRF) first proposed in [29] to facilitate for this cross-domain optimization. SFCRFs are fully-connected conditional random fields with stochastically defined cliques. Unlike traditional conditional random fields (CRF) where nodal interactions are deterministic and restricted to local neighborhoods, each node in the graph representing a SFCRF is connected to every other node in the graph, with the cliques for each node is stochastically determined based on a distribution probability. Therefore, the number of pairwise cliques might not be the same as the number of neighborhood pairs as in the traditional CRF models. By leveraging long-range nodal interactions in a stochastic manner, SFCRFs facilitate for improved detail preservation while maintaining similar computational complexity as CRFs, which makes SFCRFs particularly enticing for the purpose of improved sparse reconstruction of compressive sensing MRI. However, here the problem is to reconstruct an MRI image in the spatial domain while the available measurements are made in k-space domain. Similar to most CRF models, SFCRFs cannot be leveraged directly for this purpose. Motivated by the significant potential benefits of using SFCRFs in improving reconstruction quality of compressive sensing MRI, we extend the SFCRF model into a cross-domain stochastically fully connected conditional random field (CD-SFCRF) model that incorporates cross-domain information and constraints from k-space and spatial domains to reconstruct the desirable MRI image from sparse observations in k-space. The theory pertaining to sparse reconstruction via a cross-domain stochastically fully connected conditional random field model is detailed in Appendix 1. Implementation An implementation of the proposed CD-SFCRF framework for the purpose of sparse reconstruction from compressive sensing MRI is illustrated in Fig. 1. Here, an iterative gradient descent optimization approach is employed, and can be described as follows. First, the original compressive sensing MRI data in k-space is transformed to the spatial domain to provide an initial estimate of the reconstructed image. Second, the gradient of the unary and pairwise energy potentials is computed, where the unary data driven consistencies with respect to the original observations are enforced in the k-space, and spatial and data driven consistencies are enforced in the spatial domain. Third, the estimate of the reconstructed image is updated based on the previous estimate and the computed gradient. The second and third steps of this process are repeated until convergence. Fig. 1 Optimization framework of the proposed CD-SFCRF framework for sparse reconstruction from compressive sensing MRI Experimental setup To study the efficacy of the proposed CD-SFCRF method for the purpose of sparse reconstruction of compressive sensing MRI, experiments were performed including: i) MRI data acquired of a MRI training phantom, and ii) prostate MP-MRI data of 20 patient cases. A detailed description of the phantom data, patient data, and MRI image acquisition procedure to facilitate for the various experiments are described below. Phantom data The MRI training phantom used in the experiments, shown in Fig. 2, was a multi-modality prostate training phantom from Computerized Imaging Reference Systems Inc (CIRCS MODEL 053). The phantom is composed of a clear acrylic container with dimensions 11.5×7.0×9.5 cm with a front probe opening of 3.2 cm diameter and a rear probe opening of 2.6 cm diameter. The prostate is composed of high-scattering Blue Zerdine with dimensions 5×4.5×4.0 cm and is placed in a background gel similar to water with little backscatter attenuation (≤0.07 dB/cm−MHz). Within the prostate, there are 3 randomly placed lesions of sizes between 0.5−1.0 cm placed hypoechoic to the prostate. The urethra and rectal wall are made of low scattering Zerdine with diameter of 0.7 cm with dimensions 6×11×0.5 cm, respectively. This phantom was imaged with an inflatable Medrad eCoil ERC using DWI. The DWI MRI was acquired by a 3T GE Discovery MR750. DWI was collected at b=0 mm2/s at 3-NEX2. For the DWI data, the echo time (TE) was 71.70 ms and repetition time (TR) was 10,000.00 ms. Fig. 2 Example slice of the prostate training phantom from Computerized Imaging Reference Systems Inc (CIRCS MODEL 053) used for evaluation purposes Patient data experiments To test the efficacy of the proposed CD-SFCRF framework within a real clinical scenario, MRI data of 20 patients (17 with cancer and 3 without cancer) were acquired using a Philips Achieva 3.0T machine at Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada. All data was obtained retrospectively under the local institutional research ethics board (Research Ethics Board of Sunnybrook Health Sciences Centre). For each patient, the following MP-MRI modalities were obtained (Table 1): T2w and DWI. The patients’ age ranged from 53 to 83. Table 1 summarizes the information about the 20 patients’ datasets used in this study, which includes displayed field of view (DFOV), resolution, echo time (TE), and repetition time (TR). Table 1 Description of the prostate T2w and DWI images Modality DFOV (c m 2) Resolution (m m 3) TE (ms) TR (ms) T2w 22×22 0.49×0.49×3 110 4,687 DWI 20×20 1.56×1.56×3 61 6,178 Compressed sensing configuration In order to evaluate the efficacy of the proposed CD-SFCRF framework at different sample rates, we first acquired MRI measurements at all k-space coefficients. Based on this fully-sampled set of k-space measurements, sparse sampling was then conducted using radial sampling patterns with different numbers of radial sampling lines to achieve a desired sampling rate. For example, Fig. 3 shows a radial sampling pattern which corresponds to a sampling rate of 32 % of the k-space. Different sampling rates were tested and evaluated in this study. Fig. 3 Radial k-space sampling pattern at 32 % sampling rate Results In order to evaluate the efficacy of the proposed CD-SFCRF framework for sparse reconstruction of compressive MRI sensing, a comparative evaluation analysis was performed alongside a baseline l2 minimization (L2) reconstruction method, and a state-of-the-art homotopic l0 minimization (HL0) [15] reconstruction method. The tested methods were compared quantitatively through peak signal-to-noise ratio (PSNR) analysis, and qualitatively via visual assessment. All tested methods were implemented based on the original literature, with optimal parameters used in this study. All tested methods were run until convergence. Figure 4 shows the PSNR versus sampling rate plots for the tested methods for the phantom MRI data. Fig. 4 PSNR vs. sampling rates plots for the tested methods for the phantom MRI data at different sampling rates Tables 2, 3, and 4 show the PSNR results for the three reconstructed methods for the T2w, DWI, as well as ADC map images for the patient data experiments at different sampling rates. Table 2 PSNR ± standard deviation (stdev) for T2w images for the patient data experiments (24 images/patient for 20 patients) across different methods with P-values comparing L 2 and HL 0 methods with CD-SFCRF, respectively Sampling rate (%) L 2 (dB) P-value HL 0 (dB) P-value CD-SFCRF (dB) 17 25.56 ± 0.92 p ≪ 0.0001 26.22 ± 1.01 p ≪ 0.0001 27.34 ± 1.16 32 28.39 ± 0.98 p ≪ 0.0001 28.80 ± 1.01 p ≪ 0.0001 29.72 ± 1.16 47 30.42 ± 1.05 p ≪ 0.0001 30.80 ± 1.05 p ≪ 0.0001 31.23 ± 1.17 Bold face indicates the method with the highest performance metric Table 3 PSNR ± stdev for DWI images for the patient data experiments (24 images/b-value for 4 b-values/patient for 20 patients) across different methods with P-values comparing L 2 and HL 0 methods with CD-SFCRF, respectively Sampling rate (%) L 2 (dB) P-value HL 0 (dB) P-value CD-SFCRF (dB) 17 26.90 ± 1.86 p ≪ 0.0001 28.46 ± 2.50 p ≪ 0.0001 28.75 ± 2.22 32 31.92 ± 2.32 p ≪ 0.0001 33.39 ± 3.03 p ≪ 0.0001 33.61 ± 2.17 47 36.45 ± 2.67 p ≪ 0.0001 37.85 ± 2.48 p ≪ 0.0001 37.99 ± 2.11 Bold face indicates the method with the highest performance metric Table 4 PSNR ± stdev for ADC images for the patient data experiments (24 images/patient for 20 patient) across different methods with P-values comparing L 2 and HL 0 methods with CD-SFCRF, respectively Sampling rate (%) L 2 (dB) P-value HL 0 (dB) P-value CD-SFCRF (dB) 17 17.20 ± 0.61 p ≪ 0.0001 19.35 ± 0.58 0.88 19.50 ± 0.57 32 18.05 ± 0.48 p ≪ 0.0001 21.66 ± 0.54 0.89 21.72 ± 0.55 47 18.72 ± 0.32 p ≪ 0.0001 22.94 ± 0.39 0.89 23.16 ± 0.37 Bold face indicates the method with the highest performance metric Figures 5 and 6 shows the visual comparison between the reconstructed images produced using the proposed CD-SFCRF framework compared with that produced using the L2 and homotopic l0 minimization reconstruction methods for three cases for T2w images. Fig. 5 Sample T2w results for 3 patient images produced using CD-SFCRF, L 2, and HL 0 at 32 % sampling ratio. Compared to other methods, CD-SFCRF preserves tissue details and contrast especially in the tumourous regions. The arrow shows tumourous region in the fully sampled image (a, e, i) Fig. 6 Sample T2w results (zoomed in) for 3 patient images produced using CD-SFCRF, L 2, and HL 0 at 32 % sampling ratio. Compared to other methods, CD-SFCRF preserves tissue details and contrast especially in the tumourous regions. The arrow shows tumourous region in the fully sampled image (a, e, i) Figures 7 and 8 shows the visual comparison between the reconstructed images produced using the proposed CD-SFCRF framework compared with that produced using the L2 and HL0 methods for three patient cases for DWI (b=100s/mm2) and ADC maps. Fig. 7 Sample DWI results (b=100s/m m 2) for three patient cases produced using CD-SFCRF, L 2, and HL 0 at 32 % sampling ratio. Compared to other methods, CD-SFCRF preserves tissue details and contrast especially in the tumourous regions. The tumourous region in the fully sampled image is marked (a, e, i) Fig. 8 Sample ADC map results for three patient cases produced using CD-SFCRF, L 2, and HL 0 at 32 % sampling ratio. Compared to other methods, CD-SFCRF preserves tissue details and contrast especially in the tumourous regions. The tumourous region in the fully sampled image is marked (a, e, i) Discussion As it can be observed from Fig. 4, the proposed CD-SFCRF framework achieved noticeable PSNR improvements over the other tested methods at all tested sampling rates. The CD-SFCRF produced improvements of up to 4 dB over HL0 and 7 dB over L2 in low sampling conditions. It can also be observed that as sampling rates increase, the performance differences decrease. This is due to the fact that as the sampling rate increases, the amount of available measurements increases, and as such the level of reconstruction quality improvements that can be achieved will naturally decrease given the amount of available information becomes increasingly sufficient for high quality reconstruction. The ability of the CD-SFCRF framework to produce high quality reconstruction at very low sampling rates can be demonstrated visually as well. From additional quantitative analysis of patients MRI data presented in Tables 2, 3, and 4, it can be observed that the proposed CD-SFCRF framework achieved the greatest PSNR improvements for the lowest sampling rate (i.e., 17 %) where for T2w, CD-SFCRF improved PSNR by 1.78 dB and 1.12 dB over the L2 and HL0 methods, respectively. For DWI, CD-SFCRF improved PSNR by 1.85 dB and 0.28 B over the L2 and HL0 methods, respectively. Interestingly for ADC maps, the best improvements in PSNR were achieved for the highest sampling rate (47 %) where CD-SFCRF improved PSNR by 4.44 dB and 0.21 B over the L2 and HL0 methods, respectively. Tables 2, 3 and 4 also show the P-values calculated by comparing the proposed CD-SFCRF method with L2 and HL0 methods, respectively. As it can be seen, P-values show significant difference between CD-SFCRF and the other two methods for T2w and DWI images. For ADC maps, the proposed CD-SFCRF was significantly different than L2 method as well. The only comparison that did not show significantly different results was CD-SFCRF compared to HL0 method for ADC maps. This shows that the PSNR improvement for the proposed CD-SFCRF framework was meaningful for the majority of cases when compared to other tested methods. Comparing the results for phantom MRI data (Fig. 4) and patients MRI data shown in Tables 2, 3 and 4 shows that the proposed CD-SFCRF framework yields higher performance improvement at 10 - 20 % sampling range for phantom MRI data compared to patients MRI data. The reason for this difference on PSNR improvement is the fact that the morphological and textural properties of the phantom is significantly less complex than that of real patients’ prostates, and thus the reconstruction problem is a simpler one for the phantom and as a result, greater PSNR gains were achieved using the proposed method. Qualitative observations from Figs. 5 and 6 show that the L2 method resulted in blurry T2w images as well as noticeable radial artifacts at low sampling rates as expected due to the least squares reconstruction being prone to errors. The HL0 approach performed better than the L2 minimization and was able to noticeably reduce artifacts and provide a higher quality reconstruction. However, in comparison, the CD-SFCRF was able to better restore details and fine tissue structure in the reconstructed image when compared to HL0. This is to be expected as the CD-SFCRF takes advantage of more complete data and spatial driven consistencies in a fully connected nature, thus better modeling the underlying tissue detail and structures. Furthermore, as it can be seen in Figs. 7 and 8, the L2 method resulted in blurry DWI and ADC map images again with noticeable radial artifacts. Although the HL0 approach performed better than the L2 method, it can be observed once again that the proposed CD-SFCRF approach was able to preserve more fine tissue structure and detail in the reconstructed image when compared to the HL0 method. Nevertheless, an inherent trade-off exists between preserving fine textural granularity and reducing artifacts due to compressed sensing which can be well utilized in the proposed CD-SFCRF framework to achieve a balance between the two competing constraints. In Figs. 5, 6, 7 and 8, the tumourous regions marked by a radiologist and confirmed by pathology report (biopsy results) are shown by red arrow or white boundary. It can be seen that the proposed CD-SFCRF method preserves the separability of the cancerous and healthy tissue in all cases, which is an important measure for usability of the proposed method in practice. As it can be seen, the tumourous regions are blurred in the L2 method, which may make it difficult to detect for radiologists. Both quantitative and qualitative analysis demonstrate the potential of the proposed CD-SFCRF framework as a reliable reconstruction approach for compressive sensing in MRI. It demonstrates the ability to produce edge and tissue details at very low sampling rates. The CD-SFCRF framework better utilize available information to produce high quality reconstructed images given very limited available information. Preservation of tissue structure and detail enhancement, and noise and artifact mitigation are very important for MRI as the diagnostic quality is directly related to the image quality. This demonstrates that the CD-SFCRF framework can be a viable clinical technique as the reduction in acquisition can lead to faster acquisitions and lower patient wait times. With a lower acquisition time and hence lower patient wait time, patients can have access to the necessary treatments in a timely manner, significantly improving the patient outcome and survival rates. The compressive sensing method used to reconstruct MR images can influence the performance of the computer-aided diagnosis (CAD) tools. For example, several radiomics-based CAD algorithms have been proposed for automatic prostate cancer detection which use T2w and DWI to extract texture and morphological features fed into a classifier [31–36]. These algorithms heavily rely on the quality of regions of interests in similar cases in DWI and therefore, it is expected that a reconstructed MRI with better quality will improve the performance. As future work, we will investigate the effect of the proposed compressive sensing method on the detection accuracies of these radiomics-based CAD algorithms with respect to the L2 and HL0 methods. Moreover, recently, computational diffusion MRI (CD-MRI) has been introduced which utilizes the wealth of information in DW-MRI to computationally construct new sequences of MRI that potentially will help radiologists with more accurate and consistent diagnosis [2, 3]. The proposed CD-SFCRF framework will be integrated into CD-MRI algorithms [2, 3] to investigate whether CD-SFCRF improves the separability of cancerous and healthy tissues in prostate for these computationally generated MR sequences with respect to the L2 and HL0 methods. The limitations of the proposed CD-SFCRF method that will be addressed in the future direction of this work include the limited sample size. A larger and more diverse dataset will be used to address this limitation. Moreover, in this work, the proposed method was applied only to the prostate. Future work also includes applications of the proposed method to the MRI acquisitions of other organs such as breast or moving organs such as heart. In addition, although the CD-SFCRF can significantly decrease MRI acquisition times, because of the fully-connected nature of this method, the algorithm may require a considerable processing time to complete (although not comparable to original MRI acquisition time). As future work, we will modify the proposed method to improve processing time and the efficiency of the algorithm. Conclusions In this paper, a cross domain stochastic fully connected conditional random field (CD-SFCRF) framework for sparse reconstruction of compressive sensing MRI was presented. The proposed CD-SFCRF framework introduces constraints in both k-space and spatial domains within a stochastically fully connected graphical model to produce improved MRI reconstruction. To test the efficacy of the proposed CD-SFCRF framework, quantitative experimentation using peak signal-to-noise ratio (PSNR) analysis was performed on phantom MRI data. Quantitative and qualitative experimentations were also performed on prostate MP-MRI data of 20 patient cases at different sampling rates. The results show an improvement over other tested sparse reconstruction approaches, especially at low sampling rates. The ability to better utilize available information given very limited information demonstrates the potential of the proposed CD-SFCRF framework as a viable reconstruction algorithm for compressive sensing MRI. The proposed CD-SFCRF can significantly reduce MRI acquisition times without sacrificing quality and potential reduction in the accuracy of diagnosis. Reducing MRI acquisition time would reduce related cost significantly and lead to less patient discomfort during the MRI acquisition and more importantly, it would reduce the patient wait times considerably. A fast access to MRI would directly translate to better care given to patients who need it the most. Appendix 1 Sparse reconstruction via cross-domain stochastically fully connected conditional random field The main goal here is to reconstruct image Y given original sparsely sampled k-space observations X. We model the conditional probability P(Y|X) of the full state set Y in spatial domain given the set of sparse measurements X in k-space, which can be written as: 1 P(Y|X)=1Z(X)exp(−ψ(Y|X)) where Z(X) is the normalization function and ψ(.) is a combination of unary and pairwise potential functions: 2 ψ(Y|X)=∑i=1nψu(yi,X)+∑φ∈Cψp(yφ,X) Here yi∈Y is a single state in the set Y={yi}i=1n, yφ∈Y encodes a clique structure in the set C, and X={xj}j=1n is the observations (radially sub-sampled frequency coefficients) in the frequency domain (k-space). The unary potential ψu is enforced in the k-space while the pairwise potential ψp is applied in the spatial domain. The unary potential enforces original observations to preserve data fidelity. Since the available observations are captured in k-space in MRI, the model must be formulated in a way to be consistent in both k-space and spatial domain. The pairwise potential, on the other hand, has to be in the spatial domain to better preserve image detail since neighboring coefficients in the k-space does not contain any meaningful spatial or data consistencies to be utilized by the pairwise potential. Therefore, the optimal way to fully utilize available data within this random field model is to formulate the unary potential in the k-space and the pairwise potential in the spatial domain. One of main differences between the proposed CD-SFCRF framework from conventional CRF models is to incorporate long-range information in the model and preserve boundaries and image structural properties more effectively which is important here due to sparse available observation. To capture long-range information, CD-SFCRF assumes fully connected neighboring structure for the underlying graph which each node i has a set of neighbors 3 N(i)=j|j=1:n,j≠1 where |N(i)|=n−1 and includes all other nodes in the graph as neighbors of node i. Here the pairwise clique structures are utilized such that: 4 C=Cp(i)i=1n 5 Cp(i)=(i,j)|j∈N(i),1i,jS=1. The active cliques in the inference procedure are determined by the stochastic indicator function 1i,jS=1. The indicator function decides whether or not nodes can construct a clique, Cp(i) for node i. This stochastic indicator function combines spatial and data driven information to model the probability distribution of informative cliques which informative cliques have higher probability to participate in the inference. The set of active cliques are obtained to extract pairwise potentials in Eq. 2. As mentioned before, ψ(·) in Eq. 2 is the combination of two potential functions ψu(.), the unary potential and ψp(.), the pairwise potential. These potential functions are formulated with their corresponding weights λ, respectively as: 6 ψu(Y,X)=∑j=1KλjuFj(Y,X) 7 ψp(yφ,X)=∑{yi,yj}∈yφ,k=1K′λkpfk(yi,yj,X) where λ controls the importance of each feature function in the energy formulation and it is calculated in the training stages. Although it is possible to provide several arbitrary feature functions to model the conditional probability P(Y|X), here two feature functions are provided to formulate the image reconstruction for the purpose of sparse reconstruction from compressive sensing MRI. The conditional distribution of Y given X is trained to promote/suppress different features in both the unary and pairwise potentials. Higher λju values promotes a higher reinforcement of original observations while high λkp values promotes higher consideration of spatial and data driven neighborhood constraints. In Eq. 6, F refers to the frequency domain potential function. The unary potential is calculated in the k-space while the pairwise remains in the spatial domain. This is the novelty of the CD-SFCRF whihc facilitates for better preservation of fine tissue details and contrast in the reconstructed image. The unary potential function Fj(yi,X) can be formulated as: 8 Fj(Y,X)=∑ω=−π2π2F(Y,ω)−xω where F(·,·) is the Fourier operator and returns the k-space coefficient corresponding to frequency ω. Based on this formulation, the unary potential is enforced in the k-space and in the inferencing step, the model tries to estimate image Y to be consistent to the original k-space observation X={xω}ω=−π2π2. The pairwise function fk(yi,yj,X) can be formulated as: 9 fk(yi,yj,X)=exp−(yi−yj)2·(xi−xj)23σ2 where σ is a control variable for the amount of weighting node pairs in the clique φ={i,j}. Contrary to the unary potential, the pairwise potential is enforced in the spatial domain. Graph representation Graph G(V,E) (Fig. 9) is the realization of the CD-SFCRF where V is the set of nodes of the graph representing states Y={yi}i=1n, E is the set of edges in the graph. Observations xi∈X are made in the k-space domain. Our final state estimations Y are in the spatial domain (image). Figure 9 shows the graphical representation of how the spatial and k-space domain are incorporated to model the conditional probability P(Y|X). xi comes from sparse measurements in the k-space. In the inference procedure, the k-space observations are transformed into the spatial domain using the Fourier transform to compute the pairwise potentials. Pairwise potentials are calculated in the spatial domain and transformed into the k-space to combine with the unary potential and perform data fidelity. For different types of MRI data, different sparse sampling patterns can be used. Furthermore, pairwise connectivity can be trained for specific types of details and tissue structure. Fig. 9 Realization of CD-SFCRF graph. X i represents original observations made in the k-space, x i represents spatial domain representation of the k-space measurements and y i represent states. F denotes the Fourier operator used in transforming k-space observations into the spatial domain. Connectivity is determined based on probability distributions. Nodes with higher connectivity have solid black edges while lower probable connections are represented as dashed red lines The proposed CD-SFCRF framework utilizes consistencies from the spatial domain through the pairwise potential in conjunction with k-space information through the unary potential. A combination of the two potentials is enforced simultaneously. The unary potential utilizes original observations in the k-space, while the pairwise potential utilizes the spatial domain representation of the observation/state information and calculates pairwise potentials for nodes in the spatial domain. This allows CD-SFCRF to take advantage of the lower computational complexity introduced by the stochastically fully-connected random field model, while leveraging the original k-space observations in improving signal fidelity. Abbreviations MRIMagnetic resonance imaging CRFConditional random field CD-SFCRFCross-domain stochastically fully connected conditional random fields T2wT2-weighted DWIDiffusion-weighted imaging MP-MRIMulti-parametric MRI CDICorrelated diffusion imaging TVTotal variation ADCApparent diffusion coefficient PSNRPeak signal-to-noise ratio stdevStandard deviation CADComputer-aided diagnosis CD-MRIComputational diffusion MRI Funding This research has been supported by the Canada Research Chairs programs, Natural Sciences and Engineering Research Council of Canada (NSERC), the Ministry of Research and Innovation of Ontario, and Ontario Institute of Cancer Research (OICR). Availability of data and materials The datasets during and/or analysed during the current study available from the corresponding author on reasonable request pending the approval of the institute. Authors’ contributions EL, FK, MJS, and AW contributed to the design and implementation of the concept. EL, FK, MJS, and AW contributed to the design and implementation of the experiments, and performing statistical analysis. FK and MAH were involved in collecting and reviewing the data. All authors contributed to the writing and reviewing of 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 and consent to participate The institutional research ethics board (Sunnybrook Health Sciences Centre Research Ethics Board) approved this retrospective single institution study and waived the requirement for informed consent. ==== Refs References 1 Canadian Cancer Statistics Special topic : Predictions of the future burden of cancer in Canada. Technical Report 2015 Canadian Cancer Society http://www.cancer.ca/statistics. 2 Wong A, Glaister J, Cameron A, Haider MA. Correlated diffusion imaging. BMC Med Imaging. 2013; 13(1):26. doi:http://dx.doi.org/10.1186/1471-2342-13-26. 3 Wong A, Khalvati F, Haider MA. Dual-Stage Correlated Diffusion Imaging. In: IEEE International Symposium on Biomedical Imaging (ISBI). Brooklyn: 2015. p. 75–8. 4 Khalvati F, Wong A, Haider MA. Enhanced Dual-Stage Correlated Diffusion Imaging. In: International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Orlando: 2016. p. 5537–40. 5 Esen T, Turkbey B, Patel A, Futterer J. Multiparametric mri in prostate cancer. BioMed Research International. 2014; 2014:296810. doi:http://dx.doi.org/10.1155/2014/296810. 6 Donoho DL. Compressed sensing. IEEE Trans Inform Theory. 2006; 52(4):1289–306. doi:http://dx.doi.org/10.1109/TIT.2006.871582. 7 Candès EJ Compressive sampling Int Congr Math 2006 3 1433 52 8 Baraniuk R Compressive Sensing IEEE Signal Process Mag 2007 24 1 9 10.1109/MSP.2007.361585 9 Lustig M, Donoho D, Pauly JM. Magn Reson Med. 2007; 58(6):1182–95. doi:http://dx.doi.org/10.1002/mrm.21391. 10 Liang D, Liu B, Wang J, Ying L. Accelerating SENSE using compressed sensing. Magn Reson Med. 2009; 62(6):1574–84. doi:http://dx.doi.org/10.1002/mrm.22161. 11 Ye JC, Tak S, Han Y, Park HW. Projection reconstruction MR imaging using FOCUSS,. Magn Reson Med. 2007; 57(4):764–5. doi:http://dx.doi.org/10.1002/mrm.21202. 12 Wong A, Mishra A, Fieguth P, Clausi DA. Sparse reconstruction of breast MRI using homotopic L0 minimization in a regional sparsified domain. IEEE Trans Bio-med Eng. 2013; 60(3):743–52. doi:http://dx.doi.org/10.1109/TBME.2010.2089456. 13 Qu X, Cao X, Guo D, Hu C, Chen Z. Compressed Sensing MRI with Combined Sparsifying transforms and Smooted L0 Norm Minimization. In: IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Dallas: 2010. p. 626–9. 14 Block KT, Uecker M, Frahm J. Magn Reson Med. 2007; 57(6):1086–98. doi:http://dx.doi.org/10.1002/mrm.21236. 15 Trzasko J Manduca A Highly undersampled magnetic resonance image reconstruction via homotopic-minimization Med Imaging IEEE Trans 2009 28 1 106 21 10.1109/TMI.2008.927346 16 Chartrand R Fast algorithms for nonconvex compressive sensing: Mri reconstruction from very few data Biomedical Imaging: From Nano to Macro, 2009. ISBI’09. IEEE International Symposium On 2009 Boston IEEE 17 Yin W Osher S Goldfarb D Darbon J Bregman iterative algorithms for ∖ell_1-minimization with applications to compressed sensing SIAM J Imaging Sci 2008 1 1 143 68 10.1137/070703983 18 Figueiredo MA Nowak RD Wright SJ Gradient projection for sparse reconstruction: Application to compressed sensing and other inverse problems Selected Topics Signal Process IEEE J 2007 1 4 586 97 10.1109/JSTSP.2007.910281 19 Goldstein T Osher S The split bregman method for l1-regularized problems SIAM J Imaging Sci 2009 2 2 323 43 10.1137/080725891 20 Wang Y Yang J Yin W Zhang Y A new alternating minimization algorithm for total variation image reconstruction SIAM J Imaging Sci 2008 1 3 248 72 10.1137/080724265 21 Osher S Burger M Goldfarb D Xu J Yin W An iterative regularization method for total variation-based image restoration Multiscale Model Simul 2005 4 2 460 89 10.1137/040605412 22 Qu X. Compressed sensing MRI based on nonsubsampled contourlet transform. 2008 IEEE International Symposium on IT in Medicine and Education. 2008:693–696. doi:http://dx.doi.org/10.1109/ITME.2008.4743955. 23 Yu Y, Hong M, Liu F, Wang H, Crozier S. Compressed sensing MRI using Singular Value Decomposition based sparsity basis. Conf Proc Ann Int Conf IEEE Eng Med Biol Soc. 2011; 2011(1):5734–7. doi:http://dx.doi.org/10.1109/IEMBS.2011.6091419. 24 Aster R, Borchers B, Thurber C. Parameter estimation and inverse problems. Int Geophys. Series, 90: Elsevier; 2005. 25 Held K Kops ER Krause BJ Wells III WM Kikinis R Muller-Gartner HW Markov random field segmentation of brain mr images Med Imaging, IEEE Trans 1997 16 6 878 6 10.1109/42.650883 26 Blake A Kohli P Rother C Markov Random Fields for Vision and Image Processing 2011 Cambridge MIT Press 27 Lafferty J, McCallum A, Pereira FCN. Conditional random fields: Probabilistic models for segmenting and labeling sequence data. In: ICML (International Conference on Machine Learning): 2001. p. 282–9. 28 Nyquist H Certain topics in telegraph transmission theory Am Inst Electrical Eng Trans 1928 47 2 617 44 10.1109/T-AIEE.1928.5055024 29 Shafiee MJ, Wong A, Siva P, Fieguth P. IEEE International Conference on Image Processing (ICIP). Paris; 2014, pp. 4289–93. 30 Ljunggren S A simple graphical representation of fourier-based imaging methods J Magn Reson (1969) 1983 54 2 338 43 10.1016/0022-2364(83)90060-4 31 Khalvati F, Wong A, Haider M. Automated Prostate Cancer Detection via Comprehensive Multi-Parametric Magnetic Resonance Imaging Texture Feature Models. Biomed Central Med Imaging. 2015:15–27. doi:10.1186/s12880-015-0069-9. 32 Khalvati F, Modhafar A, Cameron A, Wong A, Haider M. A multi-parametric diffusion magnetic resonance imaging texture feature model for prostate cancer analysis. In: MICCAI 2014 Workshop on Computational Diffusion MRI. Boston: 2014. p. 79–88. 33 Cameron A Khalvati F Haider M Wong A MAPS: A Quantitative Radiomics Approach for Prostate Cancer Detection IEEE Trans Biomed Eng 2016 63 6 1145 56 10.1109/TBME.2015.2485779 26441442 34 Cameron A, Modhafar A, Khalvati F, Lui D, Shafiee M, Wong A, Haider MA. Multiparametric MRI Prostate Cancer Analysis via a Hybrid Morphological-Textural Model. In: International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Chicago: 2014. p. 3357–60. 35 Chung A, Scharfenberger C, Khalvati F, Wong A, Haider M. Textural Distinctiveness in Multi-Parametric Prostate MRI for Suspicious Region Detection. In: International Conference on Image Analysis and Recognition (ICIAR). Ontario: 2015. p. 368–76. 36 Zhang J Khalvati F Wong A Haider M Superpixel-based Prostate Cancer Detection from Diffusion Magnetic Resonance Imaging Vision Lett 2015 1 1 107 10.1016/j.visres.2015.01.015

PMC005xxxxxx/PMC5002136.txt

==== Front BMC GastroenterolBMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 52410.1186/s12876-016-0524-2Research ArticleIs endoscopic treatment beneficial in patients with clinically suspicious of common bile duct stones but no obvious filling defects during the ERCP examination? Chiang Po-Hung phchiang1@gmail.com 1Lai Kwok-Hung khlai@vghks.gov.tw 12Tsai Tzung-Jiun medfungi@gmail.com 1Lin Kung-Hung kunghunglin@vghks.gov.tw 1Wang Kai-Ming kwang@vghks.gov.tw 1Kao Sung-Shuo kosuseki@vghks.gov.tw 12Sun Wei-Chih wcsun@vghks.gov.tw 1Cheng Jin-Shiung rcheng@ms2.hinet.net 12Hsu Ping-I pihsu@vghks.gov.tw 12Tsai Wei-Lun wltsai@vghks.gov.tw 12Chen Wen-Chi wcchen@vghks.gov.tw 12Li Yun-Da ydli@vghks.gov.tw 1Wang E-Ming emwang@vghks.gov.tw 1Lin Huey-Shyan sc035@fy.edu.tw 3Chan Hoi-Hung hoihungchan@gmail.com 124561 Division of Gastroenterology, Department of Internal Medicine, Kaohsiung Veterans General Hospital, 386 Ta-Chung 1st Road, Kaohsiung, 81362 Taiwan 2 School of Medicine, National Yang-Ming University, No. 155, Sec. 2, Li-Nong Street, Pei-Tou, Taipei, 112 Taiwan 3 Department of Health-Business Administration, Fooyin University, 151 Jinxue Rd, Daliao Dist, Kaohsiung City, 83102 Taiwan 4 Department of Biological Sciences, National Sun Yat-sen University, 70 Lien-Hai Road, Kaohsiung, 80424 Taiwan 5 Department of Business Management, National Sun Yat-sen University, 70 Lien-Hai Road, Kaohsiung, 80424 Taiwan 6 College of Pharmacy and Health Care, Tajen University, 20 Weisin Road, Sin-er Village, Yanpu Township, Pingtung County 907 Taiwan 26 8 2016 26 8 2016 2016 16 1 10210 2 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 Sometimes, no definite filling defect could be found by cholangiogram (ERC) during the endoscopic retrograde cholangio-pancreatiographic (ERCP) exam; even prior images had evidence of common bile duct stones (CBDS). We aimed in estimating the positive rate of extraction of CBDS who had treated by endoscopic sphincterotomy/endoscopic papillary balloon dilation (EST/EPBD) with negative ERC finding. Methods One hundred forty-one patients with clinically suspicious of CBDS but negative ERC, who had received EST/EPBD treatments was enrolled. Potential factors for predicting CBDS, as well as the treatment-related complications were analyzed. Results Nearly half of the patients with negative ERC, had a positive stone extraction. Only patients with high probability of CBDS were significantly associated with positive stone extraction. Moreover, patients with intermediate probability of CBDS had higher rates of overall complications, including post-ERCP pancreatitis. In addition, no significant difference of post-ERCP pancreatitis was found between EST and EPBD groups in any one group of patients with the same probability of CBDS. Conclusions Regarding patients with negative ERC, therapeutic ERCP is beneficial and safe for patients present with high probability of CBDS. Moreover, under the same probability of CBDS, there was no significance difference in post-ERCP pancreatitis between EST and EPBD. Keywords ERCERCPCBDSissue-copyright-statement© The Author(s) 2016 ==== Body Background Common bile duct stone (CBDS) is an important clinical problem that can cause serious complications, such as acute cholangitis and pancreatitis [1]. Therefore, it is recommended to remove the stones endoscopically or surgically once diagnosis is established [2]. However, sometimes, early definitive diagnosis of choledocholithiasis is difficult and should be based on clinical symptoms and signs, biochemical data and image findings. Persist elevation of serum alkaline phosphatase (ALP) and alanine transaminase (ALT) were shown to correlate with the presence of CBDS even with a normal-sized CBD [3]. A recent study showed that trans-abdominal ultrasound alone is inadequate to predict the CBDS in patients presenting with acute cholecystitis [4]. Endoscopic retrograde cholangiopancreatography (ERCP) is generally believed to be the gold standard for both diagnosis and treatment of CBDS. However, inevitably, the procedure is associated with an overall complication rate of 4 ~ 10 % and mortality rate of 0.02 ~ 0.5 % [5–10]. The major complications include pancreatitis (1.3 ~ 6.7 %), infection (0.3 ~ 5.0 %), hemorrhage (0.3 ~ 2.0 %), and perforation (0.1 ~ 1.1 %) [6, 9, 11]. Others include cardiac (<0.1 %), and pulmonary events (<0.1 %) [6]. Therefore, currently, purely diagnostic ERCP is not suggested [7, 8]. Instead, relative non-invasive imaging modalities such as MRCP and EUS are preferred. In 2010, the American Society for Gastrointestinal Endoscopy (ASGE) established a general rule for the evaluation of likelihood of choledocholithiasis, in which; Patients were divided into “high probability (risk of CBDS > 50 %)”, “intermediate probability (risk of CBDS: 10 ~ 50 %)”, and “low probability (risk of CBDS < 10 %)” groups [12]. In addition, the author also pointed-out the management algorithm for patients with symptomatic choledocholithiasis [12]. However, sometimes, no obvious filling defects inside CBD could be found by cholangiogram (ERC, Fig. 1), even prior images, such as trans-abdominal ultrasound or CT scan, had demonstrated the evidence of CBDS. Regarding the possible complications, further the treatment procedures, such as endoscopic sphincterotomy (EST) and/or papillary balloon dilation (EPBD) in this situation is worthy consideration. The aim of this retrospective study was to estimate the positive rate of CBDS in patients with negative filling defects from ERC, and the factors for possible CBDS prediction, as well as the treatment-related complications (safety concern).Fig. 1 A female patient presented with epigastric pain, jaundice, while CT scan showing dilated CBD and suspicious of CBDS. In addition, no definite filling defect was found by ERC. However, EPBD was performed and a 1-cm hard stone was extracted subsequently Methods Study design, definition and patient selection This retrospective study was approved by Institutional Review Board of Kaohsiung Veterans General Hospital. Eligibility of patients includes those who were clinically classified as either intermediate or high risks for CBDS [12] according to symptoms and signs, laboratory data or image studies during the period of April 2008 to March 2014. These patients had received either EST or EPBD treatments, although no obvious filling defect was detected by ERC. Exclusion criteria include peri-ampullary tumors, hepatocellular disease, hemolytic disease, and patients who ever received endoscopic or surgical treatment for bile duct stones. In addition, positive extraction of stones was defined as stones detectable by naked eyes through the video endoscopic pictures during extracting the bile duct by using basket or balloon catheter; or presence of microlithiasis (non-visible by naked eyes) under microscopic exams of the bile. The model of “probability of CBDS” from American society for gastrointestinal endoscopy (ASGE, 2010) was applied in this study, in which high probability of CBDS includes: 1) CBDS seen on trans-abdominal ultrasonography (US) (and/or CT scan), 2) signs of acute cholangitis (people who had Charcot’s triad), 3) total bilirubin > 4 mg/dL, 4) both dilated CBD on US (>6 mm with gallbladder in situ and > 10 mm with cholecystectomy) and total bilirubin level 1.8 ~ 4 mg/dL; and intermediate probability of CBDS includes: 1) either one of these two factors: dilated CBD on US or total bilirubin level 1.8 ~ 4 mg/dL, 2) advanced age (>55 year-old), 3) elevation of a liver biochemical test other than bilirubin, and 4) gallstone pancreatitis. In addition, ERCP-related complications were defined and graded in severity according to the consensus criteria, which was adapted as (Table 1), developed by Cotton et al. [6, 7, 11].Table 1 Consensus criteria for ERCP complicationsab Mild Moderate Severe Bleeding Clinical evidence of bleeding (ie, not just endoscopic); Hb level drop <3 g; no need for transfusion. Transfusion: ≤4 units; no angiographic intervention or surgery. Transfusion: ≥5 units or intervention (angiographic or surgical). Perforation Possible, or only very slight leak of fluid or contrast dye; treatable by fluids and suction for ≤3 days. Any definite perforation treated medically for 4–10 days. Medical treatment for more than 10 days or intervention (percutaneous or surgical). Pancreatitis Clinical pancreatitis; amylase at least 3 times normal at more than 24 hours after the procedure requiring admission or prolongation of planned admission to 2–3 days. Pancreatitis requiring hospitalization of 4–10 days. Pancreatitis requiring hospitalization for more than 10 days, or hemorrhagic pancreatitis, phlegmon or pseudocyst, or intervention (percutaneous drainage or surgery). Infection (cholangitis) >38°C at 24–48 hours. Febrile or septic illness requiring >3 days of hospital treatment or endoscopic or percutaneous intervention Septic shock or surgery. ie, mild, unplanned hospital stay of 2–3 nights; moderate, 4–10 nights; and severe (>10 nights or intensive care or surgery) aFrom Ref. 6 and 11. ERCP, endoscopic retrograde cholangiopancreatography bAll other complications were graded for severity of the need for hospitalization and/or surgical treatment Endoscopic Procedures Patients were conscious for the procedure and received 10 % xylocaine spray for local anesthesia of the pharynx, intramuscular injection with 40 mg hyoscine-N-butylbromide, and intramuscular injection with 25–50 mg meperidine. ERCP was performed in the standard manner using a side-view endoscope (JF-240; Olympus Optical Corporation, Tokyo, Japan). After selective cannulation of the common bile duct by the catheter, cholangiography was performed to evaluate the presence/absence of filling defects inside CBD. A 0.035-in. guide wire was then inserted into the bile duct through the catheter. For EST group, sphincterotomy was done by using a wire-guided sphincterotome. Incision was started at the orifice of papilla and extended upward to the direction of bile duct. For EBPD, selective cannulation of the common bile duct with guide wire insertion was the same as EST. A dilating balloon (CRE balloon; Boston Scientific, Corp, Ireland) was passed via the prepositioned 0.035-in. guide wire into the bile duct. Using fluoroscopic (AXIOM, Iconos R200, Siemens AG 2002) and endoscopic guidance, the balloon was inflated with sterile saline solution up to the optimal size (at least > 6 mm in diameter) and duration (from 1.5 to 5 min) according to the patients’ condition and tolerance. In order to minimize the risk of perforation, the size of the balloon should be not exceed the size of the CBD. After the balloon and guide wire were removed, the CBDS was retrieved out using a Dormia basket or balloon-tipped catheter with or without the aid of mechanical lithotripsy (BML-4Q; Olympus Optical, Tokyo, Japan). Unnecessary cannulation or contrast injection of pancreatic duct was avoided. Statistical analysis All statistical analyses were performed using the PASW 20.0 (IBM, New York, NY, USA). Continuous valuables are expressed as mean ± SD. Chi-square analyses or Fisher’s exact tests were used for comparing categorical variables, while independent t-tests were used for comparing continuous variables between patients with final positive and negative stone extraction. Associations between the possible predictors and the positivity of CBD stones and between the possible predictors and complications were assessed by multiple logistic regressions. Results were shown as odds ratios and 95 % confidence intervals (CIs). A p-value less than 0.05 was considered statistically significant. Results Demographic data was shown in Table 2. No significant difference was found at gender, age, body mass index (BMI), initial GPT and ALP level, history of cholecystectomy, presence/absence of gallbladder stones, and juxta-papillary diverticulum (JPD), between patients with final positive or negative stone extraction. There were only initial cholangitis and high probability of CBDS significantly associated with positive stone extraction. There were total 141 (male/female: 81/60) patients, clinically suspicious of CBD stones (intermediate probability: 28, high probability: 113), undergoing successful therapeutic ERCP (EST/EPBD: 30/111) with which pre-treatment cholangiogram (ERC) showed no obvious filling defects. For the group of positive stone extraction (70 patients), there were 64 patients showed detectable (all are barely visible by naked eyes and un-measurable) stones and six patients showed microlithiasis under microscopic analysis of bile. However, there were only 21 samples of bile available for analysis (21/141 = 14.9 %). With regard to the high probability group of CBDS, 65 cases presented with evidence of CBDS at initial image, 25 with acute cholangitis, 10 with total bilirubin level >4 mg/dL, and 13 with mild elevated total bilirubin (1.8–4 mg/dL) and CBDdilatation. On the other hand, in the intermediate probability group, 16 cases presented with mild elevated total bilirubin (1.8–4 mg/dL) without CBD dilatation, six with CBD dilatation without elevated total bilirubin, two with gallstone pancreatitis, and four with age > 55 year-old. Besides, the mean length of EST was 0.91 cm (0.5 cm ~ 1.5 cm); and the mean size of dilating balloon was 0.99 cm (0.6 cm ~ 1.8 cm), depend on the relative sizes of CBD. ERCP was performed at a mean of four days after admission.Table 2 Demographic data between groups with and without stone extraction Characteristics Stone (+) Stone (−) P-value (n = 70) (n = 71) Gender (Male/Female) 40/30 41/30 0.942 Age 66.73 ± 17.93 61.63 ± 15.55 0.073 BMI 24.54 ± 3.58 24.77 ± 3.95 0.736 Cholecystectomy 8 (11.4 %) 8 (11.3 %) 0.976 GB stone 54 (77.1 %) 54 (76.1 %) 0.879 JPD 31 (44.3 %) 26 (36.6 %) 0.354 ALT at admission 278.9 ± 282.8 283.7 ± 260.1 0.917 Alk-P at admission 175.1 ± 120.7 162.5 ± 149.7 0.590 Total bilirubin at admission 3.49 ± 2.37 3.35 ± 2.20 0.726 CBDS risk (high vs. intermediate) 62 (88.6 %) 51 (71.8 %) 0.013* Cholangitis 29 (41.4 %) 14 (19.7 %) 0.005* Pancreatitis 26 (37.1 %) 22 (31.0 %) 0.440 Abbreviations: BMI body mass index, GB gallbladder, JPD juxta-papillary diverticulum, ALT aspartate transaminase, Alk-P alkaline phosphatase; CBDS, common bile duct stone *p < 0.05 By using multiple logistic regressions, only high probability of CBDS was found to be significantly associated with positive stone extraction (high vs. intermediate probability: 54.9 % vs. 28.6 %, p = 0.039) (Table 3). Moreover, there were totally 11 (7.8 %) complications found in the study (Table 4). By using multiple logistic regressions, intermediate probability of CBDS was associated with higher risk of overall complications and post-ERCP pancreatitis (p = 0.043; p = 0.007) (Tables 4 and 5). In addition, no significant difference in overall complications, including post-ERCP pancreatitis, was found between EST and EPBD groups under the same probability of CBDS, no matter high or intermediate probability. There were three (mild/moderate/severe: 1/1/1) and four (mild/moderate/severe: 1/3/0) post-ERCP pancreatitis found in EST and EPBD groups, respectively. Moreover, two mild cholangitis combined with moderate pancreatitis and two pure cholangitis (mild/moderate/severe: 1/0/1) were found in EPBD group. However, no procedure-related mortality was noted in the current study.Table 3 Risk factors of patients with stone extraction Characteristics Complication (+) Complication (−) P-value OR 95 % CI of OR (n = 70) (n = 71) CBDS probability from ASGE  CBDS risk (high vs. intermediate) 62 (88.6 %) 51 (71.8 %) 0.039* 2.670 1.050 ~ 6.790 Characteristics  Gender (M/F) 40/30 41/30 0.992 1.004 0.496 ~ 2.031  BMI 24.54 ± 3.58 24.77 ± 3.95 0.540 0.971 0.883 ~ 1.067  Cholecystectomy 8 (11.4 %) 8 (11.3 %) 0.634 1.430 0.329 ~ 6.225  GB stone 54 (77.1 %) 54 (76.1 %) 0.504 1.453 0.486 ~ 4.349  JPD 31 (44.3 %) 26 (36.6 %) 0.212 1.593 0.767 ~ 3.310 Abbreviation: OR odd’s ratio *p < 0.05 Table 4 Risk factors of the patients with complication Characteristics Complication (+) Complication (−) P-value OR 95 % CI of OR (n = 11) (n = 130) Procedure  EST/EPBD 3 (27.3 %) 27 (20.8 %) 0.909 0.919 0.215 ~ 3.920 CBDS probability from ASGE  CBDS risk (high vs. intermediate) 6 (54.5 %) 107 (82.3 %) 0.043* 0.262 0.072 ~ 0.958 Abbreviations: EST endoscopic sphincterotomy, EPBD endoscopic papillary balloon dilation *p < 0.05 Table 5 Risk factors of the patients with post-ERCP pancreatitis Characteristics Post-ERCP pancreatitis (+) Post-ERCP pancreatitis (−) P-value OR 95 % CI of OR (n = 7) (n = 134) Procedure  EST/EPBD 3 (42.9 %) 27 (20.1 %) 0.442 0.523 0.100 ~ 2.735 CBDS probability from ASGE  CBDS risk (high vs. intermediate) 2 (28.6 %) 111 (82.8 %) 0.007* 0.093 0.017 ~ 0.523 *p < 0.01; correlation between procedure and CBDS probability from ASGE showed no significance (p = 0.456) Discussion According to the current study, nearly half (49.6 %) of patients without detected filling defects in ERC, have evidence of positive stone extraction after EST or EPBD treatments. By multiple logistic regressions, only high probability of CBDS was significantly associated with positive stone extraction. Total complication rate among patients received EST or EPBD with negative filling defects from ERC was 7.80 %, and no significant difference was found between these two treatment modalities. In addition, there was no procedure-related mortality. Furthermore, intermediate probability of CBDS was associated with higher risk of overall complications, including post-ERCP pancreatitis. Therefore, endoscopic treatment (EST or EPBD) is beneficial and safe for patients with high probability of CBDS. In addition, no significant difference in overall complications, as well as post-ERCP pancreatitis, was found between EST and EPBD groups under the same probabilities of stones. The lack of important roles of liver function tests, such as GPT, ALP before ERCP in the current results, as in the previous studies [13–16] might be due to the small sample size. Moreover, bile analysis was inadequately done in this study (14.9 %). Therefore, prospective study with bile analysis of microlithiasis is crucial to elucidate the true rate of CBDS in patients with negative filling defects in ERC. In addition, endoscopic ultrasound might be done before EST and EPBD in order to minimize the ERCP-associated complications and to quickly delineate the presence of small stones or sludge in the CBD [5]. Conclusions The probability of CBDS (high vs. intermediate probability) could play a significant role in the estimation of positive stone extraction before deciding the therapeutic strategies, with the result in fewer overall complications, including post-ERCP pancreatitis after the treatment even though the negative filling defect on ERC. In addition, endoscopic treatment (EST or EPBD) is beneficial and safe to patients with high probability of CBDS. Moreover, under the same probability scores, there was no significant difference in post-ERCP pancreatitis between EST and EPBD. Future prospective study with bile analysis of microlithiasis is important to elucidate the true rate of CBDS in patients with negative filling defects in ERC. Abbreviations CBDScommon bile duct stones EPBDendoscopic balloon dilation ERCendoscopic retrograde cholangiography ERCPendoscopic retrograde cholangio-pancreatography ESTendoscopic sphincterotomy Acknowledgements The authors thank Ms Daisy Lo for checking the English grammar. Funding None. Availability of data and materials The datasets analyzed during the current study are available from the corresponding author on reasonable request. Authors’ contributions PHC, HSL, HHC, JSC and PIH designed the study and analyzed the data. PHC, HHC, and HSL were responsible for writing the manuscript and revising it critically for important intellectual content. PHC, HHC, KHL, TJT, KHL, KMW, SSK, WCS, YDL were responsible for the ERCP procedures. EMW assisted the endoscopic procedures. HHC, WLT, and WCC were responsible for patient care. All authors have read and approved the final manuscript. Authors’ information Not applicable. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The current study was approved by the Institutional Review Board of Kaohsiung Veterans General Hospital. (IRB # VGHKS13-CT9-10); consent was not required per IRB for this is a retrospective article. ==== Refs References 1. John LG, Gregory BB, Anna MD, Janet DE, Michael PF, J. Michael H, et al. National Institutes of Health Consensus Development Conference Statement on Gallstones and Laparoscopic Cholecystectomy. Am J Surg 1993, 165;(4):390–8. 2. Committee ASoP Maple JT Ikenberry SO Anderson MA Appalaneni V Decker GA The role of endoscopy in the management of choledocholithiasis Gastrointest Endosc 2011 74 4 731 744 10.1016/j.gie.2011.04.012 21951472 3. Isherwood J Garcea G Williams R Metcalfe M Dennison AR Serology and ultrasound for diagnosis of choledocholithiasis Ann R Coll Surg Engl 2014 96 3 224 228 10.1308/003588414X13814021678033 24780789 4. Boys JA Doorly MG Zehetner J Dhanireddy KK Senagore AJ Can ultrasound common bile duct diameter predict common bile duct stones in the setting of acute cholecystitis? Am J Surg 2014 207 3 432 435 10.1016/j.amjsurg.2013.10.014 24581769 5. Chan HH Wang EM Sun MS Hsu PI Tsai WL Tsai TJ Linear echoendoscope-guided ERCP for the diagnosis of occult common bile duct stones BMC Gastroenterol 2013 13 44 10.1186/1471-230X-13-44 23497328 6. Cotton PB Garrow DA Gallagher J Romagnuolo J Risk factors for complications after ERCP: a multivariate analysis of 11,497 procedures over 12 years Gastrointest Endosc 2009 70 1 80 88 10.1016/j.gie.2008.10.039 19286178 7. Cotton PB Lehman G Vennes J Geenen JE Russell RC Meyers WC Endoscopic sphincterotomy complications and their management: an attempt at consensus Gastrointest Endosc 1991 37 3 383 393 10.1016/S0016-5107(91)70740-2 2070995 8. Freeman ML Nelson DB Sherman S Haber GB Herman ME Dorsher PJ Complications of endoscopic biliary sphincterotomy N Engl J Med 1996 335 13 909 918 10.1056/NEJM199609263351301 8782497 9. Loperfido S Angelini G Benedetti G Chilovi F Costan F De Berardinis F Major early complications from diagnostic and therapeutic ERCP: a prospective multicenter study Gastrointest Endosc 1998 48 1 1 10 10.1016/S0016-5107(98)70121-X 9684657 10. Masci E Toti G Mariani A Curioni S Lomazzi A Dinelli M Complications of diagnostic and therapeutic ERCP: a prospective multicenter study Am J Gastroenterol 2001 96 2 417 423 10.1111/j.1572-0241.2001.03594.x 11232684 11. Williams EJ Taylor S Fairclough P Hamlyn A Logan RF Martin D Risk factors for complication following ERCP; results of a large-scale, prospective multicenter study Endoscopy 2007 39 9 793 801 10.1055/s-2007-966723 17703388 12. Committee ASoP Maple JT Ben-Menachem T Anderson MA Appalaneni V Banerjee S The role of endoscopy in the evaluation of suspected choledocholithiasis Gastrointest Endosc 2010 71 1 1 9 10.1016/j.gie.2009.09.041 20105473 13. Al-Jiffry BO Elfateh A Chundrigar T Othman B Almalki O Rayza F Non-invasive assessment of choledocholithiasis in patients with gallstones and abnormal liver function World J Gastroenterol 2013 19 35 5877 5882 10.3748/wjg.v19.i35.5877 24124333 14. Barkun AN Barkun JS Fried GM Ghitulescu G Steinmetz O Pham C Useful predictors of bile duct stones in patients undergoing laparoscopic cholecystectomy. McGill Gallstone Treatment Group Ann Surg 1994 220 1 32 39 10.1097/00000658-199407000-00006 7517657 15. Onken JE Brazer SR Eisen GM Williams DM Bouras EP Delong ER Predicting the presence of choledocholithiasis in patients with symptomatic cholelithiasis Am J Gastroenterol 1996 91 4 762 767 8677945 16. Peng WK Sheikh Z Paterson-Brown S Nixon SJ Role of liver function tests in predicting common bile duct stones in acute calculous cholecystitis Br J Surg 2005 92 10 1241 1247 10.1002/bjs.4955 16078299

PMC005xxxxxx/PMC5002137.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 129110.1186/s12906-016-1291-xErratumErratum to: Role of quercetin and arginine in ameliorating nano zinc oxide-induced nephrotoxicity in rats Faddah Laila M. 1Abdel Baky Nayira A. nayira@ksu.edu.sa 13Al-Rasheed Nouf M. 1Al-Rasheed Nawal M. 1Fatani Amal J. 1Atteya Muhammad 21 Pharmacology Department, Faculty of Pharmacy, King Saud University, Riyadh, Saudi Arabia 2 Anatomy Department and Stem Cell Unit, Faculty of Medicine, King Saud University, Riyadh, Saudi Arabia 3 Department of Pharmacology, Faculty of Pharmacy, Al-Azhar University, Cairo, Egypt 26 8 2016 26 8 2016 2016 16 1 32315 8 2016 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.issue-copyright-statement© The Author(s) 2016 ==== Body Erratum Following publication of the original article [1] it was brought to our attention that affiliation 1 and 3 referred to the same institution, despite the fact that these should be two distinct affiliations. Please therefore note that the correct version of affiliation 3 should read as follows: 3Department of Pharmacology, Faculty of Pharmacy, Al-Azhar University, Cairo, Egypt The online version of the original article can be found under doi:10.1186/1472-6882-12-60. ==== Refs Reference 1. Faddah LM Role of quercetin and arginine in ameliorating nano zinc oxide-induced nephrotoxicity in rats BMC Complement Altern Med 2012 12 60 10.1186/1472-6882-12-60 22551254

PMC005xxxxxx/PMC5002138.txt

==== Front Parasit VectorsParasit VectorsParasites & Vectors1756-3305BioMed Central London 173710.1186/s13071-016-1737-5ResearchCross-sectional study of Fasciola gigantica and other trematode infections of cattle in Edu Local Government Area, Kwara State, north-central Nigeria http://orcid.org/0000-0003-1255-4572Elelu Nusirat nusyelelu@yahoo.com 12Ambali Abdulganiyu aambali076@yahoo.com 1Coles Gerald C. gerald.c.coles@bristol.ac.uk 2Eisler Mark C. mark.eisler@bristol.ac.uk 21 University of Ilorin, Faculty of Veterinary Medicine, Ilorin, Kwara State Nigeria 2 University of Bristol, School of Veterinary Science, Langford, Bristol BS40 5DU UK 26 8 2016 26 8 2016 2016 9 1 4704 11 2015 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 Trematode infections of livestock are of global veterinary and public health importance causing serious economic losses. Majority of data on burden of trematode infections in Nigeria are based on abattoir surveys and there are very few data on herd level risk factors. The present study investigated the prevalence of, and herd level risk factors for, fasciolosis and other trematode infections in cattle in Edu Local Government Area (LGA). Methods A cross-sectional survey used two-stage study design to investigate cattle belonging to 65 households. Two questionnaires were administered for household-level and individual cattle-level data. Faecal and blood samples were obtained from the cattle. Logistic regression analyses were performed to determine risk factors for infections. Results Of 686 faecal samples analysed, 74.9 %, 16.1 %, 7.3 % and 1.2 % were positive for infections with Fasciola gigantica, paramphistomes, Dicrocoelium hospes and Schistosoma bovis respectively. Fasciola gigantica had higher prevalence in adult cattle (77.3 %) than weaners (62.5 %). Majority of co-infections was a combination of F. gigantica with paramphistomes 84/130 (64.6 %). Most (58.9 %) of the cattle belonged to FAMACHA© score 2. The mean packed cell volume (PCV) was 34.4 %. The sensitivity and specificity of FAMACHA© for anaemia (PCV < 24 %) were 18.2 and 96.9 %, respectively. Positive correlation was obtained between faecal egg counts for F. gigantica and paramphistomes (R = 0.15, P = 0.0001). Adult cattle were more likely to be infected with F. gigantica (odds ratio, OR: 1.94; Confidence Interval, CI: 1.19–3.16) than weaners. Cattle belonging to household heads aged between 40–59 years were more likely infected with paramphistomes (OR: 1.95; CI: 1.02–3.74) than those belonging to other age groups. Cattles from herds with size ≥ 100 were more likely infected with D. hospes than those from smaller herds (OR: 6.98; CI: 2.94–16.6). Conclusion This study revealed high prevalence of infection with F. gigantica in Kwara State. The co-infections by F. gigantica and paramphistomes with a positive correlation should be considered during anthelmintic therapy. There is a need to optimise and validate the FAMACHA© for use in cattle based on breeds and variation in colour of ocular mucous membrane. Risk factors identified could assist in tailoring control strategies for various trematode infections to particular groups of farmers and cattle. Electronic supplementary material The online version of this article (doi:10.1186/s13071-016-1737-5) contains supplementary material, which is available to authorized users. Keywords TrematodesFasciolaParamphistomesDicrocoeliumSchistosomaNigeriaKwaraPrevalenceRisk factorsFAMACHA©issue-copyright-statement© The Author(s) 2016 ==== Body Background Trematode infections cause serious economic losses to livestock globally. Many are zoonotic and thus a public health concern [1, 2]. Some of the trematode infections of cattle include species of Fasciola, Dicrocoelium, Schistosoma and paramphistomes. Fasciolosis due to F. gigantica has been reported in several parts of Africa [3–6] and Nigeria [7–10]. Other trematode infections in ruminants reported in Nigeria include species of Dicrocoelium [11], Schistosoma [12] and paramphistomes [7]. Trematode infections are known to cause clinical signs ranging from weight loss, sudden death [13] and anaemia in cattle [14, 15]. Tropical fasciolosis alone has been predicted to cause losses of about US$840 M per annum in the Africa’s 200 million cattle population [16] and this cost is likely to have increased significantly in the last sixteen years. Economic losses from fasciolosis may result directly from increased liver condemnation or indirectly from decreased livestock productivity [17]. Also about 165 million cattle are likely to be infected with Schistosoma spp. worldwide [18]. The cattle population in Nigeria is about 16 million [19] made up of predominantly humped zebu breeds (including the White Fulani, Sokoto Gudali and Red Bororo) and a limited number of hump less breeds including Keteku, Muturu and Kuri in the southwestern, southern and the northeastern parts, respectively [20]. They play a very important role in the Nigerian economy, contributing about 12.7 % of total agriculture gross domestic product (GDP) [21]. In the tropics, cattle are generally reared under the transhumance husbandry system with little supplementary feeding resulting in low productivity and high pre-weaning mortality [8]. Similarly, acute shortage of feeds during the dry season remains a common occurrence, compelling these animals to graze around water bodies that often contain large number of potential intermediate hosts of trematodes [8]. The majority of data on the burden of fasciolosis in Nigeria are based on abattoir surveys. However, there are very few data on the trematode prevalence in live cattle or on the herd level risk factors that may influence disease occurrence in Nigeria. Moreover, there are few recent data on infection of cattle with the other trematode species in Nigeria, and more recent information would be useful in formulating effective control strategies for this important group of parasites [7, 8, 22]. The present study investigated the prevalence of, and herd level risk factors for, fasciolosis and other trematode infections in cattle in the Edu Local Government Area (LGA). Methods Study location A cross-sectional study was conducted from May to August 2013 to determine the prevalence of trematode infections and herd level risk factors in cattle from 11 villages of Edu LGA, Kwara State, North-central Nigeria. Kwara State lies between 8°05′ and 10°15′N; and 2°73′ and 6°13′E (Fig. 1). It has a total area of about 34,500 square kilometres comprising rainforest in the south and wooded savannah in the larger part of the state. It has 16 local government areas. Rainfall has an annual range of 1,000–1,500 mm and average maximum temperature between 30 and 35 °C [23]. Edu LGA was selected as the study location because it has very large pastoralist settlements and is one of the largest area for cattle production in Kwara State. Rice, sugarcane and melon are the major crops planted. Because Edu LGA is bounded by the River Niger in the north, the area is often inundated with flood leading to devastating losses of livestock and farmland. The pastoralists therefore migrate uphill away from flood plains (starting in July) to neighbouring states once the rains begin. They do not return until the end of the year when the rains cease. A local informant identified 11 cattle producing villages in Edu LGA and these formed the sample population of this study.Fig. 1 Map of Kwara State showing the location of Edu Local Government Area (study location). The inset map shows Kwara State within Nigeria Study design and sampling A two-stage sampling design was carried out. The first stage determined the number of households to be selected while the second stage determined the number of cattle to be sampled in each household. The number of households to be visited was calculated based on the formula: 1.962 * Pexp (1-Pexp)/ d2 [24] with an expected prevalence (Pexp) of 50 % (no previous data on herd level prevalence in the area), 10 % desired precision and 95 % confidence interval (1.96). The second sampling stage to determine the number of cattle to be sampled per household and was applied to the two-stage sampling protocol of Cameron & Baldock [25] developed to determine presence of disease. The protocol was implemented by using the FreeCalc software version 2 [(c) Copyright 2001-Angus Cameron AusVet Animal Health Services]. The sensitivity (92.7 %) and specificity (94.9 %) of the Flukefinder ® [26], 40 % minimum expected prevalence and cattle population interval of 1–250 animals were included as parameters. A maximum of 13 animals are required to be sampled in each household. Only cattle ≥ 12 month-old were sampled to ensure they had experienced at least one complete grazing season. A total of 686 cattle in 65 households were sampled for trematode infections. While the sample size calculations indicated that 96 households were required, there were only 65 accessible households present with cattle during the survey (others had either migrated uphill away from floodplains or were inaccessible due to the flood). All study locations were georeferenced using Garmin® global positioning system (GPS). Both faecal and blood samples were collected from the animals. Data collection Two questionnaires were administered. The first focused on household level data such as farmer socio-demographic characteristics, management system, health practices, herd size, economic activities as well as knowledge and practice to control liver fluke disease. The second questionnaire was for individual cattle data and the information recorded included age, sex and breed. Both questionnaires were included in the analysis of herd level risk factors for trematode infections. Coprological analysis Faecal samples were obtained directly from the rectum of each cow into sterile plastic gloves or from the ground if seen being produced. The glove were turned inside out, carefully tied, labelled and transported under cool conditions to the laboratory for analysis. The commercially available kit FlukeFinder® Richard Dixon ID, USA (http://www.flukefinder.com/) was used to isolate the trematode eggs by differential sieving and sedimentation according to manufacturer’s instructions. The sensitivity and specificity of the Flukefinder® has been previously compared to other sedimentation methods [26]; this device has also been used previously to isolate trematode eggs [6, 27]. Two grams of individual cattle faecal sample were used and analysed according to the manufacturers’ instructions. The Flukefinder® is made up of two 2-in. wide sieves; these were washed thoroughly in between samples to prevent cross-contamination. The material was poured into a 50 mm Petri dish, three drops of methylene blue was added for contrast and then examined under a stereomicroscope at a magnification of 40×. Eggs were identified using standard keys [28]. Although the eggs of F. gigantica and paramphistomes are similar (both oval and operculated), the eggs of F. gigantica possess a distinct yellowish-brown colour and measure 156–197 μm in length and 90–104 μm in width compared with those of paramphistomes, which are clear in colour and measure 114–176 by 73–100 μm. Eggs of D. hospes are small (36–45 × 22–30 μm), oval, dark brown and operculate, with two characteristic dark “eye-spots” [28] and eggs of S. bovis are spindle-shaped with characteristic terminal spines on both sides of the non-operculate eggs [28]. Haematological analysis Blood samples for haematological analysis were collected from the jugular vein of the first three to four cattle sampled using EDTA anticoagulant (due to laboratory costs, blood samples were not obtained from all animals sampled in each household). The level of anaemia was also checked using the FAMACHA© anaemia chart to score the ocular mucous membrane [29]. The FAMCHA© chart is a low cost tool in determining anaemia status of ruminants [30]. The colour of ocular mucous membrane for each animal classified into five categories based on the FAMACHA© chart (ranging from bright red to pale) and recorded for individual animals. The packed cell volume (PCV) was also determined using the microhaematocrit method [31]. Statistical analysis Descriptive analyses were carried out on cattle data. Cattle that had at least one trematode egg count was considered positive. The overall prevalence (in %) at the animal level was the total number of cattle positive divided by the total number of cattle sampled in the study. Chi-square or Fisher’s exact tests were used to explore the relationships between the trematode prevalence, household and cattle data. Descriptive statistics for trematode co-infections are also presented. The sensitivity and specificity of the FAMACHA© score compared to PCV were determined. Animals with FS scores of 1–3 were classified as non-anaemic while those with FS scores of 4–5 were grouped as anaemic. Reference value of 24–46 % for PCV in cattle was used [32]. The PCV values of ≤24 % were classified as anaemic and those above 25 % were considered non anaemic. Pearson’s correlation coefficient was used to explore the relationships between trematode infections. In order to carry out correlation analyses, trematode egg counts were log-transformed [log(egg count +1)] to stabilise variances. The relationship between log-transformed trematode faecal egg counts and PCV was also determined. Trematode infections and herd level risk factors were investigated by using predictor variables such as the household level factors, individual cattle data and haematological indices. The outcome variable was cattle trematode infection status expressed as a binary variable (0 = negative; 1 = positive). Variables with values for P < 0.20 from the Chi-square or Fisher’s exact test obtained from the coprological analysis were included in the binary logistic regression analysis. The risk factors were explored by binary logistic regression using the forward stepwise variable selection method and the 95 % confidence interval of odds ratio was calculated for the predictors [33]. Statistical analyses were carried out using Microsoft Excel® and IBM SPSS® statistics version 21.0 software (IBM Corp. 2012, Armonk, NY, USA). Distribution of trematode infections were also represented using the QGIS® spatial software version 2.2. Results Household and cattle data A total of 65 households was surveyed (Additional file 1: Table S1). The respondents were all male with a mean age of 42.75 (SD ± 11.24, range 23–70) years. They were mostly married (96.9 %) and are of Fulani (83 %) ethnic origin. The majority had no formal education (93.5 %). They kept mainly livestock (67.7 %) comprising primarily cattle (90.5 %). The mean cattle herd size was 62.31 ± 46.26; cattle were kept mainly by pastoral/nomadic grazing system (78.5 %). A total of 686 cattle was investigated using questionnaire surveys, comprising 318 (46.8 %) males and 362 (53.2 %) females. The mean age of cattle was 3.57 (SD ± 1.38) years with minimum and maximum ages of one and ten years, respectively. The breed composition was predominantly humped zebu 655 (95.9 %) with a very small minority made up of Jersey 27 (4.0 %) and a single Friesian (0.1 %). Coprological data A total 686 faecal samples was analysed from cattle in 11 villages from which 536 (78.1 %) were positive for at least one of the parasites studied: 514 (74.9 %; 95 % CI: 72.0–79.0 %) for infections with F. gigantica, 110 (16.0 %; 95 % CI: 13.0–19.0 %) with paramphistomes, 50 (7.3 %; 95 % CI: 5.0–9.0 %) with Dicrocoelium hospes and 8 (1.2 %; 95 % CI: 0–2.0 %) with Schistosoma bovis. Of these 686 cattle sampled, 406 (75.7 %) had single species trematode infections of which 385/686 (56.1 %) were F. gigantica, 9/686 (1.3 %) paramphistome, 9/686 (1.3 %) D. hospes and 3/686 (0.4 %) S. bovis. One hundred and thirty cattle 130/386 (19.0 %) had trematode co-infections and 150 cattle were uninfected (Fig. 2). Of the 130 cattle with co-infections, 84/130 (64.6 %), 25/130 (19.2 %) and 4 (3.1 %) of the animals had co-infection of F. gigantica with paramphistomes, D. hospes and S. bovis, respectively. Fifteen cattle had co-infection with three species of trematode: 15/130 (11.5 %) with F. gigantica/ paramphistomes/ D. hospes and 1 (0.8 %) with F. gigantica/ paramphistomes/ S. bovis, but none were infected with all four species. Finally, one animal had a co-infection with paramphistomes and D. hospes 1 (0.8 %) but not with F. gigantica.Fig. 2 Proportions of single infections and co-infections by trematodes in cattle from the Edu Local Government Area of Kwara State, Nigeria Infections with F. gigantica were found in cattle in all villages at prevalences ranging from 3.6 to 100 %, with seven out of the eleven villages having prevalence greater than 70 % (Table 1). The highest prevalence was recorded in Belle (100 %), Yelwa (91.3 %), Fedudangi (85.7 %) and Ndachewoye (85.2 %) whereas the lowest prevalence was recorded in Tshonga farm (3.6 %). Paramphistome infections were also found in all eleven villages, and D. hospes infections in all but one village (Mokwagi). The highest prevalences for paramphistomes and D. hospes were 58.3 % (Belle village) and 24.0 % (Ndabata), respectively. Three out of the eleven villages studied (Fanagun, Gonandogo and Ndachewoye) were positive for infection with S. bovis with prevalence rates of less than 2 %. Distribution maps of the prevalences of the various trematode species are shown in Fig. 3.Table 1 Households and cattle trematode infections from villages studied in Edu LGA, Kwara State, Nigeria Households Cattle Village Altitude (m) F. gigantica Paramphistomes D. hospes S. bovis F. gigantica Paramphistomes D. hospes S. bovis N n N N N N n (%) n (%) n (%) n (%) Bacita 106 2 2 2 2 0 22 7 (31.8) 6 (27.3) 3 (13.6) 0 (0.0) Belle 76 5 5 5 1 0 48 48 (100) 28 (58.3) 6 (12.5) 0 (0.0) Bokungi 221 3 3 3 2 0 23 12 (52.2) 2 (8.7) 2 (8.7) 0 (0.0) Fanagun 82 14 14 7 5 3 158 122 (77.2) 12 (7.6) 9 (5.7) 3 (1.9) Fedudangi 193 3 3 1 0 0 14 12 (85.7) 2 (14.3) 1 (7.1) 0 (0.0) Gonandogo 84 10 10 4 4 2 136 100 (73.5) 2 (1.5) 3 (2.2) 2 (1.5) Mokwagi 114 1 1 1 0 0 13 5 (38.5) 1 (7.7) 0 (0.0) 0 (0.0) Ndabata 202 2 2 1 2 0 25 19 (76.0) 5 (20.0) 6 (24.0) 0 (0.0) Ndachewoye 98 21 21 17 9 2 196 167 (85.2) 45 (23.0) 13 (6.6) 3 (1.5) Tshonga Farm 175 2 1 1 2 0 28 1 (3.6) 1 (3.6) 3 (10.7) 0 (0.0) Yelwa 77 2 2 2 2 0 23 21 (91.3) 6 (26.1) 4 (17.4) 0 (0.0) Total 65 64 44 29 7 686 514 (74.9) 110 (16.0) 50 (7.3) 8 (1.2) Abbreviations: N number of samples, n number of infected, (%) prevalence Fig. 3 Distribution of trematode prevalence across villages sampled in Edu Local Government Area of Kwara State, North-central Nigeria. The size of the circles is proportional to prevalence: small circles (0 %); largest circle (100 %) Faecal egg counts ranged from 0–73 (mean 5.92) eggs per gram (epg) for F. gigantica, 0–10 (mean 0.44) epg for paramphistomes, 0–7 (mean 0.15) epg for D. hospes and 0–4 epg for S. bovis. Fasciola gigantica had higher prevalence in adult cattle (77.3 %) than in those younger than two years (62.5 %) (χ2 = 0.002, P < 0.05). The differences in prevalence between age groups were statistically significant for F. gigantica (χ2 = 0.002, P < 0.05). The results of the univariate analyses are presented in Additional file 2: Table S2). The prevalence of S. bovis infection was too low to carry out further meaningful statistical analyses. Correlations between trematode infections of cattle A significant positive correlation was obtained between log-transformed data for faecal egg counts for F. gigantica and paramphistomes (R2 = 0.023, P = 0.0001) but there were no significant correlations between faecal egg counts for F. gigantica and D. hospes (R2 = 0.00003, P = 0.906) or between faecal egg counts for paramphistomes and D. hospes (R2 = 0.004, P = 0.088). Haematological data A total of 217 whole blood samples were analysed. The packed cell volume (PCV) varied within a range of 12–48 % with a mean value of 34.9 % (SD ± 7.25 %; normal range in cattle 24–46 %) [15]. Twenty-two (10.1 %) of these cattle had PCVs less than the lower normal limit of 24 %. The frequency distribution (Table 2) of FAMACHA® score revealed that the majority 58.9 % (128/217) of the cattle studied belonged to FS2 with the least being FS4. None of the cattle studied had a FS score 5.Table 2 Mean packed cell volume (PCV) of cattle in different categories of FAMACHA© score FAMACHA© score Number of cattle Mean PCV Range PCV N (% of total) 1 56 (25.8) 35.9 22–48 2 128 (58.9) 34.9 20–48 3 23 (24.4) 30.5 20–46 4 10 (4.6) 27.7 12–41 Total number of cattle: 217 The sensitivity and specificity of the FAMACHA© ocular score in cattle were 18.2 % (4/22) and 96.9 % (189/195), respectively using PCV ≤ 24 % and FS of 4–5 as cut-off values for anaemia (Table 3).Table 3 Two-by-two contingency table of PCV and FAMACHA© score of cattle FAMACHA© Score FS 4–5 (Anaemic) FS 1–3 (Non-anaemic) Total PCV ≤ 24 % (Anaemic) 4 18 22 > 24 % (Non-anaemic) 6 189 195 Total 10 207 217 Sensitivity: 18.2 %; Specificity: 96.9 % Correlation between trematode infections and PCV in cattle The correlation analysis between log-transformed trematode faecal egg counts and PCV revealed a weak negative but statistically non-significant correlation between PCV and F. gigantica (R = -0.050, R2 = 0.0023, P = 0.463) as well as between PCV and D. hospes (R = -0.070, R2 = 0.0049, P = 0.305). Herd level risk factors of trematode infections Cattle herd sizes greater than 100 were less likely to have F. gigantica infections than those of smaller herds (odds ratio OR: 0.28; 95 % confidence interval CI: 0.14–0.58). Adult cattle (≥ 2 years), were more likely to be infected (OR: 1.94; CI: 1.19–3.16) than younger cattle. Cattle belonging to heads of households aged between 40–59 years were more likely to be infected with paramphistomes (OR: 1.95; CI: 1.02–3.74) than respondents of other age groups (20–39 and > 60 years), while those belonging to the ‘Zabaruma’ ethnic group were less likely infected (OR: 0.05; CI: 0.01–0.22) than the Fulani or Zimbabwean farmers. Cattle herd size greater than 100 were more likely to be infected with D. hospes than those in smaller herds (OR: 6.98; CI: 2.94–16.6). Logistic regression could not be performed on S. bovis infection because there were too few infected animals (Table 4).Table 4 Binary logistic regression to investigate risk factors for cattle trematode infections in Kwara State, Nigeria Parasite Risk factor Odds ratio 95 % CI P F. gigantica Cattle herd size > 100 0.28 0.14–0.58 0.001 Adult cattle (≥ 2 years) 1.94 1.19–3.16 0.008 Paramphistomes Ethnicity of head of respondents 0.05 0.01–0.22 0.001 Head of respondents (40–59 years) 1.95 1.02–3.74 0.043 D. hospes Cattle herd size > 100 6.98 2.94–16.6 0.001 Discussion This is the first reported study to determine the burden of trematode infections in live cattle in Kwara state, North-central, Nigeria. Zebu cattle were the most predominant breeds because they are the predominant breed in Nigeria kept primarily for milk and beef [34]. The few exotic breeds sampled were from the government owned intensive farm in Tshonga district. There were more female (53.5 %) animals sampled than males (46.5 %) because more female animals, are usually kept by farmers for herd growth and milk production [35]. On the basis of coprological examination, trematode infections were common in cattle in Edu LGA, occurring in 78.1 % of animals investigated. The most frequent species of trematode identified was Fasciola gigantica in 74.9 % of cattle. This is considerably higher than the 22.5 % prevalence reported in an abattoir study in the state capital [36] but similar to a study carried out in Bauchi State of northern Nigeria reporting a prevalence of 76.9 % [37]. Egg counts for F. gigantica observed in our study (range 0–73 epg) were considerably lower than reported in previous studies in zebu cattle, for example, a range of 400–1,100 epg in faeces of cattle in Bangladesh [38], a mean of 81.2 egg per 2 g of faeces in Ethiopia [39] and range of 0–167 epg in Tanzania [6]. The values from the present study were however higher than those reported in Ghana [40] and in a previous abattoir study in Nigeria [41]. Further experimental studies are recommended to determine the egg output index of adult F. gigantica in Nigeria and Africa. Fasciolosis causes losses in livestock due to losses from mortality and reduced productivity and is one of the leading causes of liver condemnation in abattoir in Kwara state [36]. Fasciolosis is also an important zoonotic disease [42]. Ndachewoye houses a large abandoned dam that was used by the Bacita sugar factory. It serves as a watering point for both cattle and humans. It is an all year round dam that could favour the life cycle of F. gigantica and other trematode species. Transmission studies on Fasciola spp. revealed that some free-floating metacercariae might be suspended in water where they can be ingested by the definitive hosts [43]. Moreover, human fasciolosis has been reported to occur from eating uncooked watercress derived from endemic areas where infected cattle range freely and probably from contaminated water [44, 45]. The other trematodes prevalent in this study were D. hospes, S. bovis and paramphistomes. The prevalence of paramphistomes recorded here were lower than in previous reports from southern Ghana [40], Zambia [46], Tanzania [6] but higher than reported in Turkey [47]. Paramphistome species has been reported in Nigeria with up to 2,000 adult parasite in cattle [48]. Only a few studies are available on D. hospes infection in Nigeria the majority representing abattoir surveys [7, 8]. Although the clinical disease when present can lead to severe anaemia, oedema and emaciation, it is usually asymptomatic [49]. A prepatent period of up to 59 days post experimental infection with Dicrocoelium dendriticum has been reported in lambs, hence the low faecal egg count (range 0–7 epg) reported in this study may not fully reflect the level of infection [50]. The prevalence (7.3 %) reported in this study is considerably lower than the 38 % previously reported in cattle at slaughter in Zaria, Nigeria [8]. Infections with S. bovis was reported in only three of the villages studied (Gonandogo, Ndachewoye and Fanagun) with low prevalences (< 2 %) similar to those observed in a recent study in Tanzania [6]. However, another study in Zambia, reported up to 22 % prevalence for bovine schistosomiasis [51]. There was a significantly higher prevalence of F. gigantica infections in adult cattle (77.3 %) than weaners (62.5 %). This is in agreement with a study in Tanzania with similar findings [6] and may be due to longer exposure of adult animals to infection [52]. Although some past studies revealed higher trematode prevalence rates in female than in male animals [40, 53], no significant difference was found in this study. This finding agrees with a previous study in Uganda that reported no significant difference in F. gigantica prevalence between sexes [27]. Out of the 686 faecal samples analysed, 130 (19.0 %) were from cattle with trematode co-infections, the majority of which were F. gigantica co-infecting with paramphistomes (64.6 %). This may reflect the similarity of the life-cycles of these parasites, which require lymnaeid snails as intermediate hosts [54, 55]. The metacercariae of both trematodes could be found in the same places and might be ingested together by the cattle [56]. This finding was also supported by a significant positive correlation between F. gigantica and paramphistomes faecal egg counts. This positive relationship between the two parasites has been reported previously [46, 51, 57]. Although the haematological results obtained from this study indicated the mean PCV of 34.5 % was within normal range for cattle, which is 24–46 % [32], anaemia due to fasciolosis has been attributed to blood-sucking by adult parasites over long period of time in chronic infections [38, 58]. Failure to demonstrate any effect of trematode infection on PCV may have been due to improved nutritional status of the cattle during the rainy season when they were sampled. The nutritional status of cattle has been previously reported to influence the effect of liver fluke disease [59]. The FAMCHA© chart is a low cost tool for determining anaemia status in ruminants [60, 61] and can be used as part of an integrated worm control program [61]. Selective treatment of animals based on anaemic status is important in preventing anthelmintic resistance [29]. The usefulness of the FAMACHA© chart been evaluated previously in small ruminants in Nigeria [62, 63]. Using scores of 4 or 5 as being indicative of anaemia, the FAMACHA© chart gave a low sensitivity (18.2 %) compared with PCV of ≤ 24 % as a gold standard indicator for anaemia. The PCV cut-off value used to represent anaemia has a significant impact on sensitivity and specificity; values of 64.1 % and 91.3 %, respectively, have been reported in sheep using FAMACHA© score 4 or 5 and PCV ≤ 19 % [30], but there were too few cattle in our study with low enough PCVs for comparable analysis. The FAMACHA© was designed specifically for use in sheep with haemonchosis, optimising the chart for use in cattle based on breeds and variation in colour of ocular mucous membrane might be rewarding. Analysis of risk factors for trematode infections indicated that cattle in large herds were significantly less likely to be infected with F. gigantica than those from smaller herds. This is because of the possibility of better herd management such as routine anthelmintic treatment. The risk of F. gigantica infection has been previously shown to be lower in medium and large sized non-dairy Danish cattle herds managed intensively [64]. This was however not the case with D. hospes infection where large cattle herds were seven times more likely to be infected than small herds. Other factors such as improper dosing of drugs, anthelmintic resistance or other management system factors may therefore predispose to D. hospes infection. The age and ethnicity of respondents proved to be important in predicting paramphistome infection, cattle belonging to respondents aged 40–59 years of age were 1.95 times more likely to have paramphistomes than cattle belonging to respondents of other age groups, and cattle of responents of Zabaruma ethnic groups were less likely to have paramphistomosis. These ethnic groups keep their cattle semi-intensively, and therefore are less likely to come in contact with infected pasture. Conclusions This study revealed the presence of infections with F. gigantica, D. hospes, S. bovis and paramphistomes in cattle sampled from Edu, Kwara State. There was a high variability in the prevalence of trematodes across villages and this may be important in successful control of the parasites. The high prevalence recorded for F. gigantica suggests that an anthelmintic resistance survey on the currently available drugs in the state would be advisable. The high positive correlation between F. gigantica and paramphistome infections and hence likelihood of co-infections should be considered when carrying out anthelmintic therapy. The drugs of choice should be effective against both parasites. There is also a need to optimise and validate the FAMACHA© chart for use in cattle based on breeds and variation in colour of ocular mucous membrane. This would improve its usefulness in identifying cattle with anaemia as a means of selective treatment and hence reduce anthelmintic use/resistance. Finally the risk factors identified in this survey such as herd size and cattle age could assist in tailoring control strategies for various trematode infections to particular group of farmers and cattle. Abbreviations EDTA, Ethylene diamine tetra acetic acid; epg, eggs per gram; FS, FAMACHA© score; GDP, Gross Domestic Product; GPS, Global positioning system; LGA, Local Government Area; PCV, Packed cell volume; QGIS, Quantum geographic information system; SD, Standard deviation; Sp, Specificity; Ss, Sensitivity; UIN, University investigation number Additional files Additional file 1: Table S1. General characterisitics of households surveyed in Edu LGA, Kwara State. (XLSX 10 kb) Additional file 2: Table S2. Univariate analysis of trematode infections in cattle based on household and cattle data in Edu, Kwara State, Nigeria. (XLSX 16 kb) Acknowledgements We are grateful to the staff of the Directorate of Veterinary services, Kwara State for their corporation in this study. Dr Gimba and Dr Deji Lawal are especially appreciated for their technical assistance. Also Mallam Idris for providing laboratory assistance. All the farmers that took part in the study are also acknowledged for their willingness to participate in the study. We also wish to thank the Universities of Ilorin, Kwara State, Nigeria and Bristol, UK for providing support for the PhD study. Funding PhD tuition was awarded to the corresponding author (NE) by Nigerian government (Tertiary education trust fund) but no specific funding was provided for the study. Availability of data and materials The data supporting the findings of this study are included within the article and its Additional files. Authors’ contributions NE carried the field sampling, analysis and writing of manuscript, AGA supervised the field sampling of cattle, GCC participated in the study design and manuscript correction, MCE participated in the study design, statistical analysis and writing of the manuscript. All authors read and approved the final version 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 study was carried out with ethical approval from the Kwara State Ministry of Agriculture, Directorate of Veterinary Services. Consent was also sought from all of the cattle farmers in the study before sampling. The study design was approved by the University of Bristol ethical review group, with university investigation number UIN/13/020. ==== Refs References 1. Chen MG Mott KE Progress in assessment of morbidity due to Fasciola hepatica infection: a review of recent literature Trop Dis Bull 1990 87 1 38 2. Wolfe MS Spurious Infection with Dicrocoelium hospes in Ghana Am J Trop Med Hyg 1966 15 2 180 182 5948725 3. Phiri AM Phiri IK Sikasunge CS Monrad J Prevalence of fasciolosis in Zambian cattle observed at selected abattoirs with emphasis on age, sex and origin J Vet Med B Infect Dis Vet Public Health 2005 52 9 414 416 10.1111/j.1439-0450.2005.00872.x 16283922 4. Abebe R Abunna F Berhane M Mekuria S Megersa B Regassa A Fasciolosis: Prevalence, financial losses due to liver condemnation and evaluation of a simple sedimentation diagnostic technique in cattle slaughtered at Hawassa Munincipal abattoir Southern Ethiopia Ethiop Vet J 2010 14 1 39 51 5. Ngole IU Ndamukong KJN Mbuh JV Internal parasites and haematological values in cattle slaughtered in Buea subdivision of Cameroon Trop Anim Health Prod 2003 35 409 413 10.1023/A:1025811428008 14620585 6. Nzalawahe J Kassuku A Stothard J Coles G Eisler M Trematode infections in cattle in Arumeru District, Tanzania are associated with irrigation Parasit Vectors 2014 7 1 107 10.1186/1756-3305-7-107 24650420 7. Schillhorn van Veen TW Folaranmi DOB Usman S Ishaya T Incidence of liver fluke infections (Fasciola gigantica and Dicrocoelium hospes ) in ruminants in northern Nigeria Trop Anim Health Prod 1980 12 2 97 104 10.1007/BF02242616 7414705 8. Ulayi BM Umaru-Sule B Adamu S Prevalence of Dicrocoelium hospes and Fasciola gigantica Infections in cattle at Slaughter in Zaria Nigeria J Anim Vet Adv 2007 6 9 1112 1115 9. Nwosu CO Srivastava GC Liver fluke infections in livestock in Borno State Nigeria Vet Quart 1993 15 182 183 10.1080/01652176.1993.9694403 10. Umar AG Nwosu CO Philip HR Seasonal prevalence and economic importance of bovine fascioliasis in Jalingo Abattoir, Taraba State Nigeria Nig Vet J 2009 30 44 50 11. Biu AA Oluwafunmilayo A Identification of some paramphistome infecting sheep in Maiduguri Nigeria Pak Vet J 2004 24 4 187 189 12. Ndifon GT Betterton C Rollinson D Schistosoma curassoni Brumpt, 1931 and S. bovis (Sonsino, 1876) in cattle in northern Nigeria J Helminthol 1988 62 01 33 34 10.1017/S0022149X00011160 3372976 13. Boray JC Experimental Fasciolosis in Australia Adv Parasitol 1969 7 95 210 10.1016/S0065-308X(08)60435-2 4935272 14. Mitchell G Update on fasciolosis in cattle In Pract 2002 24 378 385 10.1136/inpract.24.7.378 15. Radostits OM, Clive C, Douglas CB, Kenneth WH. Textbook of diseases of cattle, sheep, goats, pigs and horses. 9th ed. Book Power formerly ELST with Saunders. 2000;1380-82. 16. Spithill T, Smooker P, Copeman D. Fasciola gigantica: epidemiology, immunology and molecular biology. In: Fasciolosis. Edited by Dalton JP: CABI Publishing. 1999;465–525. 17. Taylor EL. Fasciolasis and the liver fluke., vol. 64: Food and Agriculture Organization of the United Nations Agricultural Studies; 1964. 18. De Bont J Vercruysse J Southgate VR Rollinson D Kaukas A Cattle schistosomiasis in Zambia J Helminthol 1994 68 04 295 299 10.1017/S0022149X00001516 7706674 19. FAOSTAT. Stock head/year: Nigeria [http://faostat.fao.org/site/573/DesktopDefault.aspx?PageID=573-ancor]. 20. FAO Country Forage/Forest Resource ProfileNigeria Edited by Aregheore EM 2009 FAO Rome Food and Agriculture Organization of United Nations 21. CBN. Central Bank of Nigeria, Annual Report; 1999. 22. Bogatko W Massive mixed Paramphistome cervi and F. gigantica infection in cattle herds in Northern Nigeria Med Weter 1975 31 469 470 23. Kwara State Government N: Kwara State Diary; 2012. 24. Thrusfield MV Veterinary Epidemiology 2007 3 Oxford Blackwell 25. Cameron AR Baldock FC Two-stage sampling in surveys to substantiate freedom from disease Prev Vet Med 1998 34 1 19 30 10.1016/S0167-5877(97)00073-1 9541948 26. Faria RN Cury MC Lima WS Concordância entre duas técnicas coproparasitológicas para diagnóstico de Fasciola hepatica em bovinos Arq Bras Med Vet Zootec 2008 60 1023 1025 27. Howell A, Mugisha L, Davies J, LaCourse EJ, Claridge J, Williams DJL, et al. Bovine fasciolosis at increasing altitudes: Parasitological and malacological sampling on the slopes of Mount Elgon. Uganda Parasit Vectors. 2012;5:196. 28. Soulsby EJL Helminths, arthropods and protozoa of domesticated animals 1982 London Baillière Tindall 29. Van Wyk JA Bath GF The FAMACHA© system for managing haemonchosis in sheep and goats by clinically identifying individual animals for treatment Vet Res 2002 33 5 509 529 10.1051/vetres:2002036 12387487 30. Kaplan RM, Burke JM, Terrill TH, Miller JE, Getz WR, Mobini S, et al. Validation of the FAMACHA© eye color chart for detecting clinical anemia in sheep and goats on farms in the southern United States. Vet Parasitol. 2004;123(1–2):105–20. 31. McInroy RA A micro-haematocrit for determining the packed cell volume and haemoglobin concentration on capillary blood J Clin Pathol 1954 7 1 32 36 10.1136/jcp.7.1.32 13143100 32. Radostits O, Clive C, Douglas C, Kenneth W. Textbook of diseases of cattle, sheep, goats, pigs and horses, 9th ed: Book Power formerly ELST with Saunders; 2000:1877. 33. Field A. Discovering statistics using SPSS: SAGE Publications; 2007:856. 34. Blench R. Traditional livestock breeds: Geographical distribution and dynamics in relation to the ecology of West Africa., vol. Working paper 122: Overseas development institute; 1999:1–66. 35. Majekodunmi A Fajinmi A Dongkum C Shaw A Welburn S Pastoral livelihoods of the Fulani on the Jos Plateau of Nigeria Pastoralism 2014 4 1 20 10.1186/s13570-014-0020-7 36. Adewole SO. Prevalence and pathology of Fasciola species in slaughtered cattle. J Life Sci. 2010;4(4(Serial No. 29):28–31 37. Sugun SY, Ehizibolo DO, Ogo NI, Timothy SY, Ngulukun SS. Prevalence of bovine fasciolosis in Bauchi State, Nigeria. Sahel J Vet Sci. 2010;9(2):16-20. 38. Howlader MMR Begum S Islam K Hai MA Hossain MG Further observations on the packed cell volume and haemoglobin concentration in cattle naturally infected with Fasciola gigantica Bangladesh J Vet Med 2004 2 No 2 39. Yilma JM Mesfin A Dry season bovine fasciolosis in Northwestern part of Ethiopia Rev Med Vet (Toulouse) 2000 151 6 493 500 40. Squire SA, Amafu-Dey H, Beyuo J. Epidemiology of gastrointestinal parasites of cattle from selected locations in Southern Ghana. Livest Res for Rural Dev. 2013;25(117). www.lrrd.org/lrrd25/7/squi25117.htm. 41. Schilhorn van Veen TW Ovine fascioliasis (Fasciola gigantica ) on the Ahmadu Bello University farm Trop Anim Health Prod 1979 11 3 151 156 10.1007/BF02237791 505587 42. Mas-Coma S Bargues M Valero M Fascioliasis and other plant-borne trematode zoonoses Int J Parasitol 2005 35 11–12 1255 1278 10.1016/j.ijpara.2005.07.010 16150452 43. Dreyfuss G Rondelaud D Fasciola gigantica and F hepatica : a comparative study of some characteristics of Fasciola infection in Lymnaea truncatula infected by either of the two trematodes Vet Res 1997 28 2 123 130 9112734 44. Stemmermann GN Human infestation with Fasciola gigantica Am J Pathol 1953 29 4 731 759 13065418 45. Toledo R Esteban J Fried B Current status of food-borne trematode infections Eur J Clin Microbiol Infect Dis 2012 31 1705 1718 10.1007/s10096-011-1515-4 22228314 46. Phiri AM Phiri IK Monrad J Prevalence of amphistomiasis and its association with Fasciola gigantica infections in Zambian cattle from communal grazing areas J Helminthol 2006 80 01 65 68 10.1079/JOH2005313 16469175 47. Ozdal N Gul A Ilhan F Deger S Prevalence of Paramphistomum infection in cattle and sheep in Van Province, Turkey Helminthologia 2010 47 1 20 24 10.2478/s11687-010-0003-1 48. Dube SO Obiamiwe BA Aisein MSO Descriptive studies of the Genus Paramphistomum Fischoeder , 1901 in some Nigerian cattle Discov and Innov 2005 17 3&4 186 192 49. Otranto D Traversa D Dicrocoeliosis of ruminants: a little known fluke disease Trends Parasitol 2003 19 1 12 15 10.1016/S1471-4922(02)00009-0 12488216 50. Campo R Manga-González MY González-Lanza C Relationship between egg output and parasitic burden in lambs experimentally infected with different doses of Dicrocoelium dendriticum (Digenea) Vet Parasitol 2000 87 2–3 139 149 10.1016/S0304-4017(99)00165-X 10622605 51. Yabe J Phiri IK Phiri AM Chembensofu M Dorny P Vercruysse J Concurrent infections of Fasciola , Schistosoma and Amphistomum spp. in cattle from Kafue and Zambezi river basins of Zambia J Helminthol 2008 82 373 376 10.1017/S0022149X08054904 18854056 52. Keyyu JD Monrad J Kyvsgaard NC Kassuku AA Epidemiology of Fasciola gigantica and amphistomes in cattle on traditional, small scale dairy and large scale dairy farms in the southern highlands of Tanzania Trop Anim Health Prod 2005 37 303 314 10.1007/s11250-005-5688-7 15934638 53. Adedokun OA Ayinmode AB Fagbemi BO Seasonal prevalence of Fasciola gigantica infection among the sexes in Nigerian cattle Vet Res 2008 2 1 12 14 54. FAO The Epidemiology of helminth parasites 1993 2 1 30 55. Keyyu JD, Kassuku AA, Msalilwa LP, Monrad J, Kyvsgaard NC. Cross-sectional prevalence of helminth infections in cattle on traditional, small-scale and large-scale dairy farms in Iringa district, Tanzania. Vet Res Commun. 2006;30(1):45-55. 56. Szmidt-Adjidé V Abrous M Adjidé CC Dreyfuss G Lecompte A Cabaret J Rondelaud D Prevalence of Paramphistomum daubneyi infection in cattle in central France Vet Parasitol 2000 87 2–3 133 138 10.1016/S0304-4017(99)00168-5 10622604 57. Nzalawahe J Kassuku AA Stothard JR Coles GC Eisler MC Associations between trematode infections in cattle and freshwater snails in highland and lowland areas of Iringa Rural District, Tanzania Parasitology 2015 FirstView 1 10 58. Sinclair KB Observation on the clinical pathology of ovine fasciolosis Br Vet J 1962 118 37 53 59. Hammond JA Sewell MM The pathogenic effect of experimental infections with Fasciola gigantica in cattle Bri Vet J 1974 130 5 453 465 60. Grace D Himstedt H Sidibe I Randolph T Clausen PH Comparing FAMACHA© eye color chart and Hemoglobin Color Scale tests for detecting anemia and improving treatment of bovine trypanosomosis in West Africa Vet Parasitol 2007 147 1–2 26 39 10.1016/j.vetpar.2007.03.022 17498880 61. Reynecke DP van Wyk JA Gummow B Dorny P Boomker J Validation of the FAMACHA© eye colour chart using sensitivity/specificity analysis on two South African sheep farms Vet Parasitol 2011 177 3–4 203 211 10.1016/j.vetpar.2009.08.023 19775816 62. Glaji YA Mani AU Bukar MM Igbokwe IO Reliability of FAMACHA© chart for the evaluation of anaemia in goats in and around Maiduguri Sokoto J Vet Sci 2014 12 3 9 14 10.4314/sokjvs.v12i3.2 63. Idika IK Iheagwam CN Nwobi LG Nwosu CO Evaluation of anaemia in Nigerian goats using FAMACHA© eye colour chart: a preliminary study Comp Clin Path 2013 22 4 627 630 10.1007/s00580-012-1456-z 64. Olsen A Frankena K Bodker R Toft N Thamsborg S Enemark H Halasa T Prevalence, risk factors and spatial analysis of liver fluke infections in Danish cattle herds Parasit Vectors 2015 8 1 160 10.1186/s13071-015-0773-x 25888827

PMC005xxxxxx/PMC5002139.txt

==== Front Diabetol Metab SyndrDiabetol Metab SyndrDiabetology & Metabolic Syndrome1758-5996BioMed Central London 17210.1186/s13098-016-0172-2ResearchEarly diabetes screening in women with previous gestational diabetes: a new insight Nabuco Aline +55 21 39382730Aline-alinenabuco@yahoo.com.br 13Pimentel Samara Samara-ps.samara@gmail.com 1Cabizuca Carolina A. Carolina-carolcabizuca@yahoo.com.br 1Rodacki Melanie Melanie-melanierodacki@gmail.com 1Finamore Denise Denise-denisefinamore@yahoo.com.br 2Oliveira Marcus M. Marcus-marmiranda@uol.com.br 1Zajdenverg Lenita Lenita-lenitazaj@gmail.com 11 Nutrology and Diabetes Section/Maternidade Escola, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Rio de Janeiro CEP 21941-913 Brazil 2 Maternidade Escola, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil 3 Serviço de Nutrologia e Diabetes, Hospital Universitário Clementino Fraga Filho, Rua Professor Rodolpho Paulo Rocco 255, sala 9E14, University City, CEP 21941-913 Brazil 27 8 2016 27 8 2016 2016 8 1 6129 12 2015 18 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 Gestational diabetes mellitus (GDM) is a risk factor for the development of diabetes mellitus (DM). However, there is a low return rate for this screening, so it is important to search for earlier methods for evaluation after delivery, to increase the number of pregnant women screened, so you can start the treatment or prevention of these early comorbidities. To determine the accuracy of the 75 g 2-h oral glucose tolerance test (OGTT) performed between 48–72 h after delivery for the diagnosis of DM using the OGTT after 6 weeks as the gold standard criterion, and to identify the optimal cutoff points for this exam for diabetes screening after a pregnancy complicated by GDM. Methods 82 women with previous GDM underwent an OGTT between 48–72 h postpartum and repeated the test 6 weeks after delivery. Results The prevalence of DM and prediabetes based on the first OGTT was 3.7 and 32.9 %, respectively, and 8.5 and 20.7 %, respectively, at the second OGTT. For those with DM, the area under the curve (AUC) based on the fasting plasma glucose (FPG) was 0.77 (95 % CI 0.61–0.92), and based on 2-h OGTT was 0.82 (95 % CI 0.66–0.97). For patients with prediabetes, the AUC based on the FPG was 0.73 (95 % CI 0.59–0.86) and based on the 2-h OGTT was 0.74 (95 % CI 0.61–0.87). Using a FPG cutoff value of 78 mg/dl (4.3 mmol/L) and a 2-h OGTT cutoff value of 130 mg/dl (7.2 mmol/L) for DM, the specificity was 58.7 and 60 %, the sensitivity was 71.4 and 85.7 %, the positive predictive value was 13.9 and 16.7 and the negative predictive value was 95.7 and 97.9 %, respectively. Conclusions OGTT performed early in postpartum is a useful tool for identifying women with previous GDM who must perform an OGTT 6 weeks after delivery. Keywords Gestational diabetes mellitusPostpartum screeningPuerperiumMaternidade Escola, Federal University of Rio de Janeiro.issue-copyright-statement© The Author(s) 2016 ==== Body Background The number of patients with diabetes mellitus (DM) has increased significantly in recent decades. Despite better awareness and developments in treatment and prevention of type 2 diabetes, one in two adults with diabetes is undiagnosed [1]. The increasing prevalence of overweight and obesity in both developed and developing countries are the main factors for this rise [2]. Likewise, a growing number of cases of gestational diabetes mellitus (GDM) have been described in the last decades [3]. The frequency of GDM varies between 3–14 % depending on the method used for diagnosis and the study population [4–7]. The magnitude of the risk of postpartum diabetes depends on the ethnicity, the duration of follow up and the specific criteria for GDM diagnosis. Studies have shown that 3–65 % of women with previous GDM develop type 2 diabetes within 5–16 years after the index pregnancy [8–13]. When screened 6–12 weeks postpartum, up to 10 % of women who had GDM were diagnosed with diabetes and an additional 12–36 % had impaired fasting glucose or impaired glucose tolerance [14, 15]. Postpartum screening aims to identify women that developed or have an elevated risk of developing diabetes after pregnancy. Early recognition is important because lifestyle modifications and medications can reduce the incidence of diabetes in individuals at high risk [16–18]. Additionally, the early treatment of diabetes can prevent or delay microvascular end organ complications and reduce the risk of experiencing complications in subsequent pregnancies [19–23]. Both the American Diabetes Association (ADA) and the World Health Organisation (WHO) recommend postpartum screening after 6–12 weeks, using the 75 g 2-h oral glucose tolerance test (OGTT) [24–26]. The United Kingdom’s National Institute for Health and Clinical Excellence (NICE) recommends the fasting plasma glucose (FPG) test be administered at least 6 weeks after childbirth, instead of the traditional OGTT [27, 28]. The OGTT is more sensitive, with reported sensitivities of 100 % compared with 67 % for the FPG [29]. Previous studies of postpartum diabetes screening in women with GDM-affected pregnancies have noted test completion rates that range from 14–61 % [15, 20, 30, 31]. Alternative diagnostic tools may increase the number of evaluated women with previous GDM. The ADA recommends that women with a history of GDM with a normal postpartum screening might be rescreened every 3 years, and women with impaired fasting glucose (IFG) or impaired glucose tolerance (IGT) or both (prediabetes) should be rescreened annually [18]. The purpose of this study was to determine the accuracy of the 2-h OGTT performed between 48–72 h after delivery for the diagnosis of diabetes and to identify the optimal 2-h OGTT cutoff points for screening dysglycaemia in the early postpartum period using the follow-up 2-h OGTT after 6 weeks as the gold standard criteria. Methods Study design In this prospective observational study, women with previous GDM who were recruited from a multi-ethnic population were evaluated. The diagnosis of GDM in pregnant women prior to December 2010 was made according to the Carpenter and Coustan criteria [32]; after January 2011, the diagnosis was made according to the International Association of Diabetes and Pregnancy Study Groups (IADPSG) [33]. The inclusion criteria were pregnant women diagnosed with GDM with regular follow-up in the diabetes and pregnancy outpatient clinic of the maternity school at Rio de Janeiro Federal University. Women who used medications known to affect glucose metabolism, and mothers diagnosed with GDM who were discharged before 48 h after delivery were excluded (8.1 %). All women identified with GDM underwent self-monitoring of blood glucose measurements, as well as dietary management. Insulin treatment was initiated when dietary management did not achieve the glycaemic goal (fasting blood glucose >95 mg/dl, 1 h postprandial blood glucose >140 mg/dl, or 2 h postprandial blood glucose >120 mg/dl). The standard care was to screen all pregnant women with previous GDM with the 2-h OGTT at six weeks after delivery. Data collection included a detailed clinical and obstetric history. Measurements of FPG and a 2-h OGTT were assessed using an enzymatic colorimetric method between 48–72 h and 6 weeks after delivery. Study subjects were instructed to fast overnight for at least 8 h prior to their testing day and to eat at least 150 g of carbohydrate the day prior to testing. The OGTT used a 75 g anhydrous glucose load and followed the standard WHO procedures [34]. The diagnostic categories of normal, prediabetes (i.e., impaired fasting glucose or impaired glucose tolerance) and diabetes were determined from the results of the FPG and 2-h OGTT using the WHO 2006/ADA criteria [35]. Ethical considerations All participants provided written informed consent. The Local Ethics Committee approved this study. Statistical analysis Statistical analyses were performed with SPSS version 20.0. Differences in the classifications between normal, prediabetes and diabetes using the FPG and OGTT were assessed using a non-parametric test (Wilcoxon). Receiver operating characteristic (ROC) curves were developed, and the area under the curve (AUC) with 95 % CIs was determined. The ROC curves were constructed to calculate the sensitivity, specificity, predictive value positive and predictive value negative at different cutoff values. The optimal FPG and 2 h OGTT-screening cutoff points between 48–72 h after delivery were determined by taking the greatest sum of the sensitivity and specificity for the measured FPG and 2-h OGTT values between the two diagnosed groups (diabetes and prediabetes). The positive predictive value was defined as the number of true positives divided by the total number of individuals who tested positive, whereas the negative predictive value referred to the proportion of subjects with a negative test result who were correctly diagnosed. The positive predictive value and negative predictive value were also reported for the optimal cutoff values. Results and discussion During the inclusion period, 257 women with GDM were identified; however, only 82 (31.9 %) patients met all inclusion criteria, had full laboratory data for analysis, and were included in the study. In fact, 21 mothers were discharged before 48 h (8.1 %), 49 patients did only the first OGTT (19 %), and 105 did not return to the second OGTT (40, 85 %). Perinatal features of this study cohort were as follows: the mean age was 32.2 (± 5.8) years, the mean body mass index (BMI) was = 27.7 (±5.3) kg/m2, 54.4 % were non-caucasian, 56.7 % had more than 8 years of education and 68.4 % had a relative with DM. Additionally, the mean parity of the women was 2.3 (±1.3), the mean gestational age at GDM diagnosis was 23.1 weeks (±7.4), 69.4 % had a caesarean delivery and 64.6 % required insulin treatment (Table 1).The mean FPG levels in the early period (48–72 h) after delivery were 76.7 mg/dl (4.26 mmol/L ± 0.66), whereas the mean FPG levels after 6 weeks were 92.6 mg/dl (5.1 mmol/L ± 15.7) (p < 0.0001). All patients were breastfeeding during routine postpartum OGTT. The 2 h post-load glucose was 123.6 mg/dl (6.8 mmol/L ± 2.0) and 110.0 mg/dl (6.1 mmol/L (±2.2) (p = 0.001) between 48–72 h and after 6 weeks, respectively. The prevalence of diabetes and prediabetes based on the 75 g OGTT performed at 48–72 h after birth was 3.7 and 32.9 %, respectively, whereas the prevalence based on the second OGTT was 8.5 and 20.7 %, respectively.Table 1 Patient characteristics Characteristics Mean ± SD or % Age (years) 32.2 ± 5.8 Gestational age at diagnosis (weeks) 23.1 ± 7.4 Parity (n) 2.3 ± 1.3 Weight gain until delivery (kg) 10.2 ± 6.3 Pre-pregnancy BMI (kg/m2) 27.7 ± 5.3 Gestational age at birth (weeks) 37.8 ± 2.9 Women with diagnosis of diabetes on the second 75 g OGTT (n = 7) had mean age of 33.1 (±6.4) years and the mean BMI was 29.8 (±7.2) kg/m2 respectively. In addition, the mean parity was 2.4 (±1.6), and the mean gestational age at GDM diagnosis was 19.4 (±8.7). These patients had a mean weight gain until the GDM diagnosis of 5.1 (±3.8) kg and 8.2 (±2.9) kg until delivery. Eighty-five percent were on using insulin at an average of 27.3 (±2.9) weeks, and seventy-one percent had a family history of DM. The delivery was on average 37.5 (±5.2) weeks and the neonate’s weight was 3108.3 (±957.2) g. There was no statistically significant difference between mothers with and without a diagnosis of diabetes postpartum. Using the 2-h value of 200 mg/dl (11.1 mmol/L) as the cutoff for identifying individuals with diabetes resulted in a sensitivity of 28.6 %, a specificity of 98.7 % and positive and negative predictive values of 66.7 and 93.7 %, respectively. The performance of the early 2-h OGTT was also evaluated using 140 mg/dl (7.7 mmol/L) as the cutoff for identifying individuals with prediabetes, which revealed a sensitivity of 64.7 %, a specificity of 75.4 % and positive and negative predictive values of 40.7 and 89.1 %, respectively. The AUC based on the FPG in the early period for the group with diabetes was 0.77 (95 % CI 0.61–0.92; p = 0.020) and 0.82 (95 % CI 0.66–0.97; p = 0.006) based on the 2-h OGTT (Fig. 1). The AUC based on the FPG in the early period for the group with prediabetes was 0.73 (95 % CI 0.59–0.86 p = 0.004) and 0.74 (95 %CI 0.61–0.87; p = 0.002) based on the 2-h OGTT (Fig. 1). The greatest accuracy for prediabetes was found with the cutoff values of 78 mg/dl (4.3 mmol/L) for FPG and 130 mg/dl (7.2 mmol/L) for the 2-h OGTT (specificity was 63.1 and 64.6 %, sensitivity was 70.6 and 76.5 %, the positive predictive value was 33.3 and 36.1 % and the negative predictive value was 89.1 and 91.3 %, respectively). The greatest accuracy for diabetes was found with the cutoff values of 80 mg/dl (4.4 mmol/L) for the FPG and 150 mg/dl (8.33 mmol/L) for the 2-h OGTT (the specificity was 66.7 and 80 %, the sensitivity was 71.4 % for both, the positive predictive value was 16.7 and 25 % and the negative predictive value was 96.2 and 96.8 %, respectively) (Table 2).Fig. 1 Receiver operating characteristics curve for the FPG and 2 h OGTT used for the detection of diabetes by glucose criteria. a Early FPG for DM. b Early 2-h OGTT for DM. c Early FPG for Pre DM. d Early 2-h OGTT for Pre DM Table 2 Optimal early fasting glucose and 2-h OGTT for the determination of dysglycaemia Receiver operating curve cutoff value (fasting and postprandial glucose) (mg/dl) AUC (95 % CI) Sensitivity (%) Specificity (%) Positive predictive value (%) Negative predictive value (%) Prediabetes 78 0.73 (0.59–0.86) 70.6 63.1 33.3 89.1 Prediabetes 130 0.74 (0.61–0.87) 76.5 64.6 36.1 91.3 Diabetes 130 0.82 (0.66–0.97) 85.7 60.0 16.7 97.8 Diabetes 150 0.82 (0.66–0.97) 71.4 80.0 25 96.8 Diabetes 78 0.77 (0.61–0.91) 71.4 58.7 13.9 95.7 Diabetes 80 0.77 (0.61–0.91) 71.4 66.7 16.7 96.2 The mean age of women not included in the study (n = 175) were 30.5 (±5.8) years and the mean age of pregnancy at GDM diagnosis were 23.9 (±7.4) weeks, with an average BMI of 28.4 (±5.1) kg/m2. In addition, 57.5 % were on insulin and 65.4 % had a family history of diabetes. Comparing both groups (women included and not included in the study), there was no statistically difference in continuous sociodemographic variables. Concerning the categorical variables, the history of previous GDM in the mothers included in the study was significantly higher than in the group of mothers not included (23.4 × 7.9 %, p = 0.008). There was no significant difference in other categorical variables between the two groups (Tables 3, 4).Table 3 Comparison of sociodemographic categorical variables of pregnancy and the newborn among the women included and not included in the study Included (%) Not included (%) P value Age (years) 0.085  10–20 0 5.26  21–30 40.5 43.4  31–40 50.6 48.7  41–50 8.9 2.6 Ethnicity 0.86  Caucasian 45.6 47.1  Non caucasian 54.4 52.9 Gestational age at DMG diagnosis (weeks) 0.56  ≤20 38.0 33.7  ≥21 62.0 66.3 Insulin treatment 0.35  Yes 64.6 57.5  No 35.4 42.5 Previous history of DMG 0.008  Yes 23.4 7.9  No 76.6 92.1 BMI (kg/m2) 0.46  <25 34.7 26.9  25–29.9 31.9 39.7  30–34.9 25.0 19.2  35–40 8.3 12.8  >40 0.0 1.3 Macrosomia 0.17  Yes 14.9 24.5  No 85.1 75.5 Type of delivery 0.14  Vaginal 30.7 18.2  Caesarean 69.4 81.8 DM familiar 0.69  Yes 68.4 65.4  No 31.6 34.6 Table 4 Comparison of sociodemographic continuous variables of pregnancy and the newborn among women included (n = 82) and not included (n = 175) in the study Included Not included P value Mean ± SD Median Mean ± SD Median Age (years) 32.2 ± 5.8 32 30.5 ± 5.8 31 0.075 Gestational age at diagnosis of GDM (weeks) 23.1 ± 7.4 24 23.9 ± 7.4 25 0.47 Start of insulin (weeks) 28.6 ± 7.1 30 26.9 ± 7.2 28 0.30 Parity (n) 2.3 ± 1.3 2 2.3 ± 1.4 2 0.89 Weight gain until GDM diagnosis (kg) 7.0 ± 5.5 6 8.0 ± 5.6 8.8 0.18 Weight gain until delivery (kg) 10.2 ± 6.3 8.2 11.6 ± 6.0 11.5 0.23 Pregestational BMI (kg/m2) 27.7 ± 5.3 27.1 28.4 ± 5.1 27.6 0.41 Birth weight (kg) 3250.3 ± 627.3 3257.5 3409.2 ± 652.0 3400 0.21 Gestational age at delivery (weeks) 37.8 ± 2.9 38 37.8 ± 2.0 38 0.32 We compared the influence of the type of delivery (Vaginal vs Cesarean section) on serum levels of fasting glucose and post-load 75 g, both in collecting 48–72 h as in collecting six weeks, and no statistically significant difference was found. Discussion To our knowledge, this is the first study that evaluate the accuracy and cut off values of OGTT during the early postpartum period in women still hospitalised with previous GDM for determine DM risk. Recently, a study was published and evaluated 58 women with previous GDM who agreed to perform the 75 g OGTT on the second day postpartum. These results were compared with the standard 75 g OGTT 4–12 weeks postpartum. Only 49 of the 98 women presented for routine postpartum OGTT. This study concludes that performing OGTT on the second day is feasible and should be further investigated as an alternative postpartum testing regimen in GDM [36]. In this present study, a 48–72 h OGTT was proven to be a useful tool for identifying women who must perform an OGTT at 6 weeks after delivery. We determined that a FPG of 78 mg/dl (4.3 mmol/L) or a 2-h OGTT of 130 mg/dl (7.2 mmol/L) were the optimal screening cutoff levels for prediabetes, whereas 80 mg/dl (4.4 mmol/L) or 150 mg/dl (8.3 mmol/L), respectively, were the optimal screening cutoff levels for diabetes. Thus, a FPG of 78 mg/dl (4.3 mmol/L) and a 2-h OGTT of 130 mg/dl (7.2 mmol/L) provide the optimal cutoffs to screen for dysglycemia after GDM complicated pregnancies. Additionally, we determined that the 2 h OGTT was more accurate than the FPG. It is important for screening tests to miss fewer individuals with the target disease, and therefore, screening cutoffs are typically lower compared with diagnostic cutoffs. Differences may also result from interference related to the hormonal changes that occur during this postpartum period [37–39]. A special hormonal milieu is observed between 48–72 h after birth. During pregnancy, the levels of oestrogen and progesterone increase greatly primarily because of placental production [40–43]. Following the removal of the placenta, these hormones fall sharply and reach pre-pregnancy levels on the fifth day postpartum [44]. Levels of cortisol, beta- endorphin, and β-HCG also decline during this period [45]. In women who do not breastfeed, prolactin returns to pre-pregnancy levels by 3 weeks after delivery [46]. These hormones, which are counter-regulatory to insulin, contribute to increased insulin resistance in the early postpartum period. Although the OGTT at 48–72 h did not appear to define the diagnosis of diabetes in this study, it was notable that the OGTT at 48–72 h had a high negative predictive value, with a cutoff FPG <78 mg/dl (4.3 mmol/L) and 2-h OGTT <130 mg/dl (7.2 mmol/L), thereby excluding nearly all individuals with diabetes (but not prediabetes). Moreover, diabetes-screening cutoffs that included an early FPG of 78 mg/dl (4.3 mmol/L) and a 2 h OGTT of 130 mg/dl (7.2 mmol/L) effectively identified higher-risk individuals who required a referral for additional evaluation and management. Using these cutoff points, it was found that only 37 % of women should be advised to have their glucose tolerance assessed 6 weeks after delivery. Thus, more than half of women (63 %) should be reevaluated only after 1–3 years per the ADA guidelines on the frequency of testing [18]. Furthermore, identifying women that should return at 6 weeks eases the overall burden to health-care practitioners and reduces the effort required to contact patients and, if necessary, provide a home visit by a health-care worker, which many studies have recommended [47]. The possibility of performing the screening while women are still in the hospital will most likely enhance the efficacy of detecting women at a higher risk for developing DM and will reinforce the need to return for individuals with a greater risk. Hospital sampling could overcome the barriers that currently prevent women from returning to health-care providers for investigation after 6 weeks. The characteristics of patients associated with higher rates of postpartum screening included older age, nulliparity, and higher income or education [15, 31]. Women who received prenatal care, were treated with insulin during pregnancy, or completed a postpartum visit were also more likely to receive a postpartum diabetes screening [15]. In the present study, 31.9 % of the women attended the postpartum follow-up examination, although all patients received reminders upon completion of the first OGTT. This response rate is similar to the lowest follow-up frequency [15, 31]. The most likely explanation for this result is that women in the studied population predominately belonged to low socioeconomic levels and thus, had lower levels of income and education. Sixty-four percent of patients that were included required insulin treatment, and 56 % of patients had more than 8 years of education. The problem of identifying DM after GDM, which may be of greater concern than the choice of test itself, is the poor postpartum screening rate. Although counselling for the management of chronic disease may be challenging in the postpartum period, individuals sometimes express greater interest in their health during times of illness, and opportunities for early diagnosis should not be lost. During a brief discussion, patients with elevated FPG and/or 2 h OGTT could be encouraged to partner with a provider and maintain long-term care, as well as to attempt lifestyle modifications. The concept of a “teachable moment” has been demonstrated for the case of smoking cessation, in which patients are more likely to quit smoking after health events, such as pregnancy, hospitalisation, and the diagnosis of cancer [39]. Thus, health events represent opportunities for health care providers to educate patients and encourage behaviour modifications [48–50]. The prevalence of diabetes in women after GDM in our study (8.5 %) is consistent with those previously described (5–14 %) [14, 15, 51–53]. The first OGTT identified more prediabetes than diabetes cases compared with the second OGTT (prediabetes 32.9 vs 20.7 %; DM 3.7 vs 8.5 %). This result was most likely the result of the number of patients included. Larger cohorts are necessary to recommend the OGTT at 48–72 h as a test for the early screening of postpartum women with GDM. Differences may also result from interference related to the hormonal changes that occur during this postpartum period [37–39]. A special hormonal milieu is observed between 48–72 h after birth. During pregnancy, the levels of oestrogen and progesterone increase greatly primarily because of placental production [40–43]. Following the removal of the placenta, these hormones fall sharply and reach pre-pregnancy levels on the fifty day postpartum [44]. Levels of cortisol, beta-endorphin, and β-HCG also decline during this period [45]. In women who do not breastfeed, prolactin returns to pre-pregnancy levels by 3 weeks after delivery [46]. These hormones, which are counter-regulatory to insulin, contribute to increased insulin resistance in the early postpartum period. This study had some limitations. Firstly, patients were included since 2008 and were classified according to two different criteria for GDM diagnosis (Carpenter and Coustan and the IADPSG). Ikenoue et al. suggested that the IADPSG-defined GDM of one abnormal OGTT value indicates a less severe glucose intolerance, but may still signal a risk of requiring insulin when a first-degree family history of diabetes exists [54]. Our participants had a high prevalence of a relative with diabetes (68.4 %). Moreover, women who were discharged prior to 48 h were not included in this study. These patients most likely had a milder case of GDM and lowest prevalence of postpartum DM. Additionally, the sample size is small, therefore the 95 % CIs around the AUROC curves are quite wide. Indeed, curtailing the rapidly increasing prevalence of early-onset diabetes is a formidable task for health-care practitioners. Additional efforts are necessary to identify these young women as early as possible because they are one of the best groups for which the implementation of a primary prevention strategy is most effective, not only for themselves but also for their offspring and family. Early postpartum screening for DM effectively identifies higher-risk women with previous GDM who require a referral for additional evaluation and management. Conclusions OGTT performed early in postpartum is a useful tool for identifying women with previous GDM who must perform an OGTT 6 weeks after delivery. A diabetes-screening cutoff of FPG of 78 mg/dl (4.3 mmol/L) and a 2-h OGTT of 130 mg/dl (7.2 mmol/L) effectively identified higher-risk individuals who require a referral for additional evaluation and should be further assessed in larger prospective studies. Abbreviations ADAAmerican Diabetes Association AUCarea under the curve BMIbody mass index DMdiabetes mellitus FPGfasting plasma glucose GDMgestational diabetes mellitus IFGimpaired fasting glucose IGTimpaired glucose tolerance IADPSGInternational Association of Diabetes and Pregnancy Study Groups OGTToral glucose tolerance test ROCreceiver operating characteristic NICEUnited Kingdom’s National Institute for Health and Clinical Excellence WHOWorld Health Organisation Authors’ contributions AN contributed in the, study implementation, analysis and interpretation of data and major contribution to writing drafted the manuscript. MR drafted the manuscript. CAC, DF and MO contributed in the study implementation. LZ contributed in the study design, study implementation, analysis and interpretation of data and writing drafted the manuscript. All authors read and approved the final manuscript. Acknowledgements We would like to thank Tatiana Barcelos for helping with the collection of laboratory tests. Competing interests The authors declare that they have no competing interests. Funding This study was funded by Maternidade Escola, Federal University of Rio de Janeiro. ==== Refs References 1. International Diabetes Federation. IDF Diabetes Atlas, 7th edn. Brussels, Belgium: International Diabetes Federation; 2015. http://www.diabetesatlas.org. 2. Caprio S Development of type 2 diabetes mellitus in the obese adolescent: a growing challenge Endocr Pract 2012 18 791 795 10.4158/EP12142.RA 23047931 3. Ferrara A Increasing prevalence of gestational diabetes mellitus: a public health perspective Diabetes Care 2007 30 Suppl 2 S141 S146 10.2337/dc07-s206 17596462 4. Reichelt AJ Spichler ER Branchtein L Fasting plasma glucose is a useful test for the gestational diabetes Diabetes Care 1998 21 1246 1249 10.2337/diacare.21.8.1246 9702428 5. Bowes SB Hennessy TR Umpleby AM Measurement of glucose metabolism and insulin secretion during normal pregnancy and pregnancy complicated by gestational diabetes Diabetologia 1996 39 976 983 10.1007/BF00403918 8858221 6. O’Sullivan JB Mahan CM Charles D Dandrow RV Screening criteria for high-risk gestational diabetic patients Am J Obstetr Gynecol 1973 116 895 900 7. American Diabetes Association clinical practice recommendations 1996 Diabetes Care 1996 19 Suppl 1 S1 S118 8689936 8. Philipson EH Super DM Gestational diabetes mellitus: does it recur in subsequent pregnancy? Am J Obstet Gynecol 1989 160 1324 1331 10.1016/0002-9378(89)90851-X 2735361 9. Gaudier FL Hauth JC Poist M Corbett DL Cliver SP Recurrence of gestational diabetes mellitus Obstet Gynecol 1992 80 755 758 1407910 10. Coelingh Bennink HJT Recurrence of gestational diabetes Eur J Obstet Gynecol Reprod Biol 1977 7 359 363 10.1016/0028-2243(77)90063-6 264062 11. Grant PT Oats JN Beischer NA The long-term follow-up of women with gestational diabetes Aust N Z J Obstet Gynaecol 1986 26 17 22 10.1111/j.1479-828X.1986.tb01521.x 3460570 12. Moses RG The recurrence rate of gestational diabetes in subsequent pregnancies Diabetes Care 1996 19 1348 1350 10.2337/diacare.19.12.1348 8941462 13. Damm P Gestational diabetes mellitus and subsequent development of overt diabetes mellitus Dan Med Bull 1998 45 495 509 9850811 14. Centers for Disease Control and Prevention. National diabetes fact sheet: national estimates and general information on diabetes and prediabetes in the United States, 2011. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2011. http://www.cdc.gov/diabetes/pubs/pdf/ndfs2011.pdf. 15. Tovar A Chasan-Taber L Eggleston E Oken E Postpartum screening for diabetes among women with a history of gestational diabetes mellitus Prev Chronic Dis 2011 8 A124 22005617 16. Knowler WC Barrett-Connor E Fowler SE Hamman RF Lachin JM Walker EA Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin N Engl J Med 2002 14 393 403 11832527 17. American Diabetes Association Diagnosis and classification of diabetes mellitus Diabetes Care 2007 30 Suppl 1 S42 S47 10.2337/dc07-S042 17192378 18. Metzger BE Buchanan TA Coustan DR de Leia A Dunger DB Summary and recommendations of the Fifth International Workshop Conference on gestational diabetes mellitus Diabetes Care 2007 2 30 Suppl 2 S251 S260 10.2337/dc07-s225 17596481 19. The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group. Retinopathy and nephropathy in patients with type 1 diabetes four years after a trial of intensive therapy N Engl J Med 2000 342 381 389 10.1056/NEJM200002103420603 10666428 20. Ratner R Prevention of type 2 diabetes in women with previous gestational diabetes Diabetes Care 2007 30 S242 S245 10.2337/dc07-s223 17596479 21. Chalmers J Cooper M UKPDS and the legacy effect N Engl J Med 2008 359 1618 1620 10.1056/NEJMe0807625 18843126 22. UK Prospective Diabetes Study Group Intensive blood-glucose control with suphonylureas or insulin compared with conventional treatment and risk of complication in patients with type 2 diabetes (UKPDS33) Lancet 1998 352 837 853 10.1016/S0140-6736(98)07019-6 9742976 23. The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes N Engl J Med 2005 353 2643 2653 10.1056/NEJMoa052187 16371630 24. American Diabetes Association Gestational diabetes mellitus (position statement) Diabetes Care 2004 21 Suppl 1 S88 S90 25. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus Diabetes Care 1997 20 1183 10.2337/diacare.20.7.1183 9203460 26. World Health Organisation Definition, diagnosis and classification of diabetes and its complications. Report of a WHO consultation. Part 1: diagnosis and classification of diabetes mellitus 1999 Geneva WHO 27. Nice guideline. 2015. http://www.nice.org.uk/guidance/ng3. 28. Meltzer S Leiter L Daneman D Gerstein HC Lau D Ludwig S 1998 clinical practice guidelines for the management of diabetes in Canada Can Med Assoc J 1998 159 Suppl S1 S29 9834731 29. Flack JR Payne TJ Ross GP Post-partum glucose tolerance assessment in women diagnosed with gestational diabetes: evidence supporting the need to undertake an oral glucose tolerance test Diabetes Med 2010 27 2 243 244 10.1111/j.1464-5491.2009.02892.x 30. Hunt KJ Conway DL Who returns for postpartum glucose screening following gestational diabetes mellitus? Am J Obstet Gynecol 2008 198 404 406 10.1016/j.ajog.2007.09.015 18241820 31. Clark HD Graham ID Karovitch A Keely EJ Do postal reminders increase postpartum screening of Diabetes mellitus in women with gestational diabetes mellitus? (a randomized controlled trial) Am J Obstet Gynecol 2009 200 634 19268878 32. Carpenter MW Coustan DR Criteria for screening tests for gestational diabetes Am J Obstet Gynecol 1982 144 768 773 10.1016/0002-9378(82)90349-0 7148898 33. Metzger BE Gabbe SG Persson B International Association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy Diabetes Care 2010 33 676 682 10.2337/dc10-0719 20190296 34. World Health Organisation Expert Committee on Diabetes Mellitus Diabetes mellitus 1985 Geneva World Health Organisation 35. World Health Organisation Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia 2006 Geneva WHO 36. Werner EF Has P Tarabulsi G Lee J Satin A Early postpartum glucose testing in women with gestational diabetes mellitus Am J Perinatol 2016 33 10 966 971 10.1055/s-0036-1583193 27120481 37. Xiang A Kjos S Takayanagi M Trigo E Buchanan T Detailed physiological characterization of the development of type 2 diabetes in Hispanic women with prior gestational diabetes mellitus Diabetes 2010 59 2625 2630 10.2337/db10-0521 20682697 38. Handwerger S Freemark M The roles of placental growth hormone and placental lactogen in the regulation of human fetal growth and development J Pediatr Endocrinol Metab 2000 13 343 356 10.1515/JPEM.2000.13.4.343 10776988 39. McIntyre H Chang A Callway L Cowley D Dyer A Radaelli T Hormonal and metabolic factors associated with variations in insulin sensitivity in normal pregnancy Diabetes Care 2010 33 356 360 10.2337/dc09-1196 19880583 40. Barbour LA New concepts in insulin resistance of pregnancy and gestational diabetes: long-term implications for mother and offspring J Obstet Gynaecol 2003 23 545 549 10.1080/0144361031000156500 12963518 41. Carr DB Gabbe S Gestational diabetes: detection, management and implications Clin Diabetes 1998 16 4 11 42. Jovanovic-Peterson L Peterson CM Review of gestational diabetes and low-calorie diet and physical exercise as therapy Diabetes Metab Rev 1996 122 287 308 10.1002/(SICI)1099-0895(199612)12:4<287::AID-DMR171>3.0.CO;2-E 9013073 43. Metsman JH Insulin resistance syndrome and gestational diabetes Endocr Pract 2003 9 Suppl 2 90 92 14698984 44. Godo G Morvay J Sas M Hormonal situation on parturition, early puerperium, and lactation (author’s transl) Zentralbl Gynakol 1979 101 24 1553 1561 398118 45. Mastorakos G Ilias I Maternal and fetal hypothalamic-pituitary-adrenal axes during pregnancy and postpartum Ann N Y Acad Sci 2003 997 136 149 10.1196/annals.1290.016 14644820 46. Ramos-Román MA Prolactin and lactation as modifiers of diabetes risk in gestational diabetes Horm Metab Res 2011 43 9 593 600 10.1055/s-0031-1284353 21823053 47. Gupta Y Gupta A Post-partum screening after gestational diabetes Lancet Diabetes Endocrinol 2013 1 90 91 10.1016/S2213-8587(13)70066-4 24622311 48. McBride CM Emmons KM Lipikus IM Understanding the potential of teachable moments: the case of smoke cessation Health Educ Res 2003 18 156 170 10.1093/her/18.2.156 12729175 49. Silverman RA Thakker U Ellman T Wong I Smith K Ito K Graff K Hemoglobin A1c as a screen for previously undiagnosed prediabetes and diabetes in acute-care setting Diabetes Care 2011 34 1908 1912 10.2337/dc10-0996 21775751 50. Gorin AA Phelan S Hill JO Wing RR Medical triggers are associated with better short- and long- term weight loss outcomes Prev Med 2004 39 612 616 10.1016/j.ypmed.2004.02.026 15313102 51. Kim C Newton KM Knopp RH Gestational diabetes and the incidence of type 2 diabetes: a systematic review Diabetes Care 2002 25 10 1862 1868 10.2337/diacare.25.10.1862 12351492 52. Romero Gutiérrez G Macias Rocha AL Puente Alvarez EI Prevalence of postpartum impaired glucose tolerance after gestational diabetes Ginecol Obstet Mex 2012 80 10 631 636 23240225 53. Anderberg E Landin-Olsson M Kalén J Frid A Ursing D Berntorp K Prevalence of impaired glucose tolerance and diabetes after gestational diabetes mellitus comparing different cut-off criteria for abnormal glucose tolerance during pregnancy Acta Obstet Gynecol Scand 2011 90 11 1252 1258 10.1111/j.1600-0412.2011.01214.x 21679163 54. Ikenoue S Miyakoshi K Saisho Y Clinical impact of women with gestational diabetes mellitus by the new consensus criteria: two-year experience in a single institution in Japan Endocr J 2014 61 353 358 10.1507/endocrj.EJ13-0496 24430729

PMC005xxxxxx/PMC5002140.txt

==== Front BMC GastroenterolBMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 52010.1186/s12876-016-0520-6Research ArticleSoluble ST2 is a sensitive clinical marker of ulcerative colitis evolution Díaz-Jiménez David ddiazj@u.uchile.cl 1De la Fuente Marjorie mkdelafu@gmail.com 12Dubois-Camacho Karen karenduboisc@gmail.com 1Landskron Glauben glauben.landskron@gmail.com 1Fuentes Janitza janitza.f@gmail.com 3Pérez Tamara tamaperez@hotmail.com 3González María Julieta jgonzale@med.uchile.cl 4Simian Daniela dsimian@clc.cl 2Hermoso Marcela A. +56 2 29786572mhermoso@med.uchile.cl 1Quera Rodrigo +56 2 26108048rquera@clc.cl 51 Programa Disciplinario de Inmunología, Instituto de Ciencias Biomédicas, Facultad de Medicina, Universidad de Chile, Santiago, CL 8380453 Chile 2 Subdirección de Investigación, Dirección Académica, Clínica Las Condes, Santiago, CL 7591018 Chile 3 Unidad de Hígado y Gastroenterología, Instituto Chileno-Japonés de Enfermedades Digestivas, Hospital San Borja-Arriarán, Santiago, CL Chile 4 Programa disciplinario de Biología Celular, Instituto de Ciencias Biomédicas, Facultad de Medicina, Universidad de Chile, Santiago, CL 8380453 Chile 5 Servicio de Gastroenterología, Clínica Las Condes, Santiago, CL 7591018 Chile 26 8 2016 26 8 2016 2016 16 1 1034 3 2016 14 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 ST2/IL-33 pathway has been related to ulcerative colitis (UC), and soluble ST2 (sST2), to disease severity. We tested the potential usefulness of sST2 as a predictive marker of treatment response and patients’ outcome. Methods Twenty-six patients with active UC were prospectively recruited and grouped according to an endoscopic score and therapy response. Colonoscopic biopsies were collected at baseline and 6 months or when patients showed clinical activity. The protocol was reinitiated in patients requiring rescue therapy. Blood and stool were collected at baseline, 1, 3, 6 and 12 months. Serum and mucosal ST2, and fecal calprotectin (FC) content were determined by ELISA and correlated to Mayo clinical and endoscopic subscore. Intestinal ST2 was evaluated by immunofluorescence. Wilcoxon signed rank test and Spearman correlations (Rs) were applied (p <0.05). Results Follow-up was completed in 24 patients. sST2 levels (median and range) varied from 173.5 [136.6–274.0] to 86.5 [54.6–133.2] in responders (p < 0.05), and 336.3 [211.0–403.2] to 385.3 pg/mL [283.4–517.3] in non-responders at baseline and 6 months, respectively. sST2 levels correlated with Mayo clinical and endoscopic subscore, mucosal ST2 and FC (Rs = 0.57, 0.66, 0.74 and 0.42, respectively; p < 0.0001) and showed a trend similar to that of FC in responders. Non-responders revealed an increased ST2 content, restricted to the lamina propria’s cellular infiltrate. Conclusions Consecutive sST2 measurement to follow changes in inflammatory activity of UC patients who respond or not to treatment identifies sST2, like FC, as a useful biomarker in predicting clinical outcome of UC patients. Electronic supplementary material The online version of this article (doi:10.1186/s12876-016-0520-6) contains supplementary material, which is available to authorized users. Keywords Soluble ST2Ulcerative colitisBiomarkerFecal calprotectinFondecyt11205773150328De la Fuente Marjorie Hermoso Marcela A. DA-CLCPI200803PI2013-B002Quera Rodrigo CONICYT2015issue-copyright-statement© The Author(s) 2016 ==== Body Background The two main subtypes of inflammatory bowel diseases (IBD), Crohn’s disease (CD) and ulcerative colitis (UC), are characterized by episodes of inflammatory activity and remission. To explain the multifactorial and polygenic nature of IBD, a scenario of chronic and uncontrolled activation of the mucosal immune response in genetically susceptible individuals, exacerbated by environmental factors, has been proposed [1]. Treatment of IBD seeks to induce and maintain remission, as well as reduce inflammation, promote mucosal healing and prevent complications [2], such that the choice of optimal therapy is crucial for patient recovery. In most cases, the disease can be controlled with conventional treatment, such as 5-aminosalicylate (5-ASA) derivatives, corticosteroids and immunosuppressants. However, patients who do not respond to these treatments eventually require biological therapy and, as a final alternative, colectomy. While clinical tools, such as assessment of symptoms/signs, performance of laboratory markers [3], colonoscopy/sigmoidoscopy and imaging modalities may allow monitoring the response to a selected therapy, only endoscopic studies definitively diagnose IBD and evaluate its activity through detection of intestinal lesions [4]; unfortunately this procedure is invasive and costly. Determination of disease activity remains challenging, with most clinical scores correlating poorly with the inflammatory state [4, 5]. Treatment of IBD patients has recently shifted from controlling symptoms to promoting endoscopic mucosal healing or deep remission [6], features that are now major endpoints in clinical trials and are highly efficient in predicting long-term remission and preventing hospitalization and surgery. Treatment promoting mucosal healing can slow the progression of the disease [6]. In this context, laboratory biomarkers have gained importance in evaluating and predicting the response to therapy [7, 8]. Recently, key molecules that regulate mucosal immunity and IBD pathogenesis, including those in the IL-33/ST2 signaling pathway, have emerged as suitable biomarkers in inflammatory conditions such as IBD [9, 10]. In humans, IL1RL1 expression is regulated by distal and proximal promoters that govern expression of the ST2 membrane-anchored receptor (ST2L) and a soluble isoform (sST2) generated by alternative splicing. sST2 is identical to the ST2L extracellular domain and is a decoy receptor for IL33 [11]. The IL-33/ST2 axis has been implicated in numerous other diseases, such as asthma, rheumatoid arthritis, cancer and Alzheimer’s disease [9, 12–14]. We previously proposed the sST2 variant as a potential severity biomarker, mainly in UC patients, that might allow differentiation between active and inactive disease [15]. However, sST2 behavior during the clinical course and response to therapy has not been explored. Here, we examined whether changes in serum ST2 levels can predict response to treatment, clinical activity and disease outcome when compared with other biomarkers, such as FC. Methods Patients In this observational study, we prospectively included patients with relapsed IBD from Clínica Las Condes. The diagnosis of UC was established according to international guidelines including clinical, endoscopic, histologic and radiologic criteria [16, 17]. Patients were assessed at a minimum of 3-month intervals or upon relapse. The follow-up period was 12 months in patients showing no relapse and in patients showing activity flares and requiring rescue therapy (Fig. 1a). All relapses were sufficiently severe to warrant a change in treatment.Fig. 1 Study selection and flow-chart of UC patients’ 1-year follow-up. a Follow-up scheme protocol performed in active UC patients receiving conventional therapy (baseline). Responder UC patients completed the entire follow-up period. Non-responder patients re-started the follow-up with biological therapy and were examined at 1 month, 14 weeks, 6 and 12 months. Clinical subscore, FC and serum sST2 were evaluated at 1, 3, 6 and 12 months. Endoscopic subscore and intestinal ST2 were determined at baseline, 14 weeks (w) and 6 months. b Flow-chart of patients enrolled in the follow-up Patients were excluded from the study based on the following criteria: incomplete ileo-colonoscopy (ileum not intubated), non-classifiable IBD or coexisting cardiopulmonary, renal, hepatic, celiac disease, neurologic, psychiatric, severe allergy and rheumatologic disease, a history of HIV, previous ileostomy or colostomy, and conditions associated with elevated FC levels [18, 19], i.e., non-steroidal anti-inflammatory or anti-coagulant drug use in the 6 months preceding enrollment, a history of erosive/ulcerative upper gastrointestinal disease within 2 months prior to the study or gastrointestinal infection within 4 weeks prior to the study and those patients treated with biological therapy. The source data was encrypted and the data extracted were anonymous. This study was performed in accordance with the Declaration of Helsinki and the protocol was approved by Clinica Las Condes Review Board of the Universidad de Chile, Chile. All patients gave written informed consent. Clinical activity index During clinic visits, general well-being, stool frequency, stool consistency and presence or absence of abdominal pain, tenderness, tenesmus, rectal bleeding, and mucus in stool were recorded. Clinical activity indices were scored according to data available in patient records using the Mayo clinical subscore at baseline and at 1, 3, 6 and 12 months or, in the case of flare, during the follow-up period. Three clinical variables were graded: frequency of evacuation, amount of blood in stool, and the physician’s global assessment. A Mayo clinical subscore of 0 was defined as remission, 1–3 as mildly active disease, 4–6 as moderately active disease, and ≥7 as severely active disease [20]. All clinical decisions were made independently of ST2 protein level measurements, since detections were performed only after follow-up was completed. Endoscopic assessment of disease activity Endoscopic findings were scored according to the Mayo endoscopic subscore, graded as normal (0), mild (1), moderate (2), or severe (3) disease activity. Subscore 0–1 was defined as remission and subscore ≥2 as active disease [21]. Colonoscopic biopsies were collected at baseline (colonoscopy 1) and at 6 months or if patients showed clinical signs of activity (colonoscopy 2). Patients who required rescue therapy reinitiated the protocol. Endoscopic re-evaluation in patients receiving conventional or biological therapy was conducted at 6 months or 14 weeks, respectively, or when flares were evident. Disease activity was assessed based on the results of clinical activity scoring but blinded to quantitative FC values. Clinical response was defined as a decrease from baseline in the total Mayo score by at least three points and at least 30 %, with an accompanying decrease in the subscore for rectal bleeding of at least one point or an absolute subscore for rectal bleeding of 0 or 1. Mucosal healing was defined as an absolute subscore for endoscopy of 0 or 1. Patients who had a clinical response at each time during the follow-up were considered to have a sustained clinical response [22]. Medical treatment Patients enrolled in the study were treated with conventional therapy, such as 5-ASA derivatives, corticosteroids or immunomodulators (azathioprine/6-mercaptopurine), depending on disease severity. The decision to change medical treatment was based on clinical and/or endoscopic evaluation. Physicians were blinded to results of individual quantitative FC values at the time of the patient’s visit. Rescue therapy was defined as: (a) an increased dose of any previously administered medication; or (b) a change in medication to corticosteroids, immunosuppressants (azathioprine/6-mercaptopurine) or anti-TNF (infliximab/adalimumab). Information about the drug, dose and duration of treatment was obtained from the medical history. Measurement of ST2 and IL-33 A blood sample was collected at baseline and at 1, 3, 6 and 12 months of follow-up on the day of colonoscopy or physical examination (Fig. 1a). Serum and intestinal ST2 concentrations, as well as serum IL-33 content, were measured using an enzyme-linked immunosorbent assay (ELISA) kit for human ST2 or IL-33 (DuoSet, R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions and expressed as pg/mL. Blood samples were centrifuged and serum was stored at −80 °C. For ST2 and IL-33 detection, serum samples were thawed and treated with protein A/G PLUS-agarose (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Protein extracts of colonic mucosa were obtained from each biopsy by homogenization using a lysis buffer supplemented with a protease inhibitor cocktail (Complete Mini, Roche Diagnostics, Basel, Switzerland) and subsequently disrupted by sonication. Levels of total intestinal ST2 were adjusted to total protein concentration determined by Bradford protein assay (Bio-Rad). The ST2 detection assay is stable over time, with a detection limit of 20 pg/mL, while the IL-33 detection assay is less stable over time, with a detection limit of 23.40 pg/mL, according to the manufacturer’s information. All samples were analyzed in duplicate; within-run and total coefficients of variation were ≤2.5 and ≤4.0 %, respectively. Using the cut-off for sST2 previously estimated at <74.87 pg/mL [15], we determined the specificity, sensitivity, positive predictive value and negative predictive value for our patient cohort. Fecal calprotectin measurement A fecal sample was collected at baseline and at, 1, 3, 6 and 12 months of follow-up (Fig. 1a) 24 h before colonoscopy or physical examination. Feces (50 mg) was resuspended in 500 μL of extraction buffer, spun for 10 min and centrifuged for 10 min at 10,000 rpm. The supernatant was immediately processed for rapid semi-quantitative test (Calprotectin 50 + 200, CerTest Biotec S.L. Spain) or stored at −80 °C for subsequent quantification using a standard ELISA with a solid bound specific antibody and biotinylated tracer antibody according to the manufacturer’s instructions (Hycult Biotech, HK325-02, The Netherlands). Detection levels were 1.6–100 ng/mL, corresponding to 40–2500 μg calprotectin/g of feces, adjusted for the dilution factor. Clinical decisions to change therapy were based on semi-quantitative analysis of calprotectin and independent of quantitative ELISA after completed follow-up. Endpoints The correlation of ST2 and FC measurements with clinical and endoscopic activity scoring was the primary endpoint of the study. Complete response (remission) was defined as the absence of symptoms, normal stool frequency, absence of rectal bleeding, general wellness based on the patient’s functional assessment score and endoscopy findings of subscore 0–1 [21]. Biological remission also implies healing of mucosa and normalization of inflammation biomarkers, such as serum sST2 (<74.87 pg/mL) [15] and FC, according to the cut-off obtained from our patient cohort. Mucosal ST2 detection by immunofluorescence Intestinal mucosa biopsies were reviewed by a pathologist experienced in IBD who was blinded to any clinical information about the patients. Mucosal tissue sections comprised the most inflamed segment according to chronic inflammatory infiltrate. Biopsies were obtained at baseline and at 6 months or at baseline and 14 weeks in patients treated with conventional treatment and those rescued with biological therapy, respectively. Total ST2 content was detected in 2 % PFA-fixed, paraffin-embedded UC patient biopsies cut into 4-μm sections and subjected to immunofluorescence analysis. Non-specific binding was blocked and samples were incubated with a mouse monoclonal antibody against human ST2 (MAB523, R&D Systems) followed by an Alexa 546-tagged secondary goat antibody against mouse IgG (Invitrogen/Life Technologies, Carlsbad, CA, USA) to detect ST2. Cell nuclei were stained using Hoechst 33342 (Thermo/Life Technologies, Carlsbad, CA, USA). Images were captured using an Olympus confocal laser scanning biological microscope FV10i (Olympus America Inc., Melville, NY, USA) and processed using ImageJ (NIH, Bethesda, MD USA). Negative controls were prepared under conditions identical to those described above using an isotype control. Statistical analyses Data were analyzed using GraphPad Prism5 (La Jolla, CA USA). Results of parametric numerical data are given as mean ± standard deviation (SD) and, in non-parametric distributions, as median and interquartile range (IQR). Serial assessments of change from baseline in serum ST2 and FC levels were compared using the Wilcoxon signed rank test. Differences and significance during follow-up times were analyzed by multiple comparisons using the non-parametric Kruskal-Wallis test with Dunn’s multiple comparison post-test. Correlations between sST2 levels and total intestinal ST2, FC, and clinical and activity score were analyzed using Spearman’s rank correlation coefficient (r). For each test, differences were considered significant at p ≤ 0.05. Results Serum ST2 and clinical characteristics of patients Table 1 summarizes the clinical characteristics of 26 patients included in the protocol, while Fig. 1a shows the flow-chart of the 1-year follow-up enrollment and sampling overview at each stage of the patient’s monitoring.Table 1 Clinical and demographic data of ulcerative colitis patients Number Number of patients 26 Number of patients dropped out 2 Female/male 16/10 Median [range] age at entry 34.5 [30.0–52.5] Duration of UC, years (median, range) 4.0 [1.3–11.8] Age at diagnosis, years (median, range) 32.0 [22.5–44.0] Location  Proctitis 6  Left-sided 5  Extensive 15 Clinical disease activity  Mild/ Moderate/ Severe 0/14/12 Endoscopic severity  Moderate (score 2) / Severe (score 3) 22/4 Medication used  No medication 3  5-ASA 11  Corticosteroids and 5-ASA 3  Corticosteroids, 5-asa and azathioprine 5  Corticosteroids and azathioprine 1  5-ASA and azathioprine 3 Response to therapy 21  Reactivation (after/before 6 months) 6 (1/5)  Rescued by anti-TNF-α 6 Baseline ileo-colonoscopy data were available for 26 UC patients. During the study, 24 patients finished the protocol, six of whom showed disease reactivation, once in four patients after 6 months and twice in two patients before and after 6 months, for a total of eight reactivation episodes in six presumed non-responder patients. Overall, 18 patients responded to treatment and six patients did not (Fig. 1b). Baseline sST2 levels were 189.5 pg/mL [154.1–286.9]; at 1, 3, 6 and 12 months, with levels of 87.3 [70.1–142.2], 73.25 [55.1–145.1], 95.8 [55.7–134.9] and 78.0 [59.9–99.8] pg/mL, respectively (Fig. 2a). When patients were grouped according to treatment response, baseline levels were 173.5 [136.6–274.0], whereas sST2 levels at 6 months decreased to 86.5 [54.6–133.2] pg/mL in responders and increased in the eight reactivation episodes from 336.3 [211.0–403.2] to 385.3 [283.4–517.3] pg/mL in the six non-responders (Fig. 2b). No significant differences between therapy responder and non-responder patients were found in baseline sST2 values (p = 0.1057). Quantitation of FC at baseline, at 1, 3, 6 and 12 months revealed levels of 147.0 [102.1–210.2], 107.1 [54.2–161.3], 119.5 [57.7–144.5], 85.4 [63.0–96.7] and 74.7 [48.9–1076.0] μg/g feces, respectively (Fig. 2c). According to treatment, baseline FC levels of 141.4 [92.9–204.6] decreased to 82.8 [59.2–97.8] μg/g feces in responders and from 243.9 [175.1–297.9] to 152.0 [100.1–200.8] μg/g feces in non-responders at 6 months (Fig. 2d). sST2 levels in all follow-up times correlated with the Mayo clinical subscore (Rs = 0.57, p < 0.0001); at baseline and 6 months, levels correlated with Mayo endoscopic subscores (Rs = 0.66, p < 0.0001) and total ST2 content in mucosa (Rs = 0.74, p < 0.0001) (Fig. 2e). Patients in biological remission showed normalization of the inflammation biomarker sST2, with values below the cut-off of 74.87 pg/mL [15]. Considering this cut-off value of sST2, the specificity, sensitivity, PPV and NPV were 0.44, 0.95, 0.62, and 0.90, respectively.Fig. 2 Distribution of serum ST2 according to response to therapy. a Serum ST2 levels during the follow-up period were determined in each patient and are shown as one symbol in the scatter plot. Horizontal lines indicate medians and whiskers (interquartile ranges). Differences were assessed using Kruskal-Wallis test (*** p < 0.0001). b Serum ST2 levels at baseline and 6 months in responders (R) and non-responders (NR) in relation to therapy decreased only in responders. Differences were assessed using Wilcoxon signed rank test (*** p < 0.0001). c FC levels in the 1-year follow-up and (d) at baseline and 6 months in R and NR in relation to therapy decreased in both subgroups. Panel E shows the correlation between serum ST2 levels and Mayo clinical subscore, Mayo endoscopic subscore, and total ST2 intestinal mucosa content, with a trend line for each correlation. Discontinuous line indicates cut-off value. Rs: Spearman’s rank correlation coefficient When patients were grouped according to treatment response, IL-33 baseline levels were 199.8 [51.92–372.20], decreasing at 6 months to 141.2 [44.00–352.90] pg/mL in responders, and decreasing from 629.0 [49.56–716.8] to 217.9 [50.39–647.5] pg/mL in the reactivation episodes, both without significant differences (p > 0.05) (Additional file 1: Figure S1). Relationship of ST2 to the FC biomarker Analysis showed that sST2-circulating levels directly correlated with those of FC (Rs = 0.42, p < 0.0001) in all patients (Fig. 3a). Intestinal total ST2 levels in mucosa also correlated with those of FC, although the association was less significant statistically (Rs = 0.35, p = 0.044) (Fig. 3b).Fig. 3 Correlation of ST2 and fecal calprotectin levels in patients. Serum ST2 and FC levels were determined at baseline, 1, 3, 6 and 12 months of follow-up. Total ST2 intestinal was measure at baseline and 6 months. Correlation between serum (a) and total intestinal ST2 (b) according to FC levels were assessed. Each symbol in the scatter plot represents the measurement in individual patients, with a trend line for this association. Rs: Spearman’s rank correlation coefficient (p < 0.05) sST2 response to biological therapy In our cohort, six patients were rescued with biological therapy. Immediately after infliximab infusions, sST2 levels were markedly increased, but protein content during the induction period gradually declined between weeks 3 and 24 of monitoring, reaching levels similar to those in patients responding to treatment (p < 0.01, Fig. 4a). A similar trend was observed for FC levels, indicating that fecal biomarker content also decreases gradually as a function of type and response to therapy (p < 0.05; Fig. 4b).Fig. 4 Distribution of serum ST2 and fecal calprotectin in patients rescued with biological therapy. Serum sST2 (a) and fecal calprotectin (b) levels are shown for patients non-responsive to conventional therapy or with reactivation but rescued with biological therapy. Each symbol in the scatter plot represents the measurement in individual patients; horizontal lines indicate medians and whiskers (the interquartile ranges). Differences were assessed using Kruskal-Wallis test with Dunn’s multiple comparison post-test (*p < 0.05 and **p < 0.01) Intestinal total ST2 immunoreactivity Because intestinal ST2 levels were highly correlated with sST2 concentrations (Fig. 2e), we tested whether the total mucosal ST2 content, as the source of ST2 in the periphery, might be related to healing and deep remission in responder patients. Mucosal ST2 immunoreactivity detected by immunofluorescence may allow identification of membrane-anchored ST2 and the soluble variant restricted to the intracellular compartment. In responder patients, total ST2 immunoreactivity was maintained in the cellular infiltrate of the lamina propria at 6 months of follow-up, whereas patients showing reactivation revealed increased total ST2 in inflamed mucosa but also confined to the cellular infiltrate (Fig. 5a,b). No immunoreactivity was detected in samples incubated with isotype antibody. Epithelial ST2 expression was absent in all responder and non-responder patients analyzed.Fig. 5 Immunolocalization of ST2 in colonic tissue from responder and non-responder patients during baseline and 6-month follow-up. a Total ST2 immunoreactivity was restricted to the cellular infiltrate in the lamina propria at 6 months in responder patients (right), while in patients showing reactivation, total ST2 was increased in inflamed mucosa, also confined to the cellular infiltrate (left); baseline examination revealed extensive immune cell infiltration of the intestinal mucosa and damaged tissue with loss of architecture. b Total ST2 immunoreactivity at 6 months expressed as arbitrary units (A.U.) and normalized to baseline levels in responders and non-responders (n = 4 in each group) was higher in the non-responders. Hoechst 33342/ ST2 (blue/green) (60X) Discussion In this prospective study of UC patients with active disease, we consecutively measured sST2 levels during a 1-year follow-up and compared those with FC, the canonically defined biomarker of activity in IBD patients [23–25]. Our results demonstrate for first time that sST2 is an easily detectable marker of UC evolution and effectiveness of therapy. Several studies reporting the association of the IL-33/ST2 signalling pathway with IBD pathogenesis in patients and animal models have attempted to address the actual contribution of each component of the pathway [26, 27]. In this context, the sST2 variant, acting as a decoy receptor for IL-33, is increased in UC [10, 15]. We proposed ST2 as a potential activity biomarker, with a cut-off value of 74.87 pg/mL, that might allow distinction between active and inactive disease with 83 % sensitivity and specificity [15]. At this cut-off value, we found a lower specificity (44 vs 83 %), suggestive of a poorer discrimination of the group of inactive patients or those with sST2 levels <74.87 pg/mL. The higher sST2 levels at baseline and from unresponsive patients/reactivations (endoscopic score ≥ 2) in the vast majority of all detections might explain the lower specificity. sST2 has also been considered a biomarker of diseases that affect different tissues undergoing deep inflammatory reactions and necrosis; a cut-off of 150 pg/mL and 5.8 ng/mL was determined to predict sudden cardiac death in patients with chronic heart failure and left ventricular systolic dysfunction, as well as in patients with rheumatoid arthritis, respectively [28, 29]. In our follow-up study, patients in biological remission showed normalization of sST2, with values close to the cut-off previously determined [15], suggesting that low sST2 levels predict improvement in the inflammatory state. Moreover, we find that increased sST2 levels in UC patients with active disease directly correlate with higher expression of total ST2 in intestinal mucosa, including sST2 and ST2L, consistent with previous observations [15] and suggesting that the increased total ST2 immunoreactivity in non-responder patients is mostly at the expense of newly produced sST2 by cells infiltrating the mucosa. In responder patients, we found no conclusive evidence of a decrease in total ST2 immunoreactivity, possibly reflecting the inability of the technique used to distinguish between the two protein isoforms and its limitation in assessing protein quantity. Moreover, in responder patients, ST2+ cells infiltrating the mucosa might be less effective than those present in non-responders, although additional cell markers are needed to support this hypothesis. The components of this inflammatory microenvironment, including TNF-α, IL-33 and IL-1β produced by infiltrating and resident cells (neutrophils, mast cells and fibroblasts) have been shown to upregulate sST2 expression in vitro [30], although the current lack of a differential tool to detect ST2 variants limits the interpretation of tissue immunodetection of total ST2. On the other hand, studies of intestinal inflammation in murine models have reported that IL-33 itself can play a beneficial or harmful role in IBD depending on the induction of mucosal damage (acute or chronic) or the colitogenic used, arguing against a precise role of the IL-33/ST2 axis. Indeed, IL-33 serum levels did not correlate with the activity score in UC patients [15] and in the present cohort, we found no association between IL33 serum levels with therapy response. The difficulty in evaluating disease activity in UC patients using clinical, endoscopic and histological approaches has led to greater interest in identifying new biochemical markers that directly reflect intestinal inflammation. Recently, the usefulness of FC as a biomarker has been explained by the fact that inflammation causes neutrophil activation, resulting in a proportional release-to-damage ratio. Yamamoto et al. [31] showed that consecutive measurements of FC in UC patients with proctitis were able to monitor relapse during maintenance therapy with mesalazine suppositories. A cut-off value of 55 μg/g was used to assign those patients who maintained clinical remission, although the optimal cut-off for defining activity in IBD remains controversial [32, 33]. Indeed, FC cut-off values ranging from 50 to 250 μg/g have been used [23, 34] that differ depending on the test used [35, 36]. In our study, we used quantitative ELISA and defined a cut-off of 99.26 μg/g with an AUC of 0.87 to distinguish individuals with an inactive endoscopic score (0–1) from those with activity indices (2–3) [37]. The cut-off value obtained for our cohort is similar to that previously described [38]. In 7 of 8 reactivations in non-responders from our cohort, FC levels showed a decrease despite detectable endoscopic inflammation, although FC values were above the cut-off defined in the current analysis. Unexpectedly, FC did not correlate with the endoscopic score, possibly related to the treatment received or presence of liquid stool [39]. Moreover, an inverse association between FC and neutrophil infiltration has been reported in patients under remission, with high protein levels in stool [40]. Finally, daily fluctuations of up to 40 % have been described for FC determined in four independent stool samples from active patients [41], suggesting that a single FC detection in active patients is not sufficient for a therapeutic decision. In contrast, the increased inflammatory state of the mucosa in 5 of 8 reactivations of non-responders simultaneous with augmented sST2 levels indicates that this protein can predict lack of response to the therapy used. The direct association between endoscopic score and sST2 might reflect an inflammatory process related to mucosal damage, while the production and secretion of FC requires massive infiltration of neutrophils into the lamina propria [42, 43]. While no test can replace colonoscopy with biopsy to determine the inflammatory condition of the intestinal mucosa, the ability of sST2 levels to predict inflammatory status identifies the need to incorporate analysis of sST2 in selecting patients who require control for reactivation through colonoscopy. sST2 levels have been also identified as a biomarker in heart failure, efficiently assessing cardiac remodelling and fibrosis; sST2 has been shown to efficiently monitor the effectiveness of an optimized treatment in chronic heart failure patients [44], with increased levels that may be a good predictor of cardiac decompensation as well as worsening renal function. To date, there are few reports on the behavior of sST2 according to the medication used in IBD patients. We found some association of sST2 with the use of systemic steroids [15], and others have reported the effect of infliximab on this molecule [10]. However, no study has demonstrated how ST2 levels fluctuate in serial and continuous follow-up in relation to endoscopic index. Based on our previous finding [15], we recruited patients who at baseline had a moderate or severe endoscopic index (≥2) in any analyzed section of the colon. During the follow-up, serial ST2 measurements decreased in those patients with a reduced endoscopic index at 6 months, indicating a positive response to therapy. In those patients, FC levels were also significantly decreased in direct correlation to sST2 levels. Further association studies between sST2 and FC in a larger cohort of UC patients are needed, as well as such analyses in CD and in IBD pediatric patients. Correlation studies between sST2, histologic scores and disease risk might also provide additional information about the strength of this biomarker as has been demonstrated with FC [45]. Our analyses of a one-bounded cohort of UC patients during a 1-year follow-up revealed a direct correlation between sST2 levels and inflammatory activity, strongly suggesting that sST2 is a reliable biomarker of active UC. It will be interesting to examine the predictive role of sST2 for reactivation risk in a cohort of inactive patients or undergoing clinical and endoscopic remission (Mayo subscore 0) and identifying serum levels of this biomarker in a follow-up every 4 months, a protocol in which FC was shown to predict disease recurrence [31]. Finally, it will be interesting to monitor both the behavior of sST2 in a cohort of severe UC patients and the relevance of elevated sST2 content in predicting colectomy risk in a 1-year follow-up, as shown for FC [46]. Conclusions The accuracy of sST2 in endoscopic detection of UC strongly suggests its usefulness in monitoring relapse and outcome, as well as in identifying patients likely to benefit from a particular treatment. Additional file Additional file 1: Figure S1. Distribution of serum IL33 according to response to therapy. Serum IL33 levels at baseline and 6 months in responders (R) and non-responders (NR) in relation to therapy without significant differences (p > 0.05). Differences were assessed using Wilcoxon signed rank test. (TIFF 25551 kb) Abbreviations 5-ASA5-aminosalicylate derivatives CDCrohn’s disease ELISAEnzyme-linked immunosorbent assay FCFecal calprotectin IBDInflammatory bowel diseases sST2Soluble ST2 UCUlcerative colitis We thank Fabiola Werlinger (Profesor Asistente, Departamento de Tecnología Médica, Facultad de Medicina, Universidad de Chile) and Claudio Molina (Profesor Asociado, Facultad de Odontología, Universidad Mayor, Chile) for their help in statistical analysis. Funding This study was supported by research grants from FONDECYT 1110381 (RQ), 1120577 (MAH), 3150328 (MD) and DA-CLC grants PI200803 (RQ), PI2013-B002 (RQ) and CONICYT scholarship 2015 (DDJ). Availability of data and materials All data generated or analyzed during this study are included in this published article and its supplementary information file. Authors’ contributions DDJ, participated in design of the study, data collection, acquisition, analysis, interpretation and statistical analysis of data, and drafted the manuscript. MDF and GL participated in acquisition, analysis and interpretation of immunofluorescence and confocal images. KD and MJG participated in acquisition, analysis and interpretation of data. DS and JF participated in patient enrollment, obtaining written consent from patients willing to participate in the protocol, collected the data and their interpretation. TP performed colonoscopic procedures and participated in design of the study. MAH participated in design of the study, data collection, interpretation of data, and drafted the manuscript. RQ participated in design of the study, data collection, interpretation of data, drafted the manuscript and performed colonoscopic procedures. 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 was performed in accordance with the Declaration of Helsinki and the protocol was approved by Clinica Las Condes Review Board of the Universidad de Chile, Chile. All patients gave written informed consent. ==== Refs References 1. Kaser A Zeissig S Blumberg RS Inflammatory bowel disease Annu Rev Immunol 2010 28 573 621 10.1146/annurev-immunol-030409-101225 20192811 2. Neurath MF New targets for mucosal healing and therapy in inflammatory bowel diseases Mucosal Immunol 2014 7 6 19 10.1038/mi.2013.73 24084775 3. Vermeire S Van Assche G Rutgeerts P Laboratory markers in IBD: useful, magic, or unnecessary toys? Gut 2006 55 426 31 10.1136/gut.2005.069476 16474109 4. Annese V Daperno M Rutter MD Amiot A Bossuyt P East J Ferrante M Götz M Katsanos KH Kießlich R Ordás I Repici A Rosa B Sebastian S Kucharzik T Eliakim R European evidence based consensus for endoscopy in inflammatory bowel disease J Crohns Colitis 2013 7 982 1018 10.1016/j.crohns.2013.09.016 24184171 5. Hamilton MJ The valuable role of endoscopy in inflammatory bowel disease Diagn Ther Endosc 2012 2012 467979 10.1155/2012/467979 22474403 6. Pineton de Chambrun G Peyrin-Biroulet L Lémann M Colombel J-F Clinical implications of mucosal healing for the management of IBD Nat Rev Gastroenterol Hepatol 2010 7 15 29 10.1038/nrgastro.2009.203 19949430 7. Colombel JF Rutgeerts P Reinisch W Esser D Wang Y Lang Y Marano CW Strauss R Oddens BJ Feagan BG Hanauer SB Lichtenstein GR Present D Sands BE Sandborn WJ Early mucosal healing with infliximab is associated with improved long-term clinical outcomes in ulcerative colitis Gastroenterology 2011 141 1194 201 10.1053/j.gastro.2011.06.054 21723220 8. Colombel J-F Rutgeerts PJ Sandborn WJ Yang M Camez A Pollack PF Thakkar RB Robinson AM Chen N Mulani PM Chao J Adalimumab induces deep remission in patients with Crohn’s disease Clin Gastroenterol Hepatol 2014 12 414 22.e5 10.1016/j.cgh.2013.06.019 23856361 9. Beltrán CJ Núñez LE Díaz-Jiménez D Farfan N Candia E Heine C López F González MJ Quera R Hermoso MA Characterization of the novel ST2/IL-33 system in patients with inflammatory bowel disease Inflamm Bowel Dis 2010 16 1097 107 10.1002/ibd.21175 20014018 10. Pastorelli L Garg RR Hoang SB Spina L Mattioli B Scarpa M Fiocchi C Vecchi M Pizarro TT Epithelial-derived IL-33 and its receptor ST2 are dysregulated in ulcerative colitis and in experimental Th1/Th2 driven enteritis Proc Natl Acad Sci U S A 2010 107 8017 22 10.1073/pnas.0912678107 20385815 11. Hayakawa H Hayakawa M Kume A Tominaga S -i Soluble ST2 Blocks Interleukin-33 Signaling in Allergic Airway Inflammation J Biol Chem 2007 282 26369 80 10.1074/jbc.M704916200 17623648 12. Verri W a Souto FO Vieira SM Almeida SCL Fukada SY Xu D Alves-Filho JC Cunha TM Guerrero ATG Mattos-Guimaraes RB Oliveira FR Teixeira MM Silva JS McInnes IB Ferreira SH Louzada-Junior P Liew FY Cunha FQ IL-33 induces neutrophil migration in rheumatoid arthritis and is a target of anti-TNF therapy Ann Rheum Dis 2010 69 1697 703 10.1136/ard.2009.122655 20472598 13. Cui G Qi H Gundersen MD Yang H Christiansen I Sørbye SW Goll R Florholmen J Dynamics of the IL-33/ST2 network in the progression of human colorectal adenoma to sporadic colorectal cancer Cancer Immunol Immunother 2015 64 181 90 10.1007/s00262-014-1624-x 25324197 14. Chapuis J Hot D Hansmannel F Kerdraon O Ferreira S Hubans C Maurage CA Huot L Bensemain F Laumet G Ayral AM Fievet N Hauw JJ DeKosky ST Lemoine Y Iwatsubo T Wavrant-Devrièze F Dartigues JF Tzourio C Buée L Pasquier F Berr C Mann D Lendon C Alpérovitch A Kamboh MI Amouyel P Lambert JC Transcriptomic and genetic studies identify IL-33 as a candidate gene for Alzheimer’s disease Mol Psychiatry 2009 14 1004 16 10.1038/mp.2009.10 19204726 15. Díaz-Jiménez D Núñez LE Beltrán CJ Candia E Suazo C Alvarez-Lobos M González M-J Hermoso MA Quera R Soluble ST2: a new and promising activity marker in ulcerative colitis World J Gastroenterol 2011 17 2181 90 10.3748/wjg.v17.i17.2181 21633527 16. Dignass A Van Assche G Lindsay JO Lémann M Söderholm J Colombel JF Danese S D’Hoore A Gassull M Gomollón F Hommes DW Michetti P O’Morain C Oresland T Windsor A Stange EF Travis SPL European Crohn’s and Colitis Organisation (ECCO) The second European evidence-based Consensus on the diagnosis and management of Crohn’s disease: Current management J Crohns Colitis 2010 4 28 62 10.1016/j.crohns.2009.12.002 21122489 17. Van Assche G Dignass A Panes J Beaugerie L Karagiannis J Allez M Ochsenkühn T Orchard T Rogler G Louis E Kupcinskas L Mantzaris G Travis S Stange E The second European evidence-based Consensus on the diagnosis and management of Crohn’s disease: Definitions and diagnosis J Crohns Colitis 2010 4 7 27 10.1016/j.crohns.2009.12.003 21122488 18. Limburg PJ Ahlquist DA Sandborn WJ Mahoney DW Devens ME Harrington JJ Zinsmeister AR Fecal calprotectin levels predict colorectal inflammation among patients with chronic diarrhea referred for colonoscopy Am J Gastroenterol 2000 95 2831 7 10.1111/j.1572-0241.2000.03194.x 11051356 19. Tibble J Sigthorsson G Foster R Sherwood R Fagerhol M Bjarnason I Faecal calprotectin and faecal occult blood tests in the diagnosis of colorectal carcinoma and adenoma Gut 2001 49 402 8 10.1136/gut.49.3.402 11511563 20. Gomollón F García-López S Sicilia B Gisbert JP Hinojosa J Grupo Espa˜nol de Trabajo en Enfermedad de Crohn y Colitis Ulcerosa Therapeutic guidelines on ulcerative colitis: a GRADE methodology based effort of GETECCU Gastroenterol y Hepatol 2013 36 104 14 10.1016/j.gastrohep.2012.09.006 21. D’Haens G Sandborn WJ Feagan BG Geboes K Hanauer SB Irvine EJ Lémann M Marteau P Rutgeerts P Schölmerich J Sutherland LR A review of activity indices and efficacy end points for clinical trials of medical therapy in adults with ulcerative colitis Gastroenterology 2007 132 763 86 10.1053/j.gastro.2006.12.038 17258735 22. Rutgeerts P Sandborn WJ Feagan BG Reinisch W Olson A Johanns J Travers S Rachmilewitz D Hanauer SB Lichtenstein GR de Villiers WJS Present D Sands BE Colombel JF Infliximab for induction and maintenance therapy for ulcerative colitis N Engl J Med 2005 353 2462 76 10.1056/NEJMoa050516 16339095 23. D’Haens G Ferrante M Vermeire S Baert F Noman M Moortgat L Geens P Iwens D Aerden I Van Assche G Van Olmen G Rutgeerts P Fecal calprotectin is a surrogate marker for endoscopic lesions in inflammatory bowel disease Inflamm Bowel Dis 2012 18 2218 24 10.1002/ibd.22917 22344983 24. Sipponen T Savilahti E Kärkkäinen P Kolho K-L Nuutinen H Turunen U Färkkilä M Fecal calprotectin, lactoferrin, and endoscopic disease activity in monitoring anti-TNF-alpha therapy for Crohn’s disease Inflamm Bowel Dis 2008 14 1392 8 10.1002/ibd.20490 18484671 25. Mooiweer E Fidder HH Siersema PD Laheij RJF Oldenburg B Fecal hemoglobin and calprotectin are equally effective in identifying patients with inflammatory bowel disease with active endoscopic inflammation Inflamm Bowel Dis 2014 20 307 14 10.1097/01.MIB.0000438428.30800.a6 24374878 26. Groβ P Doser K Falk W Obermeier F Hofmann C IL-33 attenuates development and perpetuation of chronic intestinal inflammation Inflamm Bowel Dis 2012 18 1900 9 10.1002/ibd.22900 22508383 27. Sedhom MAK Pichery M Murdoch JR Foligné B Ortega N Normand S Mertz K Sanmugalingam D Brault L Grandjean T Lefrancais E Fallon PG Quesniaux V Peyrin-Biroulet L Cathomas G Junt T Chamaillard M Girard J Ryffel B Neutralisation of the interleukin-33/ST2 pathway ameliorates experimental colitis through enhancement of mucosal healing in mice Gut 2013 62 1714 23 10.1136/gutjnl-2011-301785 23172891 28. Pascual-Figal D a Ordoñez-Llanos J Tornel PL Vázquez R Puig T Valdés M Cinca J de Luna AB Bayes-Genis A Soluble ST2 for predicting sudden cardiac death in patients with chronic heart failure and left ventricular systolic dysfunction J Am Coll Cardiol 2009 54 2174 9 10.1016/j.jacc.2009.07.041 19942089 29. Shen J Shang Q Wong C-K Li EK Wang S Li R-J Lee K-L Leung Y-Y Ying K-Y Yim C-W Kun EW Leung M-H Li M Li TK Zhu TY Yu S-L Kuan W-P Yu C-M Tam L-S IL-33 and soluble ST2 levels as novel predictors for remission and progression of carotid plaque in early rheumatoid arthritis: A prospective study Semin Arthritis Rheum 2015 45 18 27 10.1016/j.semarthrit.2015.02.001 25798875 30. Bandara G Beaven M a Olivera A Gilfillan AM Metcalfe DD Activated mast cells synthesize and release soluble ST2-a decoy receptor for IL-33 Eur J Immunol 2015 45 3034 44 10.1002/eji.201545501 26256265 31. Yamamoto T Shimoyama T Matsumoto K Consecutive monitoring of faecal calprotectin during mesalazine suppository therapy for active rectal inflammation in ulcerative colitis Aliment Pharmacol Ther 2015 42 549 58 10.1111/apt.13308 26140337 32. Lin J-F Chen J-M Zuo J-H Yu A Xiao Z-J Deng F-H Nie B Jiang B Meta-analysis: fecal calprotectin for assessment of inflammatory bowel disease activity Inflamm Bowel Dis 2014 20 1407 15 10.1097/MIB.0000000000000057 24983982 33. Boon GJ Day AS Mulder CJ Gearry RB Are faecal markers good indicators of mucosal healing in inflammatory bowel disease? World J Gastroenterol 2015 21 11469 80 10.3748/wjg.v21.i40.11469 26523111 34. Kristensen V Klepp P Cvancarova M Røseth A Skar V Moum B Prediction of endoscopic disease activity in ulcerative colitis by two different assays for fecal calprotectin J Crohns Colitis 2015 9 164 9 10.1093/ecco-jcc/jju015 25518057 35. Schoepfer AM Beglinger C Straumann A Trummler M Vavricka SR Bruegger LE Seibold F Fecal calprotectin correlates more closely with the Simple Endoscopic Score for Crohn’s disease (SES-CD) than CRP, blood leukocytes, and the CDAI Am J Gastroenterol 2010 105 162 9 10.1038/ajg.2009.545 19755969 36. Labaere D Smismans A Van Olmen A Christiaens P D’Haens G Moons V Cuyle P-J Frans J Bossuyt P Comparison of six different calprotectin assays for the assessment of inflammatory bowel disease United Eur Gastroenterol J 2014 2 30 7 10.1177/2050640613518201 37. Sandborn WJ Feagan BG Marano C Zhang H Strauss R Johanns J Adedokun OJ Guzzo C Colombel J-F Reinisch W Gibson PR Collins J Järnerot G Hibi T Rutgeerts P Subcutaneous Golimumab Induces Clinical Response and Remission in Patients With Moderate-to-Severe Ulcerative Colitis Gastroenterology 2014 146 85 95 10.1053/j.gastro.2013.05.048 23735746 38. García-Sánchez V Iglesias-Flores E González R Gisbert JP Gallardo-Valverde JM González-Galilea A Naranjo-Rodríguez A de Dios-Vega JF Muntané J Gómez-Camacho F Does fecal calprotectin predict relapse in patients with Crohn’s disease and ulcerative colitis? J Crohns Colitis 2010 4 144 52 10.1016/j.crohns.2009.09.008 21122498 39. Vestergaard T a Nielsen SL Dahlerup JF Hornung N Fecal calprotectin: assessment of a rapid test Scand J Clin Lab Invest 2008 68 343 7 10.1080/00365510701576198 17852834 40. Yamaguchi S Takeuchi Y Arai K Fukuda K Kuroki Y Asonuma K Takahashi H Saruta M Yoshida H Fecal calprotectin is a clinically relevant biomarker of mucosal healing in patients with quiescent ulcerative colitis J Gastroenterol Hepatol 2016 31 93 8 10.1111/jgh.13061 26212346 41. Calafat M Cabré E Mañosa M Lobatón T Marín L Domènech E High within-day variability of fecal calprotectin levels in patients with active ulcerative colitis: what is the best timing for stool sampling? Inflamm Bowel Dis 2015 21 1072 6 10.1097/MIB.0000000000000349 25793326 42. Sacco R Tanaka T Yamamoto T Bresci G Saniabadi AR Adacolumn leucocytapheresis for ulcerative colitis: clinical and endoscopic features of responders and unresponders Expert Rev Gastroenterol Hepatol 2015 9 327 33 10.1586/17474124.2014.953060 25160857 43. Saniabadi AR Tanaka T Ohmori T Sawada K Yamamoto T Hanai H Treating inflammatory bowel disease by adsorptive leucocytapheresis: a desire to treat without drugs World J Gastroenterol 2014 20 9699 715 10.3748/wjg.v20.i29.9699 25110409 44. Piper SE Sherwood R a Amin-Youssef GF Shah McDonagh T a Serial soluble ST2 for the monitoring of pharmacologically optimised chronic stable heart failure Int J Cardiol 2015 178 284 91 10.1016/j.ijcard.2014.11.097 25465308 45. Diamanti a Colistro F Basso MS Papadatou B Francalanci P Bracci F Muraca M Knafelz D De Angelis P Castro M Clinical role of calprotectin assay in determining histological relapses in children affected by inflammatory bowel diseases Inflamm Bowel Dis 2008 14 1229 35 10.1002/ibd.20472 18398894 46. Ho GT Lee HM Brydon G Ting T Hare N Drummond H Shand a G Bartolo DC Wilson RG Dunlop MG Arnott ID Satsangi J Fecal calprotectin predicts the clinical course of acute severe ulcerative colitis Am J Gastroenterol 2009 104 673 8 10.1038/ajg.2008.119 19262524

PMC005xxxxxx/PMC5002141.txt

==== Front BMC NephrolBMC NephrolBMC Nephrology1471-2369BioMed Central London 31510.1186/s12882-016-0315-6Research ArticleVitamins (A, C and E) and oxidative status of hemodialysis patients treated with HFR and HFR-Supra Palleschi Simonetta simonetta.palleschi@iss.it 1Ghezzi Paolo M. pmghezzi@gmail.com 2Palladino Giuseppe giuseppe.palladino@bellco.net 3Rossi Barbara barbara.rossi@iss.it 1Ganadu Marino anganadu@aslsassari.it 4Casu Domenica domecasu@tiscali.it 5Cossu Maria maricossu@yahoo.it 6Mattana Giovanni mattana.giovanni@tiscali.it 7Pinna Antonio Maria antonio.m.pinna@asloristano.it 8Contu Bruno brunocontu55@gmail.com 9Ghisu Tonina dialisi.dmacomer@aslnuoro.it 10Monni Alessandro mattias@tiscalinet.it 11Gazzanelli Luana lgazzanelli@aslolbia.it 12Mereu Maria Cristina mcmereu@aslsanluri.it 13Logias Franco logias@tiscali.it 7Passaghe Mario gpassaghe@aslolbia.it 14Amore Alessandro alessandro.amore@unito.it 15http://orcid.org/0000-0001-9788-2342Bolasco Piergiorgio +390706097335pg.bolasco@tin.it 16on behalf of the Sardinian Study GroupSitzia Irene Caiazzo Marialuisa Calvisi Luciangela Piras Angelo Sini Annalisa Sulis Emiliana Scalas Maria Rosaria Spiga Pierluigi Martorana Andrea Fancello Sabina Loiacono Elisa Sereni Luisa 1 Department of Hematology, Oncology and Molecular Medicine, Istituto Superiore di Sanità, Rome, Italy 2 Medical Scientific Consultant of Bellco s.r.l. Company, Mirandola, Italy 3 Bellco s.r.l Company, Mirandola, Italy 4 Ospedale A.Segni, Ozieri, Italy 5 Ospedale Civile, Alghero, Italy 6 Ospedale SS. Annunziata, Sassari, Italy 7 Ospedale S. Francesco, Nuoro, Italy 8 Ospedale S. Martino, Oristano, Italy 9 Ospedale N.S. della Mercede, Lanusei, Italy 10 ASL 3, Macomer, Italy 11 Ospedale SS. Trinità, Cagliari, Italy 12 Ospedale P. Merlo, La Maddalena, Italy 13 Ospedale N.S. di Bonaria, S. Gavino Monreale, Italy 14 Ospedale P. Dettori, Tempio Pausania, Italy 15 Ospedale Infantile Regina Margherita, Torino, Italy 16 Territorial Unit of Nephrology and Dialysis , ASL 8 of Cagliari, Cagliari, Italy 26 8 2016 26 8 2016 2016 17 1 12020 1 2016 19 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 Hemodiafiltration with on-line endogenous reinfusion (HFR) is an extracorporeal dialytic method that combines diffusion, convection and adsorption. HFR-Supra (HFR-S) is a second-generation system with increased convective permeability and adsorption capability. Previous studies suggested that HFR reduces oxidative stress compared to standard haemodialysis. The principal aim of the present study was to compare antioxidant vitamins behavior and oxidative status of hemodialysis patients treated with HFR and HFR-S. Methods The study was designed as a multicenter, randomized, crossover trial. Forty-one patients were recruited from 19 dialysis centers and after a 4-month washout stabilization period in on-line hemodiafiltration (ol-HDF), each patient was randomized to a sequence of treatments (HFR-S followed by HFR or viceversa) with each treatment applied over 6 months. Plasma levels of Advanced Oxidation Protein Products, Total Antioxidant Status, vitamins C, A and E and their ligands (Retinol Binding Protein and total lipids) were measured at baseline and at the end of each treatment period. Results Results show that the higher convective permeability of HFR-S with respect to HFR did not produce additional beneficial effects on the patients’ oxidative status, a slight decrease of both Vitamin A and Retinol Binding Protein being the only difference registered in the long-term. However, as compared to ol-HDF, both the re-infusive techniques allowed to reduce the intradialytic loss of Vitamin C and, in the long-term, improve the patients’ oxidative status and increase Retinol Binding Protein plasma values. No significant differences were found between the Vitamin C concentration of pre- and post cartridge UF neither in HFR-S nor in HFR showing that the sorbent resin does not adsorb Vitamin C. Conclusion HFR-S and HFR are almost equivalent in term of impact on antioxidant vitamins and oxidative status of hemodialysis patients. Nonetheless, as compared to ol-HDF, both treatments produced a sensible sparing of Vitamin C and may represent a new approach for reducing oxidative stress and related complications in dialysis patients. Long-term effects of re-infusive treatments on patients’ cardiovascular morbidity and mortality need to be evaluated. Trial registration ClinicalTrials.gov Identifier NCT01492491, retrospectively registered in 10 December 2011. Keywords Oxidative stressHemodiafiltrationAntioxidantsVitaminsAdsorptionHFRHigh permeabilityissue-copyright-statement© The Author(s) 2016 ==== Body Background Oxidative stress and inflammation are common occurrences in end stage renal disease (ESRD) and are believed to play an important role in the excess cardiovascular morbidity and mortality of these patients [1]. Oxidative stress in ESRD is due to both enhanced oxidative capacity and reduced antioxidant defenses, the latter including impaired enzyme activities and decreased levels of antioxidant vitamins [1–3]. Vitamin deficiencies in ESRD may originate from diet restriction, reduced absorption by medications and co-morbidities, uremia-related alterations of metabolic pathways and intradialytic losses [3]. Vitamin C (Vit C) is the most abundant and effective water-soluble antioxidant in human plasma. Vit. C deficiency is very common in dialysis patients mainly because of low dietary intake, losses during dialysis and accelerated catabolism [4–6]. Low levels of Vit C may have clinical consequences and associations with increased cardiovascular morbidity and mortality have been found [7, 8]. Vitamin E (Vit E) and Vitamin A (Vit A) are two fat-soluble plasma antioxidants and are transported in plasma by lipoproteins and retinol binding protein (RBP), respectively. Although inconsistent levels of Vit E levels have been reported in uremic patients [3], the administration of this vitamin has been suggested as a promising therapeutic strategy to limit both oxidative stress and its clinical consequences in hemodialysis patients; however, results obtained to date are contradictory [3, 9, 10]. Vit. A levels are generally increased in ESRD patients, mainly due to homeostatic dysregulation of its plasma carrier, the retinol binding protein (RBP) [3, 11]. Nonetheless, in ESRD patients relatively higher Vit. A concentrations are associated with a survival advantage [12, 13]. The molecular mechanisms underlying such an association have not yet been elucidated, albeit a role in the inhibition of inflammatory response has been hypothesized [12]. The enhanced oxidative capacity in uremia is due, at least in part, to systemic microinflammation and up-regulation of superoxide-producing enzymes [1]. This chronic condition results in structural modification of many plasma components, mainly protein and lipids [14]. These modified molecules, even of large size, might themselves become pro-inflammatory so ultimately leading to a further increase of oxidative stress. The removal of large uremic solutes, by breaking this vicious cycle, can therefore contribute to decrease oxidative stress in ESRD. It is known that large-size uremic solutes can more efficiently be removed by convective dialysis modalities, such as haemodiafiltration (HDF), than by standard haemodialysis (HD). On-line HDF (ol-HDF), performed using high-flux biocompatible membrane and high-quality ultrapure dialysate, has been reported to provide superior reduction in inflammation and oxidative stress than high-flux HD [15, 16]. However, convective techniques may also cause negative effects, including the loss of water-soluble antioxidants and amino acids, which cannot be replaced by reinfusion of the dialysate. In particular, it has been calculated that convective transport in ol-HDF is responsible of one-third of Vit. C loss during dialysis, so limiting the advantages of this techniques in decreasing oxidative stress [6]. Hemodiafiltration with endogenous reinfusion (HFR) is a kind of HDF that utilizes separated convection, diffusion and adsorption (Fig. 1) [17]. The system consist of a two-stage filter to separate convection from diffusion, combined with a sorbent cartridge to regenerate blood ultrafiltrate (UF). The regenerated ultrafiltrate, ie the plasmatic water deprived of those compounds which had bound to the sorbent resin, is reinfused back to the patient between the first and second stage of the filter. The sorbent cartridge contains styrene divinylbenzene (SDVB) hydrophobic resin which has a high affinity for different uremic toxins, cytokines and inflammatory mediators [18–21] while it does not retain nutrients like albumin and amino acids [22–24]. Hence, this technique allows the removal of medium/large uremic solutes by avoiding at the same time the loss of important nutrients. Previous studies showed that HFR has a better impact on both inflammation and oxidative stress than HD [25–28], with an efficacy similar to that of ol-HDF in the long-term reduction of inflammatory markers [29]. Notably, HFR architecture permits to increase convective membrane permeability up to a level which, due to the risk of severe nutrients losses and consequent malnutrition, is not achievable with standard HDF techniques. In the new HFR SUPRA (HFR-S) configuration, the permeability of the first convective chamber was incremented with a new polyethersulfone (PES) hyper high-flux membrane, Synclear 0.2 (Table 1). Uf proteomic analysis showed that Synclear 0.2, as compared to a PES high-flux membrane, allows a better extraction of middle-high molecular weight solutes and protein bound uremic toxins [30, 31]. Based on all the above, one can thus hypothesize that the increased depurative capacity of HFR-S might have a positive impact on the oxidative status of dialysis patients. The present study was designed to verify this hypothesis by comparing HFR and HFR-S treatments as for their short- and long-term effects on plasma levels of Advanced Protein Oxidation Products (AOPP), Total Antioxidant Status (TAS) and antioxidant vitamins C, A, E and their ligands (Retinol Binding Protein, RBP, and total lipids for Vit A and E, respectively). Moreover, since the UF re-infusion would allow sparing (micro)nutrients such as Vit C thereby potentially beneficial effects on patients (anti)oxidative status are produced, a secondary aim of the study was to compare the two re-infusive techniques with the ol-HDF, used in the run-in phase.Fig. 1 Schematic representation of HFR Table 1 Characteristics of HFR and HFR-S filters Filter section HFR HFR-S Membrane Type (surface) convective PES HF (0.7 m2) PES HHF (0.7 m2) Sterilization mode: gamma rays diffusive PES LF (1.7 m2) PES LF (1.7 m2) Kuf (mL/h/mmHg) convective 25 39 diffusive 13 13 Albumina sieving coefficient convective 0.002 0.2 diffusive <0.0001 <0.0001 Adsorbent cartridge (volume) Sterilization mode: dry steam SDVB (40 mL) SDVB (80 mL) PES Polyethersulfone (LF Low flux, HF high flux, HHF Hyper high flux) SDVB Styrene Divinylbenzene aBovin Serum Albumin Finally, we evaluated the sieving coefficient (SC-like) of vit C through the convective membranes and its adsorption on the SDVB resin cartridges at both the start and the end of HFR-S and HFR single sessions. Methods Study design and participant characteristics The study was designed as a prospective, multicenter, crossover trial with centralized randomization (Fig. 2). Between July and December 2011, forty-one ESRD patients were recruited from 19 territorial Dialysis Units of the National Public Health Care Service of Sardinia (Italy); their mean age was 67.0 ± 14.2 years (min 21, max 86); mean dialysis vintage was 101.8 ± 82.2 months. The total duration of the study was 16 months: after a 4-month run-in period in post-dilution ol-HDF by a hollow-fiber dialyzers (PHYLTHER, Bellco S.r.l., Mirandola, Italy) with 1.7 m2 PES HF membrane, each patient was randomized to a sequence of treatments (HFR-S followed by HFR or viceversa) with each treatment applied over 6 months. The randomization procedure was stratified by Center. Twenty-nine patients completed the study: two patients withdrew during the run-in period, further seven and three patients during the first and second treatment period, respectively (Fig. 2). Due to the earthquake that struck Mirandola district in 2012, some logistical problems arose in properly storing the plasma aliquots for AOPP an TAS measurements collected at the end of the second treatment period. Hence, only the values up to the first treatment period are available for these variables.Fig. 2 Study flow-chart Inclusion criteria were dialysis vintage longer than 6 months, age > 18 years and well functioning vascular access without recirculation (Qb ≥ 300 mL/min). Exclusion criteria were: residual diuresis > 300 mL/day; significant acute or chronic inflammatory comorbidities; non-renal related anemia; alcohol or drugs abuse; malignant neoplasms; hemoglobinopathy or mielopathy; pregnancy. The causes of ESRD were diabetes (21 %), glomerulonephritis (16 %), interstitial nephrits (9 %), angionephrosclerosis (30 %), polycystic kidney disease (5 %), undeterminated etiology nature (19 %). The majority of patients followed a drug regimen for the control of blood pressure, anemia and secondary hyperparathyroidism. All the subjects were warned to avoid antioxidant supplements thorough all the study. The extracorporeal dialysis treatment was performed three times per week. All 19 participating Centers made use of ultrapure, sterile and pyrogen-free dialysis solutions checked once monthly by cultural and endotoxins confirmation. Dialysate fluids were produced by a bi-osmosis and purified by daily nocturnal thermal disinfection of the piping (≥90 °C) [32–34]. Preparation of collection tubes for Vit C analysis To prevent the ex vivo degradation of Vit C, a reducing agent (dithioerythritol, DTE, 10 mmol/L final concentration in blood) was added in advance to the collection tubes. 10X DTE in physiological saline solution was prepared just before use, made sterile by filtration (0.22 μm pore size) and added to the tubes for blood and UF withdrawal (0.3 mL for each 3 mL heparinised gel-containing vacutainer and 0.2 mL for each 2 mL amber microcentrifuge tube, respectively). An insulin syringe was used to preserve the vacutainer void. The entire procedure was performed in a sterile environment and was highly reproducible (CV = 1.6 %). The tubes with the additive were stored at 4 °C until use (max 30 days). Preliminary experiments showed that the amount of DTE added did not affected blood cells integrity nor blood rheology up to at least 60 min from the collection, effectively preserving the Vit C content of either plasma or UF up to at least 90 days when specimens were stored at −25 °C [35]. Blood and UF sampling and processing Blood and pre- and post-cartridge UF samples were collected at any time of the study concomitantly with routine blood tests, both at the beginning and at the end of a first weekly dialysis session. Blood was collected in heparin gel-containing tubes (for vitamins’ assays) and EDTA tubes (for RBP, AOPP and TAS assays) and centrifuged within 30 min from the collection. Gel-containing centrifuged tubes, separated EDTA plasma and UF samples were then immediately frozen (−20 °C). Within 10 days from the collection, the samples were transferred in dry ice to the laboratory where they were stored at −80 °C until analysis (30 days max). AOPP and TAS assays Plasma AOPP levels were measured by spectrophotometry on a microplate reader (Bio-Rad 680 XR, Hercules, CA, USA) calibrated with chloramine-T solutions (Sigma-Aldrich, Saint Louis, MO, USA) in the presence of potassium iodide [36]. Absorbance of the reaction mixture was read at 340 nm against a blank solution. Results were expressed as μmol/L of chloramine-T equivalents. Plasma TAS levels were measured by a commercial colorimetric assay (TAS Randox, County Antrim, UK). Vit.s C, A, E assays The determination of Vits C, A and E was carried out by reverse-phase high performance liquid chromatography (Varian model 9010, Varian Medical Systems, Palo Alto, CA, USA) with UV and FL detection, according to previously published methods with some modifications [37, 38]. For Vit C assay, plasma and UF aliquots were extracted with an equal volume of 10 % metaphosphoric acid and the acidic extracts were then separated on a Synergy Polar-RP column (150 mm × 4.6 mm, 4 μm particle size, Phenomenex, Torrance, CA, USA), using 50 mM potassium phosphate buffer at pH 2.5 as mobile phase. For vit.s A and E assay, plasma aliquots were supplemented with the internal standard (6 μmol/L retinyl acetate) and extracted with an equal volume of a mixture of ethyl acetate-butanol (1:1). The separation of organic extracts was then performed on a fully end-capped C18 (150 mm × 4.6 mm, 5 μm particle size, Spherisorb ODS2, Waters Corporation, Milford, MA, USA) using 100 % methanol as mobile phase. Chromatographic runs were performed isocratically at a flow rate of 1.0 mL/min and the column temperature was kept constant at 28 °C. The eluted compounds were detected spectrophotometrically at 245 nm (Vit C), 292 nm (Vit E) and 325 nm (retinyl acetate) and fluorometrically at 325 nm/475 nm excitation/emission wavelengths (Vit A and retinyl acetate). Chromatograms were analysed by the Galaxie Chromatography Workstation software (version 1.9, Varian Inc.). Sample concentrations were calculated from peak areas by a linear calibration model. The limits of detection, calculated as the analyte concentration corresponding to a signal-to-noise (S/N) ratio of 3, were 0.6 μmol/L, 0.1 μmol/L and 2 μmol/L for Vits C, A and E, respectively. A correction factor was applied to Vit C results to correct for dilution by DTE. The chemicals used were of the highest purity available (Sigma-Aldrich) and the water was ultrapure (Milli-Q water filtration system, Millipore Spa, Rome, Italy). RBP and total lipids assays Plasma RBP was determined by a commercial nephelometric assay (BNII, Siemens Healthcare Diagnostics, Tarrytown, NY, USA). Cholesterol and triglycerides were determined by local clinical chemistry laboratories by means of commercial enzymatic methods. Total lipids were calculated as the sum of cholesterol and triglycerides. Data analysis The statistical power was verified by the Authors and Biostatistic Specialist of Ethic Committee. All blood values at the end of the dialysis session were corrected for hemoconcentration. Results are expressed as means and standard deviations - or medians and interquartile range depending on the distribution’s normality - for continuous data and as frequency for categorical data. The assumption of normality was checked by the Shapiro-Wilk test. Vitamin levels below the method LOD were assigned a value of ½ LOD. For all the variables but AOPP and TAS, the long-term effects of the treatments were compared by either CROS-analysis [39] (HFR-S vs HFR) or repeated measures ANOVA (HFR vs HFR-S vs run-in) using either parametric or non-parametric tests as appropriate. For AOPP and TAS, paired and unpaired t-test (or, when appropriate, their non-parametric equivalents) were used for comparison within arms and between arms, respectively. As for the acute effects, the percent reduction ratios (RR% = [start session value – end session value] *100/start session value) of the different treatments were compared by either ANOVA or ANCOVA, the latter being used in case of significant correlation between pre-dialysis levels and intra-dialytic losses. Only values ≥ LOD were entered into RR% calculation. Associations were assessed by Spearman rank correlation test. Extreme outliers (more than 3 inter-quartile ranges) were excluded from the statistical tests. P < 0.05 were considered significant and all tests were 2-tailed. Statistical analysis and data plotting were performed by using SigmaPlot ® 12.0 (Systat Software Inc.). Results and discussion Patient demographics and clinical characteristics at baseline are reported in Table 2. Study arms were well balanced and no significant biases were introduced by study dropouts at any time (data not shown). The operating conditions and adequacy parameters of the different extracorporeal treatments adopted in the study are reported in Table 3.Table 2 Personal and clinical data of patients at enrollment All Arm A Arm B n 41 21 20 Age (years) 70 (61–76) 73 (56–79) 67 (65–73) Dialysis vintage (months) 69 (35–153) 107 (53–164) 60 (29–135) Gender M = 26 (63 %) M = 12 (57 %) M = 14 (70 %) F = 15 (37 %) F = 9 (43 %) F = 6 (30 %) Albumin (mg/dL) 3.6 (3.4–3.9) 3.6 (3.4–4.0) 3.7 (3.5–3.9) Mean Body weight (kg) M = 63.5 ± 9.6 M = 63.2 ± 7.9 M = 63.8 ± 11.2 F = 56.5 ± 10.5 F = 53.6 ± 10.2 F = 60.7 ± 10.2 BMI M = 23.7 ± 3.4 M = 24.1 ± 2.3 M = 23.3 ± 4.2 F = 23.8 ± 4.4 F = 23.0 ± 4.1 F = 25.0 ± 4.8 Cardiovascular disease 59 % 62 % 55 % Hypertension 78 % 81 % 75 % Diabetes 22 % 19 % 25 % Charlson Comorbidity Score 6 (6–7) 6 (5–7) 6 (6–7) Previous dialysis treatment BHD = 15 (37 %) BHD = 8 (38 %) BHD = 7 (35 %) HDF = 10 (24 %) HDF = 4 (19 %) HDF = 6 (30 %) HFR = 16 (39 %) HFR = 9 (43 %) HFR = 7 (35 %) Continuous data are expressed either as mean ± SD or, in case of not normally distributed data, as median (IQR) Table 3 Operating conditions and adequacy parameters of the different dialysis treatments adopted in the study ol-HDF (run-in) HFR-S HFR Session length (min) 233 ± 18 235 ± 25 233 ± 16 Qb (mL/min) 313 ± 31 312 ± 31 301 ± 32 Qd (mL/min) 500 500 500 Body weight loss (kg/session) 2.8 ± 0.8 2.7 ± 1.1 2.8 ± 0.9 (Re)infusion volume (L/session) 14.9 ± 2.5 12.9 ± 2.9 13.4 ± 4.1 (Re)infusion rate (mL/min) 64.2 ± 9.6 54.9 ± 13.4 57.5 ± 13.2 eqKt/V 1.22 ± 0.25 1.17 ± 0.32 1.12 ± 0.23 ePCR (g/Kg/day) 1.4 ± 0.6 1.07 ± 0.54 1.08 ± 0.35 Data are expressed as mean ± SD (n = 41, 50 and 50 for ol-HDF, HFR-S and HFR, respectively) Patients’ biochemical data at the end of the run-in and of each treatment period are reported in Table 4. At the end of the run-in, most of the patients had low Vit. C, high Vit. A and normal Vit. E levels; RBP values were also higher than in non-uremic subjects resulting in lower-than-normal Vit. A/RBP molar ratios; AOPP values were in the normal range, while TAS values were higher-than-normal. For all parameters but AOPP, the levels we found were in agreement with those previously reported for uremic subjects [3, 8, 11, 40–42]. At difference, AOPP levels were significantly lower than it was previously reported for these patients [36]. Although this discrepancy could be due, at least in part, to the lack of standardized measurements for AOPP [43], this result suggests that our cohort was characterized by a low level of oxidative damage at baseline. Given that a significant association has been reported between malnutrition and oxidative stress in ESRD [44], the above evidence is in good agreement with and could be partially explained by the good nutritional status of enrolled patients (Table 2). Moreover, the use in all the participating centers of ultrapure dialysate locally produced by biosmosis could also have played a role in decreasing dialysis-related oxidative stress in our patients’ cohort [45].Table 4 Pre-dialysis levels of antioxidant vitamins and oxidative stress biomarkers at the end of run-in and of either HFR-S or HFR treatments End of run-in End of HFR-S End of HFR Vitamin C (μmol/L) 20.0 (11–32) 21.5 (12–43) 20.0 (10–31) Vitamin A (μmol/L) 2.8 ± 1.4 2.6 ± 1.0a 3.0 ± 1.2 Vitamin E (μmol/L) 20.2 ± 5.2 20.9 ± 5.9 20.9 ± 5.7 RBP (μmol/L) 5.8 ± 1.4 6.5 ± 1.6a, b 6.9 ± 1.6b Total lipids (mmol/L) 5.6 ± 1.2 6.1 ± 1.7 6.3 ± 1.7 Vitamin A/RBP (μmol/μmol) 0.47 ± 0.17 0.40 ± 0.11b 0.42 ± 0.13c Vitamin E/lipids (μmol/mmol) 3.6 ± 0.7 3.6 ± 0. 9 3.4 ± 0.8 AOPP (μmol/L) 28.4 (25–37) 21.7 (20–26)d - 35.8 (24–45) - 27.3 (21–34)d TAS (mmol/L) 1.88 (1.7–2.2) 1.41 (1.3–1.6)d - 1.99 (1.8–2.2) - 1.43 (1.4–1.5)d For all parameters but AOPP and TAS, only patients who completed the study were included in the analysis and cumulative data of the two treatment periods in HFR-S/HFR are represented. For AOPP and TAS only data of the first treatment period were available, hence all patients who reached this stage of the study were included in the analysis. Data are expressed either as mean ± SD or, in case of not normally distributed data, as median (IQR) a = p < 0.05 vs HFR, n = 29, Cross-analysis [39] b = p < 0.01 and c = p < 0.05 vs End of run-in, n = 29, ANOVA for repeated measures d = p < 0.01 vs End of run-in, n = 14 (HFR-S) or 16 (HFR), Wilcoxon Signed Rank Test At the end of the study, no significant differences were found between HFR-S and HFR as for their long-term effects on Vit C, Vit E, Vit E/lipids, AOPP and TAS levels (Table 4). At difference, significantly lower levels of Vit A and of its plasma carrier, RBP, were measured after HFR-S than after HFR, the Vit A/RBP ratio being not significantly different between the two treatments (Table 4, superscript asterisks). Accordingly, in the short-term, the only significant difference between HFR-S and HFR was the RBP RR% which was higher in HFR-S (Fig. 3). The present results suggest that HFR-S and HFR are equivalent in terms of effects on patients’ oxidative status. The modest difference in Vit A plasma levels did not significantly affect the plasma antioxidant capacity and is easily explained by the difference in the convective permeabilities of the two techniques: the higher convective permeability of HFR-S facilitates the removal of both RBP (MW = 21.2 kDa) and RBP/Vit A complexes from plasma leading to the decrease of the circulating levels of both the protein and its ligand. In fact, at difference of healthy subjects in whom virtually all Vit. A/RBP complexes are bound to prealbumin so preventing the renal loss of this vitamin, Vit. A/RBP complexes unbound to prealbumin are significantly increased in ESRD [40], allowing Vit. A passage trough high-flux membranes and its loss in the outflow dialysate. The evidence that the Vit A /RBP ratio was unchanged between the two treatments substantiates the above interpretation.Fig. 3 Reduction ratios of the studied variables at the end of ol-HDF (n = 41), HFR-S (n = 50) and HFR (n = 50) dialysis sessions. For AOPP and TAS the numbers of studied sessions are 30, 28 and 32 for ol-HDF, HFR-S and HFR, respectively. Each dialysis session at each time of the study was included in the analysis. Data are expressed as means ± SE (error bar). * = p < 0.01 vs ol-HDF; # = p < 0.05 vs ol-HDF; ° = p < 0.01 vs HFR-S; ANOVA (RBP, Vitamin A, Vitamin E) or ANCOVA (Vitamin C, TAS, AOPP) It has been demonstrated that the removal of middle-large uremic solutes by convective techniques contribute to decrease oxidative stress in haemodialysis patients [16]. Thus, we can conclude that the beneficial effects on oxidative stress produced by using high-flux membranes in HFR are not significantly improved by using hyper high-flux membranes in HFR-S, ie by further increasing the convective permeability. It could be argued that the low starting levels of AOPP might have decreased the power of the present study and that a higher number of subjects as well as a longer observation period would allow significant differences between HFR and HFR-S to emerge. Although we cannot rule out this possibility, the evidence that AOPP levels significantly decreased during the course of both HFR-S and HFR without significant differences between the two treatments (Table 4, superscript letters) makes this hypothesis unlikely. In fact, as compared to the values reached at the end of the run-in, both treatments induced a significant decrease of both AOPP and TAS and an increase of RBP (but not of vit. A) levels (Table 4, superscript letters). These long-term changes cannot be merely explained by differences in AOPP, TAS and RBP RR%s: actually, the RR%s measured during the run-in sessions (ol-HDF) were similar to those measured during HFR-S and HFR sessions, with the sole exception of RBP, whose RR% was significantly lower in HFR than in either HFR-S or ol-HDF, without significant differences between ol-HDF and HFR-S (Fig. 3). Thus, since RBP levels increased over time during both HFR and HFR-S treatments, it is unlikely that the differences in RR%s play a significant role in this phenomenon. Instead, the above results suggest that, as compared to ol-HDF, during both the re-infusive treatments an homeostatic response took place, consisting of 1) a progressive reduction of systemic oxidative stress accompanied by an adaptative reduction of the plasma antioxidant potential, and 2) an increase of RBP circulating levels. In our study, ol-HDF, HFR-S and HFR used the same machines, same ultrapure dialysate and same PES membranes, although that used in HFR-S had a higher convective permeability (see Table 1). Thus, in our study the main difference between ol-HDF and HFR-S/HFR consisted in the UF re-infusion. The re-infusion solution is the same plasmatic water of the patient purified on the sorbent cartridge; the composition of this solution depends on the sorbent adsorption capability which is virtually zero with respect to albumin and amino acids and significantly different from zero with respect to some uremic toxins and inflammatory citokines [18–24]. In the present study we show for the first time that the sorbent cartridge does not significantly adsorb Vit. C, since no significant differences were found between the Vit.C concentration of pre- and post cartridge UF neither in HFR-S nor in HFR (ns Wilcoxon Signed Rank Test, n = 22 and 26, respectively, Fig. 4). That means that the ultrafiltered Vit C can be entirely returned to the circuit, thereby its intra-dialytic loss is reduced. It has been reported that during a HDF session, both convective and diffusive losses of Vit. C occur, both potentially contributing to the oxidative stress in haemodialysis patients [6]. Accordingly, in the present study we show that the two re-infusive techniques allowed to reduce Vit C RR% by about 15 % as compared to ol-HDF (Fig. 3). Despite the sparing effect produced by each HFR-S/HFR session, we did not find any significant long term increase of Vit. C levels during the treatments (Table 4). This may be due, at least in part, to the high intra- and inter-subject variability of the plasma levels of this vitamin. Nonetheless, it cannot be ruled out that the sparing of Vit. C – and possibly of other small hydrophilic plasma components endowed with antioxidant/nutritional properties – may have played a role in the long-term improvement of oxidative status produced by the two re-infusive treatments.Fig. 4 Comparison between the vitamin C levels of pre- and post-cartridge UFs during HFR-S and HFR dialysis sessions. Dotted line = identity line As for RBP, the present results suggest that, as compared to ol-HDF, both HFR-S and HFR activate endogenous mechanisms of RBP plasma level modulation, likely due to the stimulation of its de novo synthesis. The RBP increase was not associated with a corresponding Vit. A increase, resulting in a net reduction of the Vit. A/RBP ratio (Table 4). This result was rather unexpected, since RBP is physiologically secreted bound to retinol. However, in uremic patients the retinol content of the liver is dramatically reduced [46], thus giving a possible explanation for the above result. At this time we can only speculate about the biological mechanisms underlying the observed increase of RBP plasma levels. From a physiological point of view, RBP behaves as a negative acute phase protein, ie both inflammation and protein malnutrition depress its hepatic synthesis [47–50]. It is known that hemodialysis patients are at higher risk of both inflammation and malnutrition, a large part of them exhibiting an activated acute phase response [51]. Hence, RBP increase might be a consequence of an improvement of patient inflammatory/nutritional status. A previous study did not find significant differences between the circulating levels of some inflammatory markers in HFR and ol-HDF, thus making this hypothesis unlikely [29]. However, it is worth underlying that in the present study RBP increase was accompanied by a significant improvement of patients’ oxidative status and that oxidative stress and inflammation share common pathogenic mechanisms [52]. From a pathological point of view, it should be taken into consideration that in ESRD free RBP and Vit A/RBP complexes are chronically increased, mainly due to the impairment of RBP renal catabolism [11, 40]. Based on an experimental model of AKI, it has been also proposed that a peripheral regulatory signal, normally removed by the kidney, might accumulate in the circulation and upregulate the hepatic release of Vit A/RBP complexes [53]. Thus one can hypothesize that this supposed regulatory signal accumulates in plasma of CKD patients, from where it might be removed by convective transport and, only in case of HFR(−S), be returned to the patient by means of UF re-infusion. In such a way, the stimulus to RBP hepatic release would be lower in ol-HDF than in either HFR-S or HFR. Anyway, regardless of the mechanisms behind the RBP level increase, higher RBP and Vit. A values in hemodialysis patients are associated with a better prognosis, lower values - even still higher than normal values - being strongly associated with higher mortality and cardiovascular risk [12, 13]. With this in mind and considering also that convective dialysis compared to purely diffusive dialysis is supposed to reduce cardiovascular mortality [54], further studies should be undertaken to i) investigate on the long-term impact of different convective techniques on patients’ survival and cardiovascular risk, and ii) better define, if any, the distinctive roles of Vit. A and RBP in the phenomenon. Finally, we found that, irrespective of the dialysis treatment adopted, i) Vit. C levels were negatively related to the patients’ age and positively related to the vit. A levels (p < 0.01), and, as expected, ii) Vit.s A and E levels were highly correlated to their respective plasma carriers, RBP and total lipids (p < 0.001,Table 5). A strong negative association between Vit. C and age in older people has been previously reported for the general population [55]. Hence, it is likely that physiological mechanisms linked to the elderly rather than to the kidney disease per se are responsible for the above association. The strong association between Vit. C and Vit. A is instead rather surprising since these vitamins are supposed to exert their antioxidant activity in different biological environments; in fact, while Vit. C is a water-soluble antioxidant, Vit. A is a lipid-soluble antioxidant, being significantly less effective than Vit. E in scavenging aqueous radicals [56]. In the general population only a weak interaction was found between Vit. C and Vit. A while a stronger association was found between Vit. C and Vit. E, likely explained by their close physiological interaction [57]. At difference, in our patients Vit. C was significantly related only to Vit. A. If mechanisms other than radical scavenging mediate Vit. C/Vit. A association in ESRD patients deserves further investigation and is beyond the aims of the present study.Table 5 Significant results of bivariate correlation analysis of antioxidant vitamins and oxidative stress biomarkers pre-dialysis levels at the end of run-in and after a six-months treatment period in HFR-S or HFR End of run-in End of HFR-S End of HFR Vitamin C vs age −0.626*** (31) −0.543** (30) −0.553** (30) Vitamin C vs Vitamin A 0.476** (31) 0.529** (30) 0.510** (30) Vitamin A vs RBP 0.841*** (32) 0.613*** (31) 0.616*** (30) Vitamin E vs total lipids 0.689*** (31) 0.642*** (27) 0.639*** (27) Analysed parameters included: age, dialysis vintage, Vit C, Vit A, Vit E, RBP, total lipids, Vit A/RBP, Vit E/total lipids, AOPP, TAS. All patients that completed at least the first treatment period were included in the analysis Data are expressed as rs (n), Spearman correlation test *p < 0.05; **p < 0.01, ***p < 0.001 Limitations of the present study were the relatively small sample size, the ethnicity of enrolled patients (all from Sardinia, an Italian island housing genetically homogenous populations with traditional living habits) and the lack of measurement of inflammation and uremic retention biomarkers. Further studies overcoming the above limitations will allow to strengthen and expand on the present results. Conclusion Kidney failure is associated with oxidative stress. Hemodialysis may contribute to increase oxidative stress of ESRD patients due to the activation of inflammatory pathways and to the loss of small water-soluble antioxidants. Convective dialysis modalities can reduce the burden of oxidative stress of hemodialysis patients by allowing the removal of large pro-oxidant uremic solutes from plasma. Unlike ol-HDF, which is a mixed convective-diffusive method, HFR-S and HFR exploit in sequence convection, adsorption and diffusion. Thanks to the adsorption step, UF is purified by toxic molecules and it can be re-infused to the patient. Since other UF components such as albumin, amino acids and Vit C are not adsorbed, HFR architecture allows to increase convective permeability without the risk of losing useful substances. In the present study, we show that the higher convective permeability of HFR-S with respect to HFR did not produce additional beneficial effects on the oxidative status of ESRD patients, the only significant difference being a long-term decrease of Vit. A due to the more efficient removal of its plasma carrier, RBP. Further studies are needed to investigate whether the more efficient removal of large molecules by HFR-S produces additional beneficial effects over HFR in term of (micro)inflammation and uremic toxins retention. In comparison to ol-HDF, both the re-infusive techniques allow to significantly reduce the intradialytic loss of Vit. C and, in the long-term, improve the patients’ oxidative status and increase RBP plasma values, all effects which might have a potentially positive impact on patients’ cardiovascular morbidity and mortality. Nutrient supplementation is often used to provide malnourished haemodialysis patients with essential aminoacids, vitamins and minerals and/or to counteract the loss of useful substances upon high-convection dialysis techniques. HFR-S and HFR, by returning his own plasmatic water back to the patient, can be a way to restore a more physiological status without additional therapies. Moreover, both the financial and technical costs of nutrient supplementation make HFR definitely more advantageous over other convective techniques with similar depurative capacity. Finally, HFR architecture paves the way for potential, further increase of convective permeability, an improvement otherwise entailing severe loss of plasma proteins which must then be restored by additional infusional therapies. Ol-HDF has been reported to produce clinical and survival benefits with respect to HD. Nonetheless ESRD patients on convective dialysis still show high levels of inflammation, oxidative stress and high cardiovascular morbidity and mortality. HFR-S and HFR could represent a new biotechnological response potentially able to reduce side effects and complications. The long-term impact of these new techniques on patients’ survival and cardiovascular risk need to be evaluated. Abbreviations AOPP, advanced oxidation protein products; DTE, dithioerythritol; ESRD, end stage renal disease; HFR, hemodiafiltration with on-line endogenous reinfusion; HFR-S, HFR-Supra; ol-HDF, on-line hemodiafiltration; PES, polyethersulfone; RBP, retinol binding protein; RR, reduction ratio; SDVB, styrene divinylbenzene; TAS, total antioxidant status; UF, ultrafiltrate; Vit A, vitamin A; Vit C, vitamin C; Vit E, vitamin E Acknowledgements We sincerely thank Mrs. Claudia Palazzari for her essential logistic contribution to the fulfillment of the study. Sardinian Study Group collaborators: Irene Sitzia 16, Marialuisa Caiazzo 3, Luciangela Calvisi 4, Angelo Piras 5, Annalisa Sini 8, Emiliana Sulis 9, Maria Rosaria Scalas 11, Pierluigi Spiga 13, Andrea Martorana 7, Sabina Fancello 14, Elisa Loiacono 15, Luisa Sereni 3. Funding Bellco S.r.l. has supported the laboratory measurements. Availability of data and materials The dataset supporting the present findings can be obtained from the study coordinator, dr. Piergiorgio Bolasco (pg.bolasco@tin.it). Authors’ contributions SP: made substantial contributions to study conception and design and to data acquisition, analysis and interpretation; performed the statistical analysis and drafted the manuscript. PMG: conceived the study and its design with PG; made substantial contributions to data analysis and interpretation; helped to draft the manuscript and to revise it for important intellectual content. GP: made substantial contributions to data acquisition, analysis and interpretation; helped to draft the manuscript. BR: made substantial contributions to vitamin data acquisition, analysis and interpretation; was involved in drafting the manuscript. MG: made substantial contributions to data acquisition. DC: made substantial contributions to data acquisition. MC: made substantial contributions to data acquisition. GM: made substantial contributions to data acquisition. AMP: made substantial contributions to data acquisition. BC: made substantial contributions to data acquisition. TG: made substantial contributions to data acquisition. AM: made substantial contributions to data acquisition. LG: made substantial contributions to data acquisition. MCM: made substantial contributions to data acquisition. FL: made substantial contributions to data acquisition. MP: made substantial contributions to data acquisition. AA: made substantial contributions to data analysis and interpretation. PB: coordinator of the Polycentric Sardinian Study Group. Conceived the study and its design with PMG; made substantial contributions to data acquisition, analysis and interpretation; helped to draft the manuscript and to revise it for important intellectual content; gave the final approval of the version to be published. All authors read and approved the final manuscript. Competing interests Paolo M. Ghezzi is medical-scientific consultant for Bellco S.r.l. (Mirandola-Italy), Giuseppe Palladino is a full time employees of Bellco S.r.l. (Scientific affairs Department and Laboratory Service). The remaining Authors declare no conflict of interest. Consent for publication Not applicable. Ethics approval and consent to participate Each enrolled patient was properly informed about the study, including the methods adopted for preserving privacy during collection and transmission of their personal and clinical data; each patient gave his/her written consent to participate. The study was approved by the Ethic Committee of the Cagliari Health Care Service and retrospectively registered at Clinical Trials.gov in 10 December 2011 (NCT01492491). ==== Refs References 1. Vaziri ND Oxidative stress in uremia: nature, mechanisms, and potential consequences Semin Nephrol 2004 24 469 473 10.1016/j.semnephrol.2004.06.026 15490413 2. Handelman GJ Vitamin C, deficiency in dialysis patients--are we perceiving the tip of an iceberg? Nephrol Dial Transplant 2007 22 2 328 331 10.1093/ndt/gfl534 17107966 3. Holden RM Ki V Morton AR Clase C Fat-soluble vitamins in advanced CKD/ESKD: a review Semin Dial 2012 25 334 343 10.1111/j.1525-139X.2012.01084.x 22607215 4. Raimann JG Levin NW Craig RG Sirover W Kotanko P Handelman G Is vitamin C intake too low in dialysis patients? Semin Dial 2013 26 1 5 10.1111/sdi.12030 23106569 5. Zhang K Dong J Cheng X Bai W Guo W Wu L Zuo L Association between vitamin C deficiency and dialysis modalities Nephrology (Carlton) 2012 17 452 457 10.1111/j.1440-1797.2012.01595.x 22404236 6. Morena M Cristol JP Bosc JY Tetta C Forret G Leger CL Convective and diffusive losses of vitamin C during haemodiafiltration session: a contributive factor to oxidative stress in haemodialysis patients Nephrol Dial Transplant 2002 17 422 427 10.1093/ndt/17.3.422 11865087 7. Dashti-Khavidaki S Talasaz AH Tabeefar H Hajimahmoodi M Moghaddam G Khalili H Plasma vitamin C concentrations in patients on routine hemodialysis and its relationship to patients’ morbidity and mortality Int J Vitam Nutr Res 2011 81 197 203 10.1024/0300-9831/a000065 22237767 8. Deicher R Ziai F Bieglmayer C Schillinger M Horl WH Low total vitamin C plasma level is a risk factor for cardiovascular morbidity and mortality in hemodialysis patients J Am Soc Nephrol 2005 16 1811 1818 10.1681/ASN.2004100850 15814831 9. Lu L Erhard P Salomon RG Weiss MF Serum vitamin E and oxidative protein modification in hemodialysis: a randomized clinical trial Am J Kidney Dis 2007 50 305 313 10.1053/j.ajkd.2007.05.001 17660032 10. Panichi V Rosati A Paoletti S Ferrandello P Migliori M Beati S A vitamin E-coated polysulfone membrane reduces serum levels of inflammatory markers and resistance to erythropoietin-stimulating agents in hemodialysis patients: results of a randomized cross-over multicenter trial Blood Purif 2011 32 7 14 10.1159/000321369 21242686 11. Vahlquist A Peterson PA Wibell L Metabolism of the vitamin A transporting protein complex. I. Turnover studies in normal persons and in patients with chronic renal failure Eur J Clin Invest 1973 3 352 362 10.1111/j.1365-2362.1973.tb00362.x 4760057 12. Espe KM Raila J Henze A Krane V Schweigert FJ Hocher B Impact of vitamin A on clinical outcomes in haemodialysis patients Nephrol Dial Transplant 2011 26 4054 4061 10.1093/ndt/gfr171 21543660 13. Kalousova M Kubena AA Kostirova M Vinglerova M Ing OM Dusilova-Sulkova S Lower retinol levels as an independent predictor of mortality in long-term hemodialysis patients: a prospective observational cohort study Am J Kidney Dis 2010 56 513 521 10.1053/j.ajkd.2010.03.031 20541302 14. Dhondt A Vanholder R Van BW Lameire N The removal of uremic toxins Kidney Int Suppl 2000 76 S47 S59 10.1046/j.1523-1755.2000.07606.x 10936799 15. den Hoedt CH Bots ML Grooteman MP van der Weerd NC Mazairac AH Penne EL Online hemodiafiltration reduces systemic inflammation compared to low-flux hemodialysis Kidney Int 2014 86 423 432 10.1038/ki.2014.9 24552852 16. Filiopoulos V Hadjiyannakos D Metaxaki P Sideris V Takouli L Anogiati A Vlassopoulos D Inflammation and oxidative stress in patients on hemodiafiltration Am J Nephrol 2008 28 949 957 10.1159/000142724 18594136 17. Wratten ML Ghezzi PM Ronco C Canaud B Aljama P Hemodiafiltration with endogenous reinfusion Contribution to Nephrology 2007 Basel Karger AG 94 102 18. De Nitti C Giordano R Gervasio R Castellano G Podio V Sereni L Choosing new adsorbents for endogenous ultrapure infusion fluid: performances, safety and flow distribution Int J Artif Organs 2001 24 765 776 11797846 19. Formica M Morosetti M Usberti M Brognoli M De Simone W Zito B Clinical experience in HFR with a new resin-only sorbent cartridge G Ital Nefrol 2004 21 Suppl 30 S71 S74 15747310 20. Wratten ML Sereni L Lupotti M Ghezzi PM Atti M Formica M Optimization of a HFR sorbent cartridge for high molecular weight uremic toxins G Ital Nefrol 2004 21 Suppl 30 S67 S70 15747308 21. Riccio E Cataldi M Minco M Argentino G Russo R Brancaccio S Evidence that p-cresol and IL-6 are adsorbed by the HFR cartridge: towards a new strategy to decrease systemic inflammation in dialyzed patients? PLoS One 2014 9 e95811 10.1371/journal.pone.0095811 24755610 22. De Simone W De Simone M De Simone A Iannaccone S Manganelli R Zito B Aspects of the on-line hemodiafiltration with regeneration and reinfusion of the ultrafiltrate (HFR): multicenter study G Ital Nefrol 2004 21 Suppl 30 S161 S167 15750977 23. Borrelli S Minutolo R De Nicola L Zamboli P Iodice C De Paola A Intradialytic changes of plasma amino acid levels: effect of hemodiafiltration with endogenous reinfusion versus acetate-free biofiltration Blood Purif 2010 30 166 171 10.1159/000320133 20924170 24. Aucella F Hemodiafiltration with endogenous reinfusion G Ital Nefrol 2012 29 Suppl 55 S72 S82 22723147 25. Gonzalez-Diez B Cavia M Torres G Abaigar P Muniz P Effect of a hemodiafiltration session with on-line regeneration of the ultrafiltrate on oxidative stress. Comparative study with conventional hemodialysis with polysulfone Blood Purif 2008 26 505 510 10.1159/000163848 18931477 26. Calo LA Naso A Carraro G Wratten ML Pagnin E Bertipaglia L Effect of haemodiafiltration with online regeneration of ultrafiltrate on oxidative stress in dialysis patients Nephrol Dial Transplant 2007 22 1413 1419 10.1093/ndt/gfl783 17237480 27. Calo LA Naso A Davis PA Pagnin E Corradini R Tommasi A Hemodiafiltration with online regeneration of ultrafiltrate: effect on heme-oxygenase-1 and inducible subunit of nitric oxide synthase and implication for oxidative stress and inflammation Artif Organs 2011 35 183 187 20946294 28. Borrelli S Minutolo R De Nicola L De Simone E De Simone W Zito B Effect of hemodiafiltration with endogenous reinfusion on overt idiopathic chronic inflammation in maintenance hemodialysis patients: a multicenter longitudinal study Hemodial Int 2014 18 758 766 10.1111/hdi.12178 24865622 29. Panichi V Manca-Rizza G Paoletti S Taccola D Consani C Filippi C Effects on inflammatory and nutritional markers of haemodiafiltration with online regeneration of ultrafiltrate (HFR) vs online haemodiafiltration: a cross-over randomized multicentre trial Nephrol Dial Transplant 2006 21 756 762 10.1093/ndt/gfi189 16303780 30. Monari E Cuoghi A Bellei E Bergamini S Caiazzo M Aucella F Proteomic analisys of protein extraction during hemofiltration with on-line endogenous reinfusion (HFR) using different polysulphone membranes J Mater Sci Mater Med 2014 25 2691 2698 10.1007/s10856-014-5290-5 25074835 31. Cuoghi A Caiazzo M Monari E Bellei E Bergamini S Sereni L New horizon in dialysis depuration: Characterization of a polysulfone membrane able to break the ‘albumin wall’ J Biomater Appl 2015 29 1363 1371 10.1177/0885328214565651 25542733 32. Alloatti S Bolasco P Canavese C Cappelli G Pedrini L Pizzarelli F Guidelines on water and solutions for dialysis. Italian Society of Nephrology G Ital Nefrol 2005 22 246 273 16001369 33. Bolasco P Contu A Meloni P Vacca D Murtas S The evolution of technological strategies in the prevention of dialysis water pollution: sixteen years’ experience Blood Purif 2012 34 238 245 10.1159/000343127 23147917 34. Bolasco P Contu A Meloni P Vacca D Galfre A Microbiological surveillance and state of the art technological strategies for the prevention of dialysis water pollution Int J Environ Res Public Health 2012 9 2758 2771 10.3390/ijerph9082758 23066395 35. Rossi B, Tittone F, Palleschi S. Set-up and validation of a convenient sampling procedure to timely and effectively stabilize Vitamin C in blood and plasma specimens stored at routine temperatures. Anal Bioanal Chem 2016, in press, doi:10.1007/s00216-016-9560-6 36. Witko-Sarsat V Friedlander M Capeillere-Blandin C Nguyen-Khoa T Nguyen AT Zingraff J Advanced oxidation protein products as a novel marker of oxidative stress in uremia Kidney Int 1996 49 1304 1313 10.1038/ki.1996.186 8731095 37. Nierenberg DW Lester DC Determination of vitamins A and E in serum and plasma using a simplified clarification method and high-performance liquid chromatography J Chromatogr 1985 345 275 284 10.1016/0378-4347(85)80165-1 4086599 38. Liau LS Lee BL New AL Ong CN Determination of plasma ascorbic acid by high-performance liquid chromatography with ultraviolet and electrochemical detection J Chromatogr 1993 612 63 70 10.1016/0378-4347(93)80368-E 8454704 39. Senn SS. Cross-over Trials in Clinical Research. 2nd ed. John Wiley & Sons; 2003. 40. Kelleher J Humphrey CS Homer D Davison AM Giles GR Losowsky MS Vitamin A and its transport proteins in patients with chronic renal failure receiving maintenance haemodialysis and after renal transplantation Clin Sci (Lond) 1983 65 619 626 10.1042/cs0650619 6354560 41. Palleschi S De Angelis S Diana L Rossi B Papa V Severini G Splendiani G Reliability of oxidative stress biomarkers in hemodialysis patients: a comparative study Clin Chem Lab Med 2007 45 1211 1218 10.1515/CCLM.2007.266 17635073 42. Gerardi G Usberti M Martini G Albertini A Sugherini L Pompella A Di Lorenzo D Plasma total antioxidant capacity in hemodialyzed patients and its relationships to other biomarkers of oxidative stress and lipid peroxidation Clin Chem Lab Med 2002 40 104 110 10.1515/CCLM.2002.019 11939481 43. Selmeci L Advanced oxidation protein products (AOPP): novel uremic toxins, or components of the non-enzymatic antioxidant system of the plasma proteome? Free Radic Res 2011 45 1115 1123 10.3109/10715762.2011.602074 21740311 44. Celik G Yontem M Cilo M Bilge M Mehmetoglu I Unaldi M The relationship between glutathione peroxidase and bioimpedance parameters in nondiabetic hemodialysis patients Hemodial Int 2012 16 274 281 10.1111/j.1542-4758.2011.00628.x 22099743 45. Susantitaphong P Riella C Jaber BL Effect of ultrapure dialysate on markers of inflammation, oxidative stress, nutrition and anemia parameters: a meta-analysis Nephrol Dial Transplant 2013 28 438 446 10.1093/ndt/gfs514 23291370 46. Stein G Schone S Geinitz D Abendroth K Kokot F Funfstuck R No tissue level abnormality of vitamin A concentration despite elevated serum vitamin A of uremic patients Clin Nephrol 1986 25 87 93 3516478 47. Kushner I The phenomenon of the acute phase response Ann N Y Acad Sci 1982 389 39 48 10.1111/j.1749-6632.1982.tb22124.x 7046585 48. Fleck A Clinical and nutritional aspects of changes in acute-phase proteins during inflammation Proc Nutr Soc 1989 48 347 354 10.1079/PNS19890050 2482490 49. Baeten JM Richardson BA Bankson DD Wener MH Kreiss JK Lavreys L Use of serum retinol-binding protein for prediction of vitamin A deficiency: effects of HIV-1 infection, protein malnutrition, and the acute phase response Am J Clin Nutr 2004 79 218 225 14749226 50. Koch A Weiskirchen R Sanson E Zimmermann HW Voigt S Duckers H Circulating retinol binding protein 4 in critically ill patients before specific treatment: prognostic impact and correlation with organ function, metabolism and inflammation Crit Care 2010 14 R179 10.1186/cc9285 20932285 51. Zimmermann J Herrlinger S Pruy A Metzger T Wanner C Inflammation enhances cardiovascular risk and mortality in hemodialysis patients Kidney Int 1999 55 648 658 10.1046/j.1523-1755.1999.00273.x 9987089 52. Pedruzzi LM Cardozo LF Daleprane JB Stockler-Pinto MB Monteiro EB Leite M Jr Systemic inflammation and oxidative stress in hemodialysis patients are associated with down-regulation of Nrf2 J Nephrol 2015 28 495 501 10.1007/s40620-014-0162-0 25585822 53. Gerlach TH Zile MH Upregulation of serum retinol in experimental acute renal failure FASEB J 1990 4 2511 2517 2335274 54. Nistor I Palmer SC Craig JC Saglimbene V Vecchio M Covic A Strippoli GF Haemodiafiltration, haemofiltration and haemodialysis for end-stage kidney disease Cochrane Database Syst Rev 2015 5 CD006258 25993563 55. Woodside JV Young IS Gilchrist SE Vioque J Chakravarthy U de Jong PT Factors associated with serum/plasma concentrations of vitamins A, C, E and carotenoids in older people throughout Europe: the EUREYE study Eur J Nutr 2013 52 1493 1501 10.1007/s00394-012-0456-8 23097178 56. Palace VP Khaper N Qin Q Singal PK Antioxidant potentials of vitamin A and carotenoids and their relevance to heart disease Free Radic Biol Med 1999 26 746 761 10.1016/S0891-5849(98)00266-4 10218665 57. Schunemann HJ Grant BJ Freudenheim JL Muti P Browne RW Drake JA The relation of serum levels of antioxidant vitamins C and E, retinol and carotenoids with pulmonary function in the general population Am J Respir Crit Care Med 2001 163 1246 1255 10.1164/ajrccm.163.5.2007135 11316666

PMC005xxxxxx/PMC5002142.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 358210.1186/s12889-016-3582-zStudy ProtocolInvestigating the patterns and determinants of seasonal variation in vitamin D status in Australian adults: the Seasonal D Cohort Study King Laura Laura.king@anu.edu.au 1Dear Keith keithdear4@gmail.com 2Harrison Simone L. simone.harrison@jcu.edu.au 3van der Mei Ingrid Ingrid.vandermei@utas.edu.au 4Brodie Alison M. abrodie1@usc.edu.au 56Kimlin Michael G. mkimlin@usc.edu.au 56Lucas Robyn M. +61 2 61253448robyn.lucas@anu.edu.au 11 National Centre for Epidemiology and Population Health, The Australian National University, Canberra, Australia 2 Duke Kunshan University, Kunshan, China 3 College of Public Health, Medical & Veterinary Sciences, James Cook, University, Townsville, Australia 4 Menzies Research Institute Tasmania, University of Tasmania, Hobart, Australia 5 Health Research Institute, University of the Sunshine Coast, Sippy Downs, Australia 6 Queensland University of Technology, Brisbane, Australia 26 8 2016 26 8 2016 2016 16 1 89223 7 2016 24 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 Vitamin D status generally varies seasonally with changing solar UVB radiation, time in the sun, amount of skin exposed, and, possibly, diet. The Seasonal D Study was designed to quantify the amplitude and phase of seasonal variation in the serum concentration of 25-hydroxyvitamin D, (25OH)D)) and identify the determinants of the amplitude and phase and those of inter-individual variability in seasonal pattern. Methods The Seasonal D Study collected data 2-monthly for 12 months, including demographics, personal sun exposure using a diary and polysulphone dosimeters over 7 days, and blood for serum 25(OH)D concentration. The study recruited 333 adults aged 18–79 years living in Canberra (35°S, n = 168) and Brisbane (27°South, n = 165), Australia. Discussion We report the study design and cohort description for the Seasonal D Study. The study has collected a wealth of data to examine inter- and intra-individual seasonal variation in vitamin D status and serum 25(OH)D levels in Australian adults. Electronic supplementary material The online version of this article (doi:10.1186/s12889-016-3582-z) contains supplementary material, which is available to authorized users. Keywords Vitamin DCohortSeasonAustraliaAdultDeterminantshttp://dx.doi.org/10.13039/501100000925National Health and Medical Research Council1023987Kimlin Michael G. issue-copyright-statement© The Author(s) 2016 ==== Body Background Vitamin D is a fat soluble steroid pre-hormone that is produced endogenously following exposure of the skin to solar ultraviolet (UV) radiation. In many regions of the world, including Australia, this endogenous synthesis is the major source of vitamin D, with only a small proportion deriving from exogenous sources such as diet, fortified foods and supplements [1]. The amount of UV radiation that is present at Earth’s surface, and the wavelength composition, varies according to distance from the Equator (latitude), the hemisphere [2], the time of day, and the time of year. That is, UV irradiance is higher closer to the Equator (lower latitude), at solar noon, and in mid-summer when the sun is most directly overhead. In addition, seasonal variation in UV radiation has a greater amplitude with increasing latitude (Fig. 1).Fig. 1 Variation in ambient UV radiation according to southerly latitude (in degrees) across the months of 2003 (data on ambient UV radiation from the TOMS satellite [18]) It is generally accepted that only shorter wavelength UVB radiation initiates vitamin D synthesis [3]. UVB photons are absorbed by the vitamin D precursor, 7-dehydrocholesterol (7-DHC), which is present in the lipid bilayer of the plasma membrane of epidermal keratinocytes and dermal fibroblasts [4]. This results in chemical rearrangement of 7-DHC and the formation of pre-vitamin D3, which is rapidly (within 2 h) converted to vitamin D3. The latter is released from the plasma membrane into the extracellular space, where is it adsorbed by vitamin D binding protein in the dermal capillary bed and introduced into the circulation. Vitamin D3 is metabolised in the liver to 25-hydroxyvitamin D [25(OH)D, the usual measure of vitamin D status], and then in the kidney (or in target tissues) to its biologically active form, 1,25-dihydroxyvitamin D3 (1,25(OH)D3) [4]. Based on the seasonal variation in ambient solar UVB radiation, seasonal variation in 25(OH)D levels, of greater amplitude at higher latitudes, is expected. But there are moderating effects, particularly the amount of time spent outdoors, clothing cover, and use of sun protection. For example, in very hot locations (typically low latitude), people may spend more time indoors during summer, while in cool locations, the warmer weather of summer leads people to spend more time outdoors [5, 6]. In addition, more skin is typically covered by clothing in the winter and this is particularly apparent at higher latitudes. In cross-sectional studies where participants have had a single blood draw that occurs at different times during the year for different participants, there is a clear seasonal pattern to the mean 25(OH)D level [1, 7]. Only a few studies have examined an individual’s 25(OH)D levels across more than one time period. In these studies, although winter and summer 25(OH)D levels were correlated (e.g. r = 0.46 in Japanese women [8]) or minimum and maximum 25(OH)D levels were (r = 0.68 in Swedish blood donors [9]), there was substantial inter-individual variation in the seasonal pattern. A similar study of 60 healthy free-living adults from southeast Queensland, Australia aged 18–87 years (70 % female) showed that 25(OH)D level increased in summer for most individuals; however, for some individuals, 25(OH)D levels decreased in summer (correlation of winter and summer 25(OH)D levels r = 0.38) [10]. A key missing component of the vitamin D and health story, recently identified from a review of studies in the United Kingdom [11], is to understand the determinants of seasonal variation in vitamin D status. Identifying what determines whether and how quickly people become vitamin D depleted during winter, and what drives the differences in the amplitude of seasonal variation between individuals, is critical to advancing knowledge in this area. There are important reasons to better understand inter- and intra-individual seasonal changes in 25(OH)D levels. First, observational studies typically test the association between the result of a single 25(OH)D test against disease risk, statistically “adjusting” for the seasonal effect based on the mean seasonal variation in 25(OH)D levels of the whole sample. The aim is to achieve measures of 25(OH)D level that are as though all participants had their blood taken on the same day. In longitudinal studies, more than one previous blood sample may be used, but in comparing either intra-individually or inter-individually, adjustment for season is usually required, as logistic considerations result in samples being taken at different times of the year. However, if there is poor correlation between winter and summer 25(OH)D, as has been previously observed [10], then an adjustment based on an individual maintaining their relationship to the mean, may not be valid. Failure to account adequately for inter-individual variation in seasonal pattern may result in spurious findings. Second, the degree of seasonal variation in 25(OH)D level itself may be a determinant of health. Increased disease risks at higher latitude may not be directly related to 25(OH)D level per se, but to the duration that a person remains vitamin D deficient in any given year, following winter (when ambient UV-B is at its lowest and clothing coverage is maximal). This possibility is difficult to investigate. Finally, understanding the amplitude and period of seasonal variation and their determinants may allow prediction of who is at risk of wintertime vitamin D deficiency based on demographic factors and 25(OH)D level measured in a different season. For example, the end of summer 25(OH)D level may allow prediction of the need for supplementation to avoid winter/early spring deficiency. Here we report the study design and cohort description for the Seasonal D Study. This is a two-centre cohort study in which participants are tracked for 12 months, in order to assess and quantify the contribution of phenotypic, behavioural and environmental factors to inter- and intra-individual seasonal variation in vitamin D status. The novelty of the present study lies in the regular objective measurement of sun exposure and skin phototype, with two-monthly standardised measurements across two locations within Australia with very different climates. Methods The Seasonal D Study was conducted in Canberra (35.3°S) which has a temperate climate with a marked summer/winter difference in temperature and ambient UV radiation; and Brisbane (27.5°S) which has a sub-tropical climate that remains relatively warm and humid year-round (Table 1).Table 1 Location, Temperature, Rainfall and UV Radiation for 2013–4 in the two Seasonal D Study Regions, Canberra and Brisbane, Australia (Source: Australian Bureau of Meteorology www.bom.gov.au, and Australian Radiation and Nuclear Safety Agency, www.arpansa.gov.au) Latitude Longitude Average noon clear sky UV Index Average Temperature °C Average Rainfall (mm) Summer (Jan) Winter (July) Summer (Jan) Winter (July) Summer (Jan) Winter (July) Brisbane 27.5 °S 153°E 10.7 3.8 30.6 22.1 212.7 26.6 Canberra 35.3°E 149°E 11.5 2.2 31.9 12.8 38.7 29.5 Recruitment The Seasonal D Study aimed to recruit 170 adult participants (18+ years) at each site with equal numbers of men and women and evenly spread across eleven age bands (18–24 years, then 5-yearly intervals between 25 and 75 years). Recruitment commenced in both locations in September 2012, and was completed in September 2013. The cohort was initially drawn from a pool of participants from the AusD Study [12] who had indicated that they would be willing to participate in follow-up studies. To reach the final recruitment targets, the remaining participants were recruited through advertisements placed around workplaces, through word of mouth and through snowball recruiting. We did not aim to recruit a random sample of the population, rather a study sample with sufficient variability in lifestyles and demographics to support investigation of the influences on seasonal variation in vitamin D status. Individuals who signalled an interest in the study were telephoned by a member of the research team to determine their eligibility and to obtain verbal consent. Individuals were eligible if they: were aged 18 years or older; had good comprehension of English; agreed to abstain from taking vitamin D supplements, multivitamins containing over 400 International Units (IU) of vitamin D, or cod liver oil, for the duration of their participation in the study; did not have a bleeding disorder; were not positive for Hepatitis B or C, or HIV; and were able to attend two-monthly on-site data collection interviews. Data collection The initial data collection interview was scheduled 10–14 days after confirmation of eligibility and verbal agreement to participate. Once the interview was scheduled, each participant was mailed a package containing materials to be used for detailed measurement of their sun exposure over the 7 days prior to their interview, to include both working and non-working days. The package included: a) seven individually packaged and coded UV-sensitive polysulphone dosimeters (with wristband) for measurement of cumulative daily personal sun exposure, and b) a self-administered diary (see Additional file 1) in which to record daily sun exposure, physical activity and sun-protection (amount and type of clothing, sunscreen and shade utilization) over the 7 days of dosimetry [12, 13]. Polysulphone dosimeters have been used extensively to assess personal UVB (vitamin D effective) exposure, including in our previous AusD study [12, 14]. In the present study, dosimeters were attached to the left wrist using the wristband supplied, and were replaced daily to avoid reaching saturation. Detailed instructions on the use of dosimeters and sun diary were included in the package, to supplement the information provided during the telephone interview. A reminder SMS was sent to participants on the day before they were to commence wearing the dosimeters and completing the sun diary. At the face-to-face interview, scheduled 1–2 days after completing 7-days of dosimetry, a research officer checked the completeness of the dosimeter and diary data, resolving any anomalies with the participant. Participants completed an online self-administered questionnaire (see Additional file 1) providing data on: date and place of birth, ancestry, current pregnancy and breastfeeding status, highest educational achievement, occupation held for the longest period, current employment status and occupation, total household income, smoking history, alcohol intake, self-rated health, current and past diagnosis of specified illnesses (including skin cancers and bone diseases), sun sensitivity (also used to assess Fitzpatrick skin type [15]), usual time spent outdoors recorded in hourly intervals for each day of the week during the previous month and usual use of sun protection when outdoors, dietary intake relevant to vitamin D, physical activity in the last 7 days [16], and usual time spent outdoors on working days and non-working days in each season. The full self-administered questionnaire was used for the baseline and end-of-study data collection, with a briefer questionnaire on factors that are likely to change over time, such as diet, for the interim data collection. Participants were also asked to bring their nutritional supplements and medications with them to the interview so that name and dose could be accurately recorded. Measurements were taken using standardised protocols, and included: height (baseline only); weight; waist and hip circumference; spectrophotometric skin reflectance (using a Minolta 2500d to record L*a*b values) on a sun-exposed body-site (the dorsum of the hand and the cheek) to record facultative skin colour and a sun-protected body-site (the upper inner arm) to record natural skin colour; blood pressure; and handgrip strength (using a portable dynamometer). A non-fasting blood sample was taken by venepuncture into a serum separator tube and, within four hours, the serum was removed and placed in 1 ml aliquots in cryotubes and frozen at −80 °C. All participants undertook the same data collection protocol (7 days of UV dosimetry and sun diary recording, followed by self-administered questionnaire, measurements and blood collection at interview) every two months over a 12-month period (i.e. 7 data collection episodes). Data collection was completed in July 2014. At the completion of the study, total serum 25(OH)D was measured using a Diasorin Liaison semi-automated chemiluminescence assay, with every tenth sample also measured by high performance liquid chromatography HPLC for quality assurance. This Diasorin Liaison assay at the Queensland University of Technology (QUT) participates in and is certified by the international Vitamin D External Quality Assessment Scheme. The inter- and intra-assay coefficients for this assay were 3–6 % and 6–9 %, respectively. Data management Data on meteorological variables were available from routine measurements from instruments located at the nearby Canberra and Brisbane airports: ambient UV radiation levels from the Australian Radiation Protection and Nuclear Safety Agency monitors and weather variables (i.e. temperature, precipitation) from the Australian Bureau of Meteorology. Questionnaire responses were entered directly into an online survey software tool (KeySurvey) then imported into Excel; all other data were scanned directly into an Access database. Dubious data were resolved by reference to the original forms. Diary days are excluded from the analysis if they had been insufficiently completed (i.e. the participant left the diary day blank), totalling 282 diary days. Sun diaries that did not have seven complete days recorded are retained in the analysis if there were at least three working days and at least one non-working day. The absorbance at 330 nm of UV dosimeters was measured prior to, and following their use, with a spectrophotometer based at QUT; daily UV radiation exposure was calculated based on the change in pre-to-post absorbance, with a seasonal calibration factor applied. Results were reported as Standard Erythemal Dose (SED; 100 J/m2) [17]. Statistical analysis In this paper we provide a brief description of the study sample and the data available for analysis. The Seasonal D Study was powered to analyse the characteristics of individually fitted seasonal curves for the variation in 25(OH)D levels, specifically their mean, amplitude and phase. We assumed that inter-individual variation would dominate the intra-individual error of estimation, and focused on modelling the mean 25(OH)D level as a function of the ambient UV radiation. The main factors influencing the study’s exposure variance, which in turn determines study power, were assumed to be location and season. On the log10 scale of UV Index, from available environmental data, the standard deviation (SD) across sites and seasons was estimated to be 0.345 log units. This approximates the study SD because there are equal numbers in each location and across time. We further assumed an error SD (in the estimated individual mean 25(OH)D level) of 25 nmol/L. The sample size to have good statistical power (90 % at p <0.05) was estimated to be n = 170, to detect a regression slope (beta) of 5 nmol/L or greater, per doubling of ambient UV radiation. As modelling the amplitude and phase the seasonal curve for 25(OH)D was likely to be less precise than modelling the mean, we represented this by a doubling of the error variance, requiring twice the sample size, or N = 340. Data on all participants with more than one data collection time point will be included in the analysis. Data from the diary and UV dosimeters will be included if there are data for at least two working days, and one non-working day, for any data collection period. Participants with missing data on other variables will be excluded from analyses that include those specific variables. We used the STROBE guidelines for reporting of observational studies to report the protocol and study sample description for the Seasonal D Study (see Additional file 2). Study progress The Seasonal D Study has recruited a total of 333 adult participants aged 18 to 79 years. This includes 168 participants from the Canberra study region and 165 participants from the Brisbane study region. Of 2124 sun diaries completed, 87 % were completed in full for the seven days (n = 1848); 8.5 % had 6 days complete (n = 180), and 4.5 % of diaries had five or fewer days complete (n = 96). Complete data for seven sun diaries is available for 255 participants. Discussion How to achieve and maintain vitamin D sufficiency, however it is defined, is a highly topical and contentious issue in many countries. Most studies investigating vitamin D status have been cross-sectional surveys that examine the prevalence of vitamin D insufficiency and deficiency, or the determinants of vitamin D status at a single time point. Seasonal variation is commonly reported, but relates to the mean 25(OH)D level of the study group who are sampled at different times of the year [1, 7]. Data from the Seasonal D Study will contribute to understanding how 25(OH)D levels vary across the year in locations with very different climates, as well as the determinants of intra-individual variation and inter-individual differences in seasonal patterns of 25(OH)D depletion/repletion. We will characterise seasonal impacts on a range of lifestyle factors, including diet, indoor/outdoor activities, physical activity, clothing cover, and climatic factors (principally temperature and ambient UV radiation), and quantify their effects on 25(OH)D levels. The Seasonal D Study was purpose-designed to analyse seasonal variation in 25(OH)D concentration and vitamin D status and their determinants. Strengths include that the study included two locations with different climates to provide considerable variability in ambient UV radiation and other seasonally-varying factors. All data collection tools and methods were standardised across the two centres and research officers were trained to use a Standard Operating Procedure. The assays of 25(OH)D concentration were completed in a single laboratory in a batched analysis at the completion of the study, with good intra- and inter-batch agreement. One limitation of the study is that cohort is older (median age 49 years) and women are over represented, relative to the Australian population. These factors limit the generalisability of the findings. Furthermore the study is set solely in Australia, a country with high levels of ambient UV radiation. While this provides the potential for high variability in levels of exposure to UV radiation (ranging from individuals who work indoors and have indoor pastimes to the outdoor worker in a high ambient UV setting), it may further limit the generalisablity of the results to other populations in less sunny locations. Current recommendations specify maintaining circulating 25(OH)D levels above 50 nmol/L throughout the year, yet are based on studies in which 25(OH)D concentration is measured at a single time point, without knowledge of, or consideration for, what may be a natural cyclical pattern of variability in 25(OH)D level by season. The Seasonal D Study will describe the typical rhythm of vitamin D production over the year for adults living in different climatic conditions. We will better understand the profile of Australian adults who are vitamin D deficient/sufficient all year round and understand which characteristics of individuals contribute to greater (or lesser) duration of time replete (above 50 nmol/L). Although many studies, in various scientific fields, have quantified seasonality (usually using “cosinor analysis”), few have explicitly investigated its determinants using statistical models. Those few that do have limited their explorations to determinants of the magnitude of seasonal variation, as summarised through statistics such as the Gini coefficient. We aim to model the detailed characteristics of seasonal patterns, and determine how many data points per person are needed to accurately estimate an individual’s seasonal vitamin D profile, so that the profile itself, not just a single 25(OH)D level, can be tested in relation to disease risks. Here we have provided the rationale, design, and methods for the Seasonal D Study which has followed 333 Australian adults over 12 months, with data collection on 7 occasions separated by approximately 2 months, providing detailed measurement of personal sun exposure and sun protection in relation to measured vitamin D status. Future work using these data will be to better understand the determinants of seasonal variation across the whole cohort, inter-individual variability within the cohort, and define intra-individual trajectories to support evidence-based sun protection and vitamin D maintenance recommendations. Additional files Additional file 1: Questionnaires and sun diary for the Seasonal D Study. (PDF 482 kb) Additional file 2: Reporting against the STROBE guidelines. (DOC 101 kb) Abbreviations 1,25(OH)D31,25-dihydroxyvitamin D3 25(OH)D25-hydroxyvitamin D IUInternational units QUTQueensland University of Technology SEDStandard Erythemal Dose SMSShort message service UVUltraviolet UVBUltraviolet radiation in the UVB wavelengths (280 nm to 315 nm) Acknowledgements The authors would like to acknowledge the contributions of the research officers who assisted with data collection for the Seasonal D Study and the study participants. Funding The Seasonal D Study was funded by the National Health and Medical Research Council (Project Grant #1023987). The funding body had no role in the design of the study, data collection, analysis and interpretation of the data or in writing the manuscript. Availability of data and materials Not applicable at this stage. Authors’ contributions RML, MGK, and KD designed the study, with assistance from SLH and IvDM; RML, MGK and AB undertook data collection; LK analysed the data; RML wrote the initial draft of the paper. All authors contributed to refining drafts and 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 Seasonal D Study was approved by the ethics committees of the Australian National University (2012/004) and Queensland University of Technology (1100001457). All participants signed written informed consent prior to participation in the study. ==== Refs References 1. Kimlin MG Lucas RM Harrison SL van der Mei I Armstrong BK Whiteman DC The contributions of solar ultraviolet radiation exposure and other determinants to serum 25-hydroxyvitamin D concentrations in Australian adults: the AusD Study Am J Epidemiol 2014 179 7 864 874 10.1093/aje/kwt446 24573539 2. McKenzie R Bodeker G Keep D Kothamp M UV Radiation in New Zealand: North-to-South differences between two sites and relationship to other latitudes Weather Climate 1996 16 1 17 26 3. McKenzie R Liley B Johnston P Scragg R Stewart A Reeder AI Small doses from artificial UV sources elucidate the photo-production of vitamin D Photochem Photobiol Sci 2013 12 9 1726 1737 10.1039/c3pp50041a 23807653 4. Holick MF The cutaneous photosynthesis of previtamin D3: a unique photoendocrine system J Invest Dermatol 1981 77 1 51 58 10.1111/1523-1747.ep12479237 6265564 5. Xiang F Harrison S Nowak M Kimlin M Van der Mei I Neale RE Weekend personal ultraviolet radiation exposure in four cities in Australia: influence of temperature, humidity and ambient ultraviolet radiation J Photochem Photobiol B 2015 143 74 81 10.1016/j.jphotobiol.2014.12.029 25600266 6. Sun J Lucas RM Harrison S van der Mei I Armstrong BK Nowak M The relationship between ambient ultraviolet radiation (UVR) and objectively measured personal UVR exposure dose is modified by season and latitude Photochem Photobiol Sci 2014 13 12 1711 1718 10.1039/C4PP00322E 25311529 7. Lucas RM Ponsonby AL Dear K Valery PC Taylor B van der Mei I Vitamin D status: multifactorial contribution of environment, genes and other factors in healthy Australian adults across a latitude gradient J Steroid Biochem Mol Biol 2013 136 300 308 10.1016/j.jsbmb.2013.01.011 23395985 8. Nakamura K Nashimoto M Yamamoto M Summer/winter differences in the serum 25-hydroxyvitamin D3 and parathyroid hormone levels of Japanese women Int J Biometeorol 2000 44 4 186 189 10.1007/s004840000067 11131290 9. Klingberg E Olerod G Konar J Petzold M Hammarsten O Seasonal variations in serum 25-hydroxy vitamin D levels in a Swedish cohort Endocrine 2015 49 3 800 808 10.1007/s12020-015-0548-3 25681052 10. Kimlin M Seasonal vitamin D UV Radiation and its Effects: an update 2010: 7–9 April 2010 2010 2010 Queenstown NIWA: National Institute for Water and Atmosphere 11. Ashwell M Stone EM Stolte H Cashman KD Macdonald H Lanham-New S UK Food Standards Agency Workshop Report: an investigation of the relative contributions of diet and sunlight to vitamin D status Brit J Nutr 2010 104 4 603 611 10.1017/S0007114510002138 20522274 12. Brodie AM Lucas RM Harrison SL van der Mei IA Armstrong B Kricker A The AusD Study: A population-based study of the determinants of serum 25-hydroxyvitamin D concentration across a broad latitude range Am J Epidemiol 2013 177 9 894 903 10.1093/aje/kws322 23524036 13. King L Xiang F Swaminathan A Lucas RM Measuring sun exposure in epidemiological studies: Matching the method to the research question J Photochem Photobiol B 2015 153 373 379 10.1016/j.jphotobiol.2015.10.024 26555640 14. Sun J Lucas RM Harrison SL van der Mei I Whiteman DC Mason R Measuring exposure to solar ultraviolet radiation using a dosimetric technique: understanding participant compliance issues Photochem Photobiol 2014 90 4 919 924 24571445 15. Fitzpatrick TB The validity and practicality of sun-reactive skin types I through VI Arch Dermatol 1988 124 6 869 871 10.1001/archderm.1988.01670060015008 3377516 16. Craig CL Marshall AL Sjostrom M Bauman AE Booth ML Ainsworth BE International physical activity questionnaire: 12-country reliability and validity Med Sci Sports Exerc 2003 35 8 1381 1395 10.1249/01.MSS.0000078924.61453.FB 12900694 17. Diffey BL Sources and measurement of ultraviolet radiation Methods 2002 28 1 4 13 10.1016/S1046-2023(02)00204-9 12231182 18. Total Ozone Monitoring Satellite, National Aeronautics and Space Administration. http://iridl.ldeo.columbia.edu/SOURCES/.NASA/.GSFC/.TOMS/; 2004. Accessed 02 June 2006.

PMC005xxxxxx/PMC5002143.txt

==== Front Biomed Eng OnlineBiomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 11910.1186/s12938-015-0119-0Book ReviewBook review of “The Biomedical Engineering Handbook” fourth edition, edited by Joseph D. Bronzino, Donald R. Peterson Koprowski Robert koprow@us.edu.pl Department of Biomedical Computer Systems, Faculty of Computer Science and Materials Science, Institute of Computer Science, University of Silesia, ul. Będzińska 39, 41-200 Sosnowiec, Poland 6 1 2016 6 1 2016 2016 15 1 123 11 2015 22 12 2015 © Koprowski. 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.This article is a review of the book “The biomedical engineering handbook”, fourth edition: four volume set (ISBN 9781439825334, 254GBP, 5430 pages) edited by Joseph D. Bronzino and Donald R. Peterson published by the CRC Press Taylor & Francis group in 2015. The content of the book and its importance for biomedical engineering have been discussed in this invited review. Keywords BiomechanicsBiomaterialsBiomedical sensorsBiomedical devicesBiomedical technologiesImage processingMedical informaticsMolecular biologyArtificial organsissue-copyright-statement© The Author(s) 2016 ==== Body Book details Title:“The biomedical engineering handbook” ISBN:9781439825334, Volume IPrint ISBN978-1-4398-2530-3 eBook ISBN978-1-4398-2531-0 Volume IIPrint ISBN978-1-4398-2527-3 eBook ISBN978-1-4398-2528-0 Volume IIIPrint ISBN978-1-4398-2525-9 eBook ISBN978-1-4398-2526-6 Volume IVPrint ISBN978-1-4398-2518-1 eBook ISBN978-1-4398-2519-8 Price:254GBP, Number of pages:5430, Number of volumes:4, Edited by:Joseph D. Bronzino and Donald R. Peterson, Published by:CRC Press Taylor & Francis group in 2015. Book review The book submitted for review entitled “The biomedical engineering handbook” was edited by Joseph D. Bronzino and Donald R. Peterson. The book consists of four hardcover volumes amounting in total to 5430 pages. This is another fourth edition. The four volumes present issues relating to the four areas of biomedical engineering: biomedical engineering fundamentals, medical devices and human engineering, biomedical signals, imaging, and informatics, molecular, cellular and tissue engineering. These volumes are divided into chapters, i.e.:The first volume (biomedical engineering fundamentals) has 5 chapters comprising 58 articles and the discussed issues concern the following areas: physiologic systems, biomechanics, bioelectric phenomena, neuroengineering, The second volume (medical devices and human engineering) has 5 chapters comprising 57 articles and the discussed issues concern the following areas: biomedical sensors, medical instruments and devices, human performance engineering, rehabilitation engineering, clinical engineering, The third volume (biomedical signals, imaging, and informatics) has 4 chapters comprising 66 articles and the discussed issues concern the following areas: biosignal processing, medical imaging, infrared imaging, medical informatics, The fourth volume (molecular, cellular and tissue engineering) has 9 chapters comprising 93 articles, the discussed issues concern the following areas: molecular biology, transport phenomena and biomimetic systems, physiological modelling, simulation and control, stem cell engineering: an introduction, tissue engineering, artificial organs, drug design, delivery systems, and devices, personalized medicine, ethics. Details of the individual volumes and changes made in subsequent editions are presented in the next section. Specific content In recent years “The biomedical engineering handbook” has been constantly expanded and extended to include further areas of medicine and engineering. The first edition was published in 1995 and won the award: “Winner of the Association of American Publishers Best New Professional/Scholarly Publication—Engineering”. The second edition was published in 2000 by CRC PRESS LLC, Springer in cooperation with IEEE Press. The 2006 edition was also published by CRC Press. The fourth, reviewed edition has been available for readers since 2015, both online and in printed version. The most obvious difference among these subsequent editions has been length. For example, the second edition has 3024 pages in two volumes, the third edition has 4232 pages (3800 b/w illustrations) in three volumes and the fourth edition has 5430 pages (2060 b/w illustrations) in four volumes. On average, therefore, each subsequent edition is published every 5 or 6 years and extended for additional volumes. According to the information given by editors in the preface to this latest edition: “More specifically, this fourth edition has been considerably updated and contains completely new sections, including.Stem cell engineering Drug design, delivery systems, and devices Personalized medicine As well as a number of substantially updated sections, includingTissue engineering (which has been completely restructured) Transport phenomena and biomimetic systems Artificial organs Medical imaging Infrared imaging Medical informatics” In fact, these sections were added and the others expanded with new articles written by new authors. In the preface to the fourth edition, the editors talk about the evolution of the modern healthcare system and then define the term “Biomedical engineering”. At the end of the preface, they distinguish several major divisions of biomedical engineering activities. This division may be controversial for the readers, which I mention later in this review describing the weaknesses of the book. The preface lists the many (570) contributors who were responsible for individual sections of this handbook. I looked up the citation counts of these contributors. In the aggregate, these 153 (43-Vol. I, 14-Vol. II, 30-Vol. III, 66-Vol. IV) authors have themselves published 100 journal articles that have attracted an aggregate number of 40,000 citations. In short, these authors are themselves productive investigators. The most prolific of the authors were David L. Kaplan, Antonios G. Mikos and Jong-Ryoul Kim. It should also be noted that the main editors (Joseph D. Bronzino and Donald R. Peterson) are very prolific authors of many articles and respected researchers in this field. The thematic scope of the articles is so wide that in this short review it is difficult to mention in detail even the key issues. Individual sections have a number of articles that range from a few to several dozen. The subjects of articles are arranged in order from history through a review of achievements in the considered biomedical engineering area and ending with specialist articles contemplating and solving a specific research problem. The first volume is devoted to the fundamentals of biomedical engineering. It contains basic information in the field of biomechanics, biomaterials, and neuroengineering. Most of the book is richly illustrated both in terms of the presented biological material as well as the schematic and block diagrams of the performed experiments. The articles in the first chapter, namely Physiologic Systems, are extremely important didactically. They provide detailed descriptions of the nervous as well as vision and hearing systems. These are the following articles:Evangelia Micheli-Tzanakou: “Nervous system”, comprising the following, more vital, sections: definitions, functions of the nervous system, representation of information in the nervous system, lateral inhibition, higher functions of the nervous system, abnormalities of the nervous system, Aaron P. Batista and George D. Stetten: “Vision system”, comprising the following, more vital, sections: fundamentals of vision research, a modular view of the vision system, eye movements, Ben M. Clopton and Herbert F. Voigt: “Auditory system”, comprising the following, more vital, sections: overview, peripheral auditory system, central auditory system, pathologies. The other chapters present a theoretical, simulation and experimental approach. The electronic version of the book enables the reader to see the selected images in higher resolution with a click of the mouse. However, apart from a few cases, no additional materials such as videos or multimedia presentations are included. These could enrich the educational value of the handbook, which is a potential use that the authors mention. The second volume is devoted to biomedical devices and human engineering. Michael R. Neuman, a very senior bioengineer, introduces the volume. He separates the field of biomedical devices into sensors, signal processors as well as display and storage devices. The subsequent articles describe various types of sensors ranging from non-electrical value sensors to electrical values sensors. The next chapter presents biomedical devices ranging from measuring amplifiers through respirators and lasers to optical monitoring. The subsequent chapters are devoted to rehabilitation and clinical engineering. They contain articles in the field of orthopaedics, ophthalmology and others. As in the first volume of the reviewed book, all the articles are very well developed in substance, logically organized, and well edited. The deserving special mention are: “Mechanical ventilation” by Khosrow Behbehani, and “Essentials of anaesthesia” by A. William Paulsen. The second volume ends with the article entitled “Applications of virtual” by Eric Rosow and Joseph Adam. The third volume is devoted to biosignals and biomedical imaging. The four chapters in this volume focus on signal and biomedical image processing, infrared radiation and computer science in biomedical engineering. These chapters contain numerous articles that are at different levels of detail. For example the chapter on signal processing includes articles on spectral analysis and data acquisition. The chapter on biomedical imaging contains articles devoted to imaging in a variety of biomedical devices. The last article entitled “Medical image search” by Thomas Deserno presents an overview of medical image repositories. The most extensive chapter is on infrared imaging. The most important articles include: “Quantitative active dynamic thermal IR-imaging and thermal tomography in medical diagnostics” by Antoni Nowakowski and others. These articles focus on dynamic thermography and technical limitations of its use. The last chapter describes software used in medicine and e-health services, both describing the software and its basic capabilities and scope. The fourth volume has 93 articles devoted to molecular, cellular and tissue engineering. In his brief introduction, Michael M. Domach outlines the scope and organization of the individual chapters. “Chapter 1 provides some historical background and basic working knowledge…. Chapter 2 is a new contribution that provides an up-to-date look at interactions and the means used to explore and quantify their nature…. Chapter 3 is a new contribution that covers DNA isolation and sequencing. Chapters 4 and 5 present applications in microbial and animal cell systems. Additions in cultivation techniques and long pathway engineering can be found in Chapter 4. RNA silencing methods and examples have been added to update Chapter 5. Chapter 6 is a new contribution that illustrates one medical impact of molecular biology.” The first article, by Nathan R. Domagalski and Michael M. Domach, entitled “Historical perspective and basics of molecular biology” describes the history of molecular biology showing the individual achievements of scientists in recent years. Other interesting articles include “Expression in mammalian cells” and “Tissue engineering bioreactors”. Strengths of the book This handbook is the largest (for comparison see [1–6]) and most recent summaries of progress in biomedical engineering that currently exists, and it spans the full range of this discipline: fundamentals, biomaterials [7], signals and biomedical images [8, 9] and molecular biology. Some of the authors are the most cited and prolific investigators in the field. All of the articles are at a high level in terms of content, logically arranged, and well edited. These articles are often mini-reviews that provide overview of selected topics. They are generally summaries of material but not critical reviews of research topics. This massive undertaking shows the strong commitment by the editors and publisher (CRC Press) to a successful project, which this certainly is. The handbook will be valuable for educating students in biomedical engineering, and also for physicians and other engineers who come into contact with medicine for their future work. It provides a valuable history of engineering in biology and medicine, and an overview of new advances in the field. Compared to earlier editions, the new material on ethics, medical imaging and molecular biology are particularly welcome. Apart from the hardback volumes (which are sold individually), the availability online will expand its reach. Individual articles can be purchased individually (for $20 per article). Weaknesses of the book Three aspects require some discussion: the purpose, proportion between the chapters and selection of authors and topics. Purpose. In the preface the authors state that “It can serve as an excellent textbook for students in areas where traditional textbooks have not yet been developed and as an excellent review of the major areas of activity in each biomedical engineering sub-discipline, such as biomechanics, biomaterials, bioinstrumentation, medical imaging, and so on.”. Certainly this handbook is far too much to ask a student to buy, and at a cost of 254GBP ($400) would be the reach of most students in any event. It is also too expensive for many faculty or biomedical engineers themselves; in much of the European Union it costs the equivalent of about half of the minimum monthly salary. However, many libraries, even in less affluent countries, purchase online access to this and other such handbooks. For educational purposes the handbook could be a valuable resource and instructors might consider using individual articles to supplement their own courses. For institutions that have not purchased online access, the question arises how to legally distribute this material to students. The lack of the usual didactic material found in textbooks (end of chapter problems, worked examples, callouts with clarifications, online videos and other material) also limit the usefulness of the handbook as a (primary) textbook. On the other hand, the level of material in most chapters is comparable to that taught in most graduate level courses in the field and should be accessible to advanced students. Proportion my second comment concerns what often seems to be a lack of balance in the volume compared to the real world of medicine and also to biomedical engineering as it is currently practiced. For example, the third volume contains an entire section, with 33 chapters on infrared imaging. This does not match hierarchically the other chapters on signal and image processing—the previous sections in this volume are on biomedical sensors and medical imaging, far broader topics. Secondly, on might ask what is so special about infrared imaging to justify 33 chapters on the topic. Why is infrared imaging more important than magnetic resonance imaging (MRI, four chapters) or ultrasound or mammography (one chapter each)? MRI is one of the most important imaging technologies to have ever been developed. It is used with countless patients on a routine basis, and is also a fundamental tool in many research programs. Infrared imaging is still a niche field with limited application in medical practice [10]. I assume that the disproportionate attention to infrared imaging reflects the interests of the authors who agreed to write chapters, and the enthusiasm of the section editor. However, the relative attention to different topics in this handbook may not be a reliable guide to the importance of particular topics in present medicine. As might be expected, the individual chapters vary in levels of detail and approach. An example is the chapter on analysis and image processing in volume III (Biomedical signals, imaging and informatics). In one group of articles in this section binarization and its consequences are discussed in detail. In other articles, it is ignored and marginalized because of its simplicity. The same basic information is duplicated in some articles. This begs the question to what extent, if at all, the editors worked with the authors to plan the detailed contents of their articles. Given the massive scale of the handbook, it may simply not have been practical to coordinate contents very precisely. Author and topic selection I looked at the citation records of the authors, and the countries in which they worked. A few of the authors are very highly cited, others had much lower visible impact in the field as judged by citation analysis. That has little directly to do with the quality of the chapters. However, a low citation count may mean that the author is working in a very small subspecialty and his/her chapter, however very well prepared, might not reflect a major effort within the field. That appears to be the case with several chapters on infrared imaging of the breast, for example. In my opinion, the editors should pay more attention to providing a balance of topics in the next edition that more accurately reflects the importance of subfields to medicine at large. One way to do this is to recruit well-known authors with high citation counts to write individual chapters. Lack of supplementary material. It would be extremely valuable to enrich the articles with supplementary online materials such as videos, presentations etc. These would enhance the usefulness of the book for education, and also allow ordinary readers to gain a better and easier understanding of the material. These comments should be taken as recommendations and suggestions for future editions. They should not be read as contradicting my overall very good opinion about this handbook. Conclusion This handbook is one of the most interesting and most complete and largest books in the field of biomedical engineering. Apart from the basic information in the field of biomechanics and biomedical devices for signal and biomedical image processing, it also covers the area of molecular biology. It will be a valuable resource for advanced undergraduate students and doctoral students as well as engineers and physicians during specialization. It is continually updated, bringing together articles written by the most famous scientists in the world. Certainly, this book should be available, either in hardcopy in libraries or online, to every student or professional involved in biomedical engineering. Whether an individual scientist should purchase a personal copy of this pricey handbook is a different matter. Acknowledgements No outside funding was received for this study. The author would like to thank Dr Alexander Lamża of the Department of Biomedical Computer Systems, University of Silesia for preparing the biometric data of all authors of the book under review. Competing interests The author declares that he has no competing interests. ==== Refs References 1. Hudak R Penhaker M Majernik J Biomedical engineering—technical applications in medicine 2012 Rijeka InTech 2. Ghista DN Biomedical science, engineering and technology 2012 Rijeka InTech 3. Fazel-Rezai R Biomedical engineering—from theory to applications 2011 Rijeka InTech 4. Serra PA, editor. Advances in bioengineering; 2015. 5. Gargiulo GD McEwan A Advanced biomedical engineering 2011 Rijeka InTech 6. Komorowska MA Olsztynska-Janus S Biomedical engineering, trends, research and technologies 2011 Rijeka InTech 7. Laskovski AN Biomedical engineering, trends in materials science 2011 Rijeka InTech 8. Gonzalez RC Woods RE Digital image processing 2007 3 New Jersey Prentice Hall: Pearson Education Inc 9. Diakides M Bronzino JD Peterson DR Medical infrared imaging: principles and practices 2013 1 USA CRC Press 10. Fitzgerald A Berentson-Shaw J Thermography as a screening and diagnostic tool: a systematic review NZ Med J 2012 125 1351 80 91

PMC005xxxxxx/PMC5002144.txt

==== Front J CheminformJ CheminformJournal of Cheminformatics1758-2946Springer International Publishing Cham 15510.1186/s13321-016-0155-1SoftwareMolmil: a molecular viewer for the PDB and beyond Bekker Gert-Jan gertjan.bekker@protein.osaka-u.ac.jp 12Nakamura Haruki harukin@protein.osaka-u.ac.jp 1Kinjo Akira R. akinjo@protein.osaka-u.ac.jp 11 Laboratory of Protein Informatics, Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871 Japan 2 Graduate School of Frontier Biosciences, Osaka University, Suita, Osaka 565-0871 Japan 26 8 2016 26 8 2016 2016 8 1 4213 4 2016 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.We have developed a new platform-independent web-based molecular viewer using JavaScript and WebGL. The molecular viewer, Molmil, has been integrated into several services offered by Protein Data Bank Japan and can be easily extended with new functionality by third party developers. Furthermore, the viewer can be used to load files in various formats from the user’s local hard drive without uploading the data to a server. Molmil is available for all platforms supporting WebGL (e.g. Windows, Linux, iOS, Android) from http://gjbekker.github.io/molmil/. The source code is available at http://github.com/gjbekker/molmil under the LGPLv3 licence. Electronic supplementary material The online version of this article (doi:10.1186/s13321-016-0155-1) contains supplementary material, which is available to authorized users. Keywords WebGLProtein structure visualizationMolecular dynamicshttp://dx.doi.org/10.13039/501100004696National Bioscience Database Centerissue-copyright-statement© The Author(s) 2016 ==== Body Background Molecular viewers are a vital tool for our understanding of protein structures and functions. The shift from regular desktop platforms such as Windows, Mac OSX and Linux, to mobile platforms such as iOS and Android in the last half-decade, however, prevents traditional online molecular viewers such as PDBj’s previously developed jV [1] and the popular Jmol [2] from running on these new platforms as these platforms do not support Java Applets. For mobile platforms a native application (i.e. an application specifically designed and optimized for each of these platforms) can be created and distributed via their respective application stores. However, with new platforms on the horizon, or already available, in addition to the already established desktop platforms, it would be a tedious and inefficient job to make a molecular viewer available on all platforms, current and future. In the same half-decade however, web browsers have also made significant advances, particularly with their JavaScript engines. Further advances in the form of CSS3 [3] and HTML5 [4] have made the web an interesting platform for developers. Mobile platforms also offer these advanced web browsers, while some newer mobile platforms such as Mozilla’s Firefox OS [5] and Ubuntu Mobile [6] are oriented around HTML5 applications. Since these advanced web browsers are available on a wide range of platforms, it seems to be the perfect platform to develop a molecular viewer for. One of the earliest attempts to switch to a Java independent molecular web viewer was done by the developers of the popular molecular viewer Jmol who developed an alternate molecular viewer named JSmol [7] to run directly within the web browser. They simply used a program known as java2script to convert Jmol’s Java code into JavaScript. JSmol however runs noticeably slower compared to Jmol for relatively small protein structures and quickly becomes unusable for larger structures. Desktop applications such as Chimera [8], VMD [9], PyMol [10] and Yasara [11] make use of GPU based hardware acceleration using OpenGL [12]. By making use of OpenGL, one can leverage the immense power of GPUs to accelerate these molecular viewers. GPUs have traditionally been designed for this purpose; i.e. to accelerate the generation of 3D images, but more recently GPUs have also been re-purposed for general purpose calculations in e.g. Molecular Dynamics [13, 14]. So wouldn’t it be great if this power were somehow also available inside web browsers? Enter WebGL. WebGL was developed by Khronos [15] to allow JavaScript applications running in the web browser to take advantage of OpenGL ES 2.0 [16] compatible GPUs which had been specifically designed for mobile devices. While WebGL has been available for several years in Chrome [17] and Firefox [18], WebGL support was only recently added to Microsoft’s Internet Explorer and Apple’s Safari, including iOS. We have developed a new web based molecular viewer, Molmil, which can take advantage of GPU hardware acceleration using WebGL. It runs on a wide range of platforms such as Windows, Linux, Mac OSX, Android and iOS. Finally, we have embedded Molmil within various services offered by Protein Data Bank Japan [1, 19–21] to demonstrate its capabilities. Implementation and capabilities Molmil has been designed as a light-weight and full-featured viewer for the PDB. As such, Molmil can load legacy PDB flat files, PDBx/mmCIF [22] and PDBML [23] formatted files. Molmil can also load a custom format which we call PDBx/mmJSON, which is a JSON [24] version of the PDBx/mmCIF data. Other formats which Molmil supports are GRO, MOL2, MDL, CCP4 [25] (for electron density maps and EM data), MyPresto’s trajectory format [14, 26], Gromacs’ TRR and XTC trajectory formats [13] and our own developed MPBF polygon format which we are using for our eF-site service [1] for large structures. Users can also load these files from their local hard drive. The PDBx/mmJSON (mmJSON for short) format was specifically developed for Molmil. In a nutshell, it is a JSON representation of the mmCIF format, made by translating the data structure based on the STAR syntax [27], into a series of associative arrays, standard arrays, integers, floats and null values. A major advantage of this format is that it can be directly interpreted by modern browsers, that is, a typed data structure is available once the JSON data has been parsed by a browser. Such is not the case for PDBML (XML) or mmCIF formats. In addition, the file size is generally smaller in the mmJSON format than in PDBML or mmCIF formats. An analysis showed that the compressed mmJSON is on average approximately 33 or 56 % smaller than a compressed mmCIF or PDBML formatted file, respectively, making it more suitable for web deployment. For large structures another derivative of mmJSON is currently under development to only include the minimal amount of information in the file required to display the backbone structure. This mmJSON variant is not yet offered as a service by PDBj, but can be previewed by using Molmil: http://gjbekker.github.io/molmil/#molmil.loadPDBlite(‘3j3q’). Implementation details regarding the parsers used to load the mmCIF data, as well as information regarding the mmJSON format and its relation with the mmCIF format is given on its Github page at http://github.com/gjbekker/cif-parsers. Displaying large entries One of the design goals of Molmil was to create a molecular viewer which can produce high quality images suitable for publications. On the other hand, it should also be able to scale to very large structures available in the Protein Data Bank (PDB) consisting of hundreds of thousands to millions of atoms. By default WebGL can only display a limited amount of polygons. However, by using an extension to WebGL (OES_element_index_uint) which is available on all modern platforms, it becomes possible to efficiently render very large or highly detailed structures. Molmil uses this extension to build high quality geometry for small and medium sized structures and dynamically scale down the quality as the size of the structure increases to gigantic proportions (such as the HIV-1 capsid; 3j3q [28], see Fig. 1). Note that to be able to load these gigantic structures, an adequate amount of memory is still required, which is often not available on smartphones, tablets and older systems. To enable high quality lighting even when using polygon models of low detail for these gigantic structures, Molmil uses tuned Phong shading [29] which can accurately calculate the lighting even for simplistic polygon models.Fig. 1 Rendering of 3j3q using Molmil. 3j3q [28] is the largest PDB entry to date (with more than 2.4 million atoms) and can be visualized using Molmil. Left: cartoon rendering with group colouring. Right: rendering of the molecular surface, coloured by electrostatic potential and hydrophobicity generated by eF-site [1] Rendering and colouring modes Molmil supports several rendering and colouring modes which can be selected using a point-and-click based menu. In the top-right corner of Molmil’s canvas the structures menu can be enabled (see also Fig. 2). This structures menu lists all the loaded structures in the current canvas. Via a hierarchical tree selection system chains and residues belonging to the structure can be explored. Right clicking on the structures, chains or residues displays a context menu via which users can change the display and colour modes. Double clicking on the residues causes Molmil to focus and jump to the selected residue. Furthermore, right-clicking on the atoms or cartoon rendering within the canvas also enables users to interactively modify the rendering and colour mode of the selected atom or the residue/chain the atom belongs to. These controls are reminiscent from Yasara [11], which uses a similar hierarchical structures menu and context menu to modify the display and colour modes of the loaded structures. Molmil currently supports VDW, ball-and-stick, stick, Cα-trace, tube, cartoon, rocket and coarse surface representations. Structures can be coloured by their secondary structure elements, CPK, group (blue-to-red gradient), chain, ABEGO [30] or custom colour assignments. Views can be outputted as high quality PNG images, suitable for publication without requiring additional plugins. An overview of Molmil’s various rendering modes can be found in Additional file 1: Appendix I.Fig. 2 Screenshot of Molmil loading the url: http://gjbekker.github.io/molmil/#molmil.loadPDB(‘1crn’). Indicated are the Main menu on the left which can be used to load files, save PNG images, modify settings and play animations if available. The structures menu on the right can be shown by clicking on the indicated button. The context menu which can be displayed by right clicking on any structures, chains or residues/ligands listed within the structures menu. The command line interface which can be shown by clicking on the indicated button Playing animations It is also possible to load PDB files with multiple models and play them as animations, by using the “Animation” control panel which is accessible from Molmil’s main menu (see Fig. 2). PDBj’s Promode Elastic service [20] uses this functionality. It is also possible to use this animation functionality using alternative formats to e.g. play Molecular Dynamics trajectories. Currently myPresto’s [26] trajectory format and Gromacs’s [13] TRR and XTC trajectory formats are supported by Molmil and support for other formats can be easily added. Although web browsers can be used to produce PNG images by Molmil, they cannot make MP4 movies. However, by using a server-side REST service it is possible to stitch the images of each frame in the trajectory together to build a movie. For obvious reasons (potential abuse) we have not made this functionality available as a public REST service, but users can obtain this application from Molmil’s Github page which they can then run on their own workstation so that their own machine will then build the movie. Using this tool, Molmil can e.g. be used to produce high quality MP4 videos of molecular simulation trajectories. Command-line interface Molmil also has an embedded command line which can be used to manually perform actions. The command line is capable of executing arbitrary JavaScript code and has access to Molmil’s application program interface (API), which can also be used by third party developers to embed and extend Molmil within their own web page/app. The command line can also be used to perform simple tasks such as selecting chains, residues or atoms and changing the display and colouring of them. Third party developers can also add support for their own custom file formats, or add additional functionality. The embedded viewers on the PDBj website have been implemented in this manner. Details and examples of the Molmil API are described on Molmil’s Github page http://github.com/gjbekker/molmil/wiki. The command line can be accessed via the “<” symbol in the lower left corner of the canvas (see also Fig. 2). Finally, the commands can also be embedded within the URL for easy sharing via email, web pages and social media (see Fig. 2 for a simple example URL). Molmil availability Molmil has seen several major revisions since its initial release in December 2013. Molmil Beta featured a prototype WebGL viewer which was implemented directly into PDBj’s web application. Molmil 0.9 was released in September 2014 and had all the basic functionalities a molecular viewer needs such as reading various formats and supporting multiple display and colour modes via a point-and-click based menu. The latest major version of Molmil, Molmil 1.0, was released in October 2015 which focused on optimizations and extending the feature set of Molmil and also added a powerful JavaScript based command line interface. Molmil has been integrated in many of PDBj’s services [21], including: Mine PDB Explorer Asymmetric Unit viewer; Mine PDB Explorer Biological Unit viewer; EDMap viewer (for electron densities); Sequence Navigator; Structure Navigator; Promode Elastic (for viewing protein mobility analysed via Normal Mode Analysis) and Chemie (PDBj’s new chemical component service). More recently, the eF-site [1] service (for viewing the electrostatic surface of proteins) was also updated with an integrated Molmil viewer and the Biological Unit viewer was also upgraded. Furthermore, high quality PNG images of all the released PDB entries on the PDBj website have also been generated by Molmil in an automated manner. More information regarding Molmil as well as use-cases and command-line examples can be found at http://github.com/gjbekker/molmil/wiki. Discussion Molmil requires the browser to support WebGL. Furthermore, for optimal use, the OES_element_index_uint must also be supported. When loading very large structures such as 3j3q, there are additional requirements to the user’s software and hardware. Since Molmil requires at least 6 GB of memory to load the structure into memory, the user’s device must be equipped with 8 GB of memory. Furthermore, to enable Molmil to make use of this large amount of memory, a 64-bit browser and operating system are required. Finally, due to limitations of Chrome’s JavaScript engine, even Chrome’s 64-bits version is incapable of loading 3j3q. Mozilla’s Firefox and Apple’s Safari however have no such problem with the 64-bit versions. Note that it will still take tens of seconds to several minutes depending on the web browser, the user’s hardware such as the processor and the user’s internet connection since 3j3q’s compressed data file is still about 40 MB large. The implementation at PDBj’s web application will automatically detect the user’s software and hardware environment and will display a link to the entry’s embedded Molmil page if the entry is compatible with the user’s system. All regular entries and most large structure entries however have no problems loading on low-end devices, as long as WebGL is supported. Although Molmil exposes some of PyMol’s commands via Molmil’s command line, support is currently limited to a small subset of commands for selection, styling and colouring of structures. Also, although Molmil can be used to load CCP4 files for electron densities and generate coarse surfaces for proteins, high quality surfaces such as MSMS and solvent-accessible surfaces are currently not yet supported, partly because of the high computational requirements involved in generating highly detailed surfaces. Since Molmil’s initial release in 2013, other WebGL-based molecular viewers have also emerged. E.g. JSmol (JSmol/WebGL) has been extended with a basic WebGL based viewer and PV [31] was developed by SWISS-MODEL and is currently also being used by the RCSB-PDB. Other WebGL-based molecular viewers are listed in Additional file 1: Appendix II. Molmil however offers high quality graphics due to using highly detailed polygon models when possible, while it can scale up to display very large structures by dynamically reducing the quality to render these large structures in combination with a finely tuned Phong shader for smooth and realistic shading even for lower detailed models. Furthermore, Molmil can also be used to easily load files from the users’ hard drive, including molecular simulation trajectories, and comes with a readily accessible command line interface, which is also programmable by embedding the commands in the URL, as shown in e.g. Fig. 2. Conclusions We have designed a versatile, high performance, high quality molecular viewer for the web. We have also deployed the molecular viewer on the PDBj web application as part of various services to aid users as well as demonstrate some of the capabilities of Molmil. The source code is available on GitHub as well as documentation on how to use and/or deploy Molmil on your own website. Questions or requests can be made either via Molmil’s GitHub page or via our contact page http://pdbj.org/contact. Availability and requirements Project name: Molmil; Project home page: http://github.com/gjbekker/molmil; Operating system(s): Platform independent; Programming language: JavaScript, CSS3, HTML5, WebGL, GLSL; Other requirements: WebGL; Licence: GNU LGPLv3; Restrictions for use by non-academics: None. Additional file 10.1186/s13321-016-0155-1 Appendices; Example renderings using Molmil & Overview of alternative WebGL based molecular viewers. Abbreviations PDBProtein Data Bank PDBjProtein Data Bank Japan CSS3Cascade Style Sheets version 3 HTML5hyper text markup language version 5 GPUgraphics processing unit RESTrepresentational state transfer EMelectron microscopy MPBFmolmil polygon binary format APIapplication programming interface VDWVan Der Waals PNGportable network graphics XMLextendible meta language Authors’ contributions HN and GB conceived the idea. GB designed and coded the software, and led the writing of the manuscript. ARK helped write the manuscript and conceived the name of the program. HN and ARK contributed to the further development of the idea and in giving suggestions for improvements to the program. All authors read and approved the manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Kinoshita K Nakamura H eF-site and PDBjViewer: database and viewer for protein functional sites Bioinformatics 2004 20 8 1329 1330 10.1093/bioinformatics/bth073 14871866 2. Hanson RM Jmol—a paradigm shift in crystallographic visualization J Appl Crystallogr Int Union Crystallogr 2010 43 5 1250 1260 3. CSS [Internet]. [Cited 2016 Mar 14]. http://www.w3.org/TR/CSS/ 4. HTML5 [Internet]. [Cited 2016 Mar 14]. http://www.w3.org/TR/html5/ 5. Firefox OS [Internet]. [Cited 2016 Mar 14]. http://www.mozilla.org/en-US/firefox/os/ 6. Ubuntu on phones [Internet]. [Cited 2016 Mar 14]. http://www.ubuntu.com/phone 7. Hanson RM Prilusky J Renjian Z Nakane T Sussman JL JSmol and the next-generation Web-based representation of 3D molecular structure as applied to proteopedia Isr J Chem 2013 53 3–4 207 216 10.1002/ijch.201300024 8. Pettersen EF Goddard TD Huang CC Couch GS Greenblatt DM Meng EC UCSF Chimera–a visualization system for exploratory research and analysis J Comput Chem 2004 25 13 1605 1612 10.1002/jcc.20084 15264254 9. Humphrey W Dalke A Schulten K VMD—visual molecular dynamics J Mol Graph 1996 14 33 38 10.1016/0263-7855(96)00018-5 8744570 10. Schrödinger L (2010) The PyMOL molecular graphics system, version ~1.3r1 11. Krieger E Vriend G YASARA view—molecular graphics for all devices—from smartphones to workstations Bioinformatics 2014 30 20 2981 2982 10.1093/bioinformatics/btu426 24996895 12. OpenGL—The industry standard for high performance graphics [Internet]. [Cited 2016 Mar 14]. http://www.opengl.org/ 13. Abraham MJ Murtola T Schulz R Páll S Smith JC Hess B GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers SoftwareX 2015 1 19 25 10.1016/j.softx.2015.06.001 14. Mashimo T Fukunishi Y Kamiya N Takano Y Fukuda I Nakamura H Molecular dynamics simulations accelerated by GPU for biological macromolecules with a non-Ewald scheme for electrostatic interactions J Chem Theory Comput 2013 9 12 5599 5609 10.1021/ct400342e 26592294 15. WebGL—OpenGL ES 2.0 for the Web [Internet]. [Cited 2015 Dec 14]. http://www.khronos.org/webgl/ 16. OpenGL ES 2_X—The standard for embedded accelerated 3D graphics [Internet]. [Cited 2016 Mar 14]. http://www.khronos.org/opengles/2_X/ 17. Chrome [Internet]. [Cited 2016 Mar 14]. http://www.google.com/chrome/ 18. Firefox [Internet]. [Cited 2016 Mar 14]. http://www.mozilla.org/en-US/firefox/new/ 19. Protein Data Bank Japan—PDB Japan—PDBj [Internet]. [Cited 2016 Mar 14]. http://pdbj.org/ 20. Wako H Endo S Normal mode analysis based on an elastic network model for biomolecules in the Protein Data Bank, which uses dihedral angles as independent variables Comput Biol Chem 2013 44 22 30 10.1016/j.compbiolchem.2013.02.006 23537565 21. Kinjo AR Suzuki H Yamashita R Ikegawa Y Kudou T Igarashi R Protein Data Bank Japan (PDBj): maintaining a structural data archive and resource description framework format Nucl Acids Res 2012 40 Database issue D453 D460 10.1093/nar/gkr811 21976737 22. Fitzgerald PMD Westbrook JD Bourne PE McMahon B Watenpaugh KD Berman HM Hall SR McMahon B The macromolecular dictionary (mmCIF) International tables for crystallography 2006 Chester International Union of Crystallography 295 443 23. Westbrook J Ito N Nakamura H Henrick K Berman HM PDBML: the representation of archival macromolecular structure data in XML Bioinformatics 2005 21 7 988 992 10.1093/bioinformatics/bti082 15509603 24. JSON [Internet]. [Cited 2016 Mar 14]. http://www.json.org/ 25. Winn MD Ballard CC Cowtan KD Dodson EJ Emsley P Evans PR Overview of the CCP4 suite and current developments Acta Crystallogr D Biol Crystallogr Int Union Crystallogr 2011 67 Pt 4 235 242 10.1107/S0907444910045749 26. Morikami K Nakai T Kidera A Saito M Nakamura H Presto (protein engineering simulator): a vectorized molecular mechanics program for biopolymers Comput Chem 1992 16 3 243 248 10.1016/0097-8485(92)80010-W 27. Hall SR The STAR file: a new format for electronic data transfer and archiving J Chem Inf Model 1991 31 2 326 333 10.1021/ci00002a020 28. Zhao G Perilla JR Yufenyuy EL Meng X Chen B Ning J Mature HIV-1 capsid structure by cryo-electron microscopy and all-atom molecular dynamics Nature 2013 497 7451 643 646 10.1038/nature12162 23719463 29. Phong BT Illumination for computer generated pictures Commun ACM 1975 18 6 311 317 10.1145/360825.360839 30. Lin Y-R Koga N Tatsumi-Koga R Liu G Clouser AF Montelione GT Control over overall shape and size in de novo designed proteins Proc Natl Acad Sci 2015 112 40 E5478 E5485 10.1073/pnas.1509508112 26396255 31. Biasini M (2015) PV [Internet]. http://github.com/biasmv/pv

PMC005xxxxxx/PMC5002145.txt

==== Front BMC GenetBMC GenetBMC Genetics1471-2156BioMed Central London 43310.1186/s12863-016-0433-yResearch ArticleGenome sequencing reveals a splice donor site mutation in the SNX14 gene associated with a novel cerebellar cortical degeneration in the Hungarian Vizsla dog breed http://orcid.org/0000-0001-7851-670XFenn Joe (+44)1707 663666jfenn@rvc.ac.uk 1Boursnell Mike mike.boursnell@aht.org.uk 3Hitti Rebekkah J. rebekkah.hitti@aht.org.uk 3Jenkins Christopher A. chris.jenkins@aht.org.uk 3Terry Rebecca L. rterry@rvc.ac.uk 2Priestnall Simon L. spriestnall@rvc.ac.uk 2Kenny Patrick J. pkenny@rvc.ac.uk 1Mellersh Cathryn S. cathryn.mellersh@aht.org.uk 3Forman Oliver P. oliver.forman@aht.org.uk 31 Department of Clinical Science and Services, Royal Veterinary College, Hawkshead Lane, North Mymms, Hertfordshire, AL9 7TA UK 2 Department of Pathology and Pathogen Biology, Royal Veterinary College, Hawkshead Lane, North Mymms, Hertfordshire, AL9 7TA UK 3 Kennel Club Genetics Centre, Animal Health Trust, Kentford, Newmarket, Suffolk, CB8 7UU UK 26 8 2016 26 8 2016 2016 17 1 1231 3 2016 24 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 Cerebellar cortical degeneration (CCD) is an increasingly recognised neurodegenerative disease process affecting many dog breeds. Typical presentation consists of a progressive cerebellar ataxia, with a variable age at onset and rate of progression between different breeds. Cerebellar histopathological findings typically consist of primary Purkinje neuronal degeneration and loss, with variable secondary depletion of the granular and molecular cell layers. Causative genes have been identified associated with CCD in several breeds, allowing screening for selective breeding to reduce the prevalence of these conditions. There have been no previous reports of CCD in Hungarian Vizslas. Results Two full-sibling Hungarian Vizsla puppies from a litter of nine presented with a history of progressive ataxia, starting around three months of age. Clinical signs included marked hypermetric and dysmetric ataxia, truncal sway, intention tremors and absent menace responses, with positional horizontal nystagmus in one dog. Routine diagnostic investigations were unremarkable, and magnetic resonance imaging performed in one dog revealed mild craniodorsal cerebellar sulci widening, supportive of cerebellar atrophy. Owners of both dogs elected for euthanasia shortly after the onset of signs. Histopathological examination revealed primary Purkinje neuron loss consistent with CCD. Whole genome sequencing was used to successfully identify a disease-associated splice donor site variant in the sorting nexin 14 gene (SNX14) as a strong causative candidate. An altered SNX14 splicing pattern for a CCD case was demonstrated by RNA analysis, and no SNX14 protein could be detected in CCD case cerebellum by western blotting. SNX14 is involved in maintaining normal neuronal excitability and synaptic transmission, and a mutation has recently been found to cause autosomal recessive cerebellar ataxia and intellectual disability syndrome in humans. Genetic screening of 133 unaffected Hungarian Vizslas revealed the presence of three heterozygotes, supporting the presence of carriers in the wider population. Conclusions This is the first report of CCD in Hungarian Vizsla dogs and identifies a highly associated splice donor site mutation in SNX14, with an autosomal recessive mode of inheritance suspected. Electronic supplementary material The online version of this article (doi:10.1186/s12863-016-0433-y) contains supplementary material, which is available to authorized users. Keywords Cerebellar cortical degenerationCerebellar abiotrophyHungarian Vizsla dogGenome sequencingissue-copyright-statement© The Author(s) 2015 ==== Body Background Cerebellar cortical degeneration (CCD) has been reported in several different dog breeds [1–15], and is also commonly referred to as cerebellar abiotrophy. CCD typically causes slowly progressive clinical signs related to diffuse cerebellar dysfunction, including a dysmetric and hypermetric cerebellar ataxia, intention tremors, menace response deficits and signs of central vestibular dysfunction, including nystagmus and loss of balance [12, 16]. Age of onset of clinical signs and the rate of progression vary between breeds, ranging from neonatal, as reported in Beagles and Rhodesian Ridgebacks [3, 13], and early onset from weeks to months of age in the majority of breeds [2, 6, 8, 9, 14, 16], up to adult-onset at months to years of age with a slower progression, seen in Old English Sheepdogs, Gordon Setters and Scottish Terriers [12, 15, 17]. Histopathological examination most often reveals a primary diffuse loss of Purkinje neurons in the cerebellar cortex, in some cases with milder secondary neuronal depletion of the granular and molecular cell layers [7, 11, 13, 14, 16]. Less commonly there have also been reports of primary granular cell loss with sparing of the Purkinje neuron layer in some dog breeds, including the Border Collie and Bavarian Mountain Dog [6, 8], as well as an immune-mediated granuloprival degeneration in the Coton de Tulear [18]. Neurodegenerative lesions in some dog breeds with CCD have been reported to also affect areas of the central nervous system outside the cerebellum, as seen in the Kerry Blue Terrier and Chinese-Crested Dog [19, 20]. An autosomal recessive mode of inheritance for CCD is suspected in most breeds, and putative causative genetic mutations have been identified for CCD in several of these [13–15, 21]. A mutation in the GRM1 gene was identified as the cause of Bandera’s neonatal cerebellar ataxia in the Coton de Tulear, which represents a functional disturbance with minimal histopathological evidence of CCD [21]. A recent study used genome-wide mRNA sequencing to identify a causative mutation in the β-III spectrin cytoskeletal protein gene, SPTBN2, in Beagle dogs with neonatal CCD [13]. Genome-wide association studies also recently identified a mutation in the SEL1L gene, encoding a component of the endoplasmic reticulum-associated protein degradation machinery, in Finnish Hounds with CCD [14], as well as a mutation associated with adult-onset CCD in Gordon Setters and Old English Sheepdogs in the autophagy gene, RAB24 [15]. This variation in clinical presentation, histopathology and molecular findings suggests that CCD represents a heterogeneous disease process with varied underlying genetic aetiologies in different dog breeds. Following the diagnosis of CCD in two Hungarian Vizsla (HV) littermates, the aim of this study was to describe the clinical and histopathological features of CCD previously unreported in this breed and to interrogate the genome of an affected dog for potential causal mutations. Methods Clinical investigation Two full-sibling HVs (Case 1 and 2) from a litter of nine were presented separately to the Royal Veterinary College (RVC), University of London for further investigation of progressive ataxia and tremors. Clinical examinations and diagnostic tests were performed at the RVC, with videos obtained of both examinations. Whole blood was taken ante-mortem from both dogs and stored for future molecular investigations. The dogs were both euthanased for welfare reasons at their owners’ request, and submitted for post mortem examination. Histopathological examination of the brain tissue was performed on both dogs. Samples of cerebellar tissue from the male sibling (Case 1) were obtained and stored in RNAlater medium to allow subsequent RNA extraction. Relevant clinical history was obtained where possible from the breeder regarding the unaffected littermates, dam and sire. Full post-mortem examination was performed in both dogs, with samples of brain tissue preserved in 10 % neutral buffered formalin followed by embedding in paraffin wax. 4 μm thick sections were cut on a microtome (Leica Jung RM2035) and stained with haematoxylin and eosin. Slides were evaluated by a board-certified veterinary pathologist (SLP), using light microscopy. Molecular investigation DNA was extracted from whole blood using the Nucleon BACC2 DNA extraction kit (GE Healthcare Life Sciences). RNA was extracted from canine cerebellum using the Qiagen RNeasy midi kit (Qiagen), with on column DNase treatment. Whole genome resequencing of a single HV with CCD (Case 2) was outsourced to the Wellcome Trust Centre for Human Genetics, University of Oxford. Paired end sequencing with a read length of 100 bp read generated a dataset of approximately 75 Gb, resulting in ~30× coverage of the genome. Reads were aligned to the CanFam3.1 canine genome build using Burrows-Wheeler Aligner (BWA) [22] and variants called using Genome Analysis Toolkit (GATK) [23]. Variants were filtered based on consequence predictions created using SNPeff [24], and by comparison with variant calls generated for 13 unrelated genome sequences of other dog breeds. A library for RNAseq was prepared using total RNA extracted from Case 1. Messenger RNA was isolated using Sera-Mag oligo(dT) magnetic beads, and libraries prepared using the NEBNext® mRNA Sample Prep Master Mix Set 1. The final RNA-seq library was quantified by qPCR using the Kapa library quantification kit (Kapa Biosystems). Paired-end sequencing of 80 bp reads was carried out on an Illumina MiSeq, producing an ~4Gb dataset. Reads were aligned to the canine reference genome (CanFam3.1) using BWA and further processed using GATK. Aligned reads were viewed using The Integrative Genomics Viewer (IGV). Genotyping of the CCD associated variant was performed using an allelic discrimination approach (Taqman), using an ABI StepOne real-time thermal cycler. Primer sequences were as follows: SNX14_F ATTACTGTATATTGTGATCTCAAAGAAATGCTAATCT; SNX14_R, AGCAAAAACGAGCAAAACAGACTTT; SNX14_V, VIC-ACATTCCAGGTATAATATT-NFQ; SNX14_M FAM-ACATTCCAGATATAATATT-NFQ. Reverse transcription PCR was carried out in two stages. Reverse transcription was performed using the Qiagen Quantitect cDNA synthesis kit. PCR was performed in 12 μl reactions consisting of 0.2 mM dNTPs (NEB), 1× PCR buffer (Qiagen), 0.5 μM of each primer (SNX14_x25F, CCGCTTGGTCTCACTCATAAC and SNX14_x29_R, CACAATGTCCAATAAAACATAAGTCA), 0.5 units HotStarTaq plus DNA polymerase (Qiagen), and template cDNA. Cycling parameters were 95 °C for 5 min, followed by 35 cycles of 95 °C for 30 s, 57 °C for 30 s and 72 °C for 60 s, and a final elongation stage of 72 °C for 10 min. Sanger sequencing was performed using BigDye v3.1 dye terminator chemistry. For western blotting, cerebellum tissue from the affected HV case and 6 unaffected control dogs (one each of the following breeds: Golden Retriever, Crossbreed, Beagle, Parson Russell Terrier, Great Dane and Labrador Retriever cross) were first homogenized in lysis radioimmunoprecipitation assay buffer (RIPA, Sigma-Aldrich, R0278). 5 μl (46–56 μg) of the total protein lysate was separated by 8 % SDS-polyacrylamide gel electrophoresis. Proteins were then transferred onto nitrocellulose membrane, and immunoblotting was performed using rabbit polyclonal anti-SNX14 antibody (1:500, Atlas Antibodies, HPA017639) and swine polyclonal anti-rabbit antibody conjugated with horseradish peroxidase (HRP) (1:1000, Daco, P0217). The protein bands were detected using Novex ECL (Thermo Fisher Scientific) HRP chemiluminescent substrate reagent kit (WP20005) with high performance chemiluminescence film (Amersham Hyperfilm ECL). To control for protein loading the same protein lysates were used in the same quantities for a separate western blot using mouse monoclonal anti-beta Actin antibody (1:1000, Abcam, mAbcam 8226) and goat polyclonal anti-mouse antibody conjugated with HRP (1:1000, Daco, P0447). Results Clinical investigation Cases 1 (male) and 2 (female) both presented aged four months old, having first demonstrated signs of ataxia between two and three months of age. Until this time they had developed the ability to walk at a similar rate to their littermates. By four months of age the ataxia had worsened and Case 1 had developed a marked head tremor, more obvious with intention. The dogs were otherwise bright and well, with normal behaviour reported by their owners and no other clinical signs reported. General physical examination was unremarkable on presentation to the RVC. On neurological examination of both dogs, mentation was alert and appropriate. They were both ambulatory with a markedly hypermetric and dysmetric generalised cerebellar ataxia, with an obvious intention tremor particularly when navigating their environment and excited, both of which were more exaggerated in Case 1. Paw positioning was intact, although hopping responses showed a delayed onset of limb protraction and exaggerated response in all limbs. Spinal reflexes were intact in all four limbs. Cranial nerve examination revealed absent menace responses bilaterally, with an intermittent positional horizontal nystagmus in Case 2 when her head was elevated. No hyperaesthesia was detected on spinal or cranial palpation. All other components of the neurological examination were within normal limits. These findings were consistent with a diffuse and symmetrical lesion affecting the cerebellum. Differential diagnoses included neurodegenerative disease, such as cerebellar cortical degeneration or lysosomal storage disorders, inflammatory/infectious disease, metabolic disease and neoplasia. Diagnostic investigations were performed for Case 1 but not Case 2, whose owners elected for euthanasia, with consent for post-mortem examination. Haematology, serum biochemistry and urinalysis in Case 1 revealed no significant abnormalities for a four month-old dog. Blood smear evaluation of leukocytes revealed no evidence of intracellular storage vacuoles. Cranial magnetic resonance imaging (MRI) revealed mildly increased width and depth of the craniodorsal cerebellar sulci, with the impression of a reduced volume of cerebellar grey matter tissue in this region (Fig. 1). The gross size of the cerebellum otherwise appeared largely normal, with a mid-sagittal cross-sectional brainstem to cerebellar area ratio of 81.1 % [25]. There were also mildly enlarged cerebral sulci and lateral ventricles, with the rest of the brain MRI otherwise unremarkable. Cerebrospinal fluid (CSF) sampled from the cerebellomedullary cistern was unremarkable, with low nucleated cellularity and protein. Based on the unremarkable blood tests, imaging findings and clinical history, a presumptive diagnosis of CCD was made. Following a deterioration in Case 1’s clinical signs in the following weeks, his owners also elected for euthanasia and consented to post-mortem examination. There were no similar clinical signs reported in the littermates or the parents of the affected HV siblings.Fig. 1 Cerebellar MRI and gross pathological specimen from an HV with CCD (Case 1). Midline sagittal T2-weighted MRI image (a), and fresh sagittal section (b), of the cerebellum of a 4-month old HV with CCD. Predominantly affecting the rostral and dorsal cerebellum, there is widening of the cerebellar sulci with increased cerebrospinal fluid (increased T2-weighted signal) between the folia and subjectively reduced cortical grey matter (asterisks). This is not evident in the caudoventral lobules, which appear grossly normal (arrows) On gross examination, in both dogs the cerebellum appeared subjectively small relative to total brain size, with bilaterally symmetrical atrophy of the cerebellar hemispheres and a relative prominence of the cerebellar vermis. On sagittal section the cerebellar folia were reduced in thickness in both dogs (Fig. 1). The resulting diffuse widening of the cerebellar sulci was most prominent in the dorsal and rostral vermis, with relative sparing of the caudoventral lobules (Fig. 1). Microscopically, in both dogs there was multifocally extensive loss of cerebellar Purkinje neurons, which was particularly severe in Case 1 (Fig. 2). Some Purkinje neurons remained in Case 2 however they were frequently degenerate (Fig. 3). In both dogs there was a marked depletion of the granular cell layer, with a diffuse mild to moderate gliosis, and atrophy of the molecular layer (Figs. 2 and 3). There were rare degenerate neurons and spheroid formation in the cerebellar nuclei of Case 1. No evidence of inflammation or neoplasia was seen and the gross and histopathological examination of both dogs was otherwise unremarkable, with no histological evidence of intracellular material consistent with a lysosomal storage product.Fig. 2 Histological findings in the cerebellar folia of an affected HV (Case 2, a) and a normal control dog (b). The affected HV (a) demonstrates marked loss of Purkinje neurons (P), with an associated secondary depletion and thinning of the granular cell layer (G). The unaffected control dog (b) cerebellar folium demonstrates the normal cellularity of the granular (G), Purkinje (P) and molecular (M) cell layers. (Haematoxylin and eosin stain, ×100 magnification) Fig. 3 Cerebellar Purkinje neuron degeneration in an affected HV case (Case 2). Higher magnification of the cerebellar cortex in an affected HV (a and b) reveals shrunken, angular and eosinophilic (degenerate) Purkinje neurons (arrows). Reduced cellularity of the granular cell layer (G) and thinning and pallor of the molecular cell layer (M) are also evident, compared to the densely populated granular cell layer and normal molecular cell layer in the unaffected control dog (c). (Haematoxylin and eosin stain, ×200 magnification) Genetic investigation The entire genome of a single HV with CCD (Case 2) was interrogated using a massively parallel sequencing approach. Paired end Illumina sequencing was performed to give a final average coverage of 30× on alignment to the canine genome sequence CanFam3.1. Variants identified in the HV were compared with variants identified for 13 additional canine breeds. Based on the hypothesis that the causal variant for CCD was rare and unique to the HV dog breed, variants were filtered to only include ones that were homozygous in the HV, and homozygous for the alternative allele in the 13 other breeds. Of the remaining variants 290 were predicted to either directly alter the coding sequence of the protein or disrupt the transcript sequence. A shortlist of 10 variants in potential candidate gene was generated by performing an automated search of the Online Mendelian Inheritance in Man database (OMIM) for the gene names implicated by the 290 variants and the key word cerebellar ataxia. The 10 shortlisted variants were individually considered by visually inspecting genome sequence read alignments, strength of the candidate gene, variant consequence (SIFT) and Case 1 genotype for each variant, established through an RNAseq dataset. All shortlisted variants, including an initially compelling variant in Ataxin1 (ATXN1), were excluded apart from an exon 26 splice donor variant (CanFam3.1, chr12:45,530,566, c.2653 + 1G > A) in the Sorting Nexin 14 (SNX14) gene; an excellent candidate causal variant (Fig. 4). A summary of the shortlisted variants is shown in the supporting information Additional file 1: Table S1. The potential causal variant was further interrogated by genotyping the second case and a set of 133 unaffected HV as controls. The other CCD case (Case 1) was also found to be homozygous for the disease-associated allele. Of the controls, three heterozygotes were identified with the remaining 130 HVs homozygous for the reference allele. Of the three heterozygotes, two were half-siblings and the other was unrelated at the parent level. Pedigree information for the CCD cases was not available, but no clinical signs were reported in either parent, supporting carrier status.Fig. 4 Splice donor mutation in CCD cases causing an altered splicing pattern. An exon 26 splice donor site mutation (GT > AT) causing exon 26 of SNX14 to extent by 275 additional nucleotides to an alternative splice donor site. The extended exon has a sequence of 15 aberrant amino acids before a premature stop codon The effect of the variant on the SNX14 transcript was investigated by RT-PCR across the exon 26–27 junction, using forward and reverse primers in exon 25 and 29 respectively and cerebellar RNA from one CCD case (Case 1) and the six control dogs. A larger than expected PCR product was seen for the CCD case (Fig. 5), with exon 26 in the case extending 275 bp to the position of an alternative splice donor site, resulting in an aberrant string of 15 amino acids and a premature stop codon (p.G885Dfs*16) (Fig. 4). Western blot analysis using an antibody targeting an amino acid sequence before the splice variant, failed to identify SNX14 protein (either full length or reduced) in cerebellum tissue of the CCD case (Fig. 6).Fig. 5 Altered splicing pattern in a CCD case. Polymerase chain reaction of cerebellar cDNA between exon 25 and 29 demonstrates a product size increase of 275 bp for the CCD case (1) in comparison to 6 control dogs (2–7) when visualised by agarose gel electrophoresis. No template control (8) Fig. 6 Western blot analysis of SNX14 protein. Western blot analysis of SNX14 failed to detect protein in cerebellar tissue of the CCD case (1). Presence of a band of expected size for SNX14 (110 kD) was detected for 6 canine control cerebellar samples (2–7). ACTB was used as a loading control Discussion In this paper we present clinical, diagnostic, pathological and genetic findings associated with, to our knowledge, the first reported cases of CCD in HV dogs. We also report the use of genome sequencing to successfully identify a disease-associated splice donor site variant in SNX14 as a strong candidate for CCD in the HV. An altered SNX14 splicing pattern for a CCD case was demonstrated by RNA analysis and no SNX14 protein could be detected in CCD case cerebellum by western blotting. The age of onset of clinical signs in the two reported HV dogs is consistent with previous reports of canine CCD in those breeds with a typically juvenile onset, with most of these cases having an onset at around 3–4 months of age, and a relatively rapid progression over several weeks to months [2, 6, 8, 9, 14, 16]. This is in contrast to breeds where a neonatal, rapidly progressive phenotype is seen, such as the Beagle and Rhodesian Ridgeback [3, 13], and those with a later onset and more slowly progressive condition typified by CCD in Gordon Setters and Old English Sheepdogs [15]. Similar to reports of canine CCD seen in other breeds [12, 16], the observed clinical signs in the HV were referable to diffuse cerebellar dysfunction, with both dogs displaying a marked hypermetric, dysmetric ataxia, with marked truncal sway and intention tremor, as well as bilateral menace deficits with normal vision. Although they remained ambulatory, both dogs demonstrated a progressive clinical course over a short period of around two months, to the point that their owners elected for euthanasia on humane grounds due to their poor quality of life. This history of progression to euthanasia is typical of CCD in most breeds, apart from those with an adult onset and slowly progressive CCD, whereby survival for years is possible [7, 12, 15]. Diagnostic investigations performed in Case 1 were consistent with CCD, with MRI revealing widening of sulci in the craniodorsal cerebellum suggesting atrophy of cerebellar grey matter in this area (Fig. 1), similar to previous MRI reports [5, 16, 25, 26]. The MRI scan performed in this HV with CCD revealed a mid-sagittal cross-sectional brainstem to cerebellar area ratio of 81.1 %, a value lower than a previously reported lower limit of 89 % for the diagnosis of cerebellar atrophy using MRI [25]. However, the 13 dogs used in that MRI study consisted mainly of breeds with an adult onset, slowly progressive form of CCD. The HV in the current study also underwent MRI early in the disease, and therefore given the variability in the time point in the disease course at which MRI is performed, comparison with dogs in the previous study is difficult. Given the uncertainty regarding the length of time required for cerebellar atrophy to be detectable on MRI in cases of CCD [25], it is possible that MRI was performed in Case 1 before sufficient atrophy to cause a reduction in MRI area ratio had developed. Histopathological examinations in both dogs revealed diffuse Purkinje neuron loss, with secondary depletion of the granular cell layer, as well as reactive changes (gliosis) likely secondary to this degeneration of the Purkinje neuron layer. These changes fit well with previous reports of CCD or cerebellar abiotrophy, where primary Purkinje neuron loss and secondary granular cell degeneration are most commonly reported [7, 11, 13, 14, 16]. Similar to previous studies, the HVs also demonstrated a characteristic distribution of cerebellar folial atrophy, with the dorsal and rostral vermis predominantly affected whilst the caudoventral regions appeared spared [10, 12, 14, 17]. Purkinje neuron loss and associated gliosis were more severe in Case 1. In addition Case 1 also exhibited early neuronal degeneration within the cerebellar nuclei, which may be associated with the progression of clinical signs in this animal prior to euthanasia. These microscopic changes were sufficient to cause mild grossly observable symmetrical atrophy of the cerebellar hemispheres and thinning of the cerebellar folia, consistent with previous reports [1, 7, 16]. Mutations in SNX14 have been reported to cause autosomal recessive cerebellar ataxia and intellectual disability syndrome in humans [27]. In the first report of an association between this syndrome and mutations in SNX14, three distinct SNX14 mutations were identified which segregated in three unrelated consanguineous families, and included a nonsense mutation located in exon 26 of SNX14. Interestingly SNX14 was investigated in the third family purely based on phenotypic similarities to the first two families. Full clinical signs included intellectual disability, cerebellar ataxia, early-onset cerebellar atrophy, sensorineural hearing loss, and the distinctive association of progressively coarsening facial features (broad face, fullness of the upper eyelid, broad nasal base and slight underdevelopment of the alae, broad and long philtrum, thick lower lip vermillion), relative macrocephaly, and the absence of seizures. Fifth finger brachycamptodactyly was also reported. There were no obvious distinctive facial phenotypic differences between CCD cases and unaffected HVs, and auditory function was normal on examination. However, electrodiagnostic tests of auditory function, such as brainstem auditory evoked potentials, could be performed in future studies to exclude a possible mild hearing loss. The intellectual disability seen in humans with SNX14 mutations was also not identified in the HVs with CCD, although this is challenging to objectively evaluate in dogs and may have developed at a later age. Marked cerebellar ataxia and cerebellar atrophy with thinning of the cerebellar folia on gross pathology were seen in HV CCD cases, consistent with human cases. SNX14 is a conserved 946 amino acid protein. In a knockdown mouse model, SNX14 has been shown to be involved in synaptic transmission and maintaining normal neuronal excitability [28]. Furthermore, using an in vivo SNX14 translation blocking in a zebrafish model, loss of neuronal tissue and reduced cerebellar area was demonstrated, which could be rescued by co-injection with the human SNX14 orthologue [29]. Cellular fractionation experiments have shown SNX14 to be associated with the lysosomal fraction [29]. SNX14 has also been shown to associate with the serotonin subtype 6 receptor, which has been implicated in cognition, anxiety and learning and memory disorders, with SNX14 acting as a negative regulator [30]. Although not directly examined, there were no reported clinical signs in the remaining seven HV siblings or the parents, supporting an autosomal recessive mode of inheritance of the SNX14 mutation in the HV, as reported in humans. Whilst we are not aware of any other cases of CCD yet identified in other HV families, the finding of three heterozygotes amongst the 133 unaffected HV samples confirms the presence of carriers in the wider HV population. As a result, genetic testing for the SNX14 mutation could be beneficial both in supporting a presumptive antemortem diagnosis of CCD in an HV with consistent clinical signs, as well as allowing breeders to identify carriers. Conclusion This paper represents the first confirmed report of cerebellar cortical degeneration in two young HV dogs and identifies a splice donor site mutation in SNX14 as the likely cause, based on RNA analysis, protein analysis and studies in other species and in vivo models. The identification of this mutation will allow the development of a genetic test to help to identify affected and carrier dogs, to help to eliminate the disease from the breed. Additional file Additional file 1: Table S1. A summary of the 10 shortlisted variants in potential candidate genes. (XLSX 14 kb) Abbreviations CCDCerebellar cortical degeneration HVHungarian Vizsla MRIMagnetic resonance imaging SNX14Sorting nexin 14 Acknowledgements The authors would like to thank all dog owners and breeders who have kindly consented to the use of samples for the purpose of this research. This manuscript was assessed according to the Royal Veterinary College Research Office’s code of good research practice (authorisation number – CSS_01208). Funding We thank the High-Throughput Genomics Group at the Wellcome Trust Centre for Human Genetics (funded by Wellcome Trust grant reference 090532/Z/09/Z) for the generation of the sequencing data. Availability of data and materials The genome sequencing dataset has been submitted to the European Nucleotide Archive (accession number: PRJEB12339) at http://www.ebi.ac.uk/ena. Authors’ contributions JF and PJK diagnosed the CCD cases, performed clinical investigations and organised sample collection. RLT and SLP performed gross and histopathological examinations. OPF, RJH and CAJ performed genetic investigations. MB and OPF analysed the genome sequencing data. JF, OPF and RLT contributed sections to the manuscript. All authors critically reviewed and approved the manuscript. Competing interests A commercial genetic test for cerebellar cortical degeneration in the Hungarian Vizsla has been made available by the Animal Health Trust based on the findings in this manuscript. The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate Ethical approval for this study was granted by the RVC Clinical Research Ethical Review Board (URN 2016 1516B). Owners of all animals included gave written and informed consent for diagnostic procedures and use of the obtained information for research purposes. ==== Refs References 1. Cummings JF de Lahunta A A study of cerebellar and cerebral cortical degeneration in miniature poodle pups with emphasis on the ultrastructure of Purkinje cell changes Acta Neuropathol 1988 75 3 261 271 10.1007/BF00690534 3348083 2. Thomas JB Robertson D Hereditary cerebellar abiotrophy in Australian kelpie dogs Aust Vet J 1989 66 9 301 302 10.1111/j.1751-0813.1989.tb13959.x 2818374 3. Chieffo C Stalis IH Van Winkle TJ Haskins ME Patterson DF Cerebellar Purkinje’s cell degeneration and coat color dilution in a family of Rhodesian Ridgeback dogs J Vet Intern Med 1994 8 2 112 116 10.1111/j.1939-1676.1994.tb03207.x 8046673 4. Steinberg HS Van Winkle T Bell JS de Lahunta A Cerebellar degeneration in Old English Sheepdogs J Am Vet Med Assoc 2000 217 8 1162 1165 10.2460/javma.2000.217.1162 11043686 5. van der Merwe LL Lane E Diagnosis of cerebellar cortical degeneration in a Scottish terrier using magnetic resonance imaging J Small Anim Pract 2001 42 8 409 412 10.1111/j.1748-5827.2001.tb02491.x 11518422 6. Sandy JR Slocombe RE Mitten RW Jedwab D Cerebellar abiotrophy in a family of Border Collie dogs Vet Pathol 2002 39 6 736 738 10.1354/vp.39-6-736 12450206 7. Olby N Blot S Thibaud JL Phillips J O’Brien DP Burr J Berg J Brown T Breen M Cerebellar cortical degeneration in adult American Staffordshire Terriers J Vet Intern Med 2004 18 2 201 208 10.1111/j.1939-1676.2004.tb00161.x 15058771 8. Flegel T Matiasek K Henke D Grevel V Cerebellar cortical degeneration with selective granule cell loss in Bavarian mountain dogs J Small Anim Pract 2007 48 8 462 465 10.1111/j.1748-5827.2006.00257.x 17663663 9. Jokinen TS Rusbridge C Steffen F Viitmaa R Syrja P De Lahunta A Snellman M Cizinauskas S Cerebellar cortical abiotrophy in Lagotto Romagnolo dogs J Small Anim Pract 2007 48 8 470 473 10.1111/j.1748-5827.2006.00298.x 17490444 10. Kent M Glass E deLahunta A Cerebellar cortical abiotrophy in a beagle J Small Anim Pract 2000 41 7 321 323 10.1111/j.1748-5827.2000.tb03210.x 10976629 11. Gandini G Botteron C Brini E Fatzer R Diana A Jaggy A Cerebellar cortical degeneration in three English bulldogs: clinical and neuropathological findings J Small Anim Pract 2005 46 6 291 294 10.1111/j.1748-5827.2005.tb00323.x 15971900 12. Urkasemsin G Linder KE Bell JS de Lahunta A Olby NJ Hereditary cerebellar degeneration in Scottish terriers J Vet Intern Med 2010 24 3 565 570 10.1111/j.1939-1676.2010.0499.x 20384950 13. Forman OP De Risio L Stewart J Mellersh CS Beltran E Genome-wide mRNA sequencing of a single canine cerebellar cortical degeneration case leads to the identification of a disease associated SPTBN2 mutation BMC Genet 2012 13 55 10.1186/1471-2156-13-55 22781464 14. Kyostila K Cizinauskas S Seppala EH Suhonen E Jeserevics J Sukura A Syrja P Lohi H A SEL1L mutation links a canine progressive early-onset cerebellar ataxia to the endoplasmic reticulum-associated protein degradation (ERAD) machinery PLoS Genet 2012 8 6 e1002759 10.1371/journal.pgen.1002759 22719266 15. Agler C Nielsen DM Urkasemsin G Singleton A Tonomura N Sigurdsson S Tang R Linder K Arepalli S Hernandez D Canine hereditary ataxia in old english sheepdogs and gordon setters is associated with a defect in the autophagy gene encoding RAB24 PLoS Genet 2014 10 2 e1003991 10.1371/journal.pgen.1003991 24516392 16. Urkasemsin G Olby NJ Canine hereditary ataxia Vet Clin North Am Small Anim Pract 2014 44 6 1075 1089 10.1016/j.cvsm.2014.07.005 25441626 17. Urkasemsin G Linder KE Bell JS de Lahunta A Olby NJ Mapping of Purkinje neuron loss and polyglucosan body accumulation in hereditary cerebellar degeneration in Scottish terriers Vet Pathol 2012 49 5 852 859 10.1177/0300985811412622 21753036 18. Tipold A Fatzer R Jaggy A Moore P Vandevelde M Presumed immune-mediated cerebellar granuloprival degeneration in the Coton de Tulear breed J Neuroimmunol 2000 110 1–2 130 133 10.1016/S0165-5728(00)00330-1 11024542 19. Vite CH DayrellHart B Lexa F Kerlin R VanWinkle T Steinberg SA Atypical disease progression and MR imaging of a Kerry Blue Terrier with cerebellar cortical and extrapyramidal nuclear abiotrophy Prog Vet Neurol 1996 7 1 12 15 20. O’Brien DP Johnson GS Schnabel RD Khan S Coates JR Johnson GC Taylor JF Genetic mapping of canine multiple system degeneration and ectodermal dysplasia loci J Hered 2005 96 7 727 734 10.1093/jhered/esi086 15958791 21. Zeng R Farias FH Johnson GS McKay SD Schnabel RD Decker JE Taylor JF Mann CS Katz ML Johnson GC A truncated retrotransposon disrupts the GRM1 coding sequence in Coton de Tulear dogs with Bandera’s neonatal ataxia J Vet Intern Med 2011 25 2 267 272 10.1111/j.1939-1676.2010.0666.x 21281350 22. Li H Durbin R Fast and accurate short read alignment with Burrows-Wheeler transform Bioinformatics 2009 25 14 1754 1760 10.1093/bioinformatics/btp324 19451168 23. McKenna A Hanna M Banks E Sivachenko A Cibulskis K Kernytsky A Garimella K Altshuler D Gabriel S Daly M The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data Genome Res 2010 20 9 1297 1303 10.1101/gr.107524.110 20644199 24. Cingolani P Platts A le Wang L Coon M Nguyen T Wang L Land SJ Lu X Ruden DM A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3 Fly (Austin) 2012 6 2 80 92 10.4161/fly.19695 22728672 25. Thames RA Robertson ID Flegel T Henke D O’Brien DP Coates JR Olby NJ Development of a morphometric magnetic resonance image parameter suitable for distinguishing between normal dogs and dogs with cerebellar atrophy Vet Radiol Ultrasound 2010 51 3 246 253 10.1111/j.1740-8261.2009.01655.x 20469545 26. Henke D Bottcher P Doherr MG Oechtering G Flegel T Computer-assisted magnetic resonance imaging brain morphometry in American Staffordshire Terriers with cerebellar cortical degeneration J Vet Intern Med 2008 22 4 969 975 10.1111/j.1939-1676.2008.0138.x 18647158 27. Thomas AC Williams H Seto-Salvia N Bacchelli C Jenkins D O’Sullivan M Mengrelis K Ishida M Ocaka L Chanudet E Mutations in SNX14 cause a distinctive autosomal-recessive cerebellar ataxia and intellectual disability syndrome Am J Hum Genet 2014 95 5 611 621 10.1016/j.ajhg.2014.10.007 25439728 28. Huang HS Yoon BJ Brooks S Bakal R Berrios J Larsen RS Wallace ML Han JE Chung EH Zylka MJ Snx14 regulates neuronal excitability, promotes synaptic transmission, and is imprinted in the brain of mice PLoS One 2014 9 5 e98383 10.1371/journal.pone.0098383 24859318 29. Akizu N Cantagrel V Zaki MS Al-Gazali L Wang X Rosti RO Dikoglu E Gelot AB Rosti B Vaux KK Biallelic mutations in SNX14 cause a syndromic form of cerebellar atrophy and lysosome-autophagosome dysfunction Nat Genet 2015 47 5 528 U137 10.1038/ng.3256 25848753 30. Ha CM Park D Kim Y Na M Panda S Won S Kim H Ryu H Park ZY Rasenick MM SNX14 is a bifunctional negative regulator for neuronal 5-HT6 receptor signaling J Cell Sci 2015 128 9 1848 1861 10.1242/jcs.169581 25795301

PMC005xxxxxx/PMC5002146.txt

==== Front BMC CancerBMC CancerBMC Cancer1471-2407BioMed Central London 274110.1186/s12885-016-2741-zResearch ArticleBelotecan/cisplatin versus etoposide/cisplatin in previously untreated patients with extensive-stage small cell lung carcinoma: a multi-center randomized phase III trial Oh In-Jae droij@jnu.ac.kr 1Kim Kyu-Sik cyberkks@jnu.ac.kr 1Park Cheol-Kyu linus214@naver.com 1Kim Young-Chul +82-61-379-7614kyc0923@jnu.ac.kr 1Lee Kwan-Ho ghlee@medical.yu.ac.kr 2Jeong Jin-Hong jhchn@medical.yu.ac.kr 2Kim Sun-Young sykim@cnu.ac.kr 3Lee Jeong-Eun vov-x@hanmail.net 3Shin Kye-Chul skc1001@wonju.yonsei.ac.kr 4Jang Tae-Won jangtw@ns.kosinmed.or.kr 5Lee Hyun-Kyung drpulmo@inje.ac.kr 6Lee Kye-Young kyleemd@kuh.ac.kr 7Lee Sung-Yong dragonett@naver.com 81 Department of Internal Medicine, Lung and Esophageal Cancer Clinic, Chonnam National University Hwasun Hospital, 322 Seoyang-ro, Hwasun-eup, Jeonnam 58128 South Korea 2 Department of Internal Medicine, Yeungnam University Medical Center, Yeungnam University College of Medicine, Daegu, Republic of Korea 3 Department of Internal Medicine, Chungnam National University Hospital, Chungnam National University School of Medicine, Daejeon, Republic of Korea 4 Department of Internal Medicine, Yonsei University Wonju Severance Christian Hospital, Yonsei University College of Medicine, Wonju, Republic of Korea 5 Department of Internal Medicine, Kosin University Gospel Hospital, Kosin University College of Medicine, Busan, Republic of Korea 6 Department of Internal Medicine, Inje University Busan Paik Hospital, Busan, South Korea 7 Department of Internal Medicine, Konkuk University Medical Center, Konkuk University School of Medicine, Seoul, Republic of Korea 8 Department of Internal Medicine, Guro Hospital, Korea University, Seoul, South Korea 26 8 2016 26 8 2016 2016 16 1 6909 6 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 No novel chemotherapeutic combinations have demonstrated superior efficacy to etoposide/cisplatin (EP), a standard treatment regimen for extensive-stage small cell lung carcinoma (ES-SCLC) over the past decade. We aimed to compare the efficacy and safety of belotecan/cisplatin (BP) and EP regimens in chemotherapy- and radiotherapy-naïve patients with previously untreated ES-SCLC. Methods We conducted a multi-center, randomized, open-label, parallel-group, phase III clinical study. A total of 157 patients were recruited at 14 centers with 147 patients meeting the inclusion/exclusion criteria and randomized to either BP (n = 71) or EP (n = 76) treatment arms. A non-inferior response rate (RR) in the BP arm, analyzed by intent-to-treat analysis according to Response Evaluation Criteria in Solid Tumors version 1.0 criteria, was used as the primary endpoint. The secondary endpoints were progression-free survival (PFS) and overall survival (OS). Results In the BP arm, one patient had a complete response, 41 had a partial response (PR), and 17 had stable disease (SD). In the EP arm, 35 patients had PR and 28 had SD. The RR in the BP arm was non-inferior to the EP regimen in patients with ES-SCLC (BP: 59.2 %, EP: 46.1 %, difference: 13.1 %, 90 % two-sided confidence interval: -0.3–26.5, meeting the predefined non-inferiority criterion of -15.0 %). No significant differences in OS or PFS were observed between the treatment arms. Hematologic toxicities, including grade 3/4 anemia and thrombocytopenia, were significantly more prevalent in the BP arm than the EP arm. Conclusions The RR to the BP regimen was non-inferior to the EP regimen in chemotherapy- and radiotherapy-naïve patients with previously untreated ES-SCLC. Hematologic toxicities were significantly more prevalent in the BP group, indicating that BP should be used with care, particularly in patients with a poor performance status. Further studies assessing PFS and OS are required to validate the superiority of the BP regimen. Trial registration ClinicalTrials.gov identifier NCT00826644. Date of Registration: January 21, 2009. Keywords Small cell lung carcinomaExtensive stage diseasePhase III studyChemotherapyFirst-lineBelotecanissue-copyright-statement© The Author(s) 2016 ==== Body Background Lung cancer is one of the leading causes of cancer-related death worldwide [1–3]. Small cell lung cancer (SCLC) accounts for up to 20 % of all new cases of lung cancer and deaths [3, 4]. Compared to non-small cell lung cancer (NSCLC), SCLC is generally more aggressive, with decreased doubling times and faster growth rates. Moreover, early widespread metastasis is a recognized feature of SCLC [5]. Extensive-stage SCLC (ES-SCLC) refers to SCLC metastasis to distant body regions. Since the mid-1980s, no significant improvement in the survival of patients with ES-SCLC has been observed; the median overall survival (OS) is estimated at approximately 10 months [6–10]. Currently, a two-drug combination of platinum and etoposide at doses associated with at least moderate toxic effects is most commonly used to treat ES-SCLC [11]. The overall response rates of 50 %–80 % and complete response rates of 0 %–30 % have been reported with this treatment approach [12, 13]. To date, a number of pharmacological agents have been developed for the treatment of NSCLC. However, no novel chemotherapeutic combinations have demonstrated superior efficacy to etoposide/cisplatin (EP), a standard treatment regimen, in the treatment of SCLC over the past decade, although irinotecan/cisplatin (IP) has been reported as an effective combination regimen [10]. Belotecan {7-[2(N-isopropylamino) ethyl]-(20S)-camptothecin} is a newly developed camptothecin analogue. According to two multi-center phase IIa studies, belotecan monotherapy is an effective modality for the treatment of SCLC in chemotherapy-naïve patients [14, 15]. Moreover, multi-center phase II studies have reported response rates (RR) higher than 70 % and OS greater than 10 months in patients with ES-SCLC receiving belotecan/cisplatin (BP) as a first-line treatment regimen [16–18]. On the basis of the above mentioned information, we conducted a multi-center, randomized, open-label, parallel-group, phase III clinical study to compare the efficacy and safety of BP and EP regimens in chemotherapy- and radiotherapy-naïve patients with previously untreated ES-SCLC. Methods Study patients Patients who met all of the following inclusion criteria were enrolled in this trial: (1) aged between 19 and 80 years, (2) histologically or cytologically proven ES-SCLC, (3) no past history of chemotherapy or radiotherapy, (4) ≥1 measurable disease according to the Response Evaluation Criteria in Solid Tumors (RECIST) criteria version 1.0, (5) Eastern Cooperative Oncology Group (ECOG) performance status (PS) of ≤2, (6) a life expectancy of ≥12 weeks, (7) adequate organ function [absolute neutrophil count (ANC) ≥1,500/mm3, platelet count ≥100,000/mm3, hemoglobin ≥9.0 g/dL, total bilirubin level ≤1.5 mg/dL, aminotransferase ≤2-fold upper limit of normal (ULN) or ≤3-fold ULN if demonstrable liver metastases, alkaline phosphatase ≤2-fold ULN, and creatinine ≤1.5 mg/dL or creatinine clearance ≥60 mL/min]. The exclusion criteria were as follows: (1) severe bacterial infection, (2) malignancies other than basal cell skin cancer or cervical carcinoma in situ, (3) brain metastases, (4) women with child-bearing potential, and (5) women who are pregnant or breast-feeding. The study was approved by the Institutional Review Board (IRB) of each medical institution. All patients provided a written informed consent. The current study was registered with ClinicalTrials.gov (Identifier: NCT00826644). Dosing rationale and schedule Patients were randomized to either EP or BP treatment arms and stratified according to ECOG PS (0-1 vs. 2) and age (<65 vs. ≥65 years) at a ratio of 1:1. A phase I study was conducted to determine the maximum tolerated dose (MTD), toxicity, and dose-limiting toxicity of BP; it showed that the MTD and recommended dose for phase II studies was 0.5 mg/m2 on days 1–4 in combination with 60 mg/m2 cisplatin on day 1 at a 3-week interval [19]. The BP regimen consisted of intravenous belotecan 0.5 mg/m2 mixed with 100 mL of 5 % dextrose over 30 min on day 1–4 and intravenous cisplatin 60 mg/m2 on day 1 of 3-week cycles. The EP regimen consisted of etoposide 100 mg/m2 on days 1, 2 and 3 and cisplatin 60 mg/m2 on day 1 of 3-week cycles. Both regimens required hydration and administration of antiemetic drugs. On day 1, the patients were treated if they showed an ANC ≥1,500/mm3, platelet count ≥100,000/mm3, and creatinine clearance ≥60 mL/min. In addition, the patients were given recombinant human granulocyte colony-stimulating factor (G-CSF) to improve ANC according to clinical judgment. Subsequent treatments were delayed on a weekly basis until recovery of ANC in cases of G-CSF treatment failure. The patients dropped out of the study if the treatment was delayed by more than 2 weeks. With respect to the dose of the subsequent treatment, the patients were given unadjusted treatment doses upon recovery of ANC to ≥1,500/mm3 and a platelet count ≥100,000/mm3. In the patients achieving a recovery of ANC to 1,000–1,500/mm3 and a platelet count to 75,000–100,000/mm3, subsequent treatment doses were reduced by 20 %. In patients with ANC <500/mm3, platelet count <25,000/mm3, or febrile neutropenia during treatment, subsequent treatment doses were reduced by 20 %. Cisplatin doses were not adjusted in patients with decreases in creatine clearance to ≥60 mL/min from baseline. Cisplatin doses were reduced by 50 % in patients with creatinine clearance of 30–60 mL/min. Cisplatin was discontinued in patients with decreases in creatinine clearance to ≤30 mL/min. Patient evaluation The patients were evaluated at baseline based on their medical history, physical examination, imaging studies, complete blood counts, and serum biochemistry. The response was assessed by computed tomography (CT) every two treatment cycles at follow-up visits. After the completion of chemotherapy, CT scans were performed every three months until evidence or suspicion of disease progression. The treatment response was centrally evaluated independently according to RECIST version 1.0 criteria as follows: (1) complete response (CR), disappearance of all clinical and radiological evidence of the tumor; (2) partial response (PR), decrease of 30 % or more in the sum of longest diameters of all target measurable lesions; (3) progressive disease (PD), increase of more than 20 % of the sum of longest diameters of all target measurable lesions or the appearance of new lesions; and (4) stable disease (SD), all other circumstances. The patients eligible for response evaluation were evaluated at a minimum interval of 4 weeks to confirm CR or PR. Adverse events (AEs) were graded according to the NCI-CTCAE, version 3.0 (http://ctep.cancer.gov/protocolDevelopment/electronic_applications/docs/ctcaev3.pdf). Efficacy endpoints Per-protocol (PP) population was defined patients who completed two cycles of chemotherapy in accordance with the protocol. Modified Intent-to-treat (mITT) population was defined as patients randomized and treated with at least one cycle of chemotherapy. mITT analysis was used to assess non-inferiority of RR in the BP arm as the primary endpoint according to RECIST 1.0 criteria. The secondary endpoints were progression-free survival (PFS) and OS. Subgroup analysis of the mITT population was performed to assess the associations between RR and ECOG PS, age, and body weight. Safety endpoints The safety population comprised all patients eligible for safety analysis. The occurrence of toxicities and adverse effects were evaluated throughout the study period through the measurement of vital signs, physical examination, and clinical laboratory tests. The safety endpoints were adverse drug reactions (ADRs), serious adverse events (SAEs), and treatment-related death (TRD). Statistical analyses To determine the sample size, we estimated an RR of 71 % for the BP arm [14, 15, 19] and 66 % for the EP arm [6, 13, 20], considering a non-inferiority margin of −15 % at a power of 80 % with a one-sided error (α) = 0.05 [21]. In addition, we assumed a dropout rate of 10 % during the follow-up period. Thus, we aimed to enroll a total of 150 patients (n = 150) in the present study. The baseline and clinical characteristics of the patients were expressed as median and range. The Student’s t-test and chi-square test were used to compare two treatment arms, as appropriate. Two-sided 90 % confidence intervals (CIs) were calculated for the difference in the RR between the two treatment arms without corrections for continuity, as described by Newcombe [21]. The Cochran–Mantel–Haenszel (CMH) test was performed in patients who met the criteria for non-inferiority with consideration of the stratification factors. Thus, we attempted to analyze the difference in RR between the two treatment arms using Fisher’s exact test. PFS was defined as the time from randomization to clinical or radiological progression or death. OS was defined as the time from randomization to death from any cause. If the patient was lost to follow up or an event (disease progression or death) did not occur until study termination, the patient was censored at the time of last contact. PFS and OS were analyzed using the Kaplan–Meier methods. Results Patient groups A total of 157 patients were recruited at 14 centers across Korea from January 2009 to January 2013. Of these, 147 patients met the inclusion/exclusion criteria of this study and were randomized to the BP (n = 71) or EP (n = 76) arms. The number of the patients in the mITT, PP, and safety populations were 71, 57, and 70 in the BP arm and 76, 63, and 77 in the EP arm, respectively (Fig. 1).Fig. 1 Disposition of the study patients. Of the 157 patients we recruited at a total of 14 centers, 147 met inclusion/exclusion criteria and then randomized to either the BP arm (n = 71) or the EP arm (n = 76). In the BP arm, the number of the patients of the mITT set, the PP set and the safety set were 71, 57 and 70, respectively. In the EP arm, these values were 76, 63 and 77 in the corresponding order Patient baseline and clinical characteristics No significant differences in the median age, median male-to-female ratio, median body surface area, or median ECOG PS were observed between the treatment arms (Table 1). A significant difference in the median body mass index was observed between the treatment arms (p < 0.05).Table 1 Patient demographics of modified intent-to-treat population Characteristic Belotecan/Cisplatin (n = 71) Etoposide/Cisplatin (n = 76) p value No. % No. % Ages, years .565  Median 67.0 66.5  Range 29–91 46–80 Sex .751  Male 62 87.3 65 85.5  Female 9 12.7 11 14.5 Body mass index, kg/m2 .004  Median 23.5 22.0  Range 15.1–30.4 15.0–30.8 Body surface area, m2 .055  Median 1.71 1.65  Range 1.30–1.99 1.26–2.06 ECOG performance status .200  0 23 32.4 20 26.3  1 37 52.1 35 46.1  2 11 15.5 21 27.6 Abbreviation: ECOG Eastern cooperative oncology group The proportion of the patients receiving more than four cycles of chemotherapy was 67.1 % in the BP arm and 63.7 % in the EP arm. The mean number of chemotherapy cycles was 3.9 in the BP arm and 4.1 in the EP arm. In the BP arm, the mean delivered dose of belotecan and cisplatin was 0.45 mg/m2 (90.0 % of planned dose) and 56.8 mg/m2 (94.6 % of planned dose), respectively. In the EP arm, the mean delivered dose of etoposide and cisplatin was 93.8 mg/m2 (93.8 % of planned dose) and 58.6 mg/m2 (97.7 % of planned dose), respectively. The relative dose intensity (RDI) was calculated by dividing the intensity of the delivered dose by that of the standard dose. A significantly lower RDI was observed in the BP arm than that in the EP arm (0.79 ± 0.14 vs. 0.86 ± 0.13, p = 0.001). Efficacy endpoints Response rates According to the mITT analysis, the RR was 59.2 % in the BP arm and 46.1 % in the EP arm (Table 2 & Fig. 2). The lower limit of the two-sided 90 % CI was greater than the non-inferiority margin (−0.3 vs. −15.0). This indicates that the BP regimen was non-inferior to the EP regimen with respect to the RR in patients with ES-SCLC.Table 2 Best overall response of modified intent-to-treat and per protocol population Modified intent-to-treat population Response BP (n = 71) EP (n = 76) Difference 90 % CI P value by CMH test No. % No. % CR 1 1.4 0 0 PR 41 57.7 35 46.1 SD 17 23.9 28 36.8 PD 4 5.6 5 6.6 NE 8 11.3 8 10.5 CR + PR 42 59.2 35 46.1 13.1 −0.3 to 26.5 .214 CR + PR + SD 59 83.1 63 82.9 0.2 −10.0 to 10.4 .826 Per protocol population Response BP (n = 57) EP (n = 63) Difference 90 % CI P value by CMH test No. % No. % CR 1 1.8 0 0 PR 40 70.2 34 54.0 SD 9 15.8 21 33.3 PD 3 5.3 3 4.8 NE 4 7.0 5 7.9 CR + PR 41 71.9 34 54.0 18.0 3.7 to 32.2 .061 CR + PR + SD 50 87.7 55 87.3 0.4 −0.95 to 10.4 .978 Abbreviation: BP belotecan/cisplatin, EP etoposide/cisplatin, CI Confidence interval, CMH Cochran-Mantel-Haenzel, CR complete response, PR partial response, SD stable disease, PD progressive disease, NE non-evaluable cases Fig. 2 Efficacy endpoints of modified intent-to-treat (mITT) and per protocol (PP) population. The BP group is not inferior to the EP group with regards to the response rate (between group difference 13.1 %, 90 % two-sided confidence interval -0.3 to 26.5, meeting the predefined non-inferiority criterion of -15.0 %) in mITT population. The response of the BP group was superior in PP population According to the PP analysis, an 18.0 % difference in the RR was demonstrated between the two treatment arms (90 % CI: 3.7–32.3). In addition, no significant differences in the RRs between the two treatment arms in either the mITT or PP groups were observed according to the CMH test. Overall survival and progression-free survival No significant differences in OS or PFS were observed between the treatment arms. The median OS was 360 days (95 % CI: 285–482) in the BP arm and 305 days (95 % CI: 232–343) in the EP arm (Log-Rank p = 0.210, Fig. 3). In addition, the median PFS was 190 days (95 % CI: 148–219) in the BP arm and 172 days (95 % CI: 144–195) in the EP arm (Log-Rank p = 0.369, Fig. 3).Fig. 3 Overall survival and progression-free survival on Kaplan–Meier analysis. (a) There were no significant differences in the OS and PFS between the two treatment arms. That is, the median OS was 360 days (95 % CI: 285–482) in the BP arm and 305 days (95 % CI: 232–343) in the EP arm (Log-Rank p = 0.210). (b) The median PFS was 190 days (95 % CI: 148-219) in the BP arm and 172 days (95 % CI: 144–195) in the EP arm (Log-Rank p = 0.369) Subgroup analysis In the BP arm, the RR was 63.3 % in patients with ECOG PS 0-1 (n = 60) and 36.4 % in those with ECOG PS 2 (n = 11). In the EP arm, the RR was 52.7 % in patients with ECOG PS 0–1 (n = 55) and 28.6 % in those with ECOG PS 2 (n = 21). In addition, the RR was 66.7 % in patients aged ≤65 years (n = 30) and 53.7 % in those aged ≥65 years (n = 41) in the BP arm. In the EP arm, the RR was 57.8 % in patients aged ≤65 years (n = 33) and 37.2 % in those aged ≥65 years (n = 43). Furthermore, the RR was 60.0 % in patients weighing less than 62.5 kg (n = 30) and 58.5 % in those weighing more than 62.5 kg (n = 41) in the BP arm. In the EP arm, the RR was 45.2 % in patients weighing less than 62.5 kg (n = 42) and 47.1 % in those weighing more than 62.5 kg (n = 34). Safety endpoints In the present study, all 147 included patients were eligible for safety analysis. No significant differences in ADRs or TRD were observed between the treatment arms (98.6 % vs. 92.2 %, p = 0.119 and 12.9 % vs. 10.4 %, p = 0.797, respectively). However, the incidence of SAEs was significantly higher in the BP arm (42/70 cases, 60.0 %) than those in the EP arm (31/77 cases, 40.3 %, p = 0.021). Hematologic toxicities were more prevalent in this study than non-hematologic toxicities. Grade 3/4 anemia (34.3 %) and thrombocytopenia (54.3 %) were significantly more prevalent in the BP arm than those in the EP arm. No significant difference in the incidence of non-hematologic toxicities was observed between the treatment arms. Grade 3/4 infection and hyponatremia occurred in approximately 10 % of the patients in both treatment arms. Nearly all the non-hematologic toxicities were of grade 1/2 severity and could be treated successfully (Table 3).Table 3 Hematologic and non-hematologic toxicities of safety population BP Group (n = 70) EP Group (n = 77) P † Gradea, n (%) Gradea, n (%) 1 2 3 4 ≥3 1 2 3 4 ≥ 3 Febrile neutropenia 0(0.0) 0(0.0) 7(10.0) 4(5.7) 11(15.7) 0(0.0) 0(0.0) 6(7.8) 0(0.0) 6(7.8) 0.196 Anemia 3(4.3) 30(42.9) 21(30.0) 3(4.3) 24(34.3) 13(16.9) 33(42.9) 10(13.0) 0(0.0) 10(13.0) 0.003 Leukopenia 3(4.3) 13(18.6) 23(32.9) 19(27.1) 42(60.0) 5(6.5) 17(22.1) 25(32.5) 10(13.0) 35(45.5) 0.098 Neutropenia 1(1.4) 3(4.3) 11(15.7) 43(61.4) 54(77.1) 2(2.6) 9(11.7) 14(18.2) 38(49.4) 52(67.5) 0.204 Thrombocytopenia 2(2.9) 14(20.0) 21(30.0) 17(24.3) 38(54.3) 13(16.9) 4(5.2) 7(9.1) 6(7.8) 13(16.9) < 0.001 Anorexia 7(10.0) 13(18.6) 1(1.4) 0(0.0) 1(1.4) 6(7.8) 9(11.7) 1(1.3) 0(0.0) 1(1.3) 1.000 Nausea 12(17.1) 12(17.1) 2(2.9) 0(0.0) 2(2.9) 10(13.0) 7(9.1) 2(2.6) 0(0.0) 2(2.6) 1.000 Vomiting 8(11.4) 5(7.1) 1(1.4) 0(0.0) 1(1.4) 4(5.2) 3(3.9) 0(0.0) 0(0.0) 0(0.0) 0.473 Weight loss 0(0.0) 0(0.0) 1(1.4) 0(0.0) 1(1.4) 0(0.0) 3(3.9) 1(1.3) 0(0.0) 1(1.3) 1.000 Anxiety 1(1.4) 2(2.9) 2(2.9) 0(0.0) 2(2.9) 3(3.9) 3(3.9) 0(0.0) 0(0.0) 0(0.0) 0.473 Diarrhea 14(20.0) 7(10.0) 2(2.9) 0(0.0) 2(2.9) 8(10.4) 2(2.6) 1(1.3) 0(0.0) 1(1.3) 0.605 Fatigue 8(11.4) 11(15.7) 4(5.2) 0(0.0) 4(5.2) 18(23.4) 12(15.6) 3(3.9) 0(0.0) 3(3.9) 0.498 Infection 0(0.0) 0(0.0) 9(12.9) 3(4.3) 13(18.6) 0(0.0) 4(5.2) 5(6.5) 0(0.0) 10(13.0) 0.577 Hepatic dysfunction 1(1.4) 5(7.1) 2(2.9) 0(0.0) 2(2.9) 4(5.2) 3(3.9) 4(5.2) 2(2.6) 6(7.8) 0.621 Hyperglycemia 0(0.0) 4(5.7) 2(2.9) 0(0.0) 2(2.9) 0(0.0) 3(3.9) 4(5.2) 1(1.3) 5(6.5) 0.446 Hyponatremia 1(1.4) 0(0.0) 6(8.6) 6(8.6) 12(17.1) 0(0.0) 1(1.3) 6(7.8) 3(3.9) 9(11.7) 0.631 Hyperkalemia 0(0.0) 0(0.0) 2(2.9) 0(0.0) 2(2.9) 0(0.0) 1(1.3) 3(3.9) 0(0.0) 3(3.9) 1.000 Hypokalemia 0(0.0) 0(0.0) 3(4.3) 0(0.0) 3(4.3) 1(1.3) 0(0.0) 1(1.3) 0(0.0) 1(1.3) 0.347 aGrade means the maximum grade of toxicity. Grade 5 toxicities were developed by infection (1 patient in BP arm and 5 patients in EP arm) and disease progression (2 patients in EP arm) †P value was calculated by Fisher’s exact test for grade ≥ 3 toxicity Discussion This is the first study to evaluate the non-inferiority of the BP regimen compared to the EP regimen as a first-line treatment in patients with previously untreated ES-SCLC. Compared to the EP regimen, the RR of the BP regimen was non-inferior in the mITT population and better in the PP population. Grade 3/4 anemia and thrombocytopenia were more prevalent in the BP arm than those in the EP arm. In addition, the RDI was significantly lower in the BP arm than that in the EP arm. However, no significant differences in OS or PFS were observed between the treatment arms. Over recent decades, EP has been considered the gold-standard treatment for ES-SCLC. In 2002, a phase III trial was conducted by the Japan Clinical Oncology Group; it demonstrated the superiority of IP over EP in patients with ES-SCLC [10]. However, two subsequent randomized phase III trials failed to confirm the superiority of IP over EP in North American and Australian populations [7, 8]. Noda et al. reported that the RR and median OS were 65 % and 12.8 months in the IP arm, respectively [10]. Our results were consistent with previous studies, demonstrating a similar efficacy based on an RR of 60 % and a median OS of 12.9 months. However, we found a higher proportion of patients with ECOG PS 2 than the study by Noda et al. (16 % vs. 8 %). Moreover, the RR in the present study was approximately 10 % lower than previous phase II trials enrolling a smaller number of patients [16–18]. To date, Lim et al. have reported the highest RR (73.8 %) with the use of the BP regimen [18]. Lee et al. reported a median PFS of 6.9 months and a median OS of 19.2 months in patients receiving the BP regimen [16]. In the current study, the degree of the difference in the RR between the two treatment arms was higher in the PP population than the mITT population (18.0 % vs. 13.1 %). Moreover, our results also showed a slightly higher RR in patients with ECOG PS 0–1 (63.3 %) and those aged ≤65 years (66.7 %). These results indicate that the BP regimen may have a greater utility in younger patients with a good performance status. In the present study, favorable rates of non-hematologic toxicities were observed in the BP arm according to the safety analysis. We found that 2.9 % of the patients in the BP arm developed grade 3/4 diarrhea; this may be considerably lower than the incidence of non-hematologic toxicities reported in the IP arm of previous trials (16 %–19 %) [8, 10]. In the BP arm, neutropenia was one of the most frequent hematologic toxicities, 77 % of which were of grade 3 or 4 severity. Grade 3/4 febrile neutropenia occurred in 15.7 % of patients in the BP arm. Of these, one patient died of pneumonia. No significant differences in the incidences of neutropenia or febrile neutropenia were observed between the treatment arms. In the EP arm, there were five cases of treatment-related deaths due to pneumonia or sepsis. However, grade 3/4 anemia (34.3 % vs. 13.0 %, p = 0.003) and thrombocytopenia (54.3 % vs. 16.9 %, p < 0.001) were significantly more prevalent in the BP arm than the EP arm. The prophylactic use of G-CSF or 5 %–10 % reductions in chemotherapeutic doses may be considered to prevent hematologic toxicities. In the present study, the mean RDI was 7 % lower in the BP arm than that in the EP arm (0.79 ± 0.14 vs. 0.86 ± 0.13). A previous phase II study found that more than half of the patients presented with grade 4 neutropenia and reported an RDI of 70.1 % in the belotecan regimen and 83.0 % in the cisplatin regimen. As a result of the findings of this trial, the recommended BP regimen dose of BP was reduced by 25 % (0.5 mg/m2 for 3 days) [17]. A first limitation of this study is the non-inferiority design of this trial. Because the experimental arm demonstrated a higher hematological toxicity, superior results are required to justify the use of the experimental treatment. Because the analysis of the PP group demonstrated a better response rate with the BP regimen than the EP regimen, further studies are required to validate the superior efficacy of the BP regimen. Second limitation is the significance level in the statistical design. Although setting the significance level for one-sided, non-inferiority trials is recommended at 0.025, we chose the alpha level at 0.05 to accomplish the trial within reasonable time span. Third, the ORR which was primary endpoint of this study may not be sufficiently correlated to patient’s outcome. We selected ORR instead of OS because this study was conducted with small sample size. Fourth, the ITT set of this study was mITT population as 9 randomized patients are excluded because they did not receive the randomly assigned regimen. They were three subjects who withdrew informed consents and six who did not meet inclusion criteria. If a subject who actually did not receive any treatment is included as a subject who received treatment, then it indicates very little about the efficacy of the treatment [22]. So we defined mITT population as patients randomized and treated with at least one cycle of chemotherapy. Fifth, the interpretation of subgroup analyses for ECOG PS score, age and body weight was limited because of the issue of multiple comparisons. Finally, as the open-label nature of this study could be regarded as a flaw in a non-inferiority study. However, because these two regimens have different infusion protocols, we were unable to design this study as a blinded trial. Conclusions The results of this trial indicate that the RR with the BP regimen is non-inferior to the EP regimen in chemotherapy- and radiotherapy-naïve patients with previously untreated ES-SCLC. However, hematologic toxicities, including anemia and thrombocytopenia, were more prevalent in the BP arm. These findings strongly suggest that clinicians should be careful in prescribing the BP regimen to elderly patients or those with a poor performance status for the purpose of preventing hematologic toxicities. Further studies evaluating PFS and OS are required to validate the superiority of the BP regimen. Abbreviations ADRAdverse drug reactions AEsAdverse events ANCAbsolute neutrophil counts BPBelotecan and cisplatin CIConfidence interval CMHCochran-mantel-haenzel CRComplete response ECOGEastern cooperative oncology group EPEtoposide and cisplatin ES-SCLCExtensive-stage small cell lung cancer G-CSFGranulocyte colony-stimulating factor HRHazards ratio IPIrinotecan and cisplatin IRBInstitutional review board mITTModified intent-to-treat MTDMaximum tolerated dose NSCLCNon-small cell lung cancer OSOverall survival PDProgressive disease PFSProgression-free survival PPPer-protocol PRPartial response PSPerformance status RDIRelative dose-intensity RECISTResponse evaluation criteria in solid tumors RRResponse rate SAESerious adverse events SCLCSmall cell lung cancer SDStable disease TRDTreatment-related death ULNUpper limit of normal We are grateful for the patients, investigators and institutions involved in the current study. The current study was conducted in the following institutions: Konkuk University Medical Center (Kye-Young Lee and Hee-Joung Kim), Korea University Medical Center (Kwang-Ho In, Se-Hwa Yoo, Sang-Yeub Lee, Sung Yong Lee, Hye-Ock Kim, and Kyoung-Ju Lee), Kosin University Gospel Hospital (Tae-Won Jang, Man-Hong Jung, and Chul-Ho Ok), Daegu Catholic University Medical Center (Kyung-Chan Kim), Dongnam Institute of Radiological and Medical Science (Sun-Hyo Park), Soonchunhyang University Cheonan Hospital (Jae-Sung Choi, Young-Hun Kim, Ki-Hyun Seo, Ju-Ock Na, and Ho-Sung Lee), Yonsei University Wonju Severance Christian Hospital (Kye-Chul Shin, Suk-Joong Yong, Won-Yeon Lee, and Joung-Hwan Kim), Yeungnam University Medical Center (Kwan Ho Lee, and Jin-Hong Chung), Wonkwang University Hospital (Sei-Hoon Yang), Inje University Busan Paik Hospital (Hyun-Kyung Lee and Young-Min Lee), Chungnam National University Hospital and Cancer Institute (Sun Young Kim, Hee-Sun Park, and Jeong Eun Lee), Hallym University Sacred Heart Hospital (Seung-Hun Jang and Yong-Il Hwang) and Chonnam National University Hwasun Hospital (Young-Chul Kim, Kyu-Sik Kim, and In-Jae Oh). Funding The current study was financially supported by the Chong Kun Dang Pharmaceutical Corp. (Seoul, Korea). But the funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Availability of data and materials Identifying/confidential patient data should not be shared. Authors’ contributions IJO and YCK conceived of the study, participated in its design and coordination, contributed to the acquisition, analysis and interpretation of data and drafted the manuscript. KSK and CKP conceived of the study, participated in its design and coordination, contributed to the acquisition, analysis and interpretation of data. KHL, JHC, SYK, JEL, KCS, TWJ, HKL, KYL and SYL conceived of the study, participated in its design and coordination and contributed to the acquisition. All authors contributed to and approved the final version 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 protocol and all recruiting materials and consent form have been approved by the Institutional Review Board (IRB) at the Chonnam National University Hwasun Hospital (IRB No. HCRI 08-010-3). All changes to the protocol and recruiting materials are continuously reviewed and approved at each IRBs of 13 centers across Korea including Konkuk University Medical Center, Korea University Medical Center, Kosin University Gospel Hospital, Daegu Catholic University Medical Center, Dongnam Institute of Radiological and Medical Science, Soonchunhyang University Cheonan Hospital, Yonsei University Wonju Severance Christian Hospital, Yeungnam University Medical Center, Wonkwang University Hospital, Inje University Busan Paik Hospital, Chungnam National University Hospital and Cancer Institute and Hallym University Sacred Heart Hospital. All participants from each institutions provided written informed consent. ==== Refs References 1. Dela Cruz CS Tanoue LT Matthay RA Lung cancer: epidemiology, etiology, and prevention Clin Chest Med 2011 32 4 605 644 10.1016/j.ccm.2011.09.001 22054876 2. Youlden DR Cramb SM Baade PD The International Epidemiology of Lung Cancer: geographical distribution and secular trends J Thorac Oncol 2008 3 8 819 831 10.1097/JTO.0b013e31818020eb 18670299 3. In KH Kwon YS Oh IJ Kim KS Jung MH Lee KH Kim SY Ryu JS Lee SY Jeong ET Lung cancer patients who are asymptomatic at diagnosis show favorable prognosis: a korean Lung Cancer Registry Study Lung Cancer 2009 64 2 232 237 10.1016/j.lungcan.2008.08.005 18809225 4. Govindan R Page N Morgensztern D Read W Tierney R Vlahiotis A Spitznagel EL Piccirillo J Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: analysis of the surveillance, epidemiologic, and end results database J Clin Oncol 2006 24 28 4539 4544 10.1200/JCO.2005.04.4859 17008692 5. Jackman DM Johnson BE Small-cell lung cancer Lancet 2005 366 9494 1385 1396 10.1016/S0140-6736(05)67569-1 16226617 6. Eckardt JR von Pawel J Papai Z Tomova A Tzekova V Crofts TE Brannon S Wissel P Ross G Open-label, multicenter, randomized, phase III study comparing oral topotecan/cisplatin versus etoposide/cisplatin as treatment for chemotherapy-naive patients with extensive-disease small-cell lung cancer J Clin Oncol 2006 24 13 2044 2051 10.1200/JCO.2005.03.3332 16648504 7. Hanna N Bunn PA Jr Langer C Einhorn L Guthrie T Jr Beck T Ansari R Ellis P Byrne M Morrison M Randomized phase III trial comparing irinotecan/cisplatin with etoposide/cisplatin in patients with previously untreated extensive-stage disease small-cell lung cancer J Clin Oncol 2006 24 13 2038 2043 10.1200/JCO.2005.04.8595 16648503 8. Lara PN Jr Natale R Crowley J Lenz HJ Redman MW Carleton JE Jett J Langer CJ Kuebler JP Dakhil SR Phase III trial of irinotecan/cisplatin compared with etoposide/cisplatin in extensive-stage small-cell lung cancer: clinical and pharmacogenomic results from SWOG S0124 J Clin Oncol 2009 27 15 2530 2535 10.1200/JCO.2008.20.1061 19349543 9. Spigel DR Townley PM Waterhouse DM Fang L Adiguzel I Huang JE Karlin DA Faoro L Scappaticci FA Socinski MA Randomized phase II study of bevacizumab in combination with chemotherapy in previously untreated extensive-stage small-cell lung cancer: results from the SALUTE trial J Clin Oncol 2011 29 16 2215 2222 10.1200/JCO.2010.29.3423 21502556 10. Noda K Nishiwaki Y Kawahara M Negoro S Sugiura T Yokoyama A Fukuoka M Mori K Watanabe K Tamura T Irinotecan plus cisplatin compared with etoposide plus cisplatin for extensive small-cell lung cancer N Engl J Med 2002 346 2 85 91 10.1056/NEJMoa003034 11784874 11. Okamoto H Watanabe K Kunikane H Yokoyama A Kudoh S Asakawa T Shibata T Kunitoh H Tamura T Saijo N Randomised phase III trial of carboplatin plus etoposide vs split doses of cisplatin plus etoposide in elderly or poor-risk patients with extensive disease small-cell lung cancer: JCOG 9702 Br J Cancer 2007 97 2 162 169 10.1038/sj.bjc.6603810 17579629 12. Pujol JL Carestia L Daures JP Is there a case for cisplatin in the treatment of small-cell lung cancer? A meta-analysis of randomized trials of a cisplatin-containing regimen versus a regimen without this alkylating agent Br J Cancer 2000 83 1 8 15 10.1054/bjoc.2000.1164 10883661 13. Roth BJ Johnson DH Einhorn LH Schacter LP Cherng NC Cohen HJ Crawford J Randolph JA Goodlow JL Broun GO Randomized study of cyclophosphamide, doxorubicin, and vincristine versus etoposide and cisplatin versus alternation of these two regimens in extensive small-cell lung cancer: a phase III trial of the Southeastern Cancer Study Group J Clin Oncol 1992 10 2 282 291 1310103 14. Lee DH Kim SW Suh C Lee JS Lee JH Lee SJ Ryoo BY Park K Kim JS Heo DS Belotecan, new camptothecin analogue, is active in patients with small-cell lung cancer: results of a multicenter early phase II study Ann Oncol 2008 19 1 123 127 10.1093/annonc/mdm437 17823384 15. Kim SJ Kim JS Kim SC Kim YK Kang JY Yoon HK Song JS Lee SH Moon HS Kim JW A multicenter phase II study of belotecan, new camptothecin analogue, in patients with previously untreated extensive stage disease small cell lung cancer Lung Cancer 2010 68 3 446 449 10.1016/j.lungcan.2009.07.002 19683359 16. Lee DH Kim SW Suh C Lee JS Ahn JS Ahn MJ Park K Na II Lee JC Ryoo BY Multicenter phase 2 study of belotecan, a new camptothecin analog, and cisplatin for chemotherapy-naive patients with extensive-disease small cell lung cancer Cancer 2010 116 1 132 136 19904804 17. Hong J Jung M Kim YJ Sym SJ Kyung SY Park J Lee SP Park JW Cho EK Jeong SH Phase II study of combined belotecan and cisplatin as first-line chemotherapy in patients with extensive disease of small cell lung cancer Cancer Chemother Pharmacol 2012 69 1 215 220 10.1007/s00280-011-1689-6 21691745 18. Lim S Cho BC Jung JY Kim GM Kim SH Kim HR Kim HS Lim SM Park JS Lee JH Phase II study of camtobell inj. (belotecan) in combination with cisplatin in patients with previously untreated, extensive stage small cell lung cancer Lung Cancer 2013 80 3 313 318 10.1016/j.lungcan.2013.02.009 23510628 19. Lee DH Kim SW Bae KS Hong JS Suh C Kang YK Lee JS A phase I and pharmacologic study of belotecan in combination with cisplatin in patients with previously untreated extensive-stage disease small cell lung cancer Clin Cancer Res 2007 13 20 6182 6186 10.1158/1078-0432.CCR-07-0534 17947485 20. Artal-Cortes A Gomez-Codina J Gonzalez-Larriba JL Barneto I Carrato A Isla D Camps C Garcia-Giron C Font A Meana A Prospective randomized phase III trial of etoposide/cisplatin versus high-dose epirubicin/cisplatin in small-cell lung cancer Clin Lung Cancer 2004 6 3 175 183 10.3816/CLC.2004.n.031 15555219 21. Newcombe RG Bender R Implementing GRADE: calculating the risk difference from the baseline risk and the relative risk Evid Based Med 2014 19 1 6 8 10.1136/eb-2013-101340 23970740 22. Gupta SK Intention-to-treat concept: A review Perspect Clin Res 2011 2 3 109 112 10.4103/2229-3485.83221 21897887

PMC005xxxxxx/PMC5002147.txt

==== Front BMC Infect DisBMC Infect. DisBMC Infectious Diseases1471-2334BioMed Central London 178810.1186/s12879-016-1788-4Research ArticleClostridium difficile infection health disparities by race among hospitalized adults in the United States, 2001 to 2010 Argamany Jacqueline R. jargamany@utexas.edu 12Delgado Andrew andrewdelgado@utexas.edu 12Reveles Kelly R. kdaniels46@utexas.edu 121 The University of Texas College of Pharmacy, 2409 University Avenue, A1900, Austin, TX 78712 USA 2 The University of Texas Health Science Center Pharmacotherapy and Education Research Center, 7703 Floyd Curl Drive, MC-6220, San Antonio, TX 78229 USA 27 8 2016 27 8 2016 2016 16 1 45410 9 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 Recognition of health disparities in Clostridium difficile infection (CDI) is an initial step toward improved resource utilization and patient health. The purpose of this study was to identify health disparities by black vs. white race among hospitalized adults with CDI in the United States (U.S.) over 10 years. Methods This was a retrospective analysis of the U.S. National Hospital Discharge Surveys from 2001 to 2010. Eligible cases included adults with an ICD-9-CM code for CDI (008.45). Patients with missing race or “other race” were excluded. The primary outcome, CDI incidence, was calculated as CDI discharges per 1,000 total discharges. Data weights were used to determine national estimates. Secondary outcomes included in-hospital mortality, hospital length of stay (LOS), and severe CDI. Comparisons were made using bivariable analyses. Race was assessed as an independent risk factor for CDI outcomes using logistic regression or proportional hazards models. Results These data represent 1.7 million CDI discharges, where 90 % of patients were identified as white and 10 % black. Blacks differed from whites with respect to all baseline characteristics (p <0.0001). CDI incidence was significantly higher in whites compared to blacks (7.7/1,000 discharges vs. 4.9/1,000 discharges, p < 0.0001). Blacks had higher mortality (7.4 % vs. 7.2 %, p < 0.0001), LOS >7 days (57 % vs. 52 %, p < 0.0001), and severe CDI (24 % vs. 19 %, p < 0.0001). In multivariable analyses, black race was a positive predictor of mortality (OR 1.12, 95 % CI 1.09–1.15) and severe CDI (OR 1.09, 95 % CI 1.07–1.11), and negative predictor for hospital LOS (OR 0.93, 95 % CI 0.93–0.94). Conclusions CDI incidence was higher for white patients; however, black race was independently associated with mortality and severe CDI. Keywords RaceDisparityClostridium difficileissue-copyright-statement© The Author(s) 2016 ==== Body Background Efforts are being made to eliminate health disparities in the United States (U.S.); however, disparities due to race, ethnicity, geographics, and socioeconomics continue to exist [1–4]. Prior studies have reported reduced access to primary healthcare services among minority patients, potentiating the likelihood of suboptimal patient outcomes [5–8]. Infectious diseases are the second-leading contributor to racial disparities in healthcare after cardiovascular disease [9]. Health disparities have been specifically reported in infectious diseases, though studies are limited. Compared to white patients, black patients have an increased incidence and severity of sepsis and receive a lower quality of care for severe infections [10–12]. Few studies have examined the impact of disparities on other infections, such as Clostridium difficile infection (CDI) [3, 13]. CDI is a common infection with increasing incidence in U.S. hospitals as well as the community [14, 15]. Symptoms of infection range from mild, uncomplicated diarrhea to more severe manifestations with complications including sepsis, renal failure, ileus, toxic megacolon, perforated intestine, or death [16–18]. The incidence of and outcomes from CDI might differ by race due to factors that influence the gastrointestinal microbiome, such as medications and diet, and socioeconomic factors, including insurance status and access to care. While research suggests black patients are less likely to receive broad-spectrum antibiotics, they are more likely to have a longer admission to a hospital emergency department and have higher rates of hospital readmission, which could impact patient health outcomes [19–22]. Despite these associations, few studies have evaluated racial health disparities in CDI. Recognition of health disparities in CDI is an initial step towards more targeted resource utilization and improved patient health. The goal of this study was to identify health disparities by black vs. white race in CDI incidence and health outcomes among hospitalized adults with CDI in the U.S. over a 10-year period. Methods Data source This study utilized data from the Centers for Disease Control and Prevention’s National Hospital Discharge Survey (NHDS). The NHDS is a national probability sample of non-federal, short-stay hospital discharges in the U.S. A complex, three-stage sampling methodology allows the user to apply data weights to derive national estimates representative of the U.S. population [23]. The survey data include patient demographics, such as age, gender, self-reported race, and marital status, as well as year of discharge, payment sources, geographic region, hospital length of stay (LOS), and hospital discharge status. Diagnoses and procedures are also reported as International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes. NHDS data have previously been used in several infectious diseases epidemiological studies, including healthcare-associated infections [14, 24, 25]. Study design This was a retrospective analysis of all patients discharged from U.S. hospitals from 2001 to 2010. Eligible cases included adults at least 18 years of age with a principal or secondary ICD-9-CM discharge diagnosis code for CDI (008.45). Patients with missing race or “other race” were excluded. Patient baseline characteristics were classified based on the categories provided in the NHDS for patient sex, hospital size (6–99 beds, 100–199 beds, 200–299 beds, 300–499 beds, or ≥500 beds), hospital ownership (proprietary, government, or nonprofit), and admission type (emergency, urgent, or elective). Other patient characteristics were classified by limited definitions designed to encompass NHDS categories: race (white, black, and other), expected primary source of payment (private, Medicare, Medicaid, self-pay, and other), and admission source (emergency room, transfer, referral, and other). Health outcomes in this study included in-hospital mortality, hospital LOS, and any severe CDI. The “Discharge Status” item of the NHDS was used to determine patient mortality. This represents all-cause, in-hospital mortality for patients with CDI. Hospital LOS was extracted from the “Days of Care” item of the NHDS and was presented as medians (interquartile ranges). In this study, severe CDI was indicated by cases with a principal or secondary ICD-9-CM code for at least one of the following: septicemia (038.x), septic shock (785.52), acute renal failure (584.x), toxic megacolon (558.2), prolonged ileus (560.1), perforated intestine (569.83), or colectomy (45.7x). Statistical analyses First, baseline patient demographics were summarized using medians (interquartile ranges) for continuous variables and counts (percentages) for categorical variables. All baseline characteristics were assessed for multicollinearity using the Spearman rank correlation. Correlation coefficients were then converted to variance inflation factors (VIF) using the following equation: VIF = 1/(1-R2). Two variables were considered highly correlated if they had a VIF >10 and were statistically significant at an alpha <0.0001. We compared baseline characteristics between races using bivariable analyses calculated using the chi-square test for categorical variables and Wilcoxon rank-sum test for continuous variables. Next, we determined the overall CDI incidence rate using CDI discharges as the numerator, as identified in our cohort, and total discharges as the denominator. Total discharges were derived from the composite NHDS data, which include all CDI and non-CDI patients. Incidence by race was calculated as CDI discharges per race category divided by total discharges per race category. Data weights were applied to derive national estimates and incidence rates were presented as CDI discharges per 1,000 total discharges. CDI incidence, patient mortality, hospital LOS, and severe CDI were characterized overall and by race. CDI incidence was compared by race using the z-test. Health outcomes were compared using the chi-square or Wilcoxon rank sum test. Risk for CDI health outcomes was analyzed using multivariable logistic regression to calculate adjusted odds ratios (aOR) and 95 % confidence intervals (CI) for mortality and any severe CDI using white race as the reference category. A Cox proportional hazards model was used to calculate the aOR and 95 % CI for hospital LOS, which allowed us to censor those CDI patients who died in the hospital. The regression models controlled for patient-specific variables, including age ≥65 years, sex, geographic region, hospital size, hospital ownership, other infectious diseases as indicated by ICD-9-CM codes (skin infection, pneumonia, bacteremia, device-related infections, urinary tract infection, and endocarditis), principal or secondary CDI diagnosis, other common principal diagnoses (e.g., cancer, congestive heart failure, chronic obstructive pulmonary disease [COPD], myocardial infarction, diabetes, cerebrovascular disease, and human immunodeficiency disease [HIV]), principal payment source, admission type, and admission source. The mortality and hospital LOS models also included severe CDI as a covariate. Statistical significance was indicated by p < 0.0001 due to the large sample size provided by the NHDS. JMP 10.0® (SAS Corp, Cary, NC) was used for all statistical comparisons. Results Baseline characteristics The patient baseline characteristics are provided in Table 1. Overall, these data represent approximately 1.7 million CDI discharges from U.S. hospitals over the study period. Of these patients, 90 % were white and 10 % were black. Approximately 500,000 patients were excluded because of a missing race code or classification as “other race.” Overall, patients were predominately female (59 %) and at least 65 years old (71 %). Black and white patients with CDI significantly differed with respect to patient age, sex, geographic region, hospital size and ownership, principal payment source, admission type and source, and principal diagnoses. White patients were a median (interquartile range) 76 (63–83) years of age, while black patients were much younger [65 (50–78) years]. Black patients were more likely than white patients to be discharged from a hospital with over 500 beds (25 % and 11 %, respectively), and were more likely to have Medicaid as their principal payment source (17 % and 5 %, respectively). Additionally, black patients were also more likely to have an emergent admission type (78 % and 65 %, respectively). No baseline characteristics had a VIF >2, indicating low multicollinearity among variables.Table 1 Baseline characteristics Demographic Overall (n = 1,676,903) White (n = 1,503,189) Black (n = 173,714) P-Valuea Age (years), median (IQR) 75 (61–83) 76 (63–83) 65 (50–78) <0.0001 Age ≥65 years, % 71 73 50 <0.0001 Female sex, % 59 59 58 <0.0001 Geographic region, % <0.0001  Northeast 32 33 28  Midwest 18 18 15  South 37 35 49  West 13 14 8 Hospital size, % <0.0001  6–99 beds 20 21 12  100–199 beds 23 24 11  200–299 beds 23 23 21  300–499 beds 21 21 31  500+ beds 13 11 25 Hospital ownership, % <0.0001  Proprietary 14 15 11  Government 10 9 15  Nonprofit 76 76 74 Principal payment source, % <0.0001  Medicare 69 71 58  Medicaid 6 5 17  Private 21 21 19  Self-pay 2 1 3  Other 2 2 3 Admission type, % <0.0001  Emergency 67 65 78  Urgent 20 21 16  Elective 13 14 6 Admission source, % <0.0001  Emergency room 59 59 67  Transfer 16 16 13  Referral 17 17 14  Other 8 8 6 Selected principal diagnoses, %   Clostridium difficile infection 32 32 31 <0.0001  Pneumonia 4 5 3 <0.0001  Urinary tract infection 3 2 5 <0.0001  Cancer 3 3 3 <0.0001  Congestive heart failure 2 2 2 <0.0001  COPD 2 2 1 <0.0001  Myocardial infarction 1 1 1 0.0294  Diabetes 1 1 2 <0.0001  Cerebrovascular disease 1 1 2 <0.0001  HIV <1 <1 2 <0.0001 IQR interquartile range, COPD chronic obstructive pulmonary disease, HIV human inmmunodeficiency virus a P-values reflect comparisons between white and black patients CDI incidence The overall CDI incidence was 7.3 CDI discharges per 1,000 total discharges. CDI incidence was significantly higher for white patients (7.7 CDI discharges per 1,000 white discharges) as compared to black patients (4.9 CDI discharges per 1,000 black discharges) (p < 0.0001). From 2001 to 2010, the incidence of CDI in black and white patients increased (Fig. 1). The incidence was highest in 2009 for white patients (9.6 CDI discharges per 1,000 white discharges), and highest in 2008 for black patients (6.7 CDI discharges per 1,000 black discharges).Fig. 1 Clostridium difficile infection (CDI) incidence per year for white and black patients, 2001–2010 Health outcomes Health outcomes are depicted in Table 2. The median (interquartile range) hospital LOS for all hospitalized patients with CDI was 8 (5–14) days and all-cause, in-hospital mortality was 7.3 %. Hospital LOS and mortality were significantly higher for black patients (9 days and 7.4 %, respectively) as compared to white patients (8 days and 7.2 %, respectively) (p < 0.0001 for each comparison).Table 2 Health outcomes Outcome Overall (n = 1,676,903) White (n = 1,503,189) Black (n = 173,714) P-Valuea CDI incidence 7.3 7.7 4.9 <0.0001 Mortality, % 7.3 7.2 7.4 <0.0001 Hospital LOS, median (IQR) 8 (5–14) 8 (4–14) 9 (5–16) <0.0001 Hospital LOS >7 days, % 52 52 57 <0.0001 Any severe CDI,b % 20 19 24 <0.0001  Sepsis 13 12 16  Renal failure 16 15 21 CDI Clostridium difficile infection, IQR interquartile range, LOS length of stay a P-values reflect comparisons between white and black patients bPercentages not shown for severe CDI outcomes present in <5 % of the population (shock, megacolon, ileus, perforated intestine, and colectomy) Black race was an independent risk factor for all health outcomes. Compared to white patients, black patients had significantly greater mortality (aOR 1.12, 95 % CI 1.09–1.15, p < 0.0001) and were at significantly greater risk for severe CDI (aOR 1.09, 95 % CI 1.07–1.11, p < 0.0001). Other significant predictors of mortality included: age ≥ 65 years (aOR 2.67, 95 % CI 2.60–2.75), female sex (aOR 1.08, 95 % CI 1.06–1.10), urinary tract infection (aOR 1.36, 95 % CI 1.35–1.37), sepsis (aOR 3.54, 95 % CI 3.47–3.60), and any severe CDI (aOR 2.12, 95 % CI 2.08–2.16). Other significant predictors of severe CDI included: age ≥ 65 years (aOR 1.39, 95 % CI 1.37–1.41), private hospital admission (aOR 1.61, 95 % CI 1.57–1.65), urinary tract infection (aOR 1.36, 95 % CI 1.35–1.37), and emergency admission (aOR 1.53, 95 % CI 1.51–1.55). After censoring for inpatient mortality, black patients had a significantly shorter hospital LOS compared to white patients (aOR 0.94, 95 % CI 0.93–0.94, p < 0.0001). Discussion In this retrospective cohort study of approximately 1.7 million hospital discharges, we identified disparities in CDI incidence and health outcomes by race. Remarkably, CDI incidence among white patients was 57 % greater than that of black patients. As CDI rates continue to increase and gaps in patient access to care continue to narrow, this disparity may change. Importantly, CDI mortality and severity were significantly worse for black patients. To our knowledge, this is the first study to find that black race is associated with poorer health outcomes in CDI. Because of the small absolute differences in mortality and severity rates between white and black CDI patients, the clinical significance of our findings is unknown, but is worth addressing in future studies due to the high cost burden associated with CDI [15]. Knowledge of healthcare disparities can be used to target resource utilization to improve patient health outcomes and eliminate disparities in the future [4]. Disparities attributed to race have been reported with respect to primary care access, as well as in surgery, sepsis, and pneumonia [5–8, 10–12, 26]. However, little research has focused on disparities associated with healthcare-associated infections [13]. Our study adds to the few prior studies examining the occurrence and recurrence of CDI where race was assessed as an independent risk factor [13, 27, 28]. In 2012, Murphy et al. published the first study to note that race and ethnicity were predictive of CDI [27]. One year later, Freedberg et al. incidentally found a higher risk of recurrent CDI in black patients in a study designed to assess the risk of CDI recurrence associated with proton pump inhibitors [28]. More recently, Bakullari et al. demonstrated significant racial and ethnic disparities in the rate of occurrence of healthcare-associated infections [13]. While the study did not find a statistically significant difference in the rate of CDI between white and black patients, it examined a much smaller population using the Medicare Patient Safety Monitoring System. The reason for such disparities in incidence and outcomes in CDI is likely multifactorial. Among patients with CDI, black patients significantly differed from white patients with respect to all patient demographics. For example, black patients with CDI were younger than white patients and had greater use of Medicaid as a principal payment source. Considering older age is a known risk factor for CDI, the younger age of black CDI patients could have contributed to the lower CDI incidence seen for blacks in this study [15, 17]. Older age is associated with more comorbidities and health care exposures, which could increase for the risk for CDI as well as increase the likelihood for early diagnosis. The significant differences in principal payment source in our study are also likely a reflection of the younger age of the black population with CDI, as compared to the typical elderly CDI population, which relied heavily on Medicare. While this study was unable to assess for relationships between all co-morbid conditions and racial differences in CDI, the chronic, co-morbid conditions disparately affecting blacks may also increase the risk for CDI and predispose this population to overall poorer health outcomes [10]. Furthermore, disparity research suggests the most effective resource use would aim at eliminating the top racial disparities, including hypertension, diabetes, and HIV [9, 10]. Factors such as health insurance, access to care, quality of care, or health-related behavior may also contribute to this disparity [9]. According to the U.S. Census Bureau in 2012, 40.6 % of blacks relied on Medicaid or other publicly-funded insurance compared to 29.3 % of non-Hispanic whites [29]. Additionally, 17.2 % of blacks were uninsured compared to 10.4 % of non-Hispanic whites [29]. Furthermore, black patients are more likely to seek care at higher-volume hospitals than white patients and often receive care at hospitals providing a lower quality of care [12, 26, 30]. According to the Agency for Healthcare Research and Quality (AHRQ) in 2010, black patients had worse access to care than white patients for one-third of core measures and were less likely to have a specific source of ongoing care (84.7 % and 86.3 %, respectively) [31]. Decreased access to care could explain the significantly increased rate of hospital admissions via emergency department for the black population seen in our study. Differences in medication use might also contribute to CDI disparities. White patients are more likely to receive prescription antibiotics, broad-spectrum antibiotics, and proton pump inhibitors, putting them at greater risk for disruption of the normal gastrointestinal flora, likely contributing to differential CDI incidence [19, 20, 28]. Surveys have also shown that white patients are more accepting of new prescription drugs than blacks [32]. Considering race is more relevant as a social construct than a biological construct in the United States, it is unlikely the disparities noted here are due to genetic differences in the population; however, this cannot be ruled out [33–35]. A previous study by Esper et al. suggested a genetic polymorphism in black patients in the human Toll-like receptor 2 (TLR2) could directly create a race-specific alteration in host response to Gram-positive pathogens [10]. Distribution differences in the polymorphic functional allele were previously found between African Americans and Caucasians, and given differential TLR2 promoter activity in response to interferon-gamma, contrastive pathogenesis of Gram-positive infectious organisms is possible between races [36]. Considering Clostridium difficile is a Gram-positive bacterium, this may warrant further research. This study has potential limitations. First, this study relied on administrative data using ICD-9-CM coding to identify CDI and other diagnoses; therefore, it may not fully capture all CDI cases. A prior study evaluating the use of the ICD-9-CM code to identify CDI found high sensitivity (78 %) and specificity (99.7 %) compared to microbiological data [37]. This study design also precluded us from assessing other markers of disease severity, such as the McCabe or Charlson Comorbidity scores. Additionally, race was self-reported and may not accurately represent the patient population. The exclusion of patients with missing race or “other race” could limit the generalizability of the results. The NHDS excludes federal hospitals and long-term care hospitals, potentially limiting the application of our results to those patient populations. Data regarding patient ethnicity, socioeconomic status, treatment(s), prior medication(s), care access, and other patient demographics were unavailable for this study. The lack of control for these factors may have influenced results. Future studies evaluating CDI disparities should consider health insurance status or funding source, education level, medication and disease history, as well as health-related behavior issues in order to evaluate the broader scope of the problem. Conclusions Despite the higher incidence of CDI among white patients, black patients had significantly higher mortality and risk for severe CDI compared to white patients. Further studies are needed to provide insights into the basis for this observation. Knowledge of existing health disparities can be used to better direct resources and improve patient care until such disparities can be eliminated or prevented. Abbreviations AHRQAgency for healthcare research and quality aORAdjusted odds ratio CDIClostridium difficile infection CIConfidence interval COPDChronic obstructive pulmonary disease HIVHuman immunodeficiency virus ICD-9-CMInternational classification of diseases, 9th rev, clinical modification LOSLength of stay NHDSNational hospital discharge survey TLR2Toll-like receptor 2 USUnited States VIFVariance inflation factor Not applicable. Funding No funding was provided for the conduct of this study, KR is supported by the National Institutes of Health/National Institute on Aging San Antonio Claude D. Pepper Older Americans Independence Center (1P30AG044271-01A1) Career Development (KL2) Program. Availability of data and materials Available for public access through the Centers for Disease Control and Prevention. Authors’ contributions KR participated in the design of the study and performed the statistical analyses. JA and AG drafted the manuscript. All authors read and approved the final manuscript. Authors’ information KR is an Assistant Professor at The University of Texas College of Pharmacy. JA and AG are graduate students at The University of Texas College of Pharmacy. 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. U.S. Department of Health and Human Services, Office of Disease Prevention and Health Promotion. Healthy People 2020. http://www.healthypeople.gov/2020/default.aspx Accessed 17 Dec 2014. 2. American Medical Association. Health and ethics policy: H-350.974 racial and ethnic disparities in health care. http://www.ama-assn.org/ad-com/polfind/Hlth-Ethics.pdf. Accessed 23 Aug 2016. 3. Centers for Disease Control and Prevention CDC health disparities and inequalities report - United States, 2013 MMWR Morb Mortal Wkly Rep 2013 62 S3 1 86 23302815 4. Nelson A Unequal treatment: confronting racial and ethnic disparities in health care J Natl Med Assoc 2002 94 8 666 8 12152921 5. Shi L Chen CC Nie X Zhu J Hu R Racial and socioeconomic disparities in access to primary care among people with chronic conditions J Am Board Fam Med 2014 27 2 189 98 10.3122/jabfm.2014.02.130246 24610181 6. Mahmoudi E Jensen GA Diverging racial and ethnic disparities in access to physician care: comparing 2000 and 2007 Med Care 2012 50 4 327 34 10.1097/MLR.0b013e318245a111 22388557 7. Levine DA Neidecker MV Kiefe CI Karve S Williams LS Allison JJ Racial/ethnic disparities in access to physician care and medications among US stroke survivors Neurology 2011 76 1 53 61 10.1212/WNL.0b013e318203e952 21084692 8. Shi L Lebrun LA Tsai J Access to medical care, dental care, and prescription drugs: the roles of race/ethnicity, health insurance, and income South Med J 2010 103 6 509 16 10.1097/SMJ.0b013e3181d9c2d8 20710132 9. Wong MD Shapiro MF Boscardin WJ Ettner SL Contribution of major diseases to disparities in mortality N Engl J Med 2002 347 20 1585 92 10.1056/NEJMsa012979 12432046 10. Esper AM Moss M Lewis CA Nisbet R Mannino DM Martin GS The role of infection and comorbidity: Factors that influence disparities in sepsis Crit Care Med 2006 34 10 2576 82 10.1097/01.CCM.0000239114.50519.0E 16915108 11. Barnato AE Alexander SL Linde-Zwirble WT Angus DC Racial variation in the incidence, care, and outcomes of severe sepsis: analysis of population, patient, and hospital characteristics Am J Respir Crit Care Med 2008 177 3 279 84 10.1164/rccm.200703-480OC 17975201 12. Mayr FB Yende S D’Angelo G Barnato AE Kellum JA Weissfeld L Do hospitals provide lower quality of care to black patients for pneumonia? Crit Care Med 2010 38 3 759 65 10.1097/CCM.0b013e3181c8fd58 20009756 13. Bakullari A Metersky ML Wang Y Eldridge N Eckenrode S Pandolfi MM Racial and ethnic disparities in healthcare-associated infections in the United States, 2009–2011 Infect Control Hosp Epidemiol 2014 35 S3 S10 S6 10.1086/677827 25222888 14. Reveles KR Lee GC Boyd NK Frei CR The rise in Clostridium difficile infection incidence among hospitalized adults in the United States: 2001–2010 Am J Infect Control 2014 42 10 1028 32 10.1016/j.ajic.2014.06.011 25278388 15. Lucado J, Gould C, Elixhauser A. Clostridium difficile infections in hospital stays, 2009. HCUP Statistical Brief 124. 2012. http://www.hcup-us.ahrq.gov/reports/statbriefs/sb124.pdf. Accessed 23 Aug 2016. 16. Dallal RM Harbrecht BG Boujoukas AJ Sirio CA Farkas LM Lee KK Fulminant Clostridium difficile: an underappreciated and increasing cause of death and complications Ann Surg 2002 235 3 363 72 10.1097/00000658-200203000-00008 11882758 17. Cohen SH Gerding DN Johnson S Kelly CP Loo VG McDonald LC Clinical practice guidelines for Clostridium difficile infection in adults: 2010 update by the society for healthcare epidemiology of America (SHEA) and the infectious diseases society of America (IDSA) Infect Control Hosp Epidemiol 2010 31 5 431 55 10.1086/651706 20307191 18. Loo VG Poirier L Miller MA Oughton M Libman MD Michaud S A predominantly clonal multi-institutional outbreak of Clostridium difficile-associated diarrhea with high morbidity and mortality N Engl J Med 2005 353 23 2442 9 10.1056/NEJMoa051639 16322602 19. Gahbauer AM Gonzales ML Guglielmo BJ Patterns of antibacterial use and impact of age, race/ethnicity, and geographic region on antibacterial use in an outpatient medicaid cohort Pharmacotherapy 2014 34 7 677 85 10.1002/phar.1425 24753176 20. Steinman MA Landefeld CS Gonzales R Predictors of broad-spectrum antibiotic prescribing for acute respiratory tract infections in adult primary care JAMA 2003 289 6 719 25 10.1001/jama.289.6.719 12585950 21. Pines JM Russell Localio A Hollander JE Racial disparities in emergency department length of stay for admitted patients in the United States Acad Emerg Med 2009 16 5 403 10 10.1111/j.1553-2712.2009.00381.x 19245372 22. Joynt KE Orav EJ Jha AK Thirty-day readmission rates for Medicare beneficiaries by race and site of care JAMA 2011 305 7 675 81 10.1001/jama.2011.123 21325183 23. Dennison C, Pokras R. Design and operation of the National Hospital Discharge Survey: 1988 redesign. Vital Health Stat. 2000;1(39):1–42. 24. McDonald LC Owings M Jernigan DB Clostridium difficile infection in patients discharged from US short-stay hospitals, 1996–2003 Emerg Infect Dis 2006 12 3 409 15 10.3201/eid1205.051064 16704777 25. Daniels KR Frei CR The United States’ progress toward eliminating catheter-related bloodstream infections: incidence, mortality, and hospital length of stay from 1996 to 2008 Am J Infect Control 2013 41 2 118 21 10.1016/j.ajic.2012.02.013 22748842 26. Lucas FL Stukel TA Morris AM Siewers AE Birkmeyer JD Race and surgical mortality in the United States Ann Surg 2006 243 2 281 6 10.1097/01.sla.0000197560.92456.32 16432363 27. Murphy CR Avery TR Dubberke ER Huang SS Frequent Hospital Readmissions for Clostridium difficile infection and the impact on estimates of hospital-associated C. difficile burden Infect Control Hosp Epidemiol 2012 33 1 20 8 10.1086/663209 22173518 28. Freedberg DE Salmasian H Friedman C Abrams JA Proton pump inhibitors and risk for recurrent Clostridium difficile infection among inpatients Am J Gastroenterol 2013 108 11 1794 801 10.1038/ajg.2013.333 24060760 29. DeNavas-Walt C Proctor BD Smith JC Income, poverty, and health insurance coverage in the United States: 2012 2013 Washington Current Population Reports. U.S. Census Bureau 60 245 30. Jha AK Orav EJ Li Z Epstein AM Concentration and quality of hospitals that care for elderly black patients Arch Intern Med 2007 167 11 1177 82 10.1001/archinte.167.11.1177 17563027 31. Disparities in healthcare quality among racial and ethnic minority groups: selected findings from the 2010 National Healthcare Quality and Dispar. October 2014. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/research/findings/nhqrdr/nhqrdr10/minority.html. Accessed 17 Dec2014 32. Groeneveld PW Sonnad SS Lee AK Asch DA Shea JE Racial differences in attitudes toward innovative medical technology J Gen Intern Med 2006 21 6 559 63 10.1111/j.1525-1497.2006.00453.x 16808736 33. Cooper R David R The biological concept of race and its application to public health and epidemiology J Health Polit Policy Law 1986 11 1 97 116 10.1215/03616878-11-1-97 3722786 34. Krieger N Shades of difference: theoretical underpinnings of the medical controversy on black/white differences in the United States, 1830–1870 Int J Health Serv 1987 17 2 259 78 10.2190/DBY6-VDQ8-HME8-ME3R 3294621 35. Williams DR The concept of race in Health Services Research: 1966 to 1990 Health Serv Res 1994 29 3 261 74 8063565 36. Yim JJ Ding L Schaffer AA Park GY Shim YS Holland SM A microsatellite polymorphism in intron 2 of human Toll-like receptor 2 gene: functional implications and racial differences FEMS Immunol Med Microbiol 2004 40 2 163 9 10.1016/S0928-8244(03)00342-0 14987735 37. Dubberke ER Reske KA McDonald LC Fraser VJ ICD-9 codes and surveillance for Clostridium difficile-associated disease Emerg Infect Dis 2006 12 10 1576 9 10.3201/eid1210.060016 17176576

PMC005xxxxxx/PMC5002148.txt

==== Front Transl NeurodegenerTransl NeurodegenerTranslational Neurodegeneration2047-9158BioMed Central London 6310.1186/s40035-016-0063-3ReviewThe biomarkers of immune dysregulation and inflammation response in Parkinson disease Chen Li victoria28@126.com 1Mo Mingshu mishmoth@outlook.com 1Li Guangning victoria28@126.com 7Cen Luan cenluan974@sina.com 2Wei Lei weilei921@163.com 13Xiao Yousheng xys135@126.com 2Chen Xiang wuliancx@163.com 1Li Shaomin shaomin_li@hms.harvard.edu 4Yang Xinling Poplar862@sohu.com 5Qu Shaogang sgq9528@163.com 6Xu Pingyi pingyixu@sina.com 121 Department of Neurology, The First Affiliated Hospital of Guangzhou Medical University, Guangdong, 510120 China 2 Department of Neurology, The First Affiliated Hospital of Sun Yat-sen University, Guangdong, 510080 China 3 Department of Neurology, The Third Affiliated Hospital of Sun Yat-sen University, Guangdong, 510082 China 4 Ann Romney Center for Neurologic Disease, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115 USA 5 Department of Neurology, The Third Affiliated Hospital of Xinjiang Medical University, Urumqi, 830011 China 6 Department of Blood Transfusion, The Fifth Affiliated Hospital, Southern Medical University, Guangzhou, Guangdong 510900 China 7 Department of Neurology, The Affiliated Huadu Hospital of Southern Medical University, Guangzhou, 510800 China 26 8 2016 26 8 2016 2016 5 1 1618 3 2016 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.Parkinson’s disease (PD) is referring to the multi-systemic α-synucleinopathy with Lewy bodies deposited in midbrain. In ageing, the environmental and genetic factors work together and overactive major histocompatibility complex pathway to regulate immune reactions in central nerve system which resulting in neural degeneration, especially in dopaminergic neurons. As a series of biomarkers, the human leukocyte antigen genes with its related proteomics play cortical roles on the antigen presentation of major histocompatibility complex molecules to stimulate the differentiation of T lymphocytes and i-proteasome activities under their immune response to the PD-related environmental alteration and genetic variation. Furthermore, dopaminergic drugs change the biological characteristic of T lymphatic cells, affect the α-synuclein presentation pathway, and inhibit T lymphatic cells to release cytotoxicity in PD development. Taking together, the serum inflammatory factors and blood T cells are involved in the immune dysregulation of PD and inspected as the potential clinic biomarkers for PD prediction. Keywords Parkinson’s diseaseα-synucleinopathyInflammationBiomarkersthe State Key Development Program for Basic Research of China2011CB510000Xu Pingyi http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China8147129281430021Xu Pingyi National Natural Science Foundation of China (CN)U1503222Yang Xinling keypoint Science Foundation of Guangdong of China2015A030311021Xu Pingyi technology project of Guangzhou 20151260Xu Pingyi assisting research project of science and technology for Xinjiang201591160Yang Xinling issue-copyright-statement© The Author(s) 2016 ==== Body Background The raised incidence of Parkinson’s disease (PD) becomes a serious issue in an aged society [1]. It is known that PD patients in Asia is nearby 11.3 % of movement disorders, slightly lower than that in north America which is up to 13.6 % and 16.6 % of Europe [1]. Based on the characteristic protein conformation and function, central nervous degenerative diseases involved in movement disorders can be clinically classed as α-synucleinopathy, Tauopathy and TDP-43 proteinopathy in which PD belongs to α-synucleinopathy [2]. According to the latest clinic diagnosis criteria of MDS (Movement Disorder Association 2015), the diagnosis of PD need to meet the following criteria: rest tremor and bradykinesia in limbs, clinic symptoms of PD are effectively improved by L-dopa, disease duration is usually accompanied by non-motor symptoms especially in early stage, and discrimination from other neural disease [3]. The various types of α-synuclein manifest mainly as monomers, oligomers, ribbons and fibrils, and a-synuclein diffuses through blood brain barrier, spread around whole nervous system, even invade other tissues or organs such as nasal mucosa, skin, kidney, liver, heart, intestinal tract and salivary glands [4, 5]. As well-known, aging is confirmed to be an important risk to PD [6]. Other risk factors include harmful environment factors and deficient genetic factors [6]. It's worth noting that environment factors such as MPTP, LPS, rotenone or other organic chemicals not only impair dopaminergic (DA) neurons directly and render the epigenetic variation of DNA methylation, but also activate the secondary inflammation/immunity reaction and significantly increase the PD morbidity [7, 8]. Immunologic dysregulation in PD Recently immunologic systems were found in central neural system (CNS) which involved with PD [9]. The classic concepts pointed out that central immune system is not directly connected with extracranial system. However, latest reports demonstrated that lymphatic ducts hind in the meninge and link to deep cervicalvenous system in which abundant T, B lymphocytes and dendritic cells are response to environment and genetic variation [10]. According to recent dynamic experiments, the lymphatic system plays an important role in the drainage of spinal fluid to lymphatic duct oriented into deep cervical venous system [10]. This discovery revealed that there is a close linkage between the meninge lymphatic system and α-synuclein transmission. The brain autopsy from PD patients and PD animal models showed that abnormal accumulation of microglia were activated in brain tissues, especially in nigrostriatal area [11]. Under the abnormal inflammation condition, macrophages pathologically diffuse from blood vessels to CNS and transform into microglias contributed to the pathological development of intracranial neural diseases [12]. Meanwhile the activated T lymphocytes were demonstrated to move out from central lymph ducts to attack central neurons, neural myelination or fibers, which resulting in multiple neuropathy as multiple sclerosis like symptoms [13]. In 6-hydroxydopamine-PD rats, there were abundant CD3+, CD4+, CD8+ T lymph cells migrated from blood vessels into substantia nigra (SN) to attack DA neurons [13, 14]. Therefore, it is reasonable to suppose these lymph cells could be the potential biomarkers for the evaluation of PD progression [14]. Immunoproteasome system in PD The α-synuclein aggregation is a critical ignition role for the immune disorder in PD [15]. In aging process, the progressive dysfunction of T, B lymph cells and macrophages were related with immune elimination which gradually gave rise to α-synuclein aggregation and neurons degeneration [16, 17]. Furthermore, due to the decreased activity of proteasome, its components as β1,β2 and β5 subunits which compose of 20S proteasome will be replaced by homogenous β1i (large multi-functional protease 2, LMP2), β2i (multi catalytic endopeptidase complex-like-1, MECL-1) and β5i (large multi-functional protease 7, LMP7) subunits and develop a new immunoproteasome system named as i-proteasome [18]. The crystal structure showed that i-proteasomes have more enzyme domains, stronger enzymatic activity, and more effective capacity to degrade the α-synuclein proteins. Particularly, i-proteasome presents the degraded peptides as antigens to special T cells [19]. These antigens combined with major histocompatibility complex molecules (MHC) to ignite another pattern of protein clearance or abnormal neural inflammation [19]. It was reported that i-proteasome was up-regulated in senile hippocampus which involves in the clearance of accumulated amyloid precursor protein (APP) in Alzheimer's disease (AD) [20]. In epilepsy, the i-proteasomes activated by oxidative damaged proteins can maintain the stability of protein metabolism in brain [21]. There are seven i-proteasome cleavage sites in the amino acid sequence of α-synuclein (sequence No.100, 8, 10, 55, 125, 45 and 18, respectively) [22]. Among them, No. 10 and 45 amino acid locate at N terminal named as KTKEGV sequence [22]. This KTKEGV sequence is vital to keep the metabolic stability of α-synuclein [22]. The i-proteasomes mainly guide the MHC molecules to mediate inflammation/immunity reaction, however we did not completely understand the detail about i-proteasomes participating in PD pathogenesis. It is known that after the α-synuclein degraded into 5–15 amino acid residues by ubiquitin proteasome system (UPS), the short antigenic peptides will be transported into endoplasmic reticulum system by the transporter associated with antigen processing (TAP) which brought the antigenic peptides to cell surfaces and the complex will be recognized by the MHC-I and MHC-II receptors. Then MHC presentation activates CD4+T cells and CD8+cytotoxic T lymphocytes (CTL) to take part in the immune regulation in CNS, respectively [23]. Human MHC molecules are also called human leukocyte antigen (HLA). DA neurons express multiple HLA molecules and present the digested peptides on the surface recognized by CD8+ CTL cells [23]. At the same time, the activated microglias secrete TNF-α, IL-1β and IL-6 inflammatory factors to attack DA neurons, or present antigens to CD4+ T cells by MHC-II pathway [24, 25]. Abundant MHC-II positive microglia and CD4+,CD8+ T cells were found in the SN of PD patients, and the microglia activity was confirmed to be related to the degeneration severity of DA neurons as well as PD progression [17]. Our data found that natural killer (NK) cells increased and CD3+T, CD3+CD4+T cells, Th1 cells decreased in the peripheral blood from PD patients [26]. These immune reactions will result in the increased secretion of TNF-α, IL-2, IL-6, IL-10, IL-1β and IFN-γ in cerebrospinal fluid (CSF) or blood, and accompanied with more than ten kinds of high sensitive and specific antibodies increased, such as elongation factor 1-alpha1 and poly (A) binding protein cytoplasmic-3 [25, 27]. The surface antigens in monocytes were up-regulated and correlated with the Unified Parkinson's Disease Rating Scale (UPDRS) III scores of PD patients, indicating the importance of central and peripheral immunity reaction occurred in PD [28]. Recent studies showed that HLA-I in DA neurons and HLA-II in activated microglia both participate the antigen presentation and α-synuclein degradation [24, 29]. When proteasome function is suppressed by intracellular α-synuclein accumulation, i-proteasome will be activated as compensatory to promote α-synuclein degradation [30]. The degraded peptides will be presented to CD8+T cells by MHC-I molecules, and the CD8+T cells release pre-inflammatory factors and activate Wnt/Nurr1 and Nuclear factor-kB signal pathways to ignite the secondary inflammation reaction [30, 31]. Meanwhile, extracellular α-synuclein integrates with HLA-DR molecules in microglia, mediate the activation of CD4+T cells by MCH-II antigen presentation pathway and T lymphocytes to participate the immune reaction in CNS [15, 25]. If the proteasomes function are impaired in DA neurons, α-synuclein accumulation could activate the i-proteasomeas as compensatory to provide much more MHC-I and MHC-II molecules to CD8+T and CD4+ T cells showing stronger immune response [29, 32]. Under the condition of HLA gene mutation, the genetic risks may profoundly affect the α-synuclein degradation, result in the immune dysregulation in CNS and finally accelerate the DA neurons apoptosis. Immunological genes involved in PD It is well known that genetic factors has high risk for PD. Much data revealed that the variation of HLA genes severely accelerate PD development [33]. Total 224 genes in HLA region are located at I, II and III region of HLA genomics, respectively [33]. Excepted the rarely expressed genes at HLA-I region, 128 genes from II and III of HLA region were found the contribution to immune regulation in CNS [33]. For example, HLA-DR at HLA II region is composed of DRA and DRB genes which encoding α and β chain respectively [33]. The α and β chains are assembled into HLA-DR molecules contributed to the immunologic response in neuron degeneration [33]. It's worth noting that HLA genes have high polymorphism among different ethnic and geographical population [34]. For example, HLA-C*0304, HLA-DRB1*0404, HLA-DRA and HLA-DQB1 were closely related to PD in European population, but no reports in Japanese [33, 34]. Our investigation firstly showed that HLA-DRB1*0301 is a suspect locus to PD in Chinese Han population (OR = 2.048,95%CI:1.455,2.884), but no meaningful polymorphism found in Chinese Uyghur population [34]. It is known that proteasome subunit beta (PSMB) genes locate at HLA-II region and adjacent to HLA-DR genes [33]. PSMB genes encodes β chain to help the α-synuclein degradation after the proteasome constitution [35]. It was speculated that PSMB may cooperate with HLA-DR to participate the immune regulation and intrinsic antigen presentation [36]. Our investigation has confirmed that female carrying rs17587 G/G at PSMB9 will significantly increase PD risk (OR = 1.851 95%CI: 1.240). The rs17587 locates at the code region of exon 60 at PSMB9 [33]. Its variation of nucleotide G to A may lead to the amino acid of arginine replaced with histidine [37]. The G/G genotype of rs17587 at PSMB9 disrupts the peptide cleavage site of proteasome resulting in the impaired activity of proteasome [37]. Other important genes are reported to involve in immunological regulation of PD, included HLA, GBA (Glucosidase Beta Acid), SNCA (Synuclein alpha), LRRK2 (Leucine-rich repeat kinase 2) and NURR1 (Nuclear receptor related 1 protein) [33, 38]. The GBA protein is liable to regulate the abnormal activation of microglia in basal ganglia [39], and the LRRK2 protein may affect the inflammatory reaction medicated by Toll-like receptor 4 [40]. Nurr1, a new anti-inflammatory factor, regulate the nuclear factor-kB-p65 signaling pathway in microglia and astrocytes through Nurr1/CoREST (RE1-silencing transcription factor co-repressor 1) complex [41]. Generally, variations in these PD-related genes may promote the antigen presentation and the activation of T and B lymphocytes, strengthen the sensitivity of DA neurons to environmental risk, and down-regulate the catecholamine metabolism conducted by MHC pathway [34]. In other words, it was suggested these genetic factors, serum immune/inflammatory molecules and blood lymphocytes may be the potential clinic biomarkers for PD prediction. Dopamine receptors in blood lymphocytes Dopamine receptors were found in the blood white cells of human kind [42]. These receptors are moderately expressed in neutrophil granulocyte or eosinophil granulocyte, but highly expressed in B lymphocytes and NK cells [42]. There are five kinds of dopamine receptors in CD3+CD4+T lymphocytes with different expression level [42]. For example, the D1 receptor expression is significantly higher than D2, and D5 receptor expression is higher than D1 in T lymphocytes and memory T lymphocytes [42]. The D2 receptor showed no significant different expression in blood cells from PD patient [42, 43]. But the D3 receptor expression declines dramatically in peripheral lymphocytes and is associated with PD severity [42]. The proportion of CD95/CD3 increases significantly in the lymphocytes from PD patients, but decreases significantly after L-dopa treatment [44]. The activation of dopamine receptors by different activators can reduce the release of TNF-α, IL-6 and other cytokines from microglia which may help to explain their neuro-protection [45, 46]. MicroRNA signatures in PD Recent studies demonstrated that some PD-related genes is modulated by 18 ~ 25 nt microRNAs (miRNAs) [47] at post-transcriptional level. Matching with target mRNA 3′ transcription terminals, miRNAs guide Dicer enzyme to degrade the mRNA targets or restrain their translation [48]. There are multiple mechanisms of miRNAs regulation on PD pathogenesis. For example, miR-7 and miR-153 modulate the mRNA level of α-synuclein [49], LRRK2 contributes to the regulation of E2F1/DP expression which is affected by miR-7 and miR-184 [50]. Recent studies mainly focused on the microglia activation and macrophage infiltration in brain. It becomes a hot spot about the mechanism of miRNAs modulating the related inflammatory disorders in CNS. Based on function classification, some microRNAs are classified as pro-inflammatory molecules, such as miR-155, miR-125 and miR-101, and some are anti-inflammatory molecules, such as miR-146, miR-21 and miR-124 [51]. It was reported that miR-146 modulates the microglia activation in CNS regulated by NF-KB and JNK-STAT (Janus kinase/signal transducers and activators of transcription) pathways, miR-124 retrains macrophage and microglia activation conducted by CCAAT-enhancer-binding proteins-α–PU.1 complex [52, 53]. All the miRNAs are involved in the inflammation/immunity system on PD pathogenesis [54]. Furthermore, the decreased level of miR-141, miR-214、miR-146b-5p and miR-193a-3p was found closely associated with the early stage of PD [55]. In our study, there were 644 abnormal miRNAs detected in A53T-PD transgenic mice using high-throughput screening. Among them, 72 miRNAs were confirmed to involve in the immune response in SN. The level of 42 miRNAs has changed significantly in midbrain, such as miR-146b-5 decreases to 0.45 fold, whereas miR-124-5p up to 0.44 fold. Several miRNAs were confirmed to be stable in blood, urine and CSF, so called circulating miRNAs [56]. We also screened out 10 circulating miRNAs in CSF samples from 40 Chinese PD patients for disease’s inflammatory evaluation [57]. Among them, miR-200a-3p and miR-542-5p were supposed to be closely related to α-synuclein aggregation, whereas miR-342-5p was proved to participate in modulating inflammatory reaction [57]. These circulating miRNAs are ideal biomarkers for the evaluation of CNS diseases. Conclusions In summary, the interaction of aging, environment risk and genetic factors results in the accumulation of α-synuclein in midbrain which ignites the special inflammatory /immune response medicated by i-proteasome. In the inflammatory/immune procession, MHC signal pathway accompanied with microglia activation is vital to DA neurons apoptosis. Taking together, α-synuclein accumulation and abnormal MHC antigenic presentation cause inflammation/immune response in CNS and provide specific biomarkers for the prediction of DA neurons degeneration and assessment of PD risk as well as PD development (Table 1).Table 1 The biomarkers of immune dysregulation and inflammation response in PD Immunological genes PSMB9-rs17587, HLA-DRA, HLA-DRB1*0404, HLA-DRB1*0301, HLA-C*0304, HLA-DQB1, etc. PD related genes PARK1, PARK2, PINK1, DJ-1, ATP13a2, GBA, RAB7L1-NUCKS1, STK39, BST1, FAM47E-SCARB2, SNCA, HLA, GPNMB, FGF20, LRRK2, GCH1, MAPT, SREBF1, DDRGK1, NURR1, etc. Proteins in CSF a-synuclein, Aβ42, Tau, Nurr1, etc. Inflammation factors TNF-α, IL-2, IL-6, IL-10, IL-1β, IFN-γ, etc. Antibodies ICAM4, PTCD2, FRMD8, CTLA-4/Fc, MYOT, HSH2D, FN1, TRIM21, Elongation factor 1-alpha 1, PABPC3, etc. Blood lymphocytes CD3+, CD4+, CD8+, CD31+ α4β7+ CD4+, CD95/CD3, NK cells, etc. PD related miRNAs miR-7, miR-153, miR-155, miR-125, miR-101, miR-146, miR-21, miR-124, miR-141, miR-214, miR-146b-5p, miR-193a-3p, miR-200a-3p, miR-542-5p, miR-342-5p, etc. Abbreviations ADAlzheimer's disease APPAmyloid precursor protein CNSCentral neural system CoRESTRE1-silencing transcription factor co-repressor 1 CSFCerebrospinal fluid CTLCytotoxic T lymphocytes DADopaminergic GBAGlucosidase beta acid HLAHuman leukocyte antigen JNK-STATJanus kinase/signal transducers and activators of transcription LMP2Large multifunctional protease 2 LMP7Large multifunctional protease 7 LRRK2Leucine-rich repeat kinase 2 MDSMovement Disorder Association MECL-1Multicatalytic endopeptidase complex-like-1 MHCMajor histocompatibility complex molecules miRNAsmicroRNAs NKNatural killer NURR1Nuclear receptor related 1 protein PDParkinson’s disease PSMBProteasome subunit beta SNSubstantia nigra SNCASynuclein alpha TAPAntigen processing UPDRSUnified Parkinson’s Disease Rating Scale UPSUbiquitin proteasome system Acknowledgements Not applicable Funding This review was supported by research grants from the State Key Development Program for Basic Research of China (2011CB510000), the National Natural Science Foundation of China (81271428, 81471292, U1503222 and 81430021) the keypoint Science Foundation of Guangdong of China (2015A030311021), a grant supported by technology project of Guangzhou (201604020152) and a grant supported by assisting research project of science and technology for Xinjiang (201591160). All founding were used for the design, collection, analysis and interpretation of data and in writing in the manuscript. Availability of data and materials Not relevant. Authors’ contributions All authors read and approved the final manuscript. LC and GL summarized the background. MM summarized the miRNA and i-proteasomes system part. LC, LW, XC and YX summarized the inflammatory cells part. LC, MM and PX conceived, designed, and performed the paper. SL, XY and SQ revised the paper. Competing interests The authors declare they have no competing interest. Consent for publication Not relevant. Ethics approval and consent to participate Not relevant. ==== Refs References 1. Pringsheim T Jette N Frolkis A Steeves TD The prevalence of Parkinson’s disease: A systematic review and meta-analysis Mov Disord 2014 29 1583 1590 10.1002/mds.25945 24976103 2. Wakabayashi K Tanji K Mori F [Pathology of basal ganglia in neurodegenerative diseases]. Brain and nerve= Shinkei Kenkyu No Shinpo 2009 61 429 439 19378813 3. Postuma RB MDS clinical diagnostic criteria for Parkinson’s disease Mov Disord 2015 30 1591 1601 10.1002/mds.26424 26474316 4. Peelaerts W Synuclein strains cause distinct synucleinopathies after local and systemic administration Nature. 2015 522 340 344 10.1038/nature14547 26061766 5. Tyson T, Steiner JA, Brundin P. Sorting Out Release, Uptake and Processing of Alpha-Synuclein During Prion-Like Spread of Pathology. J Neurochem. 2015;[Epub ahead of print]. 6. Liu WM Time trends in the prevalence and incidence of Parkinson’s disease in Taiwan: A nationwide, population-based study J Fr Med Assoc. 2015 115 531 8 10.1016/j.jfma.2015.05.014 7. Tsuboi Y Environmental-genetic interactions in the pathogenesis of Parkinson’s disease Exp Neurobiol 2012 21 123 128 10.5607/en.2012.21.3.123 23055790 8. Trinh J Guella I Farrer MJ Disease penetrance of late-onset parkinsonism: A meta-analysis JAMA Neurol 2014 71 1535 1539 10.1001/jamaneurol.2014.1909 25330418 9. Mosley RL Hutter-Saunders JA Stone DK Gendelman HE Inflammation and adaptive immunity in Parkinson’s disease Cold Spring Harb Perspect Med 2012 2 a009381 10.1101/cshperspect.a009381 22315722 10. Louveau A Structural and functional features of central nervous system lymphatic vessels Nature. 2015 533 278 10.1038/nature16999 26909581 11. Zhang P Lokuta KM Turner DE Liu B Synergistic dopaminergic neurotoxicity of manganese and lipopolysaccharide: differential involvement of microglia and astroglia J Neurochem 2010 112 434 443 10.1111/j.1471-4159.2009.06477.x 19895668 12. Wang G Microglia/macrophage polarization dynamics in white matter after traumatic brain injury J Cereb Blood Flow Metab 2013 33 1864 1874 10.1038/jcbfm.2013.146 23942366 13. Odoardi F T cells become licensed in the lung to enter the central nervous system Nature 2012 488 675 679 10.1038/nature11337 22914092 14. Wheeler CJ T-Lymphocyte Deficiency Exacerbates Behavioral Deficits in the 6-OHDA Unilateral Lesion Rat Model for Parkinson’s Disease J Neurol Neurophysiol. 2014 5 209 10.4172/2155-9562.1000209 25346865 15. Stuendl A, et al. Induction of α-synuclein aggregate formation by CSF exosomes from patients with Parkinson’s disease and dementia with Lewy bodies. Brain. 2016;139:481–94. 16. Cabezas, R. et al. Astrocytic modulation of blood brain barrier: perspectives on Parkinson’s disease. Front Cell Neurosci. (2014); 8. 17. Sanchez-Guajardo V Barnum CJ Tansey MG Romero-Ramos M Neuroimmunological processes in Parkinson's disease and their relation to α-synuclein: microglia as the referee between neuronal processes and peripheral immunity ASN Neuro 2013 5 AN20120066 10.1042/AN20120066 18. Krüger E Kloetzel P-M Immunoproteasomes at the interface of innate and adaptive immune responses: two faces of one enzyme Curr Opin Immunol 2012 24 77 83 10.1016/j.coi.2012.01.005 22296715 19. Ferrington DA Gregerson DS Immunoproteasomes: structure, function, and antigen presentation Prog Mol Biol Transl Sci 2012 109 75 10.1016/B978-0-12-397863-9.00003-1 22727420 20. Orre M Reactive glia show increased immunoproteasome activity in Alzheimer’s disease Brain 2013 136 1415 1431 10.1093/brain/awt083 23604491 21. Mishto M The immunoproteasome β5i subunit is a key contributor to ictogenesis in a rat model of chronic epilepsy Brain Behav Immun 2015 49 188 196 10.1016/j.bbi.2015.05.007 26044087 22. Dettmer U Newman AJ von Saucken VE Bartels T Selkoe D KTKEGV repeat motifs are key mediators of normal α-synuclein tetramerization: Their mutation causes excess monomers and neurotoxicity Proc Natl Acad Sci 2015 112 9596 9601 10.1073/pnas.1505953112 26153422 23. Neefjes J Jongsma ML Paul P Bakke O Towards a systems understanding of MHC class I and MHC class II antigen presentation Nat Rev Immunol 2011 11 823 836 22076556 24. Cebrián C MHC-I expression renders catecholaminergic neurons susceptible to T-cell-mediated degeneration Nat Commun. 2014 16 3633 24736453 25. More SV, Kumar H, Kim IS, Song S-Y, Choi D-K. Cellular and molecular mediators of neuroinflammation in the pathogenesis of Parkinson’s disease. Mediators of inflamm. 2013:952375. 26. Zhang S Clinical analysis of subpopulation of peripheral T and B lymphocytes in Chinese Parkinson’s disease patients Zhonghua Yi Xue Za Zhi 2014 94 47 3726 3730 25623094 27. Santiago JA Potashkin JA Current Challenges Towards the Development of a Blood Test for Parkinson’s Disease Diagnostics 2014 4 153 164 10.3390/diagnostics4040153 26852683 28. Saunders JAH CD4+ regulatory and effector/memory T cell subsets profile motor dysfunction in Parkinson’s disease J Neuroimmune Pharmacol 2012 7 927 938 10.1007/s11481-012-9402-z 23054369 29. Starossom SC Galectin-1 deactivates classically activated microglia and protects from inflammation-induced neurodegeneration Immunity 2012 37 249 263 10.1016/j.immuni.2012.05.023 22884314 30. Zhang L-M Dopamine Agonists Exert Nurr1-inducing Effect in Peripheral Blood Mononuclear Cells of Patients with Parkinson’s Disease Chin Med J 2015 128 1755 10.4103/0366-6999.159349 26112716 31. Wei L Activation of Wnt/β-catenin pathway by exogenous Wnt1 protects SH-SY5Y cells against 6-hydroxydopamine toxicity J Mol Neurosci 2013 49 105 115 10.1007/s12031-012-9900-8 23065334 32. Heneka MT Kummer MP Latz E Innate immune activation in neurodegenerative disease Nat Rev Immunol 2014 14 463 477 10.1038/nri3705 24962261 33. Mo M Association Analysis of Proteasome Subunits and Transporter Associated with Antigen Processing on Chinese Patients with Parkinson's Disease Chin Med J 2016 129 1053 10.4103/0366-6999.180513 27098790 34. Sun C, et al. HLA-DRB1 alleles are associated with the susceptibility to sporadic Parkinson’s disease in Chinese Han population. PLoS One. 2012;7(11):e48594. 35. Allcock RJN. The major histocompatibility complex: a paradigm for studies of the human genome. Immunogenetics: Methods and Applications in Clinical Practice. 2012:1–7. 36. Rhodes DA. Human Leukocyte Antigen (HLA) System and Human Disorders. eLS. 2005. 37. Camarena Á PSMB8 (LMP7) but not PSMB9 (LMP2) gene polymorphisms are associated to pigeon breeder’s hypersensitivity pneumonitis Respir Med 2010 104 889 894 10.1016/j.rmed.2010.01.014 20153157 38. Nalls MA Large-scale meta-analysis of genome-wide association data identifies six new risk loci for Parkinson’s disease Nat genet. 2014 46 989 93 10.1038/ng.3043 25064009 39. Barkhuizen M Anderson DG Grobler AF Advances in GBA-associated Parkinson’s disease–pathology, presentation and therapies Neurochem Int. 2015 93 6 25 10.1016/j.neuint.2015.12.004 26743617 40. Hsu CH MKK6 binds and regulates expression of Parkinson’s disease‐related protein LRRK2 J Neurochem 2010 112 1593 1604 10.1111/j.1471-4159.2010.06568.x 20067578 41. Smith GA A Nurr1 Agonist Causes Neuroprotection in a Parkinson’s Disease Lesion Model Primed with the Toll-Like Receptor 3 dsRNA Inflammatory Stimulant Poly (I: C) PLoS One 2015 10 e0121072 10.1371/journal.pone.0121072 25815475 42. González H Dopamine receptor D3 expressed on CD4+ T cells favors neurodegeneration of dopaminergic neurons during Parkinson’s disease J Immunol 2013 190 5048 5056 10.4049/jimmunol.1203121 23589621 43. Stevens CH Reduced T helper and B lymphocytes in Parkinson’s disease J Neuroimmunol 2012 252 95 99 10.1016/j.jneuroim.2012.07.015 22910543 44. Hurny A Michałowska-Wender G Wender M Impact of L-DOPA treatment of patients with Parkinson’s disease on mononuclear subsets and phagocytosis in the peripheral blood Folia Neuropathol 2013 51 127 131 10.5114/fn.2013.35955 23821384 45. Kumar A Sharma N Gupta A Kalonia H Mishra J Neuroprotective potential of atorvastatin and simvastatin (HMG-CoA reductase inhibitors) against 6-hydroxydopamine (6-OHDA) induced Parkinson-like symptoms Brain Res 2012 1471 13 22 10.1016/j.brainres.2012.06.050 22789904 46. Sanchez-Guajardo V Annibali A Jensen PH Romero-Ramos M α-Synuclein Vaccination Prevents the Accumulation of Parkinson Disease–Like Pathologic Inclusions in Striatum in Association With Regulatory T Cell Recruitment in a Rat Model J Neuropathol Exp Neurol 2013 72 624 645 10.1097/NEN.0b013e31829768d2 23771222 47. Mouradian MM MicroRNAs in Parkinson's disease Neurobiol Dis 2012 46 279 284 10.1016/j.nbd.2011.12.046 22245218 48. Abe M Bonini NM MicroRNAs and neurodegeneration: role and impact Trends Cell Biol 2013 23 30 36 10.1016/j.tcb.2012.08.013 23026030 49. Alvarez-Erviti L Influence of microRNA deregulation on chaperone-mediated autophagy and α-synuclein pathology in Parkinson’s disease Cell Death Dis 2013 4 e545 10.1038/cddis.2013.73 23492776 50. Chan SL Angeles DC Tan E-K Targeting leucine-rich repeat kinase 2 in Parkinson’s disease Expert Opin Ther Targets 2013 17 1471 1482 10.1517/14728222.2013.842978 24206363 51. Ponomarev ED Veremeyko T Weiner HL MicroRNAs are universal regulators of differentiation, activation, and polarization of microglia and macrophages in normal and diseased CNS Glia 2013 61 91 103 10.1002/glia.22363 22653784 52. Svahn AJ Giacomotto J Graeber MB Rinkwitz S Becker TS miR-124 Contributes to the functional maturity of microglia Dev Neurobiol 2015 76 507 18 10.1002/dneu.22328 26184457 53. Saba R MicroRNA 146a (miR-146a) is over-expressed during prion disease and modulates the innate immune response and the microglial activation state PLoS One 2012 7 e30832 10.1371/journal.pone.0030832 22363497 54. Wang H MiR‐124 Regulates Apoptosis and Autophagy Process in MPTP Model of Parkinson’s Disease by Targeting to Bim Brain Pathol. 2015 26 167 76 10.1111/bpa.12267 25976060 55. Dong H A panel of four decreased serum microRNAs as a novel biomarker for early Parkinson’s disease Biomarkers. 2016 21 129 37 10.3109/1354750X.2015.1118544 26631297 56. Jeck WR Sharpless NE Detecting and characterizing circular RNAs Nat Biotechnol 2014 32 453 461 10.1038/nbt.2890 24811520 57. Wei Y The microRNA-342-5p Fosters Inflammatory Macrophage Activation Through an Akt1-and microRNA-155–Dependent Pathway During Atherosclerosis Circulation 2013 127 1609 1619 10.1161/CIRCULATIONAHA.112.000736 23513069

PMC005xxxxxx/PMC5002149.txt

==== Front Respir ResRespir. ResRespiratory Research1465-99211465-993XBioMed Central London 42110.1186/s12931-016-0421-9ResearchThe impact of comorbidities on productivity loss in asthma patients Ehteshami-Afshar Solmaz s.ehteshamiafshar@alumni.ubc.ca 1FitzGerald J. Mark 604.875.5178Mark.Fitzgerald@vch.ca 2348Carlsten Christopher carlsten@mail.ubc.ca 235Tavakoli Hamid hamid.tavakoli@ubc.ca 4Rousseau Roxanne Roxanne.Rousseau@vch.ca 2Tan Wan Cheng wan.tan@hli.ubc.ca 6Rolf J. Douglass Douglass.Rolf@interiorhealth.ca 7Sadatsafavi Mohsen msafavi@mail.ubc.ca 2341 Experimental Medicine Program, Department of Medicine, Faculty of Medicine, The University of British Columbia, Vancouver, Canada 2 Department of Medicine, Division of Respiratory Medicine, The University of British Columbia, Vancouver, Canada 3 Institute for HEART + LUNG Health, Department of Medicine (Respiratory Division), The University of British Columbia, Vancouver, Canada 4 Centre for Clinical Epidemiology and Evaluation, The University of British Columbia, Vancouver, Canada 5 Department of Medicine, Centre for Occupational and Environmental Lung Disease, Vancouver, BC Canada 6 Centre for Heart Lung Innovation, The University of British Columbia, Vancouver, Canada 7 Kelowna Allergy & Respirology Research, Kelowna, Canada 8 Institute of Heart and Lung Health, The Lung Centre, 2775 Laurel Street, Vancouver, BC V5Z 1 M9 Canada 26 8 2016 26 8 2016 2016 17 1 10610 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 Health-related productivity loss is an important, yet overlooked, component of the economic burden of disease in asthma patients of a working age. We aimed at evaluating the effect of comorbidities on productivity loss among adult asthma patients. Methods In a random sample of employed adults with asthma, we measured comorbidities using a validated self-administered comorbidity questionnaire (SCQ), as well as productivity loss, including absenteeism and presenteeism, using validated instruments. Productivity loss was measured in 2010 Canadian dollars ($). We used a two-part regression model to estimate the adjusted difference of productivity loss across levels of comorbidity, controlling for potential confounding variables. Results 284 adults with the mean age of 47.8 (SD 11.8) were included (68 % women). The mean SCQ score was 2.47 (SD 2.97, range 0–15) and the average productivity loss was $317.5 per week (SD $858.8). One-unit increase in the SCQ score was associated with 14 % (95 % CI 1.02–1.28) increase in the odds of reporting productivity loss, and 9.0 % (95 % CI 1.01–1.18) increase in productivity loss among those reported any loss of productivity. A person with a SCQ score of 15 had almost $1000 per week more productivity loss than a patient with a SCQ of zero. Conclusions Our study deepens the evidence-base on the burden of asthma, by demonstrating that comorbidities substantially decrease productivity in working asthma patients. Asthma management strategies must be cognizant of the role of comorbidities to properly incorporate the effect of comorbidity and productivity loss in estimating the benefit of disease management strategies. Keywords AsthmaComorbiditiesProductivity lossPresenteeismAbsenteeismCollaborative Innovative Research Fund (CIRF)issue-copyright-statement© The Author(s) 2016 ==== Body Background With increasing life expectancy there has been an increase in the prevalence of many chronic diseases and the co-existence of multiple diseases [1–3]. Clinically, comorbidities are relevant given their potential effect on the index disease in terms of diagnosis, prognosis, and management [4]. Also, comorbid conditions increase the need for medication, risk of adverse effects and drug interactions, and reduce adherence to treatments, quality of life and functional status [1, 5, 6]. Patients with multiple comorbidities tend to use more medical services and impose a greater burden on the health-care system [2, 6]. Asthma is associated with several comorbidities; however, the prevalence varies across studies [5–8]. In a study from the United States, 26 and 10 % of asthma patients had at least one or ≥3 comorbidities, respectively [7]. In a study from Germany, 26 % of asthma patients had at least one other comorbidity while 17 % had 2 or more [9]. In a Canadian study, almost 60 % of asthma patients had at least one comorbidity [6], while in another study 12.5 % of adult asthma patients reported having three or more comorbidities, increasing to 20 % for adults 55 years and older [5, 6]. It has been well demonstrated that comorbidities are associated with poor outcomes in asthma patients [10]. Asthma patients with comorbidities experience more asthma exacerbations [6, 11–13] and there is a significant relationship between asthma control and the presence of comorbidities [4, 14, 15]. The reason behind this fact is unclear. It could be because the patient places a higher priority on other health conditions, which influences the adherence to asthma treatment. Also the nature of comorbidities like depression may cause the patient to pay less attention to their general health status and care less [7]. In a Canadian province, British Columbia (BC), 25 % of asthma patients have depression [5]. Also, comobidities could directly and causally affect the severity of asthma or its responsiveness to treatment; examples include rhinitis and gastroesophageal reflux disease [16]. It has been demonstrated that indirect costs of asthma accounted for the greater proportion of costs of asthma than direct costs, however most of the studies unnoticed this amount [16]. Also despite the documented burden of comorbidities in asthma, their effect on productivity loss has been overlooked in the past. One reason behind this status is that asthma patients are a relatively young population and are assumed to be free of comorbidities [15]. The general increase in longevity and the increase in the retirement age will inevitably result in more and more working asthma subjects. The aim of the present study was to evaluate the effect of comorbidities on productivity loss in a population-based sample of adult asthma patients. Methods Study design and participants This study is based on data from the Economic Burden of Asthma (EBA), a 1-year prospective cohort study with the specific aim of estimating the economic and humanistic burden of asthma (University of British Columbia Human Ethics Board H10-01542). In the EBA study, 618 patients with self-reported physician diagnosis asthma who were aged 1–85 were recruited by random digit dialing and followed up for a year. The study’s catchment areas were Vancouver and Central Okanagan census areas, the latter being in the interior of the BC Province with a large fraction of the population residing in rural areas. The details of this study have been described previously [17, 18] and the inclusion criteria of the present study are the same as the main EBA, having at least one encounter with healthcare system because of asthma in the past 5 years and having no plan to move out of the region in the next year, except it was restricted to adult (≥19 years old) patients who were employed at the baseline visit. Variables Comorbidity A comorbidity score was calculated based on the Self-administered Comorbidity Questionnaire (SCQ) administered in the last visit [19]. The recall period of the questionnaire is 12 months and thus we assumed comorbidity score was constant across the study period [19]. The SCQ score not only considers the number of the comorbidities but also their severity. Each included comorbid condition can get a maximum of three points based on the presence of disease, whether receiving treatment, and any functional limitation due to the condition. This questionnaire has been validated and has a moderately strong correlation with the widely popular Charlson comorbidity index [19]. The Charlson index is mainly designed for hospitalized patients and its evaluation needs access to medical records [19]. On the other hand, the SCQ is designed and validated for the outpatient settings by relying on patient self report as the principle source of information [19]. The original SCQ includes 13 common comorbidities, but in this study the questions related to pulmonary disorders were excluded (given that all patients had asthma), leaving the questionnaire with a maximum of 36 scores, three points for each of the 12 questions. The included comorbidities were heart disease, hypertension, diabetes mellitus, ulcer or stomach disease, kidney disease, liver disease, anemia or other blood disease, cancer, depression, osteoarthritis or degenerative arthritis, back pain, and rheumatoid arthritis. Productivity loss Productivity loss was measured at baseline by two validated questionnaires: the Work Productivity and Activity Impairment (WPAI) [20], and the Valuation of Lost Productivity (VOLP) [21]. The WPAI records patients’ absenteeism (missing work due to health conditions) and presenteeism (attending work but not being fully functional) in the last 7 days by asking about the hours they missed from work because of sick days or the times they went in late or left early due to health status and times they were not functional with limited accomplishment and unable to concentrate on their tasks due to the health status respectively [20]. The VOLP questionnaire collects information about the work environment such as time sensitivity of the job, team work, and availability of replacement, to calculate a coefficient that measures the contribution of individual to the work place (a coefficient of X indicates that each hour of a person’s absence is equivalent of X hours of work loss) [21, 22]. The monetary value of productivity loss per week was the product of three terms: amount of work time (hours) lost, the VOLP coefficient, and the hourly wage of the individual. Job titles were matched to the National Occupation Classification (NOC) codes [23] to estimate the hourly wage based on sex and age for each NOC from Statistics Canada for year 2010 [17]. The reported weekly costs are therefore in 2010 Canadian dollars (CAD). Confounders Socio-demographic data collected at the baseline visit were included in the statistical models as potential confounders (factors that can affect both comorbid level and productivity but are not on the causal pathway). They included: sex, age, household income levels (low v. high at cut-off of CAD$60,000 per year), education (low v. high at cut-off of 4-year college/university degree), type of residence (urban v. rural), place of birth (Canada v. abroad), drug insurance (having no insurance, being partially insured, or being fully insured), and the proportion of days covered (PDC) by any asthma controller medication in past 12 months as an indicator of adherence (cut-off values <50 %, 50–80 %, ≥80 %). The main analysis did not adjust for asthma control as it cannot be a confounder; rather, it is potentially being on the causal pathway (that is, part of the impact of comorbidity on productivity might be due to the comorbid conditions’ affecting the likelihood of achieving asthma control. It is also very unlikely for the current asthma control status to have an effect on comorbidities (thus being confounding factor). But a sensitivity analysis was performed to assess the effect of adjusting for control status on the outcomes. We defined control status based on Global Initiative for Asthma (GINA) 2012 definition, which included the presence of symptoms and impairment in lung function (all measured at baseline visit). Statistical analysis All analyses were performed using Stata (version 14; StataCorp, College Station, TX, USA). Two-tailed p-values at 0.05 were considered statistically significant. Descriptive analysis was performed on the baseline variables. We reported the hours and costs of both components of productivity loss (absenteeism and presenteeism), as well as total productivity loss across different levels of SCQ score. As the productivity loss data were zero-inflated, we used two-part models for statistical inference [24]. The first part was a logistic component and the second part was a generalized linear model with logarithmic link function and gamma distribution. The first part generates odds ratio (OR) associating covariates with any loss of productivity, and the second component produces relative rate (RR) associating covariates with the magnitude of productivity loss among those with any loss of productivity. For both components the dependent variable was the monetary value of productivity loss and the independent variables were the SCQ score and other covariates as previously mentioned. As there were missing values among some of the covariates, multiple imputations were first performed, creating 5 imputed datasets without missing variables; results of separate analyses on the imputed datasets were combined. To estimate the marginal effect of SCQ on productivity loss (that is, the weekly loss of productivity associated with any level of SCQ score), the OR and RR from the two components were combined, and p-values and confidence intervals were estimated using bootstrapping (500 times) as described elsewhere [25]. The procedure was conducted separately with absenteeism, presenteeism, and total productivity loss as the dependent variable. Results Figure 1 shows the flowchart of sample selection. The final sample consisted of 284 individuals whose baseline characteristics are shown in Table 1. The sample was 68 % female with a mean age of 47.8 ± 11.8 with generally high levels of education and household income. Most of the subjects (63 %) had at least one comorbid condition and the overall SCQ score was 2.47 ± 2.97, with a minimum of 0 and a maximum of 15. Only 48 % of patients reported any productivity loss, with 36 % of them reporting absenteeism and 64 % reporting presenteeism. Mean weekly hours and costs of productivity loss were 16 ± 17.6 h and $317.49 ± $858.83 respectively.Fig. 1 Flow chart of study population Table 1 Characteristics of study sample Study population = 284 Age, mean ± SDa 47.8 ± 11.8 Sex (%)  Women 193 (68)  Men 91 (32) Household income (%)  High (>60,000 CAD) 178 (62.7)  Low 96 (33.8)  Missing 10 (3.5) Educational level (%)  High 229 (80.6)  Low 55 (19.4) Place of birth (%)  Canada 207 (72.9)  Outside Canada 77 (27.1) Ethnicity (%)  Caucasian 231 (81.3)  Asian 18 (6.3)  Other 35 (12.4) Residence type (%)  Urban 260 (91.5)  Rural 24 (8.5) Asthma medication adherence (%)  PDCb < 50 % 171 (60.2)  50 % ≤ PDC < 80 % 31 (11)  PDC ≥ 80 % 81 (28.5)  Missing 1 (0.3) Asthma control level (%)  Controlled 55 (19.4)  Partially Controlled 113 (39.8)  Uncontrolled 115 (40.5)  Missing 1 (0.3) Productivity Loss (%) 136 (48) Absenteeism (%) 49 (17) Presenteeism (%) 127 (45) Hours of overall productivity loss, mean ± SD 16 ± 17.6 Costsc of overall productivity loss, mean ± SD 317.49$ ± 858.83$ Overall SCQd comorbidities score, mean ± SD 2.47 ± 2.97 Heart disease (%) 15 (5.3) Hypertension (%) 35 (12.3) Diabetes (%) 9 (3.2) Ulcer or Stomach Disease (%) 37 (13) Kidney disease (%) 3 (1.1) Liver disease (%) 2 (0.7) Anemia or other blood disease (%) 21 (7.4) Cancer (%) 6 (2.1) Depression (%) 40 (14.1) Osteoarthritis, degenerative arthritis (%) 62 (21.8) Back pain (%) 99 (34.9) Rheumatoid arthritis (%) 2 (0.7) aStandard deviation bproportions of days covered c2010 Canadian dollars dself-administered comorbidity questionnaire Unadjusted analysis Table 2 shows the results of unadjusted analysis. The hours of absenteeism increased from 1.26 to 7.14 h as the SCQ increased from 0 to 15, and for presenteeism it rose from 3.97 to 12.59 h. The costs of absenteeism increased from $50/week for SCQ of 0 to almost $300/week for SCQ of 15, while the corresponding values for presenteeism was $140/week and $734/week. The same increases were seen for the total productivity loss, from $190/per week to $1036/per week.Table 2 Unadjusted regression analysis SCQ score Hours of Absenteeism Costs of Absenteeisma Hours of Presenteeism Costs of Presenteeisma Hours of total productivity loss Costs of productivity lossa,b 0 1.26 50.35 3.97 140.20 5.22 190.18 (0.54–1.98) (19.88–80.81) (2.73–5.22) (65.27–215-13) (3.49–6.96) (97.22–283.14) 1 1.65 66.91 4.55 179.81 6.20 246.60 (0.99–2.32) (40.19–93.63) (3.48–5.62) (118.75–240.87) (4.70–7.70) (169.16–324.03) 2 2.04 83.47 5.12 219.41 7.17 303.01 (1.22–2.87) (49.80–117.14) (4.00–6.25) (145.85–292.98) (5.52–8.83) (208.30–397.72) 3 2.44 100.03 5.70 259.02 8.15 359.43 (1.33–3.54) (53.27–146.79) (4.31–7.08) (155.75–362.28) (6.04–10.26) (226.90–491.95) 4 2.83 116.59 6.27 298.62 9.13 415.84 (1.38–4.27) (54.35–178.83) (4.51–8.03) (159.03–438.21) (6.41–11.85) (237.59–594.10) 5 3.22 133.15 6.84 338.22 10.10 472.26 (1.41–5.03) (54.45–211.86) (4.65–9.04) (159.70–516.76) (6.70–13.50) (245.10–699.42) 6 3.61 149.71 7.42 377.83 11.08 528.67 (1.43–5.79) (54.06–245-36) (4.76–10.08) (159.12–596.53) (6.96–15.20) (251.09–806.25) 7 4.00 166.27 7.99 417.43 12.05 585.09 (1.44–6.56) (53.42–279.13) (4.85–11.13) (157.90–676.97) (7.19–16.91) (256.28–913.90) 8 4.39 182.83 8.57 457.04 13.03 641.50 (1.45–7.34) (52.61–313.06) (4.94–12.20) (156.28–757.79) (7.42–18.64) (260.97–1022.03) 9 4.79 199.39 9.14 496.64 14.01 697.92 (1.45–8.12) (51.70–347.09) (5.01–13.27) (154.41–838.87) (7.63–20.38) (265.36–1130.48) 10 5.18 215.96 9.72 536.24 14.98 754.33 (1.46–8.90) (50.72–381.19) (5.09–14.34) (152.38–920.11) (7.84–22.12) (269.53–1239.14) 11 5.57 232.52 10.29 575.85 15.96 810.75 (1.46–9.68) (49.69–415.35) (5.16–15.42) (150.23–1001.47) (8.05–23.86) (273.56–1347.94) 12 5.96 249.08 10.86 615.45 16.93 867.17 (1.46–10.46) (48.62–449.54) (5.23–16.50) (147.98–1082.92) (8.26–25.61) (277.48–1456.86) 13 6.35 265.64 11.44 655.06 17.91 923.58 (1.46–11.25) (47.52–483.75) (5.29–17.58) (145.68–1164.43) (8.46–27.36) (281.31–1565.85) 14 6.74 282.20 12.01 694.66 18.88 979.99 (1.45–12.03) (46.41–517.99) (5.36–18.67) (143.32–1246) (8.66–29.11) (285.08–1674.91) 15 7.14 298.76 12.59 734.26 19.86 1036.41 (1.45–12.82) (45.27–552.25) (5.42–19.75) (140.93–1327.60) (8.86–30.86) (288.80–1784.03) All the p-values <0.05 a2010 CAD bThe sum of the costs of absenteeism and presenteeism are not exactly equal to the costs of total productivity loss, because the exact distribution of error terms around each component is inevitably different in regression models Adjusted analysis The results of two-part regression model are demonstrated in Table 3. In the logistic part of the analysis, SCQ was significantly associated with higher odds of reporting absenteeism, presenteeism and total productivity loss. However, in the second part of the regression, among patients with productivity loss, SCQ was only significantly associated with the total productivity loss (RR = 1.09, CI = 1.01-1.18, P = 0.02) and not presenteeism or absenteeism separately. The other covariates were not significantly associated with productivity loss in either parts of the model.Table 3 Results of the adjusted regression analysis of productivity loss on SCQa score Absenteeism Presenteeism Total Productivity Loss Female v. male Adjusted OR for reporting productivity loss 1.72 1.08 0.99 (0.53–2.57) (0.61–2.35) (0.56–1.76) (P = 0.7) (P = 0.77) (P = 0.99) Adjusted ratio of productivity loss among those who reported productivity loss 0.42 0.88 0.79 (0.15–1.18) (0.52–2.18) (0.46–1.37) (P = 0.1) (P = 0.63) (P = 0.41) Age (per 1 year increase) Adjusted OR for reporting productivity loss 0.98 0.98 0.98 (0.95–1.01) (0.96–1.04) (0.96–1.01) (P = 0.31) P = (0.13) (P = 0.23) Adjusted ratio of productivity loss among those who reported productivity loss 0.98 0.99 0.98 (0.94–1.03) (0.96–1.04) (0.96–1.01) (P = 0.5) (P = 0.49) (P = 0.18) High v. urban education Adjusted OR for reporting productivity loss 1.16 0.68 0.52 (0.41–3.27) (0.34–2.84) (0.25–1.1) (P = 0.73) (P = 0.27) (P = 0.09) Adjusted ratio of productivity loss among those who reported productivity loss 0.92 1.50 1.44 (0.29–2.96) (0.84–2.40) (0.76–2.72) (P = 0.9) (P = 0.17) (P = 0.26) Rural residence Adjusted OR for reporting productivity loss 0.47 0.38 0.41 (0.13–1.75) (0.11–6.17) (0.13–1.35) (P = 0.26) (P = 0.11) (P = 0.14) Adjusted ratio of productivity loss among those who reported productivity loss 0.99 0.42 0.44 (0.05–20.88) (0.21–2.88) (0.16–1.15) (P = 0.99) (P = 0.1) (P = 0.1) Foreign Born v. Canadian-born Adjusted OR for reporting productivity loss 0.95 1.00 1.21 (0.41–2.19)) (0.56–2.42) (0.67–2.17) (P = 0.92) (P = 0.99) (P = 0.53) Adjusted ratio of productivity loss among those who reported productivity loss 0.41 0.84 0.72 (0.12–1.39) (0.52–2.08) (0.43–1.19) (P = 0.16) (P = 0.5) (P = 0.2) PDCb Level (Reference: PDC < 50 %)  50–80 % Adjusted OR for reporting productivity loss 2.75 1.27 1.34 (0.98–7.73) (0.54–3.66) (0.58–3.10) (P = 0.07) (P = 0.59) (P = 0.49) Adjusted ratio of productivity loss among those who reported productivity loss 1.49 2.50 2.72 (0.33–6.64) (0.70–6.77) (0.91–8.10) (P = 0.6) (P = 0.15) (P = 0.07)  >80 % Adjusted OR for reporting productivity loss 1.11 0.86 0.97 (0.48–2.59) (0.47–2.50) (0.52–1.80) (P = 0.8) (P = 0.6) (P = 0.93) Adjusted ratio of productivity loss among those who reported productivity loss 0.95 1.13 1.13 (0.40–2.29) (0.69–2.08) (0.68–1.86) (P = 0.91) (P = 0.62) (P = 0.64) Drug Insurance (Reference: full insurance)  Partial Adjusted OR for reporting productivity loss 1.36 1.03 1.11 (0.41–4.49) (0.45–3.44) (0.47–2.63) (P = 0.59) (P = 0.94) (P = 0.8) Adjusted ratio of productivity loss among those who reported productivity loss 0.84 0.60 0.73 (0.20–3.53) (0.24–3.99) (0.30–1.80) (P = 0.82) (P = 0.17) (P = 0.5)  None Adjusted OR for reporting productivity loss 0.82 0.91 0.86 (0.20–3.43) (0.36–4.07) (0.33–2.25) (P = 0.82) (P = 0.86) (P = 0.76) Adjusted ratio of productivity loss among those who reported productivity loss 1.09 0.66 0.76 (0.16–7.33) (0.24–4.58) (0.28–2.11) (P = 0.93) (P = 0.42) (P = 0.61) High v. low income Adjusted OR for reporting productivity loss 0.84 0.87 0.88 (0.37–1.88) (0.48–2.47) (0.47–1.66) (P = 0.54) (P = 0.66) (P = 0.7) Adjusted ratio of productivity loss among those who reported productivity loss 1.47 1.27 1.27 (0.55–3.93) (0.74–2.25) (0.69–2.32) (P = 0.44) (P = 0.38) (P = 0.44) SCQ (per 1 unit increase) Adjusted OR for reporting productivity loss 1.17 1.14 1.14 (1.04–1.32) (1.03–1.17) (1.02–1.28) (P = 0.01) (P = 0.01) (P = 0.01) Adjusted ratio of productivity loss among those who reported productivity loss 1.04 1.05 1.09 (0.90–1.21) (0.98–1.11) (1.01–1.18) (P = 0.56) (P = 0.14) (P = 0.02) aSelf-administered comorbidity questionnaire bproportion of days covered by medication Marginal effect of comorbidity on productivity loss The marginal effect of each level of SCQ score on total productivity loss is demonstrated in Fig. 2. In patients without any comorbidity, the productivity loss was $205 week. Total productivity loss was $1685 higher with a SCQ score of 15 in comparison to a SCQ score of zero. The margins were significant at all the levels, except SCQ score of 15 (P-value = 0.06). For Absenteeism, the costs were from $61.87/week (SD = 23.07) for SCQ score of 0 to $612.61/week (SD = 498.06) for those with the score of 15, and for presenteeism they were from $160.92/week (SD = 32.57) to $877.33/week (SD = 473) for the SCQ scores of 0 and 15, respectively. However the incremental costs for the SCQ score of ≥10 for absenteeism and SCQ score of 15 for presenteeism were not significant.Fig. 2 Incremental Costs of productivity loss based on comorbidity scores Sensitivity analysis Sensitivity analysis revealed that OR for reporting productivity loss and adjusted ratio for productivity loss among those reporting it did not change by adding control status in the model. However, the adjusted RR in the second part was no longer significant. Adding control status to the model did not have a significant impact on the estimates of the marginal loss of productivity (Appendix). Discussion In this study, we have demonstrated that as the SCQ score, a validated quantitative measure of the burden of comorbidity, increased, the hours of absenteeism and presenteeism increased significantly to almost 20 h per week. This caused almost $1685/week higher productivity loss in patients with a score of 15, the maximum score observed in our sample, in comparison to those with a zero score. The average SCQ score in the sample was 2.47 (SD 2.97). At this level, productivity loss was almost 1.5 times higher than in individuals without any comorbidity (SCQ = 0). In the full two-part regression, SCQ increased the odds of reporting productivity loss, absenteeism and or presenteeism by 14–17 %. In addition, among those with productivity loss, one-unit increase in SCQ increased productivity loss by 9 %. Overall, our results demonstrate the substantial effect of comorbidity on productivity loss in patients with asthma. Previous studies assessing the impact of comorbidities on asthma patients mostly focused on direct costs or health services use [7, 12, 15, 26]. For example, they have demonstrated that the rate of hospitalization due to asthma and Emergency Department (ED) visits in asthma patients increased in the presence of comorbidities [7, 12, 15, 26]. It has also been shown that the presence of some comorbidities increase the risk of mortality [10, 26]. The relationship between comorbidities and asthma exacerbations has also been demonstrated [6]. A study conducted in Finland showed that the presence of one and more than two comorbidities increased the risk of work disability with hazard ratios of 2.2 and 4.5, respectively [27]. In that study, work disability was defined as long-term sickness absence (≥90 days) and receiving a disability pension. Results of current study are inline with our previous study that demonstrated the presence of comorbid psychological conditions in asthma patients will increase productivity loss significantly [28]. To the best of our knowledge, there is no other study assessing the general impact of comorbidities on productivity loss in asthma patients including both absenteeism and presenteeism and transforming the productivity loss into its monetary value. The use of validated instruments enabled us to transform productivity loss time to its monetary value, incorporating the impact of the affected individual on team productivity, and the use of a robust statistical method enabled us to properly handle statistical issues around zero-inflated and skewed costs data. Besides these strengths, our study has several limitations worth mentioning. First, the final sample size (284) might have underpowered the results and our sample were mostly highly educated with high income, which could manifest the healthy volunteer bias. Second, our sample only included employed asthma patients. None of the participants in the original study reported being unemployed because of asthma. As such, we could not incorporate the loss of productivity for asthma patients who lose their job due to the asthma-related or comorbidity-related impairment. Third, self-reported physician diagnosis of asthma and self-reported comorbidities and productivity loss might reduce the accuracy of the data we used. Fourth, the percentage of patients with higher scores of SCQ was limited such that the results for the patients with SCQ scores of ≥10 should be interpreted cautiously. Ultimately, the aspect of the burden of a disease that is the most relevant for clinical practice and policy-making is the component that can be prevented by disease management. Having documented a significant association between comorbidity and productivity loss, the research agenda should move forward to studying specific comorbid conditions as well as the impact of treatment on preventing such loss of productivity. Conclusions Taking the limitations into account, our study has highlighted the important associations of comorbidities with productivity loss in working asthma patients. This is demonstrated by almost $1685/week higher productivity loss in patients with a SCQ score of 15 in comparison to those with a zero score. Productivity loss is a disregarded aspect of the economic burden of asthma [16]. Thus this study is a reminder for health care providers to pay greater attention to comorbidities in the management of asthma in order to reduce the burden of this common disease that disproportionately affects individuals in their productive years of life. Appendix Table 4 Costs of productivity loss per week (CAD)b, comparing the main model with alternative model adding control status as confounder SCQ score Main Model Alternative modela 0 205.12 ± 42.09 215.05 ± 50.51 1 244.42 ± 45.46 253.10 ± 53.85 2 290.07 ± 51.73 296.62 ± 59.65 3 342.81 ± 62.54 346.31 ± 69.40 4 403.47 ± 79.29 402.74 ± 84.45 5 472.90 ± 103.02 466.56 ± 105.91 6 552.02 ± 134.70 538.45 ± 134.69 7 641.82 ± 175.39 619.12 ± 171.69 8 743.35 ± 226.39 709.36 ± 217.94 9 857.74 ± 289.24 809.96 ± 274.69 10 986.20 ± 365.75 921.82 ± 343.39 11 1130.04 ± 458.03 1045.86 ± 425.72 12 1290.72 ± 568.50 1183.12 ± 523.60 13 1469.79 ± 699.91 1334.70 ± 639.22 14 1669.0 ± 855.37 1501.83 ± 775.04* 15 1890.27 ± 1038.37* 1685.85 ± 933.81* *P value is not significant aAlternative model: two part model regression analysis on the main model after adding control status b2010 Canadian dollars Abbreviations BCBritish columbia CADCanadian dollars EBA studyEconomic burden of asthma study EDEmergency department GINAGlobal initiative for asthma NOCNational occupation classification OROdds ratio PDCProportion of days covered (by any asthma controller medication) RRRelative rate SCQSelf-administered comorbidity questionnaire SDStandard deviation VOLPValuation of Lost Productivity WPAIWork productivity and activity impairment Acknowledgments and funding We acknowledge the financial support through the Collaborative Innovative Research Fund (CIRF), an investigator initiated, peer-reviewed competition sponsored by GlaxoSmithKline Canada. None of the sponsors played a role in the study design, data analysis, or interpretation of the results. Authors’ contribution JMF, MS, WT and DR designed the Economic Burden of Asthma (EBA) study, whose data are used in the present research. MS and RR designed the case report forms. JMF and SEA proposed the research question. SEA conceptualized the study design, was the main analyst, and wrote the first draft of the manuscript. JMF, MS, CC and HT were involved in the acquisition of the data. JMF, CC and MS critically revised the manuscript. JMF and SEA are guarantors of the paper. All authors approved the final version of the manuscript. Competing interest The authors declare that they have no competing interests. Consent for publication All authors approved the final version of the manuscript and agreed to its content and are accountable for all aspects of the accuracy and integrity of the manuscript in accordance with ICMJE criteria and agree to the terms of the BioMed Central Copyright and License Agreement, and Open Data policy. The article is original, has not already been published in a journal, and is not currently under consideration by another journal. Informed consent was obtained from all individual participants included in the study. Ethics approval and consent to participate The Human Ethics board of the University of British Columbia (University of British Columbia Human Ethics No. H10-01542) approved this study. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standard. ==== Refs References 1. Diederichs CP Wellmann J Bartels DB Ellert U Hoffmann W Berger K How to weight chronic diseases in multimorbidity indices? Development of a new method on the basis of individual data from five population-based studies J Clin Epidemiol 2012 65 6 679 685 10.1016/j.jclinepi.2011.11.006 22424984 2. Starfield B Kinder K Multimorbidity and its measurement Health Policy Amst Neth 2011 103 1 3 8 10.1016/j.healthpol.2011.09.004 3. Fortin M Lapointe L Hudon C Vanasse A Multimorbidity is common to family practice: is it commonly researched? Can Fam Physician Médecin Fam Can 2005 51 244 245 4. Ledford DK Lockey RF Asthma and comorbidities Curr Opin Allergy Clin Immunol 2013 13 1 78 86 10.1097/ACI.0b013e32835c16b6 23222157 5. Prosser R Carleton B Smith A The comorbidity burden of the treated asthma patient population in British Columbia Chronic Dis Can 2010 30 2 46 55 20302685 6. Zhang T Carleton BC Prosser RJ Smith AM The added burden of comorbidity in patients with asthma J Asthma Off J Assoc Care Asthma 2009 46 10 1021 1026 10.3109/02770900903350473 7. Patel MR Janevic MR Heeringa SG Baptist AP Clark NM An examination of adverse asthma outcomes in U.S. Adults with multiple morbidities Ann Am Thorac Soc 2013 10 5 426 431 10.1513/AnnalsATS.201302-032OC 23987919 8. Rank MA Liesinger JT Ziegenfuss JY Branda ME Lim KG Yawn BP Asthma expenditures in the United States comparing 2004 to 2006 and 1996 to 1998 Am J Manag Care 2012 18 9 499 504 23009300 9. Rank MA Shah ND Multiple chronic conditions and asthma: implications for practice and research J Allergy Clin Immunol Pract 2014 2 5 518 524 10.1016/j.jaip.2014.06.020 25213044 10. Sumino K O’Brian K Bartle B Au DH Castro M Lee TA Coexisting chronic conditions associated with mortality and morbidity in adult patients with asthma J Asthma Off J Assoc Care Asthma 2014 51 3 306 314 10.3109/02770903.2013.879881 11. Cazzola M Calzetta L Bettoncelli G Novelli L Cricelli C Rogliani P Asthma and comorbid medical illness Eur Respir J 2011 38 1 42 49 10.1183/09031936.00140310 21177843 12. Steppuhn H Langen U Scheidt-Nave C Keil T Major comorbid conditions in asthma and association with asthma-related hospitalizations and emergency department admissions in adults: results from the German National Health Telephone Interview Survey (GEDA) 2010 BMC Pulm Med 2013 13 46 10.1186/1471-2466-13-46 23849455 13. Brinke A t Sterk PJ Masclee A a M Spinhoven P Schmidt JT Zwinderman AH Risk factors of frequent exacerbations in difficult-to-treat asthma Eur Respir J 2005 26 5 812 818 10.1183/09031936.05.00037905 16264041 14. Doz M Chouaid C Com-Ruelle L Calvo E Brosa M Robert J The association between asthma control, health care costs, and quality of life in France and Spain BMC Pulm Med 2013 13 15 10.1186/1471-2466-13-15 23517484 15. Gershon AS Wang C Guan J To T Burden of comorbidity in individuals with asthma Thorax 2010 65 7 612 618 10.1136/thx.2009.131078 20627918 16. Ehteshami-Afshar S FitzGerald JM Doyle-Waters MM Sadatsafavi M The global economic burden of asthma and chronic obstructive pulmonary disease Int J Tuberc Lung Dis 2016 20 1 11 23 10.5588/ijtld.15.0472 26688525 17. Sadatsafavi M Rousseau R Chen W Zhang W Lynd L FitzGerald JM The preventable burden of productivity loss due to suboptimal asthma control: a population-based study Chest 2014 145 4 787 793 10.1378/chest.13-1619 24337140 18. Chen W Lynd LD FitzGerald JM Marra CA Rousseau R Sadatsafavi M The added effect of comorbidity on health-related quality of life in patients with asthma Qual Life Res Int J Qual Life Asp Treat Care Rehab 2015 24 10 2507 2517 10.1007/s11136-015-0995-6 19. Sangha O Stucki G Liang MH Fossel AH Katz JN The Self-Administered Comorbidity Questionnaire: a new method to assess comorbidity for clinical and health services research Arthritis Rheum 2003 49 2 156 163 10.1002/art.10993 12687505 20. Andreasson E Svensson K Berggren F PRP11 The validity of the work productivity and activity impairment questionnaire for patients with asthma (WPAIASTHMA): Results from a web-based study Value Health 2003 6 6 780 10.1016/S1098-3015(10)61986-9 21. Zhang W Bansback N Boonen A Severens JL Anis AH Development of a Composite Questionnaire, the Valuation of Lost Productivity, to Value Productivity Losses: Application in Rheumatoid Arthritis Value Health 2012 15 1 46 54 10.1016/j.jval.2011.07.009 22264971 22. Pauly MV Nicholson S Polsky D Berger ML Sharda C Valuing reductions in on-the-job illness: “presenteeism” from managerial and economic perspectives Health Econ 2008 17 4 469 485 10.1002/hec.1266 17628862 23. Government of Canada E and SDC. National Occupational Classification [Internet]. 2013. Available from: http://www30.hrsdc.gc.ca/NOC/english/NOC/2006/Welcome.aspx. Accessed 10 Mar 2016. 24. Mihaylova B Briggs A O’Hagan A Thompson SG Review of statistical methods for analysing healthcare resources and costs Health Econ 2011 20 8 897 916 10.1002/hec.1653 20799344 25. Yuan Y Multiple imputation for missing data: Concepts and new development (Version 9.0) 2010 Rockville SAS Inst Inc 49 26. Aubas C Bourdin A Aubas P Gamez AS Halimi L Vachier I Role of comorbid conditions in asthma hospitalizations in the south of France Allergy 2013 68 5 637 643 10.1111/all.12137 23573840 27. Hakola R Kauppi P Leino T Ojajärvi A Pentti J Oksanen T Persistent asthma, comorbid conditions and the risk of work disability: a prospective cohort study Allergy 2011 66 12 1598 1603 10.1111/j.1398-9995.2011.02729.x 21958351 28. Moullec G FitzGerald JM Rousseau R Chen W Sadatsafavi M Economic Burden of Asthma (EBA) study team. Interaction effect of psychological distress and asthma control on productivity loss? Eur Respir J 2015 45 6 1557 1565 10.1183/09031936.00141614 25657023

PMC005xxxxxx/PMC5002150.txt

==== Front Cancer ImagingCancer ImagingCancer Imaging1740-50251470-7330BioMed Central London 8610.1186/s40644-016-0086-0Research ArticleA method to assess image quality for Low-dose PET: analysis of SNR, CNR, bias and image noise Yan Jianhua jianhua.yan@gmail.com 123Schaefferkoette Josh dnrjds@nus.edu.sg 34Conti Maurizio maurizioconti@siemens.com 5Townsend David david_Townsend@circ.a-star.edu.sg 341 Department of Nuclear Medicine, First Hospital of Shanxi Medical University, 85 Jiefang S Rd, Yingze, Taiyuan, Shanxi 030001 China 2 Molecular Imaging Precision Medicine Collaborative Innovation Center, Shanxi Medical University, 85 Jiefang S Rd, Yingze, Taiyuan, Shanxi 030001 China 3 A*STAR-NUS, Clinical Imaging Research Center, Center for translational medicine, 14 medical drive, #B1-01, 17599 Singapore, Singapore 4 Department of Diagnostic Radiology, National University Hospital, Main Building, 5 Lower Kent Ridge Road, Level 3, 119074 Singapore, Singapore 5 Siemens Healthcare Molecular Imaging, 810 Innovation Drive, Knoxville, TN37932 USA 26 8 2016 26 8 2016 2016 16 1 2610 4 2016 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 Lowering injected dose will have an effect on PET image quality. In this article, we aim to investigate this effect in terms of signal-to-noise ratio (SNR) in the liver, contrast-to-noise ratio (CNR) in the lesion, bias and ensemble image noise. Methods We present here our method and preliminary results using tuberculosis (TB) cases. Sixteen patients who underwent 18F-FDG PET/MR scans covering the whole lung and portion of the liver were selected for the study. Reduced doses were simulated by randomly discarding events in the PET list mode data stream, and ten realizations at each simulated dose were generated and reconstructed. The volumes of interest (VOI) were delineated on the image reconstructed from the original full statistics data for each patient. Four thresholds (20, 40, 60 and 80 % of SUVmax) were used to quantify the effect of the threshold on CNR at the different count level. Image metrics were calculated for each VOI. This experiment allowed us to quantify the loss of SNR and CNR as a function of the counts in the scan, in turn related to dose injected. Reproducibility of mean and maximum standardized uptake value (SUVmean and SUVmax) measurement in the lesions was studied as standard deviation across 10 realizations. Results At 5 × 106 counts in the scan, the average SNR in the liver in the observed samples is about 3, and the CNR is reduced to 60 % of the full statistics value. The CNR in the lesion and SNR in the liver decreased with reducing count data. The variation of CNR across the four thresholds does not significantly change until the count level of 5 × 106. After correcting the factor related to subject’s weight, the square of the SNR in the liver was found to have a very good linear relationship with detected counts. Some quantitative bias appears with count reduction. At the count level of 5 × 106, bias and noise in terms of SUVmean and SUVmax are up to 10 and 20 %, respectively. To keep both bias and noise less than 10 %, 5 × 106 counts and 20 × 106 counts were required for SUVmean and SUVmax, respectively. Conclusions Initial results with the given data of 16 patients diagnosed as TB demonstrated that 5 × 106 counts in the scan could be sufficient to yield good images in terms of SNR, CNR, bias and noise. In the future, more work needs to be done to validate the proposed method with a larger population and lung cancer patient data. Keywords Low dosePET/MRPET/CTLungImage qualitySingapore Ministry of Healths National Medical Research CouncilNMRC/CIRG/1322/2012issue-copyright-statement© The Author(s) 2016 ==== Body Background PET is an important tool for in-vivo study of quantitative measurements of physiological, biochemical, or pharmacological processes. In thoracic oncology, 18F-FDG PET currently plays a major role in clinical diagnosis, staging, prognosis and assessment of response to treatment [1]. Recent advances in PET and CT technology have improved image quality while reducing radiation exposure to patients [2, 3], which open new avenues for cancer screening with PET and CT. Several studies showed that low dose CT is superior to traditional chest radiography for lung cancer screening and follow-up by detecting more nodules and lung cancers including early-stage cancers [4, 5]. In addition, low dose CT screening for subjects at high risk could reduce lung cancer mortality [6]. However, due to its limited specificity, low dose CT screening also detected more than 18 % of all lung cancers which were indolent and led to overdiagnosis in screening for lung cancer [7] although computer-aided diagnosis could improve performance of CT screening [8]. A large clinical study on CT screening of patients at risk has shown that even if cancer mortality is reduced by low dose CT screening, however, 24.2 % of the patients were tested positive, but 96.4 % of these were false positives [6]. This large number of false positives calls for imaging techniques with higher specificity, in order to avoid unneeded invasive biopsy. Additional metabolic information from 18F-FDG PET has been shown to be more specific than CT in detecting lung cancer [9]. Moreover, the combination of CT and PET demonstrated better performance in classifying solitary pulmonary nodules as benign or malignant than either PET or CT alone [8]. Thus, the synergetic effect of PET and CT could potentially improve the accuracy of screening for lung cancer [10]. Like low dose CT screening, the radiation exposure due to injected isotope should be minimized without compromising image quality of PET. The effective dose associated with 18F-FDG PET exam in this study was computed based on the reported ICRP values of 0.019 mSv/MBq for a 70 kg adult. For example, the effective dose is about 7 mSv for typical administration of 10 mCi 18F-FDG, which is much higher than that (1.5 mSv) of low dose CT protocol used in the National Lung Screening Trial [11]. The continual improvement of PET imaging, such as introduction of point spread function and time of flight technologies, could allow for lower injected activities while minimizing impact on image quality [12]. A number of studies to investigate the effect of different count levels on PET image quality with phantom have been reported [13, 14]. Our previous study demonstrated count statistics as low as 5 × 106 counts could achieve a fairly high detectability level using a data set of 18F-FDG PET images of tuberculosis (TB) patients acquired on a PET/MR scanner [15]. In this work, we aimed to assess the relationship between numbers of counts in PET scan and image quality with these data, based on image metrics such as liver signal-to-noise ratio (SNR), lesion contrast-to-noise ratio (CNR), bias relative to the “true value”, and ensemble noise in the image (lesion and normal tissue). Methods Data acquisitions Sixteen patients with TB (male: 12, female: 4) having a mean age of 45 years (range: [24–67]), a mean weight of 58.53 kg (range: [45–79]) and a mean BMI of 19.78 (range: [15.21–26.70]) underwent 18F-FDG PET/MR scans at A*STAR-NUS, Clinical Imaging Research Center using a Siemens mMR PET/MR scanner. The scanning usually started after the FDG uptake time of 60 min but were subject to the availability of the scanners. In addition, four patients had one PET/CT scan and one PET/MR scan but with one 18F-FDG injection. The interval between the two scans was around 60 min and the order of these two PET scans was random. However, only the PET/MR scan was included in this study. The PET spatial resolution of mMR has been measured, at 1 and 10 cm from the center of the field of view (FOV), as 4.3 mm and 5.2 mm transaxially, and 4.3 mm and 6.6 axially [16]. The coincidence timing window is 5.9 ns and the energy window is 435–650 keV. The detector system includes 8 rings of 56 detector blocks, each comprising an 8 × 8 matrix of lutetium oxyorthosilicate crystals (4 × 4 × 20 mm), coupled to an array of 3 × 3 APDs. The axial PET FOV is 25.2 cm, and all emission data were acquired in 3D mode. The data were organized into separate prompts and delayed sinograms. The delayed events were smoothed and used to estimate the overall random coincidences. All patients fasted at least for 8 h with serum glucose level less than 10 mmol/L. The time difference between injection and acquisition was 80.1 ± 26.2 min (range: [36.5–126.4] min) after injection of 168.6 ± 50.0 MBq (range: [118.0–260.5] MBq) 18F-FDG, with 15 min for one bed position covering the whole lung and portion of the liver. The true coincident events, after random subtraction, were 1.32 × 108 ± 3.91 × 107 (range: [8.33 × 107–2.00 × 108]). Data reconstructions The reduced doses were simulated by randomly discarding events in each list mode stream according to 10 predefined fractions of original net true counts: 5 × 10−1, 2.5 × 10−1, 1.25 × 10−1, 6.25 × 10−2, 3.33 × 10−2, 1.67 × 10−2, 5 × 10−3, 3.33 × 10−3, 1.67 × 10−3, and 5 × 10−4. The random events in the list mode data stream were discarded using the same procedure as resampling true events. Ten independent realizations at each simulated dose were generated. Each realization was reconstructed with ordinary Poisson ordered subsets expectation maximization (OP-OSEM) with a system point spread function (PSF) incorporated in the projection matrix [17]. Corrections including attenuation, randoms, and scatter were carried out for each realization. Three iterations and 21 subsets were used to produce image matrices of 172 × 172 × 127 with voxel sizes of 4.17 × 4.17 × 2.03 mm. The image volumes were then smoothed with a 5 mm Gaussian filter. VOI delineation In this study, the mean standardized uptake value (SUVmean) was calculated as 18F-FDG uptake normalized to injected dose and patient body weight and activity was decay corrected. Volume of interest (VOI) around solitary lung lesions were delineated using fixed threshold set to 40 % of the maximum standardized uptake value (SUVmax) in the lesion and followed by a manual adjustment to exclude neighboring nodes for each VOI if there was any in the VOI. This thresholding was used to investigate the effect of count level on CNR in the lesion, lesion bias with reference to SUV calculated in the images at the full statistical count level, noise across realizations and reproducibility of the SUVmean and SUVmax in the lesion. Twenty small lesions were obtained from these 16 subjects and the volume was 6.38 ± 4.75 ml (range: [1.2−17.58 ml]). The VOIs in the normal lung background and liver background were obtained by drawing spheres with diameter of 3 cm in these two regions. All of the VOIs were delineated on the images reconstructed with the original full statistics data. Image analysis For the evaluation of image quality, five metrics were used 1) SNR in the liver, 2) CNR in the lesion, 3) bias with reference to SUV calculated in the images at the full statistics count level, 4) noise in the image expressed as percentage coefficient of variation (COV) in VOIs including lesion, normal lung and liver, across 10 realizations, and 5) error or reproducibility of the SUVmean and SUVmax in the VOIs including lesion, normal lung and liver, across 10 realizations.The signal to noise ratio (SNR) was calculated as the ratio of mean value to standard deviation (SD) in the VOI: 1 MeanVOI=∑j∈VOIImeanjNVOI 2 SDVOI=∑j∈VOIImeanj−MeanVOI2NVOI 3 SNRVOI=MeanVOISDVOI where Mean(VOI) and SD(VOI) are mean value and SD value of the VOI in the mean image (Imean) across the realizations, respectively, NVOI is the number of voxels in the VOI. VOI could be in the lesion, clear lung background and liver. The SNR in the liver was widely used to quantify 18F-FDG PET image quality due to its relatively homogeneous uptake. Due to the Poisson statistics of nuclear positron emission, the SNR of PET images depends on the injected activity, the acquisition time, and the attenuation [18, 19], and it can be expressed as follows: 4 SNRliver2≈N≈Ki•D•t 5 Ki=k•g•η•1am where N is the number of detected counts, Ki is a composite sensitivity factor, D is the injected activity and t is the scan time, k is a proportionality constant, g is a noise reduction factor due to reconstruction techniques such as PSF and time-of-flight, η is the scanner sensitivity, and a(m) is the attenuation factor, a function of subject’s weight m. The corrected SNR2 in the liver can be made independent of patient dividing by Ki. The contrast to noise ratio (CNR) is a measure of the signal level in the presence of noise given by [20]: 6 CNR=Meanlesion−MeanbackgroundSDbackground where background is measured in the neighboring normal lung tissue, Mean(lesion) and Mean(background) are the mean value of the lesion and background region in the mean image of the SUV across the realizations calculated with Eq. 1, respectively and SD(background) is the SD value of the background region in the mean image across the realizations calculated with Eq. 2. The background mask was obtained by performing morphological operation on the lesion mask: 7 backgroundmask=dilationlesionmask2−lesionmask where dilation(lesionmask, 2) means dilating the lesion mask by 2 voxels. Bias describes the difference of estimated SUV from the true value, which is unknown in the clinic. In this work, the percentage difference relative to the activity uptake at full statistics count level is used to estimate bias. 8 Biasfrac%=MeanfracVOI−MeanfullVOIMeanfullVOI*100 where Biasfrac represents the bias in percent at the predefined fraction of counts, Meanfrac is the mean value of the VOI in the mean image (Imeanfrac) across the realizations at the predefined fraction count level. The Meanfull(VOI) is the mean value of the VOI in the image at the full count level (Imeanfull), typically greater than 100 × 106 counts. Both Meanfrac and Meanfull(VOI) were calculated with Eq.1. The coefficient of variation (COV) is a metric for describing ensemble noise or statistical noise in the image, and it can influence the detectability of the lesion. 9 COVfracVOI=MeanmeanfracVOIMeanSDfracVOI*100 where Meanmeanfrac(VOI) and MeanSDfrac(VOI) are the mean value of the VOI in the mean image and the SD image across the realizations at the predefined fraction count level calculated with Eq.1, respectively. Standard error (STE) is a metric for describing error on the measurement of SUV, and it can represent the reproducibility of the measurement. 10 STEfrac%=SDfracMeanfracVOIMeanfracVOI*100 where SDfrac is the standard deviation (across the realizations) of the SUV in the VOI at the predefined fraction count level, normalized to the SUV averaged across realizations. The effect of different thresholding (20, 40, 60 and 80 % of SUVmax) on CNR at different count level was investigated. In this comparison, the surrounding background mask delineated with 20 % of SUVmax thresholding was applied to the other three thresholding. Each CNR was calculated from the mean image across realizations with Eq. 6. COV of each CNR was calculated across the four thresholding to demonstrate the sensitivity of the CNR to the thresholding. The bias, COV, and STE were studied as a function of the counts in the scan, in the lung lesions, in the lung background, and in the liver, in the pooled sample of patients and lesions. For each count level we histogrammed the number of data point that pass a given threshold (in percent), which refers to either bias, COV, or STE for SUVmean and SUVmax. We used the mean of the histogram distribution to define the number of counts associated with such percent average bias (or COV, or STE). We finally plotted counts vs. percent variation for SUVmean or SUVmax bias (or COV, or STE) for TB lesions, background lungs and uniform liver region. This would allow a quantitative assessment of the number of counts needed for a given acceptable error in the measurement, systematic or statistical. Results Effect of counts on SNR in the liver The average SUVmean in the liver for the images reconstructed with the original full statistics data was 1.57 ± 0.40 (range: [0.54–2.12]). The SNR2 in the liver for the images with full statistics, for the images with fewer than 20 × 106 true counts, and for the images with fewer than 1 × 106true counts are shown in Fig. 1(a, b and c), respectively. The Ki for each subject was obtained by fitting the SNR2 with the number of counts (y = 0.11x0.69, R2 = 0.1), according to Eq. (4), and is shown in Fig. 1(d). The SNR2 corrected by the Ki are shown in Fig. 1(e and f) with true counts less than 20 × 106 and 1 × 106, respectively. One can see that the corrected SNR2 has better linear relationship with detected counts than the original SNR2.Fig. 1 a The SNR2 in the liver for all the reconstructed images; (b) The SNR2 in the liver for the images reconstructed with fewer than 20 × 106 true counts; (c) The SNR2 in the liver for the images reconstructed with fewer than 1 million true counts; (d) The composite sensitivity factor Ki for all subjects and the corresponding fitted curve; (e) The corrected SNR2 in the liver for the images reconstructed with fewer than 20 million true counts and the linearly fitted curve; (f) The corrected SNR2 in the liver for the images reconstructed with raw data less than 1 × 106 true counts and the linearly fitted curve In PET imaging, photons registered in the detectors follow a Poisson distribution and the standard deviation on the number of counts N is proportional to the square root of N, therefore, SNR2 of the counts registered in the detectors is linear with N. Iterative reconstruction method based on maximum likelihood estimate is the most popular method in the most of PET scanner, which is a nonlinear method and leads to the non-linear relationship of the SNR2 measured with the reconstructed image with N. However, in the case of high statistics, the nonlinear transformation can be approximate to a linear one. As expected, the proportionality coefficient in the mid-range of 1–20 × 106 counts is 1 (Fig. 1(e)), but at extremely low counts such proportionality does not hold anymore, and applying a linear fit the coefficient is now 1.96 (Fig. 1(f)). This represents a loss of proportionality at low counts, when Poisson statistics does not represent the data well anymore, and the relationship between counts and SNR2 is not linear. Effect of counts on CNR in the lesion Twenty six lesions were delineated by 40 % of SUVmax thresholding method and the lesions with volume less than 20 ml, corresponding to approximate 3.37 cm of diameter for spherical lesion, were found in these 16 subjects with average SUVmean of 1.92 ± 0.95 (range: [0.92–4.81]). Typical PET images with five realizations with a lesion of CNR of 3.17 and volume of 1.66 ml at the different count level are shown in Fig. 2. As expected, image quality decreases with decreasing counts. The lesion can be visually detected in the image at the count level of more than 3 × 106. Since it is challenging to reliably identify lesion with low CNR, especially at low count level, lesions with CNR less than 2 were excluded from this study. Twenty lesions were found to satisfy the requirement of volume (≤20 ml) and CNR (≥2).Fig. 2 Typical PET images (SUV: 0–6) with a lesion of CNR (=3.17) and volume (1.66 ml) at the different count level (5 realizations are displayed). Each column represents one realization. The count level corresponding to each row is 0.048 × 106, 0.16 × 106, 0.32 × 106, 0.48 × 106, 1.6 × 106, 3.2 × 106, 6 × 106, 1.2 × 107, 2.4 × 107, 4.8 × 107 Figure 3(a) shows CNR of all lesions at different count levels and Fig. 3(b) displays CNR at the count level range of 0–20 × 106. CNR decreases with the lowering of count. All CNR were normalized to the value at the full count level and are shown in Fig. 3(c and d). Normalized CNR can be fitted with count by a function (y = 1/(1 + 2.41x− 0.8), R2 = 0.91). This curve allows to predict the amount of counts required to obtain a desired normalized CNR.Fig. 3 Plot of CNR at different count levels. a CNR at the full count range, (b) CNR at counts fewer than 20 × 106, (c) normalized CNR (CNR normalized by the value at full count level) at the full count range and the fitted function, (d) normalized CNR at counts fewer than 20 × 106 and the fitted function Effect of thresholding on CNR Figure 4(a) shows the mean CNR of all lesions at different count level using different thresholding method. It can be seen that higher thresholding method produces higher CNR. In addition, all lesions’ CNR decrease with lowering counts. Figure 4(b) demonstrates the mean COV of all lesions’ CNR across the realizations, which indicates that the mean COV do not significantly vary until a count level of 5 × 106.Fig. 4 Effect of different threshold on CNR at the different count level: (a) the mean CNR of all lesions at different count level. b the mean COV of all lesions’ CNR across the realizations. The variation over the realizations is demonstrated as the error bar Effect of counts on bias and COV Figure 5(a and b) shows the mean SUVmean and SUVmax in the normal lung, liver and lesion, respectively. Figure 5(c) exhibits the CNR in the same lesion at variable count levels for the same patient. One can see that the SUVmean in these three regions for this subject do not significantly change, relative to the full count level, down to a count level of 1 × 106. For the SUVmax, the count level is around 8 × 106.Fig. 5 Example of SUV and CNR vs. counts (patient as in Fig. 4): (a) SUVmean in the normal lung (blue), liver (green) and TB lesion (red) at the different count level. b SUVmax in the normal lung (blue), liver (green) and TB lesion (red) at the different count level. c CNR in the lesion at the different count level. The variation over the realizations is demonstrated as the error bar The variation of the SUVmean and SUVmax over the realizations increases with reducing counts. All patients and lesions were pooled and analyzed in order to study relationship between counts in the scan and two key parameters: measurement bias and error. In Fig. 6, an example of histogram representing the frequency of cases with COV over 10 % at given counts in the scan, in different regions based on SUVmean and SUVmax. Each bin represents the sum of patients (Fig. 6(a, d, b and e)) or lesions (Fig. 6(c and f)) at each count level with COV larger than 10 %. Histograms were created for variable error levels, and the mean and maximum value of the distribution was computed and plotted in Fig. 7(a and d). Similarly, in Fig. 7(b and e), the STE on SUVmean and SUVmax is plotted vs. counts level. In Fig. 7(c and f), the measurement bias, defined as difference relative to the full statistics SUVmean and SUVmax, is plotted vs. counts level.Fig. 6 Frequency of cases with COV over 10 % based on SUVmean vs. counts in the scan, in different regions: (a) background liver, (b) background lung, (c) TB lesion and based on SUVmax vs. counts in the scan, in different regions: (d) background liver, (e) background lung, (f) TB lesion. Histogram bin size is 1 × 106 Fig. 7 Minimum number of counts needed to keep (a) COV for SUVmean, (b) STE of SUVmean, (c) SUVmean bias, (d) COV based on SUVmax, (e) STE of SUVmax, and (f) SUVmax below a given percent level. VOIs in different regions are shown: background lungs (blue), liver (green), and TB lesions (red) From Fig. 7(a), the counts required to keep the COV for SUVmean below 10 % for half of the cases in the sample are about 25 × 106, 25 × 106 and 5 × 106, respectively for the liver, lung, and TB lesions. The corresponding counts to keep the COV for SUVmax less than 10 % for half of the cases are about 56 × 106, 52 × 106 and 20 × 106, respectively for the liver, lung, and TB lesions from Fig. 7(d). From Fig. 7(b and e), only 1 × 106, 2.7 × 106 and 0.4 × 106 counts for SUVmean and 6.6 × 106, 8.5 × 106 and 2.8 × 106 counts for SUVmax are needed to reach the STE of 10 % for the liver, normal lung and lesion, respectively. From Fig. 7(c and f), only 2.2 × 106, 1 × 106 and 0.2 × 106 counts for SUVmean and 28.5 × 106, 24.2 × 106 and 5.2 × 106 counts for SUVmax are needed to reach the percentage bias of 10 % for the liver, normal lung and lesion, respectively. At 5 × 106 counts in the scan, bias and noise in terms of SUVmean and SUVmax are up to 10 and 20 %, respectively. The 20 lesions were categorized into two groups based on volume threshold of 5 ml and each group had 10 lesions. One subgroup has mean volume 2.71 ml [1.20–4.81 ml] and another subgroup has mean volume 9.54 ml [5.48–17.58 ml], approximately 17.3 mm [13.2 mm-21 mm] and 26.3 mm [22 mm-32.2 mm] in diameter, respectively. The counts needed to reach the same bias, STE and COV percentage for large lesions (volume ≥ 5 ml) are fewer than that for small lesions (volume ≤ 5 ml) for SUVmean, although the difference between the two groups is small. In addition, the bias and STE for SUVmax were the same for the two groups, as shown in Fig. 8.Fig. 8 Minimum number of counts needed to keep (a) COV for SUVmean, (b) STE of SUVmean, (c) SUVmean bias, (d) COV for SUVmax, (e) STE of SUVmax and (f) SUVmax bias below a given percent level, for TB lesions. Lesions are separated in two classes: smaller than 5 ml and bigger than 5 ml Discussion We evaluated image quality with objective metrics including SNR in the liver, CNR in the lesions, bias and noise in the liver, normal lung and lesions, at simulated reduced doses, using 18F-FDG PET data at various count levels from TB patients. The underlying biology of TB is different from that of lung cancer, but for a technical study such as this, the uptake levels in TB lesions will be more representative of early stage lung cancer lesions than those in more advanced lung cancer. This work will lay the foundation to determine the appropriate dose or scan time for a future prospective study with lung cancer patients PET/CT scanning. Accurate delineation of lesions is a prerequisite for quantification of FDG uptake. Although a large number of approaches have been proposed to segment tumors in PET images including threshold based, gradient based [21], and fuzzy Bayesian based methods [22], accurate tumor segmentation is still a challenging task. This is due to limited spatial resolution and the relatively high noise level in PET images, and this process evidently becomes more challenging with fewer counts (Zaidi and El Naqa, [23]). A simple thresholding method was employed here to segment the solitary lesions in the lung using the full statistics images and the resulting VOIs were copied to the images at the lower count levels. This simple thresholding method may lead to imperfect delineation of the tumor. In addition, the spill-out from the tumor to the surrounding background can lead to lower CNR. However, the inaccurate delineation will not change the behavior of the image metrics since the error will have the same effect at the different count levels, which is partially supported by the result of CNR varying with different threshold (20, 40, 60 and 80 % of SUVmax). Since we work towards low dose PET imaging for those patients at high risk who have indefinite findings with low dose CT screening, VOIs can be delineated on the CT image and copied onto registered PET images. Figure 1(a) is the plot of SNR2 in the liver for all patients at all count levels and Fig. 1(d) shows the Ki for all patients derived from the SNR2 in the liver for each patient at all count level. The SNR in the liver depends on many factors including scanner sensitivity, administered dose, scan time and patient-dependent parameter such as weight. In this study, after adjusting the effect of patient’s weight, the SNR2 was found to have a very good linear relationship with detected counts (y = 1.00 × + 0.43, R2 = 0.99), which fits with Poisson statistics of nuclear positron emission [18, 19]. At very low counts, the Poisson statistics approximation does not hold anymore, and SNR2 acquires a sharper decrease as shown in Fig. 1(b and c), and in Fig. 1(e and f). Since the SNR2 in the liver will become nonlinear at very low count levels, the slope of the fit for the count range of 0–1 × 106 will be different from that of the fit at the count range of 0–20 × 106. As expected, also the CNR in the lesion decreases with decreasing counts. After being normalized to CNR at full count data, a function can be used to predict CNR at different count level. A heuristic curve that best fits the CNR was used. In this phase, no interpretative model is proposed for the behavior of the curve, but we applied the simplest function that could fit the experimental data and allow for the prediction of CNR at low counts. A consequence of dose or count reduction is a possible bias in SUVmean or SUVmax measurement, and a larger error in the measurement, or degradation of measurement reliability, and an increase of noise in the image, which affects detectability of small lesions. This effect has been studied and a bias has been observed, as well as an increase of noise as COV, and an increase of STE of the SUVmean and SUVmax measurement in different regions of the patient. Several factors influence this including a positive bias in the cold background and negative bias in the hot regions associated with the positivity constraint of the OSEM reconstruction for SUVmean. The SUVmax is easily impacted by count reduction than SUVmean. In Fig. 5, one can observe larger error bars and SUV instability at very low counts. As shown in Fig. 7, fewer counts in the scan correspond to higher COV, STE or error of the radioactivity measurement for both SUVmean and SUVmax, as well as a higher bias relative to high count rate. In comparison with SUVmean, more counts are needed to keep the same acceptable level for COV, bias and STE for SUVmax. In terms of counts required to obtain the same percent error, fewer counts are needed for lesions than for liver and lung background. For example, if the acceptable level for COV for SUVmean is 10 %, the corresponding required counts are 5 × 106, 25 × 106, and 25 × 106, respectively for TB lesions, liver and normal lungs (Fig. 7(a)). This can be explained by the inherent “real” local variations of uptake values in large VOIs in lungs and liver, which are clearly not uniform. In addition, the local variations can explain the different appearance of histograms in the Fig. 6. The number of counts to maintain the same COV for SUVmax are 20 × 106, 56 × 106 and 52 × 106 for TB lesions, liver and normal lungs (Fig. 7(d)). The actual variations add to the statistical noise. Apparently, bias for SUVmean is less sensitive to count statistics, and minimum bias can be reached even below 1 × 106 counts. In addition, the same acceptable level for STE regarding to SUVmean can be maintained with fewer counts as compared to COV, which is demonstrated by 0.4 × 106, 2.7 × 106 and 1 × 106 counts required to reach STE threshold of 10 % for TB lesions, liver and normal lungs, respectively (Fig. 7(b)). Preliminary investigations show that the behavior of SUVmax is similar, but SUVmax is much more sensitive to noise, and comparable levels can be reached only with much higher number of counts in the scan. An additional study was done by splitting the lesion group into two subgroups: small lesions, with volume smaller than 5 ml; and large lesions, with volumes greater than 5 ml. Each subgroup has 10 lesions. As demonstrated in Fig. 7, for a given number of counts in the scan, bias, COV and STE are larger for smaller volume lesions. Regarding to SUVmean, choosing a target number of counts of about 5 × 106 counts, a noise level (COV) of 9 % is obtained for large lesions, and of 12 % for small lesions. A STE of 4 and 2 % are reached at 5 × 106 counts for small lesions and large lesions, respectively. Finally, the bias at 5 × 106 counts for both large lesions and small lesions are less than 2 %. At the same count level of 5 × 106 counts, lesion CNR is about 60 % of value at the full statistics level (data in Fig. 3(d)) and SNR in the liver is about 3 (data from Fig. 1(b)). The independent realizations at different low dose were obtained by randomly discarding the events in the list mode data stream, based on the desired count level and thus these realizations are not fully independent. In addition, we got very similar results with bootstrap resampling [24, 25], which was considered as a better method to produce independent realizations. For example, the fitted function for the SNR2 in the liver for the images reconstructed with fewer than 1 million true counts y = 2.85x + 0.19, R2 = 0.81, which is close to that function (y = 2.9x + 0.20, R2 = 0.79) with the current simulation strategy. This could be due to the fact that most of the image metrics in this work were based on the mean value across these realizations. Therefore, in order to keep in line with our previous study [15], the results with the simulations by randomly discarding the events in the list mode data stream were presented in this work. In the earlier works by Budinger, TF, et al. [26], Hoffman, EJ, et al. [27] and Strother, SC, et al. [28], the relationship between image counts and noise (root-mean-square) had been investigated. These earlier works used a different reconstruction scheme and our findings cannot be compared directly. Notwithstanding this, the statistical noise (COV) in the liver for each subject at the different count level in this study was close to the root-mean square calculated the equation found in the earlier studies up to the count level of 5 million. In this work, we chose OSEM reconstruction with 3 iterations, 21 subsets and post reconstruction Gaussian smoothing with FWHM of 5 mm which are commonly used parameters in the clinical settings. In future, we will investigate the optimization of the reconstruction parameters with prospective lung cancer patient data once the impact on image quality of reducing counts is generally understood. In addition, the impact of inaccurate attenuation map on quantitative PET lung imaging due to respiratory motion will be thoroughly explored with the data at different counts levels. Conclusions We developed a method and the tools for objectively evaluating PET images at various count levels with 18F-FDG PET data of TB patients acquired on a combined PET/MR scanner. This work allowed us to quantify the loss of SNR and CNR as a function of the counts in the scan, in turn related to dose injected. At 5 × 106 counts in the scan, the average SNR in the liver in the observed samples is about 3, and the CNR is reduced to 60 % of value at the full statistics level. At the count level of 5 × 106, bias and noise in terms of SUVmean and SUVmax are up to 10 and 20 %, respectively. This initial investigation presents the first step in a comprehensive image analysis of low dose PET imaging and lays the foundation for future study on low dose PET imaging for lung cancer, which will extend analysis to a larger and lung cancer patient sample, and assess if image quality at reduced counts is sufficient for lung cancer screening. Abbreviations CNRContrast-to-noise ratio COVCoefficient of variation OP-OSEMOrdinary poisson ordered subsets expectation maximization PSFPoint spread function SDStandard deviation SNRSignal-to-noise ratio SUVStandardized uptake value SUVmaxmMaximum SUV SUVmeanMean SUV TBTuberculosis VOIVolumes of interest The authors would like to acknowledge support for this study from Drs Nick Paton and James Molton. This research was partially supported by the Singapore Ministry of Health’s National Medical Research Council under its CS-IRG Grant NMRC/CIRG/1322/2012 and the National University of Singapore under its Start-up Grant. Funding This study was supported by the National Natural Science Foundation of China 81671775. Availability of data and materials None. Authors’ contributions All authors read and approved the final manuscript. Study concept and design: JY, JS and MC. Acquisition of data: JY and JS. Analysis and interpretation of data: JY. Drafting of the manuscript: JY and JS. Critical revision of the manuscript: JY, JS, MC and DT. Competing interests The authors declare that they have no competing interests. Consent for publication All authors read the final manuscript and consent for publication. Ethics approval and consent to participate This study was approved by the Domain Specific Review Board and Institutional Review Board of National Healthcare Group, Singapore and written consent form was obtained from each patient. ==== Refs References 1. Ambrosini V Nicolini S Caroli P PET/CT imaging in different types of lung cancer: an overview Eur J Radiol 2012 81 988 1001 10.1016/j.ejrad.2011.03.020 21458181 2. Willowson KP Bailey EA Bailey DL A retrospective evaluation of radiation dose associated with low dose FDG protocols in whole-body PET/CT Australas Phys Eng Sci Med 2012 35 49 53 10.1007/s13246-011-0119-8 22160927 3. Yanagawa M Gyobu T Leung AN Ultra-low-dose CT of the lung: effect of iterative reconstruction techniques on image quality Acad Radiol 2014 21 695 703 10.1016/j.acra.2014.01.023 24713541 4. Bach PB Mirkin JN Oliver TK Benefits and harms of CT screening for lung cancer: a systematic review JAMA 2012 307 2418 2429 10.1001/jama.2012.5521 22610500 5. Hanna WC Paul NS Darling GE Minimal-dose computed tomography is superior to chest x-ray for the follow-up and treatment of patients with resected lung cancer J Thorac Cardiovasc Surg 2014 147 30 33 10.1016/j.jtcvs.2013.08.060 24139896 6. Aberle DR Adams AM Berg CD Reduced lung-cancer mortality with low-dose computed tomographic screening N Engl J Med 2012 365 395 409 21714641 7. Patz EF Jr Pinsky P Gatsonis C Overdiagnosis in low-dose computed tomography screening for lung cancer JAMA Intern Med 2014 174 269 274 10.1001/jamainternmed.2013.12738 24322569 8. Kim SK Allen-Auerbach M Goldin J Accuracy of PET/CT in characterization of solitary pulmonary lesions J Nucl Med 2007 48 214 220 17268017 9. Schrevens L Lorent N Dooms C Vansteenkiste J The role of PET scan in diagnosis, staging, and management of non-small cell lung cancer Oncologist 2004 9 633 643 10.1634/theoncologist.9-6-633 15561807 10. Veronesi G Bellomi M Veronesi U Role of positron emission tomography scanning in the management of lung nodules detected at baseline computed tomography screening Ann Thorac Surg 2007 84 959 965 10.1016/j.athoracsur.2007.04.058 17720408 11. Kramer BS Berg CD Aberle DR Prorok PC Lung cancer screening with low-dose helical CT: results from the National Lung Screening Trial (NLST) J Med Screen 2011 18 109 111 10.1258/jms.2011.011055 22045816 12. Conti M Focus on time-of-flight PET: the benefits of improved time resolution Eur J Nucl Med Mol Imaging 2011 38 1147 1157 10.1007/s00259-010-1711-y 21229244 13. Doot RK Scheuermann JS Christian PE Karp JS Kinahan PE Instrumentation factors affecting variance and bias of quantifying tracer uptake with PET/CT Med Phys 2010 37 6035 6046 10.1118/1.3499298 21158315 14. Oehmigen M Ziegler S Jakoby BW Georgi JC Paulus DH Quick HH Radiotracer Dose Reduction in Integrated PET/MR: Implications from National Electrical Manufacturers Association Phantom Studies J Nucl Med 2014 55 1361 1367 10.2967/jnumed.114.139147 25006216 15. Schaefferkoetter JD Yan J Townsend DW Conti M Initial assessment of image quality for low-dose PET: evaluation of lesion detectability Phys Med Biol 2015 60 5543 5556 10.1088/0031-9155/60/14/5543 26134119 16. Delso G Furst S Jakoby B Performance measurements of the Siemens mMR integrated whole-body PET/MR scanner J Nucl Med 2011 52 1914 1922 10.2967/jnumed.111.092726 22080447 17. Panin VY Kehren F Michel C Casey M Fully 3-D PET reconstruction with system matrix derived from point source measurements IEEE Trans Med Imaging 2006 25 907 921 10.1109/TMI.2006.876171 16827491 18. Watson CC Casey ME Bendriem B Optimizing injected dose in clinical PET by accurately modeling the counting-rate response functions specific to individual patient scans J Nucl Med 2005 46 1825 1834 16269596 19. de Groot EH Post N Boellaard R Wagenaar NR Willemsen AT van Dalen JA Optimized dose regimen for whole-body FDG-PET imaging EJNMMI Res 2013 3 63 10.1186/2191-219X-3-63 23938036 20. Le Meunier L Slomka PJ Dey D Enhanced definition PET for cardiac imaging J Nucl Cardiol 2010 17 414 426 10.1007/s12350-010-9193-7 20151238 21. Geets X Lee JA Bol A Lonneux M Gregoire V A gradient-based method for segmenting FDG-PET images: methodology and validation Eur J Nucl Med Mol Imaging 2007 34 1427 1438 10.1007/s00259-006-0363-4 17431616 22. Hatt M Cheze le Rest C Turzo A Roux C Visvikis D A fuzzy locally adaptive Bayesian segmentation approach for volume determination in PET IEEE Trans Med Imaging 2009 28 881 893 10.1109/TMI.2008.2012036 19150782 23. Zaidi H El Naqa I PET-guided delineation of radiation therapy treatment volumes: a survey of image segmentation techniques Eur J Nucl Med Mol Imaging 2010 37 2165 2187 10.1007/s00259-010-1423-3 20336455 24. Lartizien C Aubin JB Buvat I Comparison of bootstrap resampling methods for 3-D PET imaging IEEE Trans Med Imaging 2010 29 1442 1454 10.1109/TMI.2010.2048119 20409989 25. Markiewicz PJ Reader AJ Matthews JC Assessment of bootstrap resampling performance for PET data Phys Med Biol 2015 60 279 299 10.1088/0031-9155/60/1/279 25490178 26. Budinger TF Derenzo SE Greenberg WL Gullberg GT Huesman RH Quantitative potentials of dynamic emission computed tomography J Nucl Med 1978 19 309 315 632910 27. Hoffman EJ Huang SC Phelps ME Quantitation in positron emission computed tomography: 1. Effect of object size J Comput Assist Tomogr 1979 3 299 308 10.1097/00004728-197906000-00001 438372 28. Strother SC Casey ME Hoffman EJ Measuring PET scanner sensitivity: relating counrates to image signal-to-noise ratios using noise equivalent counts IEEE Trans Nucl Sci 1990 37 783 788 10.1109/23.106715

PMC005xxxxxx/PMC5002151.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 357410.1186/s12889-016-3574-zResearch ArticleThe relationship between sports facility accessibility and physical activity among Korean adults Lee Sang Ah ivory0817@yuhs.ac 12Ju Yeong Jun joomeon@yuhs.ac 12Lee Joo Eun jooeun@yuhs.ac 12Hyun In Sun ishyun@yuhs.ac 12Nam Jin Young jynam@yuhs.ac 12Han Kyu-Tae kthan@yuhs.ac 12Park Eun-Cheol +82-2-2228-1862ecpark@yuhs.ac 231 Department of Public Health, Graduate School, Yonsei University, Seoul, Republic of Korea 2 Institute of Health Services Research, Yonsei University, Seoul, Republic of Korea 3 Department of Preventive Medicine and Institute of Health Services Research, Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul, 120-752 Republic of Korea 26 8 2016 26 8 2016 2016 16 1 89329 2 2016 23 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 benefits of physical activity on physical and mental health are well known. The accessibility of sports facilities is reported to have considerable association with the amount of physical activity a person participates in. Therefore, we investigated the association between subjectively assessed accessibility of sports facilities and physical activity among Korean adults. Methods We obtained data from the 2012 Community Health Survey. Physical activity was measured based on weekly metabolic equivalent task (MET) hours according to the International Physical Activity Questionnaire (IPAQ). Sociodemographic, economic, and health variables were used as covariates in a logistic regression model. Results A total 201,723 participants were included in this study. Participants with easy access to sports facilities participated in physical activity more often than those without easy access (OR = 1.16, 95 % CI 1.13–1.20). More physical activity was generally observed if participants had a history of depression or if participants were among the white-collar or urban subgroups. Conclusion Our results showed that the accessibility of sports facilities is associated with physical activity. Therefore, it is crucial to consider the accessibility of sports facilities when promoting an environment conducive to physical activity or designing programs for enhancing physical activity. Keywords Physical activityAccessibilityDepressionissue-copyright-statement© The Author(s) 2016 ==== Body Background The health benefits of physical activity are well known throughout the world. Participating in regular physical activity declines mortality and has positive effect on chronic diseases, such as cardiovascular disease, hypertension, and diabetes mellitus [1]. Additionally, physical activity has a beneficial effect on mental diseases such as anxiety or depression [2]. It reduces stress and depression, and increases self-confidence and emotional well-being [3]. Thus, promoting physical activity is an important part of enhancing public health. Despite these numerous benefits, physical activity levels have declined in the developed and developing countries [4]. Members of vulnerable social groups have an especially increased incidence of participating in unhealthy behaviors, including physical inactivity [5]. Additionally, those with lower levels of education or low economic status participate in physical activity less often compared to those of higher education and economic status [6]. The high rate of physical inactivity among these groups can cause significant public health problems. In addition to these groups, most people who suffer from depression also do not engage in physical activity compared to the general population [7]. According to the Epidemiology Survey of Mental Disorders in Korea 2011, the prevalence of depression is on the rise compared to other mental disorders. Additionally, lifetime prevalence and 1 year prevalence were almost 20 % higher in 2011 compared to 2006 [8]. Thus, identifying effective interventions to relieve this alarming depression rate is imperative. Considering the various benefits of physical activity, promoting physical activity may prove to be a successful intervention for alleviating depression and enhancing public health. In order to promote physical activity rates, identifying the factors associated with increased physical activity is essential. Several determinants of physical activity have been reported. Whether or not a person participates in physical activity is affected not only by individual lifestyles, but also by the environment in which a person lives or works [9, 10]. Generally, the determinants of physical activity are categorized into five factors according to the ecological model: individual, interpersonal, environmental, regional or national policy, and global factors [11]. Among these five factors, environmental factors include three dimensions: social environment, built environment, and natural environment. Ecological models stand on the basis that physical activity is conducted at the particular physical environments such as sports facilities, and these places which are designed for physical activity could have an influence on the choice of physical activity [12]. In fact, the accessibility of sports facilities (an environmental factor) has been reported in several studies to dramatically affect individual participation in physical activity [13–16]. It is believed that the presence of sports facilities is associated with participation in physical activity [14, 16–18]. Particularly, closer distances between an individual’s home and a sports facility are associated with high levels of physical activity [19]. Therefore, many countries have been investing in sports facilities over the recent decades to promote physical activity [20]. Many previous studies reporting the association of the accessibility of sports facilities with physical activity have focused solely on specific generational groups, such as adolescents, college students, and seniors [13, 21, 22]. However, considering the high depression rate in South Korea and the beneficial effects of physical activity on depression, it is necessary to determine the relationship between accessibility of sports facilities and physical activity among depressed individuals. To our knowledge, there is no study exploring the determinants of physical activity among people with history of depression, particularly in the context of the accessibility of sports facilities. Additionally, there are no studies evaluating the association of the accessibility of sports facilities with physical activity among people of diverse sociodemographic groups. Therefore, using nationally collected data, we analyzed the association between the accessibility of sports facilities and physical activity among individuals 20 years of age or older in order to investigate the association of sports facility accessibility with physical activity. Moreover, we conducted additional analyses to investigate the association between sports facility accessibility and physical activity according to history of depression and diverse sociodemographic groups based on age, income, occupation, and regional area. Methods Study participation We analyzed data from the Community Health Survey (CHS), which has been conducted by the Korean Centers for Disease Control and Prevention (KCDC) since 2008, to set and evaluate regional health plans and produce comparable regional health statistics by standardizing the survey system. We used data available from the 2012 CHS, which collected information from a total of 228,921 people aged 20 years or older. We excluded individuals with data missing for sports facility accessibility (n = 1756), physical activity (n = 806), history of depression (n = 55), BMI (n = 12,852), household income (n = 11,202), and other variables (n = 527); therefore, a final sample population of 201,723 people was selected for this study. The CHS received consent from study participants before the beginning of the study. Instruments and study processes used for the survey were approved by the KCDC Institutional Review Board (IRB #: 2012-07CON-01-2C). Study variables Sports facility accessibility To evaluate sports facility accessibility, we utilized responses to the CHS question “During the past year, was it easy to find sports facilities near your house?” Sports facilities consist of not only the place where sports equipment is available, but also the exercise environment. We classified the answers “easy to find” and “very easy to find” as easy and “difficult to find” and “very difficult to find” as difficult. Physical activity Physical activity was investigated using the CHS questionnaire data, which comprised three types of answers: vigorous, moderate, and walking. The questionnaire also requested the number of days of each activity per week (i.e., “How many days did you perform vigorous physical activity that made you feel tired or breathless during the past week?) and minutes of activity per day (i.e., “For how many minutes did you perform vigorous physical activity during the day?”). ‘Vigorous’ was defined as activity burning at least 7/kcal per minute, including activities such as jogging, running, climbing, football, baseball, intensive aerobic activity, swimming, squash, and work activities requiring running. ‘Moderate’ activity included yoga, badminton, volleyball, and work activities using both the arms and legs. Based on these definitions, we used the International Physical Activity Questionnaire short forms (IPAQ) to classify the level of physical activity engaged in by each person in this study [23]. The IPAQ suggests a metabolic equivalent task (MET) for each level of physical activity as follows:Vigorous MET-minutes/week = 8.0 * minutes of activity per day * days of activity per week Moderate MET-minutes/week = 4.0 * minutes of activity per day * days of activity per week Walking MET-minutes/week = 3.3 * minutes of activity per day * days of activity per week The IPAQ classified an individual’s activity as “Health-Enhancing Physical Activity (HEPA)” when the total score was 3000 MET-minutes (50 MET-hours)/week or more, “Active” when the total score was 600 MET-minutes (10 MET-hours) or more, and “Inactive” when the total score was below 600 MET-minutes (10 MET-hours). In this study, “HEPA” and “Active” were considered as participate, and “Inactive” was considered as non-participate [24]. Depression To identify people with a history of depression, we utilized the response to the CHS question “Have you ever been diagnosed as depressed by doctor?” Answering alternatives were binary (yes/no). Covariates We used the covariates of sex, age (under 40, 40 to 64, 65 or over), educational level (elementary school, middle school, high school, and college or higher), marital status (unmarried, married-cohabiting, married-not-cohabiting), and regional area (urban and rural) as sociodemographic variables, and monthly income (classified by quartile) and occupation (white collar, pink collar, and blue collar) as economic variables. Finally, health variables, such as the amount of sleep (less than 7 h, 7 to 8 h, and 9 h or more), self-rated stress, perceived health status (good, normal, and bad), perceived body shape (thin, normal, and obese), current drinker (yes and no), current smoker (yes and no), and history of depression (yes and no) were used as covariates. Obesity was measured by Body Mass Index (BMI: weight (kg)/height (m) 2; no: BMI <25, yes: BMI ≥25). All covariates were treated as categorical variables. Statistical analyses The purpose of this study was to analyze the factors affecting physical activity and the association of sports facility accessibility with physical activity. We performed statistical analyses of the survey data using SAS version 9.4 (SAS Institute, Cary, NC, USA). We first analyzed the distribution of each categorical variable described above to calculate the frequency and percentage of each variable and to identify significant differences between groups using the Chi-squared test. Next, we performed a multivariable logistic regression analysis to identify the relationship between sports facility accessibility and physical activity by controlling potential confounders including age, sex, monthly income, educational level, occupation, marital status, regional area, sleeping time, perceived stress rate, history of depression, perceived health status, current smoker, current drinker, perceived body shape, and obesity. Finally, we conducted subgroup analyses to investigate this association according to depression diagnosed experience, monthly house income, age, occupation, and regional area. The sampling weights were considered given that the CHS was a complex survey design. We produced adjusted odds ratios (ORs) with 95 % confidence intervals (95 % CIs). Results The general characteristics for the 201,723 individuals participating in this study are listed in Table 1. Of the 201,723 study participants, 80.9 % (n = 155,331) felt that sports facilities were easily accessible, and 19.1 % (n = 46,392) reported participating in some type of physical activity. Of the total cohort, 2.2 % (n = 4918) had been diagnosed with depression. The cohort was largely represented by those holding blue-collar jobs, with 26.3 % (n = 40,023), 14.2 % (n = 26,515), and 59.5 % (n = 135,185) of the total cohort holding white-collar, pink-collar, and blue-collar occupations, respectively.Table 1 General characteristics of the study participants Variables Total Physical activity No Yes N (%)* N (%)* N (%)* Accessibility to sports facilities  Easy 155,331 80.9 96,524 50.4 58,807 30.5  Difficult 46,392 19.1 29,965 12.8 16,427 6.3 Sex  Men 94,147 50.5 52,151 28.0 41,996 22.5  Women 107,576 49.5 74,338 35.2 33,238 14.4 Age group  Under 40 57,187 39.4 36,956 25.2 20,231 14.1  40–65 103,312 48.7 60,439 29.1 42,873 19.6  65 or over 41,224 11.9 29,094 8.9 12,130 3.1 Monthly income  Q1 (low) 53,233 16.1 35,515 11.2 17,718 4.9  Q2 50,304 24.4 31,829 15.9 18,475 8.5  Q3 55,050 31.2 33,869 19.5 21,181 11.7  Q4 (high) 43,136 28.2 25,276 16.5 17,860 11.7 Educational level  Elementary school 49,443 13.6 33,091 9.8 16,352 3.9  Middle school 23,493 9.1 14,502 5.9 8,991 3.2  High school 71,301 39.6 43,805 24.5 27,496 15.0  College or over 57,486 37.7 35,091 23.0 22,395 14.7 Occupation  White collar 40,023 26.3 24,794 16.2 15,229 10.1  Pink collar 26,515 14.2 16,497 8.8 10,018 5.4  Blue collar 135,185 59.5 85,198 38.1 49,987 21.4 Marital status  Married-cohabit 143,384 67.3 87,791 42.2 55,593 25.2  Married-non cohabit 28,414 10.7 20,312 7.7 8,102 3.0  Unmarried 29,925 22.0 18,386 13.3 11,539 8.7 Regional area  Urban 116,838 81.7 75,470 51.9 41,368 29.8  Rural 84,885 18.3 51,019 11.3 33,866 7.1 Sleeping time  Less (<7) 87,607 44.8 54,400 27.9 33,207 16.9  Normal (7–8) 105,494 51.3 65,891 32.5 39,603 18.8  Exceed (> = 9) 8,622 3.8 6,198 2.7 2,424 1.1 Perceived stress rate  Much 52,680 28.2 34,048 18.3 18,632 9.8  Little 109,873 55.7 67,814 34.7 42,059 21.0  Non 39,170 16.1 24,627 10.0 14,543 6.0 History of depression  Yes 4,918 2.2 3,383 1.5 1,535 0.7  No 196,805 97.8 123,106 61.6 73,699 36.2 Perceived health status  Good 80,996 44.2 46,660 25.4 34,336 18.7  Normal 82,126 41.9 52,291 27.5 29,835 14.4  Bad 38,601 14.0 27,538 10.2 11,063 3.7 Current smoker  Yes 43,473 24.0 25,032 14.0 18,441 10.0  No 158,250 76.0 101,457 49.1 56,793 26.9 Current drinker  Yes 135,821 74.6 81,146 45.1 54,675 29.5  No 65,902 25.4 45,343 18.1 20,559 7.4 Perceived body shape  Thin 36,884 17.1 23,969 11.2 12,915 5.8  Normal 93,003 45.0 57,094 27.9 35,909 17.1  Obese 71,836 38.0 45,426 24.1 26,410 13.9 Obesity  No (BMI <25) 153,502 76.0 97,354 48.7 56,148 27.3  Yes (BMI > =25) 48,221 24.0 29,135 14.5 19,086 9.5 201,723 100.0 126,489 63.1 75,234 36.9 *(%): Weighted percentage The odds ratios of factors associated with physical activity, and determined using a logistic regression analysis, are listed in Table 2. Those with easy access to sports facilities were 1.16 times more likely to participate in physical activity than those without easy access to sports facilities (OR = 1.16, 95 % CI 1.13–1.20). Less physical activity was also observed among low-income groups than among high-income groups. Those with white-collar occupations were less likely than those with blue-collar occupations to participate in physical activity. Those with less than 7 h of sleep per night were about as likely to participate in physical activity as those with 7–8 h of sleep per night, while those sleeping for 9 h or more each night were less likely to exercise. Finally, those with lower perceived health (normal or bad) exercised less than those with better-perceived health.Table 2 Factors associated with physical activity* Variables Physical activity Adjusted OR 95 % CI Accessibility to sports facilities  Easy 1.16 (1.13–1.20)  Difficult 1.00 - Sex  Men 1.91 (1.85–1.97)  Women 1.00 - Age group  Under 40 1.00 -  40–65 1.42 (1.38–1.47)  65 or over 0.93 (0.88–0.98) Monthly income  Q1 (low) 0.78 (0.75–0.82)  Q2 0.80 (0.77–0.83)  Q3 0.86 (0.83–0.89)  Q4 (high) 1.00 - Educational level  Elementary school 0.93 (0.88–0.97)  Middle school 0.93 (0.88–0.97)  High school 0.93 (0.90–0.96)  College or over 1.00 - Occupation  White collar 1.00 -  Pink collar 1.15 (1.10–1.21)  Blue collar 1.18 (1.14–1.22) Marital status  Married-cohabit 1.00 -  Married-non cohabit 0.92 (0.88–0.96)  Unmarried 1.21 (1.16–1.25) Regional area  Urban 1.00 -  Rural 1.23 - Sleeping time  Less (<7) 1.07 (1.05–1.10)  Normal (7–8) 1.00 -  Exceed (> = 9) 0.77 (0.72–0.82) Perceived stress rate  Much 0.94 (0.90–0.98)  Little 0.98 (0.94–1.01)  Non 1.00 - History of depression  Yes 1.14 (1.04–1.24)  No 1.00 - Perceived health status  Good 1.00 -  Normal 0.76 (0.74–0.78)  Bad 0.62 (0.59–0.65) Current smoker  Yes 0.87 (0.84–0.90)  No 1.00 - Current drinker  Yes 1.27 (1.23–1.31)  No 1.00 - Perceived body shape  Thin 1.00 -  Normal 1.16 (1.12–1.20)  Obese 1.10 (1.06–1.15) Obesity  No (BMI <25) 1.00 -  Yes (BMI > =25) 1.06 (1.02–1.10) *The result of multivariable logistic regression to investigate the association between accessibility to sports facilities and physical activity controlling for sex, age, monthly income, educational level, occupation, marital status, regional area, sleeping time, perceived stress rate, history of depression, perceived health status, current smoker, current drinker, perceived body shape and obesity The association between physical activity and access to sports facilities stratified by history of depression, monthly house income, age, occupation and regional area is shown in Table 3. The subgroup analysis showed significant differences in each group, although modifying effects were not significant in the tests for interaction except for occupation and regional area. Those who had experience of depression showed a trend towards a greater magnitude of physical activity if participants felt that they had easy access to sports facilities. There was also a trend towards a greater magnitude of physical activity if participants lived in an urban area (OR = 1.21, 95 % CI 1.16–1.26).Table 3 Subgroup analysis of physical activity with accessibility to sports facilities stratified by history of depression, monthly income, age, occupation and regional area* Variables Physical activity Difficult Easy OR OR 95 % CI History of depression  Yes 1.00 1.46 (1.21–1.76)  No 1.00 1.16 (1.12–1.19) Monthly income  Q1 (low) 1.00 1.16 (1.09–1.23)  Q2 1.00 1.17 (1.10–1.24)  Q3 1.00 1.18 (1.11–1.25)  Q4 (high) 1.00 1.14 (1.07–1.23) Age group  Under 40 1.00 1.19 (1.13–1.26)  40–65 1.00 1.14 (1.09–1.19)  65 or over 1.00 1.15 (1.08–1.24) Occupation  White collar 1.00 1.24 (1.15–1.33)  Pink collar 1.00 1.14 (1.09–1.19)  Blue collar 1.00 1.15 (1.07–1.23) Regional area  Urban 1.00 1.21 (1.16–1.26)  Rural 1.00 1.06 (1.02–1.11) *The result of subgroup analyses by using multivariable logistic regression to investigate the association between accessibility to sports facilities and physical activity controlling for sex, age, monthly income, educational level, occupation, marital status, regional area, sleeping time, perceived stress rate, history of depression, perceived health status, current smoker, current drinker, perceived body shape and obesity Discussion In this study, we analyzed the factors associated with physical activity among people aged 20 or over, focusing on the population’s access to sports facilities. Our observations show that sports facility accessibility is considerably associated with the amount of physical activity an individual participates in. Previous work suggested that the distance required of an individual to travel to a sports facility affects sports facility usage [14, 17, 18], as in the present study. This association may be explained by the fact that environmental factors affect an individual’s perception of physical activity [25]. Long distances may reduce the motivation to do physical activity [26], as inability to access appropriate facilities is reported to likely act as perceived motivational barrier [27]. Therefore, easy access to sports facilities may act as a motivator to encourage an individual to participate in physical activity [28]. For example, an individual living near a sports facility can easily access information about exercise, while those living further away from sports facilities cannot easily access this information [25]. Another study reported similar results, clarifying that an individual’s physical environment should be treated as a subsidiary determinant factor because it does not affect the frequency of physical activity as much as other factors including social support [29]. In addition, our subgroup analysis indicated that history of depression, monthly household income, age, occupation, and regional area potentially affect the association between access to sports facilities and physical activity, although the modifying effects were not significant. The results of the present study showed that easy access to sports facilities among those who had a history of depression tended to result in more physical activity than when such facilities were less accessible due to distance. Generally, depressed people have insufficient motivation to maintain an active lifestyle [30]. However, taking into account the various health and well-being benefits proffered by physical activity, it is important to encourage depressed individuals to exercise. According to a previous study, there is a clear association between autonomous types of motivation and physical activity [31]; therefore, proximity to sports facilities may encourage depressed individuals to spontaneously participate in physical activity by increasing opportunities to acquire information about exercise and access to sports equipment. The proximity of sports facilities was associated with physical activity regardless of monthly household income, age, and occupation. Regardless of monthly household income, easier access to sports facilities was associated with increased physical activity; therefore, although low income populations are usually at risk for physical inactivity [32], access to sports facilities appears to more strongly affect physical activity than does income, with easy access to sports facilities promoting physical activity even among those at the highest risk for inactivity [33]. We also observed an association between an individual’s occupation and physical activity. Without taking sports facility accessibility into account, those classified as white-collar workers exercise less than blue-collar workers. Because white-collar jobs are typically sedentary, the amount of physical activity experienced on the job by the white-collar worker is less than that experienced by the blue-collar worker, who typically does some type of physical activity while working [34, 35]. However, according to this study, when access to sports facilities was higher, white-collar workers showed higher rate of physical activity. A previous study suggested that promoting physical activity in the workplace, such as facilitating access to sports facilities, increases exercise among workers [36], reducing barriers to physical activity and promoting physical activity regardless of the risk for inactivity. The current findings should be interpreted with a degree of caution due to several limitations. First, sports facility accessibility was assessed via a single self-reported questionnaire. Therefore, we could only acquire the perceived, rather than actual, distance between an individual’s home and sports facilities. Moreover, there was the possibility of same-source bias, as the exposure and outcome variables were self-reported. Additionally, objective accessibility could not be estimated in this study. It was also difficult to identify whether participants were depressed at the time that they took the survey, as the survey only asked whether a person had ever been diagnosed with depression and did not specify whether the diagnoses was current. Moreover, this study could only be generalizable to South Korean adults. In addition, there was potential confounding influence from other area-level factors (i.e., the proximity to a sports club may be associated with other area-level factors that are associated with physical activity). Furthermore, the measurement of physical activity might have been inaccurate due to the validity of the IPAQ, which is known to overestimate physical activity relative to objectively measured data in most populations [37]. Therefore, the physical activity levels used in this study might have been overestimated. Finally, it should be noted that we used cross-sectional data; therefore, we could not exclude a bi-directional effect or relationship opposite to what was hypothesized (i.e., those who are more physically active self-select into areas with a higher density of sports clubs). Despite these limitations, this study also has several strengths. First, compared to previous studies that set a limit on particular generations, this study utilized data from adults aged 20 years to 65+ and included a cohort of those with a history of depression diagnosis. To our knowledge, this is the first attempt to investigate the association between sports facility accessibility and physical activity among those with a history of depression. Secondly, considering that perceived barriers to health-promoting behaviors are an important component of major health behavior theories [24], this study provides useful data by implementing individuals’ perceived level of access to sports facilities. Finally, we used national sample data, suggesting that our data can be widely generalized. Our results suggest that an individual’s perceived level of access to sports facilities may play an important role in physical activity of not only the general population, but also of those with a history of depression. Therefore, it is crucial to take into account sports facility accessibility when building physical activity-promoting environments or designing programs for enhancing physical activity. Additionally, this study provides a basis for future research on treating depression through physical activity. Based on these observations, further studies should investigate the association between physical activity and sports facility accessibility among cohorts with other conditions that improve with physical activity. Future studies should also use more concrete methods for investigating sports facility accessibility, such as by using both a geographic information system for investigating objective accessibility and several questionnaires rather than only one questionnaire for investigating subjective accessibility. Conclusion Our results showed that the accessibility of sports facilities is associated with physical activity. Therefore, it is crucial to consider the accessibility of sports facilities when promoting an environment conducive to physical activity or designing programs for enhancing physical activity. Abbreviations CHSCommunity health survey HEPAHealth-enhancing physical activity IPAQInternational physical activity questionnaire KCDCKorean centers for disease control and prevention METMetabolic equivalent task Acknowledgements The Korean Centers for Disease Control (KCDC) produced and provided national level data. A number of colleagues from the Department of Public Health, Graduate School of Yonsei University provided advice for this manuscript. Funding No funding was received for this study. Availability of data and materials The CHS was openly available in https://chs.cdc.go.kr/chs/index.do by submitting written oath and data utilization plan. Authors’ contributions SAL designed the study and directed its implementation. ISH, JEL participated in the design of the study and performed the statistical analysis. ECP helped supervise the whole process. JYN helped the literature review. YJJ, KTH helped revise the manuscript and performed statistical analysis. 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 The CHS received consent from study participants before the beginning of the study. Instruments and study processes used for the survey were approved by the KCDC Institutional Review Board (IRB #: 2012-07CON-01-2C). ==== Refs References 1. Wen CP Minimum amount of physical activity for reduced mortality and extended life expectancy: a prospective cohort study Lancet 2011 378 9798 1244 53 10.1016/S0140-6736(11)60749-6 21846575 2. Mammen G Faulkner G Physical activity and the prevention of depression: a systematic review of prospective studies Am J Prev Med 2013 45 5 649 57 10.1016/j.amepre.2013.08.001 24139780 3. Gauvin L Spence JC Physical activity and psychological well-being: knowledge base, current issues, and caveats Nutr Rev 1996 54 4 Pt 2 S53 65 8700454 4. Katzmarzyk PT Mason C The physical activity transition J Phys Act Health 2009 6 3 269 80 10.1123/jpah.6.3.269 19564654 5. Geronimus AT To mitigate, resist, or undo: addressing structural influences on the health of urban populations Am J Public Health 2000 90 6 867 72 10.2105/AJPH.90.6.867 10846503 6. Scheers T Philippaerts R Lefevre J Compliance with different physical activity recommendations and its association with socio-demographic characteristics using an objective measure BMC Public Health 2013 13 1 136 10.1186/1471-2458-13-136 23409982 7. Scott D Happell B The high prevalence of poor physical health and unhealthy lifestyle behaviours in individuals with severe mental illness Issues Ment Health Nurs 2011 32 9 589 97 10.3109/01612840.2011.569846 21859410 8. Cho M The epidemiological survey of mental disorders in Korea 2011 Seoul Ministry of Health and Welfare 9. Ferdinand AO The relationship between built environments and physical activity: a systematic review Am J Public Health 2012 102 10 e7 13 10.2105/AJPH.2012.300740 22897546 10. Sallis JF An ecological approach to creating active living communities Annu Rev Public Health 2006 27 297 322 10.1146/annurev.publhealth.27.021405.102100 16533119 11. Bauman AE Correlates of physical activity: why are some people physically active and others not? Lancet 2012 380 9838 258 71 10.1016/S0140-6736(12)60735-1 22818938 12. Sallis J Bauman A Pratt M Environmental and policy interventions to promote physical activity Am J Prev Med 1998 15 4 379 97 10.1016/S0749-3797(98)00076-2 9838979 13. Kim Y Kosma M Psychosocial and Environmental Correlates of Physical Activity Among Korean Older Adults Res Aging 2012 35 6 750 67 10.1177/0164027512462412 14. Limstrand T Rehrer NJ Young people’s use of sports facilities: A Norwegian study on physical activity Scand J Public Health 2008 36 5 452 9 10.1177/1403494807088455 18635728 15. Troped PJ Correlates of recreational and transportation physical activity among adults in a New England community Prev Med 2003 37 4 304 10 10.1016/S0091-7435(03)00137-3 14507486 16. Eriksson U Arvidsson D Sundquist K Availability of exercise facilities and physical activity in 2,037 adults: cross-sectional results from the Swedish neighborhood and physical activity (SNAP) study BMC Public Health 2012 12 1 1 10.1186/1471-2458-12-607 22214479 17. Ranchod YK Longitudinal associations between neighborhood recreational facilities and change in recreational physical activity in the multi-ethnic study of atherosclerosis, 2000–2007 Am J Epidemiol 2014 179 3 335 43 10.1093/aje/kwt263 24227016 18. Halonen JI Is change in availability of sports facilities associated with change in physical activity? A prospective cohort study Prev Med 2015 73 10 4 10.1016/j.ypmed.2015.01.012 25602907 19. Addy CL Associations of perceived social and physical environmental supports with physical activity and walking behavior Am J Public Health 2004 94 3 440 3 10.2105/AJPH.94.3.440 14998810 20. Ståhl T The importance of policy orientation and environment on physical activity participation—a comparative analysis between Eastern Germany, Western Germany and Finland Health Promot Int 2002 17 3 235 46 10.1093/heapro/17.3.235 12147638 21. Sallis JF Prochaska JJ Taylor WC A review of correlates of physical activity of children and adolescents Med Sci Sports Exerc 2000 32 5 963 75 10.1097/00005768-200005000-00014 10795788 22. Davison KK Lawson CT Do attributes in the physical environment influence children’s physical activity? A review of the literature Int J Behav Nutr Phys Act 2006 3 1 19 10.1186/1479-5868-3-19 16872543 23. IPAQ Research Committee. Guidelines for data processing and analysis of the International Physical Activity Questionnaire (IPAQ)–short and long forms, revised November, 2005. 2005. http://www.ipaq.ki.se. Accessed 24 Aug 2016.  24. Cerin E Perceived barriers to leisure-time physical activity in adults: an ecological perspective J Phys Act Health 2010 7 4 451 9 10.1123/jpah.7.4.451 20683086 25. Kim YH Park IK Lee HG Relationship between Social Ecological Variables and Physical Activity Korean Soc Sport Psychol 2013 24 4 61 74 10.14385/KSSP.24.4.61 26. Davison KK Jago R Change in parent and peer support across ages 9 to 15 yr and adolescent girls’ physical activity Med Sci Sports Exerc 2009 41 9 1816 25 10.1249/MSS.0b013e3181a278e2 19657287 27. Burton NW Turrell G Oldenburg B Participation in recreational physical activity: why do socioeconomic groups differ? Health Educ Behav 2003 30 2 225 44 10.1177/1090198102251036 12693525 28. Thornton LE Pearce JR Kavanagh AM Using Geographic Information Systems (GIS) to assess the role of the built environment in influencing obesity: a glossary Int J Behav Nutr Phys Act 2011 8 71 10.1186/1479-5868-8-71 21722367 29. Giles-Corti B Donovan RJ The relative influence of individual, social and physical environment determinants of physical activity Soc Sci Med 2002 54 12 1793 812 10.1016/S0277-9536(01)00150-2 12113436 30. Soundy A Barriers to and facilitators of physical activity among persons with schizophrenia: a survey of physical therapists Psychiatr Serv 2014 65 5 693 6 10.1176/appi.ps.201300276 24585134 31. Vancampfort D Adopting and maintaining physical activity behaviours in people with severe mental illness: The importance of autonomous motivation Prev Med 2015 81 216 20 10.1016/j.ypmed.2015.09.006 26386141 32. United States Department of Health Human Services. Physical activity and health: A report of the Surgeon General. Diane Publishing; 2006.  33. Brownson RC Environmental and policy determinants of physical activity in the United States Am J Public Health 2001 91 12 1995 2003 10.2105/AJPH.91.12.1995 11726382 34. Seo KM Analysis of physical activity in male office workers J Korean Community Nurs 2003 14 1 95 105 35. Steele R Mummery K Occupational physical activity across occupational categories J Sci Med Sport 2003 6 4 398 407 10.1016/S1440-2440(03)80266-9 14723390 36. Crump CE Effect of organization-level variables on differential employee participation in 10 federal worksite health promotion programs Health Educ Q 1996 23 2 204 23 10.1177/109019819602300206 8744873 37. Lee PH Validity of the international physical activity questionnaire short form (IPAQ-SF): A systematic review Int J Behav Nutr Phys Act 2011 8 1 115 10.1186/1479-5868-8-115 22018588

PMC005xxxxxx/PMC5002152.txt

==== Front Virol JVirol. JVirology Journal1743-422XBioMed Central London 60210.1186/s12985-016-0602-7Short ReportEpstein-Barr viral microRNAs target caspase 3 Harold Cecelia cecelia.harold@einstein.yu.edu 12Cox Diana diana.cox@bcm.edu 13Riley Kasandra J. kriley@rollins.edu 11 Department of Chemistry, Rollins College, Winter Park, FL 32789 USA 2 Present Address: Albert Einstein College of Medicine, Bronx, NY 10461 USA 3 Present Address: Baylor College of Medicine, Houston, TX 77030 USA 26 8 2016 26 8 2016 2016 13 1 1458 6 2016 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.The Epstein-Barr virus (EBV) is a ubiquitous herpesvirus that transforms B cells and causes several malignancies including Burkitt’s lymphoma. EBV differentially expresses at least 49 mature microRNAs (miRNAs) during latency in various infected epithelial and B cells. Recent high-throughput studies and functional assays have begun to reveal the function of the EBV miRNAs suggesting roles in latency, cell cycle control, and apoptosis. In particular, the central executioner of apoptosis, Caspase 3 (CASP3), was proposed as a target of select EBV miRNAs. However, whether CASP3 is truly a target of EBV miRNAs, and if so, which specific miRNAs target CASP3 is still under debate. Based on previously published high-throughput biochemical data and a bioinformatic analysis of the entire CASP3 3′-UTR, we identified 12 EBV miRNAs that have one or more seed binding sites in the CASP3 3′-UTR. We individually tested all 12 miRNAs for repression of CASP3 in luciferase reporter assays, and nine showed statistically significant (P < 0.001) repression of a full-length CASP3 reporter. Further, three EBV miRNAs, including BART22, exhibited repression of endogenous CASP3 protein. These data confirm that CASP3 is a direct target of specific EBV BART miRNAs. Electronic supplementary material The online version of this article (doi:10.1186/s12985-016-0602-7) contains supplementary material, which is available to authorized users. Keywords Epstein-Barr virusmicroRNACaspase 3Burkitt’s lymphomaRollins Collegenoneissue-copyright-statement© The Author(s) 2016 ==== Body Background Epstein-Barr virus (EBV; human herpesvirus 4) persists in more than 95 % of the adult human population. Initial EBV infection is either asymptomatic or manifests as self-limiting mononucleosis. However, in a subset of cases, EBV is also associated with the oncogenic transformation of cells, resulting in malignancies including Burkitt’s lymphoma (BL) and several epithelial cell cancers (reviewed in [1]). After more than 50 years of study, the role of EBV in cancer is still under investigation. In addition to ~80 proteins and two ~170 nucleotide (nt) noncoding RNAs, EBV differentially expresses at least 49 mature microRNAs (miRNAs) from 25 precursors [2–4]. MiRNAs are ~22 nt noncoding RNAs that direct the modest downregulation of specific human and viral transcripts in diverse multicellular organisms. EBV miRNA expression varies widely in different cell types, latency stages, and strains of the virus [5, 6]. Select human and EBV miRNAs are thought to contribute to the cancer phenotype in both lymphoid and epithelial tumor types and are therefore considered possible therapeutic targets [7–12]. EBV precursor miRNAs arise from two genomic clusters: three surrounding the BHRF1 gene (BHRF1-1, 1-2, and 1-3), and 22 within the BART transcripts (BamHI A rightward transcripts; BARTs1-22) [2]. The BARTs are numbered in order of discovery and not in order of their appearance in the EBV genome [2–4]. Most human and viral miRNAs undergo a standard biogenesis pathway beginning with RNA polymerase II transcription of the primary-miRNA. In two steps, the first in the nucleus and the second in the cytoplasm, RNAse III enzymes sequentially process the primary-miRNA into the intermediate precursor-miRNA hairpin and final mature miRNA. One of the two strands is bound to core protein Argonaute and a target mRNA within the RNA-induced Silencing Complex (RISC). The single-stranded mature miRNA directs RISC to specific mRNA targets, typically via limited base-pairing to the mRNA 3′-untranslated region (3′-UTR) [13]. Key to downregulation is the base pairing interaction between the mRNA and 6-8 nt of the miRNA. This “seed region,” defined minimally as base pairing including nts numbered 2-7 from the 5′ end of the miRNA ([14]; Table 1), is crucial to the contacts between the miRNA, mRNA, and Argonaute protein [15]. Different patterns of base pairing can yield different degrees of repression; multiple sites are additive or synergistic, depending on their orientation [16]. Because the function of a given miRNA is determined by the mRNAs it targets, we and others conducted global biochemical and bioinformatic analyses predicting interactions between EBV and human miRNAs and human transcripts [11, 17, 18]. These studies provided thousands of predictions for miRNA-mRNA interactions, but further biochemical experimentation is required to confirm bona fide interactions between EBV miRNAs and human transcripts.Table 1 Tested interactions between EBV miRNAs and the full-length CASP3 3’UTR Human Caspase 3 3’UTR (chr4:184627696-184629271; GRCh38/hg38 assembly) was searched for seed base pairing (nt 2-7 highlighted in gray, A1 position underlined), categorized by standard definitions [27]. “Non-canon” refers to the non-canonical site proposed by Veriede et al. [25]. All miRNA sequences are from MirBASE v. 21 [28]. Mutated sites in reporters are noted in bold lowercase letters Three types of studies have demonstrated the important and complex role of EBV miRNAs in apoptosis. Phenotypic experiments implicated the BHRF1 miRNAs in apoptotic inhibition and promotion of the cell cycle during the early phase of infection of human primary B cells [12, 19]. Expression of clusters of BART miRNAs [20] or select BART miRNAs alone (BARTs 3, 6, 8, 16, and 22; [11]) is sufficient to exert an anti-apoptotic phenotype in AGS gastric carcinoma cells. Two different high-throughput studies of EBV miRNA targeting showed statistically significant enrichment of transcripts associated with apoptosis in B cells infected with different strains of EBV [17, 18]. In our published study of Jijoye cells [18], a BL cell line that expresses all known EBV miRNAs, 1664 human 3′-UTRs were identified as targets of the 12 most abundant EBV miRNAs, and approximately eight percent of these (132 transcripts) were associated in some way with apoptosis. Direct targets of EBV miRNAs were identified using standard reporter assays to validate base pairing between one or more EBV miRNAs and mRNA targets, including pro-apoptotic BBC3/PUMA [21], BCL2L11/BIM [20], BCL2L8/BAD [22], and anti-apoptotic BID [8]. Thus, the BART miRNAs directly downregulate both pro-and anti-apoptotic mediators that may vary by cell type. Perhaps the most compelling mechanism for EBV miRNA inhibition of apoptosis is the potential repression of Caspase 3 (CASP3), the central proteolytic executioner of apoptosis [23], first identified as a potential direct EBV miRNA target in our high-throughput study [18]. Multiple groups have attempted to illuminate the role of EBV miRNAs in the regulation of CASP3. In an indirect assay for CASP3 regulation by EBV miRNAs, B cells infected by EBV lacking BART miRNAs expressed higher CASP3 and were more resistant to mitochondria-mediated apoptosis-inducing drugs relative to cells infected with intact EBV [24]. In a direct assay for miRNA repression, slight (~20 %) repression of a CASP3-luciferase reporter was observed after transfection of synthetic BARTs 1-3p and 16. Further, CASP3 protein levels decreased upon ectopic BART expression and increased upon depletion of EBV from S1-1 BL cells [25]. However, significant repression of a CASP3-luciferase reporter was not observed for BARTs 3, 6, 8, 16, or 22, when each was expressed from a lentiviral vector transfected into HEK293T cells, in spite of seed sequence sites in the CASP3 3′-UTR for each of these miRNAs [11]. Differences in experimental approach have likely led to these inconsistent results. Regardless, other groups have found that CASP3 protein is consistently lower in BL and epithelial cells expressing clusters of the BART miRNAs relative to cells without BARTs [20, 24, 25]. An individual study of each EBV miRNA with a potential binding site in CASP3 has not yet been undertaken, so it is unclear if single EBV miRNAs directly target CASP3. If single miRNAs are used as therapeutic targets, it is important to identify the unique targets/function of each. We employed systematic testing with standard reporter assays and Western blots to address our initial observation that CASP3 is a target of multiple EBV BART miRNAs [18] and clarify the role of EBV miRNAs in the regulation of CASP3. Nine EBV miRNAs bind to 13 specific locations on the CASP3 3′-UTR; three of the nine miRNAs detectably repress endogenous CASP3 protein independently. This is a critical step in detailing the mechanism by which EBV regulates apoptosis and could provide clues for future rational drug design. Methods Bioinformatic analyses Our previously published crosslinking and immunoprecipitation/high-throughput sequencing (HITS-CLIP) experiment identified CASP3 as a putative target of one or more EBV miRNAs, but we had not examined the CASP3 3′-UTR in single nucleotide resolution [18]. High-confidence HITS-CLIP sequence fragments were mapped onto the human genome (hg 18) as in [18], and the UCSC Genome Browser (https://genome.ucsc.edu) was used to generate an image from our BED file to overlay with the 3′-UTR of CASP3 (Fig. 1). The CASP3 3′-UTR was searched by hand for the reverse complement of canonical seed binding sites (7mer-A1, 7mer-m8, and 8mer) for all 49 known EBV miRNAs, and genomic coordinates of potential EBV miRNA binding sites were confirmed using the BLAST-Like Alignment Tool (BLAT) on the UCSC Genome Browser (Fig. 1, Table 1). All miRNA-mRNA pairs that met the minimal energetic requirements for base pairing were visualized using RNA22 [26]. Sites of potential human miRNA seed binding were identified by TargetScanHuman 7.0 [27].Fig. 1 Caspase 3 has 16 predicted sites for EBV miRNA base pairing, 15 of which align with HITS-CLIP data. HITS-CLIP high-throughput sequencing reads (0 or 1 mismatch; ≥25 nt long) from six biological replicates of Ago-bound RNAs in Jijoye BL cells (unique reads, one color per biological replicate) mapped to scale on the Caspase 3 3′-UTR (1576 nt; RefSeq ID: NM_004346.3). Predicted sites of miRNA seed sequence binding for EBV miRNAs (black) and human miRNAs (purple) are noted below, with non-verifiable EBV miRNA sites in gray. When two sites for a given miRNA are present, they are labeled “A” and “B” from 5′ to 3′, respectively. Sites for BARTs 4, 7, and 22B met our initial HITS-CLIP reproducibility thresholds (three or more experimental replicates, peak height of five or more; [18]) Constructs The full length 3′-UTR of CASP3 (hg18, Chr4:185785817-185787419) was PCR amplified from genomic DNA and inserted downstream of firefly luciferase between NheI and XhoI in the pmiRGLO dual luciferase vector (Promega; Additional file 1: Table S1). Mutants were generated by site-directed mutagenesis (Stratagene) at the positions indicated in Table 1 with primers provided in Additional file 1: Table S1. All plasmids were verified by sequencing (Eurofins Genomics). Synthetic miRNAs (IDT) were synthesized according to the mature miRNA sequences in miRBase 21 [28] with a 5′-phosphate on the functional strand only, a mismatch at nt 2, and 2-nt 3′ overhangs, as in previous studies [14, 18, 29]. Pairs were annealed and confirmed on 3 % agarose gels according to [30]. The control miRNA “CTL” (5′-CTAGTATGACTAGTATGATCCGG top strand) has been used in previous studies [18]. CTL is a scrambled sequence with the same proportion of each nucleotide as hsa-miR-17, and it has no known biological function, seed sequence similarity to a known viral/human miRNA, or canonical targeting of CASP3. Luciferase assays HEK293T cells (ATCC) in 24-well plates were co-transfected with 40 nM synthetic miRNA duplex and 2.5 ng pmiRGLO reporter plasmid using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Cells were lysed 24 h post-transfection, and the Dual-Luciferase Assay Reporter System (Promega) manufacturer’s protocol was performed with the GloMax Junior Luminometer (Promega). Firefly luciferase activity was normalized to the control Renilla luciferase activity in each well, and this ratio was then normalized to the control construct appropriate for each experiment. At least four independent transfections for each condition were measured. Excel was used to calculate the standard deviation for each data point, and two-tailed Student’s t-tests were used to compare samples as noted. Western blots HEK293T cells were transfected in triplicate with the denoted synthetic miRNAs (40 nM) in the figures using Lipofectamine 2000 (Invitrogen). Protein lysates were prepared from cells, and Western blot analyses were performed as described (Wade and Allday, 2000) using NuPAGE 4-12 % gels and associated buffers (Invitrogen). Primary antibodies were anti-Caspase 3 (8G10; #9665) and anti-alpha-tubulin #2125 (Cell Signaling Technology). Western blots were quantified using ImageQuant software, with CASP3 levels normalized to the tubulin loading controls and these values normalized to the control transfection condition. Results Argonaute HITS-CLIP experiments yield genome-wide maps of likely Argonaute and miRNA binding. Our previous HITS-CLIP data revealed potential interactions between EBV miRNAs and 132 host apoptotic mRNAs, including CASP3 [18]. We examined CASP3 because of its central role in apoptosis and the high reproducibility of the HITS-CLIP sequencing tags—sites of Argonaute binding—that overlapped with predicted seed binding sites for BARTs 4, 7, and 22 (Fig. 1). Because our HITS-CLIP experiment was performed in BL cells expressing all EBV miRNAs to varying degrees, we conducted a bioinformatic search for binding sites for all 49 EBV miRNAs and uncovered six more EBV miRNAs with one or more binding sites as well as a possible site for BART20 that did not overlap with our HITS-CLIP Argonaute binding sites (Fig. 1). In total, there are 14 putative miRNA binding sites for nine EBV miRNAs on the CASP3 3′-UTR. We also included two non-canonical sites for comparison: the previously published BART16 [25], which has a wobble base pair in the seed, and BART10, which has a 6mer site predicted in CASP3 (Fig. 1, Table 1). In order to validate the bioinformatic and HITS-CLIP predicted interactions, the full-length 3′-UTR of CASP3 was cloned into a dual-luciferase reporter vector and co-transfected into HEK293T cells with individual synthetic miRNAs. Significant repression of the full-length CASP3 3′-UTR reporter occurred after transfection of synthetic BART 4, 3, 13, 8, 7, 18, 1, 2, or 22 (Fig. 2a). BART22 incited the greatest repression, and BARTs 16, 10, and 20 did not detectably repress CASP3.Fig. 2 EBV miRNAs repress a full-length luciferase-Caspase 3 3′-UTR reporter. HEK293T cells were co-transfected with the designated luciferase-Caspase 3 reporters and either total synthetic control miRNA duplex (CTL) or an EBV miRNA predicted to base pair with the Caspase 3′-UTR (Table 1 includes all WT and mutant sequences). Firefly/Renilla luciferase ratios were normalized to the same reporter transfected with the negative control miRNA (CTL). a Wild type (WT) Caspase 3 is repressed significantly by nine EBV miRNAs. b WT repression is compared to loss of repression of the relevant mutant for each of the miRNAs that showed the greatest WT repression. In all luciferase assays, mean values were from at least four independent transfections. Error bars, standard deviation; P values from two-tailed Student’s t-tests of noted pairs, *P < 0.0001, **P < 0.01 The specificity of downregulation by CASP3 at the each of the most statistically significant and repressed (>20 %) binding sites was further confirmed by comparing wild-type 3′-UTR repression to the repression of a corresponding reporter with 2-3 point mutations in each seed binding region (Fig. 2b, Table 1). Three of the miRNAs that showed significant repression each had two possible sites (Fig. 1), so we individually mutated each site. The HITS-CLIP data show far less Argonaute binding at the BART8A, BART2A, and BART22A sites relative to their partner B sites (Fig. 1). In the case of BART2, the “B” site appeared to be more functional than the A. Mutation of the BART 22 “A” site alone was sufficient to de-repress the reporter almost completely (Fig. 2b), but mutation of the “B” site also showed significant derepression, suggesting that these two sites may work synergistically even though they are 312 nt apart (Table 1), because loss of either single site eliminates repression. In contrast, the weak BART8 repression via two sites appears to be roughly equal and additive (Fig. 2b). In total, our luciferase assay validates direct, reproducible, varying repression of the CASP3 3′-UTR reporter by BARTs 4, 3, 13, 8, 7, 1, 2, and 22. Given that these miRNAs significantly repress protein made from a reporter construct, we hypothesized that we might also see an effect on endogenous levels of CASP3 protein. Thus, we transfected individual synthetic EBV miRNAs into HEK293T cells, which lack endogenous EBV miRNAs, to confirm downregulation of endogenous CASP3 protein by Western blotting. CASP3 was downregulated most significantly by BART22 but also weakly by BARTs 8, 7, 1, and 2 (Fig. 3).Fig. 3 Select EBV miRNAs repress endogenous Caspase 3 protein. HEK293T cells were transfected with the denoted control or EBV miRNA. Western blots of extracts prepared 24 h post-transfection were probed for endogenous proteins with anti-Caspase 3 or anti-tubulin antibodies. Normalized CASP3 levels from triplicate experiments are reported below with the S.E.M Discussion In the time since the publication of our high-throughput experiment that proposed CASP3 as a target of EBV miRNAs, several groups have published conflicting evidence of this interaction [11, 18, 22, 24, 25]. Here we confirmed that human CASP3 is a direct target of nine EBV BART miRNAs in luciferase assays by testing all canonical, predicted sites. Three of these miRNAs detectably, independently repressed endogenous CASP3 in HEK293T cells. To complement our reporter assays, we used high levels of individual synthetic EBV miRNAs to demonstrate detectable repression of endogenous CASP3 protein on the part of select EBV miRNAs (Fig. 3). The levels of miRNA used in this proof-of-principle experiment are probably not physiologically attainable, but others have documented the effect of EBV miRNA expression on CASP3 protein levels and downstream effects on its substrate PARP (a measure of CASP3 activity) in both B cells and epithelial cells that are models for EBV infection. For example, CASP3 and cleaved PARP increase when AGS epithelial cells are treated with an inhibitor of BART20-5p. While BART20-5p is not likely a direct repressor of the CASP3 transcript, it represses BAD, a pro-apoptotic protein upstream of CASP3 [22]. Similarly, in EBV-infected B cells, CASP3 protein increases when all of the BARTs are deleted from the M81 strain [24]. Because the BART miRNAs target so many apoptotic transcripts [11, 17, 18], the effect of each individual EBV miRNA on CASP3 protein levels may vary. Thus, the luciferase assay is perhaps the best way to probe direct targeting of a given transcript. The dual luciferase reporter assay undertaken in a cell line lacking relevant endogenous miRNAs remains the gold standard in the validation of direct targeting by miRNAs. Kang et al. (2015) recently conducted a reporter assay experiment similar to ours but only tested five miRNAs that had exerted an anti-apoptotic phenotype when expressed alone in their epithelial (gastric carcinoma) cell model (BARTs 3, 6, 8, 16, 22). The authors concluded that none of these miRNAs significantly repressed their CASP3 luciferase reporter in HEK293T cells. This conclusion may be explained by two key differences in our experimental designs: (1) the differing identity of our negative controls, and (2) the origin of our miRNAs. Cognizant that ectopic expression/transfection of both miRNAs and luciferase reporters skews the natural stoichiometry within a cell [31], and synthetic miRNAs accumulate well above physiologically attainable levels [32], all of our data are normalized to repression by a equal amounts of a transfected synthetic miRNA that does not repress our target of interest rather than a “mock” or “empty vector” transfection. Our synthetic miRNA duplexes may have accumulated to higher levels relative to the reporter than the miRNAs expressed from transduced lentiviral vectors employed by Kang et al, since those miRNAs presumably required standard biogenesis processing. The BART miRNAs are expressed together in large clusters and likely operate in tandem [3, 19, 33]. Multiple miRNA binding sites in a single 3′-UTR are additive, so major targets of EBV miRNAs would have several miRNA binding sites, as appears to be the case for CASP3. Indeed, expression of BART miRNA clusters in B cells and AGS epithelial cells correlates with CASP3 depletion [20, 24]. If miRNA seed binding sites are 8 ~ 40 nt separated, repression is likely synergistic [14]. In CASP3 we noted one such potential case: the seed site for BART4 is 10 nt away from the seed site for human miRNA let-7. Further, while not highly expressed, let-7 is present in EBV+ Jijoye cells at approximately the same level as EBV-miR-BART5 or BART20-3p [18]. Human and viral miRNAs have been previously shown to co-target transcripts [17, 18], and it is possible that the effectiveness of BART4 repression is enhanced by its proximity to the let-7 site. TargetScan Human identified one other conserved but weaker potential binding site for a human miRNA: hsa-miR-138 (Fig. 1) [27]. MiR-138 is expressed at levels comparable to BART5 and let-7 in Jijoye cells [18], but our HITS-CLIP data do not overlap with this site, and no others have validated its direct regulation of CASP3. Consistent with this is the recent finding that miR-138 inhibits TNF-α-induced apoptosis in the human intervertebral disc, where an inhibitor of miR-138 dramatically suppressed the expression of cleaved CASP3 [34]. A previous study showed downregulation of CASP3 by BARTs 1-3p and 16 [25]. While our luciferase assay showed CASP3 repression upon treatment with BART1, we, like others [11], were unable to confirm their findings in the case of BART16. This is probably explained by the absence of true seed binding sites for BART16. The potential binding sites on the 3′-UTR of CASP3 contain a non-canonical Watson-Crick base pair in one nucleotide position of the seed region (Table 1). Non-canonical binding has been reported between human miRNAs and mRNAs, including seed regions binding with a G:U wobble base; however, the repression by these sites are not as significant as that of canonical sites [35, 36]. Similarly, the weaker 6mer site for BART10 did not exhibit repression of CASP3. Interestingly, the bioinformatically predicted site for BART20 was not validated by either HITS-CLIP data nor luciferase assays in spite of a predicted 7mer-A1 binding site, underscoring the need to move beyond bioinformatic predictions into the realm of biochemical testing before drawing conclusions. Kaposi’s sarcoma herpesvirus (KSHV), a human gamma-herpesvirus responsible for primary effusion lymphoma (PEL) and Kaposi’s sarcoma (KS), also expresses miRNAs that repress apoptotic targets including CASP3 [37, 38]. Though KSHV and EBV miRNA sequences are not conserved and show functional differences with respect to the latent-to-lytic switch, targeting the same apoptotic transcript potentially shows a similar method for controlling apoptosis in the two viruses [39, 40]. A decrease in apoptotic activity of EBV-positive cells could promote malignancy in conjunction with the deactivation of other important cellular pathways by EBV miRNAs. The 100+ other apoptotic transcripts with putative EBV miRNA binding sites should be studied to elucidate the mechanism by which EBV controls host cell apoptosis and further our understanding of how the virus causes malignancy. While the EBV BART miRNAs work in tandem to repress CASP3, therapeutically targeting a large number of miRNAs is likely more challenging and may have more off-target consequences. Knowing that single miRNAs can repress a key target may help us better conduct future rational drug design regarding EBV and BL. Additional file Additional file 1: Table S1. Oligonucleotides used in this study. (DOC 60 kb) Abbreviations EBVEpstein-Barr virus CASP3Caspase 3 ntnucleotide 3′-UTR3′-untranslated region BLBurkitt’s lymphoma RISCRNA-induced silencing complex miRNA(s)microRNA(s) BARTBamHI A rightward transcripts Acknowledgements We thank Joan Steitz, Daifei Liu, Therese Yario, other members of the Steitz lab, and Jennifer So, for constructs, preliminary studies, and consultations. Thanks to Susan Walsh, Anne Murdaugh, and Divya Kolakada for critical comments on the manuscript. K.J.R. was supported by Critchfield Research Grants (Rollins College) and Cornell New Faculty funding. C.H. and K.J.R. were supported by the Student-Faculty Collaborative Scholarship Program, Edward W. and Stella C. Van Houten Memorial Fund, and Herbert E. Hellwege Research Fund (Rollins College). Funding K.J.R. was supported by Critchfield Research Grants and Cornell New Faculty funding (Rollins College). C.H. and K.J.R. were supported by the Student-Faculty Collaborative Scholarship Program (Rollins College). These funding bodies played no role in the design of the study, writing of the manuscript, or collection, analysis, or interpretation of data. Availability of data and materials The datasets supporting the conclusions of this article are available in the NCBI Sequence Read Archive repository, SRA accession SRP068881, http://www.ncbi.nlm.nih.gov/sra/SRP068881. Authors’ contributions KJR conceived the study, performed all experiments, and wrote the manuscript. DC and CH helped with pilot experiments and cloning, and CH assisted in manuscript authorship. All authors critically reviewed 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. Houldcroft CJ Kellam P Host genetics of Epstein-Barr virus infection, latency and disease Rev Med Virol 2015 25 2 71 84 10.1002/rmv.1816 25430668 2. Pfeffer S Zavolan M Grässer FA Chien M Russo JJ Ju J John B Enright AJ Marks D Sander C Tuschl T Identification of virus-encoded microRNAs Science 2004 304 734 6 10.1126/science.1096781 15118162 3. Cai X Schäfer A Lu S Bilello JP Desrosiers RC Edwards R Raab-Traub N Cullen BR Epstein-Barr virus microRNAs are evolutionarily conserved and differentially expressed PLoS Pathog 2006 2 e23 10.1371/journal.ppat.0020023 16557291 4. Grundhoff A Sullivan CS Virus-encoded microRNAs Virology 2011 411 325 43 10.1016/j.virol.2011.01.002 21277611 5. Nourse JP Crooks P Keane C Nguyen-Van D Mujaj S Ross N Jones K Vari F Han E Trappe R Fink S Gandhi MK Expression profiling of Epstein-Barr virus-encoded microRNAs from paraffin-embedded formalin-fixed primary Epstein-Barr virus-positive B-cell lymphoma samples J Virol Methods 2012 184 46 54 10.1016/j.jviromet.2012.05.005 22609801 6. Yang H-J Huang T-J Yang C-F Peng L-XL-X Liu R-Y Yang G-D Chu Q-Q Huang J-L Liu N Huang H-B Zhu Z-Y Qian C-N Huang B-J Comprehensive profiling of Epstein-Barr virus-encoded miRNA species associated with specific latency types in tumor cells Virol J 2013 10 314 10.1186/1743-422X-10-314 24161012 7. Shin VY Chu K-M MiRNA as potential biomarkers and therapeutic targets for gastric cancer World J Gastroenterol 2014 20 10432 9 10.3748/wjg.v20.i30.10432 25132759 8. Shinozaki-Ushiku A Kunita A Isogai M Hibiya T Ushiku T Takada K Fukayama M Profiling of virus-Encoded MicroRNAs in Epstein-Barr virus-associated gastric carcinoma and their roles in gastric carcinogenesis J Virol 2015 89 5581 91 10.1128/JVI.03639-14 25740983 9. Linnstaedt SD Gottwein E Skalsky RL Luftig M a Cullen BR Virally induced cellular microRNA miR-155 plays a key role in B-cell immortalization by Epstein-Barr virus J Virol 2010 84 11670 11678 10.1128/JVI.01248-10 20844043 10. Qiu J Smith P Leahy L Thorley-Lawson DA The Epstein-Barr Virus Encoded BART miRNAs potentiate tumor growth In Vivo PLOS Pathog 2015 11 e1004561 10.1371/journal.ppat.1004561 25590614 11. Kang D Skalsky RL Cullen BR EBV BART MicroRNAs target multiple pro-apoptotic cellular genes to promote epithelial cell survival PLOS Pathog 2015 11 e1004979 10.1371/journal.ppat.1004979 26070070 12. Seto E Moosmann A Grömminger S Walz N Grundhoff A Hammerschmidt W Micro RNAs of Epstein-Barr virus promote cell cycle progression and prevent apoptosis of primary human B cells PLoS Pathog 2010 6 e1001063 10.1371/journal.ppat.1001063 20808852 13. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004;281–297. 14. Grimson A Farh KK-H Johnston WK Garrett-Engele P Lim LP Bartel DP MicroRNA targeting specificity in mammals: determinants beyond seed pairing Mol Cell 2007 27 91 105 10.1016/j.molcel.2007.06.017 17612493 15. Schirle NT Sheu-Gruttadauria J MacRae IJ Structural basis for microRNA targeting Science 2014 346 608 13 10.1126/science.1258040 25359968 16. Bartel DP MicroRNAs: target recognition and regulatory functions Cell 2009 136 215 33 10.1016/j.cell.2009.01.002 19167326 17. Skalsky RL Corcoran DL Gottwein E Frank CL Kang D Hafner M Nusbaum JD Feederle R Delecluse H-J Luftig MA Tuschl T Ohler U Cullen BR The Viral and Cellular MicroRNA Targetome in lymphoblastoid cell lines PLoS Pathog 2012 8 e1002484 10.1371/journal.ppat.1002484 22291592 18. Riley KJ Rabinowitz GS Yario TA Luna JM Darnell RB Steitz JA EBV and human microRNAs co-target oncogenic and apoptotic viral and human genes during latency EMBO J 2012 31 2207 2221 10.1038/emboj.2012.63 22473208 19. Bernhardt K Haar J Tsai M-H Poirey R Feederle R Delecluse H-J A Viral microRNA cluster regulates the expression of PTEN, p27 and of a bcl-2 Homolog PLoS Pathog 2016 12 e1005405 10.1371/journal.ppat.1005405 26800049 20. Marquitz AR Mathur A Nam CS Raab-Traub N The Epstein-Barr virus BART microRNAs target the pro-apoptotic protein Bim Virology 2011 412 392 400 10.1016/j.virol.2011.01.028 21333317 21. Choy EY-W Siu K-L Kok K-H Lung RW-M Tsang CM To K-F Kwong DL-W Tsao SW Jin D-Y An Epstein-Barr virus-encoded microRNA targets PUMA to promote host cell survival J Exp Med 2008 205 2551 60 10.1084/jem.20072581 18838543 22. Kim H Choi H Lee SK Epstein-Barr Virus MicroRNA miR-BART20-5p Suppresses Lytic induction by inhibiting BAD-Mediated caspase-3-dependent apoptosis J Virol 2015 90 1359 68 10.1128/JVI.02794-15 26581978 23. Green DR Llambi F Cell death signaling Cold Spring Harb Perspect Biol 2015 7 a006080 10.1101/cshperspect.a006080 26626938 24. Lin X Tsai M-H Shumilov A Poirey R Bannert H Middeldorp JM Feederle R Delecluse H-J The Epstein-Barr Virus BART miRNA cluster of the M81 strain modulates multiple functions in primary B cells PLoS Pathog 2015 11 e1005344 10.1371/journal.ppat.1005344 26694854 25. Vereide DT Seto E Chiu Y-F Hayes M Tagawa T Grundhoff A Hammerschmidt W Sugden B Epstein-Barr virus maintains lymphomas via its miRNAs Oncogene 2014 33 1258 64 10.1038/onc.2013.71 23503461 26. Miranda KC Huynh T Tay Y Ang Y-S Tam W-L Thomson AM Lim B Rigoutsos I A Pattern-based Method for the Identification of MicroRNA binding sites and their corresponding Heteroduplexes Cell 2006 126 1203 1217 10.1016/j.cell.2006.07.031 16990141 27. Agarwal V, Bell GW, Nam J-W, Bartel DP. Predicting effective microRNA target sites in mammalian mRNAs. Elife. 2015;4. 28. Kozomara A Griffiths-Jones S miRBase: annotating high confidence microRNAs using deep sequencing data Nucleic Acids Res 2014 42 Database issue D68 73 10.1093/nar/gkt1181 24275495 29. Guo YE Riley KJ Iwasaki A Steitz JA Alternative capture of Noncoding RNAs or protein-coding genes by Herpesviruses to alter host T cell function Mol Cell 2014 54 67 79 10.1016/j.molcel.2014.03.025 24725595 30. Tuschl T. Annealing siRNAs to produce siRNA duplexes. CSH Protoc. 2006;2006. 31. Riley KJ Steitz JA The “Observer Effect” in genome-wide surveys of protein-RNA interactions Mol Cell 2013 49 601 604 10.1016/j.molcel.2013.01.030 23438856 32. Riley KJ Yario TA Steitz JA Association of argonaute proteins and microRNAs can occur after cell lysis RNA 2012 18 1581 1585 10.1261/rna.034934.112 22836356 33. Kanda T Miyata M Kano M Kondo S Yoshizaki T Iizasa H Clustered microRNAs of the Epstein-Barr virus cooperatively downregulate an epithelial cell-specific metastasis suppressor J Virol 2015 89 2684 97 10.1128/JVI.03189-14 25520514 34. Wang B Wang D Yan T Yuan H MiR-138-5p promotes TNF-α-induced apoptosis in human intervertebral disc degeneration by targeting SIRT1 through PTEN/PI3K/Akt signaling Exp Cell Res 2016 345 199 205 10.1016/j.yexcr.2016.05.011 27207584 35. Seok H Ham J Jang E-S Chi SW MicroRNA target recognition: insights from Transcriptome-wide non-canonical interactions Mol Cells 2016 39 375 81 10.14348/molcells.2016.0013 27117456 36. Loeb GB Khan AA Canner D Hiatt JB Shendure J Darnell RB Leslie CS Rudensky AY Transcriptome-wide miR-155 binding map reveals widespread noncanonical microRNA targeting Mol Cell 2012 48 760 70 10.1016/j.molcel.2012.10.002 23142080 37. Haecker I Gay LA Yang Y Hu J Morse AM McIntyre LM Renne R Ago HITS-CLIP expands understanding of Kaposi’s sarcoma-associated herpesvirus miRNA function in primary effusion lymphomas PLoS Pathog 2012 8 e1002884 10.1371/journal.ppat.1002884 22927820 38. Suffert G Malterer G Hausser J Viiliäinen J Fender A Contrant M Ivacevic T Benes V Gros F Voinnet O Zavolan M Ojala PM Haas JG Pfeffer S Kaposi’s Sarcoma Herpesvirus microRNAs target Caspase 3 and regulate apoptosis PLoS Pathog 2011 7 e1002405 10.1371/journal.ppat.1002405 22174674 39. Kincaid RP Sullivan CS Virus-encoded microRNAs: an overview and a look to the future PLoS Pathog 2012 8 e1003018 10.1371/journal.ppat.1003018 23308061 40. Boss IW Plaisance KB Renne R Role of virus-encoded microRNAs in herpesvirus biology Trends Microbiol 2009 17 544 53 10.1016/j.tim.2009.09.002 19828316

PMC005xxxxxx/PMC5002153.txt

==== Front Allergy Asthma Clin ImmunolAllergy Asthma Clin ImmunolAllergy, Asthma, and Clinical Immunology : Official Journal of the Canadian Society of Allergy and Clinical Immunology1710-14841710-1492BioMed Central London 15010.1186/s13223-016-0150-0ResearchTwo-weeks-sustained unresponsiveness by oral immunotherapy using microwave heated cow’s milk for children with cow’s milk allergy Takahashi Masaya takahasm@hirakata.kmu.ac.jp Taniuchi Shoichiro (+81)-72 -804-0101taniuchi@hirakata.kmu.ac.jp Soejima Kazuhiko soejimak@hirakata.kmu.ac.jp Hatano Yasuko hatanoy@hirakata.kmu.ac.jp Yamanouchi Sohsaku yanoius@hirataka.kmu.ac.jp Kaneko Kazuanri kanekok@hirakata.kmu.ac.jp Department of Pediatrics, Kansai Medical University, 2-5-1 Shin-machi, Hirakata, Osaka Japan 26 8 2016 26 8 2016 2016 12 1 444 12 2015 12 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 Several studies have reported that oral immunotherapy (OIT) is effective for children with cow’s milk (CM) allergy. These studies reported the efficacy of OIT in terms of desensitization, but did not describe sustained unresponsiveness to CM. The aim of this study was to evaluate the efficacy of the OIT protocol using microwave heated cow’s milk (MH-CM) in terms of 2-weeks-sustained unresponsiveness (2-weeks-SU) and safety. Methods Forty-eight children were enrolled in this study. Thirty-one children agreed to receive rush OIT using MH-CM (the OIT group) and another 17 children who did not want to receive rush OIT formed the untreated group. Rates of desensitization and 2-weeks-sustained unresponsiveness were compared between the two groups at 1 year after the start of OIT. We followed up these rates and safety data for another year and for longer in the OIT group. Results No children in the untreated group did not pass an open food challenge to CM. Of the 31 children in the OIT group, 14 (P = 0.002) achieved desensitization, and 8 (P = 0.036) achieved 2-weeks-SU to CM at 1 year from the start of OIT. Two years after the start of OIT, both the rate of desensitization and the rate of 2-weeks-SU in the OIT group significantly increased compared with the rates at 1 year (P = 0.025 and P = 0.008 respectively). Conclusions The rush OIT protocol using MH-CM was effective at inducing 2-weeks-SU s to CM and had a good safety profile in children with CM allergy. Trial registration Approval number: 324, Registered 3 February 2009 Keywords Food allergyMicrowave heated cow’s milkCow’s milk specific IgESustained unresponsivenessOral immunotherapyMami Mizutani Fundationissue-copyright-statement© The Author(s) 2016 ==== Body Background Cow’s milk (CM) is a common food ingested by children, and is the second most common immediate-type food allergy in Japanese children [1]. Allergen avoidance is the basic approach for the management of food allergy until clinical tolerance is confirmed. Although 50 % of children have tolerance to CM by 5 years of age and increasing to 75 % by their early teenage years [2], some children experience persistent allergic reactions [3, 4]. Oral immunotherapy (OIT) for food allergy has often been used for young children with CM allergy [6–15]. However, efficacy of OIT was only previously described in terms of desensitization and also there are no reports of sustained unresponsiveness and efficacy of following cow’s milk oral immunotherapy (CM-OIT) using microwave heated cow’s milk (MH-CM). Therefore, the true efficacy of OIT for CM allergy is unclear. Kim et al. previously showed that the addition of baked milk to the diet of children tolerating such foods appears to accelerate the development of unheated-milk tolerance compared to strict avoidance [16]. An explanation of this observation might be that the high temperature reduces allergenicity by destroying conformational epitopes of milk proteins that subsequently it promotes unheated CM tolerance. Another recent study reported that MH-CM reduced allergic responses in a mouse allergic model [17]. No study has been performed to investigate OIT for children with CM allergy using MH-CM. The aim of the present study was to evaluate the sustained unresponsiveness and safety of an OIT protocol using MH-CM in children with CM allergy compared with an untreated group. Patients and methods Study design A prospective, longitudinal, intervention study was performed from June 2010 to March 2015 in the Department of pediatrics of Kansai Medical University Hospital, Osaka, Japan. The study protocol was not registered in the official internet registration system in Japan, but was approved by the Institutional Ethics Committee of Kansai Medical University, which included defining the outcomes prospectively. Patients Inclusion criteria were as follows: aged 5–18 years and persistent CM allergy within the 2 weeks before the start of CM-OIT that was confirmed by (1) positive clinical history, (2) positive food challenge test to CM and (3) level of CM specific immunoglobulin E (sIgE, CAP-Phadia, Uppsala, Sweden) to CM above 0.35 kUA/L. We ensured that families had adequate information regarding the study and understood the implications of participation. Signatures designating informed consent were obtained. Exclusion criteria were as follows: severe atopic dermatitis and uncontrolled asthma (baseline FEV1 < 80 % of predictive value) according to Japanese pediatric guideline for the treatment and management of bronchial asthma [18]. We enrolled 48 children who met these criteria. Those who agreed to treatment were assigned to the treatment group (OIT group), and those who did not want to receive treatment were assigned to the control group (the untreated group). This allocation was not done in a random manner and was based on the decision of the children and parents. The patients in the untreated group were followed up at 10–14 months, and oral food challenge (OFC) and blood examinations were repeated. Oral food challenge All CM challenges were open challenges, and were performed at hospital settings and supervised by physicians. Clinical features of a reaction to CM were investigated for clinical purposes via an OFC as described previously [19]. A double-blind placebo-control food challenge is the gold standard for clinical studies, but is time-consuming in general practice. We could not assess subjective symptoms by OFC. Therefore, if patients had subjective allergic symptoms such as nausea, abdominal pain, sore throat, or itching, we increased the loading dose until the objective symptoms appeared. During the challenge, full emergency equipment was at hand. Prior to enrollment in the study, the children’s parents provided informed consent. Patients were asked to avoid anti-histamine use for 72 h before the OFC, but topical steroids were allowed. Patients were admitted to our day clinic in the morning. The challenge material for the OFC was fresh CM. The procedure is described in detail in Table 1. The challenge was interrupted if children demonstrated unambiguous clinical reactivity or after the administration of 40 mL of CM. Following the cessation of CM feeding, all children were observed for at least 3 h. In cases of obvious allergic symptoms (such as rash, coughing, vomiting, or diarrhea) to loading doses of less than 40 mL CM positivity was indicated; otherwise, CM negativity was indicated. The threshold dose was defined as the last dose in OFC at which only objective symptoms occurred and the levels of severity of allergy to CM were assessed by Sampson’s grading score [20].Table 1 Oral food challenge and Rush OIT protocol Oral food challenge Procedure Open food challenge Materials Fresh CM or MH-CM (heated for 100 s in 550 W microwave oven) or Initial dose 0.1 mL 1 mL (if no anaphylaxis is suspected) Subsequent dose 0.2, 0.5, 1,2, 5. 10, 20 and 40 mL 2, 5, 10, 20 and 40 mL (if no anaphylaxis is suspected) Interval Every 30 min Threshold dose Defined as the last dose in OFC at which only objective symptoms occurred and the levels of severity of allergy to CM [20] Rush OIT Initial dose Starting approximately at a tenth of the threshold dose determined at OFC using MH-CM 1.2-fold for each patient Increase in dose Starting with MH-CM, the dose was increased approximately 1.2-fold every time Ingest MH-CM 2–4 times a day at 2-h intervals If gradea 2 or 3 develops, the same dose or the previous tolerated dose was repeated If gradea 4 develops, OIT was stopped and the previous tolerated dose was repeated on the next day When reaching 200 mL of MH-CH, the ingestion was once a day No further increases in dosage because of repeated adverse events, after the highest tolerated dose was continued for 3 consecutive days without any allergic reaction, the rush OIT was terminated Maintenance Ingest the maintenance dose of 200 mL of MH-CM everyday for maintenance at home If the subject did not reach the target dose of 200 mL of MH-CM, the loading dose was gradually increased by 1 mL per day until the target dose of 200 mL was reached and the dose was continued MH-CM to fresh CM In cases where no adverse reactions were observed for 2 months during a daily intake of 200 mL of MH-CM, the time spent heating the CM in the microwave oven was gradually shortened by 10 s every week OIT oral immunotherapy, CM cow’s milk, OFC open food challenge, MH-CM microwave heated cow’s milk aGrade of anaphylaxis according to Sampson’s score [20] We also used MH-CM for OFC before OIT to define the initial dose of MH cow’s milk oral immunotherapy. Fresh CM was warmed in a microwave oven at 550 W for 100 s and then cooled to room temperature before the OFC. The MH-CM food challenge was only performed with the OIT group immediately before the start of OIT. Oral immunotherapy OIT was performed within 2 weeks of the OFC using fresh CM. All the patients were admitted to our hospital. The OIT was performed following a previously described method [19]. The OIT protocol is described in detail in Table 1. The OIT consisted of two phases: (1) a rush phase, and (2) a maintenance phase. After patients had undergone the maintenance phase followed by the rush phase, we evaluated whether patients in the OIT group were desensitized or had sustained unresponsiveness to CM at 1-, 2-, 3- and 4-year follow-up. When patients ingested 200 mL of fresh CM daily for 2 months without adverse events, they were considered desensitized to CM. If there were no adverse events for 2 months with a daily intake of 200 mL of fresh CM, a CM-OFC was performed at 2 weeks after discontinuation of the daily administration of 200 mL of fresh CM. During the 2 weeks the patient continues strictly elimination of CM products. Briefly, the CM-OFC was undertaken using the total 200 mL of fresh CM. The initial dose was set at 15 mL, and the increased at 30, 60 mL, and finally 95 mL (total 200 mL) every 30 min. If the result of the CM-OFC was negative, the patient was defined as having 2-weeks-sustained unresponsiveness (2-weeks -SU) for CM. Additionally, after the positivity of CM-OFC, the patient continued a daily intake of 200 mL of fresh-CM. After patients showed, 2-weeks-SU, the patients were free to ingest CM or CM-containing products. Twenty of the OIT patients were followed for 4 years after the initiation of CM-OIT and 30 were followed for 3 years. In comparison, in the untreated group, patients continued the elimination diet of CM and CM products after their OFCs. One year after the initial OFC, OFC with fresh CM was performed for the untreated group, and the results were compared with their previous OFC results. Laboratory tests Blood samples were collected before starting the rush OIT and at 1 and 2 years after the start of OIT. Analyses Two statistical analyses were performed: (1) comparison of the rates of desensitization and 2-weeks-SU at 1 year in the OIT and the untreated groups at 1 year, and (2) comparison of the rates of desensitization and 2-weeks-SU in the OIT group at 1- and 2-year follow-ups. Furthermore, we statistically analyzed the same factors indicated in Table 2 among children in the OIT group, to determine whether the baseline characteristics differed between the patients who achieved 2-weeks-SU to 200 mL of fresh CM at 2 years after the start of OIT and those who failed to achieve 2-weeks-SU. Table 2 Baseline characteristics of patients in the study group Characteristic Group P value OIT (N = 31) Untreated (N = 17) Gender (male) 23 14 0.396 Age at the challenge test  Median 9 7 0.179  Range (5–17) (5–16) Initial total IgE level (IU/mL)  Median 871 935 0.940  Range (90–21966) (122.7–30725) Initial CM-specific IgE level (kUA/L)  Median 25 29 0.407  Range (0.9–5730) (0.56–6290) Initial casein-IgE level (kUA/L)  Median 21.5 28.4 0.693  Range (1.1–7210) (1.12–8350) Initial β-lactoglobulin-IgE level (kUA/L)  Median 2 1.05 0.425  Range (0–61.8) (0.05–75.3) Presence of other food allergiesa  No 10 (32 %) 3 (18 %) 0.229  Yes 21 (68 %) 14 (82 %) Frequency of anaphylaxis at accidental ingestion  Never 9 (29 %) 2 (12 %) 0.109  Once 12 (39 %) 6 (35 %)  2–10 times 10 (32 %) 9 (53 %)  >10 times 0 0 Atopic dermatitis 13 18 0.48 Asthma  Severity of asthmaa 14 (80 %) 10 (59 %) 0.376  Intermittent 3 (45 %) 2 (12 %)  Moderate 11 (35 %) 8 (47 %)  Persistent 0 0 Threshold dose of CM at OFC (mL)  Median 2 2 0.337  Range (0.1–40) (0.5–10) Grade of allergic reaction at OFCb  1 8 (26 %) 2 (12 %) 0.823  2 13 (42 %) 12 (71 %)  3 6 (19 %) 3 (18 %)  4 4 (13 %) 0 (0 %) CM cow’s milk, OIT oral immunotherapy, OFC open food challenge a Severity of asthma according to Japanese Pediatric Guideline for the treatment and management of bronchial asthma [21] b Grade of anaphylaxis according to Sampson’s score [20] Statistics We statistically evaluated the clinical outcome of the OIT group (31 children) and the untreated group (17 children) at 1 year, using a 2-sided alpha level of 0.05, to detect a significant difference between the 2 groups. Using Fisher’s exact test and the Mann–Whitney U-test tested baseline characteristics of the patients. Fisher’ exact test was used to evaluate between-group differences with regard to achieving desensitization and 2-weeks-SU to 200 mL of CM at 1-year follow-up. We also evaluated the differences in achieving a remission stage at the 1- and 2-year follow-ups within the OIT group using the Wilcoxon sign test. All analyses were performed with Excel Statcel 3 (Microsoft Inc. USA). Results Baselines characteristics The baseline characteristics of these subjects are shown in Table 2; no significant differences were noted between the 2 groups. Clinical outcomes Study enrollment and outcomes are shown in Fig. 1 and Table 3. Fourteen of 31 children (42 %) in the OIT group were categorized as desensitized to fresh CM at the 1-year follow-up, whereas none in the untreated group passed an oral food challenge. In addition, there was a statistically significant difference between the 2 groups (P = 0.002). In the untreated group, the threshold doses (median 1 mL and range 0. 5–10 mL) and severity levels of allergy (median 2 and range 1–3) to CM in the initial OFC was not significantly different to the threshold doses (median 1.75 mL and range 0.1–40 mL) and severity levels of allergy (median 2 and range 1–4) in the second OFC (P = 0.913, P = 0.688), respectively. The repeated OFC in the untreated group was performed at the 1-year interval (median: 12 months and range: 10–14 months). In addition, 7 of 31 children (21 %) in the OIT group showed a significant induced 2-weeks-SU to CM in contrast to 0 of 17 children in the untreated group at 1-year follow-up (P = 0.036) (Table 3). By the 2-year follow-up, both the rates of desensitization and 2-weeks-SU sustained in the OIT group were significantly increased compared with the rates at the 1-year follow-up (P = 0.025 and P = 0.008, respectively, by Wilcoxon rank test). Both the rates of desensitization and 2-weeks-SU gradually increased every year and reached 85 and 70 %, respectively, by 4 years (Table 3). Two to 3 years after achieving, 2-weeks-SU, we followed the patients and confirmed that the patients did not experience any adverse events including oral itch after the ingestion of CM or CM products, including CM dependent exercise-induced anaphylaxis. Patients did not experience any adverse events in the follow-up period.Fig. 1 Study enrollment and outcomes of OIT oral immunotherapy with CM cow’s milk. (Hash) After the achievement of 2-weeks-sustained unresponsiveness, the patients are followed for 2–3 year and they are able to ingest CM and CM products freely without any adverse events. OFC open food challenge Table 3 The rate of desensitization and 2-weeks-sustained unresponsiveness for cow’s milk allergy in the oral immunotherapy group and the untreated group OIT (N = 31) Untreated group (N = 17) Desensitization Two-weeks-sustained unresponsiveness Pass OFC At 1-year follow-up 14/31 (45 %)*, # 7/31 (21 %)**, ## 0/17 (0 %)a At 2-year follow-up 18/30 (60 %)# 14/30 (47 %)## At 3-year follow-up 21/30 (70 %) 16/30 (53 %) At 4-year follow-up 17/20 (85 %) 14/20 (70 %) OIT oral immunotherapy, OFC open food challenge * P = 0.002 by Fisher’s exact test ** P = 0.036 by Fisher’s exact test # P = 0.025 by Wilcoxon signed rank test ## P = 0.008 by Wilcoxon signed rank test aPatients in untreated groups continued complete elimination of cow’s milk and a year later were performed open food challenge using fresh cow’s milk was performed The levels of CM-sIgE decreased significantly in the OIT group (median CM-sIgE level, 25.0–12.2 kUA/L) during 1-year follow-up. The levels of CM-sIgE also decreased in the untreated group (median CM-sIgE level, 29.2–27 kUA/L) between the initial OFC and the OFC at 1 year-interval. Threshold dose and grade of anaphylaxis using MH-CM in OFC Before the rush OIT using MH-CM, OFC using MH-CM was performed. The threshold dose (median 10 mL and range 1–40 mL) was significantly higher for patients using MH-CM than for the patients using fresh CM (median 2 mL and range 0.1–40 mL) (P = 0.014) (Fig. 2). In addition, the severity of the grade of allergic reaction was milder for patients who received MH-CM compared with that for patients receiving fresh CM (P = 0.036) (Fig. 2).Fig. 2 Changes of the threshold dose and the severity of the grade of anaphylaxis by Samson’s grading score [20] at OFC open food challenge using fresh CM cow’s milk and microwave heated cow’s milk (MH-CM). The MH-CM food challenge was only performed with the OIT group at the start of OIT. Fresh CM was heated in a microwave oven at 550 W for 100 s and the cooled to room temperture before the OFC Skin prick test Before the rush OIT, a skin prick test (SPT) was performed in 10 of 21 patients in the OIT group using MH-CM and fresh CM. The median wheal size of SPT (median 10.5 mm and range 7–16 mm) using MH-CM was not statistically different compared with that of SPT using fresh CM (median 11.5 mm and range 7–16 mm) (P = 0.878 by Wilcoxon rank test). Baseline characteristics in the OIT group according remission and continued intolerance There were no significant differences of threshold dose and severity of the grade of allergic reaction to CM at OFC between the patients who achieved 2-weeks-SU and those who failed to achieve 2-weeks-SU by the 1-year follow up. However, the following factors were significantly different between the patients who achieve 2-weeks-SU and those who did not (Table 4): initial serum CM-sIgE level (P = 0.0006), casein-sIgE level (P = 0.0005), β-lactoglobulin-sIgE level (P = 0.017), and severity of asthma (P = 0.03).Table 4 Baseline characteristics of patients in the intolerance group and the 2-weeks -sustained unresponsiveness group Characteristic Group P value Intolerance (N = 17) Two-weeks-sustained unresponsiveness (N = 14) Gender (male) 12 10 0.637 Age at the challenge test  Median 7 10 0.469  Range (5–17) (6–15) Initial total IgE level (UA/mL)  Median 1016 497 0.165  Range (202–21,966) (90–6337) Initial CM-specific IgE level (kUA/L)  Median  Range 87.4 (0.9–5730) 5.88 (1.78–55) <0.001 Initial casein-IgE level (kUA/L)  Median 83.9 7.445 <0.001  Range (1.1–7210) (1.12–86.9) Initial β-lactoglobulin-IgE level (kUA/L)  Median 8.6 1.66 0.017  Range (0.2–23.8) (0.0–17.4) Presence of other food allergiesa  No 5 (29 %) 5 (36 %) 0.503  Yes 12 (71 %) 9 (64 %) Frequency of adverse events/child/dose (%)  Total events 17 7.3 0.503  Grade 1 8 3.6  Grade 2 3.6 0  Grade 3 2.4 0  Grade 4 0 0 Atopic dermatitis 3 (18 %) 1 (7 %) 0.378 Asthma  Severity of asthmaa  Intermittent 1 (6 %) 3 (21 %) 0.030  Moderate 9 (53 %) 7 (7 %)  Persistent 0 0 Threshold dose of fresh CM at OFC (mL)  Median 1.5 5 0.790  Range (0.1–40) (1–20) Threshold dose of heated CM at OFC (mL)  Median 10 15 0.091  Range (1–40) (5–40) Grade of allergic reaction at OFC (fresh CM)b  1 3 (18 %) 5 (36 %) 0.415  2 7 (41 %) 6 (43 %)  3 3 (18 %) 3 (21 %)  4 4 (23 %) 0 (0 %) Grade of allergic reaction at OFC (heated CM)b 8 (41 %) 9 (64 %) 0.415  1 6 (35 %) 3 (21 %)  2 2 (12 %) 1 (7 %)  3 1 (6 %) 1 (7 %) CM cow’s milk, OFC open food challenge aSeverity of asthma according to Japanese Pediatric Guideline for the treatment and management of bronchial asthma [21] bGrade of anaphylaxis according to Sampson’s score [20] Safety data during the OIT period Table 5 shows the average frequencies of adverse events that occurred per dose in each subject either at hospital or at home. In the rush phase of OIT (in hospital), the average of adverse events per subject per dose was 0.167. According to Sampson’s classification [20], 0.113 of subjects had grade 1 adverse events, and 0.035 was grade 2, 0.017 were grade 3 and 0.002 was 4. At home during the following year, the average of adverse events per subject per dose was 0.085 and 0.041 of subjects had grade 1 adverse events, 0.035 were grade 2, 0.01 were grade 3 or 0.0004 were 4. At home between years 1 and 2, the average of adverse events per subject per dose was 0.044 and 0.024 of subjects had grade 1 adverse events, 0.013 were grade 2, 0.0005 were grade 3 or 0.0002 were 4.Table 5 Averages of reactions and treatments per dose per child in the rush phase of oral immunotherapy (in hospital), at 1 year (at home) and at 1–2 years (at home) Reactions per dose per child Total Grade 1a Grade 2a Grade 3a Grade 4a Rush phase 0.167 0.113 0.035 0.017 0.002 1 year 0.085 0.041 0.035 0.01 0.0004 1–2 years 0.044 0.024 0.013 0.0005 0.0002 Treatments per dose per child Anti-histamines Nebulized epinephrine Nebulized β2 agonists Oral steroids Intravenous steroids Epinephrine injection Rush phase 0.039 0.02 0.02 0.04 0.003 0.003(6)b 1 year 0.02 0 0.017 0.027 0.0009 0.0006(9)b 1–2 years 0.01 0 0.005 0.02 0 0.0001(1)b Each data express averages of reactions and treatments per dose per child aGrade of anaphylaxis according to Sampson’s score [20] bTotal numbers of injection For the treatment of these symptoms, 21 of 31 (64 %) children received medication at hospital. The averages of those receiving treatments per subject per dose are shown in Table 2. The number of patients receiving oral anti-histamines nebulized β-2 agonists, nebulized epinephrine agonists, oral dexamethasone, and intravenous steroids gradually decreased over the 2 study years. However, in the intolerant group, mild and moderate adverse events occurred and were treated by anti-histamines or nebulized β-2 agonists. Three patients each received a single epinephrine injection and one received three epinephrine injections at hospital. Four patients each received a single epinephrine injection, one received two epinephrine doses and one received three doses of epinephrine at the 1-year follow-up. In addition, these children did not achieve desensitization by 2 years after the start of OIT. Discussion The present study described the efficacy and safety of OIT using MH-CM for children with severe CM allergy compared with an untreated group. Three randomized control trials of CM-OIT previously reported significant differences of the threshold dosage and desensitization rates between patients who underwent OIT and those who maintained an elimination diet [5, 8, 9]. The success rate for desensitization ranged from 36 to 67 %. In our study, the desensitization rates were 45, 60, 70 and 85 % at 1, 2, 3 and 4 years after starting OIT, respectively. The three previous studies did not describe the results of CM sustained unresponsiveness [5, 8, 9], although one study described the rate of sustained unresponsiveness CM-OIT [21] in which 75 % of patients were successfully desensitized; however, patients who passed the OFC 2 weeks after ceasing the OIT included 27.1 % of patients on CM and those patients might have achieved sustained unresponsiveness. The published rates for sustained unresponsiveness are similar to those obtained in our study, but the rates of achieving desensitization were higher than in our study. This difference might be explained by the different protocols used or by the different study populations. The period following the discontinuation of OIT was relatively short (2 weeks) to assess sustained unresponsiveness. Although this approach was selected to be in line with NIAID–FDA recommendations food allergy clinical trial design at the time the study was designed and registered [22], it is acknowledged that a longer period of at least 4 weeks after discontinuation of treatment would now be advised. We plan to perform a longer period follow-up study (5 or more years after OIT) in which subjects will undertake an OFC after 4–8 weeks of secondary CM elimination to assess prolonged sustained unresponsiveness. MH-CM is prepared by heating in a microwave oven for 100 s. The temperature of MH-CM is almost 60 °C and is not aggregated. Before the intake of CM, MH-CM was cooled to room temperature. The threshold for MH-CM-OFC significantly increased compared with that of fresh CM-OFC. In addition, the severity of allergic symptoms in the MH-CM-OFC was significantly less than that for the fresh CM-OFC. Three children had a decreased threshold to MH-CM, two of which showed decrease in their grading score (4-1 and 2-1). The third child showed no change in grading score. The grading score of the child receiving fresh CM and MH-CM had scores of 2 and 2, respectively, and showed sporadic cough. A reason why there was no difference in grading score and decreased threshold to MH-CM between the two OFCs might be unclear. In a study by Mecherfli et al. to examine whether the observed microwave effects were dependent upon temperature, peptic hydrolysis was studied with conventional heating using the same temperatures as those attained with different microwave wattages: 43.2, 51.7 and 63.5 °C for 50 W (5 min), 100 W (3 min) and 200 W (3 min) irradiation, respectively [17]. They found a significant effect on peptic hydrolysis of purified β-lactoglobulin at 200 W (3 min) with 52 % protein degradation as estimated by electrophoretic densitometry relative to conventional heating at an identical temperature. Thus, microwave effects on peptic hydrolysis were significant compared with conventional heating to the same temperature. We measured the temperature of MH-CM at 550 W for 100 s, and found the value to be 60 °C. The effect on our microwave-heated method at 550 w (100 s) was supposed to be equal to that of the method at 200 w (3 min). Mecherfli et al. also confirmed this hydrolysis effect by pepsin using the sera of young patients allergic to bovine whey proteins using an anti-IgE immunoblotting assay [23]. These results suggest that microwave-heated treatment reduces the allergenicity of cow milk proteins containing β-lactoglobulin by enzymatic digestion in the intestine. Therefore, in the MH-CM-OFC, patients had higher threshold doses and fewer allergic symptoms than those in the fresh CM-OFC, though there were no significant differences in the wheal diameter by SPT between heated CM and fresh CM groups. A significant difference between the two OFCs was observed, but this was small. Therefore, future investigations should use a larger number of the two OFCs or double blind placebo control study of OIT using both MH-CM and fresh CM to elucidate whether MH-CM is a better intervention than fresh CM. Adverse events in the MH-CM-OIT group were also mild. It is difficult to compare the frequencies of adverse events during OIT with other studies because the methods of calculating adverse reactions and the classification of adverse reactions are different among studies. The only way to assess the adverse events is by using the percentage of patients requiring epinephrine. The percentage of patients requiring epinephrine in previous studies was 6.7 [6], 13.3 [8] 16.7 [5], and 30.8 % [9], whereas in our study it was 12.9 %. The two studies that had lower rates of epinephrine usage [6, 8] enrolled different age populations and use a different dosing method compared with our study. The study design and patient populations of the two studies with higher rates of epinephrine usage [5, 9] were similar to those of our study. The percentages of epinephrine treatment in these two studies were higher than that in our study. Based on these findings, MH-CM-OIT may be relatively safe compared with fresh CM-OIT. We statistically analyzed the factors in Table 4 for children in the OIT group to determine whether there was a difference in baseline characteristics between the patients who achieved 2-weeks-SU to 200 mL of fresh CM and those who failed to do so. There were several significant differences between the patients who achieved 2-weeks-SU and those who failed to achieve 2-weeks-SU including initial serum CM-sIgE level (P = 0.0006), casein-sIgE level (P = 0.0005), β-lactoglobulin-sIgE level (P = 0.017), and severity of asthma (P = 0.03) (Table 2). Similar results using fresh CM-OIT protocol were reported by Vázquez-Ortiz et al. [24]. Although they analyzed the baseline characteristics of the two groups by safety data. Cox proportional hazard multivariate regression model identified three variables (CM-sIgE, CM-skin prick test and Sampson’s severity grades at baseline food challenge) as independent risk factors for the persistence of allergic reactions. It is very interesting to note that the factors predicting prognosis are serum CM-sIgE level than the threshold dose by OFC. There were several limitations in our study. First, the study was not randomized. The observational period of the untreated group was limited to 1 year and we could not extend it because all the patients wanted to start either slow-type home-based OIT or rush type OIT after 1-year of observation. Therefore, we could not set a control group for the whole study period. Second, the immunological markers assayed were limited to IgE levels, and we were not able to include serum IgG4 levels or the SPT. Third, multiple regression analysis was used to identify independent factors to predict tolerance and reactivity. However the sample size was too small to determine whether the factors were significant. Fourth, the OIT and untreated groups are not equivalent in terms of their demographics, because there was a trend in the control group toward increased atopy (slightly higher total and CM-sIgE levels, increased percentage with other food allergies, increased percentage with moderate asthma), an increased percentage who had multiple prior episodes of anaphylaxis after accidental ingestion, and an increased percentage with grade 2 or higher anaphylaxis during initial OFC. These factors may be considered to cause a self-selection bias in the untreated group. Finally, we already had performed fresh CM-OIT in ten children before the start of the present study. We could not compare the study group using fresh CM-OIT, because the study was different from the present protocol. However, our study is the first study to assess 2-weeks-SU in CM-OIT compared with a control group using a MH-CM material. Conclusions In this study, we found that a rush OIT protocol using MH-CM was effective in a significant percentage of children with CM allergy and showed a good safety profile. Furthermore, serum CM-sIgE levels might be useful to predict 2-weeks-SU and safety of CM-OIT. Abbreviations CMCow’s milk OIToral immunotherapy MH-CMmicrowave heated cow’s milk 2-weeks-SU2-weeks-sustained unresponsiveness sIgEspecific immunoglobulin E OFCoral food challenge SPTskin prick test Authors’ contributions Conception and design, MT and ST; acquisition of data, YS, ST, MT, SK, and YH analysis and interpretation of data, YS, ST and MT drafting of the manuscript, ST and MT; critical revision, SK and YH; statistical analysis, YS; and supervision, ST and KK. All authors read and approved the final manuscript. Acknowledgements Not applicable. Competing interests The authors declare that they have no competing interests. Availability of data and materials Yes. Ethics approval and consent to participate This study was approved by the Institutional Ethic Committee at Kansai Medical University. The experiments were performed in accordance with the approved guidelines. An informed, written consent was obtained from each patient. The project was registered as a clinical study at the Institutional Ethics Committee of Kansai Medical University (Approval Number: 324 February 3, 2009). Funding This study was supported by the Mami Mizutani Foundation. ==== Refs References 1. Uris A Ebisawa M Ito K Aihara Y Ito S Mayumi M Japanese guideline for food allergy 2014 Allergol Int 2014 63 399 419 10.2332/allergolint.14-RAI-0770 25178179 2. Spergel JM Natural history of cow’s milk allergy J Allergy Clin Immunol 2013 131 813 814 10.1016/j.jaci.2013.01.027 23452903 3. Skripak JM Matsui EC Mudd K Wood RA The natural history of IgE-mediated cow’s milk allergy J Allergy Clin Immunol 2007 120 1172 1177 10.1016/j.jaci.2007.08.023 17935766 4. Santos A Dias A Pinheiro JA Predictive factors for the persistence of cow’s milk allergy Pediatr Allergy Immunol 2010 21 1127 1134 10.1111/j.1399-3038.2010.01040.x 20444157 5. Longo G Barbi E Berti I Meneghetti R Pittalis A Ronfani L Specific oral tolerance induction in children with very severe cow’s milk-induced reactions J Allergy Clin Immunol 2008 121 343 347 10.1016/j.jaci.2007.10.029 18158176 6. Martorell A De la Hoz B Ibáñez MD Bone J Terrados MS Michavila A Oral desensitization as a useful treatment in 2-year-old children with cow’s milk allergy Clin Exp Allergy 2011 41 1297 1304 10.1111/j.1365-2222.2011.03749.x 21481024 7. Morisset M Moneret-Vautrin DA Guenard L Cuny JM Frentz P Hatahet R Oral desensitization in children with milk and egg allergies obtains recovery in a significant proportion of cases. A randomized study in 60 children with cow’s milk allergy and 90 children with egg allergy Eur Ann Allergy Clin Immunol 2007 39 12 19 17375736 8. Pajno GB Caminiti L Ruggeri P De Luca R Vita D La Rosa M Oral immunotherapy for cow’s milk allergy with a weekly up-dosing regimen: a randomized single-blind controlled study Ann Allergy Asthma Immunol 2010 105 376 381 10.1016/j.anai.2010.03.015 21055664 9. Skripak JM Nash SD Rowley H Cuny JM Frentz P Hatahet R A randomized, double blind, placebo-controlled study of milk oral immunotherapy for cow’s milk allergy J Allergy Clin Immunol 2008 122 1154 1160 10.1016/j.jaci.2008.09.030 18951617 10. Patriarca G Buonomo A Roncallo C Del Ninno M Pollastrini E Milani A Oral desensitization in cow milk allergy: immunological findings Int J Immunopathol Pharmacol 2002 15 53 58 12593788 11. Patriarca G Schiavino D Nucera E Schinco G Milani A Gasbarrini GB Food allergy in children: results of a standardized protocol for oral desensitization Hepatogastroenterology 1998 45 52 58 9496487 12. Patriarca G Nucera E Pollastrini E Roncallo C De Pasquale T Lombardo C Oral specific desensitization in food-allergic children Dig Dis Sci 2007 52 1662 1672 10.1007/s10620-006-9245-7 17245630 13. Reche M, Valbuena T, Fiandor A, Padial A, Caete A, Pascual C. Early induction of oral tolerance protocol (OTI) in children with cow’s milk allergy. J Allergy Clin Immunol. 2011,127:AB24. 14. Rodriguez-Alvarez M Fernandez RM Robeldo ET Vazquez-Cortes S Martinez-Cocera C Follow up of desensitized patients and immunological changes after specific oral tolerance induction to milk Allergy 2009 64 481 482 15. Meglio P Bartone E Plantamura M Arabito E Giampietro PG A protocol for oral desensitization in children with IgE-mediated cow’s milk Allergy Allergy 2004 59 980 987 10.1111/j.1398-9995.2004.00542.x 15291907 16. Kim JS Nowak-Węgrzyn A Sicherer SH Noone S Moshier EL Sampson HA Dietary baked milk accelerates the resolution of cow’s milk allergy in children J Allergy Clin Immunol 2011 128 125 131 10.1016/j.jaci.2011.04.036 21601913 17. Mecherfi KE Olivier Rouaud O Curet S Negaoui H Chobert JM Kheroua O Peptic hydrolysis of bovine beta-lactoglobulin under microwave treatment reduces its allergenicity in an ex vivo murine allergy model Int J Food Sci Technol 2015 50 356 364 10.1111/ijfs.12653 18. Hamasaki Y Kohno Y Ebisawa M Kondo N Nishima S Nishimuta T Japanese pediatric guideline for the treatment and management of bronchial asthma 2012 Pediatr Int 2014 56 441 450 10.1111/ped.12389 25252046 19. Okamoto S Taniuchi S Sudo K Hatano Y Nakano K Shimo T Predictive value of IgE/IgG4 antibody ratio in children with egg allergy Allergy Asthma Clin Immunol 2012 8 9 10.1186/1710-1492-8-9 22676477 20. Sampson HA Anaphylaxis and emergency treatment Pediatrics 2003 111 1601 1608 12777599 21. Sato S Yanagida N Ogura K Asaumi T Okada Y Koike Y Immunotherapy in food allergy: towards new strategies Asian Pac J Allergy Immunol 2014 32 195 202 10.1016/j.jaci.2013.12.700 25268336 22. Plaut M Sawyer RT Fenton MJ Summary of the 2008 National Institute of allergy and infectious diseases–US food and drug administration workshop on food allergy clinical trial design J Allergy Clin Immunol 2009 124 671 678 10.1016/j.jaci.2009.05.027 19560803 23. El Mecherfi KE Saidi D Kheroua O Ghazalia Boudraa G, Touhami M, Rouaud O, Curet S. Combined microwave and enzymatic treatments for β-lactoglobulin and bovine whey proteins and their effect on the IgE immunoreactivity Eur Food Res Technol 2011 233 859 869 10.1007/s00217-011-1581-y 24. Vázquez-Ortiz M Alvaro-Lozano M Alsina L Garcia-Paba MB Piquer-Gibert M Giner-Muñoz MT Safety and predictors of adverse events during oral immunotherapy for milk allergy: severity of reaction at oral challenge, specific IgE and prick test Clin Exp Allergy 2013 43 92 102 10.1111/cea.12012 23278884

PMC005xxxxxx/PMC5002154.txt

==== Front Malar JMalar. JMalaria Journal1475-2875BioMed Central London 149510.1186/s12936-016-1495-zResearchAcceptability of malaria rapid diagnostic tests administered by village health workers in Pangani District, North eastern Tanzania http://orcid.org/0000-0003-0762-9189Mushi Adiel K. adiel.mushi@gmail.com 12Massaga Julius J. JJMassaga@gmail.com 12Mandara Celine I. drceline2003@yahoo.com 3Mubyazi Godfrey M. gmmubyazi@yahoo.co.uk 12Francis Filbert ffrancis8@gmail.com 3Kamugisha Mathias kamugishauk@yahoo.co.uk 3Urassa Jenesta jurassa@yahoo.com 2Lemnge Martha marthalemnge@yahoo.com 3Mgohamwende Fidelis fsmgoha@yahoo.com 4Mkude Sigbert mkudesigsbert@yahoo.com 4Schellenberg Joanna Armstrong joanna.schellenberg@lshtm.ac.uk 51 Centre for Enhancement of Effective Malaria Interventions, 2448, Barack Obama Drive, P.O. Box 9653, Dar es Salaam, Tanzania 2 National Institute for Medical Research, HQ, 3 Barack Obama Drive, 11101 Dar es Salaam, Tanzania 3 National Institute for Medical Research, Tanga Centre, P.O. Box 5004, Tanga, Tanzania 4 National malaria Control Programme, Ministry of Health and Social Welfare, 6 Samora Machel Avenue, 11478 Dar es Salaam, Tanzania 5 Faculty of Infectious and Tropical Disease, London School of Hygiene & Tropical Medicine, Keppel Street, London, UK 27 8 2016 27 8 2016 2016 15 1 43924 2 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 Malaria continues to top the list of the ten most threatening diseases to child survival in Tanzania. The country has a functional policy for appropriate case management of malaria with rapid diagnostic tests (RDTs) from hospital level all the way to dispensaries, which are the first points of healthcare services in the national referral system. However, access to these health services in Tanzania is limited, especially in rural areas. Formalization of trained village health workers (VHWs) can strengthen and extend the scope of public health services, including diagnosis and management of uncomplicated malaria in resource-constrained settings. Despite long experience with VHWs in various health interventions, Tanzania has not yet formalized its involvement in malaria case management. This study presents evidence on acceptability of RDTs used by VHWs in rural northeastern Tanzania. Methods A cross-sectional study using quantitative and qualitative approaches was conducted between March and May 2012 in Pangani district, northeastern Tanzania, on community perceptions, practices and acceptance of RDTs used by VHWs. Results Among 346 caregivers of children under 5 years old, no evidence was found of differences in awareness of HIV rapid diagnostic tests and RDTs (54 vs. 46 %, p = 0.134). Of all respondents, 92 % expressed trust in RDT results, 96 % reported readiness to accept RDTs by VHWs, while 92 % expressed willingness to contribute towards the cost of RDTs used by VHWs. Qualitative results matched positive perceptions, attitudes and acceptance of mothers towards the use of RDTs by VHWs reported in the household surveys. Appropriate training, reliable supplies, affordability and close supervision emerged as important recommendations for implementation of RDTs by VHWs. Conclusion RDTs implemented by VHWs are acceptable to rural communities in northeastern Tanzania. While families are willing to contribute towards costs of sustaining these services, policy decisions for scaling-up will need to consider the available and innovative lessons for successful universally accessible and acceptable services in keeping with national health policy and sustainable development goals. Keywords Malaria rapid diagnostic testsVillage health workersAcceptabilityRural TanzaniaMalaria Capacity Development Consortium (MCDC)1A05Mushi Adiel K. issue-copyright-statement© The Author(s) 2016 ==== Body Background The world has achieved tremendous progress with a 47 % decline in malaria mortality rates and an estimated 4.3 million averted deaths between 2001 and 2013 [1]. Despite this progress, the malaria burden remains heaviest in the African continent where an estimated 90 % of global malaria deaths occur. Such deaths could have been saved if the tools currently available had reached those who needed them [1, 2]. More efforts are urgent to improve access to prompt malaria case diagnosis and management especially in rural African villages where the highest malaria prevalence rate occurs in the midst of inadequate access to health facilities and required human resources for health [3]. Several resource-constrained countries report a positive effect on access to services for malaria and other childhood illnesses after implementing integrated community case management of malaria (iCCM). The iCCM strategy recommended by WHO and UNICEF since June 2012 has successfully used trained community-based service providers to complement health systems in improving access to essential treatment for children, including facilitating prompt management of malaria within 24 h of onset of symptoms [4–11]. Malaria rapid diagnostic tests (RDTs) have the potential to reduce unnecessary and inappropriate treatment of malaria by targeting anti-malarials to only those with confirmed malaria [12–18]. Nevertheless, delivery of known health interventions through the existing health infrastructure remains inadequate particularly in rural areas [1, 19]. Formalization of iCCM with trained village health workers (VHWs) can strengthen and extend the scope of public health services, including diagnosis and management of uncomplicated malaria in areas with insufficient health infrastructure and qualified health workers [1, 18–20]. VHWs, popularly known as community health workers or community owned persons, are a cadre that relies on community members who are trained in delivering malaria case management and other health services in many endemic countries. VHWs deserve attention in the era of sustainable development goals, which include achieving universal access to quality and essential healthcare services [51]. They substantially contribute to improved access to increased demand for basic essential health services, including managing preventive and therapeutic care for malaria and other basic services in compliance with Millennium Development Goal targets [21–23]. VHWs have greatly assisted in overcoming access barriers and increase early treatment-seeking behaviour for malaria symptoms, prompt diagnosis and appropriate prescription of anti-malarials through RDTs, besides facilitating adherence to treatment and hence contributing to reduction in malaria-related mortality in the context of iCCM. Having established convincing evidence in management of malaria and other childhood illnesses, some malaria-affected countries, such as Zambia, Democratic Republic of Congo, Sudan, and Uganda, have reached advanced stages in considering formalization of VHWs [24–29]. Besides treatment, trained VHWs facilitate early referral of malaria parasite-negative cases to health facilities for management by qualified health personnel [19, 20]. Accordingly, countries, such as Ethiopia and Rwanda, have integrated iCCM in their health systems, leading to improved access to essential health supplies, including artemisinin-based combination therapy (ACT), RDTs and antibiotics for pneumonia in line with national guidelines [24–29]. These countries are in the same region and similarly prone to malaria as is Tanzania. Tanzania is among 16 countries that account for 80 % of malaria deaths globally, as well as being among the 18 countries that account for 80 % of global malaria cases [1]. This situation has prevailed despite long-term initiatives inspired by recognizable political will to fight malaria in 2006 and 2013 in Abuja, Nigeria. Tanzania and other African states resolved to control, eliminate and subsequently eradicate malaria on the continent. The country is among the founders of the African Leaders’ Malaria Alliance (ALMA), a platform that supports accountability mechanisms against malaria in malaria-prone African countries. However, Tanzania has not yet implemented iCCM which is recommended by WHO and listed as one of ALMA’s performance indicators for measuring progress and accountability [30, 31]. The national malaria treatment guidelines in Tanzania recommend confirmatory diagnosis prior to treatment. Tanzania initiated a phased implementation of RDTs in 2009, starting at hospital level, through to health centres and to the dispensary level. Although nearly half (44 %) of the Tanzanian population lives between 2 and 5.9 km from the nearest health centre or dispensary, the prevailing geographical inequalities and resources shortage challenge access to health services in the country in general [32, 33]. The existing ratio of 1.3 health workers per 1000 persons means the country has a shortage of qualified personnel in all cadres, particularly in remote rural areas [34]. Consequently, sub-optimal access to prompt treatment of malaria is likely to persist if service delivery remains through existing health facilities alone [35, 36]. According to lessons from other countries as presented above, formalization of trained VHWs has potential to expand access to malaria diagnosis using RDTs and treatment at community level, in line with policies currently recommended by WHO. VHWs in Tanzania are usually local people, selected by their own communities and widely recognized as providers of simple, curative and preventive care following swift training/crash courses and subsequent refresher programmes [37]. Tanzania has long experience of community-based health workers who are popularly termed in Swahili, the county’s national language, as Wahudumu wa afya vijijini, which translates as village health workers (VHWs). The popularity of VHWs in Tanzania stems from a Tanzanian Government initiative, in collaboration with UNICEF, which introduced the Child Survival, Protection and Development (CSPD) programme during the early 1990s to help conduct village censuses, check child nutritional status and provide support, including nutritional counselling. Since then, VHWs have been involved in broader health service provision than initially envisaged. For example, in southern Tanzania, VHWs have been found working at a government dispensary either together with or sometimes on behalf of trained medical assistants and nurses, in drug prescription and dispensing, besides child immunization and deworming campaigns, child growth monitoring, and community-based family planning services. Some VHWs participating in the present study reported being trained and equipped to administer deworming medicines and provide wound management services in their villages. The unresolved issues around recruitment, training, motivation, incentives, management, and acceptability have previously challenged the sustainability of VHWs in Tanzania [37–40]. It is encouraging to note that VHWs received some consideration in the previous and current national malaria control programme strategic plans [41, 42]. While implementation has not yet started, evidence about the role of VHWs in iCCM in Tanzania is on the rise. Studies conducted in northern and eastern Tanzania have demonstrated the capacity of VHWs to not only apply knowledge received in presumptive treatment of malaria [43], but also to use RDTs to inform decisions to treat positive cases and facilitate referral of those with negative tests to a higher level of care provision. VHWs have proven ability to reduce unnecessary treatment with anti-malarials in parts of Tanzania [44, 45]. Acceptability of a service delivery strategy within targeted communities is a key requirement for scaling up sustainable programmes, including RDTs delivered by VHWs [46, 47]. Here is evidence from rural northeastern Tanzania of community acceptability of RDTs delivered by VHWs. Methods Study design A mixed methods, cross-sectional study was conducted between March and May 2012 on community acceptability of RDTs used by VHWs in eight villages within Pangani district, northeastern Tanzania [43]. Study area and population Pangani district is located between 5°15.5–6° S and 38°35–39°00 E on the southern coast of Tanga region along the Indian Ocean coastline. Covering an area of 1803.8 km2, this rural district had a total population of 54,025 inhabitants in 2012 [34]. The district is characterized by a long rainy season from March to June, and short rains between November and December. There is perennial transmission of malaria, a leading cause of morbidity and mortality among children aged 2–9 years [43]. The main source of income includes subsistence farming, fishing, small-scale livestock rearing, and small trade. A few residents worked on sisal plantations [24]. At the time of conducting this study, Pangani district had a total of 22 dispensaries (19 public and three private), one health centre, one district hospital (public/government), and 15 accredited drug-dispensing outlets (ADDO). In Mwera division where this study took place, there were five health facilities including one health centre, four dispensaries and five ADDOs. Within the same division, there were 15 community-owned resource persons (CORPs) and 48 mother coordinators (MCs). The latter had been introduced in 2007 and equipped with supplies and skills to provide presumptive treatment of malaria to young children within a framework of research on home management of malaria (HMM) [43]. With a grant from a malaria capacity development consortium, a new component was nested within the same framework of the HMM project in Pangani district to study the feasibility of implementing RDTs through existing CORPs and MCs, jointly referred in this paper as VHWs [48]. Under this new component VHWs initially attended a 3-day training course at Pangani district headquarters, followed by attachment to their nearest dispensary or health centre for a period of 1 month. This training equipped them with practical skills to diagnose and treat positive malaria cases confirmed by RDTs, and/or refer negative and sick children to nearby health facilities, as well as record keeping and handling of RDTs before, during and after attending clients. All VHWs attended this training before they received supplies of RDTs and ACT to start services in their villages under close supervision by project team members in collaboration with a field coordinator and those in charge of nearby health facilities. Data collection The quantitative data were collected from eight of the 14 host HMM project villages of Mwera division. These eight villages were purposively selected; half of them had no health facilities to draw a convenient sample of 346 households with children aged under 5 years. In those households, mothers or caregivers of eligible children were interviewed using a structured questionnaire containing questions on knowledge, attitudes, practices, and future prospects of RDTs used by VHWs. Qualitative data were collected through focus group discussions (FGDs) held in four of the eight study villages. Two villages where FGDs took place had a health facility, the remainder did not. In each village, two FGDs were held, one with mothers or caregivers, and the other with VHWs to investigate awareness, willingness to pay, perceptions, practices, acceptability, and future prospects regarding RDTs used by VHWs. Demonstration of RDTs Prior to data collection, it was possible for respondents to see RDTs during FGDs, household surveys and concurrent community malaria surveys. Some respondents had also seen RDTs at nearby local health facilities or in use by VHWs in the study communities. Language used Swahili, the national language of Tanzania, is usually spoken in the area and was used throughout data collection. Data management and analysis Quantitative data were double-entered in Epi-info version 6.1 (CDC) software to check logical completeness and consistencies before analysis was performed according to a pre-defined analytical plan, using Stata version12 (Stata Corp LP, College Station, Texas, USA). Chi square tests were used to assess the association between categorical variables. p value was considered significant at <0.05. The FGDs were recorded using MP3 voice-recorders and transcribed verbatim alongside daily debriefing notes. All transcripts were manually scrutinised to code-related text under corresponding themes that generated the qualitative excerpts presented in this paper. Ethical considerations This study received ethical approval certificate number NIMR/HQ/R.8a/Vol.IX/1670 from the Medical Research Coordinating Committee of the National Institute for Medical Research (NIMR) in Tanzania. The district and village authorities were sensitized about the study before granting permission to work in their areas. Written informed consent was obtained from all study participants. Results Information on the type of data collected, data collection techniques, areas covered, and total data collected per technique is summarized in Table 1. In total, 346 mothers/caregivers of under-5 children with median age of 29 years (IQR 22–35) participated in the household survey. A total of four separate FGDs, each with eight participants was held with 32 FGDs in four villages. Equal numbers of mothers of under-5 children and VHWs from villages with and without VHWs participated in these sessions. The majority of household respondents had completed primary education (Table 2).Table 1 Data collection summary Type of data collected Data collection techniques Participants Number of villages covered Total data collected Quantitative Household questionnaire Mothers of children aged less than 5 years 8 out of 14 HMM study villages (4 with, and 8 without health facilities) 346 household questionnaires administered Qualitative Focus group discussions Separate sessions for mothers of children aged less than five years and village health workers 4 out of 8 villages where mRDTs component took place (2 with, and 2 without health facilities) 8 FGD sessions (two separate FGDs in each village, each with mothers/VHWs) The host HMM project that relied on presumptive treatment of malaria covered 14 villages. The current study with mRDTs component took place in 8 out of the 14 HMM study villages Table 2 Demographic characteristics of mothers/caretakers Variables (n = 346) n % Sex  Male 23 6.7  Female 323 93.3  Median age (IQR) in years 29 22–35 Age distribution in years  14–24 years 114 33  25–34 years 145 42  35+ years 87 25 Education  Primary 277 80  Secondary 21 6  Informal 3 1  None 45 13 Marital status (N = 345)  Single 45 13  Married 272 79  Separated 13 4  Widow 4 1  Divorced 11 3 Caregivers  Village with health facility 161 46  Village without health facility 183 53 Awareness of caregivers on different types of RDTs The respondents were firstly asked several questions on RDTs in order to situate RDTs in their existing understanding and experience. A total of 346 caregivers responded on awareness of RDTs used for HIV screening, RDTs and other diseases. No evidence was found of a difference in awareness of RDTs used for HIV and those used for malaria (54 vs. 46 %, p = 0.134), with very few reports of RDTs used for other health purposes (Fig. 1). In FGDs, mothers/caregivers more often mentioned RDTs for HIV used at health facilities than RDTs, based on the knowledge they gained particularly from the district hospital. Some participants confidently reported personal experience of being screened for HIV using similar tools and approach as they had seen used in a community malaria survey.Fig. 1 Awarenes of different types of rapid diagnosis test As an indication of positive attitudes towards RDT use in HIV testing, some FGD participants expressed a wish for similar tests for malaria in their community setting:We encourage the government to introduce RDTs in health facilities and to our VHWs as they did for HIV RDTs because this will help us to know if our children have malaria, to what extent. If not we can be advised on where else to seek care instead of receiving malaria treatment presumptively. (FGD, Mothers, village 2) There is no one who does not know about RDT for HIV AIDS these days. If you go to Pangani district hospital for mother and child health services don’t get surprised to be invited for a blood check-up using RDT together with your sexual partner. It is a nice thing and if introduced for malaria I will go for that personally and my child because malaria parasites will be checked and results obtained within short time. (Mother, FGD, village 2) Caregivers’ readiness to accept RDTs if administered by VHWs Caregivers were sequentially asked to express readiness to accept the RDTs administered by VHWs for themselves, or for their children in case of malaria symptoms. The majority (96 %) of 346 interviewed caregivers stated readiness to accept RDTs in both scenarios. As shown (Table 3), readiness to accept RDTs was significantly higher among caregivers with formal education 97 % (95 % CI 95–98 %) compared to those without formal education 90 % (95 % CI 81–99 %, p = 0.02). There was no evidence of an association between readiness to accept RDTs and either (i) the distance from home to the nearest health care facility; or, (ii) gender, age or marital status. Similarly, the majority (96 %) of caregivers indicated readiness to accept the RDTs for their own child if administered by VHWs.Table 3 Willingness to accept mRDTs if administered by VHWs Variables Willingness to accept mRDTs p value Yes No Availability of health facility, n (%)  Community with health facility 156 (97) 5 (3) 0.407  Community without health facility 176 (95) 9 (5) Age groups in years, n (%)  14–24 years 110 (97) 4 (4) 0.454  25–34 years 137 (95) 8 (6)  35+ years 85 (98) 2 (2) Sex, n (%)  Male 23 (100) 0 (0) 0.308  Female 309 (6) 14 (4) Marital status, n (%)  Single 43 (96) 2 (4)  Married 262 (96) 10 (4) 0.836  Separated 12 (92) 1 (8)  Widow 4 (100) 0 (0)  Divorced 10 (91) 1 (9) Education level, n (%)  Formal (primary school or above) 289 (97) 9 (3) 0.02  Informal (never been to school) 43 (90) 5 (10) Perceived benefits of RDTs administered by VHWs Respondents’ perceived benefits of RDTs used by VHWs included confirmation of presence or absence of malaria parasites before prescription (51 %), brings diagnostic services closer to community (36 %) and receiving results more quickly (9 %). The majority of respondents (91 %) among the 346 caregivers in the household survey felt that RDTs would increase utilization of services provided by VHWs. However, 7 % of respondents reserved their opinion while (2 %) avoided RDTs used by VHWs due to perceptions that VHWs might not be competent enough to perform such services. Mothers in this study thought that introduction of RDTs would add value to the services rendered by VHWs and would complement those provided at health facilities, which were initially perceived as inadequate due to a lack of diagnostic services. Mothers supported implementation of RDTs through VHWs with expectations for improved services.We already know them [VHWs]. Mothers have been presenting their children to receive Mseto [i.e. ACT], only based on their thermometer and weighing scale which provides weight to guide on the amount of required drug. So, adding RDTs will increase people’s respect to their services because anyone would like to see her child getting the right treatment for a confirmed problem. (Mother, FGD, village 1) What we dislike is how our doctors at dispensaries and health centres endanger lives by giving anti-malarials to children before screening them to know problem. You just see them opening a box with Mseto [ACT] they give you a dose with instructions to follow when administering to a child. If symptoms persist, and you take the child there again they will give quinine, and in case of recurring symptoms they instruct to go further to the Pangani district hospital. This makes us doubt about quality of the malaria treatment services given to our children. None of us would say no to RDTs if introduced to our VHWs because that will give us and our children what health facilities do not have. (Mother, supported by all the members in the group, FGD, village 2) Some of us initially doubted these VHWs when we heard that they would be testing blood to know if a child deserves anti-malarials or not. It is not a joke when it comes to being able to diagnose malaria and give medicines. But now, I wish they will continue. I am saying so because what they are doing is not common for someone without advanced secondary school or nursing qualification. I will support if we can get even more VHWs in our sub villages as long as they will handles like the one we know. We have seen them receiving practical training before they come to community. And our doctors and nurses have assured us that the VHWs had qualified training. Various people who went to dispensary also spread news that the VHWs were practicing like experienced health workers. The same doctors also refer us to VHWs with our children sometimes for malaria diagnosis or medicines during stock outs at the dispensaries. (Mother, supported by all the members in the group, FGD, village 2) Decision-making on the use of RDTs from VHWs Over half (58 %) of all respondents in the household survey thought that mothers could lead decisions in seeking services, including those involving the use of RDTs from VHWs, although at times fathers could also do so while 30 % of them said that the head of a household or other guardian could make a decision to take a sick child to the VHW depending on the perceived urgency. Very few (12 %) thought that others in the family could make such a decision in the absence of the parents. Prerequisites for accepting RDTs administered by VHWs The caregivers were asked to comment on “what should be done to ensure successful use of RDTs if administered by VHWs?” The most common response (38 %) was the need for training of the VHWs on the administration of this technology, followed by those suggesting the need for ensuring reliable availability of anti-malarials and RDTs (29 %). Others recommended supportive supervision (9 %) of VHWs coupled with community sensitization (8 %). The respondents suggested communication channels about RDTs administered by VHWs should include the use of a village crier, in Swahili Mpiga Mbiu, (48 %), a village assembly (28 %), radio (7 %), posters, leaflets and religious leaders (20 %). Caregivers’ perceived willingness to pay for RDTs administered by VHWs The overwhelming majority (92 %) of household respondents (n = 346; 95 % CI 89–95 %) said they were willing to pay for RDTs administered by VHWs. In a question about what they would be willing to contribute towards the costs of RDTs used by VHWs, the median cost mentioned was Tshs 6000 (approx US$3), ranging from Tshs 500 (US$0.4). FGD participants also gave their views in favour of contributing to costs of RDTs used by VHWs:The test [RDT] is more important than ACT which we have been receiving from VHWs and hospitals without knowing the exact problems in our children. There are plenty of such medicines even at local shops at very low price (Tshs 500). What is most important is the diagnostic test and not relying on the thermometer only, and therefore we better have it even if it will be done at a small charge. (Mother, FGD, village 2) Discussion Can rural communities in Tanzania accept RDTs if implemented by VHWs? This study took place when Tanzania had a policy for using RDTs, not only for malaria screening at hospitals but for other diseases, including HIV and syphilis [49]. Hence, in order to situate discussions in existing realities, interviewing was extended to explore local knowledge and perceptions of rapid diagnostic tests used in HIV and syphilis besides malaria. Taking into consideration the broader notion of fears and stigma attached to HIV/AIDS testing in the past [50], and similarity of RDTs and RDTs used for HIV screening, it was considered necessary to understand community perceptions on such tests. Accounts from the majority of mothers and VHWs who shared positive personal experience of RDTs in HIV screening, did not only allow continuation of in-depth discourses that revealed high readiness of community members to accept RDTs, but also paved the way to adapting questions and discussions about RDTs used by VHWs into local understanding and practices about similar health interventions. This initial entry point was considered necessary. Knowledge about local perceptions of local health services is a key to scaling up malaria interventions, including RDTs [51]. Moreover, implementation of new interventions or clinical guidelines may not be successful, based on evidence, if it does not consider the crucial role of local knowledge and context [52, 53]. The community in this study expressed acceptance of RDTs after hearing that VHWs attended training at district headquarters and later at local health facilities, or in direct encounter during practical training at a nearby dispensary or health centre. Such acceptance might not have been possible in this study where villagers did not expect someone without secondary school education or nursing background to be able learn and perform sensitive tasks such as using RDTs and giving medication to children. Hence, what appears as believing and accepting after seeing, community attitudes changed towards VHWs, inspired not only by knowledge about VHWs having attended training at district headquarters but also by actual experience of their services during subsequent practical training at local health facilities before they started applying RDTs in their villages. The same VHWs were already accepted in provision of presumptive malaria management services in the same areas, and equipping them with skills and supplies transparently contributed to acceptability of RDTs. The findings have also shown approval of VHWs from health workers at local dispensaries and health centres. This together with situations in which VHWs referred sick children to health facilities after negative results from RDTs and vice versa, in the case of RDTs and ACT stock-outs at health facilities, portrayed good relationship between formal health workers and VHWs. Health facility staff are trusted as reliable source of health information in Tanzania [54]. Their approval and support, and their good relationship with VHWs are potential elements that boost community acceptance of RDTs used by VHWs has also been shown to inspire acceptance of tuberculosis therapy in Tanzania [55]. Availability of medication usually results in increased client satisfaction with health services in national community health worker programmes [38, 56]. In this study, community members appreciated reliable availability of RDTs and anti-malarials from VHWs. Moreover, the community reported convenient access to VHWs as an aspect that promoted mutual respect and understanding of service provision and utilization. A few mothers/caregivers in this study indicated that they might not take their children to VHWs even if they were equipped with RDTs, because of perceived limited ability of fellow villagers to master professional tasks. This raises questions such as, what might happen if a child felt sick in an area far from a health facility but where there is a VHW with RDTs? Would those parents necessarily avoid RDTs used by VHWs and opt for the distant health facility if they suspected malaria? The answer to this health seeking puzzle might link to existing knowledge of similar services. Refusals among parents/caregivers of children aged under 5 in consenting for earlier diagnosis has been documented in health facility-based studies conducted in Tanzania and other countries. This could happen due to perceptions that malaria symptoms are easily recognized and that anti-malarials are known to them and readily accessible for self-treatment. Perceived ability to recognize malaria symptoms among parents/caregivers might lead them to put pressure on VHWs as they have earlier done with health facility staff on prescription of anti-malarials. Parents may as well justify avoiding RDTs for their children if the costs are higher than alternative sources for managing “fake malaria” [57–60]. Regardless of whether RDTs are to be used by VHWs or not, there is a need to consider tailoring behaviour change communication on RDTs to convince potential users on where to get them affordably as well as the benefits of not relying on symptoms alone to diagnose malaria [58]. Promisingly, on balance, the perceived benefits outweighed the negative perceptions as the majority of respondents in this study expressed readiness to accept RDTs administered by VHWs for themselves, or for their children, in case of malaria symptoms. Presence of VHWs with RDTs and ACT in villages without health facilities with reduced travel to health facilities is consistent with the call for improving access to fever case-management through RDTs [61]. As in other places, mothers and VHWs in this study commented on the benefits of RDTs used by VHWs in terms of reducing the budget associated with over-use of ACT in children with negative malaria results, as well as controlling attendance to already resource-constrained health facilities, which could be costly to reach due to travel costs [16, 31, 34, 35, 62]. Mothers also linked RDTs used by VHWs with reducing drug costs which could be incurred based on clinical judgment as reported from studies conducted on the routine healthcare system [18, 19]. Tanzanian health workers have shared views in favour of RDTs to avoid over-use of ACT, both for reasons of reducing cost and for fear of anti-malarial resistance [58]. Lessons for implementation The Ministry of Health in Tanzania requires evidence on the applicability of RDTs at community level [42]. Lessons from Uganda suggest the necessity of strategies to improve drug supply, community support and feedback provision from the formal health system for better performance by community health workers [63]. As well as training and regular supplies, the findings have highlighted a need for close supportive supervision of VHWs in order for them to provide acceptable services [64], otherwise, community members may lose faith or trust in poorly trained VHWs when it comes to medical competence [65]. In Uganda, the iCCM became acceptable to caregivers after they learned that community-based health workers had received training on how to use RDTs and make appropriate decisions [66]. When communities hear about VHWs using RDTs they are likely to apply a ‘seeing is believing’ approach in making judgement about trusting this strategy in the knowledge that they have attended appropriate training, including practical sessions at local health facilities. Approval from health workers based at local health facilities and good relationship with VHWs also boosted community acceptability of VHWs and RDTs that that they use in the villages. Success of RDTs relies on the efforts to ensure community comprehension and trust in the skills of VHWs alongside initiatives to motivate these practitioners beyond the long-term tradition of regarding them as volunteers [67]. Children aged under 5 years are exempted from user fees for health services in the Tanzanian national health policy [68]. Parents/caregivers were not required to meet the costs of RDTs and ACT for children targeted in this study [47, 58]. Nevertheless, a previous multi-country study that included Tanzania reported parents’ willingness to pay not more than Tshs 500 (approx US$0.4) on ACT [58]. This price is consistent with the lowest price that this study recorded on RDTs. The highest reported willingness to pay on RDTs in this study, Tshs 6000 (approx US$3), was twice that of the typical RDT sold in a pharmacy in Dar es Salaam, the commercial capital of Tanzania, of Tshs 3000 (approx US$1.5). To make RDTs accessible to around 90 % of Tanzanian consumers, the recommended price was Tshs 250 (approx US$0.20–0.25) [58]. Willingness to pay may not necessarily imply actual ability to pay [26], and given the existence of socio-cultural barriers, may discourage utilization, and therefore the reported willingness to pay in this study should be viewed cautiously [57–60]. Government and donor willingness to commit funds is an option to make RDTs affordable if the intervention is to be used to significantly reduce the burden of malaria, especially for the poorest consumers and vulnerable members of the society, most of whom live in rural villages [58]. Policy decisions on sustainability of RDTs used by VHWs are expected to be aligned with Sustainable Development Goal 3, which includes a target to achieve universal health coverage through equitable access to affordable, essential medicine and vaccines. This will require integration of individual and population-level health promotion and preventative efforts with curative services [8, 69]. The lessons from possible barriers associated with willingness to pay and other socio-cultural barriers may inspire further in-depth, socio-economic, effectiveness studies on how to ensure sustainable financing, availability, increased coverage, and utilization of RDTs, not only through VHWs, but also by government and private healthcare providers [12, 47]. Recognition of services delivered by VHWs and integration of these services into the national health delivery policy could pave the way to supply chain and management in line with available guidelines. In this study, they complied with provision of free services to the target group of children aged under 5 years according to national policy. Nevertheless, Tanzania could learn from past local experience and best practices from countries where VHWs are involved in iCCM. Innovative sustainability strategies will be required to better address critical issues such as alignment of VHW programmes with existing public health system, effective management strategies, motivation, incentives, support from health workers based at nearby health facilities, uninterrupted availability of supplies, affordability to targeted users, monitoring and evaluation, and advocacy to promote the use of RDTs [4–6, 9, 10, 67]. Strengths and limitations The study included areas located far from health facilities in a district where people are generally highly aware of malaria as a public health problem. Health facilities in the study area had limited supplies and service providers, and some were open for 12 h or less each day. These factors might have led to the high acceptability of RDTs reported to a level that might not be duplicated in areas with different coverage of healthcare services. The present study was nested within a framework of a long-term research on HMM, which is not typical of Tanzania. As for the representativeness of the national situation, a study done in only one district is likely to provide limited data and evidence for generalization for the whole of Tanzania. There is a room to argue on limited generalizability of the study findings from a quantitative point of view. Qualitatively, this approach may not be sufficient to facilitate in-depth understanding, which might need more time and a different approach, such as participant observation [70, 71]. However, the study attempted to minimize limitations by involving a competent multidisciplinary team at all stages. It exposed community appreciation of the role of VHWs in malaria case management both presumptively and by using RDTs in under-served populations according how they received training and supplies. Conclusions RDTs implemented by VHWs are acceptable to rural communities in north eastern Tanzania. While families are willing to contribute towards the costs of sustaining these services, policy decisions for scaling-up will need to consider the available and innovative lessons for successful universally accessible and acceptable services in keeping with national health policy and sustainable development goals. Abbreviations ACTartemisinin-based combination therapy ADDOaccredited drug dispensing outlets ALMAAfrican Leaders’ Malaria Alliance HMMhome management of malaria CORPscommunity-owned resource persons CSPDChild Survival, Protection and Development programme FGDfocus group discussion iCCMintegrated community case management of malaria MCDCMalaria Capacity Development Consortium MCsmother coordinators RDTrapid diagnostic test UNICEFUnited Nations Children’s Emergency Fund VHWsvillage health workers WHOWorld Health Organization Authors’ contributions AKM: study design, drafting and refining data collection tools, data collection and analysis and first draft of the manuscript. JM, CIM, and GM: technical input on the study design/research protocol development, design and refinement of the data collection tools, data collection contributed to data analysis and writing the manuscript. FF and MK: statistical support throughout the study and commented on the manuscript. JU: supported qualitative data collection and processing and commented on the manuscript. FM and FM: technical support from national malaria policy perspective and commented on the manuscript. JAS and ML: study design and commented on the manuscript. All authors read and approved the final manuscript. Acknowledgements We thank the study participants, community members and leaders in villages where this study took place. Special thanks to health facility staff who trained and mentored the VHWs under enthusiastic field coordinator, Mr. Mgaya. Special thanks to Council Health Management Team in Pangani district for great support of our research team. We appreciate the committed data collection assistants for their hard work during training and fieldwork. Special thanks to Mr. Obeid Ole Kaondo, Mr. Beda Kusare and Ms. Azama Maneno at Centre for Enhancement of Effective Malaria Interventions in Tanzania for great administrative and logistic support. We acknowledge invaluable support from Dr. Mohamed, the manager of Tanzania National Malaria Control programme to this study. We are thankful to Dr. Mwelecele N. Malecela, the Director of National NIMR, Tanzania, for pronounced leadership and granting permission to publish this manuscript. Competing interests The authors declare that they have no competing interests. Availability of data and materials The datasets generated during the current study are available from the corresponding author on reasonable request. Disclaimer The findings and conclusions presented in this manuscript are those of the authors and do not necessarily reflect the official position of their affiliated institutions. Ethics approval and consent to participate This study received ethical approval certificate number NIMR/HQ/R.8a/Vol.IX/1670 from the Medical Research Coordinating Committee of the National Institute for Medical Research (NIMR) in Tanzania. The district and village authorities were sensitized about the study before granting permission to work in their areas. Written informed consent was obtained from all study participants. Funding This study was made possible through support provided by Malaria Capacity Development Consortium (MCDC) Award number 1A05. ==== Refs References 1. WHO Global technical strategy for malaria 2016–2030, adopted by the World Health Assembly in May 2015 2015 Geneva World Health Organization 2. Bryce J Victora CG Black RE The unfinished agenda in child survival Lancet 2013 382 1049 1059 10.1016/S0140-6736(13)61753-5 24054535 3. de Castro MC Fisher MG Is malaria illness among young children a cause or a consequence of low socioeconomic status? Evidence from the united Republic of Tanzania Malar J 2012 11 161 10.1186/1475-2875-11-161 22571516 4. Amouzou A Hazel E Shaw B Miller NP Tafesse M Mekonnen Y Effects of the integrated community case management of childhood illness strategy on child mortality in Ethiopia: a cluster randomized trial Am J Trop Med Hyg 2016 94 596 604 10.4269/ajtmh.15-0586 26787148 5. English L Miller JS Mbusa R Matte M Kenney J Bwambale S Erratum to: monitoring iCCM referral systems: Bugoye integrated community case management initiative (BIMI) in Uganda Malar J 2016 15 343 10.1186/s12936-016-1375-6 27377403 6. English L Miller JS Mbusa R Matte M Kenney J Bwambale S Monitoring iCCM referral systems: bugoye integrated community case management initiative (BIMI) in Uganda Malar J 2016 15 247 10.1186/s12936-016-1300-z 27129920 7. Ferrer BE Webster J Bruce J Narh-Bana SA Narh CT Allotey NK Integrated community case management and community-based health planning and services: a cross sectional study on the effectiveness of the national implementation for the treatment of malaria, diarrhoea and pneumonia Malar J 2016 15 340 10.1186/s12936-016-1380-9 27371259 8. Phiri TB Kaunda-Khangamwa BN Bauleni A Chimuna T Melody D Kalengamaliro H Feasibility, acceptability and impact of integrating malaria rapid diagnostic tests and pre-referral rectal artesunate into the integrated community case management programme. A pilot study in Mchinji district, Malawi Malar J 2016 15 177 10.1186/s12936-016-1237-2 27000034 9. Shaw B Amouzou A Miller NP Tafesse M Bryce J Surkan PJ Access to integrated community case management of childhood illnesses services in rural Ethiopia: a qualitative study of the perspectives and experiences of caregivers Health Policy Plan 2016 31 656 666 10.1093/heapol/czv115 26608585 10. Silva R Amouzou A Munos M Marsh A Hazel E Victora C Can community health workers report accurately on births and deaths? Results of field assessments in Ethiopia, Malawi and Mali PLoS One 2016 11 e0144662 10.1371/journal.pone.0144662 26731544 11. Integrated community case management of malaria. http://www.who.int/malaria/areas/community_case_management/overview/en/. Accessed 28 Jul 2016. 12. Uzochukwu BS Onwujekwe OE Uguru NP Ughasoro MD Ezeoke OP Willingness to pay for rapid diagnostic tests for the diagnosis and treatment of malaria in southeast Nigeria: ex post and ex ante Int J Equity Health 2010 9 1 10.1186/1475-9276-9-1 20148118 13. Ansah EK Epokor M Whitty CJ Yeung S Hansen KS Cost-effectiveness analysis of introducing RDTs for malaria diagnosis as compared to microscopy and presumptive diagnosis in central and peripheral public health facilities in Ghana Am J Trop Med Hyg 2013 89 724 736 10.4269/ajtmh.13-0033 23980131 14. Reyburn H Mbakilwa H Mwangi R Mwerinde O Olomi R Drakeley C Rapid diagnostic tests compared with malaria microscopy for guiding outpatient treatment of febrile illness in Tanzania: randomised trial BMJ 2007 334 403 10.1136/bmj.39073.496829.AE 17259188 15. Nankabirwa J Zurovac D Njogu JN Rwakimari JB Counihan H Snow RW Tibenderana JK Malaria misdiagnosis in Uganda—implications for policy change Malar J 2009 8 66 10.1186/1475-2875-8-66 19371426 16. Mosha JF Conteh L Tediosi F Gesase S Bruce J Chandramohan D Cost implications of improving malaria diagnosis: findings from north-eastern Tanzania PLoS One 2010 5 e8707 10.1371/journal.pone.0008707 20090933 17. Ansah EK Narh-Bana S Epokor M Akanpigbiam S Quartey AA Gyapong J Rapid testing for malaria in settings where microscopy is available and peripheral clinics where only presumptive treatment is available: a randomised controlled trial in Ghana BMJ 2010 340 c930 10.1136/bmj.c930 20207689 18. African Union Commission Abuja actions toward the elimination of HIV and AIDS, tuberculosis and malaria in Africa by 2030 2013 Abuja African Union Commission 19. WHO World malaria report 2013 2013 Geneva World Health Organization 20. African Union Commission Universal access to HIV/AIDS, tuberculosis and malaria services by a United Africa by 2010 2006 Abuja African Union Commission 21. Freeman P Perry HB Gupta SK Rassekh B Accelerating progress in achieving the millennium development goal for children through community-based approaches Glob Public Health 2012 7 400 419 10.1080/17441690903330305 19890758 22. Haines A Sanders D Lehmann U Rowe AK Lawn JE Jan S Achieving child survival goals: potential contribution of community health workers Lancet 2007 369 2121 2131 10.1016/S0140-6736(07)60325-0 17586307 23. WHO Global experience of community health workers for delivery of health related millennium development goals: a systematic review, country case studies, and recommendations for integration into National Health Systems 2010 Geneva World Health Organization 24. http://www.who.int/malaria/areas/community_case_management/en/. Accessed 28 Jul 2016. 25. WHO/RBM&TDR WHO/TDR manual on HMM: Scaling up home-based management of malaria. From research to implementation 2000 Geneva World Health Organization 26. WHO The roll back malaria strategy for improving access to treatment through home management of malaria. WHO/HTM/MAL/2005.1101 2005 Geneva World Health Organization 27. WHO/RBM&TDR Scaling up home-based management of malaria. From research to implementation. WHO/HTM/MAL 2004. 1096, TDR/IDE/HMM 04.1 2004 Geneva World Health Organization 28. Kidane G Morrow RH Teaching mothers to provide home treatment of malaria in Tigray, Ethiopia: a randomised trial Lancet 2000 356 550 555 10.1016/S0140-6736(00)02580-0 10950232 29. Sirima SB Konate A Tiono AB Convelbo N Cousens S Pagnoni F Early treatment of childhood fevers with pre-packaged antimalarial drugs in the home reduces severe malaria morbidity in Burkina Faso Trop Med Int Health 2003 8 133 139 10.1046/j.1365-3156.2003.00997.x 12581438 30. ALMA. African leaders malaria alliance scorecard for accountability and action. 2014. http://alma2030.org/. Accessed 28 Nov 2015. 31. ALMA. African leaders malaria alliance scorecard for accountability and action- United Republic of Tanzania ALMA Quarterly Report Quarter Two, 2015. 32. D’Alessandro U Talisuna A Boelaert M Editorial: should artemisinin-based combination treatment be used in the home-based management of malaria? Trop Med Int Health 2005 10 1 2 10.1111/j.1365-3156.2004.01375.x 15655007 33. Williams HA Causer L Metta E Malila A O’Reilly T Abdulla S Dispensary level pilot implementation of rapid diagnostic tests: an evaluation of RDT acceptance and usage by providers and patients—Tanzania, 2005 Malar J. 2008 7 239 10.1186/1475-2875-7-239 19019233 34. NBS Basic facts and figures, 2012 on human settlements in Tanzania mainland 2012 Dar-es-Salaam Statististics NBS 35. Kadobera D Sartorius B Masanja H Mathew A Waiswa P The effect of distance to formal health facility on childhood mortality in rural Tanzania, 2005–2007 Glob Health Action 2012 5 1 9 23151364 36. Armstrong Schellenberg JR Adam T Mshinda H Masanja H Kabadi G Mukasa O Effectiveness and cost of facility-based Integrated Management of Childhood Illness (IMCI) in Tanzania Lancet 2004 364 1583 1594 10.1016/S0140-6736(04)17311-X 15519628 37. Gilson L National community health worker programmes World Health Forum 1990 11 85 86 2206244 38. Gilson L Walt G Heggenhougen K Owuor-Omondi L Perera M Ross D National community health worker programs: how can they be strengthened? J Public Health Policy 1989 10 518 532 10.2307/3342522 2621254 39. Mayombana C Jenkins J de Savigny D Tayari S Lubomba G Burnier E Training of village health workers in Tanzania; a comparison of two approaches Trop Doct 1990 20 63 67 2363189 40. Mushi AK Reaching the poorest children in rural southern Tanzania socio-cultural perspectives for delivery and uptake of preventive child health interventions 2009 London London School of Hygiene and Tropical Medicine 41. NMCP National malaria medium term strategic plan 2002–2007 2003 Dar es Salaam United Republic of Tanzania 42. NMCP National malaria medium term strategic plan 2008–2013 2009 Dar es Salaam United Republic of Tanzania 43. Massaga JJ Mushi AK Kamugisha M Francis F Barongo V Mubyazi G Performance of community-based drug providers in malarial fever management with artemether–lumefantrine in Pangani district, North-eastern Tanzania 2007 Dar es Salaam NIMR 44. Ishengoma DS Lwitiho S Madebe RA Nyagonde N Persson O Vestergaard LS Using rapid diagnostic tests as source of malaria parasite DNA for molecular analyses in the era of declining malaria prevalence Malar J 2011 10 6 10.1186/1475-2875-10-6 21226910 45. Mubi M Kakoko D Ngasala B Premji Z Peterson S Bjorkman A Malaria diagnosis and treatment practices following introduction of rapid diagnostic tests in Kibaha District, Coast Region, Tanzania Malar J 2013 12 293 10.1186/1475-2875-12-293 23977904 46. Druetz T Ridde V Kouanda S Ly A Diabate S Haddad S Utilization of community health workers for malaria treatment: results from a 3-year panel study in the districts of Kaya and Zorgho, Burkina Faso Malar J 2015 14 71 10.1186/s12936-015-0591-9 25889306 47. Pallas SW Minhas D Perez-Escamilla R Taylor L Curry L Bradley EH Community health workers in low- and middle-income countries: what do we know about scaling up and sustainability? Am J Public Health 2013 103 e74 e82 10.2105/AJPH.2012.301102 23678926 48. http://www.mcdconsortium.org/members/adiel-mushi. Accessed 12 July 2016. 49. Tucker JD Bu J Brown LB Yin YP Chen XS Cohen MS Accelerating worldwide syphilis screening through rapid testing: a systematic review Lancet Infect Dis 2010 10 381 386 10.1016/S1473-3099(10)70092-X 20510278 50. Katz IT Ryu AE Onuegbu AG Psaros C Weiser SD Bangsberg DR Impact of HIV-related stigma on treatment adherence: systematic review and meta-synthesis J Int AIDS Soc 2013 16 18640 10.7448/IAS.16.3.18640 24242258 51. Bastiaens GJ Bousema T Leslie T Scale-up of malaria rapid diagnostic tests and artemisinin-based combination therapy: challenges and perspectives in sub-Saharan Africa PLoS Med 2014 11 e1001590 10.1371/journal.pmed.1001590 24465186 52. Harrison MB Legare F Graham ID Fervers B Adapting clinical practice guidelines to local context and assessing barriers to their use CMAJ 2010 182 E78 E84 10.1503/cmaj.081232 19969563 53. Mickan S. Adapting knowledge to local context. Centre for Evidence-based Medicine, 2014. http://www.cebm.net/wp-content/uploads/2014/07/Adapting-knowledge-to-local-context.pdf. Accessed 28 July 2016. 54. Pool R Mushi A Schellenberg JA Mrisho M Alonso P Montgomery C The acceptability of intermittent preventive treatment of malaria in infants (IPTi) delivered through the expanded programme of immunization in southern Tanzania Malar J 2008 7 213 10.1186/1475-2875-7-213 18939971 55. Wandwalo E Robberstad B Morkve O Cost and cost-effectiveness of community based and health facility based directly observed treatment of tuberculosis in Dar es Salaam, Tanzania Cost Eff Resour Alloc 2005 3 6 10.1186/1478-7547-3-6 16018806 56. Zgierska A Miller M Rabago D Patient satisfaction, prescription drug abuse, and potential unintended consequences JAMA 2012 307 1377 1378 10.1001/jama.2012.419 22474199 57. Chandler CI Whitty CJ Ansah EK How can malaria rapid diagnostic tests achieve their potential? A qualitative study of a trial at health facilities in Ghana Malar J 2010 9 95 10.1186/1475-2875-9-95 20398262 58. Rennie W Lugo L Rosser E Harvey SA Willingness to use and pay for a new malaria diagnostic test for children under 5: results from Benin, Peru, and Tanzania 2009 Bethesda Center for Human Services 59. Muela SH Mushi AK Ribera JM The paradox of the cost and affordability of traditional and government health services in Tanzania Health Policy Plan 2000 15 296 302 10.1093/heapol/15.3.296 11012404 60. Muela SH Ribera JM Tanner M Fake malaria and hidden parasites-the ambiguity of malaria Anthropol Med 1998 5 43 61 10.1080/13648470.1998.9964548 26868738 61. SDSN. Indicators and a monitoring framework for sustainable development goals: launching a data revolution for the SDGs. Sustainable Development Solutions Network. 2015. 62. Uzochukwu BS Obikeze EN Onwujekwe OE Onoka CA Griffiths UK Cost-effectiveness analysis of rapid diagnostic test, microscopy and syndromic approach in the diagnosis of malaria in Nigeria: implications for scaling-up deployment of ACT Malar J 2009 8 265 10.1186/1475-2875-8-265 19930666 63. Bagonza J Kibira SP Rutebemberwa E Performance of community health workers managing malaria, pneumonia and diarrhoea under the community case management programme in central Uganda: a cross sectional study Malar J 2014 13 367 10.1186/1475-2875-13-367 25231247 64. Skarbinski J Ouma PO Causer LM Kariuki SK Barnwell JW Alaii JA Effect of malaria rapid diagnostic tests on the management of uncomplicated malaria with artemether–lumefantrine in Kenya: a cluster randomized trial Am J Trop Med Hyg 2009 80 919 926 19478249 65. Mubyazi GM Mushi AK Shayo E Mdira K Ikingura J Mutagwaba D Local primary health care committees and community-based health workers in Mkuranga District, Tanzania: does the public recognize and appreciate them? Ethno Med 2007 1 27 35 66. Mukanga D Tibenderana JK Peterson S Pariyo GW Kiguli J Waiswa P Access, acceptability and utilization of community health workers using diagnostics for case management of fever in Ugandan children: a cross-sectional study Malar J 2012 11 121 10.1186/1475-2875-11-121 22521034 67. Ruizendaal E Dierickx S Peeters Grietens K Schallig HD Pagnoni F Mens PF Success or failure of critical steps in community case management of malaria with rapid diagnostic tests: a systematic review Malar J 2014 13 229 10.1186/1475-2875-13-229 24924295 68. Mubyazi G Massaga J Kamugisha M Mubyazi JN Magogo GC Mdira KY User charges in public health facilities in Tanzania: effect on revenues, quality of services and people’s health-seeking behaviour for malaria illnesses in Korogwe district Health Serv Manag Res 2006 19 23 35 10.1258/095148406775322061 69. Schmidt H Gostin LO Emanuel EJ Public health, universal health coverage, and sustainable development goals: can they coexist? Lancet 2015 386 928 930 10.1016/S0140-6736(15)60244-6 26138141 70. Launiala A How much can a KAP survey tell us about people’s knowledge, attitudes and practices? Some observations from medical anthropology research on malaria in pregnancy in Malawi Anthropol Matters J 2009 11 1 13 71. Mushi AK Armstrong Schellenberg J Mrisho M Manzi F Mbuya C Mponda H Development of a behaviour change communication strategy for a vaccination-linked malaria control tool in southern Tanzania Malar J 2008 7 191 10.1186/1475-2875-7-191 18823531

PMC005xxxxxx/PMC5002155.txt

==== Front Malar JMalar. JMalaria Journal1475-2875BioMed Central London 149710.1186/s12936-016-1497-xResearchReviewing South Africa’s malaria elimination strategy (2012–2018): progress, challenges and priorities Raman Jaishree jaishreer@nicd.ac.za 123Morris Natashia natashia.morris@mrc.ac.za 4Frean John johnf@nicd.ac.za 12Brooke Basil basilb@nicd.ac.za 12Blumberg Lucille lucilleb@nicd.ac.za 5Kruger Philip philip.kruger@dhsd.limpopo.gov.za 6Mabusa Aaron aaronm@mpuhealth.gov.za 7Raswiswi Eric eric.raswiswi@kznhealth.gov.za 8Shandukani Bridget shandm@health.gov.za 9Misani Eunice misiae@health.gov.za 9Groepe Mary-Anne groepem@who.int 10Moonasar Devanand moonad@health.gov.za 391 Centre for Opportunistic, Tropical and Hospital Infections, National Institute for Communicable Diseases, Johannesburg, South Africa 2 Wits Research Institute for Malaria, University of Witwatersrand, Johannesburg, South Africa 3 Institute for Sustainable Malaria Control, University of Pretoria, Pretoria, South Africa 4 Health GIS Centre, South African Medical Research Council, Durban, South Africa 5 Division of Public Health Surveillance and Response, National Institute for Communicable Diseases, Johannesburg, South Africa 6 Department of Health and Social Welfare, Limpopo Provincial Government, Tzaneen, South Africa 7 Department of Health and Social Services, Mpumalanga Provincial Government, Nelspruit, South Africa 8 Department of Health KwaZulu-Natal, KwaZulu-Natal Provincial Government, Jozini, South Africa 9 Malaria Directorate, National Department of Health, Pretoria, South Africa 10 WHO Intercountry Support Team, Pretoria, South Africa 27 8 2016 27 8 2016 2016 15 1 4389 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 With a sustained national malaria incidence of fewer than one case per 1000 population at risk, in 2012 South Africa officially transitioned from controlling malaria to the ambitious goal of eliminating malaria within its borders by 2018. This review assesses the progress made in the 3 years since programme re-orientation while highlighting challenges and suggesting priorities for moving the malaria programme towards elimination. Methods National malaria case data and annual spray coverage data from 2010 until 2014 were assessed for trends. Information on surveillance, monitoring and evaluation systems, human and infrastructure needs and community malaria knowledge was sourced from the national programme mid-term review. Results Malaria cases increased markedly from 6811 in 2013 to 11,711 in 2014, with Mpumalanga and Limpopo provinces most affected. Enhanced local transmission appeared to drive malaria transmission in Limpopo Province, while imported malaria cases accounted for the majority of cases reported in Mpumalanga Province. Despite these increases only Vhembe and Mopani districts in Limpopo Province reported malaria incidences more than one case per 1000 population at risk by 2014. Over the review period annual spray coverage did not reach the recommended target of 90 % coverage, with information gaps identified in parasite prevalence, artemether-lumefantrine therapeutic utilization, asymptomatic/sub-patent carriage, drug efficacy, vector distribution and insecticide resistance. Conclusions Although South Africa has made steady progress since adopting an elimination agenda, a number of challenges have been identified. The heterogeneity of malaria transmission suggests interventions in Vhembe and Mopani districts should focus on control, while in KwaZulu-Natal Province eliminating transmission foci should be prioritized. Cross-border initiatives with neighbouring countries should be established/strengthened as a matter of urgency since malaria importation poses a real threat to the country’s elimination efforts. It is also critical that provincial programmes are adequately resourced to effectively conduct the necessary targeted elimination activities, informed by current vector/parasite distribution and resistance data. More sensitive methods to detect sub-patent infections, primaquine as a transmission-blocking drug, and alternative vector control methods need to be investigated. Knowledge gaps among malaria health workers and affected communities should be identified and addressed. Keywords MalariaEliminationSouth AfricaVector controlCase managementSurveillanceChallengesPrioritieshttp://dx.doi.org/10.13039/501100001322South African Medical Research Councilissue-copyright-statement© The Author(s) 2016 ==== Body Background Malaria is endemic to only three of South Africa’s nine provinces and is currently restricted to the low altitude border regions of these three provinces (Fig. 1) [1]. A robust malaria vector control and surveillance strategy dating back to the early 1940s ensured decades of effective malaria control resulting in the near elimination of malaria in South Africa by 1970 [2]. Unfortunately, favourable climatic factors (high rainfall and flooding) resulted in malaria resurging during the 1972/1973 malaria season. Heightened surveillance, prompt treatment with chloroquine together with dichloro-diphenyl-trichloroethane (DDT)-based indoor residual spray (IRS) operations brought malaria back under control until the mid-1980s [2]. Fig. 1 Malaria risk map for South Africa (Source South African Medical Research Council). Cross-hatched areas are game reserves The establishment of chloroquine-resistant parasites in the region caused a minor spike in malaria cases during the mid- to late-1980s, which was rapidly reversed by the replacement of chloroquine with sulfadoxine-pyrimethamine as first-line treatment [3]. Malaria remained under control until the mid-1990s when malaria case numbers began increasing once again, culminating in the country’s worst malaria epidemic to date, with over 60,000 reported cases at the height of the outbreak in 2000 [4]. Favourable climatic factors (elevated temperatures and rainfall), increased population movement across the country’s borders, together with the increased prevalence of drug-resistant parasites and insecticide-resistant vectors, particularly Anopheles funestus [5], have been identified as major contributing factors [6]. Since the 1999/2000 malaria outbreak, enhanced cross-border malaria control [7], the introduction of artemisinin-based combination therapy (ACT) for the treatment of uncomplicated malaria [8] and the re-introduction of DDT for IRS operations [9] have contributed to a marked and sustained decline in malaria prevalence. With fewer than 6000 cases notified in 2007 [10], the Malaria Directorate of the South African Department of Health initiated an internal dialogue on malaria elimination in accordance with the World Health Organization (WHO) recommendations [11]. These deliberations concluded with South Africa formally adopting an elimination strategy in 2012, aimed at halting local malaria transmission within the country’s borders by 2018 [12]. The elimination strategy consists of four clearly-defined key objectives, namely:To strengthen passive and active surveillance and monitoring and evaluation systems so that 100 % of districts report promptly and routinely on key malaria indicators by 2015 To ensure that all levels of the malaria programme have sufficient capacity to coordinate and implement malaria interventions by 2016 To ensure that 100 % of the population has adequate knowledge, attitudes and practices on malaria by 2018 through appropriate information, education and communication (IEC), social mobilization and advocacy To effectively prevent malaria infections and eliminate all parasite reservoirs in South Africa by 2018. Embedded within these broad objectives are three critical district level milestones, namely:Three districts in Limpopo (Capricorn, Waterberg, and Sekhukhune) and two districts in KwaZulu-Natal (Zululand and uThungulu) with <0.1 case per 1000 population at risk in 2012, reaching zero local cases by 2014 Three additional districts within the three malaria endemic provinces achieving zero local cases by 2016, and finally, All malaria endemic districts in South Africa reporting zero local cases by 2018. A mid-term review of the national malaria elimination strategic plan was conducted in July 2015 to assess progress made toward the successful completion of these critical milestones [13]. This paper reviews the progress made in the 3 years since South Africa’s official transition from malaria control to malaria elimination. It highlights the technical, policy and operational challenges that the country faces in achieving its elimination milestones while identifying opportunities that must be leveraged to ensure South Africa’s 2018 goal of malaria elimination becomes a reality. Methods Country setting Malaria is seasonal in South Africa, occurring between September and May with cases generally peaking after the Christmas and Easter holidays. Over 90 % of the reported cases are a result of Plasmodium falciparum infections with An. arabiensis, the most likely mosquito vector [14]. At present malaria is restricted to the low-altitude border regions of KwaZulu-Natal, Mpumalanga and Limpopo with approximately 10 % of the country’s population residing within a malaria risk area. Within KwaZulu-Natal three municipal districts, uMkhanyakude, uThungulu and Zululand, are currently classified as malaria endemic. In the recent past imported malaria has accounted for the majority of the cases reported from these three districts, placing them within the elimination phase of the WHO elimination continuum [6]. Of the three districts in Mpumalanga, namely Ehlanzeni, Gert Sibande, and Nkangala, at present only Ehlanzeni is classed as malaria endemic. Despite being the major contributor to Mpumalanga’s malaria burden, Ehlanzeni is considered to be in the elimination phase of the WHO elimination continuum as most of the cases reported in the district are imported [6]. In Limpopo, there are five malaria endemic municipal districts, namely Capricorn, Greater Sekhukhune, Mopani, Vhembe and Waterberg. Vhembe contributed to more than 60 % of the national malaria burden and is considered to be in the control phase of the elimination continuum, with the remaining districts either in the elimination or prevention of reintroduction phase [6]. In accordance with the national malaria treatment guidelines [15], all fever cases presenting at health facilities must be tested for malaria by P. falciparum specific rapid diagnostic test (RDT) or microscopy. Uncomplicated falciparum infections are treated with artemether-lumefantrine (Coartem®) while either IV quinine or IV artesunate are used to treat complicated malaria cases. Morbidity and mortality data Health-care workers at health facilities within the three malaria endemic provinces enter confirmed malaria case data into clinic or hospital registers as well as report cases to the district health office telephonically. In addition individual malaria case records are entered onto malaria notification forms, which are forwarded weekly to the provincial malaria control programmes (MCP). At the MCP offices individual patient demographic and case management data are captured onto a computerized malaria information system (MIS) [16]. A case report is then generated by the MIS, which is issued to the surveillance agent or malaria case investigator for follow-up and investigation. Once the follow-up investigation is completed the completed forms are returned to MCP offices, where any new information pertaining to the case is entered into the MIS using the patient’s unique identifier, thereby ensuring the new data are linked to the original patient record. Individual and aggregate malaria case data from each provincial MIS are transferred on a monthly basis to a national integrated MIS developed by the South African Medical Research Council and housed at the National Institute for Communicable Diseases (NICD) in Johannesburg, South Africa. Malaria case definition All confirmed malaria cases are classified as either local or imported [17]. A local (autochthonous) case is defined as a malaria infection acquired within a malaria-receptive area of South Africa where there is no history of travel to another malaria endemic country and where local transmission cannot be disproven. An imported malaria case is defined as an infection whose origin can be traced to a known malarious area outside of South Africa to which the individual has travelled. In instances where local transmission is unlikely but the malaria patient cannot be traced to verify travel history, the case is categorized as unclassified. Indoor residual spraying Currently in South Africa generalized IRS operations are conducted in the malaria-affected areas of the three malaria-endemic provinces, using a mosaic strategy comprising pyrethroids and DDT and, in certain instances carbamates, irrespective of malaria incidence. At the beginning of each malaria season provincial MCPs determine the number of structures to be sprayed, guided by the number of structures within the malaria endemic area, availability of insecticide and available insecticide resistance data. Spray personnel record the number of rooms and structures sprayed with insecticide on spray cards. The spray card data is verified by an IRS team leader and then submitted to the provincial MCP where the spray data are entered into a spraying information system (SIS). The SIS allows for the rigorous monitoring and evaluation of IRS operations. Mid-term review Data on the surveillance, monitoring and evaluations systems, programme human and infrastructure capacity, as well as community knowledge of malaria, were sourced from the report produced following the mid-term review of the national malaria elimination strategy [13]. Statistical analysis Descriptive statistical analysis was conducted for all variables identified as important. Statistical significance was set at 5 % with data analysis carried out using Stata version 13.1 (State Cooperation, College Station, TX, USA). Results National malaria-related morbidity and mortality data Although malaria case numbers increased marginally from 6548 in 2010 and to 7104 in 2011, they declined markedly to 5065 in 2012 (Fig. 2a). Unfortunately since 2012 the number of reported malaria cases has increased annually, peaking at 11,432 in 2014. Over the review period malaria case numbers from all three malaria endemic provinces mirrored the national trend (Fig. 2b) with KwaZulu-Natal consistently contributing the least to the national malaria burden. The major contributor to the national malaria burden from the endemic provinces alternated between Limpopo and Mpumalanga, with Limpopo accounting for majority of the cases reported in 2014 (Fig. 2b).Fig. 2 a Annual number of malaria cases (local, imported and unclassified) and deaths in South Africa between 2010 and 2014. b Annual number of malaria cases (local, imported and unclassified) by province in South Africa between 2010 and 2014. c Annual number of malaria cases (local and imported) in under- and over-5-year-old individuals by gender and province in South Africa between 2010 and 2014. KZN KwaZulu-Natal, LIM Limpopo, MPN Mpumalanga. d Contribution of local, imported and untraceable malaria cases to each province during the period 2010–2014 Between 2010 and 2012 national malaria-related deaths almost halved from 63 in 2010 to 34 in 2012, but have subsequently increased annually, mirroring the trend in malaria cases (Fig. 2a). Over the review period malaria deaths increased significantly from 63 in 2010 to 133 by 2014 (OR 1.09; 95 % CI 1.01–1.17; p = 0.024; Fig. 2a). Univariate analysis revealed that while females were marginally more likely to contract malaria (OR 1.02; 95 % CI 1.00–1.03; p = 0.038; Fig. 2c), pregnancy did not increase the risk of infection (OR 0.74; 95 % CI 0.63–0.86; p < 0.0001). The odds of children under the age of 5 years contracting malaria remained unchanged during the study period (OR 1.01; 95 % CI 0.99–1.03; p = 0.286; Fig. 2c). A closer inspection of the morbidity and mortality data revealed that the major source of the malaria infection varied greatly between the provinces (Fig. 2c, d). Local cases accounted for approximately 35 % of the cases reported in KwaZulu-Natal in 2010, and this value declined significantly to 18 % by 2014 (OR 0.79; 95 % CI 0.74–0.84; p < 0.0001; Fig. 2c, d). In contrast local cases accounted for the majority of Limpopo’s malaria burden, increasing significantly from 58 % in 2010 to 71 % in 2014 (OR 1.16; 95 % CI 1.14–1.18; p < 0.0001; Fig. 3). While the contribution of local cases to Mpumalanga’s malaria burden declined significantly over the study period (OR 0.94; 95 % CI 0.92–0.97; p < 0.0001; Fig. 2d), an increase in local cases from 12 to 21 % between 2012 and 2014 was noted (Fig. 2c, d).Fig. 3 Malaria incidence in South Africa by municipal district for the period 2010–2014 Provincial malaria morbidity and mortality data KwaZulu-Natal Province Over the past 5 years KwaZulu-Natal Province has consistently contributed to <1 % of the South Africa’s national malaria burden. During the review period, the provincial district of uMkhanyakude notified the highest number of malaria cases (over 95 %) with Zululand reporting the least. By 2014 malaria incidence in uMkhanyakude, uThungulu and Zululand districts was 0.10, 0.02 and 0.01 per 1000 population at risk, respectively (Fig. 3). The number of imported cases increased significantly across three endemic provincial districts from 92 in 2010 to 320 in 2014 (OR 1.27; 95 % CI 1.19–1.35; p < 0.0001), with females less frequently associated with imported malaria cases (OR 0.62; 95 % CI 0.51–0.75; p < 0.0001; Fig. 2c). Although malaria-related mortality remained unchanged over the study period (OR 0.76; 95 % CI 0.54–1.07; p = 0.113), individuals with locally-acquired malaria infections were four times less likely to survive compared to individuals with imported malaria (OR 4.14; 95 % CI 1.43–11.98; p = 0.009). Children under 5 years of age were more likely to acquire a local malaria infection compared to all other age groups (OR 1.36; 95 % CI 1.06–1.74; p = 0.017; Fig. 2c). Limpopo Province In contrast to KwaZulu-Natal, Limpopo Province has contributed significantly to South Africa’s malaria burden, accounting for most of the reported cases in 2010, 2011 and 2014 (Fig. 2b). Of the five provincial malaria endemic districts, Vhembe is the highest burdened district, followed by Mopani, with the remaining three districts mainly reporting imported cases. By 2014 malaria incidence in Vhembe was 2.4 and 1.7 in Mopani (Fig. 3). The number of locally-acquired cases increased significantly over the study period from 2478 in 2010 to 4116 in 2014 (OR 1.16; 95 % CI 1.14–1.18; p < 0.0001). Although unclassified cases decreased markedly from 1175 in 2010 to 181 in 2012, they began increasing thereafter, reaching 1053 by 2014. Despite this sharp increase in unclassified cases since 2012, the odds of a case remaining unclassified declined significantly over the study period (OR 0.88; 95 % CI 0.86–0.90; p < 0.0001). As seen in KwaZulu-Natal, females (OR 1.84; 95 % CI 1.73–1.97; p < 0.0001; Fig. 2c) and children under the age of five (OR 1.80; 95 % CI 1.66–1.95; p < 0.0001; Fig. 2c) in Limpopo Province were more likely to locally acquire malaria infections. Individuals who acquired an infection locally were two times less likely to survive than individuals who acquired the infection outside of South Africa (OR 2.35; 95 % CI 1.65–3.34; p < 0.0001). Mpumalanga Province While Mpumalanga Province, like Limpopo Province, contributes significantly to South Africa’s malaria burden, the majority of the cases reported in Mpumalanga Province are classified as imported with almost all cases notified in the district of Ehlanzeni. The number of imported cases increased over the study period from 1530 in 2010 to 4168 in 2014 (OR 1.06; 95 % CI 1.03–1.08; p < 0.0001). Once again females (OR 1.21; 95 % CI 1.11–1.30; p < 0.0001; Fig. 2c) and children under the age of five (OR 1.88; 95 % CI 1.71–2.07; p < 0.0001; Fig. 2c) were more likely to acquire malaria infections locally. Individuals who acquired an infection within Mpumalanga Province were three times less likely to survive the malaria infection compared to those infected with malaria outside of South Africa (OR 3.40; 95 % CI 2.33–4.97; p < 0.0001). Imported cases Of the 36,712 cases reported over the review period 17,511 (47.6 %) were classified as imported cases. Most of the imported cases originate from other African countries (Table 1) with the occasional case imported from Asia (data not shown). Mozambique accounted for 87.3 % (15,287/17,511) of all imported cases followed by Zimbabwe (5.8 %, 1025/17,511) (Table 1).Table 1 Ten most common African source countries of imported malaria reported in South Africa for the period 2010–2014 Country Number of cases Mozambique 15,287 Zimbabwe 1025 Ethiopia 403 Somalia 304 Swaziland 130 Malawi 110 Zambia 44 Democratic Republic of Congo 35 Tanzania 27 Congo 18 Spray coverage Both Mpumalanga and Limpopo provinces achieved an average annual IRS coverage of 85 % and above (Table 2) during the review period. Although KwaZulu-Natal generally achieved a spray coverage of above 80 %, only 69 % of the targeted structures were sprayed during IRS operations in 2012 (Table 2).Table 2 Number of structures sprayed and spray coverage achieved in South Africa by province from 2010 until 2014 Year Province Structures targeted Structures sprayed Spray coverage (%) 2010 KwaZulu-Natal 287,639 263,819 91.7 Limpopo 1,080,859 967,502 89.5 Mpumalanga 699,105 653,007 93.4 2011 KwaZulu-Natal 265,771 238,829 89.9 Limpopo 1,226,731 1,064,396 86.8 Mpumalanga 563,587 519,500 92.2 2012 KwaZulu-Natal 324,690 224,036 69.0 Limpopo 1,295,671 1,140,588 88.0 Mpumalanga 537,862 507,226 94.3 2013 KwaZulu-Natal 309,004 256,335 83.0 Limpopo 1,290,513 1,104,446 85.6 Mpumalanga 490,396 448,468 91.5 2014 KwaZulu-Natal 322,010 290,505 90.2 Limpopo 1,501,929 1,298,445 86.5 Mpumalanga 696,264 632,574 90.9 Surveillance, monitoring and evaluation systems Case reporting While most malaria cases detected at primary health care facilities were reported to the district and provincial malaria offices, this reporting generally did not take place within the required 24 h. These delays in case notification impeded both prompt reactive case investigations and the monitoring of malaria case data in real time at the provincial and national levels. In an attempt to improve 24-h case reporting, the National Malaria Directorate together with the Clinton Health Access Initiative developed a cellular application, MalariaConnect, which allows for immediate case reporting using cellular devices at no cost to the end user. Since the phased roll-out of the application commenced in August 2015, 305 facilities across five districts (Vhembe, Mopani, Ehlanzeni, uMkhanyakude and uThungulu) within the three endemic provinces have received training on the MalariaConnect application. Although all 305 facilities have begun using MalariaConnect, only 62 % of all confirmed cases at these facilities were being reported through MalariaConnect. Encouragingly however, 85 % of all cases reported using MalariaConnect was notified within 24 h. This has resulted in a marked improvement in case investigation response time, from an average of 6 days to an average of 3 days. Currently each malaria endemic province has its own MIS where all malaria case data are captured. Although data collected by these three information systems are not uniform, a core set of essential data variables are maintained and captured by all three systems. Captured provincial case data are transferred to the national integrated MIS on a regular basis using a web-based platform. Epidemic preparedness and response systems As an interim measure, while a national malaria early warning system (MEWS) is being developed, the national malaria control programme established epidemic thresholds based on 5 years of retrospective national case data to support provincial elimination efforts. In addition, certain provincial control programmes developed their own epidemic thresholds using retrospective provincial case data. While provincial response plans in the event of a threshold being breached are in place, financing of these responses remains a challenge. Entomological and insecticide resistance monitoring Varying levels of entomological surveillance and insecticide resistance monitoring are being undertaken in all three provinces, depending on available resources (human, infrastructural and financial). The NICD has assisted both Mpumalanga and KwaZulu-Natal with routine entomological surveillance and insecticide resistance monitoring. Detailed information on insecticide susceptibility status by vector species and province can be found in Brooke et al. [14, 18]. Human and financial resource capacity Both the national and provincial MCP are experiencing a severe shortage of technical experts at all levels. This lack of capacity continues to impact negatively on every facet of the elimination programme. The limited funds available for effective implementation of the highly resource-dependent elimination agenda is placing increased strain on already financially over-stretched provincial control programmes. The Malaria Directorate has attempted to garner funds from external sources but thus far funds raised have been insufficient to fill the identified resource gaps. Discussion In the 3 years since adopting the elimination agenda, South Africa has taken some positive steps towards achieving its 2018 elimination target. The potential for onward transmission has been significantly reduced by the increased access to point-of-care malaria diagnostics, prompt reactive case investigations facilitated through the implementation of a 24-h case reporting system, and improved surveillance for vectors and insecticide resistance. In addition the collection of more detailed travel and behavioural data during case investigations has enabled more rigorous case verification and more accurate case classification. This improved case classification helped inform appropriate intervention implementation while providing an indication of progress towards elimination. Notwithstanding these advancements, numerous challenges have been identified, which have the potential to jeopardize South Africa’s elimination goals if not adequately addressed. Like most of the malaria endemic countries within southern Africa, South Africa experienced an upsurge in malaria cases in 2014 [19]. This regional increase appears to have been driven primarily by favourable climatic conditions (optimum rainfall and ideal temperatures). Encouragingly, despite the regional increase in 2014, 10 of the 12 South African malaria-endemic districts maintained a malaria incidence of <1 case per 1000 population at risk. The two districts, Vhembe and Mopani, where the WHO elimination threshold [11] was exceeded, are located in Limpopo Province, the highest-burdened malaria-endemic province. Vhembe has been, and continues to be, the most affected malaria district in South Africa, accounting for over 60 % of all cases reported annually. As malaria transmission intensity decreases, malaria incidence becomes more heterogeneous [11], resulting in a stratification of intervention needs. The elevated malaria incidence in both Vhembe and Mopani districts implies a need for the implementation of enhanced integrated generalized control measures in which options for controlling malaria importation are explored. In total contrast to Limpopo Province, all three malaria endemic districts in KwaZulu-Natal had reached the minimum elimination threshold by 2010 [6] and remained there through the review period. Although these districts did not achieve the critical milestone of zero local cases by 2014, the sustained low incidence suggests local elimination is possible if hotspots (areas of persistent residual transmission) are eliminated [20, 21]. Hotspots fuel onward transmission, therefore targeting them reduces transmission intensity, positively impacting both the most affected households and the community as a whole. As optimum coverage of the targeted intervention is essential for hotspot elimination [20], selecting the appropriate combination of interventions targeting the vector and/or parasites is vitally important. To facilitate hotspot elimination and possibly fast-track malaria elimination in KwaZulu-Natal, a needs assessment for hotspot identification and elimination should be prioritized. Despite contributing significantly to South Africa’s national malaria burden, over 80 % of the cases reported in Mpumalanga are classified as imported [22]. Reactive case-detection data revealed South Africa’s high-burden neighbours, namely Mozambique and Zimbabwe, are the major source of imported malaria. Mathematical models [23] support the view that malaria elimination cannot be realized in the presence of sustained malaria importation [24–26]. In its recently launched global technical strategy (GTS) for malaria [27], the WHO acknowledged the importance of reducing malaria importation in an elimination setting by including a supportive strategy dedicated to cross-border collaborations. In line with the GTS recommendations, South Africa is engaging in the Elimination Eight (E8) regional initiative whose core objective is to strengthen regional collaborations to eliminate malaria in eight participating countries [28]. In addition, building on the successes of the Lubumbo Spatial Development Initiative, a cross-border collaboration between South Africa, Swaziland and Mozambique [7], South Africa is currently engaging with Mozambique and Swaziland on new cross-border initiative, called the Mozambique, South Africa and Swaziland (MOSASWA) malaria cross-border initiative. The overarching goal of MOSASWA is to achieve zero local transmission in Swaziland, South Africa and Maputo Province, Mozambique, by 2020 and pre-elimination status in southern Mozambique (Maputo and Gaza Provinces) by 2025, through harmonized collaborative efforts. Underpinning all these cross-border initiatives is the need for a surveillance system sensitive enough to track mobile and migrant populations, diagnostics that accurately detect asymptomatic individuals, sub-patent carriers and gametocyte carriers, transmission-blocking anti-malarials, novel vector surveillance and control methods, as well as appropriately skilled personnel. The accurate detection of all malaria carriers (symptomatic and asymptomatic) is a fundamental requirement of an elimination agenda. However, as transmission intensity continues to decline, the commonly used diagnostic tools, light microscopy and RDTs, lack the sensitivity required to detect sub-patent infections [29, 30]. Novel tools such as ultra-sensitive polymerase chain reaction (uPCR) and loop-mediated isothermal amplification (LAMP), which have been shown to be more sensitive in low transmission settings [31–33], need to be assessed for cost-effectiveness in a rural South African setting. Similar to other countries where malaria transmission intensity has decreased markedly, adults, as opposed to children and pregnant women, bear the higher malaria risk [34–36]. This heightened risk is most likely driven by social, behavioural and/or occupational factors that increase the odds of adults being exposed to malaria-infected vectors [34, 36]. In contrast to previous studies, adult females were slightly more at risk of contracting malaria, particularly if locally transmitted, compared to adult males. The reason for this is unclear but one possible contributing factor could be that adult males who travel for work and/or recreation are acting as asymptomatic parasite reservoirs and are sustaining local transmission [34]. A rather concerning finding was the increased risk of a negative outcome across all age groups if malaria was contracted within South Africa. One possible reason for this is a low index of suspicion among health-care workers and the general public in light of South Africa’s perceived low malaria risk. Malaria awareness campaigns aimed at improving health-seeking behaviours and case management practices of the general public and health-care workers, respectively, should be prioritized. All malaria endemic countries neighbouring South Africa have implemented a single-dose primaquine policy as a means of reducing onwards transmission [19], in accordance with a WHO recommendation [37]. As this anti-malarial drug is currently not registered in South Africa, this policy cannot be implemented locally at present. The scientific basis of a single-dose primaquine policy for malaria elimination in South Africa needs to be carefully evaluated. In addition to the maintenance and improvement of current IRS-based vector control interventions, South Africa needs to explore alternate vector control strategies such as larval source management and technologies that target outdoor-resting adult mosquito vectors. Finally, as community and malaria health worker engagement in, and support of, the elimination agenda is fundamental to its success, knowledge gaps, if any, need to be identified and appropriately addressed. Based on the findings of this review, key operational issues that should be prioritized to further South Africa’s elimination goals are listed below:Implementation of foci of transmission identification and elimination in the three malaria endemic districts of KwaZulu-Natal, as a means of realising the lapsed 2014 milestone of zero local cases in these districts. Maintaining generalized control intervention with blanket coverage in Vhembe and Mopani districts, Limpopo Province. Operationalisation of cross-border initiatives, particularly with Mozambique, to reduce the importation of malaria. Development of an enhanced surveillance system that allows for the tracking of mobile/migrant populations as well as proactive and reactive case detection. Reactive case detection should become routine in all districts where the WHO elimination threshold of <1 case per 1000 population at risk, has been met. Ideally as part of this investigation either a day 3 or 7 follow-up filter-paper blood sample should be collected from all individuals treated for malaria to allow for the assessment of artemether-lumefantrine efficacy. Evaluation of evidence for a single-dose primaquine policy. Assessment of novel techniques capable of detecting sub-patent and gametocyte carriers. Additional vector control measures, especially those targeting out-door resting vectors, need to be evaluated. Entomological surveillance activities, including routine insecticide resistance monitoring, need to be scaled up. Knowledge gaps among the affected communities and malaria health workers must be regularly assessed and addressed. Appropriate messaging that targets high-risk groups need to be developed. Conclusions Despite the marked increase in local malaria case numbers reported in 2014, South Africa has made considerable progress in implementing its elimination agenda. A 24-h malaria reporting system has been implemented in 305 facilities within the malaria endemic regions, enhanced surveillance for vectors and insecticide resistance has commenced and improved case management measures have been implemented. The sustained implementation of effective interventions has decreased transmission intensity causing malaria to become more heterogeneous. This heterogeneity calls for a stratification of interventions implemented. In areas of high transmission intensity, such as Vhembe and Mopani districts, generalized activities focussed at control should continue. In areas nearing elimination, such as KwaZulu-Natal, targeted activities aimed at identifying and eliminating foci of transmission must become a priority. If South Africa is to eliminate malaria by 2018, surveillance must be enhanced to allow for the timely identification and elimination of foci of transmission, using the appropriate targeted interventions at optimum coverage. In addition, systems to identify and appropriately treat asymptomatic/sub-patent malaria carriers must be implemented, the importation of malaria tackled by strengthening cross-border initiatives, a primaquine policy to reduce the likelihood of onwards transmission should be investigated, anti-malarial drug therapeutic efficacy assessed regularly, malaria awareness campaigns conducted frequently and the supply of chemoprophylaxis for travellers considered. South Africa’s experiences to date emphasize the need for an intensification of parasite and vector surveillance and implementation of a high-resolution MIS that enables active case detection and management as well as the targeting of priority focus areas, in countries attempting to eliminate malaria. Abbreviations ACTartemisinin-based combination therapy DDTdichloro-diphenyl-trichloroethane GTSglobal technical strategy IECinformation, education and communication IRSindoor residual spray MCPmalaria control programme NICDNational Institute for Communicable Diseases RDTrapid diagnostic test Authors’ contributions JR and DM defined the review, reviewed literature, wrote the first draft and edited all subsequent versions of the manuscript. JR performed the statistical analyses. NM generated the risk and incidence maps, assisted with data collection, manuscript drafting and editing. BB, PK, AM, ER, BS, EM, BB, MG were involved in data collection and manuscript editing. LB and JF contributed to the drafting and editing the manuscript. All authors read and approved the final draft of the manuscript. Acknowledgements The authors wish to acknowledge the assistance of Mr. Bongani Ngwenyama, Ms Rebecca Graffy, the provincial malaria control programme staff and malaria mid-term review team with data collection and data cleaning. Competing interests The authors declare that they have no competing interests. Availability of the data and material Data presented in this manuscript is the property of the Malaria Directorate, South African National Department of Health and can be requested from them. Consent for publication Approval for the publication of this manuscript was received from the CEO of the NICD, the Director of the Malaria Directorate, South African National Department of Health and all the co-authors. Ethical considerations Approval for this review and access to the national malaria case database was received from the Malaria Directorate, South African National Department of Health. This review was in compliance with ethical guidelines. Funding A South African Medical Research Council Malaria Collaborating Centre Research Grant to JR, JF, BB was used to cover data collection and publication costs. ==== Refs References 1. Morris N Frean J Baker L Ukpe IS Barnes KI Kruger P Re-defining the extent of malaria transmission in South Africa: implications for chemoprophylaxis S Afr Med J 2013 103 861 864 10.7196/SAMJ.7533 24148174 2. Coetzee M Kruger P Hunt RH Durrheim DN Urbach J Hansford CF Malaria in South Africa: 110 years of learning to control the disease S Afr Med J 2013 103 770 778 10.7196/SAMJ.7446 24079632 3. Ukpe IS Moonasar D Raman J Barnes KI Baker L Blumberg L Case management of malaria: treatment and chemoprophylaxis S Afr Med J 2013 103 793 798 10.7196/SAMJ.7443 24079636 4. Maharaj R Raman J Morris N Moonasar D Durrheim DN Seocharan I Epidemiology of malaria in South Africa: from control to elimination S Afr Med J 2013 103 779 783 10.7196/SAMJ.7441 24079633 5. Hargreaves K Koekemoer LL Brooke BD Hunt RH Mthembu J Coetzee M Anopheles funestus resistant to pyrethroids in South Africa Med Vet Entomol 2000 103 793 798 6. Maharaj R Morris N Seocharan I Kruger P Moonasar D Mabuza A The feasibility of malaria elimination in South Africa Malar J. 2012 11 423 10.1186/1475-2875-11-423 23253091 7. Sharp BL Kleinschmidt I Streat E Maharaj R Barnes KI Durrheim DN Seven years of regional malaria control collaboration—Mozambique, South Africa and Swaziland Am J Trop Med Hyg 2007 76 42 47 17255227 8. Barnes KI Durrheim DN Little F Jackson A Mehta U Allen E Effect of artemether-lumefantrine policy and improved vector control on malaria burden in KwaZulu-Natal, South Africa PLoS Med 2005 11 e330 10.1371/journal.pmed.0020330 16187798 9. Maharaj R Mthembu DJ Sharp BL Impact of DDT re-introduction on malaria transmission in KwaZulu-Natal S Afr Med J 2005 95 11 10. South African National Department of Health National Malaria Statistics 2014 Pretoria South Africa National Department of Health 11. WHO. Malaria elimination: a field manual for low and moderate endemic countries. Geneva: World Health Organization; 2007. Available: http://apps.who.int/iris/bitstream/10665/43796/1/9789241596084_eng.pdf. Accessed 17 May 2016. 12. South Africa National Department of Health Malaria elimination strategy for South Africa 2012–2018 2012 Pretoria South Africa National Department of Health 13. South Africa National Department of Health. Mid-term review of the national malaria elimination strategic plan 2012–2018. 2015. Technical report, South Africa National Department of Health. 14. Brooke BD Koekemoer LL Kruger P Urbach J Misiani E Coetzee M Malaria vector control in South Africa S Afr Med J 2013 103 784 788 10.7196/SAMJ.7447 24079634 15. South Africa National Department of Health. Guidelines for the treatment of malaria in South Africa. 2010. Technical report, South Africa National Department of Health. 16. Martin C Curtis B Fraser C Sharp BL The use of a GIS-based malaria information system for malaria research and control in South Africa Health Place 2002 8 227 236 10.1016/S1353-8292(02)00008-4 12399212 17. South Africa National Department of Health. Surveillance guidelines for malaria elimination and prevention of re-introduction for South Africa. 2012. Technical report, South Africa National Department of Health. 18. Brooke BD Robertson L Kaiser ML Raswiswi E Munhenga G Venter N Insecticide resistance in the malaria vector Anopheles arabiensis in Mamfene, KwaZulu-Natal S Afr J Sci 2015 111 1 3 10.17159/sajs.2015/20150261 19. Southern African Developmental Community, SADC Malaria Status by 2014 Report, 2015. 20. Bousema T Griffin JT Sauerwein RW Smith DL Churcher TS Takken W Hitting hotspots: spatial targeting of malaria for control and elimination PLoS Med 2012 9 e1001165 10.1371/journal.pmed.1001165 22303287 21. Dolgin E Targeting hotspots of transmission promises to reduce malaria Nat Med 2010 16 1055 10.1038/nm1010-1055 20930727 22. Silal SP Little F Barnes KI White LJ Predicting the impact of border control on malaria transmission: a simulated focal screen and treat campaign Malar J 2015 14 268 10.1186/s12936-015-0776-2 26164675 23. Silal SP Little F Barnes KI White LJ Towards malaria elimination in Mpumalanga, South Africa: a population-level mathematical modelling approach Malar J 2014 13 297 10.1186/1475-2875-13-297 25086861 24. Li S Yin S Wang J Li X Feng J Shifting from control to elimination: analysis of malaria epidemiological characteristics in Tengchong County around the China–Myanmar border, 2005–2014 Malar J. 2016 15 45 10.1186/s12936-016-1089-9 26823183 25. Bradley J Lines J Fuseini G Schwabe C Monti F Slotman M Outdoor biting by Anopheles mosquitoes on Bioko Island does not currently impact on malaria vontrol Malar J 2015 14 170 10.1186/s12936-015-0679-2 25895674 26. Hassan IM Shahly A Alzahrani MH Alhakeem RF Alhelal M Alhogail A Progress towards malaria elimination in Jazan Province, Kingdom of Saudi Arabia: 2000–2014 Malar J 2015 14 444 10.1186/s12936-015-0858-1 26552387 27. WHO. Global technical strategy for malaria 2016–2030. Geneva: World Health Organization; 2015. Available: http://who.int/malaria/publications/atoz/9789241564991/en/. Accessed 17 May 2016. 28. Elimination 8 (2015) E8 strategic plan (2015-2020). Elimination 8 Secretariat, Windhoek. Available: http://www.shrinkingthemalariamap.org/files/content/resource/attachment/E8%20Strategic%20Plan%20(2015-2020).pdf. Accessed 17 May 2016. 29. Britton S Cheng Q McCarthy JS Novel molecular diagnostic tools for malaria elimination: a review of options from the point of view of high-throughput and applicability in resource limited settings Malar J 2016 15 88 10.1186/s12936-016-1158-0 26879936 30. Cook J Xu W Msellem M Vonk M Bergstrom B Gosling R Mass screening and treatment on the basis of results of a Plasmodium falciparum -specific rapid diagnostic test did not reduce malaria incidence in Zanzibar J Infect Dis 2015 211 476 483 10.1093/infdis/jiu655 31. Morris U Khamis M Aydin-Schmidt B Abass AK Msellen MI Nassor MH Field deployment of loop-mediated isothermal amplification for centralized mass-screening of asymptomatic malaria in Zanzibar: a pre-elimination setting Malar J 2015 14 205 10.1186/s12936-015-0731-2 25982190 32. Imwong M Stepniewska K Tripura R Peto TJ Lwin KM Vihokhern B Numerical distribution of parasite densities during asymptomatic malaria J Infect Dis 2015 213 1322 1329 10.1093/infdis/jiv596 26681777 33. Imwong M Nguyen TN Tripura R Peto TJ Lee SJ Lwin KM The epidemiology of subclinical malaria infections in South-East Asia: findings from cross-sectional surveys in Thailand–Myanmar border areas, Cambodia, and Vietnam Malar J 2015 14 381 10.1186/s12936-015-0906-x 26424000 34. Cotter C Sturrock HJ Hsiang MS Lui J Phillips AA Hwang J The changing epidemiology of malaria elimination: new strategies for new challenges Lancet 2013 382 900 911 10.1016/S0140-6736(13)60310-4 23594387 35. Littrell M Sow GD Ngom A Ba M Mboup BM Dieye Y Case investigation and reactive case detection for malaria elimination in northern Senegal Malar J 2013 12 331 10.1186/1475-2875-12-331 24044506 36. Zhou S Zhongjie L Cotter C Zheng C Zhang Q Li H Trends of imported malaria in China 2010–2014: analysis of surveillance data Malar J 2016 15 39 10.1186/s12936-016-1093-0 26809828 37. WHO. Policy brief on single-does primaquine as a gametocyctocide in Plasmodium falciparum malaria. Geneva: World Health Organization; 2015. Available: http://who.int/malaria/publications/atoz/policy-brief-single-dose-primaquine-pf/en/. Accessed 17 May 2016.

PMC005xxxxxx/PMC5002156.txt

==== Front BMC MedBMC MedBMC Medicine1741-7015BioMed Central London 67510.1186/s12916-016-0675-6CommentaryThe efficiency of chronic disease care in sub-Saharan Africa Geldsetzer Pascal pgeldset@hsph.harvard.edu 1Ortblad Katrina katrina.ortblad@mail.harvard.edu 1Bärnighausen Till tbaernig@hsph.harvard.edu 1231 Department of Global Health and Population, Harvard T.H. Chan School of Public Health, 665 Huntington Avenue, Building 1, Boston, MA 02115 USA 2 Institute of Public Health, Heidelberg University, Im Neuenheimer Feld 324, Heidelberg, 69120 Germany 3 Africa Health Research Institute, P.O. Box 198, Mtubatuba, 3935 South Africa 26 8 2016 26 8 2016 2016 14 1 1272 8 2016 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.The number of people needing chronic disease care is projected to increase in sub-Saharan Africa as a result of expanding human immunodeficiency virus (HIV) treatment coverage, rising life expectancies, and lifestyle changes. Using nationally representative data of healthcare facilities, Di Giorgio et al. found that many HIV clinics in Kenya, Uganda, and Zambia appear to have considerable untapped capacity to provide care for additional patients. These findings highlight the potential for increasing the efficiency of clinical processes for chronic disease care at the facility level. Important questions for future research are how estimates of comparative technical efficiency across facilities change, when they are adjusted for quality of care and the composition of patients by care complexity. Looking ahead, substantial research investment will be needed to ensure that we do not forgo the opportunity to learn how efficiency changes, as chronic care is becoming increasingly differentiated by patient type and integrated across diseases and health systems functions. Please see related article: http://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-016-0653-z Keywords EfficiencyDifferentiated careIntegrated careChronic diseasesAntiretroviral therapyHIVNon-communicable diseasesissue-copyright-statement© The Author(s) 2016 ==== Body Evolving health systems for chronic care in sub-Saharan Africa The disease burden of chronic non-communicable diseases (NCD) is increasing in sub-Saharan Africa (SSA), while the burden attributable to acute infectious diseases is on the decline [1]. In the midst of this epidemiological transition, increased access to effective antiretroviral therapy (ART) is transforming human immunodeficiency virus (HIV) infection from a disease that cuts life expectancy by decades to a chronic affliction that does not substantially reduce life expectancy [2, 3]. As a result, the number of people needing chronic disease care continues to grow in the region [1, 4], calling into question the capacity of the generally weak health systems in SSA to effectively treat and support all patients needing chronic care [5, 6]. As ART continues to be scaled up with expanded eligibility [7], this strain on health systems in SSA will likely continue to increase over the coming decades [7]. The health system challenges posed by the continuously increasing number of patients with HIV have spurred efforts to assess the efficiency of current models of HIV care as well as to develop new, more efficient models of care. HIV and NCD share many health system challenges in terms of the successful implementation of primary prevention, screening, and early linkage to care; monitoring of therapeutic success; and the need for long-term medication adherence. Thus, research on the efficiency of health systems in delivering HIV care will also provide crucial lessons for health systems to help them successfully address the expected rise in the burden of NCD. This commentary discusses the recent article by Di Giorgio et al. in BMC Medicine [8], and future areas of research that emerge from this work. Untapped capacity in health systems in sub-Saharan Africa The study by Di Giorgio et al. suggests that the health systems in Kenya, Uganda, and Zambia have considerable capacity to deliver HIV care to more patients without adding labor and capital inputs [8]. In fact, the authors estimate that the current systems could deliver care to an additional 459,000 patients with HIV– a 40 % increase in the current HIV patient volume. These findings imply that many healthcare facilities in the study countries have untapped capacity to deal with not only higher volumes of HIV patients, but also increased demand for other types of chronic disease care. Thus far, policy and research in the field of healthcare delivery in SSA has largely been concerned either with inducing demand for healthcare, or, more recently, improving the quality of care on the supply side. The study by Di Giorgio et al. is an important contribution to this field because it points to a research and policy area that has received little attention thus far: supply-side interventions at the level of the healthcare facility to increase the efficiency of care. We see two important methods improvements to increase the robustness of this research: controlling for quality of care and controlling for patient composition when measuring efficiency of care. We also see the emergence of a new focus in health systems research: experiments and quasi-experiments to establish the causal impact of changing models of care on efficiency. Efficiency and quality of care As healthcare facilities increase the quantity of care with a fixed level of inputs, quality of care may deteriorate. For instance, healthcare workers may have less time available to spend per patient, resulting in lower diagnostic accuracy and reduced patient satisfaction. One important area of future research will thus be to carefully assess the degree to which the volume of chronic care patients can be increased without adversely affecting quality of care and patient outcomes. Owing to data constraints, Di Giorgio et al. were only able to include structural indicators of care quality provided by healthcare facilities (e.g., the availability of certain medications). Future research in this area should control for outcome quality of care in the comparison of efficiency of care across facilities, including objective outcome quality (e.g., viral load suppression for HIV care, hemoglobin A1c for diabetes care) and subjective outcome quality (e.g., overall patient satisfaction with care). In addition, time-and-motion studies can provide insights into process aspects of quality of care by enumerating the health worker activities during the patient encounter and the time spent per activity [9]. Patient composition Patients in advanced disease stages or those with comorbidities are likely to demand more human resource inputs (in terms of both health worker time and skill level) and other inputs (e.g., blood tests) than healthier patients. The composition of the patient population – that is, the distribution of patients across ‘complex’ and ‘simple’ patient types – is therefore likely to have profound effects on the quantity of patients any given number of health workers can provide with high-quality care. Because patient composition is likely to vary between healthcare facilities, and at facilities over time, it will be important for future research on chronic care efficiency in SSA to adjust efficiency estimates for patient composition. Indicators of patient complexity will need to be carefully constructed to avoid systematic misclassification of complexity by quality of care. For instance, CD4-cell count is both an indicator of quality of care for patients with HIV, as well as a measure of HIV disease severity and thus patient complexity. While adjustments for patient composition will require additional data collection and research investments, we would argue that these adjustments are essential to accurately determine which healthcare facilities operate at the production possibility frontier and which are less efficient. New models of care Recently, there has been a proliferation of new HIV care models, which provide varying levels of intensity and types of care to different categories of patients. Such models may deliver increased levels of facility-based care to some patients, such as adherence support for those who first present to HIV programs while feeling well [7]. However, for those who are clinically stable on ART, differentiated HIV care models usually aim to reduce patient volumes at healthcare facilities. Examples of such models are less frequent clinic appointments, the delivery of antiretroviral drugs (ARV) to patients’ homes, peer educator-led ARV refill groups, and community ARV distribution points [7, 10]. The finding by Di Giorgio et al. of untapped capacity in the current standard facility-based models of HIV care should be viewed in the context of these emerging models, which often aim to increase the efficiency of HIV care by decongesting healthcare facilities. Many of the new models are still in the development stage and have yet to be evaluated rigorously and at scale [10]. Measuring quality of care will be particularly critical in these evaluations because, in the new models of HIV care, highly qualified or specialized (and thus expensive) healthcare resources – such as nurses – are often replaced by less qualified or specialized (and thus less expensive) resources – such as community health workers. If the quality of HIV care using the new models of HIV care is at least equal to the quality of care delivered using the current models, the new models will be more efficient and should thus become the preferred approach. Another change in HIV care in SSA is the joint delivery of HIV and NCD treatment in integrated models of chronic care [11]. Such integration is likely to affect efficiency of care in complex ways whose net effects are hard to predict and thus need to be empirically determined. From the perspective of the patient, integrated models of chronic care are likely efficiency-enhancing because they create a ‘one-stop shop’ for multiple chronic care needs (such as for HIV, diabetes, hypertension, and depression [12, 13]), lowering the number of healthcare visits each patient has to make and thus reducing average out-of-pocket expenditures and lost time from work to seek care [14]. The potential for such patient-centered efficiency gains is likely to increase over the coming decades as continuously rising ART coverage will increase the co-prevalence of HIV and NCD. This increase will be driven by a new epidemiological transition in SSA through several important pathways [2] (Geldsetzer et al. in press): ART allows patients with HIV to survive into the older ages when NCD become very common [3]; the aging process in people surviving on ART is accelerated by chronic HIV inflammation that is not eliminated by ART [15–17]; and several ARVs have side effects that cause NCD [18]. From the perspective of the health system, the impact of integrating HIV care with other health system functions is less clear. On the one hand, integration of care could reduce clinicians’ efficiency, because it decreases specialization and increases task switching costs, as suggested by classic division labor theory [19–21]. On the other hand, integration could also improve the efficiency of the health system through the joint utilization of fixed factors of production (e.g., clinics) and sharing of health system functions (e.g., monitoring and evaluation systems) [22]. In addition, it has been argued that integration will improve the quality of care through a shift from disease- to more person-centered healthcare [6], resulting in more trusting relationships between patient and health worker, and health workers who are more familiar with a patient’s multiple healthcare needs. Conclusions The increasing demand for chronic diseases is a formidable challenge to health systems in SSA. Finding ways to improve efficiency of chronic disease care while maintaining or even increasing quality of care will be crucial. The empirical literature on this topic is astoundingly sparse; yet, the answers to many of the key questions on the efficiency of novel models for chronic care are far from obvious. Thus, we agree with Davies and colleagues who have recently concluded in a commentary on the future of chronic disease care in SSA that “[t]he lynchpin of a successful effort to grow health systems that can deal with communicable diseases and NCDs equally effectively is research” [23]. Acknowledgements TB is funded by the Wellcome Trust as well as NIH (NICHD R01-HD084233, NIAID R01-AI124389 and R01-AI112339, and NIA P01 AG041710). PG, KO, and TB received funding from the International Initiative for Impact Evaluation (3ie), and PG and TB were also funded by the Clinton Health Access Initiative (CHAI). Authors’ contributions PG, KO and TB jointly conceived and wrote this commentary. All authors read and approved the final manuscript. Competing interests The authors declare they have no competing interests. ==== Refs References 1. Global Burden of Disease Study 2013 Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;386(9995):743–800. 2. Bärnighausen T Welz T Hosegood V Bätzing-Feigenbaum J Tanser F Herbst K Hill C Newell M-L Hiding in the shadows of the HIV epidemic: obesity and hypertension in a rural population with very high HIV prevalence in South Africa J Hum Hypertens 2008 22 236 9 10.1038/sj.jhh.1002308 18046436 3. Hontelez JA de Vlas SJ Baltussen R Newell ML Bakker R Tanser F Lurie M Bärnighausen T The impact of antiretroviral treatment on the age composition of the HIV epidemic in sub-Saharan Africa AIDS 2012 26 Suppl 1 S19 30 10.1097/QAD.0b013e3283558526 22781175 4. Wang H Wolock TM Carter A Nguyen G Kyu HH Gakidou E Hay SI Mills EJ Trickey A Msemburi W Estimates of global, regional, and national incidence, prevalence, and mortality of HIV, 1980–2015: the Global Burden of Disease Study 2015 Lancet HIV 2016 3 8 e361 87 10.1016/S2352-3018(16)30087-X 27470028 5. Ullrich A Ott JJ Vitoria M Martin-Moreno JM Atun R Long-term care of AIDS and non-communicable diseases Lancet 2011 377 9766 639 40 10.1016/S0140-6736(11)60233-X 21334535 6. Atun R Jaffar S Nishtar S Knaul FM Barreto ML Nyirenda M Banatvala N Piot P Improving responsiveness of health systems to non-communicable diseases Lancet 2013 381 9867 690 7 10.1016/S0140-6736(13)60063-X 23410609 7. World Health Organization Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection - Recommendations for a public health approach 2016 2 Geneva World Health Organization 480 8. Di Giorgio L Moses MW Fullman N Wollum A Conner RO Achan J Achoki T Bannon KA Burstein R Dansereau E The potential to expand antiretroviral therapy by improving health facility efficiency: evidence from Kenya, Uganda, and Zambia BMC Med 2016 14 1 108 10.1186/s12916-016-0653-z 27439621 9. Sando D Geldsetzer P Magesa L Lema IA Machumi L Mwanyika-Sando M Li N Spiegelman D Mungure E Siril H Evaluation of a community health worker intervention and the World Health Organization’s Option B versus Option A to improve antenatal care and PMTCT outcomes in Dar es Salaam, Tanzania: study protocol for a cluster-randomized controlled health systems implementation trial Trials 2014 15 359 10.1186/1745-6215-15-359 25224756 10. Lazarus JV Safreed-Harmon K Nicholson J Jaffar S Health service delivery models for the provision of antiretroviral therapy in sub-Saharan Africa: a systematic review Trop Med Int Health 2014 19 10 1198 215 10.1111/tmi.12366 25065882 11. Hope R Kendall T Langer A Bärnighausen T Health systems integration of sexual and reproductive health and HIV services in sub-Saharan Africa: a scoping study J Acquir Immune Defic Syndr 2014 67 Suppl 4 S259 70 10.1097/QAI.0000000000000381 25436826 12. Mills EJ Bärnighausen T Negin J HIV and aging--preparing for the challenges ahead New Engl J Med 2012 366 14 1270 3 10.1056/NEJMp1113643 22475591 13. Negin J Bärnighausen T Lundgren JD Mills EJ Aging with HIV in Africa: the challenges of living longer AIDS 2012 26 Suppl 1 S1 5 10.1097/QAD.0b013e3283560f54 22713477 14. Chimbindi N Bor J Newell ML Tanser F Baltusen R Hontelez J de Vlas S Lurie M Pillay D Bärnighausen T Time and money: the true costs of health care utilization for patients receiving ‘free’ HIV/TB care and treatment in rural KwaZulu-Natal J Acquir Immune Defic Syndr. 2015 70 e52 60 10.1097/QAI.0000000000000728 26371611 15. Guimaraes MM Greco DB Figueiredo SM Foscolo RB Oliveira AR Jr Machado LJ High-sensitivity C-reactive protein levels in HIV-infected patients treated or not with antiretroviral drugs and their correlation with factors related to cardiovascular risk and HIV infection Atherosclerosis 2008 201 2 434 9 10.1016/j.atherosclerosis.2008.02.003 18359028 16. De Luca A de Gaetano DK Colafigli M Cozzi-Lepri A De Curtis A Gori A Sighinolfi L Giacometti A Capobianchi MR D’Avino A The association of high-sensitivity c-reactive protein and other biomarkers with cardiovascular disease in patients treated for HIV: a nested case–control study BMC Infect Dis 2013 13 414 10.1186/1471-2334-13-414 24004495 17. Nou E Lo J Hadigan C Grinspoon SK Pathophysiology and management of cardiovascular disease in patients with HIV Lancet Diabetes Endocrinol 2016 4 7 598 610 10.1016/S2213-8587(15)00388-5 26873066 18. Bavinger C Bendavid E Niehaus K Olshen RA Olkin I Sundaram V Wein N Holodniy M Hou N Owens DK Risk of cardiovascular disease from antiretroviral therapy for HIV: a systematic review PLoS One 2013 8 3 10.1371/journal.pone.0059551 23555704 19. Smith A The wealth of nations 1776 London Methuen & Co., Ltd 20. Chaney T Ossa R Market size, division of labor, and firm productivity J Int Econ 2013 90 1 177 80 10.1016/j.jinteco.2012.11.003 21. Skaugset LM Farrell S Carney M Wolff M Santen SA Perry M Cico SJ Can you multitask? Evidence and limitations of task switching and multitasking in emergency medicine Ann Emerg Med 2016 68 2 189 95 10.1016/j.annemergmed.2015.10.003 26585046 22. Sweeney S Obure CD Maier CB Greener R Dehne K Vassall A Costs and efficiency of integrating HIV/AIDS services with other health services: a systematic review of evidence and experience Sex Transm Infect 2012 88 2 85 99 10.1136/sextrans-2011-050199 22158934 23. Davies J Bennet N Van Epps H HIV and NCDs: the need to build stronger health systems Lancet Diabetes Endocrinol 2016 4 549 50 10.1016/S2213-8587(16)30110-3 27290621

PMC005xxxxxx/PMC5002157.txt

==== Front J Transl MedJ Transl MedJournal of Translational Medicine1479-5876BioMed Central London 99810.1186/s12967-016-0998-2ResearchIntra-articular injection of two different doses of autologous bone marrow mesenchymal stem cells versus hyaluronic acid in the treatment of knee osteoarthritis: multicenter randomized controlled clinical trial (phase I/II) Lamo-Espinosa José M. jlamodeespi@unav.es 1Mora Gonzalo gonzalomora@me.com 1Blanco Juan F. juanfblanco@telefonica.net 23Granero-Moltó Froilán fgranero@unav.es 1345Nuñez-Córdoba Jorge M. jnunezco@unav.es 567Sánchez-Echenique Carmen msancheze@unav.es 1Bondía José M. jbondia@unav.es 8Aquerreta Jesús Dámaso jdaquerret@unav.es 8Andreu Enrique J. andreu@unav.es 34Ornilla Enrique eornilla@unav.es 9Villarón Eva M. emvillaron@saludcastillaleon.es 31012Valentí-Azcárate Andrés avalenti@unav.es 1Sánchez-Guijo Fermín ferminsg@usal.es 31012del Cañizo María Consuelo concarol@usal.es 31012Valentí-Nin Juan Ramón jrvalenti@unav.es 1Prósper Felipe +34 948255400fprosper@unav.es 345111 Department of Orthopaedic Surgery and Traumatology, Clínica Universidad de Navarra, Pamplona, Spain 2 Department of Orthopaedic Surgery and Traumatology, IBSAL-Hospital Universitario de Salamanca, Salamanca, Spain 3 TerCel (Spanish Cell Therapy Network, Spanish National Institute of Health Carlos III), Madrid, Spain 4 Cell Therapy Area, Clínica Universidad de Navarra, Pamplona, Spain 5 Navarra Institute for Health Research (IdiSNA), Pamplona, Spain 6 Division of Biostatistics, Research Support Service, Central Clinical Trials Unit, Clínica Universidad de Navarra, Pamplona, Spain 7 Department of Preventive Medicine and Public Health, Medical School, University of Navarra, Pamplona, Spain 8 Department of Radiology, Clínica Universidad de Navarra, Pamplona, Spain 9 Department of Rheumatology, Clínica Universidad de Navarra, Pamplona, Spain 10 Department of Hematology, IBSAL-Hospital Universitario de Salamanca, Salamanca, Spain 11 Department of Hematology, Clínica Universidad de Navarra, Avenida Pío XII 36, 31009 Pamplona, Navarra Spain 12 Centro en Red de Medicina Regenerativa y Terapia Celular de Castilla y León, Castilla y León, Salamanca, Spain 26 8 2016 26 8 2016 2016 14 1 24622 6 2016 2 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 Mesenchymal stromal cells are a promising option to treat knee osteoarthritis. Their safety and usefulness must be confirmed and the optimal dose established. We tested increasing doses of bone marrow mesenchymal stromal cells (BM-MSCs) in combination with hyaluronic acid in a randomized clinical trial. Materials A phase I/II multicenter randomized clinical trial with active control was conducted. Thirty patients diagnosed with knee OA were randomly assigned to intraarticularly administered hyaluronic acid alone (control), or together with 10 × 106 or 100 × 106 cultured autologous BM-MSCs, and followed up for 12 months. Pain and function were assessed using VAS and WOMAC and by measuring the knee motion range. X-ray and magnetic resonance imaging analyses were performed to analyze joint damage. Results No adverse effects were reported after BM-MSC administration or during follow-up. BM-MSC-administered patients improved according to VAS during all follow-up evaluations and median value (IQR) for control, low-dose and high-dose groups change from 5 (3, 7), 7 (5, 8) and 6 (4, 8) to 4 (3, 5), 2 (1, 3) and 2 (0,4) respectively at 12 months (low-dose vs control group p = 0.005 and high-dose vs control group p < 0.009). BM-MSC-administered patients were also superior according to WOMAC, although improvement in control and low-dose patients could not be significantly sustained beyond 6 months. On the other hand, the BM-MSC high-dose group exhibited an improvement of 16.5 (12, 19) points at 12 months (p < 0.01). Consistent with WOMAC and VAS values, motion ranges remained unaltered in the control group but improved at 12 months with BM-MSCs. X-ray revealed a reduction of the knee joint space width in the control group that was not seen in BM-MSCs high-dose group. MRI (WORMS protocol) showed that joint damage decreased only in the BM-MSC high-dose group, albeit slightly. Conclusions The single intraarticular injection of in vitro expanded autologous BM-MSCs together with HA is a safe and feasible procedure that results in a clinical and functional improvement of knee OA, especially when 100 × 106 cells are administered. These results pave the way for a future phase III clinical trial. Clinical Trials.gov identifier NCT02123368. Nº EudraCT: 2009-017624-72 Electronic supplementary material The online version of this article (doi:10.1186/s12967-016-0998-2) contains supplementary material, which is available to authorized users. Keywords Bone marrow-mesenchymal stromal cellsKnee osteoarthritisNon-surgical managementStem cell therapyMINECO-Instituto de Salud Carlos IIIPI13/01633Granero-Moltó Froilán http://dx.doi.org/10.13039/501100004587Instituto de Salud Carlos IIIRD12/0019/0017RD12/0019/0031del Cañizo María Consuelo Prósper Felipe issue-copyright-statement© The Author(s) 2016 ==== Body Background Osteoarthritis (OA) is a chronic disease involving progressive degeneration of the articular cartilage and subchondral bone, accompanied by synovitis [1]. Due to its avascular nature and the limited self-renewal capacity of chondrocytes, adult articular cartilage presents limited repair capability [2]. Current treatment options for articular cartilage injury and osteoarthritis are aimed to relieve inflammation and pain, but have no effect on the natural progression of the disease [3]. To date, in severe cases of knee OA, knee replacement is the only therapeutic option [4]. During the last two decades focal cartilage defects have been treated using cell therapy and tissue engineering approaches. In this context, autologous chondrocyte implantation (ACI) or matrix-induced autologous chondrocyte (MACI) implantation techniques have been applied with promising results, although the large non-contained cartilage defects found in OA and its own pathogenesis cannot be treated using ACI or MACI [5–7]. The use of intraarticular injections of mesenchymal stromal cells (MSCs) may represent some advantages over chondrocytes in patients with OA. First, because of their ability for self-renewal, the number of cells that can be obtained is increased without cartilage donor site morbidity and with reduced cost [6–8]. Second, MSCs are responsible for the normal turnover and maintenance of adult mesenchymal tissues, including cartilage, and has been suggested that the number of MSCs present in the subchondral bone decreases with age and OA grade, suggesting that such MSCs deficit could prime the degenerative process [9–12]. It has also been proposed that during tissue injury MSCs migrate to participate in the reparative process, giving MSCs a potential therapeutic value when added exogenously [13, 14]. Additionally, cultured MSCs induce in vitro chondrocyte proliferation and extracellular matrix protein synthesis, including aggrecan and type II collagen, which support their critical role in cartilage tissue repair [15, 16]. There is an increasing number of reports on the treatment of OA using MSC, but these are methodologically heterogeneous in dose, cell source, coadjuvants and cell processing methods, which makes it difficult to compare the different studies [17]. In many cases, treatments consist of the administration of bone marrow concentrates as a source of MSCs. However, it is well known that only 0.001 % of the mononuclear cells found in the bone marrow could be considered as MSCs as defined by the ICRS in 2006. Therefore, their number in a bone marrow concentrate is very limited compared to that obtained upon culturing MSCs [18–20]. Only a few studies using MSCs produced by good manufacture practices (GMP) such as advanced cell-therapy products have been reported [21–24]. In addition, there is a need to explore the effect of different cell doses in a randomized way to gain insight into the ideal conditions for knee OA patients to take advantage of MSC therapy. For these reasons, the purpose of this study was to randomly assess the safety, feasibility and efficacy of the intra-articular injection of two different doses of GMP-produced autologous bone marrow MSCs (BM-MSCs) with hyaluronic acid (HA) in patients with knee OA. Methods Participants and study design This is a phase I/II randomized clinical trial with active control conducted between August 2012 and October 2014, involving the Clínica Universidad de Navarra (Pamplona, Spain) and IBSAL-Hospital Universitario de Salamanca (Salamanca, Spain). All the procedures were approved by the Institutional Review Board of Navarra and the Spanish Agency of Medicines and Medical Devices (Nº EudraCT: 2009-017624-72, Clinical Trials.gov identifier: NCT02123368). All participants provided written informed consent. Criteria for eligibility of patients Inclusion criteria were as follows: males and females aged 50–80, diagnosis of knee OA according to American College of Rheumatology criteria, visual analogue scale (VAS) joint pain ≥2.5, Kellgren–Lawrence radiological classification scale ≥2, body mass index between 20 and 35 kg/m [2], and availability to be followed during the study period; exclusion criteria were: previous diagnosis of polyarticular disease, severe mechanical extra-articular deformation (>15° varus/15° valgus), systemic autoimmune rheumatic disease, arthroscopy or intraarticular infiltration in the last 6 months, chronic treatment with immunosuppressive or anticoagulant drugs, corticosteroids treatment in the 3 last months, nonsteroidal anti-inflammatory drugs therapy in the last 15 days, bilateral knee OA requiring treatment in both knees, poorly controlled diabetes mellitus, blood dyscrasias, and allergy to HA or bird proteins. Treatment groups Participants were assigned to comparison groups by an unblinded computer-generated list, based on unrestricted randomization, which was maintained centrally by staff with no clinical involvement in the trial so no center knew the treatment allocation of any patient until the patient had been recruited into the trial. Three groups were created: Control group, constituted by patients who received a single intra-articular injection of 60 mg HA (Hyalone®) in a final volume of 4 ml. Low-dose BM-MSCs group, constituted by patients who received a single intra-articular injection of 10 × 106 autologous cultured BM-MSC in 1.5 ml Ringer’s lactate solution, followed by an intraarticular injection of 4 ml HA. High-dose BM-MSCs group, constituted by patients who received a single intra-articular injection of 100 × 106 autologous cultured BM-MSCs in 3 ml Ringer’s lactate solution, followed by an intraarticular injection of 4 ml HA. Sample size calculation We estimated that a sample size of ten patients per group was required to detect an effect size of 0.6 with a power of 80 %, assuming a balanced allocation to treatment groups, and a 5 % type I error probability. Cell culture BM-MSCs were generated under good manufacturing practice conditions (GMP) with standard operating procedures. Briefly, bone marrow (100 ml) was harvested from the pelvic bone (iliac crest) under sterile conditions. The mononuclear cell fraction was isolated by Ficoll density gradient centrifugation (Ficoll-Paque, GE Healthcare Bio-Sciences AB, Uppsala, Sweden). Cells, ranging between 20 × 106 and 60 × 106, were subsequently seeded in 175 cm2 flasks with growth medium, which consisted of αMEM without ribonucleosides (Gibco, Life Technologies, Carlsbad, CA, USA) supplemented with 5 % platelet lisate, 2 units/ml heparin, penicillin–streptomycin at 1 % (Gibco) and 1 ng/ml human fibroblast growth factor (bFGF) (Sigma-Aldrich, St. Louis, MO, USA). The flasks were maintained in culture at 37 °C in 5 % CO2 atmosphere. The growth medium was changed every 3–4 days. About 10–15 days later, colonies were formed and the cells were split with TrypLE Select™ (Life Technologies) and seeded at 3000–5000 cells/cm2. Once 70–80 % confluence was reached, cells were split again and cultured until they were available at the amounts required to be administered to patients. Finally, cells were harvested with TrypLE Select™, washed three times with PBS and resuspended in Ringer’s lactate buffer (Grifols, Barcelona, Spain) containing 1 % human albumin (Grifols), to be administered within 24 h of harvesting of the cells. Cells were characterized according to ISCT criteria. Cells were then analyzed by flow cytometry (FACSCalibur, BD Biosciences, San José, CA, USA) with the appropriate antibodies (BD Biosciences) to confirm expression of surface markers CD90, CD73 and CD44, as well as absence of CD34 and CD45. Cell injection Cell injection was performed without radiographic guidance through a lateral patellar approach by three different orthopaedic surgeons from both involved centers (Additional file 1: Figure S1), 3–4 weeks after the iliac crest biopsy had been performed. In 90 % of the patients, cells were administered within the first hour after being harvested. For this purpose, a 19 G needle was used in two consecutive intraarticular injections. In the first one, 10 × 106 (low dose) or 100 × 106 (high dose) BM-MSCs were administered in 1.5 and 3 ml doses respectively. Subsequently, 4 ml HA (Hyalone®) were injected using the same via. Outcomes of interest The occurrence of complications and/or adverse effects during the study was registered. In addition, the response to the intra-articular infusion of HA with or without BM-MSCs was assessed using the following procedures: A goniometer-based evaluation of the articular range of motion at baseline i.e. before treatment administration, and 3, 6 and 12 months after treatment. Two scale-based methods Visual Analog Scale (VAS) [25] and the Likert version of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) [26, 27], evaluated at baseline and 3, 6 and 12 months after treatment, to clinically assess pain and function. VAS ranges from 0 (maximum relief, i.e., no pain) to 10 (no relief, i.e., maximal pain). WOMAC comprises three subscores: pain, which includes 5 items; stiffness, with 2 items; and physical function, with 17 items. According to previous literature, patients were considered WOMAC responders when they reported an improvement of 20 % in at least two items together with an improvement of ten points in the overall scale [28]. Rosenberg X-ray projections at baseline and 6 and 12 months afterwards to provide a radiographic assessment of the joint space width. A custom methacrylate patient positioner was used to achieve a comparative view (Additional file 2: Figure S2). A magnetic resonance imaging (MRI) study at baseline and 6 and 12 months after treatment. Two experienced radiologists evaluated MRI images in a blinded manner by assessing the number and location of the lesions, cartilage thickness, signal intensity, and subchondral bone alteration and volume, following the Whole-Organ Magnetic Resonance Imaging Score (WORMS) protocol, in which higher score values indicate more damage [29]. 3T Magnetom TRIO equipment (Siemens, Erlangen, Germany) was used following a protocol which included an axial T1 weighted image (WI) with slice thickness of 5 mm, coronal T1 WI (4 mm), sagittal T1 WI (4 mm), sagittal T2 FS WI (4 mm) and sagittal gradient echo 3D (DESS) (2 mm). Statistics The analyses were performed according to treatment assignment, and all available data from all patients were included in the analyses, following the intention-to-treat principle. Descriptive data summaries are presented as median [interquartile range (IQR)] or percentages. Within each group, the comparison of each clinical and radiographic endpoint between the value obtained at 6 or 12 months and the baseline value, i.e. the one obtained immediately before the administration of the treatment, was performed using the Mann–Whitney U test. Changes in the same end points over time were determined calculating the differences between the measurements collected at the 6 or 12-month follow-up visit and the baseline visit. Subsequently, comparisons between treatment groups were carried out using the Kruskal–Wallis test and the Mann–Whitney U test. All tests were two-tailed. A p value of 0.05 was considered to indicate statistical significance, without adjustment for multiple testing. All analyses were performed using Stata 14 (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP) and IBM SPSS Statistics 20 (IBM Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp). Results Demographics of patients Thirty-two patients were assessed for eligibility, and were consecutively randomized to treatment groups (Fig. 1). Two patients who had been randomly assigned to the control group withdrew consent and were excluded from the trial. All the groups showed similar baseline characteristics of age and body mass index. Patients in the three groups showed an uneven distribution according to the Kellgren–Lawrence scale but without statistical significance (p = 0.585, Table 1).Fig. 1 Study flow diagram. Patients were screened in the two participating centers by using the inclusion and exclusion criteria Table 1 Baseline characteristics of patients Control BM-MSCs Low-dose High-dose N 10 10 10 Age (years) 60.3 (55.1, 61.1) 65.9 (59.5, 70.6) 57.8 (55.0, 60.8) Males, n (%) 7 (70) 4 (40) 8 (80) BMI (kg/m2) 29.6 (26.2, 30.8) 27.1 (24.4, 31.2) 28.5 (25.8, 31.0) Time since OA diagnosis (years) 6 (2, 8) 9 (4, 12) 10 (7, 15) K-L 2, n (%) 4 (40) 1 (10) 3 (30) K-L 3, n (%) 2 (20) 2 (20) 3 (30) K-L 4, n (%) 4 (40) 7 (70) 4 (40) Unless specified, data are presented as median [interquartile range (IQR)]. OA osteoarthritis, K–L Kellgren and Lawrence grading scale of severity of knee OA Safety No serious adverse events or complications derived from the procedures or treatments were noted. There were no clinically important trends in the results of physical examination, vital signs and laboratory tests during the study. Articular pain requiring anti-inflammatory treatment during the first 24 h after infiltration was observed in 1, 3 and 6 patients in the control, low-dose BM-MSC and high-dose BM-MSC groups respectively. All patients recovered completely without sequelae and no treatment group-dependent differences were detected in the dose of required anti-inflammatory drug or in the time that passed until recovery. Clinical assessment of pain and function VAS and WOMAC clinical scores were used in order to obtain the best picture of how patients perceived their own evolution. Evaluations were performed before the administration of treatment and 3, 6 and 12 months afterwards, and the results are summarized in Fig. 2, Additional file 3: Table S1 (VAS) and Table 2 (WOMAC). The patients that were solely given HA did not show changes during follow up in their pain status according to VAS (Fig. 2; Additional file 3: Table S1). Furthermore, although they initially perceived some improvement according to the WOMAC pain and physical function subscores, this perception was not significantly sustained in the long term (Table 2). Inatraarticular delivery of BM-MSCs, specially when used at high dose, enabled patients to perceive an improvement in their perception of pain in their daily activity. On one hand, the VAS score value was significantly reduced upon treatment with low and high BM-MSC doses at all follow-up times (Fig. 2; Additional file 3: Table S1). Furthermore, treatment with 100 × 106 cells was associated with a significant improvement in all WOMAC subscores at 12 months (Table 2). It is important to note that, when the overall WOMAC value at 12 months was subtracted from the baseline value in each patient, the median decrease in the score, i.e. the relief of the symptoms, was notably larger if patients had been treated with BM-MSCs [−6.5 (−19, 4), −14 (−27, 4), and −14 (−15, −8), median (IQR), for control, low-dose and high-dose BM-MSCs groups respectively]. Thus, only the patients who had been treated with BM-MSCs met criteria to be considered WOMAC responders in the long term.Fig. 2 VAS scores along the study. The median values of VAS in the three groups before administration of treatments and 3, 6 and 12 months afterwards are presented. *p < 0.05; **p < 0.01 with respect to the baseline value of the same group Table 2 WOMAC score before administration of treatments and 3, 6 and 12 months afterwards WOMAC Time Control BM-MSCs Low-dose High-dose Pain Baseline 5.5 (5, 6) 7.5 (5, 9) 4.5 (4, 5) 3 months 3 (1, 3)* 3.5 (3, 7) 3 (2, 5) 6 months 2.5 (1, 5)* 3.5 (3, 7) 3.5 (2, 5) 12 months 2 (1, 6) 3.5 (3, 5) 2.5 (2, 4)* Stiffness Baseline 2 (1, 3) 4 (2, 5) 2.5 (2, 4) 3 months 2 (1, 2) 2 (0, 4) 2 (1, 2) 6 months 0.5 (0, 2) 1.5 (1, 3)* 2 (1, 3) 12 months 2 (1, 2) 2 (1, 2)* 2 (1, 2)* Function Baseline 21 (13, 24) 26.5 (23, 32) 19 (12, 25) 3 months 9 (7, 11)* 17.5 (8, 26) 10 (7, 18) 6 months 7.5 (2, 13)* 18 (10, 23) 14.5 (8, 17) 12 months 9.5 (5, 23) 17 (10, 20) 11 (9, 14)* Overall Baseline 29 (19, 38) 37 (32, 42) 28 (16, 34) 3 months 12 (11, 14)* 25.5 (11, 37) 13 (11, 26)* 6 months 10 (4, 20)* 24 (13, 31) 20 (13, 23) 12 months 13.5 (8, 33) 21.5 (15, 26) 16.5 (12, 19)** The values of each one of the three WOMAC subscales as well as the overall WOMAC score at baseline and 3, 6 and 12 months afterwards are presented. Data are the median (IQR) of each group. Function means physical function. *p < 0.05, **p < 0.01 with respect to the baseline value of the same group Effect of treatments on the range of knee motion The knee flexion and extension ranges of motion were significantly improved in the patients who were treated with BM-MSCs and the effect was seen earlier in patients receiving the higher doses of BM-MSC. No improvement was seen in patients receiving HA alone (Fig. 3; Additional file 4: Table S2).Fig. 3 Knee range of motion along the study. The median values expressed in degrees of the goniometric measurements of the knee flexion (top) and extension (bottom) ranges of motion before administration of treatments and 3, 6 and 12 months afterwards are presented. *p < 0.05; **p < 0.01 with respect to the baseline value of the high-dose group. #p < 0.05 with respect to the baseline valued of the low-dose group Radiological and MRI findings The analysis of the knee joint space by X-rays during follow up showed a borderline reduction in the control group (p = 0.05 at 12 months), which was not observed in patients treated with high dose BM-MSC (Table 3; Additional file 5: Table S3). The assessment in the low dose group was not possible because the baseline value was 0. These results suggest that BM-MSC may halt the progressive loss of cartilage observed in patients with OA despite the use of HA.Table 3 X-ray measurement of the evolution of the knee articular interline at 6 and 12 months after the administration of treatments Time Control BM-MSCs Low-dose High-dose 6 months −3 (−6, 0) 0 (−1, 0) 0 (−1, 1) 12 months −4 (−18, 0) 0 (0, 3) 0 (−1, 2) For each group of treatment, variation for knee joint space width, which was measured in mm, was calculated by subtracting, for each patient of the group, the value at 6 or 12 months from the baseline value. Data are presented as the median (IQR) of each group Consistent with the X-Ray results, the analysis of the MRI following the WORMS protocol showed a reduction in the score value during follow up (Table 4). Patients treated with high dose BM-MSCs showed a median improvement of 4 points at 12 months, with an improvement of 22 points in 25 % of patients, while there were no signs of improvement either in the control or in the low BM-MSC group.Table 4 WORMS score before administration of treatments and 6 and 12 months afterwards Time Control BM-MSCs Low-dose High-dose Baseline 79 (41, 94) 75 (64, 107) 60 (53, 84) 6 months 78 (34, 107) 70 (57, 126) 53 (51, 90) 12 months 83 (25, 95) 90 (67, 140) 53 (46, 82) 12 months evolution −0.5 (−16, 15) 2.5 (−3, 25) −4 (−22, 2) The overall WORMS scores at baseline and 6 and 12 months afterwards are presented as the median (IQR) of each group. The evolution within each treatment group at 12 months is also presented, and was calculated by subtracting for each patient the values at 12 months from the corresponding baseline values. Data are the median (IQR) of each group Discussion The interest in the clinical use of MSCs for the treatment of knee OA has recently grown. However, the optimal dose and source of cells, as well as the use of coadjuvants, are not yet established. In the present clinical trial we used two single doses of BM-MSCs, 10 and 100 × 106 cells, coadministered with HA, and compared their effects with the single administration of HA in patients with knee OA. We found that the use of BM-MSCs resulted in a significant relief of pain symptoms in the long term. According to the VAS scores, when BM-MSCs had been administered, an improvement was seen from the earliest evaluation and was maintained until the last one, at 12 months, at which time point the highest effect was observed. Interestingly, this pain reduction was independent of the dose of BM-MSCs administered. On the other hand, no significant changes in VAS were detected in the control group, and the value at 12 months was similar to the one registered before the administration of the treatments. Accordingly, the analysis of the information provided by WOMAC score confirmed that BM-MSCs induced relief of pain symptoms. It is interesting to note that, although treatment with HA alone was able to reduce the WOMAC score during the first 6 months, this improvement was not sustained in the long term, but when patients received BM-MSCs, a significant reduction in WOMAC score was detected at 12 months. In addition, unlike what was observed with VAS, only the high dose of BM-MSCs showed an efficient reduction in the WOMAC score. Furthermore, it is notable that only patients treated with high-dose BM-MSCs met the criteria to be considered WOMAC responders [28]. The effect of MSCs on pain improvement in knee OA is controversial and the literature provides differing accounts. One metaanalysis and a comprehensive review have been recently published on this topic. Xia et al. [17] performed a metaanalysis by managing the results of seven clinical trials, concluding that cell treatments were not able to reduce pain scores. Unfortunately, the heterogeneity in the methodology used in the different studies, with different cell production methods and dosage, precludes these authors from drawing solid conclusions. On the other hand, Rodríguez-Merchán [30] reviewed 25 articles that reported the use of intra-articular injection of MSCs in knee OA, finding that MSCs induce pain relief and functional improvement in three randomized clinical trials which, however, were not comparable to ours methodologically. One of them used bone marrow concentrate, another used peripheral blood progenitor cells, while the third one used cultured autologous BM-MSCs together with a high tibial osteotomy, which is a surgical treatment with a well-known impact on pain relief [31–33]. The number of clinical randomized trials comparing different treatment and dosage is limited. In an interesting study, Orozco et al. [24, 34] reported an improvement in pain and function with the use of a single intra-articular injection of 40 × 106 cultured autologous MSCs in twelve patients. In a more comparable randomized clinical trial, using allogenic MSCs, Vega et al. reported good clinical outcomes in pain control and function when comparing the use of a single intra-articular injection of 40 × 106 cultured allogenic MSCs against a single intraarticular injection of HA [23]. Osteoarthritis is not considered a classical inflammatory arthropathy due to the absence of neutrophils in the synovial fluid and the lack of systemic manifestations of inflammation [35]. However, it is frequently associated with inflammation signs and symptoms such as joint pain, swelling and stiffness, leading to significant functional impairment and disability [36]. The improvement in pain scores together with the mild effect on function and MRI scores suggests that the positive effect of BM-MSCs that we have observed may rely on their paracrine function. In support of this notion, MSC antiinflammatory properties have been correlated with pain reduction elsewhere [37–40]. In addition, the reduction in pain scores may explain the positive changes in flexion and extension. Although such changes are small, it must be noted that a limitation of only a few degrees in flexo-extension may severely compromise the daily functional activity. These improvements together with the findings in the image analyses, suggest that MSC-based therapies may be indicated in asymptomatic patients with mild OA grade, in whom the injected MSCs could be more effective through their paracrine function when a healthier cartilage is still present. The maintenance of the knee joint space width has been related to an appropriate cartilage thickness [41]. Unlike what happened in the patients that were treated with HA only, who experienced a reduction of this space over the time of the study, the space width was preserved when BM-MSC were also administered, even though the results obtained in the patients that had received the low dose must be taken cautiously since the baseline value in 25 % of them was already zero, which precludes suitable follow-up. Nevertheless, a difference could be observed between the high dose and control groups, which did exhibit comparable baseline values. This finding is consistent with MRI observations and is in agreement with previous reports that also investigated the role of cultured MSCs or MSCs embedded in scaffolds in knee OA [23, 24, 42–44]. The required dose of MSCs to treat knee OA efficiently is a topic of active research. Recently Jo et al. [12] performed a pilot study comparing three doses of cultured adipose tissue-derived MSCs (1 × 106, n = 3; 50 × 106, n = 3; and 100 × 106, n = 3). They found a significant reduction in the VAS score only in the high dose group at 6 months, in spite of the small number of patients included. Since results were better with the highest dose, they focused on this in a second phase of the study 100 × 106 (n = 9), with promising results. Our findings also suggest that it is preferable to administer 100 × 106 rather than 10 × 106 cells. However, we have to bear in mind that, despite randomization, the OA degree at recruitment was more severe in the patients who received only 10 × 106 cells, which may obscure our interpretation of this result. It is accepted that OA patients have a MSC deficit that leads to a degenerative process, and the number, in vitro proliferation and differentiation potential of BM-MSCs present in the subchondral bone decreases with age and OA grade [10, 11, 45]. However, we were able to obtain a sufficient amount of BM-MSCs in osteoarthritis patients, regardless of their age or grade of disease [46–49]. We have not identified any problems during the process of production of autologous BM-MSCs, achieving the number of autologous BM-MSCs proposed, even though the mean age of patients was around 60 years. The present study is not exempt from limitations. First, ethical issues prevented us from performing a double-blinded trial. In order to minimize this inconvenience, subjective clinical scores were contrasted with objective measures to minimize bias. In addition, two independent radiologists carried out the MRI analyses in a blinded manner. Second, the relatively short duration of the study prevented us from analyzing the efficiency of the treatments beyond 1 year after the administration of the treatments. Finally, as anticipated, the severe initial condition of a portion of patients who were going to be administered the low dose of cells may have stopped these exerting more beneficial effects. Conclusions Our study shows that the single intraarticular injection of in vitro expanded autologous BM-MSCs together with HA is a safe and feasible procedure that results in a clinical and functional improvement of knee OA, especially when 100 × 106 cells are administered. These results pave the way for a future phase III clinical trial. Additional files 10.1186/s12967-016-0998-2 Pattern of treatment administration. BM-MSCs (bottom right inset) were administered in two consecutive intraarticular injections with a 19 G needle using a lateral patellar approach. 10 × 106 or 100 × 106 cells were injected in 1.5 and 3 ml respectively and subsequently 60 mg hyaluronic acid were administered in 4 ml. Patients randomized to the control group received solely the second injection. 10.1186/s12967-016-0998-2 A–C, methacrylate patient positioner to permit a correct caption of Rosenberg X-ray projections. The X ray tube is placed behind the patient, at the level of the knee and at an angle of 10° with respect to the horizontal in order to evaluate the knee articular width. D, examples of the X-ray images, obtained at baseline and 6 and 12 months afterwards, of the knees of three of the recruited patients are shown. For each patient, images are comparable to each other, which makes it possible to obtain a valid and comparable value of the articular line. 10.1186/s12967-016-0998-2 VAS before administration of treatments and 3, 6 and 12 months afterwards. 10.1186/s12967-016-0998-2 Goniometric measurements of the knee flexion and extension ranges of motion before administration of treatments and 3, 6 and 12 months afterwards. 10.1186/s12967-016-0998-2 X-ray measurement of the knee articular interline before administration of treatments and 6 and 12 months afterwards. Abbreviations αMEMalpha minimum essential medium bFGFfibroblast growth factor BM-MSCsbone marrow mesenchymal stromal cells GMPgood manufacture practices HAhyaluronic acid ISCTInternational Society for Cellular Therapy K-LKellgren and Lawrence scale MRImagnetic resonance imaging MSCsmesenchymal stromal cells OAosteoarthritis VASvisual analogue scale WOMACWestern Ontario and McMaster Universities Osteoarthritis Index WORMSWhole-Organ Magnetic Resonance Imaging Score José M. Lamo-Espinosa and Gonzalo Mora contributed equally to this work Authors’ contributions Study design: JML-E, GM, JM-NC, MCC, FP. Provision of study materials or patients: JML-E, GM, JFB, AV-A, EMV, GS-G, JRV-N, EA, FP. Data collection and assembly: JML-E, GM, EA, JFB, JMN-C, FG-M, CS-E, JMB, JD-A. Obtaining of funding: FG-M, MCC and FP. Drafting manuscript: JML-E, FG-M, FP. JML-E, FG-M, JMN-C, FP take responsibility for the integrity of the data analysis. All authors read and approved the final manuscript. Acknowledgements Not applicable. Competing interests The authors declare that they have no competing interests. Availability of data and materials All the data presented is available upon request. Ethics approval and consent to participate All the procedures were approved by the Institutional Review Board of Navarra and the Spanish Agency of Medicines and Medical Devices. Funding This work has been partially supported by Grants PI13/01633 (MINECO through Instituto de Salud Carlos III to FG-M) and RD12/0019/0017 (to MCC) and RD12/0019/0031 (to FP) from Instituto de Salud Carlos III (red TerCel). EMV is supported by Centro en Red de Medicina Regenerativa y Terapia Celular de Castilla y León, Consejería de Sanidad, Junta de Castilla y León. ==== Refs References 1. Ishiguro N Kojima T Poole AR Mechanism of cartilage destruction in osteoarthritis Nagoya J Med Sci 2002 65 3–4 73 84 12580533 2. Mazor M Lespessailles E Coursier R Daniellou R Best TM Toumi H Mesenchymal stem-cell potential in cartilage repair: an update J Cell Mol Med 2014 18 12 2340 2350 10.1111/jcmm.12378 25353372 3. Simon LS Osteoarthritis Curr Rheumatol Rep 1999 1 1 45 47 10.1007/s11926-999-0024-2 11123014 4. Buckwalter JA Saltzman C Brown T The impact of osteoarthritis: implications for research Clin Orthop Relat Res 2004 427 S6 S15 10.1097/01.blo.0000143938.30681.9d 15480076 5. Steinert AF Ghivizzani SC Rethwilm A Tuan RS Evans CH Noth U Major biological obstacles for persistent cell-based regeneration of articular cartilage Arthritis Res Ther 2007 9 3 213 10.1186/ar2195 17561986 6. Knutsen G Drogset JO Engebretsen L Grontvedt T Isaksen V Ludvigsen TC Roberts S Solheim E Strand T Johansen O A randomized trial comparing autologous chondrocyte implantation with microfracture. Findings at five years J Bone Joint Surg Am 2007 89 10 2105 2112 10.2106/JBJS.G.00003 17908884 7. Brittberg M Lindahl A Nilsson A Ohlsson C Isaksson O Peterson L Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation N Engl J Med 1994 331 14 889 895 10.1056/NEJM199410063311401 8078550 8. Nejadnik H Hui JH Feng Choong EP Tai BC Lee EH Autologous bone marrow-derived mesenchymal stem cells versus autologous chondrocyte implantation: an observational cohort study Am J Sports Med 2010 38 6 1110 1116 10.1177/0363546509359067 20392971 9. Caplan AI Review: mesenchymal stem cells: cell-based reconstructive therapy in orthopedics Tissue Eng 2005 11 7–8 1198 1211 10.1089/ten.2005.11.1198 16144456 10. Caplan AI Adult mesenchymal stem cells for tissue engineering versus regenerative medicine J Cell Physiol 2007 213 2 341 347 10.1002/jcp.21200 17620285 11. Murphy JM Dixon K Beck S Fabian D Feldman A Barry F Reduced chondrogenic and adipogenic activity of mesenchymal stem cells from patients with advanced osteoarthritis Arthritis Rheum 2002 46 3 704 713 10.1002/art.10118 11920406 12. Jo CH Lee YG Shin WH Kim H Chai JW Jeong EC Kim JE Shim H Shin JS Shin IS Ra JC Oh S Yoon KS Intra-articular injection of mesenchymal stem cells for the treatment of osteoarthritis of the knee: a proof-of-concept clinical trial Stem Cells 2014 32 5 1254 1266 10.1002/stem.1634 24449146 13. Campagnoli C Roberts IA Kumar S Bennett PR Bellantuono I Fisk NM Identification of mesenchymal stem/progenitor cells in human first-trimester fetal blood, liver, and bone marrow Blood 2001 98 8 2396 2402 10.1182/blood.V98.8.2396 11588036 14. Karp JM Leng Teo GS Mesenchymal stem cell homing: the devil is in the details Cell Stem Cell 2009 4 3 206 216 10.1016/j.stem.2009.02.001 19265660 15. Horie M Choi H Lee RH Reger RL Ylostalo J Muneta T Sekiya I Prockop DJ Intra-articular injection of human mesenchymal stem cells (MSCs) promote rat meniscal regeneration by being activated to express Indian hedgehog that enhances expression of type II collagen Osteoarthr Cartil 2012 20 10 1197 1207 10.1016/j.joca.2012.06.002 22750747 16. Gupta PK Das AK Chullikana A Majumdar AS Mesenchymal stem cells for cartilage repair in osteoarthritis Stem Cell Res Ther 2012 3 4 25 10.1186/scrt116 22776206 17. Xia P Wang X Lin Q Li X Efficacy of mesenchymal stem cells injection for the management of knee osteoarthritis: a systematic review and meta-analysis Int Orthop 2015 39 12 2363 2372 10.1007/s00264-015-2785-8 25944079 18. Qi Y Feng G Yan W Mesenchymal stem cell-based treatment for cartilage defects in osteoarthritis Mol Biol Rep 2012 39 5 5683 5689 10.1007/s11033-011-1376-z 22183306 19. Dominici M Le Blanc K Mueller I Slaper-Cortenbach I Marini F Krause D Deans R Keating A Prockop D Horwitz E Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement Cytotherapy 2006 8 4 315 317 10.1080/14653240600855905 16923606 20. Simmons PJ Torok-Storb B Identification of stromal cell precursors in human bone marrow by a novel monoclonal antibody, STRO-1 Blood 1991 78 1 55 62 2070060 21. Burger SR Current regulatory issues in cell and tissue therapy Cytotherapy. 2003 5 4 289 298 10.1080/14653240310002324 12944234 22. Fekete N Rojewski MT Furst D Kreja L Ignatius A Dausend J Schrezenmeier H GMP-compliant isolation and large-scale expansion of bone marrow-derived MSC PLoS One 2012 7 8 e43255 10.1371/journal.pone.0043255 22905242 23. Vega A Martin-Ferrero MA Del Canto F Alberca M Garcia V Munar A Orozco L Soler R Fuertes JJ Huguet M Sanchez A Garcia-Sancho J Treatment of knee osteoarthritis with allogeneic bone marrow mesenchymal stem cells: a randomized controlled trial Transplantation 2015 99 8 1681 1690 10.1097/TP.0000000000000678 25822648 24. Orozco L Munar A Soler R Alberca M Soler F Huguet M Sentis J Sanchez A Garcia-Sancho J Treatment of knee osteoarthritis with autologous mesenchymal stem cells: two-year follow-up results Transplantation 2014 97 11 e66 e68 10.1097/TP.0000000000000167 24887752 25. Huskisson EC Measurement of pain Lancet 1974 2 7889 1127 1131 10.1016/S0140-6736(74)90884-8 4139420 26. Bellamy N Outcome measurement in osteoarthritis clinical trials J Rheumatol Suppl 1995 43 49 51 7752137 27. Escobar A Quintana JM Bilbao A Azkarate J Guenaga JI Validation of the Spanish version of the WOMAC questionnaire for patients with hip or knee osteoarthritis. Western Ontario and McMaster Universities Osteoarthritis Index Clin Rheumatol 2002 21 6 466 471 10.1007/s100670200117 12447629 28. Escobar A Gonzalez M Quintana JM Vrotsou K Bilbao A Herrera-Espineira C Garcia-Perez L Aizpuru F Sarasqueta C Patient acceptable symptom state and OMERACT-OARSI set of responder criteria in joint replacement. Identification of cut-off values Osteoarthr Cartil 2012 20 2 87 92 10.1016/j.joca.2011.11.007 22155074 29. Peterfy CG Guermazi A Zaim S Tirman PF Miaux Y White D Kothari M Lu Y Fye K Zhao S Genant HK Whole-organ magnetic resonance imaging score (WORMS) of the knee in osteoarthritis Osteoarthr Cartil 2004 12 3 177 190 10.1016/j.joca.2003.11.003 14972335 30. Rodriguez-Merchan EC Intra-articular injections of mesenchymal stem cells for knee osteoarthritis Am J Orthop 2014 43 12 E282 E291 25490014 31. Wong KL Lee KB Tai BC Law P Lee EH Hui JH Injectable cultured bone marrow-derived mesenchymal stem cells in varus knees with cartilage defects undergoing high tibial osteotomy: a prospective, randomized controlled clinical trial with 2 years follow-up Arthroscopy 2013 29 12 2020 2028 10.1016/j.arthro.2013.09.074 24286801 32. Varma HS Dadarya B Vidyarthi A The new avenues in the management of osteo-arthritis of knee–stem cells J Indian Med Assoc 2010 108 9 583 585 21510531 33. Saw KY Anz A Jee CSY Merican S Ng RC Roohi SA Ragavanaidu K Articular cartilage regeneration with autologous peripheral blood stem cells versus hyaluronic acid: a randomized controlled trial Arthroscopy 2013 29 4 684 694 10.1016/j.arthro.2012.12.008 23380230 34. Orozco L Munar A Soler R Alberca M Soler F Huguet M Sentis J Sanchez A Garcia-Sancho J Treatment of knee osteoarthritis with autologous mesenchymal stem cells: a pilot study Transplantation 2013 95 12 1535 1541 10.1097/TP.0b013e318291a2da 23680930 35. Goldring MB Goldring SR Osteoarthritis J Cell Physiol 2007 213 3 626 634 10.1002/jcp.21258 17786965 36. Felson DT Clinical practice. Osteoarthritis of the knee N Engl J Med 2006 354 8 841 848 10.1056/NEJMcp051726 16495396 37. Uccelli A Moretta L Pistoia V Mesenchymal stem cells in health and disease Nat Rev Immunol 2008 8 9 726 736 10.1038/nri2395 19172693 38. Doorn J Moll G Le Blanc K van Blitterswijk C de Boer J Therapeutic applications of mesenchymal stromal cells: paracrine effects and potential improvements Tissue Eng Part B Rev 2012 18 2 101 115 10.1089/ten.teb.2011.0488 21995703 39. Salgado AJ Reis RL Sousa NJ Gimble JM Adipose tissue derived stem cells secretome: soluble factors and their roles in regenerative medicine Curr Stem Cell Res Ther 2010 5 2 103 110 10.2174/157488810791268564 19941460 40. Abumaree M Al Jumah M Pace RA Kalionis B Immunosuppressive properties of mesenchymal stem cells Stem Cell Rev 2012 8 2 375 392 10.1007/s12015-011-9312-0 21892603 41. Buckland-Wright JC Macfarlane DG Lynch JA Jasani MK Bradshaw CR Joint space width measures cartilage thickness in osteoarthritis of the knee: high resolution plain film and double contrast macroradiographic investigation Ann Rheum Dis 1995 54 4 263 268 10.1136/ard.54.4.263 7763102 42. Centeno CJ Busse D Kisiday J Keohan C Freeman M Karli D Increased knee cartilage volume in degenerative joint disease using percutaneously implanted, autologous mesenchymal stem cells Pain Physician 2008 11 3 343 353 18523506 43. Emadedin M Aghdami N Taghiyar L Fazeli R Moghadasali R Jahangir S Farjad R Baghaban Eslaminejad M Intra-articular injection of autologous mesenchymal stem cells in six patients with knee osteoarthritis Arch Iran Med 2012 15 7 422 428 22724879 44. Kim YS Choi YJ Lee SW Kwon OR Suh DS Heo DB Koh YG Assessment of clinical and MRI outcomes after mesenchymal stem cell implantation in patients with knee osteoarthritis: a prospective study Osteoarthr Cartil 2015 45. Chua KH Zaman Wan Safwani WK Hamid AA Shuhup SK Mohd Haflah NH Mohd Yahaya NH Retropatellar fat pad-derived stem cells from older osteoarthritic patients have lesser differentiation capacity and expression of stemness genes Cytotherapy 2014 16 5 599 611 10.1016/j.jcyt.2013.08.013 24290076 46. Im GI Jung NH Tae SK Chondrogenic differentiation of mesenchymal stem cells isolated from patients in late adulthood: the optimal conditions of growth factors Tissue Eng 2006 12 3 527 536 10.1089/ten.2006.12.527 16579686 47. Dudics V Kunstar A Kovacs J Lakatos T Geher P Gomor B Monostori E Uher F Chondrogenic potential of mesenchymal stem cells from patients with rheumatoid arthritis and osteoarthritis: measurements in a microculture system Cells Tissues Organs 2009 189 5 307 316 10.1159/000140679 18562787 48. Kafienah W Mistry S Dickinson SC Sims TJ Learmonth I Hollander AP Three-dimensional cartilage tissue engineering using adult stem cells from osteoarthritis patients Arthritis Rheum 2007 56 1 177 187 10.1002/art.22285 17195220 49. Scharstuhl A Schewe B Benz K Gaissmaier C Buhring HJ Stoop R Chondrogenic potential of human adult mesenchymal stem cells is independent of age or osteoarthritis etiology Stem Cells 2007 25 12 3244 3251 10.1634/stemcells.2007-0300 17872501

PMC005xxxxxx/PMC5002158.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 131310.1186/s12906-016-1313-8Research ArticleImpact of 2, 3, 5, 4′-tetrahydroxystilbene-2-O-β-D-glucoside on cognitive deficits in animal models of Alzheimer’s disease: a systematic review Sheng Chenxia 12Peng Weijun 1Chen Zeqi 2Cao Yucheng 1Gong Wei 1Xia Zi-an 1Wang Yang 2Su Nanxiang 1Wang Zhe +86-73185295242ericwangzhe@126.com 11 Department of Integrated Traditional Chinese & Western Medicine, The Second Xiangya Hospital, Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011 People’s Republic of China 2 Department of Integrated Traditional Chinese & Western Medicine, Xiangya Hospital, Central South University, Changsha, 410008 China 26 8 2016 26 8 2016 2016 16 1 32017 12 2015 23 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 efficacy of 2, 3, 5, 4′-tetrahydroxystilbene-2-O-β-D-glucoside (TSG) treatment on cognitive decline in individuals with Alzheimer’s disease (AD) has not been investigated. Therefore, we systematically reviewed the effect of TSG on cognitive deficits in a rodent model of AD. Methods We identified eligible studies published from January 1980 to April 2015 by searching seven electronic databases. We assessed the study quality, evaluated the efficacy of TSG treatment, and performed a stratified meta-analysis and meta-regression analysis to assess the influence of study design on TSG efficacy. Results Among a total of 381 publications, 18 fulfilled our inclusion criteria. The overall methodological quality of these studies was poor. The meta-analysis revealed a statistically significant benefit of TSG on acquisition memory (standardized mean difference [SMD] = −1.46 (95 % CI: −1.81 to −1.10, P < 0.0001) and retention memory (SMD =1.93 (95 % CI: 1.40 to 2.46, P < 0.0001) in experimental models of AD. The stratified analysis revealed a significantly higher effect size for both acquisition and retention memory in studies that used mixed sex models and a significantly higher effect size for acquisition memory in studies that used transgenic models. Conclusions Our meta-analysis highlights a significantly better treatment effect in rodent AD models that received TSG that in those that did not. These findings indicate a potential therapeutic role of TSG in AD therapy. However, additional well-designed and detailed experimental studies are needed to evaluate the safety of TSG. Keywords Alzheimer’s diseaseCognitive deficitsSystematic review2, 3, 5, 4′-tetrahydroxystilbene-2-O-β-D-glucosidehttp://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China81373705Wang Zhe http://dx.doi.org/10.13039/501100004735Natural Science Foundation of Hunan Province13JJ3030Wang Zhe the Young Scientists Fund of National Natural Science Foundation of China81603670Peng Weijun Hunan Province Administration of Traditional Chinese Medicine201346Sheng Chenxia issue-copyright-statement© The Author(s) 2016 ==== Body Background Alzheimer’s disease (AD) is a major public health problem and a leading cause of disability [1, 2]. The number of people affected by AD is increasing rapidly worldwide, and more than 35 million people currently have AD. By 2050, the prevalence of AD is expected to quadruple to 1 in 85 people, of which 43 % are expected to need a high level of care. World Alzheimer Report 2015 showed that the total estimated worldwide cost of dementia is $818 billion, and it will reach the trillion dollar mark by 2018 [3]. The clinical characteristics of AD are memory loss and impairment of at least one other cognitive domain [4]. Memory dysfunction is generally the first symptom of AD, and it is generally the most severe cognitive impairment. Mounting evidence indicates that the severity of memory dysfunction correlates strongly with the presence of beta-amyloid plaques and intracellular tau and neocortical neurofibrillary tangles [5, 6]. Despite massive research effort to elucidate the causes and mechanisms underlying AD, including recent advances in our understanding of its molecular pathology, effective treatment remains elusive, and none of the existing drugs are able to halt its progression [7, 8]. Consequently, there is a growing interest in new therapeutic strategies for the treatment of AD [9]. Polygonum multiflorum Thunb (PM) is a traditional Chinese herb that has been used widely as an anti-aging drug in the Orient since ancient times. TSG (2, 3, 5, 4′-tetrahydroxystilbene-2-O-β-D-glucoside), a monomer of stilbene, is one of the main components extracted from the root of PM [10]. TSG can cross the blood–brain barrier and has protective effects on hippocampal synaptic plasticity in vitro [11, 12] and in vivo [13]. Recent studies have also shown that TSG reduces the overexpression of amyloid precursor protein (APP) [14], inhibits reactive oxygen species generation [15], and attenuates cognitive impairment in several animal models of AD, including age-advanced rats [11], APP transgenic mice [16], amyloid-β1–42-injected rats [12], and aluminium-exposed rats [14]. No systematic studies have investigated the effect of TSG on cognition in humans with AD. Thus, in the absence of systematic studies investigating TSG in humans, it is not appropriate to state that studies are needed to confirm the benefits of TSG because there are no findings to confirm. It is more appropriate to state that studies must be conducted to identify a potential benefit. Systematic reviews of animal studies synthesize the available evidence in an unbiased manner to provide evidence for the potential translational value of effective therapeutic interventions in animal models to humans [17], contribute to models of clinically relevant problems, and facilitate decisions regarding the design and conduct of subsequent human clinical trials [18]. Therefore, the aim of the current study was to perform a robust systematic review and meta-analysis of all available experimental evidence concerning the effects of TSG on cognitive impairment in animal models of AD and to provide an evidence-based foundation for future clinical trials. Methods Literature search On April 3, 2015 we searched seven electronic databases (PubMed, Web of Science, MEDLINE, Google Scholar, Embase, CNKI, and Wanfang data). All searches were restricted to literature published between January 1980 and April 2015. The following terms were included in the searches: “Alzheimer’s disease” (or “Alzheimer disease”, “dementia”, “Alzheimer”, “Alzheimers” or “Alzheimer’s”) and “tetrahydroxystilbene glucoside” (or “Polygonum multiflorum Thunb”, “Radix Polygonum Multiform”, “tetrahydroxy stilbene glucoside”, “2, 3, 5, 4′-tetrahydroxystilbene-2-O-β-D-glucoside”, or “TSG”). We limited our search results to animal studies. Additional relevant publications were identified from the reference lists of the resulting research articles and reviews. Bias was prevented by the a priori defined inclusion and exclusion criteria described in Table 1. Two investigators (SC and PW) assessed the titles and abstracts of the studies and obtained copies of the articles describing controlled studies of TSG or its analogues in animal models of AD.Table 1 Criteria for study inclusion/exclusion Inclusion criteria Exclusion criteria (1) TSG were administered. (1) TSG were not administered. (2) Experimental AD was induced in rodents. (2) Other types of animals (e.g., sheep, cats, and dogs) were used. (3) treatment group was treated with TSG, and control group was administered a placebo. (3) Treatment group was administered another neuroprotective agent in addition to TSG. (4) Cognitive function was measured by the MWM, passageway water maze, passive avoidance task, Y maze experiment etc. (4) Treatment group was administered another Chinese Traditional Medicine in addition to TSG. (5) Article was published in English or Chinese language. (5) Only biochemical or physiological outcomes of treatment efficacy were assessed. (6) No control group was used. (7) Duplicate publications or data presented in duplicate by additional publications. Data extraction The following information was extracted from each included study by two investigators: animal species; sex; type of AD model; sample size; dose, method, and timing of TSG administration; main experimental groups; intervention regime (i.e., administration route and number of injections); and cognitive outcome assessments. Any studies that reported effects of TSG on learning and memory abilities using an animal model of AD were included. The cognitive outcomes were assessed by the Morris water maze, passageway water maze, passive avoidance task, and Y maze experiment, among others, which are commonly used to evaluate spatial learning/memory in both mice and rats [19, 20]. The details of the individual study characteristics were extracted from each publication. When a single publication included groups with different TSG doses or different AD models across groups, these data were extracted and considered independent experiments. Because the learning trials to assess memory function were conducted over 5 days, the final test indicates the learning ability of rats/mice [21]. Therefore, we extracted the data for the final time point only when memory function was assessed at a different time point. If any information was missing, then the study investigators attempted to obtain the information from the study authors. If these data were not available, then we excluded the study from the analysis. If the data were presented in graphical form only, then we contacted the authors to request the numerical values. If numerical values could not be obtained, then the numerical values were estimated from the graphs using digital ruler software. Methodological study quality The methodological quality of the studies was assessed based on a checklist of the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES), as previously described, with minor modifications [22]. One point was assigned for written evidence of each of the criteria described in Table 2.Table 2 The CAMARADES quality items Authors & Year ➀ ➁ ➂ ➃ ➄ ➅ ➆ ➇ ➈ ➉ Quality score Zhang et al. 2006 [17] √ √ √ √ √ √ √ 7 Zhou et al. 2012 [13] √ √ √ √ √ √ √ 7 Zhang et al. 2006 [27] √ √ √ √ √ √ 6 Xing et al. 2006 [28] √ √ √ √ √ √ 6 Xie et al. 2005 [29] √ √ √ √ 4 Chu et al. 2005 [30] √ √ √ √ √ √ 6 Huang et al. 2008 [31] √ √ √ √ √ 5 Huang et al. 2008 [32] √ √ √ √ √ 5 Liu et al. 2012 [33] √ √ √ √ √ √ 6 Chu et al. 2004 [34] √ √ √ √ √ 5 Ye et al. 2003 [35] √ √ √ √ √ 5 Ye et al. 2005 [36] √ √ √ √ √ √ 6 Wang et al. 2007 [12] √ √ √ √ √ 6 Luo et al. 2009 [15] √ √ √ √ √ √ √ 7 Hou et al. 2011 [40] √ √ √ √ √ √ √ 7 Luo et al. 2010 [37] √ √ √ √ √ 5 Zhao et al. 2004 [38] √ √ √ √ √ √ √ 7 Luo et al. 2012 [39] √ √ √ √ √ 5 (1) peer reviewed publication; (2) presence of randomization of subjects into treatment groups; (3) assessment of dose–response relationship; (4) blinded assessment of behavioral outcome; (5) monitoring of physiological parameters such as body temperature; (6) calculation of necessary sample size to achieve sufficient power; (7) statement of compliance with animal welfare regulations; (8) avoidance of anesthetic agents with marked intrinsic neuroprotective properties (e.g., ketamine); (9) statement of potential conflict of interests; (10) use of a suitable animal model Although a large number of tools is currently used to assess the quality of animal studies, most instruments assess study quality and internal and external validity simultaneously [23]. No tools that have been identified that are able to assess internal validity alone. Therefore, in addition to the modified CAMARADES checklist, we used another previously described checklist [24, 25] to assess study quality based on the study characteristics, such as the age, species, and sex of the animals used, and the dose and duration of TSG supplementation (Table 3). The quality of all studies was assessed independently by two reviewers (PW and SC).Table 3 The second quality items Study Quality: Zhang et al. 2006 [17] Zhou et al. 2012 [13] Zhang et al. 2006 [27] Xing et al. 2006 [28] Xie et al. 2005 [29] Chu et al. 2005 [30] Huang et al. 2008 [31] Huang et al. 2008 [32] Liu et al. 2012 [33] Chu et al. 2004 [34] Ye et al. 2003 [35] Ye et al. 2005 [36] Wang et al. 2007 [12] Luo et al. 2009 [15] Hou et al. 2011 [40] Luo et al. 2010 [37] Zhao et al. 2004 [38] Luo et al. 2012 [39] Research question specified and clear? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Outcome measures relevant for AD research √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Are the characteristics of study population clear? Species √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Background/generation √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Sex (and distribution) √ √ √ √ √ N √ N √ √ √ √ √ √ √ √ √ √ Age √ N √ √ √ √ N √ N √ √ √ √ N √ N N √ Presence and correct control group? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Where the groups similar at baseline (if not randomized think of weight and sex etc.)? √ √ √ √ N √ √ √ √ √ N √ ? √ √ √ √ √ Is the experiment randomized? √ √ √ √ N √ √ √ √ √ N √ √ √ √ √ √ √ Kind of supplement mentioned (TSG)? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Age when supplementation started mentioned? √ N √ √ √ √ N √ N √ √ √ √ N √ N N √ Duration of supplementation clear and specified? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Amount of TSG mentioned √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Administration route specified √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Is the timing of the supplementation during the day specified and similar in both groups? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Methods used for outcome assessment the same in both groups ? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Drop outs described for each group separately? N N N N N N N N N √ N N N N N N N N Blinded outcome assessment? √ N N N N N N N N N N N N N N N N N Was the outcome assessment randomized across the groups? N N N N N N N N N N N N N N N N N N Total number of animals included in statistical analyses clear? √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Age of sacrificing animals mentioned? √ N √ √ √ √ N √ N √ N √ √ N √ N N √ Quality score (items√) 19 15 18 18 16 17 16 17 16 19 15 18 18 15 15 15 15 18 Statistical analysis The global estimated effect of TSG treatment on cognitive outcomes was calculated using the standardized mean difference (SMD) and 95 % confidence intervals (CI), which is used as a summary statistic in meta-analyses when studies assess the same outcome but measure the outcome in different ways, based on the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions [26]. The SMD is equal to the difference in mean outcomes between groups divided by the standard deviation of outcomes among participants and is reported in units of standard deviation. Negative SMD effect sizes indicate a positive efficacy for acquisition memory, whereas positive SMD effect sizes indicate a positive efficacy for retention memory. Within- and between-study heterogeneity was evaluated using Cochran’s Q-statistic, P < 0.10, to indicate heterogeneity among studies [25]. The statistical heterogeneity across studies was assessed using the I2 statistic, with values of 75, 50, and 25 % representing high, moderate, and low heterogeneity, respectively. A value ≥50 % suggests unacceptable heterogeneity among the studies [27]. A random-effects model was used to pool the SMD when the heterogeneity was significant (I2 ≥ 50 %); otherwise, a fixed effects model was applied. Subgroup analyses were also used to identify associations between relevant study characteristics, such as species, sex, TSG dose, and study quality, as possible sources of heterogeneity. Heterogeneity and the x2 distribution with n-1° of freedom (df), where n equals the number of groups, was used to assess the differences in mean effect sizes between groups. To adjust the values for multiple comparisons, we used Bonferroni’s correction methods (declared significance =1 − (1 − denoted significance)∧(1/number of comparisons)), which was appropriate for the number of analyses conducted [28]. The denoted significance level was set at P < 0.05. The declared P values for this study were 0.0017 for acquisition memory and 0.0037 for retention memory. Finally, meta-regression analyses were conducted to reveal potential sources of heterogeneity in the efficacy of TSG when high heterogeneity was present. The following variables were included in the meta-regression analyses: species, sex, TSG dose, and study quality. To allow for multiple comparisons, the significance was set at P < 0.01. All statistical analyses were performed using the Stata software package (version 13.0) and Review Manager (version 5.3). Results Study inclusion A total of 381 publications were identified, of which 18 met our inclusion criteria [11, 12, 14, 16, 29–42]. Our meta-analysis is based on these 18 studies, which include 39 comparisons of acquisition memory and 15 comparisons of retention memory (Fig. 1).Fig. 1 Flow diagram of the study search process Study characteristics Of the 18 included studies (Table 4), 13 were published in Chinese academic journals and the remainder were published in English. The characteristics of these studies are presented in Table 1. A total of 10 studies used mice (3 Balb/c mice [31, 32, 36], 2 Kunming mice [33, 38], 3 PDAPPV717I transgenic mice [16, 29, 30], and 2 senescence accelerated prone mice/8 [34, 35]), 7 studies used Sprague-Dawley rats [11, 12, 14, 37, 39, 41, 42] and 1 study used Wistar rats [40]. Female animals were used in 3 studies [31, 32, 36], male animals were used in 11 studies [11, 12, 14, 34, 35, 37–42], and both males and females were used in 4 studies [16, 29, 30, 33]. Five studies used a transgenic model [16, 29, 30, 34, 35], 3 studies used a D-galactose-induced model [31–33], 2 studies used a cholinergic damage model [37, 38], 2 studies used an age-advanced model [11, 42], 5 studies used an amyloid-β1-42-injected model [12, 36, 39, 41, 42], 1 study used an aluminium chloride-exposed model [14], and 1 study used a hypercholesterolemia model [40]. To assess learning and memory, 14 studies used the Morris water maze test, and all of these studies used a hidden platform during the probe phase [12, 16, 29–38, 40, 42]. One study used a passageway water maze [11], 1 study used passive avoidance task [14], and 2 studies used a Y maze experiment [39, 41].Table 4 Characteristics of included studies Authors & Year Species & sex (no.) Sex Type of mode Drug (treated/control) Main experimental groups Dose of administration Method/Time of TSG administration Quality score Outcome Zhang et al. 2006 [17] PDAPPV717I transgenic mouse (72) Female & Male Transgenic model TSG/water 1) AD plus water 120 (TSG -L), & 240 (TSG-H) mg/kg body wt Intragastrically/4 months 7 MWM 2) AD plus TSG -L 3) AD plus TSG-H Zhou et al. 2012 [13] Sprague-Dawley rat (n = 12) Male Aβ infused rats TSG/NS 1) AD plus NS 25 mg/kg body wt Intragastrically/4 months 7 MWM 2) AD plus TSG Zhang et al. 2006 [27] PDAPPV717I transgenic mouse (53) Female & Male Transgenic model TSG/water 1) AD plus water 50 mg (TSG -L), 100 mg (TSG -M),& 200 mg (TSG-H) g/kg body wt Intragastrically/4 months 6 MWM 2) AD plus TSG -L 3) AD plus TSG -M 4) AD plus TSG-H Xing et al. 2006 [28] PDAPPV717I transgenic mouse (46) Female & Male Transgenic model TSG/water 1) AD plus water 0.05 (TSG -L), & 0.20 (TSG-H) g/kg body wt Intragastrically/6 months 6 MWM 2) AD plus TSG -L 3) AD plus TSG-H Xie et al. 2005 [29] BALB/c mouse (n = 12) Female D-galactose infused mice TSG/water 1) AD plus water 0.05 g/kg body wt Intragastrically/2 months 4 MWM 2) AD plus TSG Chu et al. 2005 [30] Balb/c mouse (n = 52) Female D-galactose infused mice TSG/water 1) AD plus water 33 mg (TSG-L), 100 mg (TSG-M),& 300 mg (TSG-H) g/kg body wt Intragastrically/2 months 6 MWM 2) AD plus TSG -L 3) AD plus TSG -M 4) AD plus TSG-H Huang et al. 2008 [31] Kunming mouse (n = 40) Female & Male D-galactose Infused mice TSG/water 1) AD plus water 33 mg (TSG-L), 100 mg (TSG-M),& 300 mg (TSG-H) g/kg body wt Intragastrically/2 months 5 MWM 2) AD plus TSG –L 3) AD plus TSG –M 4) AD plus TSG-H Huang et al. 2008 [32] SMAP mouse Male Transgenic model TSG/NS 1) AD plus NS 33 mg (TSG-L), 100 mg (TSG-M),& 300 mg (TSG-H) g/kg body wt Intragastrically/50 days 5 MWM 2) AD plus TSG –L 3) AD plus TSG –M 4) AD plus TSG-H Liu et al. 2012 [33] SMAP mouse Male Transgenic model TSG/NS 1) AD plus NS 33 mg (TSG-L), 100 mg (TSG-M),& 300 mg (TSG-H) g/kg body wt Intragastrically/50 days 6 MWM 2) AD plus TSG –L 3) AD plus TSG –M 4) AD plus TSG-H Chu et al. 2004 [34] Balb/c mouse (n = 52) Female Aβ infused mice TSG/NS 1) AD plus NS 0.1 g/kg body wt Intragastrically/8 weeks 5 MWM 2) AD plus TSG Ye et al. 2003 [35] Sprague-Dawley rat (n = 43) Male ibotenic acid infused rats TSG/NS 1) AD plus water 30 mg (TSG-L), 60 mg (TSG-M),& 120 mg (TSG-H) g/kg body wt Intraperitoneally/1 month 5 MWM 2) AD plus TSG -L 3) AD plus TSG -M 4) AD plus TSG-H Ye et al. 2005 [36] Sprague-Dawley rat (n = 29) Male scopolamineinfused rats TSG/NS 1) AD plus water 33 mg (TSG-L) & 100 mg (TSG-H) g/kg body wt Intraperitoneally/2 months 6 MWM 2) AD plus TSG -L 3) AD plus TSG-H Wang et al. 2007 [12] Sprague-Dawley rat Male Aged rats TSG/water 1) AD plus water 30 mg(TSG-L), 60 mg (TSG-M) g/kg body wt Intragastrically/12 W 6 PWM 2) AD plus TSG -L 3) AD plus TSG-H Luo et al. 2009 [15] Sprague-Dawley rat Male aluminum chloride exposure rats TSG/water 1) AD plus water 4000 mg g/kg body wt Intragastrically/20 W 7 PAT 2) AD plus TSG Hou et al. 2011 [40] Sprague-Dawley rat Male Aged rats TSG/NS 1) AD plus NS 50 mg g/kg body wt Intragastrically/20 W 7 MWM 2) AD plus TSG Luo et al. 2010 [37] Sprague-Dawley rat Male Aβ infused rats TSG/NS 1) AD plus NS 100 mg g/kg body wt Intragastrically/3 W 5 YEM 2) AD plus TSG Zhao et al. 2004 [38] Wistar rat Male Hypercholestero-lemia rats TSG/NS 1) AD plus NS 30 mg (TSG-L), 60 mg (TSG-M),& 120 mg (TSG-H) g/kg body wt Intragastrically/10 W 7 MWM 2) AD plus TSG -L 3) AD plus TSG -M 4) AD plus TSG-H Luo et al. 2012 [39] Sprague-Dawley rat Male Aβ infused rats TSG/NS 1) AD plus NS 50 mg g/kg body wt Intragastrically/3 W 4 YEM 2) AD plus TSG MWM Morris water maze test, PWM Passageway water maze, PAT Passive avoidance task, YEM Y maze experiment Study quality According to the modified CAMARADES checklist, the median quality score for the 18 included studies was poor (5.692; interquartile range: 5–6), with scores ranging from 4 to 7. No study received a score of 0 or 10. Five studies received scores indicating high quality [12, 14, 16, 39, 42]. One study reported monitoring of physiological parameters [12]. One study mentioned allocation concealment [16]. Two studies [31, 37] did not report randomization of animals into treatment groups. Ten studies [16, 29, 30, 32–35, 37, 38, 40] assessed dose-response relationships. Four studies [12, 14, 39, 42] stated no potential conflicts of interest. Unfortunately, no studies described the calculation of the sample size required to achieve sufficient power to detect differences. According to our secondary criteria, the average quality score of the included studies was 16.74, with scores ranging from 15 to 19. Six studies [12, 14, 37, 39, 40, 42] received a score of 15, and two studies received a score of 19 [16, 36]. Six studies did not report the age of the animals [12, 14, 37, 39, 40, 42]. Only one study [16] reported blinded outcome assessments. No studies mentioned any dropouts. No studies mentioned whether the order of the outcome assessments was randomized across groups. Overall efficacy For acquisition memory, the global estimated effect of TSG was −1.46 (95 % CI: −1.81 to −1.10, P < 0.0001), with significant heterogeneity among studies (heterogeneity: x2 = 216.17, df = 38, P < 0.00001, I2 = 82 %; Fig. 2a). For retention memory, the global estimated effect of TSG was 1.93 (95 % CI: 1.40 to 2.46, P < 0.0001), with significant heterogeneity among studies (x2 = 56.97, df = 14, P < 0.0001; I2 = 75 %; Fig. 2b).Fig. 2 Effects of TSG on acquisition memory (a) and retention memory (b). The horizontal lines represent the mean estimated effect sizes and 95 % CIs for each comparison. The vertical grey bars represent the 95 % CIs of the pooled estimated effect sizes Stratified meta-analysis Subgroup analyses were conducted to assess the degree to which the methodological differences between trials may have systematically influenced differences observed in the primary treatment outcomes. The results of the stratified analyses are described in Table 5.Table 5 The results of stratified meta-analysis Subgroups Acquisition memory Retention memory Studies Participants Effect size [95 % CI] Subgroup differences Dtudies Participants Effect size [95 % CI] Subgroup differences Animal species  APP mice 11 349 −1.84 [−2.62, −1.05] P = 0.0002 4 144 2.84 [1.35, 4.32] P < 0.00001  Balb/c mice 5 124 −0.45 [−0.80, −0.09]  Kunming mice 5 98 −0.84 [−1.27, −0.42] 3 60 1.12 [0.56, 1.67]  SMP8 mice 6 120 −0.79 [−1.17, −0.41] 6 120 1.36 [0.95, 1.77]  SD rats 9 208 −2.78 [−4.06, −1.51] 2 48 3.60 [2.63, 4.57]  Wistar rats 3 105 −1.35 [−1.17, −0.41] Sex  Male 20 471 −1.59 [−2.12, −1.06] P < 0.00001 8 168 1.61 [1.17, 2.45] P = 0.65  Female 5 124 −0.45 [−0.80, −0.09]  Female & Male 14 409 −1.68 [−2.31, −1.05] 7 204 2.06 [1.15, 2.98] Model  Untransgenic 22 535 −1.46 [−1.94, −0.98] P = 0.99 5 108 1.98 [0.88,3.08] P = 0.93  Transgenic 17 469 −1.46 [−2.00, −0.91] 10 264 1.92 [1.28, 2.56] Dose  Less 100 mg 21 582 −1.92 [−2.52, −1.32] P = 0.04 8 216 2.23 [1.34, 3.11] P = 0.15  100 mg 11 270 −0.91 [−1.37, −0.45] 3 60 1.25 [0.69, 1.82]  200 mg 2 64 0.92 [−1.46, −0.39]  300 mg 4 88 −1.01 [−1.46, −0.56] 4 96 1.98 [0.92, 3.05] Quality  4 2 60 −2.10 [−5.27, −1.08] P < 0.00001 1 12 3.86 [1.66, 6.05] P = 0.006  5 15 361 −1.26 [−1.64, −0.81] 6 120 1.15 [0.76, 1.55]  6 15 387 −0.63 [−0.83, −0.42] 3 60 1.56 [0.96, 2.16]  7 7 196 −3.82 [−4.41, −3.23] 5 180 2.97 [1.71, 4.22] Table 6 Metaregression analysis to identify sources of bias associated with study characteristics Covariates Coef. Std. err. t P > |t| [95 % conf. interval] (a) Acquisition memory  Quality score −.8748308 .2568104 −3.41 0.002 −1.400067 −.3495946  sex .6720335 .4156478 1.62 0.117 −.1780618 1.522129  dose .1827889 . 2101048 0.87 0.391 −.2469237 .6125015  special .24532 .1328595 1.85 0.075 −.0264082 .5170483  model .8326842 .7477604 1.11 0.294 −.8588674 2.524236 (b) Retention memory  Quality score .4047783 .3156139 1.28 0.224 −.2828853 1.092442  sex .7864006 .291455 2.70 0.019 1.421426 .1513747  dose −.4543919 .3441781 −1.32 0.211 −1.204292 .2955078  special .7864006 .291455 2.70 0.019 .1513747 1.421426 We examined the protective effects of TSG on different rodent species. For both acquisition and retention memory, the effect size was significantly higher in studies that used Sprague-Dawley rats than in studies that used other species (Fig. 3a and b); P = 0.0002 for acquisition memory and P < 0.00001 for retention memory, respectively). The effect size was −2.78 (95 % CI: −4.06 to −1.51) for acquisition memory and 3.60 (95 % CI: 2.63 to 4.57) for retention memory in studies that used Sprague-Dawley rats.Fig. 3 Effect size stratified by animal species for (a) acquisition memory and (b) retention memory, according to animal gender. Effect size stratified by gender for (c) acquisition memory and (d) retention memory stratified by the model method for (e) acquisition memory and (f) retention memory. Grey bands represent the 95 % CIs for the global estimated effect sizes We also examined the effect size of TSG on acquisition memory and retention memory in studies that used male, female, or mixed sex animals. The effect size on acquisition memory was significantly higher in studies that used mixed sex animals than in those that used male or female animals only (x2 = 18.45, df = 2, P < 0.00001; Fig. 3c). The effect size on retention memory was examined in studies that used mixed sex or male animals only because limited data were available from studies with female animals only. The effect size was higher in studies that used mixed sex animals (−2.06, 95 % CI: 1.15 to 2.98) than in those that used male animals only, but this difference was not significant (x2 = 0.21, df = 1, P = 0.65; Fig. 3d). A significant effect size for acquisition memory was observed in both transgenic models (−1.46, 95 % CI: −1.94 to −0.98, P < 0.0001) and non-transgenic models (−1.46, 95 % CI: −2.00 to −0.91, P < 0.0001); no significant difference was observed between models (x2 = 0.00, df = 1, P = 0.99; Fig. 3e). A slightly higher effect size for retention memory was observed in non-transgenic models than in transgenic models, but no significant differences were observed between models (x2 = 0.001, df = 1, P = 0.93; Fig. 3f). Next, we analysed the efficacy of different doses of TSG on cognitive performance. For both acquisition and retention memory, significant beneficial effects were found for all doses of TSG, with a maximum effect at the lowest dose for both acquisition memory (−1.92, 95 % CI: −2.52 to −1.32) and retention memory (2.23, 95 % CI: 1.34 to 3.11). However, no significant differences among doses were detected for either acquisition memory (x2 = 8.48, df = 3, P = 0.04; Fig. 4a) or retention memory (x2 = 3.86, df = 2, P = 0.15; Fig. 4b).Fig. 4 Effect size stratified by the dose of TSG for (a) acquisition memory and (b) retention memory according to animal gender. Effect size stratified by quality score for (c) acquisition memory and (d) retention memory. Rey bands represent the 95 % CIs for the global estimated effect sizes The effect sizes for acquisition and retention memory were also examined relative to the study quality score. For acquisition memory, the effect size was significantly higher in studies with a quality score of 7 (−3.82, 95 % CI: −4.41 to −3.23) than in those with a quality score of 4, 5, or 6 (x2 = 101.37, df = 3, P < 0.00001); Fig. 4c). No significant differences in effect size were observed relative to study quality for retention memory (x2 = 12.51, df = 3, P = 0.006; Fig. 4d); however, the effect size was highest for studies with a quality score of 4 (3.86, 95 % CI: 1.66 to 6.05). Meta-regression analyses A multivariate random-effects regression with species, sex, model, TSG treatment dose, and study quality score was conducted to further explore the heterogeneity among studies regarding acquisition and retention memory. For acquisition memory, the study quality score was a significant source of heterogeneity (P < 0.05). For retention memory, heterogeneity was independent of all tested factors (Table 6). Finally, we analysed the combined data for acquisition and retention memory to determine whether the study quality score was a significant source of heterogeneity. However, the results showed that the study quality score was not a significant source of heterogeneity for the combined data (coef: −.1396835; 95 % CI: −1.896979-1.61761; t: −0.16). Discussion Systematic reviews of animal studies synthesize the available evidence in an unbiased manner to provide evidence for the potential translational value of effective therapeutic interventions in animal models to humans [17], contribute to models of clinically relevant problems, and facilitate decisions regarding the design and conduct of subsequent human clinical trials. Systematic reviews of animal studies synthesize existing evidence in an unbiased manner to facilitate decisions regarding the design and conduct of subsequent human clinical trials [18]. To the best of our knowledge, this is the first systematic review and meta-analysis to examine the efficacy of TSG in animal models of AD. This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses [43] flow diagram (Fig. 1). Although small study effects and statistical heterogeneity were present among the included studies, we found that TSG may improve cognitive outcomes relevant to AD [44]. Subgroup analyses of stratified characteristics were assessed to examine the variation in the effects of the intervention, which would suggest that the stratifying characteristic is a crucial factor for heterogeneity and may affect the treatment efficacy. Based on current guidelines, which recommend at least 10 studies per characteristic to stratify subgroups [45], we were able to conduct subgroup analyses of potential sex and species differences, which revealed a higher effect of TSG on acquisition and retention memory in Sprague-Dawley rats than in other species, and less acquisition and retention memory loss following TSG supplementation in studies that used mixed sex groups than in those that used only male or female animals. In addition, TSG treatment was similarly neuroprotective for acquisition and retention memory in both transgenic and non-transgenic models. However, we also found that the lowest dose of TSG provided the greatest benefit in terms of acquisition and retention memory, which is not consistent with a previously described dose-linear response curve [46]. This finding suggests that effect size has been overstated in studies that used lower doses of TSG. The meta-regression analysis revealed that the heterogeneity was not due to the variables included in the model. Sex, species, animal model, and TSG dose did not affect the heterogeneity of either acquisition or retention memory. The study quality score may explain the heterogeneity in acquisition memory, but the heterogeneity could not be explained by the findings for retention memory or the combined data for acquisition and retention memory. These results may be a consequence of the small sample sizes in those studies and the limited number of studies, reducing the reliability of the analysis. Therefore, we could not conclude that the study quality is dependent on the outcome. There are some limitations to our present meta-analyses. First, our conclusions are limited by the availability of published trials. We did not include unpublished data in our study. Although we attempted to identify all relevant studies from both Western and Eastern countries, all included studies were conducted within China, which may limit generalizations based on our findings. In addition, it has been reported that some Asian countries, including China, publish unusually high proportions of positive results [27, 47]. Of the studies included in this meta-analysis, most did not report negative findings. We conducted an extensive search of unpublished material in an attempt to obtain negative results, but no unpublished negative studies were identified. We cannot exclude the possibility that studies with negative findings remain unpublished because significant positive findings are more likely to be published than non-significant findings. A meta-analysis based on the published literature may overestimate the efficacy of an intervention [48]. Therefore, publication bias may exist in our meta-analysis, although it seems unlikely that the direction or significance of our findings would be modified by unacknowledged trials. Second, we observed significant heterogeneity among the study results. Although we used accepted techniques for the meta-regression analysis to identify factors associated with variability in the benefits of TSG treatment, the statistical power of these analyses was relatively low given the number of available trials. Unfortunately, for retention memory, the adjusted R2 was 29.72 % due to the limited number of studies. In addition, the covariates included in the model could not explain the heterogeneity more than would be expected by chance. Therefore, it was impossible to accurately determine whether the observed heterogeneity was independent of these factors. The presence of heterogeneity highlights the need for caution in interpreting the present findings [49]. Third, no trial exceeded 6 months in duration, which is relatively short given that patients with AD may require treatment with TSG for decades. Long-term treatment may lead to adverse events or persistent or significant disability/incapacity. Furthermore, we focused on only the effect of TSG on cognitive deficits in animal models of AD, largely due to insufficient data regarding the effect of TSG on neuropathological changes (i.e., β-amyloid plaques and neurofibrillary tangles) in AD. Fourth, we assessed the methodological quality of studies in accordance with previously described standards for the preclinical development of neuroprotective drugs, with minor modifications [18]. Overall, we found that the quality of the included studies was poor. Many of the studies failed to report blinded outcome assessments, which is recommended for open-label trials to reduce bias. Patient, clinician, and/or assessor awareness of the treatment assignment may influence outcome reporting or measurements and introduce bias [50]. Moreover, although it is important to judge the efficacy of a new drug or therapy, no study reported sample size calculations [51], which should be calculated during the planning phase of the study to evaluate the accuracy of a priori estimates and assist in the design of future experiments [52]. Furthermore, lower quality studies showed a trend towards better retention memory outcomes. Therefore, the global estimated effect of TSG on cognition may be overstated in low quality studies. In addition, studies that included female animals failed to describe their hormonal cycle, which may influence behaviour, body physiology, and cognitive and learning-related performance, and should be accounted for in the experimental design (e.g., by increasing power and/or balancing the randomization of animals across groups) [24, 53]. Fifth, an increasing number of reports on adverse effects and hepatotoxicity of PMT products have been reported in patients [54]. TSG, the main water-soluble active component of PMT, was considered the major cause of hepatotoxicity [55, 56]. Nonetheless, no study reported any data on the safety and toxicity of TSG perhaps due to the perception that herbal agents are safe because they are natural products and have a long history of use. Along with the medical use and researches of herbal medicines increased, toxicity and safety of those medicinal materials had become the crucial concerns [57]. It is essential to design additional well-designed and detailed experimental studies to evaluate the safety of TSG before human clinical studies and application. Conclusions Despite its limitations, this systematic review and meta-analysis demonstrates that TSG may reduce cognitive deficits in animal models of AD and indicate a potential therapeutic role of TSG in AD therapy. However, additional scientific experimental studies are needed to evaluate the safety of TSG before human clinical studies and application. Abbreviations ADAlzheimer’s disease APPAmyloid precursor protein CAMARADESCollaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies CIConfidence intervals dfDegrees of freedom PMPolygonum multiflorum Thunb SMDStandardized mean difference TSG2, 3, 5, 4′-tetrahydroxystilbene-2-O-β-D-glucoside Acknowledgements We would like to express our gratitude and thanks to American Journal Experts (Durham, North Carolina, USA) for editing of English. Funding This work was supported by the National Science Foundation of China (No. 81373705), the Young Scientists Fund of National Natural Science Foundation of China (81603670), the Natural Science Foundation of the Hunan Province of China (No. 13JJ3030), and the Hunan Province Administration of Traditional Chinese Medicine (No. 201346). Availability of data and materials All the data supporting the findings is contained within the manuscript. Authors’ contributions All authors have contributed to this article. C.S. and W.P. searched the databases, extracted the data, screened trials and appraised the quality of included trials. C.S. reformed the tables and drafted the full text. Z.C., Y.C.,W.G.,S.C.,Z.X.,Y.W. and N.S. were responsible for editing, Z.W. acted as an arbitrator and conceived the article. All authors 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. Ethics approval and consent to participate Not applicable. ==== Refs References 1. G. B. D. Mortality and Causes of Death Collaborators Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013 Lancet 2014 385 9963 117 71 25530442 2. Marchalant Y Baranger K Wenk GL Khrestchatisky M Rivera S Can the benefits of cannabinoid receptor stimulation on neuroinflammation, neurogenesis and memory during normal aging be useful in AD prevention? J Neuroinflammation 2012 9 10 10.1186/1742-2094-9-10 22248015 3. Prince M Wimo A Guerchet M Ali GC Wu YT Prina M World alzheimer report 2015. the global impact of dementia. an analysis of prevalence, incidence, cost and trends 2015 4. Castellani RJ Rolston RK Smith MA Alzheimer disease Dis Mon 2010 56 9 484 546 10.1016/j.disamonth.2010.06.001 20831921 5. Nelson PT Alafuzoff I Bigio EH Bouras C Braak H Cairns NJ Castellani RJ Crain BJ Davies P Del Tredici K Correlation of Alzheimer disease neuropathologic changes with cognitive status: a review of the literature J Neuropathol Exp Neurol 2012 71 5 362 81 10.1097/NEN.0b013e31825018f7 22487856 6. Kortvelyessy P Gukasjan A Sweeny-Reed C Heinze HJ Thurner L Bittner DM Progranulin and amyloid-beta levels: Relationship to neuropsychology in frontotemporal and Alzheimer’s disease J Alzheimers Dis 2015 46 2 375 80 10.3233/JAD-150069 25777512 7. Theofilas P Polichiso L Wang X Lima LC Alho AT Leite RE Suemoto CK Pasqualucci CA Jacob-Filho W Heinsen H A novel approach for integrative studies on neurodegenerative diseases in human brains J Neurosci Methods 2014 226 171 83 10.1016/j.jneumeth.2014.01.030 24503023 8. Wallace TL Ballard TM Pouzet B Riedel WJ Wettstein JG Drug targets for cognitive enhancement in neuropsychiatric disorders Pharmacol Biochem Behav 2011 99 2 130 45 10.1016/j.pbb.2011.03.022 21463652 9. Boada M, Ramos-Fernandez E, Guivernau B, Munoz FJ, Costa M, Ortiz AM, Jorquera JI, Nunez L, Torres M, Paez A. Treatment of Alzheimer disease using combination therapy with plasma exchange and haemapheresis with albumin and intravenous immunoglobulin: Rationale and treatment approach of the AMBAR (Alzheimer Management By Albumin Replacement) study. Neurologia. 2014. doi:10.1016/j.nrl.2014.02.003 10. Qin R Li X Li G Tao L Li Y Sun J Kang X Chen J Protection by tetrahydroxystilbene glucoside against neurotoxicity induced by MPP+: the involvement of PI3K/Akt pathway activation Toxicol Lett 2011 202 1 1 7 10.1016/j.toxlet.2011.01.001 21237255 11. Wang R Tang Y Feng B Ye C Fang L Zhang L Li L Changes in hippocampal synapses and learning-memory abilities in age-increasing rats and effects of tetrahydroxystilbene glucoside in aged rats Neuroscience 2007 149 4 739 46 10.1016/j.neuroscience.2007.07.065 17935895 12. Zhou L Hou Y Yang Q Du X Li M Yuan M Zhou Z Tetrahydroxystilbene glucoside improves the learning and memory of amyloid-beta((1)(−)(4)(2))-injected rats and may be connected to synaptic changes in the hippocampus Can J Physiol Pharmacol 2012 90 11 1446 55 10.1139/y2012-121 23181273 13. Wang T Yang YJ Wu PF Wang W Hu ZL Long LH Xie N Fu H Wang F Chen JG Tetrahydroxystilbene glucoside, a plant-derived cognitive enhancer, promotes hippocampal synaptic plasticity Eur J Pharmacol 2011 650 1 206 14 10.1016/j.ejphar.2010.10.002 20951128 14. Luo HB Yang JS Shi XQ Fu XF Yang QD Tetrahydroxy stilbene glucoside reduces the cognitive impairment and overexpression of amyloid precursor protein induced by aluminum exposure Neurosci Bull 2009 25 6 391 6 10.1007/s12264-009-0601-4 19927176 15. Li X Li Y Chen J Sun J Li X Sun X Kang X Tetrahydroxystilbene glucoside attenuates MPP + −induced apoptosis in PC12 cells by inhibiting ROS generation and modulating JNK activation Neurosci Lett 2010 483 1 1 5 10.1016/j.neulet.2010.07.027 20643188 16. Zhang L Xing Y Ye CF Ai HX Wei HF Li L Learning-memory deficit with aging in APP transgenic mice of Alzheimer’s disease and intervention by using tetrahydroxystilbene glucoside Behav Brain Res 2006 173 2 246 54 10.1016/j.bbr.2006.06.034 16901557 17. van Luijk J Leenaars M Hooijmans C Wever K de Vries R Ritskes-Hoitinga M Towards evidence-based translational research: the pros and cons of conducting systematic reviews of animal studies ALTEX 2013 30 2 256 7 10.14573/altex.2013.2.256 23665810 18. Peng W Yang J Yang B Wang L Xiong XG Liang Q Impact of statins on cognitive deficits in adult male rodents after traumatic brain injury: a systematic review BioMed Res Int 2014 2014 261409 25157352 19. Vorhees CV Williams MT Assessing spatial learning and memory in rodents ILAR J 2014 55 2 310 32 10.1093/ilar/ilu013 25225309 20. Wolf A Bauer B Abner EL Ashkenazy-Frolinger T Hartz AM A comprehensive behavioral test battery to assess learning and memory in 129S6/Tg2576 mice PLoS One 2016 11 1 e0147733 10.1371/journal.pone.0147733 26808326 21. Vorhees CV Williams MT Morris water maze: procedures for assessing spatial and related forms of learning and memory Nat Protoc 2006 1 2 848 58 10.1038/nprot.2006.116 17406317 22. Macleod MR O’Collins T Howells DW Donnan GA Pooling of animal experimental data reveals influence of study design and publication bias Stroke 2004 35 5 1203 8 10.1161/01.STR.0000125719.25853.20 15060322 23. Hooijmans CR Rovers MM de Vries RB Leenaars M Ritskes-Hoitinga M Langendam MW SYRCLE’s risk of bias tool for animal studies BMC Med Res Methodol 2014 14 43 10.1186/1471-2288-14-43 24667063 24. Hooijmans CR Pasker-de Jong PC de Vries RB Ritskes-Hoitinga M The effects of long-term omega-3 fatty acid supplementation on cognition and Alzheimer’s pathology in animal models of Alzheimer’s disease: a systematic review and meta-analysis J Alzheimers Dis 2012 28 1 191 209 22002791 25. Sheng C Peng W Xia ZA Wang Y Chen Z Su N Wang Z The impact of ginsenosides on cognitive deficits in experimental animal studies of Alzheimer’s disease: a systematic review BMC Complement Altern Med 2015 15 1 386 10.1186/s12906-015-0894-y 26497388 26. Zeng X Zhang Y Kwong JS Zhang C Li S Sun F Niu Y Du L The methodological quality assessment tools for preclinical and clinical studies, systematic review and meta-analysis, and clinical practice guideline: a systematic review J Evid Based Med 2015 8 1 2 10 10.1111/jebm.12141 25594108 27. Higgins J, Green S. Cochrane handbook for systematic reviews of interventions, version 5.1.0. Cochrane Collaborations. Chichester, UK: John Wiley & Sons, Ltd. 2011. 28. Vesterinen HM Currie GL Carter S Mee S Watzlawick R Egan KJ Macleod MR Sena ES Systematic review and stratified meta-analysis of the efficacy of RhoA and Rho kinase inhibitors in animal models of ischaemic stroke Syst Rev 2013 2 33 10.1186/2046-4053-2-33 23687965 29. Zhang L Xing Y Ye CF Ai HX Wei HF Li L Learning memory deficit in APP transgenic model of Alzheimer’s disease and intervention by Shen- wu capsule and its effective component Chin J Behav Med Sci 2006 15 3 193 6 30. Xing Y Zhang L Li L Effects of tetrahydroxy-stilbene-glucoside on learning and memory abilities and cerebral β-amyloid expression in APP transgenic mice Chin J New Drugs 2006 15 7 511 3 31. Xie WJ Zhang L Wei HF Chu J Effects of tetrahydroxy-stilbene-glucoside on gene expression in hippocampus in mice with dementia induced by D- galactose Chin J Pharmacol Toxicol 2005 19 1 24 8 32. Chu J Ye CF Li L Zhang L Effects of tetrahydroxystilbene glucoside on learning and memory abilities and neurotrophic factor of brain aging model mice induced by D-galactose China Pharmacy 2005 16 1 510 33. Huang ZS Xie HY Nong S Li T Zhang SQ Effects of TSG on learning and memory abilities and expressions of APP and Aβ in hippocampus of mice with dementia induced by D- galactose J Youjiang Med Coll Nationalities 2008 30 5 713 6 34. Huang R, C., Huang ZS, Li Y, Lai S, Huang J. Ameliorative effects of TSG on learning and memory abilities and free radical scavenging in SMP8 mice. Chin J New Clin Med. 2008;2(9):893–896. 35. Liu L Lai S Li Y Huang ZS Eeffects of tetrahydroxy-stilbene-glucoside on learning and memory abilities in SMP8 mice Mod Chin Med 2012 14 10 12 4 36. Chu J Ye CF Li L Effect of tetrahydroxy-stilbene-glucoside on learning and memory and imflammatory reaction Tradis Chin Drug Res Clin Pharmacol 2004 15 4 235 7 37. Ye CH Zhang L Li B Ai HX Effects of 2,3,5,4-tetrahydroxystilbene-2-O-β-d-glycoside on learning and memory function and excitability in Alzheimer’s disease rat induced by model cholinergic damage Chin J Rehabil Theory Pract 2003 9 10 593 5 38. Ye CH Wei HF Zhang L Zhang L Li L Ameliorative effect of 2,3,5,4′-tetrahydroxystilbene-2-O-β-D-glucoside on learning and memory disorder induced by scopolamine in mice Chin J Clin Rehabil 2005 9 18 190 2 39. Luo HB Yang JS Shi XQ Effect of tetrahydroxy stilbene glycoside on the memory and senile plaque in rat model of Alzheimer disease Prog Mod Biomed 2010 10 05 872 4 40. Zhao L Li YL Zhang L Li L Effect of extract of polygonum multiflorum on learning and memory and blood lipids in rats with hypercholesterolemia Chin Tradit Herb Drugs 2004 35 11 1277 9 41. Luo HB Shi XQ Yang JS Fu XF Yang QD Expression of autophagy-associated protein induced by p-amyloid and effect of tetrahydroxy stilbene glucoside on it in hippocampus Chin J Neurol 2012 45 2 96 101 42. Hou Y Yang Q Zhou L Du X Li M Yuan M Zhou Z Li Z Tetrahydroxystilbene glucoside improves learning and (or) memory ability of aged rats and may be connected to the APP pathway Can J Physiol Pharmacol 2011 89 11 801 9 22032649 43. Stovold E Beecher D Foxlee R Noel-Storr A Study flow diagrams in Cochrane systematic review updates: an adapted PRISMA flow diagram Syst Rev 2014 3 54 10.1186/2046-4053-3-54 24886533 44. Thompson PA Wright DE Counsell CE Zajicek J Statistical analysis, trial design and duration in Alzheimer’s disease clinical trials: a review Int Psychogeriatr 2012 24 5 689 97 10.1017/S1041610211001116 21910950 45. Amos T Stein DJ Ipser JC Pharmacological interventions for preventing post-traumatic stress disorder (PTSD) Cochrane Database Syst Rev 2014 7 CD006239 25001071 46. Dubois B Zaim M Touchon J Vellas B Robert P Murphy MF Pujadas-Navines F Rainer M Soininen H Riordan HJ Effect of six months of treatment with V0191 in patients with suspected prodromal Alzheimer’s disease J Alzheimers Dis 2012 29 3 527 35 22330824 47. Mimouni M Krauthammer M Gershoni A Mimouni F Nesher R Positive results bias and impact factor in ophthalmology Curr Eye Res 2014 40 8 858 61 10.3109/02713683.2014.957777 25198281 48. Schmucker C Bluemle A Briel M Portalupi S Lang B Motschall E Schwarzer G Bassler D Mueller KF von Elm E A protocol for a systematic review on the impact of unpublished studies and studies published in the gray literature in meta-analyses Syst Rev 2013 2 24 10.1186/2046-4053-2-24 23634657 49. Gagnier JJ Morgenstern H Altman DG Berlin J Chang S McCulloch P Sun X Moher D Ann Arbor Clinical Heterogeneity Consensus Group Consensus-based recommendations for investigating clinical heterogeneity in systematic reviews BMC Med Res Methodol 2013 13 106 10.1186/1471-2288-13-106 24004523 50. Kahan BC Cro S Dore CJ Bratton DJ Rehal S Maskell NA Rahman N Jairath V Reducing bias in open-label trials where blinded outcome assessment is not feasible: strategies from two randomised trials Trials 2014 15 456 10.1186/1745-6215-15-456 25416527 51. Gaye A Burton TW Burton PR ESPRESSO: taking into account assessment errors on outcome and exposures in power analysis for association studies Bioinformatics 2015 31 16 2691 6 10.1093/bioinformatics/btv219 25908791 52. Rutterford C Taljaard M Dixon S Copas A Eldridge S Reporting and methodological quality of sample size calculations in cluster randomized trials could be improved: a review J Clin Epidemiol 2014 68 6 716 23 10.1016/j.jclinepi.2014.10.006 25523375 53. Robertson DA Savva GM Kenny RA Frailty and cognitive impairment--a review of the evidence and causal mechanisms Ageing Res Rev 2013 12 4 840 51 10.1016/j.arr.2013.06.004 23831959 54. Lei X Chen J Ren J Li Y Zhai J Mu W Zhang L Zheng W Tian G Shang H Liver damage associated with Polygonum multiflorum Thunb.: a systematic review of case reports and case series Evid Based Complement Alternat Med 2015 2015 459749 10.1155/2015/459749 25648693 55. Yu J Xie J Mao XJ Wang MJ Li N Wang J Zhaori GT Zhao RH Hepatoxicity of major constituents and extractions of Radix Polygoni Multiflori and Radix Polygoni Multiflori Praeparata J Ethnopharmacol 2011 137 3 1291 9 10.1016/j.jep.2011.07.055 21840387 56. Bounda GA Feng YU Review of clinical studies of Polygonum multiflorum Thunb. and its isolated bioactive compounds Pharmacognosy Res 2015 7 3 225 36 10.4103/0974-8490.157957 26130933 57. Kale OE Awodele O Safety evaluation of Bon-sante cleanser(R) polyherbal in male Wistar rats BMC Complement Altern Med 2016 16 1 188 10.1186/s12906-016-1188-8 27387026

PMC005xxxxxx/PMC5002159.txt

==== Front BMC BioinformaticsBMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 120110.1186/s12859-016-1201-8Research ArticleIdentification of DNA-binding proteins using multi-features fusion and binary firefly optimization algorithm Zhang Jian 1Gao Bo 1Chai Haiting 1Ma Zhiqiang 1Yang Guifu guifuyang.nenu@gmail.com 121 School of Computer Science and Information Technology, Northeast Normal University, Changchun, 130117 People’s Republic of China 2 Office of Informatization Management and Planning, Northeast Normal University, Changchun, 130117 People’s Republic of China 26 8 2016 26 8 2016 2016 17 1 32321 1 2016 24 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 DNA-binding proteins (DBPs) play fundamental roles in many biological processes. Therefore, the developing of effective computational tools for identifying DBPs is becoming highly desirable. Results In this study, we proposed an accurate method for the prediction of DBPs. Firstly, we focused on the challenge of improving DBP prediction accuracy with information solely from the sequence. Secondly, we used multiple informative features to encode the protein. These features included evolutionary conservation profile, secondary structure motifs, and physicochemical properties. Thirdly, we introduced a novel improved Binary Firefly Algorithm (BFA) to remove redundant or noisy features as well as select optimal parameters for the classifier. The experimental results of our predictor on two benchmark datasets outperformed many state-of-the-art predictors, which revealed the effectiveness of our method. The promising prediction performance on a new-compiled independent testing dataset from PDB and a large-scale dataset from UniProt proved the good generalization ability of our method. In addition, the BFA forged in this research would be of great potential in practical applications in optimization fields, especially in feature selection problems. Conclusions A highly accurate method was proposed for the identification of DBPs. A user-friendly web-server named iDbP (identification of DNA-binding Proteins) was constructed and provided for academic use. Electronic supplementary material The online version of this article (doi:10.1186/s12859-016-1201-8) contains supplementary material, which is available to authorized users. Keywords DNA-binding proteinsBinary firefly algorithmFeature selectionParameters optimizationFundamental Research Funds for the Central Universities14ZZ2240Zhang Jian issue-copyright-statement© The Author(s) 2016 ==== Body Background DNA-binding proteins (DBPs) are fundamental in many biological processes, such as recognition of specific nucleotide sequence, regulation of gene, transcription and translation, and DNA replication and repair [1, 2]. Thus, it is highly desirable to develop effective DBP identification methods. Traditionally, experimental techniques, which include filter binding assays [3], X-ray crystallography [4] and genetic analysis [5], are used to identify DBPs. Although these techniques can produce detailed information and provide confident assertion of the DBPs, they are both expensive and time-consuming. This spurred the development of computational methods to tackle this problem. These computational methods can be divided into two categories: structure-based methods [6–8] and sequence-based methods [9–15]. Many of the early methods are structure based. Gao et al. [6] developed a knowledge-based method named DNA-binding Domain Hunter for identifying DBPs and associated binding sites using structural comparison. Zhao et al. [7] proposed a template-based prediction method by employing both structural similarity and binding affinity. Nimrod et al. [8] recruited random forests to identify DBPs by detecting evolutionarily conserved regions and using electrostatic features. However, the number of proteins with well annotation and good resolution structure are very limited. The structure-based methods may break down when homogeneous structures of a query protein is not available. Hence, many sequence-based methods had been proposed to deal with this problem. Kumar et al. [9] utilized various SVM modules and evolutionary information to forge the DNA-binder method. Kumar et al. [10] employed random forest to predict DBPs. Lin et al. [11] proposed the iDNA-Prot predictor by incorporating the features into the general form of pseudo amino acid composition that were extracted from protein sequence via the grey model and adopting the random forest operation engine. Song et al. [12] and Xu et al. [13] both applied the ensemble learning technique combined with hybrid features to predict DBPs. Zou et al. [14] conducted a comprehensive feature analysis of four categories of protein properties and three different feature transformation methods to find an optimal prediction model. Lou et al. [15] predicted DBPs by performing feature ranking with random forest and feature selection with forward best-first strategy. The features comprised properties from primary sequence, predicted structures and sequence alignment. Although many efforts were put on the computational identification of DBPs, the prediction performance was still far from satisfactory. There are some possible reasons: (i) structure-based methods can provide reliable results in recognizing specific proteins. However, the insufficiency in known DBP structures leads to limited applications of these methods. Sequence-based methods are featured by their widely application, while the performance of these predictors are usually not as good as expected; (ii) the complexity of DBPs. The DBPs span over many protein families from enzymes to transcription factors [16], which makes it very difficult to describe DBPs discriminatively using mathematical models; (iii) A common approach to describe a protein in DBP prediction is by forming a feature vector, but the redundancy and contradiction among these features may seriously deteriorate the predication and generalization ability of the model. In light of the aforementioned problems, we proposed a novel sequence-based predictor, named iDbP (identification of DNA-binding Proteins), to identify DBPs in this study. Firstly, instead of developing a narrow-application structured-based method, we focused on the challenge of sequenced-based methods. Secondly, a number of discriminative features, including evolutionary conservation, secondary structure motifs and physicochemical properties, were constructed to encode the proteins. These informative features have been proved to be associated with DNA binding interactions. Thirdly, a novel improved binary firefly algorithm (BFA) was introduced to remove redundant and noisy features as well as select optimal parameters for the classifier. In the proposed BFA, we used normalized Hamming distance to calculate attractiveness for fireflies, which greatly improved the converging rate. We also added a dynamic mutation operator to increase the diversity of fireflies. Based on the effective BFA, our predictor produced promising performance on the main dataset and two benchmark datasets. Tests on an independent testing dataset collected from PDB and a large-scale DBP dataset collected from UniProt database demonstrated the good generalization ability of iDbP. Methods Datasets In this study, experimentally verified DBPs were collected from the Protein Data Bank (PDB, http://www.rcsb.org) by specifying keyword “DNA binding protein” and release date “before 2015-05-01” through “Advanced Search”, and 1248 sequences were obtained. Then, these sequences were pre-processed through the following procedures: (1) Sequences which contained unknown residues were discarded. (2) Sequences with less than 50 amino acid residues or belonged to fragments were removed [17]. (3) Sequences with multi-bindings were removed to avoid other influences. (4) Sequence similarity among the dataset was reduced to less than 30 % by using PISCES [18]. As a result, 455 experimentally verified DBPs were obtained as positive samples. Similarly, 455 experimentally verified non-binding proteins were also extracted from PDB with “Does not contain: DNA binding protein” as key words with less than 30 % identity. Finally, a main dataset was obtained by combining the 455 DBPs and 455 non-DBPs. This main dataset was used to find the optimal feature subset and train the iDbP prediction model. To construct the training dataset, 355 sequences were randomly picked from positive and negative samples of the main dataset, respectively. The remaining positive and negative samples were used for testing. In order to ensure unbiased and objective results, the process of under-sampling was performed 20 times. The final performance was the average prediction results of 20 experiments on different training and testing datasets. To evaluate the effectiveness of the proposed method as well as to perform fair comparisons with previous methods [9–15], two benchmark training and testing datasets were adopted: (i) PDB594 and PDB186 [15]. The training dataset PDB594 contained 297 DBPs and 297 non-DBPs, and the testing dataset PDB186 contained 93 DBPs and 93 non-DBPs. Both PDB594 and PDB186 shared sequence similarity of less than 25 %; (ii) DNAdset and DNAiset [14]. DNAdset included 231 DBPs and 231 non-DBPs, and DNAiset contained 80 DBPs and 192 non-DBPs. The sequence similarity in DNAdset and DNAiset was less than 30 %. In real life, the number of DBPs is much less than that of non-DBPs. To further test the generalization ability of our method, a new-compiled independent testing dataset (named DBP189) was introduced in this work. All the predictors that we compared with in this research were built before May 2015. Therefore, proteins released in PBD after May 2015 would be less likely to be used to train these models. DBP189 contained 21 DBPs and 167 non-DBPs, which were deposited in PDB between 2015-05-01 and 2016-05-01. None of these proteins shared more than 30 % sequence similarity with the main dataset. The main dataset and DBP189 were provided in Additional file 1. Feature vector Evolutionary conservation profile Highly conserved regions are often required for basic cellular function, stability or reproduction. Thus, evolutionary conservation analysis are often indicative of structural or functional importance [19, 20]. The position specific scoring matrix (PSSM), which carries evolutionary information of proteins, was widely used in various bioinformatics researches. In this study, the PSSM of each protein was generated by using PSI-BLAST [21] to search against the non-redundant database (ftp://ftp.ncbi.nlm.nih.gov/blast/db/nr.tar.gz) through 3 iterations with E-value of 0.0001. A L × 20 PSSM was generated for each protein, where L was the length of the sequence. 1 PSSM=E1,1E1,2⋯E1,20E2,1E2,2⋯E2,20⋮⋮⋯⋮EL,1EL,2⋯EL,20 Each score in PSSM represents whether the related substitution exceed or beneath expected frequency, and indicates whether this substitution would be favored in the process of evolution. Here, these preferences are statistical classified and analyzed by using the following formula: 2 Pm,n=∑m=1LEm,n×δδ=1,Rm=anδ=0,Rm≠an where Rm indicates the m-th (mϵ{1, 2, …, L}) residue in the protein sequence, and an (nϵ{1, 2, …, 20}) indicates the type of amino acid. To eliminate the influences of sequence length, Pm,n is normalized into the [0, 1] interval by using logistic function: 3 ERi→ai=11−e−Pm,n Finally, feature vector ERi→ai|R∈1L,i∈12…20 was generated to construct the features of evolutionary conservation profile. Secondary structure motifs Secondary structure plays an important role in the function of DBPs [22]. Many DBPs show obvious preference of certain secondary structure motifs, such as helix-turn-helix and coil-helix-coil. These structures are usually solvent exposed and hydrophilic, which grant high probabilities in interaction with DNA segments [23]. Shown in Fig. 1 are the examples of DBP complexes. The secondary structure motifs repeat regularly in DBPs, and this phenomenon could be utilized to discriminate DBPs from non-binding proteins. Figure 2 shows the distributions of the secondary structure motifs on the main dataset. The over-expression of “CXC”, “HCX” and “ECX” confirms the experimental observation of enrichments of a series of helices or strands in DBPs.Fig. 1 An example that illustrates the preferences of certain secondary structure motifs of a protein complex. Panel (a) is a TATA-binding protein (PDB ID: 1AIS_A). The binding surface is composed of strands (red) while the outer region is composed of helices (green). The general secondary structure pattern of this protein is strand-helix-strand-helix-strand-helix-strand-helix. Panel (b) is a transcription initialization protein (PDB ID: 1AIS_B) that is mainly composed of helices (green) and turns (blue) Fig. 2 The distribution of secondary structure motifs To obtain secondary structure motifs, firstly, the predicted secondary structure for each residue was calculated as a probability matrix using PSIPRED [24] (Eq. (4)). 4 ssprobMarix=P1→HP1→EP1→CP2→HP2→EP2→C⋮⋮⋮PL→HPL→EPL→C where Pi → H/E/C (iϵ{1, 2, …, L}) is the probability of the i-th residue to be part of a helix (H), strand (E) or coil (C). Next, max(Pi → H/E/C) for each position would be selected as the corresponding secondary structure, and secondary structure segments were generated to represent the secondary structure distribution for the protein. Then, the secondary structure motifs were obtained from the segments: 5 ssmotif=∑segαsegβsegγ where segα/β/γ indicates continuous secondary structure segments of the same type and α, β, γ ∈ {H, E, C}. Finally, a protein was encoded by a 12-dimentional feature vector. Physicochemical properties Physiochemical properties reveal macroscopic phenomena among atoms and molecules such as motions, energy, force and dynamics [25]. For instance, Surendra et al. [26] pointed out that hydrophobic and polar residues contributed the bonds across the interfaces and binding residues were strongly correlated with exposed surface area. Solvation free energy [27] and transfer free energy [28], which helped to form small paths, were vital free energy to the hot spots. In addition, graph shape also played an important role in deciding the functional sites on the protein surface [29]. In this study, fourteen physiochemical properties, namely net charge [30], hydrophobicity [31], hydrophilicity [27], polarity [32], polarizability [33], solvation free energy [27], graph shape index [34], transfer free energy [28], amino acid composition [35], correlation coefficient in regression analysis [36], residue accessible surface area [37], partition coefficient [38], entropy of formation [39], and pKa values of side chain [40], were collected and used. In this encoding scheme, each property were first calculated by taking the sum of its value over the residues on the whole sequence. Then, the summarized value of each property was divided by the length of the sequence [41]. Support vector machine Support vector machine (SVM) is a machine learning technique derived from statistical learning theory first proposed by Vapnik [42]. It was successfully applied in many bioinformatics problems and yielded promising results. In this study, we utilized the LIBSVM toolset [43] and chose Radial Basis Function (RBF) as the kernel function. Two parameters c and γ of SVM were optimized using BFA. All feature descriptors were normalized into the [0, 1] interval by using logistic function. The proposed binary firefly algorithm Continuous firefly algorithm The continuous Firefly Algorithm (FA) is a swarm-intelligence and meta-heuristic optimization algorithm developed by Xin-She Yang in 2007 [44]. FA is based on the idealized behavior of the flashing characteristics of the fireflies. It is featured by its efficiency as well as robustness. As a novel meta-heuristic algorithm, FA has been proved to be able to find almost optima in continuous problems [45]. In essence, the idea of FA can be abstracted into the following three rules [46]:(i) Every firefly has its own lightness and could be attracted by other fireflies; (ii) The brightness and distance determine the attractiveness. That is, a brighter firefly will always attract its adjacent less bright ones. The attractiveness will decline if the distance between two fireflies increases. If a firefly cannot find a brighter firefly within the designated distance, it will make random movements; (iii) The brightness of a firefly is referred as light intensity (I), which is defined as: 6 I=Ffx,β where f(x) is the objective function. The attractiveness β is proportional to I, and is defined as: 7 β=β0e−γr2 where β0 is the attractiveness at r = 0; γ denotes the light absorption coefficient; and r represents the distance between any two fireflies. The movement of a firefly xi attracted to another firefly xj is defined as: 8 xi=xi+βxj−xi+αεi where α is the randomization parameter, and εi is an element of a vector drawn from random Gaussian or uniform distributions. Binary firefly algorithm The original FA is designed for continuous problems, which means that the outcome of the objective function (i.e. the brightness of a firefly) must lie in a continuous interval. Recently, several BFA were developed to solve discrete problems, such as scheduling, timetabling and combination. Compared with the original FA, BFA obeyed similar fundamental principles while redefined distance, attractiveness, or movement of the firefly [47–49]. Palit et al. [47] applied BFA to discover the plaintext from the cipher text. Sayadi et al. [48] defined a new firefly position and applied BFA to manufacture cell formation. Poursalehi et al. [49] introduced a new form of movement of fireflies to global best in each iteration, and applied BFA on fuel reload design of nuclear reactors. In this study, a novel improved BFA was proposed for feature selection as well as parameter optimization. The feature selection task is a typical combination problem in essence. That is, to select an optimal combination of features from a given feature space. By using this optimal subset, the machine learning algorithm could produce the best predictive performance. Every feature must be either in or not in this subset. Theoretically, for an n-dimensional feature space, there will be 2n possible solutions (NP-hard problem). Empirically, meta-heuristic algorithms will perform better than traditional filter or wrapper methods [50]. In BFA, every firefly represents a subset of the feature space and a group of parameters (i.e., a possible solution for the problem). The effectiveness of BFA is determined by two factors: the ability to converge to the potential global optimum rapidly and the capability of jumping out of local optima. In this work, normalized Hamming distance was used to calculate attractiveness and improve converging rate in feature selection; dynamic mutation operator was introduced to increase the diversity of fireflies. The pseudo code of BFA is provided in Algorithm 1. Firefly representation In BFA, a binary string is used to encode a firefly. Every element in the string is either 0 or 1, the length and interpretation for the string are both problem specific. That is, a firefly X is defined as the following: 9 X=x1x2x3…xnwherexi∈01 Figure 3 shows an instance of the definition of a firefly X with a length of n. The string is divided into three parts. The first part (t elements) and second part (t elements) are used to represent the values of parameters c and γ of SVM, respectively. The third part represents the features. Its length w is the same as the dimension of the feature space. In this part, 1 denotes the corresponding feature is selected, and 0 indicates the opposite.Fig. 3 The coding scheme for a firefly b. The attractiveness of a firefly Similar to FA, a firefly in BFA is also attracted by brighter fireflies. However, the attractiveness is not only determined by the brightness but also greatly affected by the similarity between fireflies. In BFA, the attractiveness β between a pair of fireflies is defined as β=β0e−γr2. Here, γ controls the impact of β in the movement function; r determines the stride of the firefly movement. For two fireflies Xi and Xj, r is defined based on the similarity ratio of the two fireflies (or the normalized Hamming distance of two vectors) as follows: 10 r=1−∑k=1nXik⊕Xjkn where ⊕ denotes the XOR operation, n is the length of X. Mathematically, the less identical bits two fireflies share, the greater stride a firefly would take and the more likely it would move towards the brighter one. β is the probability of a hetero-bit in the moving firefly changes to the corresponding bit in the brighter firefly (0 → 1 or 1 → 0). Compared with Cartesian distance and Euclidean distance, the normalized Hamming distance performs best in keeping good feature as well as removing bad ones, and also made the algorithm converge fast. Figure 4 demonstrates an example of calculating parameter r.Fig. 4 An Example of calculating parameter r. Firefly X = {1 0 0 1 1 1 1 0 0 0}, Firefly Y = {1 0 1 1 0 1 1 1 0 0}. The distance or difference is calculated by X ⊕ Y operation and equals {0 0 1 0 1 0 0 1 0 0}. Finally, the similarity ratio of between X and Y is r. -(3/10) = 0.7 c. The movement of a firefly When a firefly moves, every bit in its representation string will make a decision to move (change its value) or not. The decision is determined after two actions: the attraction, which is regulated by the attractiveness (β); and the mutation, which is controlled by a parameter (α). The movement of a bit Xik in firefly Xi moving towards the corresponding bit Xjk in firefly Xj is defined as follows: 11 Xik=gfXikXjkβ,α 12 fXikXjkβ=Xjk,Xik,ifXik≠Xjkandrand01<βotherwise 13 gXikα=1−Xik,Xik,ifrand01<αotherwise 14 α=0.5−0.5×IterationMaxIteration where the inner function f(x,y,x) of (Eq.11) regulates the attracted movement of bit Xik to Xjk, and the outer function g(x, α) regulates the random moving behavior (mutation) of Xik. It should be noted that an attracted movement would incur only when the two corresponding bits are different, while the mutation might occur on every bit with the same probability. The introduction of dynamic mutation operator grants the firefly the ability to escape from a local optimum and check nearby regions while flying. In this work, parameter α controls the probability of mutation. The mutation probability is high in initial iterations, which makes BFA focus on exploration. As the number of iteration increases, the mutation probability will decrease, and BFA will accelerate its converging pace gradually. Figure 5 demonstrates an example of firefly movement. If a firefly is attracted by another, each different bit in the attracted firefly would change with probability β. Then each bit in the new firefly mutates with probability α.Fig. 5 An example of movement and mutation for a firefly Statistic inference and performance evaluation Five indices were employed to measure the performance of our method. These indices included sensitivity (SN), specificity (SP), accuracy (ACC), and Matthews’s correlation coefficient (MCC): 15 SN=TPTP+FN 16 SP=TNTN+FP 17 ACC=TP+TNTP+TN+FP+FN 18 MCC=TP×TN−FN×FPTP+FN×TP+FP×TN+FP×TN+FN where TP, FP, TN, and FN were the abbreviations of true positive, false positive, true negative, and false negative, respectively. The area under the receiver operating characteristic curve (ROC-AUC) was carried out when we assessed our method with other feature selection methods. The performance was evaluated by using leave-one-out cross-validation on the main dataset and selected optimal feature subset and parameters. Finally, the workflow of our method is shown in Fig. 6.Fig. 6 The flowchart of proposed method Results and discussion The performance of the proposed method The proposed method was implemented by combining informative features and optimizing parameters using BFA based on SVM. The settings of BFA were tuned as the following: the number of fireflies was set to 30; the visibility γ was set to 1; and the maximum iterations was set to 500. The light intensity was defined as follows: 19 I=ω×MCC+1−ω×1−nN where n was the number of selected features, N was the total number of features, and ω was the weighting coefficient that controlled the trade-off between the prediction accuracy and the selected features. Usually, the weighting coefficients of an algorithm are determined empirically. In our research, ω was set as 0.55. Here, MCC was used as the key criterion to evaluate the performance of a feature subset, as it could provide balanced and unbiased measurement of the prediction ability of the model. nN was used to assess the number of selected features. This experiment was repeated 20 times. The final performance was the average of the 20 results. The experiment with the medium value of MCC were chosen and the corresponding optimal feature subset and parameters were used to build the iDbP prediction model. The following experiments were all based on the selected optimal feature subset and parameters. Finally, the proposed method achieved a promising performance with the mean MCC of 0.595, ACC of 0.795, SN of 0.863, SP of 0.726 on the main dataset. Comparison with other feature selection techniques Feature selection is an important technique in predictive modeling. By removing redundant features, it can considerably improve the prediction accuracy. In this section, we compared BFA with several popular feature selection techniques: binary particle swarm optimization (BPSO) [50], genetic algorithm (GA) [51], minimum redundancy maximum relevance [52] combined with incremental feature selection (mRMR + IFS) [41], the original FA [44], and the straightforward method with all features. PSO is a meta-heuristic algorithm that optimizes a problem by searching optimal particle (candidate solution). The position and velocity of the particle vary in each iteration to approach the best position (global optimum). BPSO is the binary version of PSO. GA is a classic intelligent algorithm that emulates genetic evolution. It uses binary representation in nature and is good at discrete optimizations. mRMR + IFS is a combined feature selection scheme. It firstly sorts the features with criteria of minimum redundancy maximum relevance. Then, it iteratively uses the first n ranked features to build models to find the best feature subset. For the original FA, which should only be used in continuous problems, the binary string of the feature vector was transferred to decimal values. All these methods were embedded with SVM and run 20 times on the main dataset using exactly the same procedure. The final performance for each method were the average performance of 20 results. Table 1 lists the detailed results of five feature selection methods and the straightforward method with all features. Compared with simple feature fusion or filter feature selection, the meta-heuristic algorithms were more effective in selecting the optimal feature subsets. In addition, the FA produced an unsatisfactory performance, which proved that it was not suitable for discrete problem. Among the three meta-heuristic algorithms, BFA outperformed other methods with the highest MCC of 0.595.Table 1 Comparison of BFA with different feature selection methods Method SN SP ACC MCC BFA 0.863 0.726 0.795 0.595 BPSO 0.830 0.710 0.770 0.544 GA 0.840 0.680 0.760 0.527 FA 0.720 0.610 0.665 0.332 mRMR + IFS 0.790 0.640 0.715 0.435 All features 0.680 0.760 0.600 0.365 To assess the robustness of our BFA, we further drawn ROC curves for each method using the leave-one-out cross-validation on the main dataset. With all features, the predictor gave an AUC of 0.727. The mRMR + IFS scheme gave an AUC of 0.767. Additionally, the heuristic feature selection algorithms achieved better performance, an AUC of 0.747 for FA, an AUC of 0.768 for GA and an AUC of 0.779 for BPSO (Fig. 7). The newly proposed BFA produced an AUC values of 0.791, which was the highest among these feature selection methods. In our research, the BFA takes about 90 min to complete one entire experiment on a PC with a 3.20 GHz Intel Xeon CPU and 8GB RAM. Further improvement can be achieved by parallel computation, which is almost 4 times faster by computing 6 fireflies concurrently.Fig. 7 ROC curves of different feature selection methods Comparison with existing methods Comparison with other predictors on benchmark datasets In recent years, several methods were proposed to identify DBPs. These methods included DNAbinder [9], iDNA-Prot [11], enDNA-Prot [13], nDNA-Prot [12], DBPPred [15], DBD-Threader [53] and Zou’s method [14]. Among these methods, DNAbinder, iDNA-Prot, enDNA-Prot, nDNA-Prot, DBPPred and Zou’s method were sequence-based methods. To ensure a fair comparison with previous studies, the training dataset PDB594 of DBPPred was adopted to train iDbP and the independent testing dataset PDB186 was used to evaluate our predictor and compare with previous studies. Listed in Table 2 are the results of the comparison. Our iDbP achieved the highest SN of 0.894, ACC of 0.809 and MCC of 0.625. Additionally, we also compared the AUC value of iDbP with these predictors. As the AUC scores for iDNA-Prot, DNA-Prot, enDNA-Prot, nDNA-Prot, and DBD-Threader were unavailable, the comparisons were performed among DBPPred, DNAbinder, DNABIND and iDbP. The DBPPred, DNAbinder, DNABIND produced the AUC scores of 0.791, 0.607 and 0.694. Our iDbP yielded the highest AUC score of 0.803, which was slightly better than DBPPred.Table 2 Comparison of iDbP with existing methods on dataset PDB186 Method SN SP ACC MCC iDbP 0.894 0.722 0.809 0.625 DBPPred 0.796 0.742 0.769 0.538 iDNA-Prot 0.677 0.667 0.672 0.344 nDNA-Prot 0.710 0.623 0.667 0.335 enDNA-Prot 0.602 0.699 0.651 0.303 DNA-Prot 0.699 0.538 0.618 0.240 DNAbinder 0.570 0.645 0.608 0.216 DBD-Threader 0.237 0.957 0.597 0.279 Similarly, the training dataset DNAdset from Zou’s method was adopted to train iDbP and the independent testing dataset DNAiset was used to evaluate iDbP and compare with previous studies. As the services of DBPPred and DBDThreader were not availiable. The comparison on Zou’s benchmark dataset was performed among iDNA-Prot, DNAbinder, enDNA-Prot, nDNA-Prot, Zou’s method and our iDbP. As shown in Table 3, the iDbP yielded the best performance with the SN of 0.908, SP of 0.911, ACC of 0.910 and MCC of 0.803.Table 3 Comparison of iDbP with existing methods on dataset DNAiset Method SN SP ACC MCC iDbP 0.908 0.911 0.910 0.803 Zou’s method 0.890 0.828 0.900 0.753 iDNA-Prot 0.875 0.798 0.837 0.709 nDNA-Prot 0.779 0.887 0.851 0.664 enDNA-Prot 0.760 0.868 0.832 0.623 DNAbinder 0.717 0.642 0.863 0.473 Theoretically, protein structures could provide more information than primary sequences. However, our experiments showed that the sequence-based method could produce approximate or even better results. In general, the sequence-based methods are significant supplements for the structure-based methods, especially when the high-resolution 3D structures or the homology templates of the query proteins are hard to obtain. Comparison with other predictors on DBP189 dataset To demonstrate the generalization ability of our iDbP, we performed further comparisons with previous methods on DBP189. Three DBP prediction tools, namely DNA-Prot, iDNA-Prot and DNAbinder, still provided online or local prediction services. The prediction results (shown in Table 4) on the DBP189 dataset indicated that our method still characterized by good predictive performance on imbalanced testing dataset. Among these methods, our iDbP achieved the highest MCC of 0.5996, which was about 5 % more than the second highest method DNA-Prot.Table 4 Comparison of predictive quality on the DBP189 dataset Method SN SP ACC MCC iDbP 0.7619 0.9162 0.8989 0.5996 DNA-Prot 0.7143 0.9042 0.8830 0.5415 iDNA-Prot 0.6190 0.8563 0.8298 0.3960 DNAbinder 0.5714 0.8263 0.7979 0.3234 Application to large-scale DBP prediction In real-life application, computational tools are often used to identify possible candidate proteins in large scale. To simulate this scenario, we collected 15,413 DBPs from five most popular organisms (human, A. thaliana, mouse, S. cerevisiae and fruit fly) in UniProt database. After removing incomplete segments and unannotated proteins, we finally obtained a large-scale testing dataset with 2859 DBPs (Provided in Additional file 2). As shown in Table 5, by using our iDbP, nearly 59 % of human proteins, 53 % of A. thaliana proteins, 54 % of Mouse proteins, 61 % of S. cerevisiae proteins, and 59 % of Fruit fly proteins were successfully predicted as DBPs. In summary, about 56 % proteins were successfully recognized. The results showed that iDbP could be a reliable tool in large-scale applications.Table 5 Number of annotated and recognized DBPs in UniProt database Category Number of proteins Proteins with complete DNA binding annotations Number of predicted SN DBPs Human 6,813 1,049 613 58 % A. thaliana 3,378 929 489 53 % Mouse 2,514 424 232 54 % S. cerevisiae 1,545 314 191 61 % Fruit fly 1,163 143 84 59 % Summary 15,413 2,859 1609 56 % Conclusion In this work, we proposed a new method, named iDbP, to predict DBPs from primary sequence. Multiple informative features, which derived from evolutionary conservation profile, secondary structure motifs and physiochemical properties, were used to discriminate DBPs from non-binding proteins. Next, a novel improved BFA was forged to perform feature selection and parameter optimization. The experimental results of our predictor on two benchmark datasets outperformed many state-of-the-art predictors, which revealed the effectiveness of our method. Moreover, the promising performance on an independent testing dataset and large-scale proteins from UniProt database proved the good generalization ability of our method. In addition, the novel improved BFA would be of a powerful algorithm which could find widely applications in discrete optimization problems. The web-server is available for academic research at http://59.73.198.144:8080/iDbP/. Additional files Additional file 1: The main dataset and DBP189 used in this study. (PDF 751 kb) Additional file 2: The large-scale testing dataset compiled from UniProt. (PDF 477 kb) We thank Fundamental Research Funds for the Central Universities (Northeast Normal University) for proving the funding for this work. Funding This work was supported by the Fundamental Research Funds for the Central Universities (Grant No. 14ZZ2240). Availability of data and material All the datasets used in this research can be found in Additional files associated with this paper. Authors’ contributions JZ conceived the idea and was in charge of the iDbP implementation. BG, HTC and ZQM optimized the algorithm and participated in the development and validation of the Web server. BG and GFY drafted the first version of the manuscript. JZ and BG designed experiments, gathered test data, and were in charge of the experiments. ZQM and GFY supervised the progress of the whole project and critically checked the first draft. GFY was in charge of the whole process of final revision. All authors have read and approved the final manuscript. Authors’ information Not applicable. 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. Langlois RE Lu H Boosting the prediction and understanding of DNA-binding domains from sequence Nucleic Acids Res 2010 15 gkq061 2. Sarai A Kono H Protein-DNA recognition patterns and predictions Annu Rev Biophys Biomol Struct 2005 34 379 98 10.1146/annurev.biophys.34.040204.144537 15869395 3. Parola M Bellomo G Robino G Barrera G Dianzani MU 4-Hydroxynonenal as a biological signal: molecular basis and pathophysiological implications Antioxid Redox Signal 1999 1 3 255 84 10.1089/ars.1999.1.3-255 11229439 4. Chou CC Lin TW Chen CY Wang AH Crystal structure of the hyperthermophilic archaeal DNA-binding protein Sso10b2 at a resolution of 1.85 Angstroms J Bacteriol 2003 185 14 4066 73 10.1128/JB.185.14.4066-4073.2003 12837780 5. Freeman K Gwadz M Shore D Molecular and genetic analysis of the toxic effect of RAP1 overexpression in yeast Genetics 1995 141 4 1253 62 8601471 6. Gao M Skolnick J DBD-Hunter: a knowledge-based method for the prediction of DNA–protein interactions Nucleic Acids Res 2008 36 12 3978 92 10.1093/nar/gkn332 18515839 7. Zhao H Yang Y Zhou Y Structure-based prediction of DNA-binding proteins by structural alignment and a volume-fraction corrected DFIRE-based energy function Bioinformatics 2010 26 15 1857 63 10.1093/bioinformatics/btq295 20525822 8. Nimrod G Szilágyi A Leslie C Ben-Tal N Identification of DNA-binding proteins using structural, electrostatic and evolutionary features J Mol Biol 2009 387 4 1040 53 10.1016/j.jmb.2009.02.023 19233205 9. Kumar M Gromiha MM Raghava GP Identification of DNA-binding proteins using support vector machines and evolutionary profiles BMC Bioinformatics 2007 8 1 1 10.1186/1471-2105-8-463 17199892 10. Kumar KK Pugalenthi G Suganthan PN DNA-Prot: identification of DNA binding proteins from protein sequence information using random forest J Biomol Struct Dynam 2009 26 6 679 86 10.1080/07391102.2009.10507281 11. Lin WZ Fang JA Xiao X Chou KC iDNA-Prot: identification of DNA binding proteins using random forest with grey model PLoS One 2011 6 9 e24756 10.1371/journal.pone.0024756 21935457 12. Song L Li D Zeng X Wu Y Guo L Zou Q nDNA-prot: identification of DNA-binding proteins based on unbalanced classification BMC Bioinformatics 2014 15 1 1 10.1186/1471-2105-15-298 24383880 13. Xu R, Zhou J, Liu B, Yao L, He Y, Zou Q, Wang X. enDNA-Prot: identification of DNA-binding proteins by applying ensemble learning. BioMed Res Int. 2014. 14. Zou C Gong J Li H An improved sequence based prediction protocol for DNA-binding proteins using SVM and comprehensive feature analysis BMC Bioinformatics 2013 14 1 1 10.1155/2013/191586 23323762 15. Lou W Wang X Chen F Chen Y Jiang B Zhang H Sequence based prediction of DNA-binding proteins based on hybrid feature selection using random forest and Gaussian naive Bayes PLoS One 2014 9 1 e86703 10.1371/journal.pone.0086703 24475169 16. Rohs R Jin X West SM Joshi R Honig B Mann RS Origins of specificity in protein-DNA recognition Annu Rev Biochem 2010 79 233 10.1146/annurev-biochem-060408-091030 20334529 17. Chou KC Shen HB Recent progress in protein subcellular location prediction Anal Biochem 2007 370 1 1 16 10.1016/j.ab.2007.07.006 17698024 18. Wang G Dunbrack RL PISCES: recent improvements to a PDB sequence culling server Nucleic Acids Res 2005 33 suppl 2 W94 8 10.1093/nar/gki402 15980589 19. Zhang J Chen W Sun P Zhao X Ma Z Prediction of protein solvent accessibility using PSO-SVR with multiple sequence-derived features and weighted sliding window scheme BioData Min 2015 8 1 1 15 10.1186/s13040-014-0034-0 25621011 20. Zhang J, Zhao X, Sun P, Gao B, Ma Z. Conformational B-cell epitopes prediction from sequences using cost-sensitive ensemble classifiers and spatial clustering. BioMed Res Int. 2014. 21. Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 402 10.1093/nar/25.17.3389 9254694 22. Bochman ML Paeschke K Zakian VA DNA secondary structures: stability and function of G-quadruplex structures Nat Rev Genet 2012 13 11 770 80 10.1038/nrg3296 23032257 23. Greive SJ, Fung HK, Chechik M, Jenkins HT, Weitzel SE, Aguiar PM, Brentnall AS, Glousieau M, Gladyshev GV, Potts JR, Antson AA. DNA recognition for virus assembly through multiple sequence-independent interactions with a helix-turn-helix motif. Nucleic Acids Res. 2016;44(2):776-789. 24. Jones DT Protein secondary structure prediction based on position-specific scoring matrices J Mol Biol 1999 292 2 195 202 10.1006/jmbi.1999.3091 10493868 25. Liu ZP Wu LY Wang Y Zhang XS Chen L Prediction of protein–RNA binding sites by a random forest method with combined features Bioinformatics 2010 26 13 1616 22 10.1093/bioinformatics/btq253 20483814 26. Bordner AJ Abagyan R Statistical analysis and prediction of protein–protein interfaces Proteins Struct Funct Bioinf 2005 60 3 353 66 10.1002/prot.20433 27. Jayaram B McConnell KJ Dixit SB Beveridge DL Free energy analysis of protein–DNA binding: the EcoRI endonuclease–DNA complex J Comput Phys 1999 151 1 333 57 10.1006/jcph.1998.6173 28. Chaires JB Satyanarayana S Suh D Fokt I Przewloka T Priebe W Parsing the free energy of anthracycline antibiotic binding to DNA Biochemistry 1996 35 7 2047 53 10.1021/bi952812r 8652545 29. Liu S Liu S Zhu X Liang H Cao A Chang Z Lai L Nonnatural protein–protein interaction-pair design by key residues grafting Proc Natl Acad Sci 2007 104 13 5330 5 10.1073/pnas.0606198104 17372228 30. Ahmad S Sarai A Moment-based prediction of DNA-binding proteins J Mol Biol 2004 341 1 65 71 10.1016/j.jmb.2004.05.058 15312763 31. Landschulz WH Johnson PF McKnight SL The leucine zipper: a hypothetical structure common to a new class of DNA binding proteins Science 1988 240 4860 1759 64 10.1126/science.3289117 3289117 32. Ip YT Kraut R Levine M Rushlow CA The dorsal morphogen is a sequence-specific DNA-binding protein that interacts with a long-range repression element in Drosophila Cell 1991 64 2 439 46 10.1016/0092-8674(91)90651-E 1988156 33. Yu X Cao J Cai Y Shi T Li Y Predicting rRNA-, RNA-, and DNA-binding proteins from primary structure with support vector machines J Theor Biol 2016 240 2 175 84 10.1016/j.jtbi.2005.09.018 16274699 34. Slattery M Riley T Liu P Abe N Gomez-Alcala P Dror I Zhou T Rohs R Honig B Bussemaker HJ Mann RS Cofactor binding evokes latent differences in DNA binding specificity between Hox proteins Cell 2011 147 6 1270 82 10.1016/j.cell.2011.10.053 22153072 35. PSORT I PSORT: a program for detecting sorting signals in proteins and predicting their subcellular localization J Mol Biol 1997 266 594 600 10.1006/jmbi.1996.0804 9067612 36. Mukherjee S Berger MF Jona G Wang XS Muzzey D Snyder M Young RA Bulyk ML Rapid analysis of the DNA-binding specificities of transcription factors with DNA microarrays Nat Genet 2004 36 12 1331 9 10.1038/ng1473 15543148 37. Ahmad S Gromiha MM Sarai A Analysis and prediction of DNA-binding proteins and their binding residues based on composition, sequence and structural information Bioinformatics 2004 20 4 477 86 10.1093/bioinformatics/btg432 14990443 38. Goodsell D Dickerson RE Isohelical analysis of DNA groove-binding drugs J Med Chem 2004 29 5 727 33 10.1021/jm00155a023 2422377 39. Chaires JB A thermodynamic signature for drug–DNA binding mode Arch Biochem Biophys 2006 453 1 26 31 10.1016/j.abb.2006.03.027 16730635 40. Nowak MW Kearney PC Saks ME Labarca CG Silverman SK Zhong W Thorson J Abelson JN Davidson N Nicotinic receptor binding site probed with unnatural amino acid incorporation in intact cells Science 1995 268 5209 439 42 10.1126/science.7716551 7716551 41. Zhang J Sun P Zhao X Ma Z PECM: Prediction of extracellular matrix proteins using the concept of Chou’s pseudo amino acid composition J Theor Biol 2014 363 412 8 10.1016/j.jtbi.2014.08.002 25123433 42. Vapnik V The nature of statistical learning theory 2013 43. Chang CC Lin CJ LIBSVM: a library for support vector machines ACM Trans Intell Syst Technol 2011 2 3 27 10.1145/1961189.1961199 44. Yang XS. Firefly algorithms for multimodal optimization. In Stochastic algorithms: foundations and applications. Springer Berlin Heidelberg; 2009: 169–178. 45. Hashmi A Goel N Goel S Gupta D Firefly algorithm for unconstrained optimization IOSR J Comput Eng 2013 11 1 75 8 10.9790/0661-1117578 46. Yang XS He X Firefly algorithm: recent advances and applications Int J Swarm Intell 2013 1 1 36 50 10.1504/IJSI.2013.055801 47. Palit S Sinha SN Molla MA Khanra A Kule M A cryptanalytic attack on the knapsack cryptosystem using binary firefly algorithm In Int Conf Comput Commun Technol (ICCCT) 2011 2 428 32 48. Sayadi MK Hafezalkotob A Naini SGJ Firefly-inspired algorithm for discrete optimization problems: an application to manufacturing cell formation J Manuf Syst 2013 32 1 78 84 10.1016/j.jmsy.2012.06.004 49. Poursalehi N Zolfaghari A Minuchehr A A novel optimization method, Effective Discrete Firefly Algorithm, for fuel reload design of nuclear reactors Ann Nuclear Energy 2015 81 263 75 10.1016/j.anucene.2015.02.047 50. Chuang LY Chang HW Tu CJ Yang CH Improved binary PSO for feature selection using gene expression data Comput Biol Chem 2008 32 1 29 38 10.1016/j.compbiolchem.2007.09.005 18023261 51. Chandrashekar G Sahin F A survey on feature selection methods Comput Electr Eng 2014 40 1 16 28 10.1016/j.compeleceng.2013.11.024 52. Ding C Peng H Minimum redundancy feature selection from microarray gene expression data J Bioinform Comput Biol 2005 3 02 185 205 10.1142/S0219720005001004 15852500 53 Gao M Skolnick J A threading-based method for the prediction of DNA-binding proteins with application to the human genome PLoS Comput Biol 2009 5 11 e1000567 10.1371/journal.pcbi.1000567 19911048

PMC005xxxxxx/PMC5002160.txt

==== Front Malar JMalar. JMalaria Journal1475-2875BioMed Central London 149210.1186/s12936-016-1492-2ResearchPrevalence of malaria in pregnancy in southern Laos: a cross-sectional survey Briand Valérie 0033153731527valerie.briand@ird.fr 12Le Hesran Jean-Yves Jean-Yves.Lehesran@ird.fr 12Mayxay Mayfong mayfong@tropmedres.ac 345Newton Paul N. paul.newton@tropmedres.ac 35Bertin Gwladys gwladys.bertin@ird.fr 12Houzé Sandrine sandrine.houze@bch.aphp.fr 126Keomany Sommay drsommay@yahoo.com 7Inthavong Yom yominthavong@yahoo.com 8Vannavong Nanthasane anandafet@gmail.com 9Chindavongsa Keobouphaphone chinda07@gmail.com 10Hongvanthong Bouasy cmpelao@gmail.com 10Fievet Nadine fievet@ird.fr 121 Faculté de Pharmacie, Institut de Recherche pour le Développement (IRD), Mère et enfant face aux infections tropicales (UMR216), 4 avenue de l’Observatoire, 75006 Paris, France 2 COMUE Sorbonne Paris Cité, Faculté de Pharmacie, Université Paris Descartes, Paris, France 3 Lao-Oxford-Mahosot Hospital-Wellcome Trust Research Unit (LOMWRU), Microbiology Laboratory, Vientiane, Lao People’s Democratic Republic 4 Faculty of Postgraduate Studies, University of Health Sciences, Vientiane, Lao People’s Democratic Republic 5 Centre for Tropical Medicine and Global Health, Nuffield Department of Clinical Medicine, University of Oxford, Oxford, UK 6 Laboratoire de Parasitologie, CNR du Paludisme, AP-HP, Hôpital Bichat, Paris, France 7 Salavan Provincial Hospital, Salavan, Salavan Province Lao People’s Democratic Republic 8 Vapi District Hospital, Vapi, Salavan Province Lao People’s Democratic Republic 9 Champasack Provincial Health Office, Champasack, Champasack Province Lao People’s Democratic Republic 10 Center of Malariology, Parasitology and Entomology, Vientiane, Lao People’s Democratic Republic 26 8 2016 26 8 2016 2016 15 1 4369 6 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 There are no data on the burden of malaria in pregnancy (MiP) in Laos, where malaria still remains prevalent in the south. Methods Two cross-sectional surveys were conducted in 2014 to assess the prevalence of MiP in Vapi District, Salavan Province, southern Laos: the first consisted of screening 204 pregnant women during pregnancies [mean (95 % CI) gestational age: 23 (22–25) weeks] living in 30 randomly selected villages in Vapi District; the second was conducted among 331 pregnant women, who delivered during the study period in Vapi and Toumlane District Hospitals and in Salavan Provincial Hospital. Peripheral and placental malaria was detected using rapid diagnostic tests (RDT), thick blood smears (TBS) and real-time quantitative polymerase chain reactions (RT-qPCR). Factors associated with low birth weight (LBW) and maternal anaemia were assessed. Results In the villages, 12/204 women (5.9 %; 95 % CI 3.1–10.0) were infected with malaria as determined by RT-qPCR: 11 were Plasmodium vivax infections and 1 was mixed Plasmodium vivax/Plasmodium falciparum infection, among which 9 were sub-microscopic (as not detected by TBS). History of malaria during current pregnancy tended to be associated with a higher risk of MiP (aIRR 3.05; 95 % CI 0.94–9.88). At delivery, two Plasmodium falciparum sub-microscopic infections (one peripheral and one placental) were detected (4.5 %; 0.6–15.5) in Vapi District. In both surveys, all infected women stated they had slept under a bed net the night before the survey, and 86 % went to the forest for food-finding 1 week before the survey in median. The majority of infections (94 %) were asymptomatic and half of them were associated with anaemia. Overall, 24 % of women had LBW newborns. Factors associated with a higher risk of LBW were tobacco use (aIRR 2.43; 95 % CI 1.64–3.60) and pre-term delivery (aIRR 3.17; 95 % CI 2.19–4.57). Factors associated with a higher risk of maternal anaemia were no iron supplementation during pregnancy, Lao Theung ethnicity and place of living. Conclusions The prevalence of MiP in this population was noticeable. Most infections were asymptomatic and sub-microscopic vivax malaria, which raises the question of reliability of recommended national strategies for the screening and prevention of MiP in Laos. Electronic supplementary material The online version of this article (doi:10.1186/s12936-016-1492-2) contains supplementary material, which is available to authorized users. Keywords MalariaPregnancyLow birth weightEpidemiologyMolecular biologyLaosFrance Expertise Internationale (FEI)13INI144Fievet Nadine issue-copyright-statement© The Author(s) 2016 ==== Body Background In 2007, an estimated 125 million women became pregnant in areas of malaria transmission. Of them 62 %, corresponding to 77 million women, were living in the Asia–Pacific region where malaria transmission is low or unstable [1]. Malaria in pregnancy (MiP) is responsible for poor maternal and child outcomes, including maternal anaemia, low birth weight (LBW), due to intra-uterine growth retardation and pre-term birth, which are important contributors to infant morbidity and mortality [2]. In these settings, MiP is also associated with spontaneous abortion, stillbirth and maternal death [3]. A single infection in pregnancy, even asymptomatic or sub-microscopic, can be detrimental to a mother and her fetus, in particular in early pregnancy, when most of the women do not attend antenatal care (ANC) [3]. Few studies have reported on the prevalence and consequences of MiP in low or unstable malaria transmission areas, and particularly in Southeast Asia, compared to high transmission areas [3]. In a recent review, the proportion of women with parasitaemia during pregnancy in the Asia–Pacific region has been estimated to be 15 % (range 1.2–40.8) based on cross-sectional surveys, and as high as 36.5 % (range 6.0–64.0) based on longitudinal studies [3]. In Laos, malaria is endemic throughout most of the country, but intensity of transmission is variable, with more intense transmission in remote and forested areas particularly in the south [4]. The distribution of insecticide-treated nets (ITNs) at household level, free screening and case management of clinical malarial infections and the large-scale distribution of artemisinin-based combination therapy (ACT), have contributed to reducing the incidence of malaria. However, clinical malaria still remains a serious public health problem in the five southern provinces of the country, where a high number of clinical cases are still reported in ‘hot spots’ of malaria transmission (Centre of Malariology, Parasitology and Entomology, CMPE Malaria Information System, unpublished observations). Most patients are adults living in villages who work in forest-related occupations or who are migrating adults workers, sometime accompanied by their families, according to work availability [5]. In Laos, malaria screening for pregnant women is only performed when they present with fever. Therefore, asymptomatic and sub-microscopic infections as well as symptomatic women who do not access health services are neither detected nor reported. The lack of precise information on the burden of MiP in Southeast Asia, and particularly in Laos, has hampered effective lobbying for the inclusion of malaria preventive strategies during pregnancy [6]. In Laos, interventions against malaria that are recommended during pregnancy are the same as for the general population, they consist of use of ITNs and effective treatment of symptomatic malaria [7]. In the context of persistent malaria transmission in southern Laos, the prevalence of MiP, including asymptomatic and sub-microscopic infection, and its related morbidity in this population were assessed. Epidemiological methods that minimize selection bias as well as highly sensitive methods of detection for malaria were used. Methods The study was conducted in Salavan Province, which is one of the five southern provinces where the reported number of clinical malaria cases has been the highest since 2011 (CMPE Malaria Information System, unpublished observations). In this area, malaria is perennial with one peak during the rainy season, which occurs between May and October. Two cross-sectional surveys were carried out simultaneously. One was conducted at facility level to investigate MiP at delivery, at Salavan Provincial Hospital (latitude 15.817107 N, longitude 106.252214 E, altitude 183 masl), which is a primary-secondary hospital, and in Vapi (latitude 15.66526 N, longitude 105.906837 E, altitude 136 masl) and Toumlane (latitude 15.993951 N, longitude 106.191913 E, altitude 197 masl) District Hospitals, which are located in more rural areas (Fig. 1).Fig. 1 Study sites, Salavan Province, Laos, 2014. The blue box corresponds to Toumlane District Hospital, the red box to Salavan Provincial Hospital and the black box to Vapi District Hospital. Source: World Health Organization, Lao PDR The second survey was conducted at community level in the villages of Vapi District as a one-time point survey of pregnant women. Data were collected during the rainy season (i.e., malaria transmission period) from July to October 2014. Both surveys were part of a larger programme (PALULAO, FEI n°13INI144), which aimed to assess the prevalence of asymptomatic and sub-microscopic malarial infections in the general population, including pregnant women, adults and children. This programme was led by the CMPE in Laos, in collaboration with the Salavan Provincial Department of Health, the Research Institute for Development (IRD) and the Lao-Oxford-Mahosot Hospital Wellcome Trust Research Unit (LOMWRU). Study populations and designs At facility level, all women who delivered in one of the three participating maternity clinics between August and October 2014 were included after providing a written informed consent. Women were excluded from the study if they had abortion, complicated/at risk pregnancy requiring a pre-labour emergency or elective caesarean section. The following information was recorded on admission by specifically trained midwives: maternal characteristics (age, ethnicity, place of living, education level), obstetrical history (gravidity, number of spontaneous abortions, stillbirths, and live births), clinical data related to the current pregnancy (number and place of ANC visits, obstetrical complication, iron and folic acid supplementation, tobacco use, malaria history, and bed net use the night before admission), and working in the forest during the current pregnancy. At delivery, gestational age, blood pressure, temperature, and delivery characteristics were recorded. Newborn anthropometric data (weight, height and head circumference), gender, Apgar score and congenital abnormalities were recorded. Newborn weight was measured within 1 h after birth using an electronic digital scale with an accuracy of 100 g (SECA Ltd). In addition, maternal peripheral, placental and umbilical cord blood samples were drawn for malaria diagnosis using rapid diagnostic tests (RDT), thick and thin blood smears (TBS) and real-time quantitative polymerase chain reactions (RT-qPCR). Haemoglobin (Hb) level, and proteinuria and glycosuria [using a urine dipstick test (Cybow 2GP®)] were determined. At the community level, the cross-sectional survey was carried out in 30 randomly selected villages out of a total of 53 villages in Vapi District. The day before the survey the head of the village made an announcement inviting all pregnant women to participate in the study. Moreover, women with suspicion of pregnancy were encouraged to visit the study team, which performed a urinary pregnancy test. Maternal characteristics, obstetrical history, clinical data related to the current pregnancy (including bed net use the night before the survey) and activities in the forest were collected. Women were screened for malaria using RDT, TBS and PCR whether or not they were symptomatic. Hb level and proteinuria and glycosuria were determined. In both surveys, data quality was monitored by study coordinators on a weekly basis. Women with positive RDTs for malaria as well as anaemic women were treated according to the national guidelines. Laboratory investigations RDT (Malaria antigen Pf/Pv, SD Bioline 05FK80, Europ Continents Ltd) was performed on all peripheral blood samples. Blood collected on filter papers was dried and conserved at room temperature until DNA extraction using Chelex [8]. RT-qPCR was performed on peripheral, placental and umbilical cord blood samples for sub-microscopic infection detection. A duplex real-time PCR assay, using genus-specific and species-specific primers and probes for gene encoding the small sub-unit (18S) of Plasmodium rRNA, as described by Diallo et al. [9], was used. Samples underwent 40 cycles of amplification using the ViiA™ 7 real-time PCR system (Applied Biosystem) and were quantified using a DNA standard range made from a suspension of in vitro culture of the 3D7 P. falciparum line (MR4). PCR were performed in the Microbiology Laboratory of Mahosot Hospital, Vientiane, Laos. A quality control for PCR was performed in Hôpital Bichat, France. DNA extracts of 30 % randomly selected PCR-positive samples and 13 % randomly selected PCR-negative samples were analysed in blind with the FTD (Fast-Track Diagnostics) Malaria PCR, a multiplex PCR for detection of Plasmodium spp. DNA. In case of positive result, the PCR FTD Malaria differentiation was performed to identify the species of Plasmodium. Agreement between both methods for PCR was high (Cohen’s kappa coefficient: 0.92). Agreement was complete for the 48 samples from pregnant women which were analysed. TBSs were stained with Giemsa. All placental blood smears as well as a sub-sample of peripheral blood smears (including all PCR-positive samples) were read, and parasitaemia was quantified by the Lambaréné method [10]. TBS reading was performed on site by CMPE staff and Salavan Provincial Hospital staff. Infections were considered sub-microscopic if malaria parasites were detected by qPCR but absent by microscopy. Hb level was determined by HemoCue® (community-level survey) or coulter counter (HUMACOUNT 60TS) (facility-level survey). Statistical analysis Maternal characteristics were described using proportions (95 % confidence interval) for categorical variables and means (95 % CI) and medians (IQR) for continuous variables. The proportion and its exact Poisson 95 % CI of women infected with malaria using RDT and PCR were calculated. Maternal characteristics associated with malaria during pregnancy (community-level study) were assessed using a modified Poisson regression model [11]. For each covariate, the incidence rate ratio (IRR) and its 95 % CI were estimated. Factors associated with malaria at delivery and its related maternal and child morbidity (facility-level study) could not be assessed because of the low number of malarial infections. However, maternal sociodemographic and obstetrical characteristics associated with low birth weight (LBW) and maternal anaemia at delivery were assessed. Maternal anaemia was defined as a haemoglobin level less than 11 g/dL. Mean Hb level at delivery, mean birth weight and small-for-gestational age (SGA)—as a proxy for intra-uterine growth restriction—were considered as secondary outcomes. SGA was defined as a birth weight below the 10th percentile of the birth weight-for-gestational age according to INTERGROWTH charts [12]. Gestational age and how it had been determined (ultrasound scan, date of last menstrual period (LMP) or fundal height measurement) was recorded from the maternity clinic registers. Twins and stillbirths were excluded from birth weight and SGA analyses. For each outcome, univariate analyses were first performed using linear regression and modified Poisson models depending on the outcome analyzed. Variables for which P value was less or equal to 0.25 were further included in a subsequent multivariate model. The variables were then introduced together in a multivariate regression and were eliminated step by step using the backward method. At this level, only variables for which the P value was less than 0.05 were kept. We considered the following covariates: ethnicity, place of living, maternal age (transformed into a four class variable corresponding to the quartiles), gravidity (primigravidae or multigravidae), tobacco use, forest activity during the current pregnancy, gestational hypertension, use of bed net the night before admission, number of ANC visits (zero, one–three, ≥four visits) iron and folic acid supplementation during the current pregnancy (intake the day before admission, intake but not the day before admission, no intake), maternal anaemia at delivery (for birth weight analyses only), duration of pregnancy (<37, 37–38, ≥39 weeks of gestation) and sex of the newborn. Bed net use was forced in all multivariate models. Data entry and analysis were performed using Epidata (version 3.1) and Stata (version 13), respectively. Results Community survey At community-level, 205 pregnant women (median (IQR) of six women (five) per village) were included (Table 1). Women were primi-, secondi- and multigravidae in 28.8, 30.7 and 40.5 % of cases, respectively. Mean (95 % CI) gestational age was 23 (22–25) weeks as reported by women. It had been determined from LMP (67.5 %), ultrasound scan (20.2 %) or by fundal height measurement (12.3 %). Gestational hypertension without proteinuria was detected in only one woman. History of malaria during pregnancy was reported by 8.3 % of women. Only two women out of a total of 205 (1.0 %; 95 % CI 0.1–3.5) were infected with malaria using RDTs and these two persons were asymptomatic. Malaria was detected using RT-qPCR in 204/205 women, for one woman venous blood was not available. Twelve women out of a total of 204 (5.9 %; 95 % CI 3.0–10.3) were infected with malaria using PCR. Nine (9/12, 75 %) of these infections were sub-microscopic (as not detected by microscopy). Eleven of these infections were vivax malaria, and one was mixed P. vivax/P. falciparum infection (Table 2). All infected women were asymptomatic, except for one woman who had history of fever the day before the survey. All women except one stated that they slept under a bed net the night before the survey. Ninety-two percent of infected women (vs 82 % of uninfected women) had been to the forest within the previous 2 weeks to find food.Table 1 Maternal characteristics during pregnancy, Salavan Province, Laos, 2014 Community survey N = 205 Hospital survey N = 331 n (%) or mean (95 % CI) Maternal age (years) 24.9 (24.0–25.7) 25.0 (24.4–25.7) Gravidity  G1–G2 122 (59.5) 220 (66.5)  G3+ 83 (40.5) 111 (33.5) Ethnicity  Lao Loum 204 (99.5) 266 (80.4)  Lao Theung 1 (0.5) 65 (19.6) Place of living/delivery  Vapi 205 (100.0) 44 (13.3)  Toumlan 40 (11.8)  Salavan 248 (74.9) Educationa NA  No formal education 25 (12.8)  Primary school 103 (52.5)  Secondary school or higher 68 (34.7) Number of ANC visits during this pregnancy 1.4 (1.1–1.6) 3.8 (3.6–4.1) Iron supplementationa 103 (50.2) 310/330 (93.9) Folic acid supplementationa 0 (100.0) 258/327 (78.9) Gestational hypertensiona 1/204 (0.5) 11 (3.3) Tobacco use NA 38 (11.5) Went to the forest during the current pregnancy 170 (82.9) 141 (42.6) Bed net possession 204 (99.5) 307 (92.8) Bed net use the night before the survey/admission 204 (99.5) 305 (92.1) Any malaria episode during pregnancya 17/204 (8.3) 7/330 (2.1) Gestational age (weeks of gestation)a 23 (22–25) 38 (38-38) Maternal Hb level (g/dL)a 10.8 (10.6–11.0) 11.4 (11.3–11.6) Maternal anaemia (Hb < 11 g/dL)a 106 (51.7) 112/327 (34.3) ANC antenatal care, Hb haemoglobin aN = 331 women at delivery, N = 205 during pregnancy, or indicated otherwise. Blood pressure was not recorded for one woman (community-level survey), history of malaria could not be determined in two women (one during the community-level survey and one during the facility-level survey), education level was not recorded in nine women who were recruited in the first three investigated villages in Vapi District, iron and folic acid supplementation could not been determined in, respectively, one and four women during the facility-level survey, and Hb level could not be measured in four women in Toumlane District Hospital because of a defective device Table 2 Prevalence of peripheral malaria in pregnancy, Salavan Province, Laos, 2014 Number tested n (%) RDT+ n (%) PCR+ Pf+ Pf + Pv+ Pv+ Total Community survey (during pregnancy) Vapi 205 2 (1.0) 0 (0.0) 1 (0.5) 11 (5.4) 12/204 (5.9)a Hospital survey (at delivery) Vapi 44 0 (0.0) 2 (4.5) 0 (0.0) 0 (0.0) 2 (4.5) Salavan 248 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Toumlane 39 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Pv P. vivax, Pf P. falciparum aRT-qPCR was performed in 204/205 women (venous blood was not available for one woman) In univariate analysis factors significantly associated with MiP were history of malaria during the current pregnancy and a low education level. In multivariate analysis only history of malaria remained associated with MiP, although this association was only marginally significant (aIRR 3.05; 95 % CI 0.94–9.88). The prevalence of maternal anaemia was 51.7 % (106/205) with overall mean (95 % CI) Hb of 10.8 (10.6–11.0) g/dL (Table 1). Hospital survey A total of 331 pregnant women were included at delivery: 248 (74.9 %) in Salavan, 44 (13.3 %) in Vapi and 39 (11.8 %) in Toumlane Hospitals (Table 1). Women were primi-, secondi- and multigravidae in 38.4, 28.1 and 33.5 % of cases, respectively. Their mean (95 % CI) frequency of ANC visits was 3.8 (3.6–4.1) during their pregnancies. Gestational age had been determined using ultrasound scan in 84 % of women, from the LMP in 3.6 % and by fundal height measurement in 12.4 %. Gestational hypertension was detected in 3.3 % of women with no pre-eclampsia. Glycosuria was detected in only two women, for whom no further biological tests could be performed to confirm gestational diabetes. History of malaria during current pregnancy was reported by 2.1 % of women. Bed net use the night before admission was reported by 92.1 % of women. Looking for food in the forest during current pregnancy was reported by 42.6 % of women. The median (IQR) days of the last stay in the forest before delivery was seven (24). At delivery, sub-microscopic malarial infections were detected in 2/44 (4.5 %; 95 % CI 0.6–16.4) women living in Vapi District, representing two asymptomatic P. falciparum infections (Table 2). These women were primi- and secondigravidae. One of these infections was detected in the placenta. No infections were detected by RDTs. The prevalence of maternal anaemia at delivery was 34.3 % (112/327), and the mean (95 % CI) Hb level was 11.4 (11.3–11.6) g/dL (Table 3). The mean (95 % CI) birth weight of the children was 2793 g (2745–2842 g) and 24 % of women had LBW newborns. The frequency of pre-term birth was 8.5 %. Twenty-three percent of newborns (72/313 born between 33 and 43 weeks of gestation) were small-for-gestational age according to INTERGROWTH charts. Among them, 81 % had a LBW and 19 % had a normal birth weight.Table 3 Delivery and newborn characteristics, Salavan Province, Laos, 2014 Hospital survey N = 331 n (%) or mean (95 % CI) Multiple birth 8/331 (2.4) Gestational age (weeks of gestation) 38 (38–38) Delivery outcome  Live-birth 325 (98.2)  Stillbirth 6 (1.8) Maternal Hb level (g/dL)a 11.4 (11.3–11.6) Maternal anaemia (Hb < 11 g/dL)a 112/327 (34.3) Preterm birth 28 (8.5) Sex ratio (m/f) 0.98 Birth weight (grams)a,b 2793 (2745–2842) LBW (<2500 g)a,b 76/317 (24.0) SGAa,b 72/313 (23.0) Hb haemoglobin aN = 331 women at delivery or indicated otherwise. Hb level was missing for four women in Toumlane District Hospital because of a defective device bMean birth weight and proportions of LBW and SGA were calculated in live-singletons only. SGA was defined using the INTERGROWTH charts in newborns born from 33 to 43 weeks of gestation [12] Factors associated with low birth weight, SGA and mean birth weight The initial statistical analysis plan included the assessment of the effect of malaria at delivery on birth weight and maternal haemoglobin level. Because of the very low prevalence of malaria at delivery, this factor could not be included in the analyses. In univariate analysis, women who smoked, were of Lao Theung ethnicity, who did not use a bed net the night before admission, who had no ANC visits and those who had a pre-term delivery were significantly more likely to have a LBW newborn (Table 4). In multivariate analysis, tobacco use (aIRR 2.43; 95 % CI 1.64–3.60) and pre-term delivery (aIRR 3.17; 95 % CI 2.19–4.57) remained significantly associated with a higher risk of LBW. In addition, women who did not use a bed net the night before admission tended to have a higher risk of LBW than those who slept under a bed net (aIRR 1.56; 95 % CI 0.92–2.64).Table 4 Factors associated with low birth weight. Salavan Province, Laos, 2014 Univariate analysisa P value Multivariate analysisa P value Crude IRR (95 % CI) Adjusted IRR (95 % CI) Age (years) (20–24 = Ref) 0.36  <20 1.54 (0.92–2.60)  25–27 1.16 (0.65–2.06)  ≥28 1.01 (0.60–1.68) Primigravidity 1.26 (0.85–1.87) 0.25 Tobacco use 2.50 (1.68–3.72) <10−3 2.43 (1.64–3.60) <0.001 Ethnicity Lao Theung (Ref = Lao Loum) 1.57 (1.03–2.39) 0.04 Place of living (Ref = Salavan) 0.17  Vapi 0.66 (0.31–1.44)  Toumlane 1.51 (0.93–2.44)  Others 0.74 (0.26–2.13) Went to the forest during the current pregnancy 1.22 (0.83–1.82) 0.30 No bed net use 2.07 (1.27–3.36) 0.003 1.56 (0.92–2.64) 0.10 Number of ANC visits (≥4=Ref) 0.01 1–3 1.41 (0.94–2.13) 0 2.53 (1.34–4.77) Gestational hypertension 1.26 (0.48–3.33) 0.64 Moderate anaemia at delivery 1.25 (0.84–1.87) 0.28 Duration of pregnancy (weeks gestation) (37–38 = Ref) <10−3 <0.001  ≥39 0.49 (0.25–0.99) 0.49 (0.25–0.97)  <37 3.47 (2.51–4.81) 3.17 (2.19–4.57) Female 1.15 (0.77–1.70) 0.50 IRR incidence rate ratio aThe analysis was conducted using a modified Poisson regression. Only live-singletons were included. The multivariate analysis was performed on 313 women. The final model was obtained after a backward selection procedure, bed net use was forced in the final model SGA and mean birth weight were considered as secondary outcomes. In univariate analysis, SGA was significantly associated with the same factors as for LBW (tobacco use, bed net use, and ethnicity) (Additional file 1). In contrast, there was no association between SGA and the number of ANC visits and duration of pregnancy. In multivariate analysis, no bed net use and tobacco use remained significantly associated with a higher risk of SGA after a backward selection procedure. When considering birth weight as a continuous variable, the same factors as for LBW as well as primigravidity and a mother’s young age (<20 years) were significantly associated with a reduction in birth weight in univariate analysis (Additional file 2). In addition, there was a marginally significant association between birth weight and gestational hypertension and visit to the forest during current pregnancy. In multivariate analysis, young age (−135 g; 95 % CI −263; −7), tobacco use (−148 g; 95 % CI −286; −10), gestational hypertension (−254 g; 95 % CI −494; −14), a low number of ANC visits (−89 g; 95 % CI −235; −33, in women with one to three visits compared to women with ≥four visits), and pre-term birth (−661 g; 95 % CI, −824; −497) remained significantly associated with a reduction in birth weight. Factors associated with maternal anaemia and mean haemoglobin level at delivery In univariate analysis, women from Toumlane (compared to those from Salavan), of Lao Theung ethnicity, women who did not use a bed net the night before admission, those who made no ANC visits during pregnancy (compared to women who made ≥four ANC visits) and those who did not have iron supplementation during current pregnancy (compared to women who stated that they had iron supplementation the day before admission) were significantly more likely to be anaemic at delivery. In contrast, primigravidae and women living in Vapi District were less likely to be anaemic (Table 5). In multivariate analysis, Lao Theung ethnicity (aIRR 1.49; 95 % CI 0.99–2.25), no iron supplementation (aIRR 1.58; 95 % CI 1.08–2.32) and living in Vapi District (aIRR 0.45; 95 % CI 0.21–0.98) remained significantly associated with anaemia.Table 5 Factors associated with maternal moderate anaemia. Salavan Province, Laos, 2014 Univariate analysisa P value Multivariate analysisa P value Crude IRR (95 % CI) Adjusted IRR (95 % CI) Age (years) (20–24 = Ref) 0.69  <20 1.18 (0.78–1.79)  25–27 1.07 (0.70–1.65)  ≥28 0.90 (0.61–1.33) Primigravidity 0.76 (0.54–1.04) 0.10 Tobacco use 1.31 (0.87–1.96) 0.19 Ethnicity Lao Theung (Ref = Lao Loum) 2.07 (1.56–2.74) <0.001 1.49 (0.99–2.25) 0.05 Place of living (Ref = Salavan) <0.001 0.02 Vapi 0.41 (0.19–0.88) 0.45 (0.21–0.98) Toumlane 1.92 (1.44–2.58) 1.42 (0.94–2.16) Others 0.49 (0.17–1.39) 0.46 (0.16–1.33) Went to the forest during the current pregnancy 1.12 (0.83–1.51) 0.47 No bed net use 1.80 (1.24–2.63) 0.002 Number of ANC visits (≥4 = Ref) 0.007  1–3 1.33 (0.98–1.82)  0 2.09 (1.29–3.39) Gestational hypertension 0.79 (0.30–2.10) 0.64 Iron supplementation (intake the day before admission = Ref) 0.03 0.06  Intake, but not  before admission 1.02 (0.58–1.78) 1.00 (0.56–1.79)  No intake during  pregnancy 1.75 (1.16–2.62) 1.58 (1.08–2.32) Folic acid supplementation (intake the day before  admission = Ref) 0.34  Intake, but not before admission 1.25 (0.76–2.05)  No intake during  pregnancy 0.80 (0.53–1.22) Duration of pregnancy (weeks gestation) (37–38 = Ref) 0.55  ≥39 0.83 (0.56–1.24)  <37 1.11 (0.68–1.83) IRR incidence rate ratio aThe analysis was conducted using a modified Poisson regression. The multivariate analysis was performed on 326 women. The final model was obtained after a backward selection procedure When considering haemoglobin level as a continuous variable, there was a significant reduction in mean Hb level in women living in Toumlane District (compared to those living in Salavan District) and those with a low number of ANC visits (one to three compared to ≥four visits) in multivariate analysis (Additional file 3). In contrast, women living in Vapi District had a significantly higher Hb level. The association between Hb level and iron and folic acid supplementation did not remain significant after adjustment for the number of ANC visits. Discussion This survey assessed the burden and epidemiology of MiP in southern Laos, where malaria transmission remains relatively high despite the large-scale distribution of both ITNs and ACT. In Vapi villages, malaria was detected by PCR in almost 6 % (12/204) of pregnant women. The majority of these infections were asymptomatic and sub-microscopic and, therefore, would not have been detected by recommended national screening strategies [7]. All 12 infections were due to P. vivax, among which one was mixed P. vivax/P. falciparum. Only two women had RDT-positive infections. PCR detected more malarial infections than RDT (frequency of malaria 5.9 vs 1.0 % using PCR vs RDT, respectively). The low proportion of RTD-positive infections is likely to be due to the low sensitivity of the test for low parasitaemia infections [13]. This may also explain the relatively large difference in malaria frequency between PCR and RDT compared to what has been reported in areas of high malaria transmission [14]. Only a single prior study had evaluated MiP in Laos, in Vientiane in 1998. This study was conducted in pregnant women with history of fever, who were recruited during ANC visits. Among 68 women, 16 (23.5 %) were diagnosed with P. falciparum using TBS [15]. The higher prevalence of malaria compared to the present survey may be explained by the selection of symptomatic women at a time when malaria transmission was higher. In addition, the present survey was conducted at community-level in order to reduce selection bias that arises when only those women who attend antenatal clinics are recruited. In the villages almost 6 % of pregnant women were infected with malaria. This proportion may have been higher if longitudinal follow-up had been used, providing a cumulative proportion of infections during pregnancy. For instance, on the Thai-Burmese border, while the mean malaria prevalence during weekly cross-sectional surveys was 4 % using TBS, the cumulative proportion of women infected with malaria during pregnancy was as high as 22.8 % [3]. A history of malaria during current pregnancy was a risk factor for current malarial infections in pregnancy. These infections may correspond to a relapse of a previous P. vivax infection, or they may reflect a higher level of exposure to malaria of these women, compared to those who did not report previous infections. At delivery, there was a lower prevalence of sub-microscopic infections, which varied according to antenatal clinic. In Vapi Hospital, the prevalence (4.5 %) was comparable to what had been found in the villages. However, malaria was not detected in pregnant women from Toumlane, although the sample size was very low, nor in women who delivered in Salavan Hospital, which is located in a more urban area than Vapi and where malaria transmission is probably lower. The comparison of these results with those from studies conducted in the Asia–Pacific region is difficult since malaria transmission is highly variable in this region and most of these studies included both symptomatic and asymptomatic pregnant women. On the Thai-Burmese border, where malaria transmission and women’s characteristics may be the most comparable to those in Laos, the prevalence of peripheral and placental malaria at delivery using TBS or PCR ranged from 3.0 to 10.9 % [3]. In the present study, a single sub-microscopic placental infection was detected. This may be explained by the fact that most infections (86 %) were due to P. vivax, which is less likely to sequester in the placenta than P. falciparum. In both surveys, only two out of a total of 14 infections were caused by P. falciparum, and Plasmodium knowlesi was not detected. This is in accordance with the increasing ratio between P. vivax and P. falciparum infections, which is observed in the region [3]. The higher prevalence of P. vivax compared to P. falciparum infections may also be explained by the inclusion of asymptomatic women, P. falciparum infections being more likely to be symptomatic. Plasmodium vivax infection during pregnancy has been associated with poor maternal and child outcomes [6, 16], although its deleterious effects are lower than those observed with P. falciparum. A single episode of vivax or falciparum malaria can have deleterious consequences for the foetus [6, 17]. Asymptomatic and sub-microscopic infections have also been shown to be detrimental [6, 17, 18]. In this study, the direct effect of malaria on birth weight could not be assessed because of the very low number of infected women at delivery. Women who did not sleep under a bed net the night before the survey tended to have a higher risk of LBW and newborns with reduced birth weight, compared to women who used a bed net. The reduction in birth weight (−140 g) was of the same magnitude as that reported for MiP elsewhere in the region [3], making lack of bed net use a possibly proxy for exposure to malaria in this study. However, it has been shown that mosquitoes in this region have a propensity to feed outdoors and at times when women are still active and not under a bed net [19, 20]. Therefore, the association between no bed net use and a higher risk of birth weight may be explained by confounding factors, which were not taken into account in the analysis, such as a lower socio-economic status, a lower education level and worse health practices in women who did not use a bed net. A high proportion of women (24 %) had LBW newborns, mostly due to IUGR. Indeed, newborns were classified as SGA, as a proxy for IUGR, in 23 % of cases with 80 % of LBW babies who were SGA. This is in accordance with the prevalence of LBW and the proportion of LBW due to IUGR in Asia [2]. In multivariate analysis, factors significantly associated with LBW were pre-term birth and smoking. Women who reported cigarette smoking during current pregnancy had SGA and LBW rates 2.0 and 2.4 times greater, which correspond to what has been reported elsewhere in developed countries [21]. Since 11.5 % of women reported smoking during pregnancy, the attributable risk for LBW due to smoking was as high as 14 % in this population. In addition, mothers of young age (<20 years old), a low number of ANC visits (one to three compared to ≥four visits), gestational hypertension and pre-term birth were associated with a significant reduction in birth weight (−135, −89, −254 and −661 g, respectively). These associations were found after adjustment for gravidity, which is likely correlated with these factors. Indeed, gestational hypertension and maternal age have been recognized as independent risk factors for LBW-IUGR [22, 23]. However, the association between maternal age and LBW may also have been partially confounded by maternal under-nutrition, which was not adjusted for in the present analysis. Thirty-four per cent of women were anaemic at delivery. Women who did not take iron supplementation during pregnancy had a 58 % higher risk of anaemia compared to women who stated that they took iron the day before the survey. These results highlight the importance of the prevention and treatment of gestational hypertension, smoking and iron deficiency, which are preventable risk factors for IUGR and anaemia. Women from Lao Theung ethnicity were more likely to be anaemic than women from Lao Loum ethnicity. In addition, women living in Vapi District had a lower risk for anaemia than those living in Salavan. This risk differences may be due to residual confounding, such as food habits, maternal nutritional and socio-economic status, which were not taken into account in the analysis. The strength of the present analysis is the use of population-based data to assess the prevalence of malaria during pregnancy, limiting selection bias that arises when women are recruited at facility-level. However, the participation in the study of all eligible women in the villages cannot be warranted. Moreover, fewer women than anticipated were recruited both in the villages and at delivery, which may have led to estimating the prevalence of malaria with some inaccuracy. In Vapi and Toumlane District Hospitals all women were included at delivery, which was not the case in Salavan Provincial Hospital where a very few number of women could not be included before an emergency delivery. Finally, the high variability of malaria prevalence in south Laos limits the generalization of our results [4]. The effect of MiP on maternal anaemia and birth weight could not be assessed because of the low number of malarial infections at delivery. This will require either a higher sample size at delivery or a longitudinal follow-up of pregnant women with repeated malaria screening during pregnancy. Another weakness of this study is the limit of confounders available. Indeed, the analysis was adjusted for the available confounders, but there still remains the possibility of residual confounding. Indeed, controlling for socio-economic characteristics and maternal nutritional status may probably have attenuated the risk of LBW and anaemia with young maternal age, ethnicity, place of living and bed net use. In addition, some variables such as bed net use, history of malaria and gestational age during pregnancy were self-reported by the women and could not be verified, leading to possible misclassifications. Conclusions This survey contributes a first step towards the assessment of MiP in Laos, where the prevalence of P. vivax sub-microscopic infections in this population was noticeable. However, these findings cannot be generalized beyond the present study area since malaria transmission is highly heterogeneous within the country [4]. Further studies in southern Laos are needed to draw a more representative picture of MiP, including both asymptomatic and sub-microscopic Plasmodium infections that have deleterious consequences during pregnancy and contribute to malaria transmission [13]. Systematic detection and treatment of malaria in women during ANC visits, regardless of whether they are symptomatic or not, may reduce the consequences of MiP [24]. In addition, the use of ITNs throughout pregnancy should remain a priority. Additional preventive strategies such as intermittent preventive treatment in pregnancy (IPTp) as recommended in high malaria transmission areas, may contribute to reduce the consequences of sub-microscopic infections, as well as infections which occur between ANC visits. Artemisinin-based combination therapies have been shown to be safe during pregnancy [25] and efficacious against both P. falciparum and P. vivax in such areas [26]. In particular, IPTp with dihydroartemisinin-piperaquine may be an interesting option, as recently reported in Africa [27]. Additional files 10.1186/s12936-016-1492-2 Factors associated with small-for-gestational age (SGA). Salavan Province, Laos, 2014. Maternal characteristics associated with small-for-gestational age in both univariate and multivariate analysis using a modified Poisson regression model. 10.1186/s12936-016-1492-2 Factors associated with mean birth weight. Salavan Province, Laos, 2014. Maternal characteristics associated with mean birth weight in both univariate and multivariate analysis using a linear regression model. 10.1186/s12936-016-1492-2 Factors associated with mean haemoglobin level at delivery. Salavan Province, Laos, 2014. Maternal characteristics associated with mean haemoglobin level at delivery in both univariate and multivariate analysis using a linear regression model. Abbreviations MiPmalaria in pregnancy LBWlow birth weight ANCantenatal care ITNinsecticide-treated net IPTpintermittent preventive treatment in pregnancy ACTartemisinin-based combination therapy Hbhaemoglobin TBSthick blood smear RDTrapid diagnostic test RT-qPCRreal-time quantitative polymerase chain reaction IUGRintrauterine growth retardation IRRincidence rate ratio aIRRadjusted incidence rate ratio CIconfidence interval IQRinterquartile range CMPECentre of Malariology, Parasitology and Entomology PALULAOprevalence of malaria in South Lao PDR (France Expertise Internationale funded programme) Authors’ contributions VB drafted the manuscript and participated in the design and coordination of the PALULAO study; JYLH, NF, PNN, and MM helped to draft the manuscript; JYLH, NF, MM, BH, KC, and SK participated in the design and coordination of study; YI and NV supervised the field work; GB, NF and PNN supervised the realisation of the RT-qPCRs; SH performed the RT-qPCR quality control in France. All authors read and approved the final manuscript. Acknowledgements We acknowledge the valuable contribution of the women, midwives, nurses, and assistants who participated in the study. Competing interests The corresponding author had full access to all data in the study and had final responsibility for the decision to submit for publication. None of the authors in this paper declare any competing interests, including financial interests and relationships and affiliations relevant to the subject of the manuscript. Availability of data and materials The dataset supporting the conclusions of this article is available in the Open Science Framework repository, in https://www.osf.io/9g2ku/files/. Ethics approval and consent to participate The Lao National Ethical Committee for Health Research (NECHR) and the Lao Ministry of Health approved the study protocol. Both oral and written communication was used to provide information to the women who were interested in participating in the study. Those agreeing to participate were asked to provide a signed (or thumb printed for illiterate women) written informed consent. Funding This work supported by France Expertise Internationale (now called Expertise France) contract number: 13INI144. PN and MM are supported by the Wellcome Trust. ==== Refs References 1. Dellicour S Tatem AJ Guerra CA Snow RW ter Kuile FO Quantifying the number of pregnancies at risk of malaria in 2007: a demographic study PLoS Med 2010 7 e1000221 10.1371/journal.pmed.1000221 20126256 2. Katz J Lee AC Kozuki N Lawn JE Cousens S Blencowe H Mortality risk in preterm and small-for-gestational-age infants in low-income and middle-income countries: a pooled country analysis Lancet 2013 382 417 425 10.1016/S0140-6736(13)60993-9 23746775 3. Rijken MJ McGready R Boel ME Poespoprodjo R Singh N Syafruddin D Malaria in pregnancy in the Asia-Pacific region Lancet Infect Dis 2012 12 75 88 10.1016/S1473-3099(11)70315-2 22192132 4. Jorgensen P Nambanya S Gopinath D Hongvanthong B Luangphengsouk K Bell D High heterogeneity in Plasmodium falciparum risk illustrates the need for detailed mapping to guide resource allocation: a new malaria risk map of the Lao People’s Democratic Republic Malar J 2010 9 59 10.1186/1475-2875-9-59 20181273 5. Edwards HM Canavati SE Rang C Ly P Sovannaroth S Canier L Novel cross-border approaches to optimise identification of asymptomatic and artemisinin-resistant Plasmodium infection in mobile populations crossing Cambodian borders PLoS One 2015 10 e0124300 10.1371/journal.pone.0124300 26352262 6. McGready R Lee SJ Wiladphaingern J Ashley EA Rijken MJ Boel M Adverse effects of falciparum and vivax malaria and the safety of antimalarial treatment in early pregnancy: a population-based study Lancet Infect Dis 2012 12 388 396 10.1016/S1473-3099(11)70339-5 22169409 7. WHO Global Malaria Programme. World Malaria Report 2015. Geneva: World Health Organization; 2015. http://www.who.int/malaria/publications/world-malaria-report-2015/report/en/. Accessed 1 April 2016. 8. Plowe CV Djimde A Bouare M Doumbo O Wellems TE Pyrimethamine and proguanil resistance-conferring mutations in Plasmodium falciparum dihydrofolate reductase: polymerase chain reaction methods for surveillance in Africa Am J Trop Med Hyg 1995 52 565 568 7611566 9. Diallo A Ndam NT Moussiliou A Dos Santos S Ndonky A Borderon M Asymptomatic carriage of Plasmodium in urban Dakar: the risk of malaria should not be underestimated PLoS One 2012 7 e31100 10.1371/journal.pone.0031100 22363558 10. Planche T Krishna S Kombila M Engel K Faucher JF Ngou-Milama E Comparison of methods for the rapid laboratory assessment of children with malaria Am J Trop Med Hyg 2001 65 599 602 11716121 11. Zou G A modified Poisson regression approach to prospective studies with binary data Am J Epidemiol 2004 159 702 706 10.1093/aje/kwh090 15033648 12. Papageorghiou AT Ohuma EO Altman DG Todros T Cheikh Ismail L Lambert A International standards for fetal growth based on serial ultrasound measurements: the Fetal Growth Longitudinal Study of the INTERGROWTH-21st Project Lancet 2014 384 869 879 10.1016/S0140-6736(14)61490-2 25209488 13. Imwong M Nguyen TN Tripura R Peto TJ Lee SJ Lwin KM The epidemiology of subclinical malaria infections in South-East Asia: findings from cross-sectional surveys in Thailand-Myanmar border areas, Cambodia, and Vietnam Malar J 2015 14 381 10.1186/s12936-015-0906-x 26424000 14. Walker-Abbey A Djokam RR Eno A Leke RF Titanji VP Fogako J Malaria in pregnant Cameroonian women: the effect of age and gravidity on submicroscopic and mixed-species infections and multiple parasite genotypes Am J Trop Med Hyg 2005 72 229 235 15772312 15. Sychareun V Phengsavanh A Kitysivoilaphanh B Prabouasone K Viriyavejakul P Krudsood S A study of anemia in pregnant women with Plasmodium falciparum at district hospitals in Vientiane, Lao PDR Southeast Asian J Trop Med Publ Health 2000 31 Suppl 1 91 98 16. Nosten F McGready R Simpson JA Thwai KL Balkan S Cho T Effects of Plasmodium vivax malaria in pregnancy Lancet 1999 354 546 549 10.1016/S0140-6736(98)09247-2 10470698 17. Rijken MJ Papageorghiou AT Thiptharakun S Kiricharoen S Dwell SL Wiladphaingern J Ultrasound evidence of early fetal growth restriction after maternal malaria infection PLoS One 2012 7 e31411 10.1371/journal.pone.0031411 22347473 18. Cottrell G Moussiliou A Luty AJ Cot M Fievet N Massougbodji A Submicroscopic Plasmodium falciparum infections are associated with maternal anemia, premature births, and low birth weight Clin Infect Dis 2015 60 1481 1488 25694651 19. Muenworn V Sungvornyothin S Kongmee M Polsomboon S Bangs MJ Akrathanakul P Biting activity and host preference of the malaria vectors Anopheles maculatus and Anopheles sawadwongporni (Diptera: Culicidae) in Thailand J Vector Ecol 2009 34 62 69 10.1111/j.1948-7134.2009.00008.x 20836806 20. Tananchai C Tisgratog R Juntarajumnong W Grieco JP Manguin S Prabaripai A Species diversity and biting activity of Anopheles dirus and Anopheles baimaii (Diptera: Culicidae) in a malaria prone area of western Thailand Parasit Vectors 2012 5 211 10.1186/1756-3305-5-211 23009133 21. Blatt K Moore E Chen A Van Hook J DeFranco EA Association of reported trimester-specific smoking cessation with fetal growth restriction Obstet Gynecol 2015 125 1452 1459 10.1097/AOG.0000000000000679 26000517 22. Kramer MS Intrauterine growth and gestational duration determinants Pediatrics 1987 80 502 511 3658568 23. Kozuki N Lee AC Silveira MF Sania A Vogel JP Adair L The associations of parity and maternal age with small-for-gestational-age, preterm, and neonatal and infant mortality: a meta-analysis BMC Publ Health 2013 13 Suppl 3 S2 10.1186/1471-2458-13-S3-S2 24. Nosten F McGready R Mutabingwa T Case management of malaria in pregnancy Lancet Infect Dis 2007 7 118 125 10.1016/S1473-3099(07)70023-3 17251082 25. Moore KA, Simpson JA, Paw MK, Pimanpanarak M, Wiladphaingern J, Rijken MJ,et al. Safety of artemisinins in first trimester of prospectively followed pregnancies: an observational study. Lancet Infect Dis. 2016. pii: S1473-3099(15)00547-2. (Epub ahead of print). 26. Poespoprodjo JR Fobia W Kenangalem E Lampah DA Sugiarto P Tjitra E Dihydroartemisinin-piperaquine treatment of multidrug resistant falciparum and vivax malaria in pregnancy PLoS One 2014 9 e84976 10.1371/journal.pone.0084976 24465458 27. Desai M Gutman J L’lanziva A Otieno K Juma E Kariuki S Intermittent screening and treatment or intermittent preventive treatment with dihydroartemisinin-piperaquine versus intermittent preventive treatment with sulfadoxine-pyrimethamine for the control of malaria during pregnancy in western Kenya: an open-label, three-group, randomised controlled superiority trial Lancet 2015 386 2507 2519 10.1016/S0140-6736(15)00310-4 26429700

PMC005xxxxxx/PMC5002161.txt

==== Front BMC GastroenterolBMC GastroenterolBMC Gastroenterology1471-230XBioMed Central London 51310.1186/s12876-016-0513-5Research ArticleDistribution and clinical significance of hepatitis B virus genotypes in Pakistan http://orcid.org/0000-0001-6680-3541Mahmood Majid majid1133@gmail.com 12Anwar Muhammad Asim asimanwar73@yahoo.com 3Khanum Azra azrakhanum@uaar.edu.pk 4Zaman Nasib nasibzaman@gmail.com 5Raza Abida abida_rao@yahoo.com 61 Department of Zoology, The University of Poonch, Rawalakot, Azad Jammu and Kashmir 12350 Pakistan 2 Department of Zoology, Pir Mehr Ali Shah Arid Agriculture University, Rawalpindi, 46300 Pakistan 3 Department of General Medicine, Pakistan Atomic Energy Commission (PAEC) General Hospital, Islamabad, 44000 Pakistan 4 Barani Institute of Management Sciences (BIMS), Rawalpindi, 46300 Pakistan 5 Center for Biotechnology & Microbiology, University of Swat, Swat, 19130 Pakistan 6 Diagnostic Department, Nuclear Medicine, Oncology and Radiotherapy Institute (NORI), Islamabad, 44000 Pakistan 26 8 2016 26 8 2016 2016 16 1 10419 4 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 Hepatitis B virus (HBV) genotype and its role in disease progression and patients’ response to antiviral treatment, is not well studied in Pakistan. This comprehensive study was aimed to determine the distribution of HBV genotypes in Pakistan and their possible association with phases of HBV infection. Methods A total of 840 HBsAg positive samples was collected and tested for HBV DNA quantity. Samples below 100 IU/ml were excluded from the study. A total of 715 samples representing all the six parts of the country were genotyped by type specific primer PCR method. Clinical data of only 384 patients was compared as the remaining 332 were either receiving antiviral treatment or their infection phase was not confirmed. Results Genotype D was found in 509 samples (71.2 %), genotype A in 55 samples (7.7 %) and mixed infection with genotypes A and D in 124 samples (17.3 %). Genotypes B, C and E were identified in less than 1 % of the total samples. Genotype A, D and their mixture (A + D) were compared for severity of HBV infection. Significant differences were not found in distribution of HBV genotypes among different disease stages. Conclusion HBV genotype D was the predominant infection in all study areas of Pakistan followed by mixed genotypes infection (A + D) whereas genotype A has 10 times lower prevalence than genotype D. Genotypes B, C, E and F altogether make only 1.5 % of the prevalence. Genotype do not appears to show the severity of liver disease. Keywords Hepatitis B virusHBV genotypesLiver disease progressionPCRhttp://dx.doi.org/10.13039/501100004681Higher Education Commission, PakistanPIN: 074-2673-Bm4-145Mahmood Majid issue-copyright-statement© The Author(s) 2016 ==== Body Background Hepatitis B virus (HBV) infection has been a major health problem as about 400 million people carry hepatitis B surface antigen worldwide [1], out of which, nearly 1 million die every year [2]. The infection rate of HBV has decreased significantly in developed countries like United States of America [3], where the acute HBV infection rate has fallen by 78 % during 1990–2005 [4]. Unfortunately, in developing and underdeveloped countries including Pakistan, the infection rate seems to have not decreased, even to any appreciable level. The failure in tackling the infection has been related to the high cost of antiviral drugs, lack of vaccination and public awareness about infection and the mismanagement to address the problem in Pakistan [5]. Eight HBV genotypes have been established on the basis of 8 % or more nucleotide divergence in the genome. These genotypes are named as A, B, C, D [6], E, F [7, 8], G [9] and H [10] which are common [11], while genotype I [12, 13] and J [14] are also introduced as new genotypes but their status is questionable. HBV genotypes have specific pattern of distribution in different geographic regions and ethnic groups of the world. Genotype A is prevalent in Brazil, USA, Canada, Northwest Europe, South Asia, Central African countries, Tunisia and Benin [15–17]. Genotype B is common in Japan, Taiwan, Philippines, Hong Kong, China, Vietnam, Thailand, Indonesia and United States of America. Genotype C occurs in Australia, Polynesia, Melanesia, Micronesia, Indonesia, China, Hong Kong, Vietnam, Thailand, Japan, Korea, Taiwan, India, Solomon Islands, Brazil and USA. Genotype D is predominant in Mediterranean region, Spain, Albania, Czech Republic, Russia, Turkey, Middle East, Iran, Afghanistan, South Asia, Solomon Islands, Tunisia, Polynesia, Melanesia, Micronesia, Brazil and USA [17–19]. Genotype E is endemic to Africa where it occurs only in some countries of the Western part of the continent while genotype F is widely distributed in new world. It has been reported from Alaska, Argentina, South America, Central America and Polynesia [17, 20]. Genotypes G has been recorded from North America, France and Germany, genotype H from Central America, South America and Mexico while genotypes I and J were reported only from Vietnam and Japan respectively [17]. HBV genotypes have been reported to have significant association with progression of liver diseases, risk of cirrhosis, development of hepatocellular carcinoma (HCC), viral load, HBsAg sero-clearance, HBeAg sero-conversion and response to antiviral therapy [21]. Co-infection with two different virus genotypes has been reported to be associated with worse prognosis of the disease. Some studies showed that co-infection with HBV/B and HBV/C is associated with high viral load and more severe liver disease as compared to single genotype infection [22, 23]. The patients with HBV/A have been reported to be more sensitive to interferon α treatment as compared to those infected with HBV/D [24] while the patients having infection with HBV/B have a higher response rate to interferon α treatment as compared to the patients infected with HBV/C [25]. Moreover, infection with genotype C alone was also found to be associated with significantly higher risks of liver cirrhosis and hepatocellular carcinoma as compared to genotype B infection [26, 27] while HBV/B shown an earlier HBsAg sero-clearance rate as compared to HBV/C [28]. In another study, HBV/A was found to have significantly higher sero-conversion rate of HBeAg as compared to HBV/B, HBV/C, HBV/D and HBV/F [29]. From all the studies mentioned above it’s quite clear that genotypes influence disease condition as well as response to treatment. The picture regarding distribution of HBV genotypes in Pakistan and its association with disease stage has not been clearly understood yet. In fact, all of the available studies are mostly confined to specific cities/areas like Karachi (Sind) and Lahore (Punjab) and none of them contain data from other parts of the country. A vast range of areas from different regions of Pakistan still remains unexplored for HBV genotypes. Moreover, the existing state of knowledge regarding the distribution and prevalence of HBV genotypes in Pakistan seems to be sketchy and questionable too. For example, four studies [5, 30–32] are in agreement that genotype D infection is most prevalent in the country while three studies [18, 33, 34] reported that genotype C is the most common ones in Pakistan. Yet in an another study, genotype C and D have been cited as most common HBV genotypes in Pakistan [35]. In view of the presence of the inadequate data, it was felt that a comprehensive study is required to determine the current distribution of HBV genotypes in all the six provinces of Pakistan and their possible association with different phases of HBV infection. Methods Patients Eight hundred and forty (840) samples from HBsAg positive patients were collected during October 2010 to March 2013 from different locations of six regions of Pakistan which include Punjab and Federal Capital Territory of Islamabad (n = 359); Khyber Pakhtoonkhwa along with Federally Administered Tribal Areas (n = 115); Azad Jammu and Kashmir (n = 81); Gilgit-Baltistan (n = 111); Sind province including Karachi city (n = 107); and Baluchistan (n = 67). The sampling sites and the details of patients are given in Fig. 1 and Table 1 respectively. All the patients were diagnosed as chronic HBV carriers. Serum samples were acquired from the different hospitals and diagnostic laboratories of the above regions where they were originally collected for routine diagnostic purposes. All the available data of clinical and demographic importance were collected at the time of sample collection by filling out a questionnaire. All the samples were then brought to Nuclear Medicine, Oncology and Radiotherapy Institute (NORI), Islamabad and stored at −25 °C for further analysis. The samples were tested for positivity and quantity of HBV DNA. Only the samples having more than 100 IU/ml of viral load were subjected to HBV genotyping (Table 1). The samples with lower than 100 IU/ml of viral load were excluded because their genotyping was not successful in most of the cases. All the routine tests including HBsAg, ALT, and anti-HBe were performed in Central Laboratories, NORI using commercially available kits (Abott Diagnostics, USA).Fig. 1 The map shows the sampling sites in all regions of Pakistan. Sites are indicated with numerals. Sites: 1. Mirpur; 2. Kotli; 3. Palandari; 4. Rawalakot; 5. Bagh; 6. Kahuta; 7. Muzaffarabad; 8. Neelum; 9. Astor; 10. Sikardu; 11. Diamir; 12. Gilgit; 13. Hunza; 14. Chitral; 15. Dir; 16. Swat; 17. Kohistan; 18. Batgram; 19. Mansehra; 20. Abottabad; 21. Haripur; 22. Nowshera; 23. Mardan; 24. Peshawar; 25. Kohat; 26. Hangu; 27. Bannu; 28. LakkiMarwat; 29. D I Khan; 30. Wana; 31. Parachinar; 32. Islamabad; 33. Rawalpindi; 34; Chakwal; 35. Jhelum; 36. Gujrat; 37. Sialkot; 38. Gujranwala; 39. Sargodha; 40. Mianwali; 41. Bhakkar; 42. Jhang; 43. Faisalabad; 44. Lahore; 45. Sahiwal; 46. Multan; 47. D.G. Khan; 48. Bahawalpur; 49. Rahim Yar Khan; 50. Sukhar; 51. Khairpur; 52. Nawabshah; 53. Hyderabad; 54. Mirpur Khas; 55. Karachi; 56. Turbat; 57; Khuzdar; 58. Sibi; 59. Loralai; 60. Quetta. Regions: 1–8, Azad Jammu & Kashmir; 9–13, Gilgit-Baltistan; 14–31, Khyber Pakhtunkhwa; 32, Federal Capital Territory of Islamabad (FCT); 33–49, Punjab; 50–55, Sind; 56–60, Baluchitan Table 1 Samples collected and selected for HBV genotyping from different regions of the country Total no of samples (No of male and female samples) Punjab and FCT KP and FATA AJK Gilgit-Baltistan Sind Baluchistan Total Total collected 359 (243 + 116) 115 (78 + 37) 81 (50 + 31) 111 (83 + 28) 107 (68 + 39) 67 (39 + 28) 840 (561 + 279) Negative HBV DNA 29 (18 + 11) 7 (5 + 2) 6 (4 + 2) 8 (6 + 2) 13 (7 + 6) 6 (4 + 2) 69 (44 + 25) <100 IU/ml 19 (11 + 8) 9 (6 + 3) 7 (4 + 3) 10 (6 + 4) 10 (7 + 3) 1 (0 + 1) 56 (34 + 22) Total Genotyped 311 (214 + 97) 99 (67 + 32) 68 (42 + 26) 93 (71 + 22) 84 (54 + 30) 60 (35 + 25) 715 (483 + 232) Abbreviations: FCT federal capital territory, KP Khyber Pkhtoonkhwa, FATA federally administered tribal areas, AJK Azad Jammu and Kashmir The stage of HBV infection was primarily assigned by the local physician and further confirmed by the status of different HBV markers and biochemical tests. Only 384 patients were included in the study of genotypes influence on disease stage. All the patients (331) who had already received antiviral treatment or had co-infection with HIV, HCV or HDV were excluded from this part of study. The liver disease stage was primary assigned by classified gastroenterologists and then the disease phase was further confirmed on the base of patients’ HBV viral load, ALT levels, HBeAg status and the presence or absence of anti HBe antibodies (Table 3). The study was approved by institutional ethics committee of PMAS Arid Agriculture University Rawalpindi and each patient gave written consent. HBV DNA extraction and quantification HBV DNA was extracted from the plasma according to manufacturer’s protocol using “Instant Virus DNA kit” (AJ Roboscreen, Analytikajena Biosolutions, GmbH, Germany) while the DNA was quantified using RoboGene® HBV Quantification kit (AJ Roboscreen, analytikajena Biosolutions, GmbH, Germany) according to manufacturer’s instructions. HBV genotyping The genotyping was performed using PCR method with genotype specific primers described by Naito et al. [36] with some modifications in cycling profile and PCR constituents. For regular PCR, 10 μl of GoTaq® Green Master mix (Promega, USA), 1 μl each of universal primers P1A and S1-2, 2 μl of ddH2O and 6 μl of diluted template (DNA) were used to make 20 μl of reaction volume. The thermal profile was; 10 min at 95 °C, then 30 cycles of; 94 °C for 20 s, 55 °C for 20 s and 72 °C for 1 min, then another 7 min at 72 °C (ABI 9700, USA). Two mixes were used in nested PCR. Mix 1 had the forward primers for genotype A, B and C with a common reverse primer while mix two had reverse primers for genotypes D, E and F with common forward primer. Two μl of product from 1st round of PCR were added to each mix as DNA template with 8 μl of GoTaq® Green Master mix, 8 μl of ddH2O and 1 μl of each primer to make 22 μl of final volume. The cycling profile for nested PCR was; 95 °C for 10 min, then 40 cycles of 94 °C for 45 s, 63 °C for 20 s, and 72 °C for 60 s, and then another 7 min at 72 °C (ABI 9700, USA). Products were run on 2 % agarose gel which was stained in a solution of ethidium bromide, and then observed under ultraviolet fluorescence (BioRad Gel Doc-XR, USA). Amplified products in mix 1 were 68 bp, 281 bp and 122 bp in size for genotypes A, B and C respectively while the product sizes in mix 2 were 119 bp, 167 bp and 97 bp for genotypes D, E and F respectively. The product’s size was estimated with reference to a 50-bp DNA ladder (GeneRuler™ Fermentas) and the genotype was determined for each sample according to the sizes of DNA product obtained in both mixes (Fig. 2). The term untypable was assigned to those samples which had higher than 100 IU/ml of viral load but were not successfully genotype or their genotype was not identified by the method.Fig. 2 Amplified PCR products on 2 % gel showing bands for four genotypes. Lanes 2, 4, 6, 8, 10, 12, 14 and 16: the product bands of mix 1. Lanes 3, 5, 7, 9, 11, 13, 15 and 17: the product bands of mix 2. Lanes 1 and 18: 50 bp DNA ladder. lanes 2 + 3: genotype D. lanes 4 + 5: genotype A. lanes 6 + 7: genotype A + D. lanes 8 + 9: genotype A + B + D. lanes 10 + 11: genotype C and a nonspecific band. lanes 12 + 13: genotype B + D. lanes 14 + 15: genotype A. lanes 16 + 17; Genotype A + D Statistical analysis Chi square test was applied to evaluate the relationship HBV genotype and HBV infection phase with each other as well as with patients’ gender while one way analysis of variance (ANOVA) was performed to analyze the relationship of HBV genotype and HBV infection phase with the age of patients. A value of p less than 0.05 was considered as statistically significant. SPSS version-16.0 was used for all the analyses. Results Out of total 840 samples, 125 were excluded from the study as HBV DNA was not detected in them. The remaining 715 samples were further analyzed (Table 1). Regional distribution of HBV genotypes Out of the 715 samples taken for genotyping, 699 were successfully genotyped while 16 (2.2 %) remained untypable. HBV genotypes A, B, C, D and E were detected in 55 (7.7 %), 4 (0.6 %), 6 (0.8 %), 509 (71.2 %) and 1 (0.1 %) samples respectively (Table 2). A total of 124 (17.3 %) patients were infected with more than one genotype i.e. mix genotype infection. The mix genotype infection comprised six combinations in different proportions which were - A + B (0.3 %), A + D (13.7 %), B + D (1.9 %), C + D (1.0 %), E + D (0.1 %) and A + B + D (0.3 %).Table 2 Distribution of HBV genotypes and their combinations in the samples taken from different regions of Pakistan. The numbers and proportions (%) are given against each genotype or the genotypic combination Genotype KP and FATA AJK Gilgit-Baltistan Sind Punjab and FCT Baluchistan Total A 7 (7) 3 (4.4) 3 (3.2) 9 (10.7) 31 (10) 2 (3.3) 55 (7.7) B 2 (2) 0 1 (1.1) 0 1 (0.3) 0 4 (0.6) C 3 (3) 1 (1.5) 0 0 1 (0.3) 1 (1.7) 6 (0.8) D 67 (68) 56 (82.3) 75 (80.6) 61 (72.6) 197 (63.3) 53 (88.3) 509 (71.2) E 0 0 0 1 (1.2) 0 0 1 (0.1) Mix 18 (18) 8 (11.8) 11 (12.8) 11 (13) 73 (23.5) 3 (5) 124 (17.3)  A + B 2 (2.0) 0 0 0 0 0 2 (0.3)  A + D 9 (9.1) 5 (7.3) 4 (4.3) 9 (10.7) 68 (21.8) 3 (5.0) 98 (13.7)  B + D 2 (2.0) 0 6 (6.4) 2 (2.3) 4 (1.3) 0 14 (1.9)  C + D 4 (2.0) 3 (4.4) 0 0 0 0 7 (1.0)  E + D 0 0 0 0 1 (0.3) 0 1 (0.1)  A + B + D 1 (1.0) 0 1 (1.1) 0 0 0 2 (0.3) Untypable 2 (2.0) 0 3 (3.2) 2 (2.3) 8 (2.5) 1 (1.7) 16 (2.2) Total 99 68 93 84 311 60 715 Concerning the regions included in this study, genotype D was always the predominant one with the prevalence ranging from 63.3 to 88.3 % (Table 2). Genotype A and the mix genotype combination A + D was also present in all regions. The prevalence of genotype A, D and mix genotypes respectively was 10 %, 63.3 % and 23.5 % in Punjab and FCT region; 7 %, 68 % and 18 % in KP and FATA region; 4.4 %, 82.3 % and 11.8 % in AJK region; 3.2 %, 80.6 % and 12.8 % in Gilgilt-Baltistan region; 10.7 %, 72.6 % and 13 % in Sind region; 3.3 %, 88.3 % and 5 % in Baluchistan region (Fig. 3).Fig. 3 Genotype prevalence of hepatitis B virus in all regions of the country. UT stands for un-typable and mix means all the infections with more than one genotypes Multiple genotype infections presented the second more prevalent group of genotypes with 17.3 % of the total samples infected with 6 different combinations of genotypes (Table 2). Mix infection with A + D genotype was the major combination found among multiple genotype infections which was 98/124 of the samples. Other mix genotype combinations identified were B + D, C + D, A + B, A + B + D and E + D which were 1.9 %, 1.0 %, 0.3 %, 0.3 % and 0.1 % of the samples respectively. None of the combinations except A + D was found in more than 2 % of the samples. This combination was detected from all parts of the Pakistan, though it was highest in Punjab and FCT region (21.8 %) as compared to all other parts (Table 2). Distribution HBV genotypes in different phases of disease All statistical comparisons performed for genotypes A, D and A + D as the other genotypes and their combinations in mix infections were rare (Table 2). Out of total 715 patients included in the study, 331 (46.3 %) had either already received antiviral treatment or had co-infection with HIV, HCV or HDV or had genotype other than A and D. Thus the remaining 384 (53.7 %) patients who were not exposed to any treatment till the time of sample collection were classified in four different phases of liver disease: 56 (14.6 %) were in immune control phase, 89 (23.2 %) in immune tolerant phase, 153 (39.8 %) in immune clearance phase and 86 (22.4 %) in immune escape phase (Table 3). None of the patients had liver cirrhosis or hepatocellular carcinoma in this study. The mean age of the patients and gender ratio was statistically equal among all the disease phases. Genotype D (n = 279) was the most prevalent in all groups of patients with 36 patients in immune control group, 67 in immune tolerant group, 112 in immune clearance and 64 in immune escape group. The second more prevalent was the mix infection A + D (n = 62) with 13, 12, 26 and 11 cases in immune control, immune tolerant, immune clearance and immune escape groups respectively. Genotype A (n = 43) was the least prevalent with 7, 10, 15 and 11 cases in immune control, immune tolerant, immune clearance and immune escape phases respectively. The differences in distribution of genotypes in disease phases were not statistically significant. Similarly, there were no statistical differences of age and gender among the three genotype groups compared (Table 3).Table 3 Characteristics of patients infected with different HBV genotypes (n = 384) Characteristics Genotype A Genotype D Genotype A + D Sig. Number of patients (%) 43 (11.2) 279 (72.7) 62 (16.1) Age (Mean ± SD) 32.4 ± 11.7 34.1 ± 12.1 35.1 ± 12.8 NSa Gender (M/F) 32/11 202/77 40/22 NSb Immune Control (N = 56) normal ALT, HBeAg negative, anti HBe positive, low viral load 7 (16.3 %) 36 (12.9 %) 13 (21.0 %) NSb Immune Tolerant (N = 89) normal ALT, HBeAg positive, anti HBe negative, high viral load 10 (23.2 %) 67 (24.0 %) 12 (19.3 %) NSb Immune Clearance (N = 153) elevated ALT, HBeAg positive, high viral load 15 (34.9 %) 112 (40.1 %) 26 (41.9 %) NSb Immune Escape (N = 86) Normal or elevated ALT, HBeAg negative, anti HBe positive, High HBV DNA, 11 (25.6 %) 64 (22.9 %) 11 (17.7 %) NSb aOne way Analysis of Variance, bPearson Chi Square Test, NS for non significant Discussion The present study is the first of its kind which included a cohort of 715 patients with chronic HBV infection from all six regions of Pakistan. The study analyzed the patients for HBV genotypes and their possible associations with liver disease severity in 384 consecutive chronic HBV carriers. According to the results, genotypes A, B, C, D, E and mix genotype infections with 6 different genotypic combinations are present in Pakistan, but the majority of patients (92.6 %) were infected with either genotype D, or genotype A or a mixture of both (A + D). The pre-dominant genotype in all regions was D followed by genotype A and mix genotype infection with A + D. Genotypes B, C, E and F were rare and collectively form only 1.5 % of the total prevalence. HBV genotypes were not studied exclusively from all over the Pakistan so far. Table 4 summarizes some available data on the topic from Pakistan along with the results of present study. Most of the available data are from Karachi and Lahore or Punjab and Sind while none of the study has explored the regions of FATA, Gilgit-Baltistan, AJK, some parts of Khyber Pakhtoonkha province and most of the Baluchistan province. Though three studies carried out by Awan et al. [18], Alam et al. [30] and Idrees et al. [34] investigated the provinces of KP and Baluchistan additionally but the number of samples taken were not satisfactory to conclude the genotype distribution in these provinces (Table 4). None of the studies included the regions of Gilgit-Baltistan, AJK and FATA. Two of these three [18, 34] studies reported genotype C as predominant while one [30] found genotype D as predominant one. Genotype B which was not reported from Pakistan in any other study has also been reported in all three of these studies from 18 to 25 % of the samples. These results are contradictory to these studies.Table 4 The prevalence of different HBV genotypes in our study compared to previously published studies Source Place of sampling Sample size Proportion of genotype as percent of total samples A B C D E F Mix Mixed combinations UT Hanif et al., 2013 [5] Karachi, Islamabad/Rawalpindi 200 (40, 160) 10 - - 59 - - 31 A + D - Awan et al., 2010 [18] Punjab, KP, Sind, Baluchistan 300 (222, 36, 26, 15) 14 18 28 13 0.6 1.3 16 A + B + D, A + D + F, A + C, A + D, A + E, A + F, B + C, B + E, C + D 10.3 Ahmed et al., 2009 [39] Punjab, Sind 236 0.8 - 5.9 93.2 - - - - - Baig et al., 2009 [32] Karachi 315 20 - - 70 - - 10 A + D - Noorali et al., 2008 [31] Karachi 180 - - - 84 - - 16 B + D - Hakim et al., 2008 [38] Karachi 180 - - - 84 - - 16 B + D - Alam et al., 2007 [30] Punjab, KP, Sind Baluchistan 110 (30, 28, 25, 18) 4.5 24.5 - 60 - - 2.7 A + D, B + D 8.2 Abbas et al., 2006 [37] Karachi 109 - - - 98.2 - - 1.8 A + D - Idrees et al., 2004 [34] Punjab, KP, Sind Baluchistan 112 (Details not available) 21.4 17.9 41.1 8.0 - - 7.1 Details not available 4.5 Naaz, 2001 [33] Lahore 12 - - 75 25 - - - - - This study Punjab, KP, Sind, Baluchistan, AJK, Gilgit-Baltistan 715 (311, 99, 84, 60, 68, 93) 7.7 0.6 0.8 71.2 0.1 - 17.3 A + D, B + D, C + D, A + B, A + B + D, E + D 2.2 UT stands for Un-typable All the studies confined to the samples from Punjab and Sind reported genotype D as predominant except one [33] which reported prevalence of genotype C and D as 75 and 25 % respectively. The remaining six studies [5, 31, 32, 37–39] reported genotype D as the commonest one in Pakistan with percentage distribution of 59, 84, 70, 98.2, 84 and 93.2 % respectively (Table 4). The results of present study representing all parts of the country with significant number of samples is in agreement with the later group [5, 31, 32, 37–39] that reported genotype D as the most prevalent HBV genotype in Pakistan. Genotype A was found to be the second more prevalent single genotype infection with 7.7 %. Genotype A was also reported in most of the previous studies [5, 18, 30, 32, 34, 39] in different proportions ranging from 0.8 to 21.4 % and a mean prevalence of 11.7 % (Table 4). Genotype B was present only in 0.6 % of present study samples while it was reported in proportions ranging from 18 to 24.5 % by previous studies from Pakistan [18, 30, 34]. None of the other studies reported genotype B infection, however mix infections involving genotype B were reported by some studies [18, 30, 31, 38] as high as 16 % of the samples (Table 4). Genotype C was detected in only 0.8 % of the samples in our study while three previous studies [18, 33, 34] have reported it as predominant genotype in Pakistan, whereas one study [39] reported it as second more prevalent genotype in the country. Genotype C was not reported by any of the other six studies (Table 4). Genotype E and F were not reported by any previous study except [18] which reported them in the proportions of 0.6 and 1.3 %. As already mentioned, genotypes E and F are endemic to Africa and the New World respectively [17, 20] and the fact that they were found in only 1 case (genotype E) was quite expected. Present study thus supports the conclusion that not only genotypes E and F but also genotypes B and C were rare in Pakistan (Table 4). The relationship of HBV genotypes with liver disease is still not very clear. Although some studies have reported that HBV genotype influence the severity of liver disease and course of chronic HBV infection [26, 28, 40–44]. Some other studies showed that HBV genotype had no influence on course of HBV infection [43, 45, 46]. Present study is in concordance with the later group [43, 45, 46]. Most of the studies compared genotype B and C in Asia and reported that the latter is more associated to severe liver disease than the former. On the other hand the genotype B is associated with an early HBeAg sero-conversion, more sustained remission after HBeAg sero-conversion, less active hepatic necroinflammation, a slower rate of progression to cirrhosis, and a lower rate of hepatocellular carcinoma (HCC) development as compared to genotype C [26, 28, 41, 43, 44]. A study from Pakistan compared genotype A, D and A + D for their association with complex liver diseases and concluded that genotype A is associated with more complex liver disease [47]. However, present study did not find any association of genotypes with disease stage. The relatively high prevalence of mix infection with genotypes A + D is a confirmation of the data reported by several previous findings about HBV genotyping in Pakistan [5, 18, 30, 32, 34, 37]. Conclusions In conclusion, it has been observed in the present study that the distribution of HBV genotypes was almost similar in all the regions of Pakistan. Genotype D was the most common one (71.2 %) followed by mix infection with genotypes A and D (17.3 %) and genotype A (7.7 %). Although genotypes B, C, E and F were also detected but their prevalence was very low (1.5 %). Furthermore, genotype was not found to have any significant influence on liver disease progression in Pakistan. Abbreviations AJKAzad Jammu and Kashmir ANOVAAnalysis of Variance bpBase Pairs ddH2ODouble Distilled Water DNADeoxy ribo nucleic acid FATAFederally Administered Tribal Areas FCTFederal Capital Territory HBsAgHepatitis B Surface Antigen HBVHepatitis B Virus HCCHepatocellular Carcinoma HECHigher Education Commission IU/mlInternational Unit Per Milli Liter KPKhyber Pakhtoonkha NORINuclear Medicine, Oncology and Radiotherapy Institute PCRPolymerase Chain Reaction PMASPir Mehr Ali Shah SPSSStatistical Package for Social Sciences USAUnited States of America °CCentigrade Acknowledgements The authors are thankful to Professor Ferruccio Bonino and Professor Mirza Azhar Beg for critically reviewing the article and setting its language. Funding The study was partially funded by Higher Education Commission of Pakistan (HEC) in the form of indigenous PhD fellowship (PIN: 074-2673-Bm4-145). There was no role of the funding body in design of the study, data collection, analysis and publication. Availability of data and materials Data sharing not applicable to this article as no datasets were generated or analyzed during the current study. Original data may be obtained by email to corresponding author. Authors’ contributions MM, AR and AK conceived and designed the study and experiments. MM, MAA and NZ performed the experiments and collected the data. MM and NZ analyzed the data. MM prepared the initial draft of manuscript. AK, AR, and MAA critically reviewed and revised the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. The study is a part of PhD thesis of Majid Mahmood. Consent for publication Not applicable. Ethics approval and consent to participate The study was approved by institutional ethics committee of PMAS Arid Agriculture University Rawalpindi and each patient gave written consent before the start of study. ==== Refs References 1. Bonino F. Introduction to hepatitis B virus heterogeneity – a means to personalized care. Antivir Ther. 2010;15 Suppl 3:1–2. 2. Boesecke C, Wasmuth JC. Hepatitis B. In: Mauss S, Berg T, Rockstroh J, Sarrazin C, Wedemeyer H, editors. Hepatology a clinical text book. 3rd ed. Wuppertal: Flying Publishers; 2012. p. 32–35. 3. Rantala M, Van de Laar MJ. Surveillance and epidemiology of hepatitis B and C in Europe – a review. Eurosurveillance. 2008;13(21):1-8. 4. Wasley A Grytdal S Gallagher K Surveillance for acute viral hepatitis--United States MMWR Surveill Summ 2008 57 1 24 18354374 5. Hanif M Zaidi P Habib S Ahmed A Raza A Study of genotypes and subgenotypes of Hepatitis B virus prevalent in big cities of Pakistan Afr J Microbiol Res 2013 3 152 157 6. Okamoto H Tsuda F Sakugawa H Sastrosoewignjo RI Imai M Typing hepatitis B virus by homology in nucleotide sequence: comparison of surface antigen subtypes J Gen Virol 1988 69 2575 2583 10.1099/0022-1317-69-10-2575 3171552 7. Naumann H Schaefer S Yoshida CFT Gaspar AMC Repp R Identification of a new hepatitis-B-virus (HBV) genotype from Brazil that expresses HBV surface antigen subtype adw4 J Gen Virol 1993 74 1627 1632 10.1099/0022-1317-74-8-1627 8345355 8. Norder H Courouce AM Magnius LO Complete genomes, phylogenetic relatedness, and structural proteins of six strains of the hepatitis B virus, four of which represent two new genotypes Virol 1994 198 489 503 10.1006/viro.1994.1060 9. Stuyver L De Gendt S Van Geyt C Zoulim F Fried M A new genotype of hepatitis B virus: complete genome and phylogenetic relatedness J Gen Virol 2000 81 67 74 10.1099/0022-1317-81-1-67 10640543 10. Arauz-Ruiz P Norder H Robertson BH Magnius LO Genotype H: a new Amerindian genotype of hepatitis B virus revealed in Central America J Gen Virol 2002 83 2059 2073 10.1099/0022-1317-83-8-2059 12124470 11. Kay A Zoulim F Hepatitis B virus genetic variability and evolution Virus Res 2007 127 164 176 10.1016/j.virusres.2007.02.021 17383765 12. Tran TT Trinh TN Abe K New complex recombinant genotype of hepatitis B virus identified in Vietnam J Virol 2008 82 5657 5663 10.1128/JVI.02010-07 18353958 13. Olinger CM Jutavijittum P Hubschen JM Yousukh A Samountry B Possible new hepatitis B virus genotype, southeast Asia Emerg Infect Dis 2008 14 1777 1780 10.3201/eid1411.080437 18976569 14. Tatematsu K Tanaka Y Kurbanov F Sugauchi F Mano S A genetic variant of hepatitis B virus divergent from known human and ape genotypes isolated from a Japanese patient and provisionally assigned to new genotype J J Virol 1995 83 10538 10547 10.1128/JVI.00462-09 19640977 15. Kidd-Ljunggren K Oberg M Kidd AH The hepatitis B virus X gene: analysis of functional domain variation and gene phylogeny using multiple sequences J Gen Virol 1995 76 2119 2130 10.1099/0022-1317-76-9-2119 7561749 16. Bowyer SM Staden LV Kew MC Sim JG A unique segment of the hepatitis B virus group A genotype identified in isolates from South Africa J Gen Virol 1997 78 1719 1729 10.1099/0022-1317-78-7-1719 9225049 17. Bonino F, Teerha P, Brunetto MR, Yun-Fan L. Diagnostic markers of chronic hepatitis B infection and disease. Antivir Ther. 2010;15 Suppl 3:35–44. 18. Awan Z Idrees M Amin I Butt S Afzal S Pattern and molecular epidemiology of Hepatitis B virus genotypes circulating in Pakistan Infect Genet Evol 2010 8 1242 1246 10.1016/j.meegid.2010.08.006 20727423 19. Datta S An overview of molecular epidemiology of hepatitis B virus (HBV) in India Virol J 2008 5 156 10.1186/1743-422X-5-156 19099581 20. Nakano T Lu L Hu X Mizokami M Orito E Characterization of hepatitis B virus genotypes among Yucpa Indians in Venezuela J Gen Virol 2001 82 359 365 10.1099/0022-1317-82-2-359 11161274 21. Cao GW Clinical relevance and public health significance of hepatitis B virus genomic variations World J Gastroenterol 2009 46 5761 5769 10.3748/wjg.15.5761 19998495 22. Toan NL le Song H Kremsner PG Duy DN Binh VQ Impact of the hepatitis B virus genotype and genotype mixtures on the course of liver disease in Vietnam Hepatology 2006 43 1375 1384 10.1002/hep.21188 16729315 23. Yin J Zhang H Li C Gao C He Y Role of hepatitis B virus genotype mixture, subgenotypes C2 and B2 on hepatocellular carcinoma: compared with chronic hepatitis B and asymptomatic carrier state in the same area Carcinogenesis 2008 29 1685 1691 10.1093/carcin/bgm301 18192693 24. Erhardt A Blondin D Hauck K Sagir A Kohnle T Response to interferon alfa is hepatitis B virus genotype dependent: genotype A is more sensitive to interferon than genotype D Gut 2005 54 1009 1013 10.1136/gut.2004.060327 15951551 25. Wai CT Chu CJ Hussain M Lok AS HBV genotype B is associated with better response to interferon therapy in HBeAg(+) chronic hepatitis than genotype C Hepatology 2002 36 1425 1430 12447868 26. Chan HL Hui AY Wong ML Tse AM Hung LC Genotype C hepatitis B virus infection is associated with an increased risk of hepatocellular carcinoma Gut 2004 53 1494 1498 10.1136/gut.2003.033324 15361502 27. Chan HL Wong GL Tse CH Chim AM Yiu KK Hepatitis B Virus Genotype C Is Associated With More Severe Liver Fibrosis Than Genotype B Clin Gastroenterol Hepatol 2009 7 1361 1366 10.1016/j.cgh.2009.08.004 19683072 28. Chu CJ Hussain M Lok AS Hepatitis B virus genotype B is associated with earlier HBeAg seroconversion compared with hepatitis B virus genotype C Gastroenterology 2002 122 1756 1762 10.1053/gast.2002.33588 12055581 29. Livingston SE Simonetti JP Bulkow LR Homan CE Snowball MM Clearance of hepatitis B e antigen in patients with chronic hepatitis B and genotypes A, B, C, D, and F Gastroenterology 2007 133 1452 1457 10.1053/j.gastro.2007.08.010 17920063 30. Alam MM Zaidi SZ Malik SA Shaukat S Naeem A Molecular epidemiology of Hepatitis B virus genotypes in Pakistan BMC Infect Dis 2007 7 115 10.1186/1471-2334-7-115 17922910 31. Noorali S Hakim ST McLean D Kazmi SU Bagasra O Prevalence of Hepatitis B virus genotype D in females in Karachi, Pakistan J Infect Dev Ctries 2008 2 373 378 19745506 32. Baig S Siddiqui A Chakravarty R Moatter T Hepatitis B virus subgenotypes D1 and D3 are prevalent in Pakistan BMC Res Notes 2009 2 1 10.1186/1756-0500-2-1 19121226 33. Naaz A. Variability analysis in immunogenic region of surface gene of Pakistani isolates of Hepatitis B Virus. PhD thesis. University of Punjab 2001, pp. 68–113 34. Idrees M Khan S Riazuddin S Common genotypes of hepatitis B virus J Coll Physicians Surg Pak 2004 14 344 347 15233887 35. Ali M Idrees M Ali L Hussain A Rehman IU Hepatitis B virus in Pakistan: A systematic review of prevalence, risk factors, awareness status and genotypes Virol J 2011 8 102 10.1186/1743-422X-8-102 21375760 36. Naito H Hayashi S Abe K Rapid and Specific Genotyping System for Hepatitis B Virus Corresponding to Six Major Genotypes by PCR Using Type-Specific Primers J Clin Microbiol 2001 39 362 364 10.1128/JCM.39.1.362-364.2001 11136801 37. Abbas Z Muzaffar R Siddiqui A Naqvi SAA Rizvi SAH Genetic variability in the precore and core promoter regions of hepatitis B virus strains in Karachi BMC Gastroenterol 2006 6 20 10.1186/1471-230X-6-20 16863587 38. Hakim ST Kazmi SU Bagasra O Seroprevalence of Hepatitis B and C Genotypes Among Young Apparently Healthy Females of Karachi-Pakistan Lib J Med 2008 3 66 70 10.4176/071123 39. Ahmed CS Wang ZH Bin Z Chen JJ Hou JL Hepatitis B virus genotypes, subgenotypes, precore, and basal core promoter mutations in the two largest provinces of Pakistan J Gastroenterol Hepatol 2009 24 569 573 10.1111/j.1440-1746.2008.05642.x 19368634 40. Lindh M Hannoun C Dhillon AP Norkrans G Horal P Core promoter mutations and genotypes in relation to viral replication and liver damage in East Asian hepatitis B virus carriers J Infect Dis 1999 179 775 782 10.1086/314688 10068571 41. Kao JH Chen PJ Lai MY Chen DS Hepatitis B genotypes correlate with clinical outcomes in patients with chronic hepatitis B Gastroenterology 2000 118 554 559 10.1016/S0016-5085(00)70261-7 10702206 42. Orito E Mizokami M Sakugawa H Michitaka K Ishikawa K A case-control study for clinical and molecular biological differences between hepatitis B viruses of genotypes B and C Hepatology 2001 33 218 223 10.1053/jhep.2001.20532 11124839 43. Sumi H Yokosuka O Seki N Arai M Imazeki F Influence of hepatitis B virus genotypes on the progression of chronic type B liver disease Hepatology 2003 37 19 26 10.1053/jhep.2003.50036 12500184 44. Chu CM Liaw YF Genotype C hepatitis B virus infection is associated with a higher risk of reactivation of hepatitis B and progression to cirrhosis than genotype B: a longitudinal study of hepatitis B e antigen-positive patients with normal amino-transferase levels at baseline J Hepatol 2005 3 411 417 10.1016/j.jhep.2005.03.018 16006001 45. Gandhe SS Chadha MS Arankalle VA Hepatitis B virus genotypes and serotypes in western India: lack of clinical significance J Med Virol 2003 69 324 330 10.1002/jmv.10292 12526041 46. Zeng G Wang Z Wen S Jiang J Wang L Geographic distribution, virologic and clinical characteristics of hepatitis B virus genotypes in China J Vir Hepat 2005 12 609 17 10.1111/j.1365-2893.2005.00657.x 47. Baig S Siddiqui A Ahmed W Qureshi H Arif A The association of complex liver disorders with HBV genotypes prevalent in Pakistan Virol J 2007 4 128 10.1186/1743-422X-4-128 18042293

PMC005xxxxxx/PMC5002162.txt

==== Front BMC Med EducBMC Med EducBMC Medical Education1472-6920BioMed Central London 74110.1186/s12909-016-0741-yResearch ArticleBuilding blocks for social accountability: a conceptual framework to guide medical schools http://orcid.org/0000-0003-4700-1521Preston Robyn +61 7 4781 3182robyn.preston@jcu.edu.au 13Larkins Sarah Sarah.larkins@jcu.edu.au 134Taylor Judy Judy.taylor@jcu.edu.au 134Judd Jenni Jenni.judd@jcu.edu.au 12341 College of Medicine and Dentistry, Division of Tropical Health and Medicine, James Cook University, 1 James Cook Drive, Townsville, QLD 4811 Australia 2 Division of Tropical Health and Medicine, James Cook University, Townsville, Australia 3 Anton Breinl Research Centre for Health Systems Strengthening, James Cook University, Townsville, Australia 4 Australian Institute of Tropical Health and Medicine, James Cook University, Townsville, Australia 26 8 2016 26 8 2016 2016 16 1 2272 2 2016 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 This paper presents a conceptual framework developed from empirical evidence, to guide medical schools aspiring towards greater social accountability. Methods Using a multiple case study approach, seventy-five staff, students, health sector representatives and community members, associated with four medical schools, participated in semi-structured interviews. Two schools were in Australia and two were in the Philippines. These schools were selected because they were aspiring to be socially accountable. Data was collected through on-site visits, field notes and a documentary review. Abductive analysis involved both deductive and inductive iterative theming of the data both within and across cases. Results The conceptual framework for socially accountable medical education was built from analyzing the internal and external factors influencing the selected medical schools. These factors became the building blocks that might be necessary to assist movement to social accountability. The strongest factor was the demands of the local workforce situation leading to innovative educational programs established with or without government support. The values and professional experiences of leaders, staff and health sector representatives, influenced whether the organizational culture of a school was conducive to social accountability. The wider institutional environment and policies of their universities affected this culture and the resourcing of programs. Membership of a coalition of socially accountable medical schools created a community of learning and legitimized local practice. Communities may not have recognized their own importance but they were fundamental for socially accountable practices. The bedrock of social accountability, that is, the foundation for all building blocks, is shared values and aspirations congruent with social accountability. These values and aspirations are both a philosophical understanding for innovation and a practical application at the health systems and education levels. Conclusions While many of these building blocks are similar to those conceptualized in social accountability theory, this conceptual framework is informed by what happens in practice - empirical evidence rather than prescriptions. Consequently it is valuable in that it puts some theoretical thinking around everyday practice in specific contexts; addressing a gap in the medical education literature. The building blocks framework includes guidelines for social accountable practice that can be applied at policy, school and individual levels. Electronic supplementary material The online version of this article (doi:10.1186/s12909-016-0741-y) contains supplementary material, which is available to authorized users. Keywords Social accountabilityMedical schoolsMedical educationConceptual frameworkAustraliaThe Philippineshttp://dx.doi.org/10.13039/501100000937Department of Education, Employment and Workplace Relations, Australian GovernmentAustralian Postgraduate AwardPreston Robyn http://dx.doi.org/10.13039/501100001792James Cook UniversityCollege of Medicine and DentistryGraduate Research SchoolPreston Robyn issue-copyright-statement© The Author(s) 2016 ==== Body Background Medical schools aspiring towards social accountability deem themselves responsible for meeting the health needs of the communities that they serve1 [1]. They orientate their education, research and service to the health needs of the population [2]. From the 1980s to 2000s Dr Charles Boelen and other colleagues at the World Health Organization (WHO) fostered the theory of socially accountable medical education and socially accountable medical schools [3]. These authors then developed the conceptualization, production and utilization grid, the values of social accountability (relevance, equity, cost-effectiveness and quality) and the Towards Unity for Health partnership pentagram [1–10]. From the 2000s to 2010s there was re-interest in measuring and evaluating socially accountable medical schools; including a number of initiatives at individual schools and collectively. These encompassed the Lancet’s Independent Commission on Health Professional Education for the 21st Century [11] and the call for socially accountable medical education to be aligned with accreditation systems. Despite the advancement of conceptual frameworks, many medical schools aspiring to social accountability had developed independently from the theoretical concepts of socially accountable medical education. There are gaps in the literature around the key factors, both externally and internally, influencing medical schools advancing towards social accountability and the way in which they operate. While the theory of social accountability identifies key principles, how these play out in different schools is the topic of this paper. We present a conceptual framework, derived from empirical research from practice, describing the building blocks for socially accountable medical education. Methods A social justice perspective, drawing upon constructivism and critical theory, was utilized throughout this research. This perspective is linked with the phenomena of social accountability and the epistemological position of the researchers [12, 13]. The voices of all participants had to be heard and given equal weight in contributing to new knowledge. This was important as this research included cases from more financially and less financially resourced countries. Recent literature in medical education has criticized global medical school partnerships as post-colonialist or neo-imperialist enterprises, with imbalances in power relationships and benefits [14, 15]. Furthermore, the powerful profession of ‘the medical doctor’ was investigated. Issues of power and professionalism are played out between and within health professionals [16]. One way in which socially accountable medical education has been defined and implemented is in the context of medical dominance and the changing role of ‘the doctor’ as a professional and ‘leader’ in the health world [16–21]. The authors attempted to actively acknowledge and confront that research can reproduce or enforce existing power structures. Researchers aimed to be self-reflective and self-critical of the values and interest that underlie social theory [22] and explicitly declare their own assumptions [23]. This project utilized multiple case studies [13, 24–26] of four medical schools in Australia and in the Philippines.2 Cases were sampled on the basis of maximizing the likelihood of demonstrating a range of contextual influences [13] and membership of the Training for Health Equity Network (THEnet), a collaboration formed in 2008 of 11 health professional schools aspiring towards social accountability3 [27–32]. Seventy-five purposively sampled faculty, students, health professionals and community members were interviewed (Table 1) on their understanding of social accountability and what contextual factors had influenced their schools. From a review of the literature these were hypothesized to be:Table 1 Individual Participants [3] Type Description Participants Case 1 FUSOM Case 2 JCUSMD Case 3 ADZUSOM Case 4 UPMSHS Total Staff/Faculty “Champions at schools.” Any staff members who know about social accountability and/ or the history of the school, including leaders and former leaders. 11 11 7 (including 2 former students) 9 (including 2 former students) 38 Students Student at any year level who are interested in social accountability. In Australia, students who were involved in the rural and/or international health student groups. 4 4 5 4 17 Health Sector A person holding a position involved in policy, medical education at a state, regional or provincial level, who has had involvement in the school. A preceptor or someone from a hospital or the health department, who has been involved with the school. At two schools (UPMSHS and ADZU) this included Graduates of the school who may also have a teaching or preceptor role. 2 (including 1 preceptor) 2 (including 1 joint JCUSOM appointment with Department of Health) 6 (all had teaching roles and all were former students) 4 (including 3 former students; all involved in teaching students) 14 Community “Champions at schools.” Any community person involved in the school in teaching or research or at rural placement sites. 1 1 2 2 6 Total 18 18 20 19 75 Profile of the local health workforce; Partnerships with the local, state and national health system; and Partnerships between the medical school and its ‘community’. (Additional file 1 outlines examples of the interview guides). Site visits were undertaken by the principal researcher (RP) to all four schools, including a rural area at each location. Documents analysed included peer reviewed papers, meeting reports, websites and policies [33]. Data analysis was abductive and involved deductive and inductive iterative thematic analysis of qualitative data [34], both within case and across cases [25, 35]. All interviews were transcribed verbatim and any changes identified by participants during verification were incorporated into the transcripts. Field notes, and notes from documentary analysis were transcribed and NVivo 10 was used to store, organize and analyze data [36]. Within case analysis was undertaken by abductive coding; the combination of a predefined deductive coding framework (set by the questions and the three contextual issues obtained from the literature review) and open coding that recorded emerging ideas [3, 37]. Themes were developed and finalized through the constant comparison method, drawing out amalgamated ideas from the codes. Initial similar themes across cases were developed from data immersion. Themes were refined and connections between themes mapped using site-order descriptive matrixes for each of these questions [38]. Coder triangulation was undertaken and negative instances that challenged initial assumptions and interesting outliers were identified to ensure that conclusions were drawn from the data [39]. Themes that emerged consistently across all data were categorized as core themes. Concepts were then further developed in relation to the research questions and from the literature. There were consistent themes across all cases, regardless of the country or other factors. This paper reports on the cross-case analysis. Results Local and national workforce issues and health needs were important external contextual factors that influenced all four socially accountable medical schools:So workforce was the reason the government funded it [the medical school] so absolutely if you move back from government funding [it] was workforce, workforce, workforce… (FU1_staff). Workforce needs were defined as shortages of doctors in underserved areas and the mal-distribution between urban and rural areas. All institutions had geographic proximity to a defined region of workforce need, or sent their students there for a lengthy period, often coexisting with populations with the poorest health indicators or greatest health needs. Programs and schools were established to meet these workforce and health care needs and there were some innovations in both the Philippines and in Australia. The workforce and health system needs influenced how students were trained and the desired outcomes of their training. Schools produced graduates who would be well-equipped to meet the workforce and health system needs. Placements ensured that graduates were personally and professionally equipped to work in areas of workforce need. An important contextual influence came from the communities the medical schools served. These differed:At the population level: Associated with the influence of workforce, at some schools the population of communities and their different health needs had influenced the establishment and orientation of the schools; As communities of place and interest with a formal or informal community engagement policy: Some schools had formal community engagement policies and informal ways to work with communities of place and interest; sometimes informed by the wider university policy; and, As students becoming part of the communities: Students undertake placements in underserved communities and are integrated into community life:“That’s why you are sent to communities that do not have water, that do not have electricity, to see that not everyone is the same. So that opens our eyes. This is the social situation of our country, not everyone is privileged, not everyone can go to school, and not everyone has the chance to eat three square meals a day. This is our situation in our region and you’re here, you’re privileged to study medicine, it’s for you to help these people” (ADZU2_student). Government policy was an influence in the Australian case studies because workforce need was a political driver. A number of Australian government policies and programs were designed to address rural workforce issues such as rural placements [40–43]:In fact you could say that the Commonwealth [government] now has [become], to some extent, our main champion because without them we would not [have] the money and the policy wouldn’t have developed the way it has and so on (FU8_Health). In the Philippines, due to the prevailing culture of export driven medical and health professional education, government policy and programs were not highlighted as influencing socially accountable medical schools. Local politics and local politicians were a powerful contextual issue that influenced social accountability at all four medical schools. While the local government sector had more influence in the Philippines, in both countries participants saw the local or first level of government as being part of the community. In the Philippines the health system is devolved with local officials and local politics influencing the budget and this could either impede or enhance student selection, placements, and graduate job placements:We need to have a strong partnership or bond with the Local Government Unit (LGU) because it is the LGU that identifies which health workforce they need, so we train the students to become the health workers that they need. (UPMSHS16_staff). In Australia, while the local government did not have as much influence over the health system, support was provided through other initiatives that have involved local councils and medical education such as the provision of housing and social support on placements and in some cases, local councils owning primary health care services. The values and missions of affiliated universities, when they were aligned with social accountability, were an important contextual influence. Many of the staff’s own beliefs, including social justice and a religious association, were consistent with overall values. In addition, some schools were part of universities that had a wider institutional culture that encouraged going against the norm to develop innovative programs. For example, difference and freedom from tradition gave FUSOM an opportunity to focus on “innovation” or being able to have “… a different approach to the way they did business [so there] was a fair degree of risk taking on their behalf, innovation, opportunity, that perhaps came with being a new kid on the block” (FU8_health). Financial, human and infrastructural resources were also a contextual influence in two ways. In some contexts lack of resources has spurred innovation [30, 32]. While both the Filipino schools had received international support at establishment and had continued partnerships with schools in more resourced countries; a lack of resources also promoted self-reliance and a culture of independence and empowerment. There was also an awareness that locals were better equipped to develop the schools; and a creative ability to “… make do with what we have” (UPMSHS5_staff). In Australia, government funding was essential for socially accountable programs, such as rural placements, to be established and sustained. Indeed these programs were accurately deemed more expensive than conventional medical education programs due to decentralization to rural areas. Leaders and individual “champions” of social accountability influenced the organizational culture of the schools and the type of people attracted to working in the schools and drove innovation. Staff and students were drawn to working with leaders and at institutions that had these values. These leaders’ values, medical experience in the region and personal and professional connections established, re-orientated or sustained the school. Champions from across the health, education and political sectors lobbied for the establishment of new schools or programs:…they’re not just people who look at the problem in front of their nose, they take a step back and think. That sort of controlled rage really, is a drive to change things. Because it takes a lot of effort to change the way things are done, so when I say controlled rage I suppose you know the fire in the belly (JCU12_staff). All schools were inaugural members of the Training for Health Equity Network (THEnet) and therefore had articulated a common understanding of social accountability [28–32]. In general, THEnet legitimized what schools were doing; membership and interaction with this group clarified existing practices and connected the schools with a global movement. All schools had an understanding of resisting traditional models of medical education. Belonging to this group also helped members interpret the high level WHO policies or directives into practical guidelines confirming existing practice. Through THEnet leaders and staff were able to represent their schools on the WHO and other global forums and groups interested in social accountability. Leaders also developed alliances in medical education that assisted with their own professional development. All schools had an understanding of resisting traditional models of medical education. An underlying assumption that was not fully explored by the participants was the idea of being different and defending this difference. This was articulated at JCUSOM as the health and community needs of the geographical north being different. At FUSOM participants discussed the difference and innovation of the university and school. Respondents knew that ADZU SOM had always been against the mainstream medical education. At UPMSHS participants noted they were doing what they could with little resources. The schools naturally challenge orthodoxies, and are seen as alternative models for health professional education [40] but have also faced “institutional isolation and skepticism from more traditional medical schools” [28] (p.340). There was a shared understanding that while the schools may be critiqued for persisting against opposition, subverting dominant paradigms or doing that which was thought impossible by more traditional schools, there was also a defiant confidence that resistance to current models of medical education is ‘the right thing to do’ [27, 44]. Discussion Conceptual framework: building blocks for socially accountable medical schools The building blocks conceptual framework (Fig. 1) was created by analyzing the internal and external factors influencing the selected medical schools and demonstrates what ‘building blocks’ or conditions might be necessary to assist medical schools moving towards social accountability. These building blocks are classified as environmental (macro), school (meso) or people (micro) factors. The ‘bedrock’ of social accountability applies at all levels: shared values congruent with social accountability. While many of these building blocks are similar to those conceptualized in social accountability theory, this conceptual framework is informed by what happens in practice - empirical evidence rather than prescriptions. Consequently it is valuable in that it puts some conceptual thinking around everyday practice. These prescriptions are based on data from the four cases and cannot hope to cover all the nuances of practice in widely varying contexts.Fig. 1 Building Blocks for Socially Accountable Medical Schools Environmental factors Environmental factors or macro factors are those outside the medical school. They include the local, regional or national workforce and health needs as well international, national and local government policies and resourcing. Responding to workforce and health needs Workforce was a major influence of schools aspiring to be socially accountable as the main focus of socially accountable medical schools should be meeting workforce needs and understanding the health issues of their region. This finding concurs with the literature on socially accountable medical schools [1, 29, 45]. Schools in areas of workforce need have a more holistic or “whole of school” approach to social accountability as the health system needs are very apparent and hard to ignore. What this means in practice To understand workforce issues the curriculum of the school should include an analysis of the local workforce situation as well as key health issues. Leaders and staff need to have professional experiences and connections in areas of health and workforce need. Responsive and appropriate international, national and local government policy with funding and resourcing The second building block for socially accountable medical education is the partnership that schools form with government, the health sector, and health care policy makers in health professional education. Governments play different roles depending on the country but the governmental and political context is always important. In Canada, socially accountable medical schools have been mandated in Health Canada policy since 2001 [45, 46]. In more financially resourced countries, such as Canada or Australia, socially accountable programs could not persist without government funding. In contrast, government policies that have restricted social accountability in the Philippines were overcome due to other factors such as strong leadership and people within innovative schools. The differences in political organization, or between a more centralized and decentralized government, might explain some of the differences in Australia and the Philippines. The decentralized Filipino health system may mean that local initiatives are more influential than government policy. What this means in practice Medical schools and their leaders need to lobby for medical education and health policies that support the principles of social accountability. For example, for government policies and funding programs that support medical schools in regional and rural settings and in areas where there is workforce need. Medical schools can also adapt and promote government workforce policy or aspects that are socially accountable (for example rural placements). The need for these policies should be promoted and explained in terms of graduate outcomes and long-term impact on health needs of communities. Medical schools should also advocate for a balance of adequate resources with room for innovation and adaptation to the local contexts. Accreditation can motivate medical schools towards social accountability. Medical schools should call for national accreditation systems to include assessment of social accountability. Indeed progress has started on moving the World Federation for Medical Education accreditation standards in this direction [47]. School factors School or meso factors include those factors within medical schools including leaders, people (staff/faculty/students) and policies. An explicit policy of community participation and informal linkages with a defined community Community placements, and connections between underserved or rural communities and their health needs, enable students to experience the realities of professional and community life [44, 48–50]. The types of connections between schools and communities vary again depending on the context. The idea of service and links with the community were stronger in the Philippines where there was a sense of personal responsibility. Participants in the Philippines saw the students on placement as ‘helping’ with health and community services. In Australia however the imperative for workforce is more political or organizational and government resources are assigned to this endeavor through programs such as the Rural Clinical Teaching Program [51]. While all schools were influenced by communities, the community may not have appreciated or fully understood the importance of this influence. In addition, the connections may be limited and there might be a lack of authentic engagement and/or partnership. In a critical review Hunt et al. noted that university-community engagement was mainly frequently understood as outreach or service and no articles discussed mutually beneficial or collaborative partnerships with communities [52]. Student placements are one of the major ways through which links are made between rural or underserved populations and the medical school [53]. Community placements were also responsible for changing or confirming student attitudes and aspirations. Long term community-based medical education that develop significant relationships enhances student learning [54]. What this means in practice In order to work effectively with their community a school needs to have a well-defined understanding of the communities, or the defined population that they serve, and an explicit policy about the types of community participation that are appropriate. There will be a need for engaging, partnering or connecting with the community to be undertaken in diverse ways, for example for community to be involved in strategic planning. The community participation strategy of the school needs to be well resourced, achievable and assessed or evaluated. Generally, schools should partner with non-government or community-based or civil society organizations that already work in communities of need to provide teaching, research, service or health projects. To enable successful partnerships, communities need to be made more aware of their power and influence and community organizations can help facilitate this process. Leadership and champions with shared values and political influence and willingness to challenge dominant paradigms Strong leadership and champions can transform medical education [30, 32, 55] and this is a building block of social accountability. In a study of Deans as “spiritual leaders”, Evans [56] states that leaders should “personify and embody” the values of medicine as a profession and a vocation and “they must remind us of those values and inspire us to embrace them and be guided by them” (p. 655). In promoting social accountability leaders must have not only lived experience as a health professional but “have an iron will to succeed [as] personal experience and integrity are important in the difficult debates that will ensue” [28] (p.10). Institutional or policy barriers to change, including curriculum reform, at medical schools are likely [57–60]. Strong leaders will continue to challenge the dominant paradigms of medical education when they are not supportive of social accountability [61]. What this means in practice In practice, leadership means three things: inspiring socially accountable values; institutionalizing them; and creating a shared understanding of social accountability at the individual, organizational, and community level. The values, medical experience in the region and personal and professional connections of leaders help establish, reorientate or sustain schools. Part of the role of the leader is to institutionalize values and support and to model the culture of social accountability. This is a sustainability step enabling these Schools to embed these values within their culture so that they are not lost when a leader moves on. There needs to be explicit acknowledgment of the type of leadership schools aspiring to be social accountable desire to institutionalize the qualities of leaders. Mentoring aspiring leaders can also ensure socially accountable leadership. Membership of a coalition of institutions aspiring to be socially accountable The role of a network or coalition as a building block to support social accountability is important. The Training for Health Equity Network is a community of practice [40, 62, 63]:“Communities of practice are groups of people who share a concern, a set of problems, or a passion about a topic, and who deepen their knowledge and expertise in this area by interacting on an ongoing basis” [63] (p. 4). This community of practice collaborates to solve problems, shares knowledge, creates tools, and has a body of common knowledge, personal relationships, and sense of identity [63]. The influence of THEnet, and other groups of medical schools is that it “can drive change from within” [40] (p. 4) The schools were brought together due to their common goals and challenges and work collaboratively. What this means in practice Medical schools should join or create a coalition of institutions aspiring to be socially accountable. Coalitions can help foster the professional development of leaders into international medical education leadership roles. Connection with a coalition helps legitimize and develop social accountability, particularly when schools have alternative models of medical education. Other staff and students can be part of a community of practice and receive peer support from colleagues at other schools. A community of practice can help translate or interpret global policies and directives to the local level. These groups can provide opportunities for individual schools to highlight their work. These coalitions should continue to engage schools in the wider social accountability movements such as the WHO and Global Consensus on Social Accountability (GCSA). People factors People or micro factors are the values of people who make up medical schools and their wider institution including staff, students, leaders, health professionals and students. This building block is fully explored in Preston et al. [3]. An explicit understanding of social accountability As explored by Preston et al. [3] the difference in perceptions of social accountability indicates that the term may not be universally understood and appreciated [64–70]. Furthermore, the term has danger of losing meaning if schools do not continue to critically appraise and debate the term and what it means in their own unique context. What this means in practice All stakeholders need to be involved in debating and developing and operationalizing a shared understanding of social accountability to ensure it is socio-culturally appropriate to the context of the school [3]. The school community may argue that the term “social accountability” does not fully capture their mission and practice; particularly in non-English speaking settings. A local term or phase may be developed that captures the nuances and essence of the school culture. Staff students, health sector and community members could be asked “if our school was socially accountable what would it look like?” [3] In addition, as discussed by Ritz et al. schools should adopt a “critically reflexive social accountability” [69], (p. 155) that questions underlying assumptions and discourses. Dissenting views and debate should be encouraged as part of this process [3]. The personal values and beliefs of staff, leaders and student influence the organizational culture of the school. People and their values can help drive organizational change. Schools need to foster a values based approach to social accountability. This may involve personalizing social accountability so it means something to all staff, students and others involved in the school, including the health and community sectors. Strategies to influence the organizational culture of the school involve including an understanding of “social accountability” as part of the staff recruitment process (for example in selection criteria or interview questions). Applicants could be asked to describe their understanding of social accountability and how they feel they could contribute to the aspirations of the school. Furthermore, new staff could be inducted into the values of the medical school, by undertaking an orientation on socially accountable practice. Faculty/staff need to be given opportunities to reflect and act on their own practices; to “walk the talk”. Faculty development could incorporate practical projects on how to engage with underserved communities. Staff in a biomedical or laboratory research role may feel alienated from social accountability. These staff could be given opportunities to learn how social accountability applies to their work. For example, they could develop professional activities that link them with the social accountability agenda of the school. These could include projects with high schools, communities, and research focused on the priority health needs of the communities. However, schools need to appreciate and accept that not all staff will or can have socially accountable practice at the forefront of their work. The bedrock: Shared values and aspirations congruent with social accountability The bedrock of social accountability, that is, the foundation for all building blocks, is shared values and aspirations congruent with social accountability. These values and aspirations are both a philosophical understanding for innovation and a practical application at the health systems and education levels. The values are held by people: the community; leaders; students; faculty and health professionals. These actors or stakeholders aspire to apply these values, often in defiance of ‘mainstream’ health and education systems, by “walking the talk” of social accountability in their everyday work and life. Conclusions The nature of development of social accountability is contextually dependent, politically, historically, socially, spiritually, and economically influenced. Some schools did not term their practice as such until legitimized in this movement for socially accountable medical education. Consequently, development of practice and theory should be seen as iterative rather than prescriptive. Through the process of critical analysis to identify the most important key influences of socially accountable medical education, the building blocks become clear. These are: environmental, including workforce; funding and partnerships; internal to the school, including leadership and community engagement, and about the values of people. There are also examples given in this paper of the actions that medical schools might take to move towards social accountability. The actions are very different depending on the context but if actions are thoughtfully applied and consistent with the building blocks then this might support further movement towards socially accountable medical education. Additional file Additional file 1: Examples of Interview guides. Examples of interview guides for faculty/staff, student, health professional and community representative interviews. (DOCX 18 kb) Abbreviations ADZU SOMAteneo de Zamboanga University School of Medicine FUSOMFlinders University School of Medicine GCSAGlobal Consensus for Social Accountability HRECHuman Research Ethics Committees JCUJames Cook University JCUSOMJames Cook University School of Medicine LGULocal Government Unit THEnetTraining for Health Equity Network UPMSHSUniversity of the Philippines Manila, School of Health Sciences WHOThe World Health Organization 1 “Aspiring towards social accountability” is used for schools with an explicit social accountability mandate as social accountability is recognized to be an aspirational ideal. 2 Flinders University, Australia; James Cook University, Australia; Ateneo de Zamboanga University, The Philippines; University of the Philippines, Manila, School of Health Sciences, The Philippines. 3 “A shared understanding of social accountability, in the context of health professional education, was a key inclusion criterion for the cases. If schools and their staff/faculty did not have a shared understanding of social accountability the design flaw would have been “but what is social accountability?” The term could have meant anything and been interpreted in divergent ways. By being members of THEnet schools, the institution had declared that they were aspiring to be socially accountable. Data collection was not distracted by philosophical discussions on what was social accountability and was the school aspiring to be socially accountable” [3] [p.3]. Acknowledgements Thank you to the Training for Health Equity Network (THEnet) Colleagues who were research mentors at three sites. Thank you to participants who invited RP into their workplaces and homes and gave their time and their knowledge to this research. Funding Funding for this project from the Australian Postgraduate Award; College of Medicine and Dentistry, James Cook University and the Graduate Research School, James Cook University. Apart from approving the project proposal, the funding sources have not been involved in the project in any way. Availability of data and materials The datasets generated during the current study are not publicly available due to confidentiality. An interview guide for the faculty/staff, student, health professional and community interviews are attached to this publication (see Additional file 1). Authors’ contributions RP conceived and designed the study, collected and analysed the data and drafted the first manuscript. JT, SL and JJ made substantial contributions to conception and design of the study and the interpretation of data. All authors have been involved in revising the manuscript critically for important intellectual content. All authors have given final approval of the version to be published. All authors have agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Authors’ information Dr Robyn Preston BA (DevS) (Hons), PGCertDisasRefugHlth, MHSc (HealthProm), PhD is a Lecturer in the Discipline of General Practice and Rural Medicine at the College of Medicine and Dentistry; a Cohort Doctoral Studies Mentor with the Australian Institute of Tropical Health and Medicine and a member of the Anton Breinl Research Centre for Health Systems Strengthening at James Cook University, Townsville, Australia. Prof Sarah Larkins, MBBS, BMedSci, MPH&TM, PhD, FRACGP, is an academic general practitioner, Associate Dean, Research, College of Medicine and Dentistry, Co-Director Anton Breinl Research Centre for Health Systems Strengthening, Australian Institute of Tropical Health and Medicine James Cook University. Adjunct Associate Professor Judy Taylor PhD is researching community development with the College of Medicine and Dentistry, James Cook University, Townsville Australia and is a member of the Anton Breinl Research Centre for Health Systems Strengthening and the Australian Institute of Tropical Health and Medicine. Dr Jenni Judd, DHSc MPH MEd DipHPE is an Adjunct Associate Professor-Principal Research Fellow College of Medicine and Dentistry, Division of Tropical Health and Medicine and Anton Breinl Research Centre for Health Systems Strengthening, and the Australian Institute of Health and Tropical Medicine, James Cook University and an Adjunct Associate Professor, School of Public Health and Social Work, Queensland University of Technology. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The James Cook University (JCU) Human Research Ethics Committee (HREC) granted ethics approval (approval number H4245). All participants were reassured that no information recorded in the interviews would in any way identify individuals. Potential participants were provided with a participant information sheet which explained the aims of the project and information about their participation in the interviews. Participants completed a written consent form prior to participation. Interviews were held at a time and in a place convenient for participants. ==== Refs References 1. Boelen C Heck J Defining and measuring the social accountability of medical schools 1995 Geneva World Health Organisation 2. Boelen C Woollard R Social accountability: The extra leap to excellence for educational institutions Med Teach 2011 33 8 614 619 10.3109/0142159X.2011.590248 21774646 3. Preston R, Larkins S, Taylor J, Judd J. From personal to global: Understandings of social accountability from stakeholders at four medical schools. Medical Teacher. 2016:1-8. 10.3109/0142159X.2015.1114596. 4. Boelen C A new paradigm for medical schools a century after Flexner’s report Bull World Health Organ 2002 80 7 592 593 12163925 5. Boelen C Building a socially accountable health professions school: Toward unity for health Education for Health 2004 17 2 223 231 10.1080/13576280410001711049 15763765 6. Boelen C: Social accountability and the future of medical education. Wilson Centre for Research in Education, University of Toronto, Centre de pédagogie appliquée aux sciences de la santé, Université de Montréal, Health Canada; 2008 [http://www.afmc.ca/fmec/pdf/National%20Literature%20Reviews.pdf]. 7. Boelen C Social accountability and excellence Educ méd 2009 12 4 199 205 10.4321/S1575-18132009000500001 8. Boelen C Social accountability: The mark of excellence in health professional education Teaching to cure - Teaching to Care. Balancing the dynamics in health sciences education.: 30 June - 2 July 2011 2011 South Africa Faculty of Health Sciences, North-West University, Potchefstroom Campus 9. Boelen C Dharamsi S Gibbs T The social accountability of medical schools and its indicators Educ Health 2012 25 180 194 10.4103/1357-6283.109785 10. Boelen C Woollard R Social accountability and accreditation: A new frontier for educational institutions Med Educ 2009 43 9 887 894 10.1111/j.1365-2923.2009.03413.x 19709014 11. Bhutta Z Chen L Cohen J Crisp N Evans T Fineberg H Frenk J Garcia P Horton R Ke Y Education of health professionals for the 21st century: A global independent commission Lancet 2010 375 9721 1137 1138 10.1016/S0140-6736(10)60450-3 20362799 12. Carter S Little M Justifying knowledge, justifying method, taking action: Epistemologies, methodologies, and methods in qualitative research Qual Health Res 2007 17 10 1316 1328 10.1177/1049732307306927 18000071 13. Patton M Qualitative research and evaluation methods 2002 3 London Sage 14. Bleakley A Brice J Bligh J Thinking the post-colonial in medical education Med Educ 2008 42 3 266 270 10.1111/j.1365-2923.2007.02991.x 18275413 15. Crane J Scrambling for Africa? Universities and global health Lancet 2011 377 1388 1390 10.1016/S0140-6736(10)61920-4 21074254 16. Wolinsky F The professional dominance perspective, revisited Milbank Q 1988 66 Supplement 2 33 47 10.2307/3349913 3075009 17. Germov J Germov J Challenges to medical dominance Second opinion 2002 South Melbourne Oxford University Press 18. Turner B Medical power and social knowledge 1995 2 London Sage Publications 19. Wartenberg T The forms of power 1990 Philadelphia Temple University Press 20. Hafferty F Theories at the crossroads: A discussion of evolving views on medicine as a profession Milbank Q 1988 66 Supplement 2 202 225 10.2307/3349923 3251139 21. Nugus P Greenfield D Travaglia J Westbrook J Braithwaite J How and where clinicians exercise power: Interprofessional relations in health care Soc Sci Med 2010 71 5 898 909 10.1016/j.socscimed.2010.05.029 20609507 22. Kellner D Critical theory and the crisis of social theory Sociol Perspect 1990 33 1 11 33 10.2307/1388975 23. Kincheloe J McLaren P Denzin N Lincoln Y Rethinking critical theory and qualitative research The Sage handbook of qualitative Research 2005 Thousand Oaks Sage Publications Inc 303 342 24. Thomas G How to do your case study: A guide for students and researchers 2011 London Sage Publications Inc. 25. Stake RE The art of case study research 1995 Thousand Oaks Sage Publications 26. Baxter P Jack S Qualitative case study methodology: Study design and implementation for novice researchers Qual Rep 2008 13 4 544 559 27. Strasser R Neusy A-J Context counts: Training health workers in and for rural and remote areas Bull World Health Organ 2010 88 777 782 10.2471/BLT.09.072462 20931063 28. Pálsdóttir B Neusy AJ Global health: Networking innovative academic institutions Infect Dis Clin N Am 2011 25 2 337 346 10.1016/j.idc.2011.02.001 29. Pálsdóttir B Neusy A-J Reed G Building the evidence base: Networking innovative socially accountable medical educational programs Educ Health 2008 21 2 1 6 30. Neusy A-J Pálsdóttir B A roundtable of innovative leaders in medical education MEDICC Rev 2008 10 4 20 24 21483332 31. Pálsdóttir B Neusy AJ Transforming medical education: Lessons learned from THEnet: A paper prepared for The Commission on Education of Health Professionals for the 21st Century 2010 32. Neusy A-J Pálsdóttir B Roundtable: Revisiting innovative leaders in medical education MEDICC Rev 2011 13 3 6 11 33. Yin R Enhancing the quality of case studies in health services research Health Serv Res 1999 34 5 Part II 1209 1224 10591280 34. Liamputtong P Qualitative research methods 2009 3 South Melbourne Oxford University Press 35. Abma TA Stake RE Science of the particular: an advocacy of naturalistic case study in health research Qual Health Res 2014 24 8 1150 1161 10.1177/1049732314543196 25028158 36. QSR International: NVivo 10. In.: QSR International Pty Ltd; 2012. Computer Software. 37. Silver C Lewins A Using software in qualitative research: a step by step guide 2014 2 London Sage Publications Inc 38. Miles M Huberman A Qualitative data analysis: An expanded sourcebook 1994 2 Thousand Oaks Sage Publications Inc 39. Seale C Qualitative methods: Validity and reliability Eur J Cancer Care 2001 10 2 133 134 10.1046/j.1365-2354.2001.0253b.x 40. Murray R Larkins S Russell H Ewen S Prideaux D Medical schools as agents of changes: Socially acountabile medical education Med J Aust 2012 196 10 653 10.5694/mja11.11473 22676883 41. Biggs J Wells R The social mission of Australian medical schools in a time of expansion Aust Health Rev 2011 35 4 424 429 22126944 42. World Health Organization Policy brief on regulation of health professions education 2013 Geneva World Health Organization 43. World Health Organization Policy brief on accreditation of institutions for health professional education 2013 Geneva World Health Organization 44. Howe A Twelve tips for community-based medical education Med Teach 2002 24 1 9 12 10.1080/00034980120103423 12098450 45. Rourke J Social accountability in theory and practice Ann Fam Med 2006 4 s1 S45 48 10.1370/afm.559 17003163 46. Health Canada Social accountability: A vision for Canadian medical schools 2001 Ottawa Health Canada 47. Lindgren S Karle H Social accountability of medical education: Aspects on global accreditation Med Teach 2011 33 8 667 672 10.3109/0142159X.2011.590246 21774655 48. Sen Gupta T, Murray R, McDonell A, Murphy B, Underhill A. Rural internships for final year students: Clinical experience, education and workforce. Rural Remote Health. 2008;827. 49. Dussault G Franceschini M Not enough there, too many here: Understanding geographical imbalances in the distribution of the health workforce Hum Resour Health 2006 4 1 12 10.1186/1478-4491-4-12 16729892 50. Towle A Godolphin W Dornan T Mann K Scherpbier A Spencer J Patient involvement in health professional education Medical education theory and practice 2011 Edinburgh Churchill Livingstone Elsevier 65 77 51. Australia Department of health: rural clinical training and support (RCTS) 2011-2014 - operational framework: attachment a – rural clinical training and support (RCTS) program parameters 2012 Canberra Health Workforce Australia 52. Hunt J Bonham C Jones L Understanding the goals of service learning and community-based medical education: A systematic review Acad Med 2011 86 2 246 251 10.1097/ACM.0b013e3182046481 21169780 53. Holley KA Rural minority student engagement with a healthcare pipeline program J Res Rural Educ (Online) 2013 28 4 1 11 54. Kelly L Walters L Rosenthal D Community-based medical education: Is success a result of meaningful personal learning experiences? 2014 55. Woollard R Caring for a common future: Medical schools’ social accountability Med Educ 2006 40 4 301 313 10.1111/j.1365-2929.2006.02416.x 16573665 56. Evans C The dean as spiritual leader Acad Med 1998 73 6 645 648 10.1097/00001888-199806000-00009 9653401 57. Jippes M Driessen EW Majoor GD Gijselaers WH Muijtjens AMM van der Vleuten CPM Impact of national context and culture on curriculum change: A case study Med Teach 2013 35 8 661 670 10.3109/0142159X.2013.785629 23627360 58. Kayyal M Gibbs T Managing curriculum transformation within strict University governance structures: An example from Damascus University Medical School Med Teach 2012 34 8 607 613 10.3109/0142159X.2012.689033 22830318 59. Cohen J Dannefer E Seidel H Weisman C Wexler P Brown T Brieger G Margolis S Ross L Kunitz S Medical education change: A detailed study of six medical schools Med Educ 1994 28 5 350 360 10.1111/j.1365-2923.1994.tb02545.x 7845253 60. Guilbert J Why is it taking so long for healthcare professional education to become relevant and effective? What can be done? 2014 61. Boelen C Medical education reform: The need for global action Acad Med 1992 67 11 745 749 10.1097/00001888-199211000-00007 1418252 62. Ghoshal S, BonTempo J. Modern communities of practice: Recommendations for building, maintaining and measuring impact. Health Communication Capacity Collaborative, USAID, John Hopkins University; 2014 63. Wenger E McDermott R Snyder W Cultivating communities of practice: A guide to managing knowledge 2002 Boston Harvard Business School 64. Galukande M Nakasujja N Sewankambo N Social accountability: A survey of perceptions and evidence of its expression at a Sub Saharan African university BMC Med Educ 2012 12 1 96 10.1186/1472-6920-12-96 23075090 65. Reddy AT Lazreg SA Phillips RL Bazemore AW Lucan SC Toward defining and measuring social accountability in graduate medical education: A stakeholder study J Grad Med Educ 2013 5 3 439 445 10.4300/JGME-D-12-00274.1 24404308 66. Puschel K Rojas P Erazo A Thompson B Lopez J Barros J Social accountability of medical schools and academic primary care training in Latin America: Principles but not practice Fam Pract 2014 67. Rudolf M Reis S Gibbs T Murdoch Eaton D Stone D Grady M Berlin A Blair M Essa-Hadad J Spitzer-Shohat S How can medical schools contribute to bringing about health equity? Israel J Health Policy Res 2014 3 1 17 10.1186/2045-4015-3-17 68. Jamieson J Snadden D Dobson S Frost H Voyer S Health disparities, social accountability and postgraduate medical education: A paper commissioned as part of the environmental scan for the Future of Medical Education in Canada Postgraduate Project 2011 69. Ritz S Beatty K Ellaway R Accounting for social accountability: Developing critiques of social accountability within medical education 2014 70. McCrea ML Murdoch-Eaton D How do undergraduate medical students perceive social accountability? Med Teach 2014 36 10 867 875 10.3109/0142159X.2014.916784 25072172

PMC005xxxxxx/PMC5002163.txt

==== Front Biomed Eng OnlineBiomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 2675888511810.1186/s12938-015-0118-1ResearchFusion of detected multi-channel maternal electrocardiogram (ECG) R-wave peak locations Yu Qiong piaoliyin@163.com 1Guan Qun tsy8088@sina.com 2Li Ping zyyliping@sina.com 2Liu Tie-Bing njbull@163.com 2Huang Xiao-Lin huang_xiaolin@163.com 1Zhao Ying yingzhao@nju.edu.cn 1Liu Hong-Xing njhxliu@nju.edu.cn 1Wang Yuan-Qing yqwang@nju.edu.cn 11 School of Electronic Science and Engineering, Nanjing University, Xianlin Campus, Nanjing, 210023 China 2 Nanjing General Hospital of Nanjing Military Command, Nanjing, 210002 China 8 1 2016 8 1 2016 2016 15 1 417 6 2015 22 12 2015 © Yu et al. 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 Almost all promising non-invasive foetal ECG extraction methods involve accurately determining maternal ECG R-wave peaks. However, it is not easy to robustly detect accurate R-wave peaks of the maternal ECG component in an acquired abdominal ECG since it often has a low signal-to-noise ratio (SNR), sometimes containing a large foetal ECG component or other noises and interferences. This paper discusses, under the condition of acquiring multi-channel abdominal ECG signals, how to improve the robustness of maternal ECG R-wave peak detection. Methods On the basis of summarising the current single channel ECG R-wave peak detection methods, the paper proposed a specific fusion algorithm of detected multi-channel maternal ECG R-wave peak locations. The proposed entire algorithm was then tested using two databases; one database, created by us, was composed of 343 groups of 8-channel data collected from 78 pregnant women, and the other one, called the challenge database, was from the Physionet/Computing in Cardiology Challenge 2013, including 175 groups of 4-channel data. When using these databases, each group of data was classified into two parts, called the training part and the validation test part respectively; the training part was the first 8.192 s of each group of data and the validation test part was the next 8.192 s. Results To show the results, three evaluation parameters—sensitivity (Se), positive predictive value (PPV) and F1—are used. The validation test results for the database we collected are Se = 99.93 %, PPV = 99.98 %, and F1 = 99.95 %, while the results for the challenge database are Se = 99.91 %, PPV = 99.86 %, and F1 = 99.88 %. Conclusion The results of the test show that the robustness of our proposed whole fusion algorithm was superior to that of other outstanding algorithms for maternal R-wave detection, and is much better than that of single channel maternal R-wave detection algorithms. Keywords Multi-channelAbdominal ECG signalR-wave detectionFusionthe National Natural Science Foundation of China61271079Liu Hong-Xing issue-copyright-statement© The Author(s) 2016 ==== Body Background Hospitals have strived to include foetal ECG measurements in foetal monitoring, acquiring the foetal ECG waveform and calculating the foetal heart rate. A lot of studies have been conducted to extract foetal ECG from recorded pregnant abdominal ECG signals [1–5]. However, recent revelations have indicated that the key step for a successful foetal ECG extraction is to locate the maternal ECG R-wave peaks accurately, after which, the maternal ECG component can be accurately estimated and cleanly cancelled [3–5]. It is very challenging to robustly detect accurate R-wave peaks of the maternal ECG component in an acquired abdominal ECG. Although the maternal ECG component is usually much greater than the foetal ECG component in an abdominal ECG signal, there are still some cases in which the foetal ECG component is just as great as or greater than the maternal ECG component. In these cases, a lot of incorrect maternal ECG R-wave peak detections will inevitably appear when using a single channel ECG R-wave peak detection method. This is a major obstacle for implementing foetal ECG monitoring, although it has yet to gain widespread attention. Multi-point measurements provide an opportunity to improve the robustness of maternal ECG R-wave peak detection. One easy way is to synchronously collect one thoracic ECG signal when recording abdominal ECG signals to help locate the R-wave peaks of the maternal ECG component in recorded abdominal ECG signals [6]. However, in this method, a disadvantage is that an extra thoracic electrode is needed, which is undesirable in clinical application. On the other hand, noises and disturbances in the thoracic ECG signal sometimes will also result in incorrect R-wave peak detection when using a single channel ECG R-wave peak detection method. Alternatively, we can try to combine the acquired multi-channel abdominal ECG signals to produce a pure maternal ECG signal as a substitute for the thoracic ECG signal acquisition, based on a principal components analysis (PCA) or independent components analysis (ICA), but it is usually difficult to automatically select the pure maternal ECG signal from PCA or ICA outputs [2, 7–9]. In the Physionet/Computing in Cardiology Challenge 2013, there appear two outstanding maternal R-wave detection methods based on ICA [8, 9], called ICA$_{ACC}$ [8] and ICA$_{SMI}$ [9], respectively. The main steps of the ICA$_{ACC}$ algorithm are: (1) using FASTICA to produce N independent components for the N-channel signals, (2) initial detection on both the inputs and outputs of ICA, (3) calculating the ACC measure for each output and choosing the purest maternal ECG from the N outputs using ACCs, and (4) further processing the MQRS detections on this chosen channel using kernel density estimation, a matched filter and RR correlation to obtain more accurate MQRS detections as the final output. For the ICA$_{SMI}$ algorithm, the main steps are: (1) performing ICA on N-channel signals to obtain N independent components, (2) initial detection on both the inputs and the outputs of ICA, (3) calculating a parameter called SMI for each input and output, and (4) choosing the maternal QRS time series as the final output from the inputs and outputs using SMIs. Noticeably, in previous research [10], an effective and simple scheme was used to improve the robustness of foetal R-wave detection after removing maternal components from the abdominal ECG signals. Its main idea is: first, to detect the foetal R-wave peaks of each estimated foetal ECG signal using a single channel ECG R-wave peak detection method, and then, to vote with the detected results of all channels’ estimated foetal ECG signals to correct the incorrect peak detections. Referencing above, this paper will apply the new thought of fusion to the robust detection of maternal R-wave peaks of abdominal ECG signals. The proposed whole fusion algorithm consists of the first modification of the initial detection, channel selection, second modification, and voting. The following chapters are arranged as follows. In the “Methods” chapter, first we will introduce, in detail, the two databases used for our research. Second, in the following subsections, each step of our proposed fusion algorithm will be thoroughly described and a block diagram for the algorithm will be provided. Finally, we will describe the parameter training and validation test method for the whole fusion algorithm. In the “Results” chapter, we will present the validation test results for the whole fusion algorithm, as well as for the two ICA-based maternal R-wave detection algorithms (ICA$_{ACC}$, ICA$_{SMI}$) for comparison. The final two chapters are the “Discussion” and “Conclusion”, respectively. Methods The block diagram for our proposed fusion algorithm can been found in Fig. 1. Our entire algorithm consists of five blocks: initial detection, first modification, channels selection, second modification and voting. In the following subsections, each individual block will be thoroughly explained.Fig. 1 Block diagram for the entire algorithm Data description Database 1 (collected ourselves) The authors collected actual ECG data from a hospital for our research. The methods of collecting the data can be found in the literature [11]. We conducted the collecting experiment using 78 pregnant women in the 38–40th week of gestation in the Obstetrics Branch of Nanjing General Hospital of Nanjing Military Command. The data collecting system was composed of a standard 12-lead ECG machine and a PC machine. The standard 12-lead ECG machine, made by Ni-hon Kohden Corporation with the model No. 1350p, was connected to the PC machine with a dedicated USB cable. The ECG machine acted as a data acquisition module. The parameters of this machine must be preset; the sample frequency was set to 500 Hz, the cut-off frequency of anti-aliasing filter was set to 75 Hz, and the switches for baseline drift suppression and EMG interference suppression were turned on. Previous research [11] has described the placement of the electrodes on the abdominal surface more clearly. We collected 343 groups of data in all, and every group of data was comprised of 8-channel abdominal ECG signals with a duration of 24 s. For every pregnant woman, we collected three to six groups of data. Database 2 (challenge data) The challenge data were gathered from the Physionet/Computing in Cardiology Challenge 2013. The datasets used for the challenge were obtained from five different sources (including real and simulated data), yielding a total of 447 records. All records were formatted to have a 1 kHz sampling frequency, 1-min duration, and four channels of non-invasive abdominal maternal ECG leads. The whole records were divided into three datasets for the challenge: Set A (75 records, both records and reference for FQRS locations were made public), Set B (100 records, only the records were made public), and Set C (272 records, both records and reference for FQRS locations were withheld from the public). The more detailed information about the database can be seen in the literature [12]. In our test, we only used the public datasets A and B for the test. Initial detection A single channel ECG R-wave peak detection algorithm was the basis of the fusion algorithm for multi-channel ECG signals. Researchers have proposed a variety of single channel ECG R-wave peak detection algorithms, and all of these algorithms can be generalised as three steps: (1) taking a pre-processing stage for the original single channel ECG signal to improve its signal-to-noise ratio, (2) detecting the R-wave peaks on the enhanced signal obtained through the pre-processing stage, and (3) taking the initially detected R-wave peak positions of the enhanced signal as benchmarks and searching the real R-wave peaks within their neighbourhoods on the original wave. In the above three steps, the common methods for the first step are the difference method [13], the Hilbert transform method [14], the template matching method [15], the wavelet transform method [16, 17] and so on. For the second step, the common methods include all kinds of threshold setting and local searching methods [18, 19]. Among the pre-processing methods mentioned above, the robustness of the wavelet transform method was much stronger when compared to the others [20]. As long as the mother wavelet has been chosen appropriately, the wavelet transform method will suppress the baseline drift and high frequency noise at a high level. In detail, there are two ways for the wavelet transform method to enhance the original single channel ECG signal for R-wave detection: one, the modulus maxima method [21], computes the continuous wavelet transform coefficients of the original signal based on a selected wavelet and a certain scale and then takes the modulus of wavelet transform coefficients as the enhanced signal for the R-wave detection; the other was to compute the wavelet transform coefficients of the signal, followed by taking some scales’ coefficients to reconstruct an enhanced ECG signal with larger SNR [22]. This paper will apply the above first usage of the wavelet transform to do the pre-processing. In addition, before the wavelet transform, some pre-filtering methods are applied to each channel of the multi-channel data. First, a bandpass FIR filter between 0.5–80 Hz was used to suppress some noises, such as baseline wander, muscular artefacts and so on. Second, to remove the power line, two notch filters (at 50 and 60 Hz) were used. Third, to further suppress the baseline wander, a two-order smooth filter (0.2 s window) was used to obtain it and then it was subtracted from the signal. First modification The first modification was used to prepare for the following channel selection, since there are usually many errors in the initial detection results and these errors will inevitably increase the difficulties in correct channel selection. The modification algorithm we used here was based on comparing each single signal’s RR interval value with the median value of all RR interval values of the channel. The main steps are as follows:Step 1:After initial detection, for each single channel of the multi-channel signals, compute its RR interval values and mark them as RRs. Step 2:Compute the median value of the RRs and mark it asRRm. Step 3:Compute the difference value between each RR interval value andRRm; if the difference value is smaller than one threshold (named thres1_1; see “Results”), one of the corresponding two detected R-wave peaks needs to be removed; if the difference value is greater than one threshold (named thres1_2; see “Results”), an additional R-wave peak needs to be detected between the corresponding two detected R-wave peaks. Channel selection Channel selection was an important step to guarantee the robustness of the whole fusion algorithm, since once the number of channels with erroneous single channel maternal R-wave peak detection was greater than or equal to half of the number of all channels, we cannot vote to get accurate maternal R-wave peak positions (described in next subsection). To make the right channel selection, we used two kinds of indices to determine whether we should accept a signal. One index, named index1, was calculated based on the template method and the other, index2, was calculated using the RR interval values of the signal. The main steps of channel selection are as follows:Step 1:After the first modification (above), re-calculate RR intervals on each channel of the multi-channel signals, then compute the median value of RR interval values of each channel and mark them as RRm1, RRm2,…, RRmi, …, RRmn, respectively (n refers to the number of channels of the multi-channel signals). Step 2:Exclude the channels whose medians are lower than 0.4 s or higher than 1.6 s and mark the medians of the remaining channels as RLm1, RLm2, …, RLmi, …, RLmk (k is the number of remaining channels). Step 3:Compute the median of RLm1, RLm2, …, RLmi, …, RLmk and mark it as RLmm1. Step 4:Perform further exclusions. For each remaining channel, as we previously mentioned, first compute two indices, index1 and index2. Second multiply index1 by index2 and mark the result as index. Third, compute the maximal value of all indexes of the remaining channels and mark it as ma. Finally, compare each index with ma and if an index is greater than or equal to 0.5*ma, the corresponding channel is selected. The method of calculating index1 of each remaining channel is as follows. For each channel’s first-modified signal, first we need to make a maternal wave complex template. In our paper, a maternal wave complex, whose duration is assigned as RLmm1, is defined as 0.5*RLmm1 before and 0.5*RLmm1 after the detected maternal R-wave peak location. Averaging all the maternal wave complexes synchronised on their R-wave peaks results in the maternal wave complex template. Second, compute the correlation coefficient between the template and each actual maternal wave complex. Third, compute the number of correlation coefficients that are greater than or equal to one threshold (named thres2_1; see “Results”) and then compute the ratio of this number to the number of all maternal R-wave peaks obtained by the first modification. The ratio value was taken as the index1of this channel. Regarding the method of calculating index2, for each remaining channel, first compute the absolute difference value between itsRLmi and RLmm1, marking the absolute difference value as med_diff. Second, compute the minimum of all med_diffs of the remaining channels and mark it as mi. Finally, for each med_diff, if the difference value between it andmi is lower than one threshold (named thres2_2; see “Results”), the index2 of the corresponding channel is set to 0.3, and if the different value is between the thres2_2 and 2*thres2_2, the index2 of the corresponding channel is set to 0.2, while in other situations, the index2 is set to 0.1. Second modification This section is also very important because it corrects the error detections on the selected channels more carefully than the first modification does, reducing the pressure to the voting and improving the robustness of the whole fusion algorithm to some degree. The second modification includes two parts: the first part is a modification using the template method and the second is a modification through the RR interval values of the signal. The steps of the first part are as follows. First, compute the median value of all of the RR interval values of each selected channel. Then, compute the median value of all of the selected channels’ median values and mark it as RRmm2. Second, for each selected channel, make a maternal wave complex template whose duration is RRmm2 and then compute the correlation coefficient between the template and each actual maternal wave complex. Finally, remove the detected maternal R-wave peaks whose correlation coefficients are lower than one threshold (named thres3_3; see “Results”). The fundamental idea of the second part is just like that of the first modification. It is also based on the difference values between the RR interval values and the median value of the RR intervals of each selected channel. If there is a difference value which is less than one threshold (named thres3_1; see “Results”), one of the corresponding two detected R-wave peaks will be removed; if the different value is greater than one threshold (named thres3_2; see “Results”), another R-wave peak needs to be detected between the corresponding two detected R-wave peaks. However, there is an important detail that is different from the first modification: for different groups of data, the thresholds (thres3_1 and thres3_2) are different and are assigned according to one parameter (named med_cha) related to the actual heart rate variability (HRV) of the group of data. The method of calculating the parameter, med_cha, is described as follows: After finishing the first part of the second modification, for each selected channel of the multi-channel signals, first, we compute its RR interval values, named RRs, and the median value of them, named RRmedian. Second, we exclude the RR interval values that are greater than 1.4*RRmedian or less than 0.6*RRmedian. Third, we group the remaining RR interval values according to the original sequence and make each group consist of five values (each group can partially overlap). The number of groups can be computed using the formula ceil(length(N)/5), where (N) is the number of all remaining RR interval values of this channel. Finally, we calculated the absolute difference value between the maximal value and the minimum value of each group, and the median value of all absolute difference values for each selected channel. After finishing all of the above steps, the minimum of the median values for all selected channels was the parameter, med_cha, we needed. Thres3_1 is set to -a*med_cha and thres3_2 is set to a*med_cha (the parameter a was a constant value). At the same time, the absolute value of thres3_1 and thres3_2 must be restricted to between 0.12 s and 0.25 s. Voting Voting is the final step of our fusion algorithm and further corrects the error detections. Noticeably, even if the channel selection is not totally correct, we can obtain the correct results through the voting, as long as the number of error selected channels is lower than half of the number of all selected channels. The voting part of the whole proposed fusion algorithm is based on the theory of clustering, and its basic steps are: (1) taking the detected and corrected maternal R-wave peak positions of each selected channel of the multi-channel pregnant abdominal ECG signals, (2) projecting or drawing the all detected R-peak positions of selected channels on a time axis and regarding them as samples, (3) clustering all of the samples according to their distances and labelling their classes, (4) determining whether or not the number of samples in a clustered class is greater than or equal to half of the number of all channels (if so, the class and its samples will be kept, otherwise the class and its samples will be deleted), and (5) for each remaining class, taking the median value of its samples as its representative R-wave peak position, and then searching for the peak in the neighbourhood of the representative position on each original signal, resulting in the all exact maternal R-wave peak positions of the multi-channel signals. In Step 3 above, if the distance between two samples was less than 30 ms, we assume that they belong to the same class; otherwise we assume that they belong to different classes [11]. In Step 4, we removed the classes that have few samples since we assume that they are false classes clustered with incorrectly detected peaks. For explicitness, the voting is illustrated in Fig. 2, taking the 4-channel signals as an example and letting red circles represent the maternal R-wave peak positions of each signal. The pseudo code of the voting is described as follows: Fig. 2 Illustration of voting. The red circles refer to the detected maternal R-wave peak locations. a The initial maternal R-wave detection results of the four-channel signals; b the results after projecting the initial detection results on the same axis; c cluster results for all locations; d remaining classes after removing error classes; e final detection results after back projecting the remaining classes to the real four-channel signals Parameter training The parameter training was used for more accurately determining the parameters’ values in our algorithm. For each group of data in each database, the first 8.192 s of the data are used as the training data. The training can be further divided into two kinds—offline training and online training. In our paper, both thres3_1 and thres3_2 training belong to online training since we must train the two parameters for each group of data and different individual data have different HRVs. Regarding the other parameters, their trainings belong to offline training, which means that once the parameters are adjusted to be good with some groups of data, the trained parameters will stay invariant for other groups of data. All of the values of the trained parameters are shown in the "Results" section. Validation test method The validation test part involves three algorithms (our fusion algorithm, ICA$_{ACC)$ and ICA$_{SMI}$), as well as two above-mentioned databases (Database 1 and Database 2). Every algorithm will be tested using the two databases to compare with each other. In the validation test for each group of data in each database, only the second 8.192 s of the data are used, since we need to determine the correctness of detections through our observations and there would be a significant burden to the eyes if we used more sample points for each signal in the multi-channel signals. Additionally, though we just decided the correctness of the detections through our observations, we believe that the decisions are reliable because the data used are all multi-channel and in almost all conditions, there is at least one channel whose MQRS peaks are clearly visible. To show the results, we use three measurement parameters—Se, PPV, and F1 [9]. The Se, PPV, and F1 are obtained by first counting the number of correctly detected maternal R-wave peaks (TP), the number of extra falsely detected maternal R-wave peaks (FP), and the number of missed maternal R-wave peaks (FN) of all groups of data used, and then calculating the three parameters’ values according to previous research [9]. In the comparison of three algorithms, the average time spent after the initial detection for each group of data will be also included. Results In the test, first, we detected the maternal R-wave peaks on each signal of the multi-channel ECG signals respectively using the wavelet-based single channel R-wave detection method. The parameters of the wavelet-based method for pre-processing are set as follows: the selected mother wavelet is bior1.5 [23], the selected decomposition scale for Database 1 is 2^5 and for Database 2, it is 2^6 according to the frequency characteristics of the data (the sampling frequency). Next, we conducted the parameter training of the fusion algorithm with Database 1, and through the training part, the parameters were finally assigned as follows: in the first modification, thres1_1 = −0.30 s and thres1_2 = 0.30 s; in the channel selection, thres2_1 = 0.6, thres2_2 = 0.06 s; in the second modification, thres3_3 = 0.5, a = 2.5. Since thres3_1 and thres3_2 are usually different for different groups of data, their values are not included in this manuscript. After finishing the training, we conducted the test with Database 1 for our fusion algorithm. We also tested the ICA$_{ACC}$, ICA$_{SMI}$, respectively, with Database 1 to compare them to our whole fusion algorithm. The results are shown in Table 1.Table 1 Results of the direct single channel maternal R-wave detection algorithm and the results of the fusion algorithm, ICA$_{ACC}$ and ICA$_{SMI}$, with Database 1 Se (%) PPV (%) F1 (%) t (s) Initial detection 97.08 93.79 95.41 ICA$_{ACC}$ algorithm 99.37 99.82 99.59 0.9232 ICA$_{SMI}$ algorithm 99.81 99.55 99.68 0.0073 Our fusion algorithm 99.93 99.98  99.95 0.0911 Similar to the above, next we conducted the test for the fusion algorithm and the other two algorithms with Database 2, and the results are shown in Table 2.Table 2 Results of the direct single channel maternal R-wave detection algorithm and the results of the fusion algorithm, ICA$_{ACC}$ and ICA$_{SMI}$, with Database 2 Se (%) PPV (%) F1 (%) t (s) Initial detection 94.15 89.12  91.57  ICA$_{ACC}$ algorithm 98.93 99.41  99.17  1.7973 ICA$_{SMI}$ algorithm 99.74 98.98  99.36  0.0016 Our fusion algorithm 99.91 99.86 99.88  0.0704 In addition, in Fig. 3, the record ‘b20’ in Database 2 was taken as an example to show the performance of the three methods. In the figure, ‘★’ refers to the maternal R wave peak location obtained by the ICA$_{ACC}$ algorithm, ‘□’ means the location obtained by the ICA$_{SMI}$ algorithm, and the red lines represent the locations of the R-wave peaks detected using our fusion algorithm. Observing it carefully, we can find that the fusion algorithm has detected all of the maternal R-wave peaks correctly while the other two algorithms have not.Fig. 3 Example showing the performance of the three methods. Filled black star refers to the maternal R-wave peak location obtained with the ICA$_{ACC}$ algorithm, white square means the location obtained by the ICA$_{SMI}$ algorithm, and the red lines represent the locations of the R-wave peaks detected by our fusion algorithm Through the results of test for the fusion algorithm, we can see that it is a large improvement for the single channel maternal R-wave algorithms. Additionally, considering the results of the comparison test between the three algorithms, we also found that though our algorithm does not have the fastest speed, its detection results are the best and are superior to the other two algorithms. Discussion Our proposed fusion algorithm and other two comparative fusion algorithms are much better than single channel maternal R-wave detection algorithms. Let’s take an example to show this. The data for this example was the first channel signal of the 88th group of data in Database 1. Although the data have a low signal-to-noise ratio (containing a large foetal ECG component) and we obtained bad maternal R-wave detection results when directly using the single channel R-wave detection methods, shown in Fig. 4a, the three fusion algorithms in this paper were all able to give perfect maternal R-wave detection results, shown in Fig. 4c. Thus, with the correct maternal R-wave detection results, the ideal final foetal ECG were extracted using the foetal ECG extraction method in previous research [24], shown in Fig. 4d, which was in contrast with the bad foetal extraction result based on the single-channel R-wave detection results using the same extraction method [24], shown in Fig. 4b.Fig. 4 Foetal extraction results based on one single channel detection method and our fusion method. The red circles indicate the detected maternal R-wave peak locations. a The maternal ECG R-peak detection results obtained by initial single channel R-wave detection method; b the remaining component after maternal ECG cancelling based on the single channel R-wave detection results; c the maternal ECG R-peak detection results obtained by our whole fusion algorithm; d the remaining component after maternal ECG cancelling based on the detection results of our fusion algorithm The comparative fusion algorithm ICA$_{SMI}$, involving choosing the channel with the best maternal QRS peak detection, was simple and cost the least amount of time, but it was found to be very sensitive to the initial detection, since it did not modify the initial detection results, and once the initial detection results on the selected output was incorrect due to other reasons, such as low quality, this algorithm was eventually not able to obtain the correct detection results. Regarding another comparative fusion algorithm of ICA$_{ACC}$, although it is less sensitive to the initial detection step than ICA$_{SMI}$, it was found that a few times its channel selecting for the maternal component after doing ICA was incorrect. Let’s take the record ‘b76’ in Database 2 as an example. Both its initial detection results on pre-processed signals and the detection results on ICA outputs were shown in Fig. 5, where the red circles refer to the detected maternal R-wave peak locations. In testing, the ICA$_{ACC}$ algorithm selected the wrong maternal component (the second independent component separated, IC2), since IC2 had the highest ACC value, resulting in the incorrect final detection results, shown in Fig. 6. Nevertheless, when we applied our proposed fusion algorithm to this group of data, the correct results are obtained; in Fig. 6, the red circles refer to the maternal R-wave peak locations detected by ICA$_{ACC}$ and the red lines represent the locations of the R-wave peaks detected by our fusion algorithm. In addition to above, the fusion algorithm of ICA$_{ACC}$ was much more complex and cost the most time.Fig. 5 Example showing the ICA$_{ACC}$’s failure selecting for the maternal component from ICA outputs. The red circles refer to the detected maternal R-wave peak locations; IC2 was selected incorrectly for its highest ACC value because the agreement between the detection results of IC2 and the initial ch2 was greater than the others Fig. 6 Incorrect final detection Results of ICA$_{ACC}$ due to its failure selecting for the maternal component. The red circles refer to the maternal R-wave peak locations detected by ICA$_{ACC}$ and the red lines represent the locations of the R-wave peaks detected by our fusion algorithm Comparing our whole fusion algorithm with other two algorithms, the time required was a little more than ICA$_{SMI}$, but much less than ICA$_{ACC}$. In comparison with the ICA$_{SMI}$, our fusion algorithm’s dependence on the initial detection was much lower since it has done a lot of work on the modification. In comparison with ICA$_{ACC}$, the voting of our fusion algorithm was slightly similar to that, but the whole fusion algorithm also had many other different elements, such as channel selection and so on. In addition, since our fusion algorithm was not based on ICA, it avoided some problems that ICA-based methods may have, such as the incorrect selection of the separated independent maternal component, and the low-quality independent component outputs of ICA due to the failed global optimal solution search when solving the ICA problem. Although, in general, the proposed whole fusion algorithm was superior to others, it still did not work well for some groups of data in Databases 1 and 2. The 254th group of data in Database 1 was used as an example, shown in Fig. 7. Obviously, the second R-wave peak of each channel, which is labelled by red lines, was ignored by our algorithm, since there was a sudden HRV on the signal and the parameters of our algorithm was not able to adapt it.Fig. 7 Example showing our fusion algorithm’s failure. The red circles indicate the detected maternal R-wave peak locations, and the red line represents the ignored maternal R-wave peak location Future work on our algorithm we think is how to better set the parameters of the wavelet-based single channel maternal R-wave detection algorithm, such as the selection of the mother wavelet, the selection of the decomposition scale and so on. The setting of these parameters in our paper was from previous research [23]. Perhaps they can be further optimised in the future so as to improve the initial R-wave detection results of multi-channel signals. Conclusion Through the above analysis and comparison, we can conclude that though the single channel maternal R-wave detection algorithms are also useful algorithms, the detection results when directly used to detect the maternal R-wave peaks of the signal with low SNR may be not satisfactory. A lot of multi-channel maternal R wave detection algorithms based on them can greatly improve the correct rate of maternal R-wave detection in the condition of multi-channel abdominal ECG signals, especially our proposed whole fusion algorithm when compared with the other two outstanding ICA-based maternal R-wave peak detection algorithms. In summary, the proposed complete fusion algorithm based on the single channel ECG R-wave peak detection algorithm is promising for use in improving the robustness and accuracy of the maternal R-wave detections in the condition of multi-channel abdominal ECG signals. Abbreviations ECGelectrocardiogram SNRsignal-to-noise ratio PCAprincipal components analysis ICAindependent components analysis HRVheart rate variability Sesensitivity PPVpositive predictive value Authors’ contributions YQ was responsible for writing the codes and carrying out the experiments. LHX conducted the experiments and designed the algorithm. All of the authors contributed to the general concept and the writing of the paper. All authors read and approved the final manuscript. Acknowledgements The National Natural Science Foundation of China (No. 61271079) supported this paper. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Ghaffari A Atyabi SA Mollakazemi MJ Niknazar M Niknami M Soleimani A A novel noninvasive technique to recognize fetal QRS complexes from noninvasive fetal electrocardiogram signals Computing Cardiol 2013 40 293 296 2. Varanini M Tartarisco G Billeci L Macerata A Pioggia G Balocchi R An efficient unsupervised fetal QRS complex detection from abdominal maternal ECG Physiol Meas 2014 35 1607 1619 10.1088/0967-3334/35/8/1607 25069520 3. Kuzilek J Lhotska L Advanced signal processing techniques for fetal ECG analysis Computing Cardiol 2013 40 177 180 4. Rodrigues R Fetal ECG detection in abdominal recordings: a method for QRS location Computing Cardiol 2013 40 325 328 5. Di Maria C Duan W Bojarnejad M Pan F King S Zheng D Murray A Langley P An algorithm for the analysis of fetal ECGs from 4-channel non-invasive abdominal recordings Computing Cardiol 2013 40 305 308 6. Jian-Bao Z, Huan W, Qing-Shui K. Electrocardio R-wave crest real-time locating method assisted by chest impedance second order difference image: China, CN103083011 A[P], 2013.5.8, http://www.google.com/patents/CN103083011A?cl=zh&hl=zh-CN. 7. Sameni R Jutten C Shamsollahi MB Multichannel electrocardiogram decomposition using periodic component analysis IEEE Trans Biomed Eng 2008 55 8 1935 1940 10.1109/TBME.2008.919714 18632355 8. Andreotti F Riedl M Himmelsbach T Wedekind D Wessel N Stepan H Schmieder C Jank A Malberg H Zaunseder S Robust fetal ECG extraction and detection from abdominal leads Physiol Meas 2014 35 1551 1567 10.1088/0967-3334/35/8/1551 25071095 9. Behar J Oster J Clifford GD Combining and benchmarking methods of foetal ECG extraction without maternal Physiol Meas. 2014 35 1569 1589 10.1088/0967-3334/35/8/1569 25069410 10. Almeida R Gonçalves H Bernardes J Rocha AP Fetal QRS detection and heart rate estimation: a wavelet-based approach Physiol Meas 2014 35 1723 1735 10.1088/0967-3334/35/8/1723 25070210 11. Jie-Min Z Qun G Li-Ming T Tie-Bing L Hong-Xing L Xiao-Lin H Jun-Feng S Position difference regularity of corresponding R-wave peaks for maternal ECG components from different abdominal points Chin Phys B 2014 23 1 018701 10.1088/1674-1056/23/1/018701 12. Clifford GD Silva I Behar J Moody GB Non-invasive fetal ECG analysis Physiol Meas 2014 35 1521 1529 10.1088/0967-3334/35/8/1521 25071093 13. Pan J Tompkins WJ Real-time QRS detection algorithm IEEE Trans Biomed Eng 1985 BEM-32 230 236 10.1109/TBME.1985.325532 3997178 14. Nygards M Sommo L Delineation of the QRS complex using the envelope of the e.c.g Med Biol Eng Compu 1983 21 5 538 547 10.1007/BF02442378 15. Krasteva V Jekova I QRS template matching for recognition of ventricular ectopic beats Ann Biomed Eng 2007 35 12 2065 2076 10.1007/s10439-007-9368-9 17805974 16. Golbayani AH, Ghasemi M, A new mathematical based QRS detector using continuous wavelet transform. Comput Electri Eng, Elsevier Science Ltd, 2008: 81–91. 17. Kadambe S, Murray R, Boudreaux-Bartels GF, Wavelet transform-based QRS complex detector. IEEE Trans Biomed Eng. 1999;46(7). 18. Chouhan VS Mehta SS Detection of QRS complexes in 12-channel ECG using adaptive quantized threshold JCSNS Int J Comput Sci Network Secur 2008 8 1 155 163 19. Elgendi M, Mahalingam S, Jonkman M, Boer FD. A robust QRS complex detection algorithm using dynamic thresholds, Proceedings of the International Symposium on Computer Science and its Applications, Hobart, Australia 2008:153–158. 20. Ma Y, Tian Y, Ma Y, Zhang Y, Zhan K. A new real-time R-wave detection algorithm based on integral projection function, IEEE EMBS International Conference on Biomedical Engineering and Sciences 2012. 21. Legameta IR Addison PS Grubb N Clegg GR Robertson CE Fox KAA Watson JN R-wave detection using continuous wavelet modulus maxima Computing Cardiol 2003 30 565 568 22. Castillo E Morales DP Botella G García A Parrilla L Palma AJ Efficient wavelet-based ECG processing for single channel FHR extraction Digit Signal Proc 2013 23 1897 1909 10.1016/j.dsp.2013.07.010 23. Wu S Shen Y Zhou Z Lin L Zeng Y Gao X Research of fetal ECG extraction using wavelet analysis and adaptive filtering Comput Biol Med 2013 43 1622 1627 10.1016/j.compbiomed.2013.07.028 24034754 24. Martens SMM Rabotti C Mischi M Sluijter RJ A robust fetal ECG detection method for abdominal recordings Physiol Meas 2007 28 373 388 10.1088/0967-3334/28/4/004 17395993

PMC005xxxxxx/PMC5002164.txt

==== Front Ital J PediatrItal J PediatrItalian Journal of Pediatrics1824-7288BioMed Central London 28810.1186/s13052-016-0288-xEditorialProgress in pediatrics in 2015: choices in allergy, endocrinology, gastroenterology, genetics, haematology, infectious diseases, neonatology, nephrology, neurology, nutrition, oncology and pulmonology Caffarelli Carlo carlo.caffarelli@unipr.it 1Santamaria Francesca santamar@unina.it 2Di Mauro Dora dora.dimauro@hotmail.com 1Mastrorilli Carla carla.mastrorilli@icloud.com 1Mirra Virginia virginia.mirra@gmail.com 2Bernasconi Sergio sbernasconi3@gmail.com 31 Clinica Pediatrica, Department of Clinical and Experimental Medicine, Azienda Ospedaliera-Universitaria, University of Parma, Parma, Italy 2 Department of Translational Medical Sciences, Federico II University, Naples, Italy 3 Pediatrics Honorary Member University Faculty, G D’Annunzio University of Chieti-Pescara, Chieti, Italy 27 8 2016 27 8 2016 2016 42 1 7515 7 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.This review focuses key advances in different pediatric fields that were published in Italian Journal of Pediatrics and in international journals in 2015. Weaning studies continue to show promise for preventing food allergy. New diagnostic tools are available for identifying the allergic origin of allergic-like symptoms. Advances have been reported in obesity, short stature and autoimmune endocrine disorders. New molecules are offered to reduce weight gain and insulin-resistance in obese children. Regional investigations may provide suggestions for preventing short stature. Epidemiological studies have evidenced the high incidence of Graves’ disease and Hashimoto’s thyroiditis in patients with Down syndrome. Documentation of novel risk factors for celiac disease are of use to develop strategies for prevention in the population at-risk. Diagnostic criteria for non-celiac gluten sensitivity have been reported. Negative effect on nervous system development of the supernumerary X chromosome in Klinefelter syndrome has emerged. Improvements have been made in understanding rare diseases such as Rubinstein-Taybi syndrome. Eltrombopag is an effective therapy for immune trombocytopenia. Children with sickle-cell anemia are at risk for nocturnal enuresis. Invasive diseases caused by Streptococcus pyogenes are still common despite of vaccination. No difference in frequency of antibiotic prescriptions for acute otitis media between before the publication of the national guideline and after has been found. The importance of timing of iron administration in low birth weight infants, the effect of probiotics for preventing necrotising enterocolitis and perspectives for managing jaundice and cholestasis in neonates have been highlighted. New strategies have been developed to reduce the risk for relapse in nephrotic syndrome including prednisolone during upper respiratory infection. Insights into the pathophysiology of cerebral palsy, arterial ischemic stroke and acute encephalitis may drive advances in treatment. Recommendations on breastfeeding and complementary feeding have been updated. Novel treatments for rhabdomyosarcoma should be considered for paediatric patients. Control of risk factors for bronchiolitis and administration of pavilizumab for preventing respiratory syncytial virus infection may reduce hospitalization. Identification of risk factors for hospitalization in children with wheezing can improve the management of this disease. Deletions or mutations in genes encoding proteins for surfactant function may cause diffuse lung disease. Keywords AllergyEndocrinologyGastroenterologyGeneticsOnco-haematologyInfectionsNeonatologyNeurologyNutrition and pulmonologyissue-copyright-statement© The Author(s) 2016 ==== Body Background This review highlights main advances in allergy, endocrinology, gastroenterology, genetics, haematology, infectious diseases, neonatology, nephrology, neurology, nutrition, oncology and pulmonology in childhood. Papers were primarily selected from the most accessed articles published in Italian Journal of Pediatrics and in international journals in 2015. Review Allergy Food allergy is a frequent public health complaint [1–3] with potentially life threatening reactions. Several observational studies have suggested that early introduction of certain solid foods in infants may be associated with a decline of food allergy. Interventional studies were developed to more definitively ascertain these outcomes. In infants with atopic eczema, egg introduction at 4 months resulted in a nonsignificant lower rate of egg allergy at 12 months of age in comparison with a control group who went on egg-free diet [2]. Du Toit et al. [1] introduced peanut at a median age of 7.8 months, in children at risk for peanut allergy because of atopic eczema o egg sensitization. They found that peanut consumption was effective in reducing the prevalence of peanut allergy both in children who were not sensitized to peanuts (primary prevention) and in those who were sensitized (secondary prevention) in comparison with controls. Consequently, a consensus communication [4] recommended early introduction of peanut to prevent peanut allergy in children at risk. However, the term early seems to be inappropriate and it may be misinterpreted, since findings by Du Toit et al. [1] do not imply that current weaning plan should be modified [5]. Findings of interventional studies on inducing food tolerance by the oral route raise the question whether cutaneous exposure to large amounts of food allergens [6] or inhalation of food particles [7] may be also used to prevent food allergy. In the pediatric population, allergic-like symptoms are one of the most common causes for consultation in a primary care setting. Oral food challenge remains the gold standard for diagnosis of food allergy. This happens also when other concomitant precipitating factors are required to trigger an allergic reaction [8]. Regarding other investigations, Caglayan et al. [9] showed that atopy patch test for the evaluation of egg and cow’s milk allergy should not be routinely used. Fiocchi et al. [10] aimed to clarify the sensitization rate in children with common allergic symptoms. They determined serum specific IgE levels to foods and inhalants in 532 children (<15 years of age) with at least one allergy-like symptom, from 21 primary care centers in two geographic areas of Italy and Spain. They found 267 (50.2 %) atopic children with at least one positive serum specific IgE antibody. Multiple sensitizations were common: 14 were monosensitized, 37 were sensitized to 2–3 allergens and 49 % to more than 3 allergens. This point out that molecular diagnosis would be often necessary to distinguish between polysensitized children who are sensitized to genuine allergens and those sensitized to cross-reactive allergens [11]. The average number of symptoms in the atopic group was 3.3 vs 2.8 in the non-atopic group. The prevalence of sensitization to single allergens was highest for grass, ragweed and house-dust mites (19–28) and among tree allergens was for olive trees (16.5 %). Cow’s milk and egg white were the most sensitizing foods. Food allergen sensitization predominated in younger children whereas the inverse occurred with inhalant allergens. A significant positive correlation between patient age and number of sensitizations was found. This study highlights that more specific means than symptoms alone, are needed to reduce unnecessary specific IgE measurement in primary care. However, it should be also taken into account that when measurement of serum specific IgE levels and not skin prick tests to allergen extracts are performed, some allergic sensitizations may be missed. Along this line, it has been recently shown that this happens also with molecular allergens. Asero et al. [12] found that results of specific IgE to Phl p 12 (grass profilin) and skin prick tests to natural profilin (Pho d 2), purified from date palm extract gave a concordant response in 91 % of cases. In asthmatic patients, exhaled breath biomarkers [13] may be useful for identifying allergic patients. Among them fractional exhaled nitric oxide (FeNO) is significantly correlated with specific IgE antibodies levels to house dust mite in atopic asthmatic patients [14]. Furthermore, FeNO values greater that 21 ppb are associated with airway eosinophilia in corticosteroid-naïve patients [15] and are helpful in predicting loss of asthma control [16]. It has been found increased levels of oxidative stress markers such as pH [17] and hydrogen peroxide [18], that are sometimes related to decrease lung function in asthmatics. Hydrogen peroxide induces apoptosis of eosinophils and enhances the resolution of allergic inflammation [19]. Longitudinal studies are warranted to establish whether such markers may be helpful in diagnosis and control of asthma. Exercise induced bronchoconstriction (EIB) is commonly experienced by asthmatic children. Johansson et al. [20] showed that the prevalence of EIB was 19.2 % in a general population of adolescents and 42.4 %. In subjects with exercise-induced dyspnea, less than 50 % of subjects with EIB had ever had asthma. In agreement, it has been shown that EIB is a feature of not only of asthma but also of allergic rhinitis and atopic eczema [21]. Furthermore, 10.8 % of subjects with exercise-induced dyspnea presented exercise-induced laryngeal obstruction. This can be an important differential diagnosis in adolescents breathing problems during exercise [20]. Endocrinology Advances have been reported in obesity, short stature and autoimmune endocrine disorders. Childhood obesity and overweight have increased worldwide during the last decades. Lobstein et al. [22] highlight the need for obesity prevention programme in high-income countries. Comprehensive interventions rather than single-component interventions, focused on meals, classroom activities, sports, and play activities are essential to reverse the obesity trend. Interventions should be conducted at kindergarten, school, home, community. Childhood obesity prevention interventions have been demonstrated to be cost effective. Another question is how early prevention programme should start. The NOURISH randomized controlled trial [23] showed that in infancy, complementary feeding practices promoting self-regulation of intake and preference of healthy foods had effects that were sustained up to 5 years of age. However, no effect was noted for anthropometric measures or the prevalence of overweight/obesity. Obesity prevention should take into account the industrial interest in creating overweight [22]. So, regulatory steps should be taken to protect children from exposure to advertising that may influence dietary preferences. Specific nutrient standards for products are required for improving nutrition. Children and adolescents who are obese have risk factors for cardiovascular disease, prediabetes, bone and joint problems, sleep apnea, and social and psychological problems and to be obese as adults. Little is known whether there are subgroups of obese children who are at greater risk for developing cardiovascular diseases. Chang et al. [24] studied lipid profile, insulin, glucose, leptin and circulating levels of inflammatory markers, such as tumor necrosis factor-α, interleukin-6, monocyte chemoattractant protein-1, and the high-sensitive C-reactive protein, in male children, 19 obese, 10 overweight and 16 controls. Obese and overweight subjects had insulin resistance, significant higher inflammatory cytokine levels including high-sensitive C-reactive protein, PAI-1, tumor necrosis factor-α and leptin levels. Lipid profile was abnormal only in obese children. A factor analysis identified three domains that explained 74.08 % of the total variance among the obese children (factor 1:lipid, factor 2: obesity inflammation, factor 3: insulin sensitivity domains). These domains might be applied to predict the onset of cardiovascular diseases in adulthood. The treatment of childhood obesity is time consuming, problematic, and costly. Two studies addressed the role of various dietary interventions and physical activity in children. Stagi et al. [25] assessed the effect of various treatments on body mass index (BMI) and basal insulin levels in obese children and adolescents with family history of obesity and type 2 diabetes mellitus (T2DM). In a randomised study, 133 children were divided into three arms. Arm A was treated with low glycaemic index diet and new complex of polysaccharidic macromolecules (cellulose, hemicellulose, pectin, mucilages) that slows down the rate of carbohydrate and fat absorption. Arm B was treated with low glycaemic index diet diet and Arm C treated with energy restricted diets. After 1 year, BMI reduction was significant in individuals subjected to low glycaemic index diet but not to energy restricted diets diet. Furthermore, Arm A subjects have significantly reduced BMI-SDSs compared to Arm B and Arm C subjects. Insulin and glucose levels were significantly reduced in Arm A, whereas in Arm B, we observed a significant reduction in insulin but not in glucose levels. In Arm C, insulin and glucose levels did not differ from baseline. The benefit of a complex of polysaccharidic macromolecules (cellulose, hemicellulose, pectin, mucilages) added to low glycaemic index diet should be confirmed with further studies that include larger sample with longer follow-up. Verduci et al. [26] evaluated whether a 1-year intervention based on normocaloric diet and physical activity may impact the BMI status, blood lipid profile and glucose metabolism indicators in 90 obese children. At the end of intervention, children showed a reduction in BMI z-score, triglycerides and triglyceride glucose index, and an increase in HDL cholesterol. Prevalence of insulin resistance declined from 51.8 to 36.5 and prevalence of metabolic syndrome from 17.1 % to 4.9 %. Several insights on short stature have been reported. National studies on the prevalence of short stature are important. However, the prevalence of short stature may have significant regional variations in relation to race, climate, psychosocial problems and malnutrition. Wang et al. [27] assessed height in 12,009 students aged from 7 to 18 years, in Anhui province (China). Short stature was defined according to the criteria of the Chinese Medical Association as follows: height under 2SD of average height in same race, gender and age; height below the third percentile of average height in same race, gender and age. The prevalence of short stature was 3.16 % and it was similar to that reported in other regions of China. The rate was higher in rural areas than in urban areas, higher in economically backward areas than in economically developed areas. They concluded that a local government plan is needed to prevent short stature. It should carry out measures such as health education, reasonable diet, fitting exercise, adequate sleep and improving life style. Among children with idiopathic short stature effective treatment modalities are still under evaluation. Recombinant growth hormone (rGH) or insulin-like growth factor-1 (rIGF-1) can be assumed as an alternative in the treatment of these children. Recently, rGH has been approved for therapy in idiopathic short stature but reports on this treatment are limited. Siklar et al. [28] recruited 21 ISS children of unknown etiology and with IGF-1 levels below the −2 SDs of normal levels with the aim of determining short and long-term effects of rGH treatment. The height of the patients improved from −3.16 ± 0.46 SD score (SDS) to −1.9 ± 0.66 SDS. At the end of the follow-up period (5.42 ± 1.67-year), mean height SDS was −1.72. Almost 40 % of patients reached their target height. A female preponderance was noted in the responder group. The rGH treatment was safe. Response to GH therapy is variable in subjects and depends on individual factors including age, BMI, genetic and gender. The exon 3-deleted GH receptor (GHRd3) has been showed to be associated with better growth response to r-GH treatment in children with idiopathic short stature and children born small for gestational age. Valsesia et al. [29] analyzed the clinical and genetic data generated in the PREDICT study and in the PREDICT long-term follow-up study to investigate association between GHRd3 and growth response to r-hGH over 3 years in relation to severity of GH deficiency. GHRd3 carriers with higher peak GH level had better growth (C2.7 cm; C0.2 SDS) than those with low peak GH level. Similar patterns were observed for GH-dependent biomarkers. Gene expression profiles were significantly different between groups, indicating that the interaction between GH status and GHRd3 carriage can be identified at a transcriptomic level. GH treatment is frequently prescribed in Prader Willi Syndrome (PWS) at early infancy or childhood and leads to improve changes in body composition, physical activity and growth velocity. Butler et al. [30] provided new growth charts of weight, height, head circumference and BMI for non-GH treated subject with PWS aged 3 to 18 years to be used in clinical setting. Anthropometric measures were obtained from 120 PWS subjects and standardized growth charts representing 7 percentile ranges were developed along with the normative third, 50th, and 97th percentiles from national and international data. PSW-specific growth standards must be used in subjects with PWS to monitor growth patterns and nutritional status and plan individual medical care, diet intervention and physical activities. Regarding autoimmune endocrine disorders disease, only few reports have found that in childhood, Graves’ disease (GD) and Hashimoto’s thyroiditis (HT) may follow one another in the same individuals. These autoimmune thyroid disorders are more frequent in patients with Turner syndrome or Down syndrome. Aversa et al. [31] retrospectively analysed the sequential phenotypic conversion from HT to GD and the subsequent evolution of GD in a series of 12 children and adolescents with Down syndrome. All were prepubertal except one. Patients with HT diagnosis differed in thyroid hormone levels and in use of L-thyroxine. Time interval between HT diagnosis and GD onset ranged from 0.7 to 6.5 years (median 4.2). There was no correlation between duration of L-thyroxine treatment and time interval HT – GD. After metimazole withdrawal, one third of patients exhibited a spontaneous change of thyroid function, from hyperthyroidism to hypothyroidism and needed L-thyroxine treatment. These children did not significantly differ from the other ones in thyroid function tests, serum autoantibody levels at HT diagnosis or in thyrotropin receptor autoantibodies values at GD presentation. These data indicate that children with Down syndrome might manifest over time a phenotypic metamorphosis from HT to GD and they may subsequently fluctuate from hyperthyroidism to hypothyroidism. The association of autoimmune thyroid disease with autoimmune type 1 diabetes mellitus is well established [32]. Balsamo et al. [33] enrolled 152 children and adolescents at type 1 diabetes mellitus onset who had positive antithyroid antibody in 11 % of cases. They observed that thyroid function at type 1 diabetes mellitus onset is impaired mainly in association with metabolic derangement, irrespective of thyroid autoimmunity. Patients with antithyroid antibodies at type 1 diabetes mellitus onset were at higher risk of developing hypothyroidism over time. Gastroenterology Several studies on celiac disease (CD) and non-celiac gluten sensitivity have added insights into pathophysiology and treatment. Sarno et al. [34] recognized that according to family studies, 87 % of celiac disease (CD) heritability could be explained. Among genetic factors associated with CD, they described risk loci, including genetic human leukocyte antigen HLA class II region, HLA-DQ2 or HLA-DQ8 haplotypes, major histocompatibility complex class I region and non-HLA genes. They reviewed several studies that evaluated the role of environmental factors for triggering CD in the population with genetic predisposition. They noted that neither breastfeeding nor timing of gluten introduction reduced the risk for CD development. They point out that CD risk might be reduced by avoidance of high amounts of gluten in the first year of life [35]. Finally early infections, especially due to adenovirus and rotavirus, may have an impact on the risk of CD by inducing inflammatory T cell responses. Future studies on reduction of immunogenic gluten and vaccination may be helpful to develop an effective prevention of CD. Non-celiac gluten sensitivity [36, 37] is a recently documented syndrome caused by ingestion of gluten whose symptoms should be mainly distinguished from irritable bowel syndrome [38]. A consensus conference [39] has proposed new criteria for identifying non-celiac gluten sensitivity after excluding celiac disease or allergy to wheat. Non-celiac gluten sensitivity should be suspected when there is a decrease of >30 % from the baseline score of a standardized questionnaire for at least 50 % of a 6-week strict gluten-free diet. The definitive mean for the diagnosis is a positive double-blind placebo-controlled challenge. The challenge consists in a 1-week gluten diet (8 g) followed by a 1-week gluten-free diet and then a second 1-week challenge. Response to challenge should be assessed by the questionnaire whose results are positive when there is >30 % score variation between the gluten and the placebo challenge. As recognized by Authors [39], the threshold of 30 % increment in symptoms needs to be validated. Genetics Sex chromosomal aneuploidies in males are rare diseases with an overwhelming involvement of endocrinological and auxological issues. The most common of them is Klinefelter syndrome (KS) [40]. Brain magnetic resonance imaging (MRI) studies on KS have provided evidence that sex-chromosome polysomy exerts specific effects on brain development. Milani et al. [41] described a patient with a 48,XXXY/49,XXXXY mosaicism who showed some unusual neuroradiological features. The patient is a 20-months-old boy, who at birth had facial dysmorphysms, inverted nipples, and ventricular septal defect. He was first visited because of developmental delay, congenital heart defect and hypogenitalism. An MRI showed an asymmetric dysmorphic appearance of the posterior cranial fossa, that had small size, with dysmorphism of cranio-cervical junction and reduced visualization of cerebrospinal fluid spaces, that were confirmed at the computerized tomography. Brain abnormalities are frequently found in KS variants [42, 43], but also neurocognitive abnormalities, like as auditory and motor systems that often in KS are selectively affected [44]. Therefore, the supernumerary X chromosome seems to have a negative effect on white matter and central nervous system development. The patient described is the first case of a cranio-cervical junction malformation associated with 48,XXXY/ 49,XXXXY syndrome. It is important to underline that skeletal abnormalities are not exclusively related to the limbs, but also to the axial structures (such as the cervical spine), thus suggesting that the neuroradiological assessment is potentially useful in the diagnostic approach to patients with 48,XXXY and 49,XXXXY syndrome. Rubinstein-Taybi syndrome (RSTS) is an extremely rare multiple congenital anomaly/intellectual disability syndrome, with an estimated prevalence of one case per 125,000 live births. Given the complexity and rarity of RSTS, there are still numerous unanswered questions about it. In this review Milani et al. summarized the clinical features and genetic basis of RSTS, and highlighted the areas for future studies on appropriate diagnostic protocol and follow-up care for RSTS [45]. RSTS is characterized by slow development of height and weight, microcephaly, dysmorphic facial features, broad thumbs, and big toes. Over 90 % of individuals survive to adulthood, and healthcare for these patients is particularly complex, time-consuming, and often not standardized in specific guidelines. The gene most frequently involved is CREBBP [46]. Alterations in the E1A-binding protein p300 have also been detected, but many cases have been diagnosed only on a clinical basis [47]. Little is known about genotype-phenotype correlations in RSTS. Recurrent mutations maybe a key tool in addressing genotype-phenotype correlations in patients sharing the same defects and specific clinical signs, as demonstrated in two cases in a clinical cohort of 46 RSTS patients [48]. Nevertheless, a severe phenotype has been reported in RSTS patients who show large gene deletions. Novel genetic and epigenetic therapies may be promising, but there is still an urgent need to improve and personalize the standard follow up protocol [49]. In order to improve it, the authors drafted their follow-up proposal. Haematology Children with immune trombocytopenia and sickle-cell anemia (SCA) have been intensively investigated. Childhood immune thrombocytopenia is frequently a self-limiting disease, however in 13–36 % of cases it lasts more than 12 months and it is classified as chronic. Current treatments include immunosuppressors, such as corticosteroids and rituximab, immunoglobulins, and splenectomy, with high side-effects and potential short-term and long-term risks. Eltrombopag is an oral non-peptide thrombopoietin receptor agonist that binds to the transmembrane domain of the thrombopoietin receptor, leading to signal transduction through various pathways, including Jak/STAT and MAPK, which results in proliferation and differentiation of megakaryocytes and increased platelet production. Eltrombopag is an effective therapy for chronic immune thrombocytopenia, severe aplastic anemia, and chronic hepatitis C-associated thrombocytopenia in adults. The effect of eltrombopag in children with chronic immune thrombocytopenia have been investigated by two parallel randomised, multicentre, placebo controlled trials, PETIT trial [50], as a phase 2 trial, and PETIT2 study [51] as the phase 3 trial. In PETIT study, eltrombopag produced a sustained platelet response of 50x109 per L in 62 % of patients vs 32 % in placebo group, from week 1 to week 6; in PETIT2 study, in 40 % of children vs 3 % in placebo group, in the double-blind period of 13 weeks. During the open-label phase, about 80 % of patients achieved a platelet count of 50 × 109 per L or more at least once in both studies. Reduction in bleeding was also observed. Adverse events were mostly mild and serious adverse events (reversible increased alanine aminotransferase concentrations, pneumonia, aseptic meningitis, neutropenia, and anemia) were infrequent. At variance from adults, no malignancy and thrombosis occurred. Further data on cost, availability of more effective options, assessment of patient-reported outcomes and patient adherence to long-term daily medication are warranted to understand whether eltrombopag is a suitable therapeutic option for children with chronic immune thrombocytopenia in clinical practice [52]. Children with SCA are prone to sleep pathology because of smaller bladder capacity during sleep, hypostenuria, increased arousal thresholds, sleep-disordered breathing (obstructive sleep apnea syndrome). Mascarenhas et al. [53] performed a retrospective study comparing polysomnography results in 65 children with SCA versus 65 children without SCA with suspected OSAS. Mean SpO2 and minimum SpO2 were significantly lower in SCA patients. No significant difference was found between groups in efficiency, latency and percentage of sleep phases. SCA group was at higher risk for nocturnal enuresis (35.4 % vs 6.2 %, p < 0.01). Eneh et al. [54] performed a questionnaire-based, case–control study evaluating nocturnal enuresis and possible risk factors among 70 children, aged 5 to 11 years, with SCA and 70 age- and sex-matched controls. Although, there was no difference in nocturnal enuresis between the 2 groups, the prevalence of nocturnal enuresis was significantly higher in males with SCA than in male controls and in subjects whose parents had a childhood history of enuresis. They found no relationship with socio-economic status. Their findings show that reduced responsiveness to toilet training in boys and familiar predisposition may play a role in nocturnal enuresis. This study point out that school children with SCA should be frequently evaluated for nocturnal enuresis, particularly if they are males and have parental history of nocturnal enuresis in childhood. Infectious diseases Epidemiologic studies continue to document changes in the post-vaccination era. In the United States MacNeil et al. [55] found that Neisseria meningitidis remains an important cause of infectious disease in children <1 year, despite the use of pneumococcal vaccines in infants. They collected data from active, population- and laboratory-based surveillance, including cultures from blood, CSF, joint or pleural fluid for N. meningitidis, during 2006 through 2012. They estimated 113 cases annually of meningococcal disease among infants aged <1 year, for an overall incidence of 2.74 per 100 000 infants and a death rate of 6 children per year. Serogroup B was responsible for 64 % of cases, followed by serogroup Y and C responsible for 16 and 12 % of infant cases respectively. Authors suggest that future meningococcal disease vaccination strategies should be targeted on serogroup B meningococcal disease. Furthermore, they may include a maternal vaccination program to protect infants aged <1 year as they would be too young to have received the minimum 2 or 3 doses of vaccine that are needed to prevent the disease. Azzari et al. [56] studied the burden of bacteremia and invasive disease (ID) in 920 children less than 5 years old with a fever of 39 °C or greater in a prospective, multi-centre, hospital-based study. They found ID in 225 children, sepsis in 38 and non-invasive disease in 629. Twenty-two (9.8) children with ID, 2 (5.3) with sepsis and 10 (1.6 %) with a clinical diagnosis of non-invasive disease were bacteremic, having at least one positive sample, detected either by molecular assessment (PCR), culture, or both. In children with bacteremia, the most common diagnoses were community-acquired pneumonia (15/34), pleural effusion (4/34) and meningitis (4/34). It was noted a higher sensitivity of molecular versus cultural techniques for detecting bacteremia. Among bacteremic cases, Streptococcus pneumoniae was detected in 85.3 % (29/34) of cases, Haemophilus influenzae in 3 (2 non-typeable and 1 capsulated), Escherichia Coli in 1, and Neisseria meningitidis in 1. The most commonly detected Streptococcus pneumoniae serotype was 19A, detected in four cases. Community-acquired pneumonia was due to Streptococcus pneumoniae in 14 children, 3 due to serotype 3 and 3 due to serotype 14. Pleural effusion were always due to Streptococcus pneumoniae, meningitis were due to Streptococcus pneumoniae in 3 patients and to Neisseria meningitides in 1. The mean direct medical cost of bacteremic cases was 3306 euro. This study confirms that Streptococcus pyogenes is an important global pathogen, causing considerable morbidity in the paediatric population with high costs and it indicates the need for preventing pneumococcal infection by vaccination. In the past, the belief that measles, mumps and rubella vaccination may cause autism had induced some parents to not immunize their children. Using an administrative claims database associated with a large US health plan, Jain et al. [57] confirmed that measles, mumps and rubella vaccine was not associated with an increased risk of autistic spectrum disorder at any age in 95 727 children. They found that children with an older sibling with autistic spectrum disorder, more frequently develop autistic spectrum disorder, compared with those with siblings without autistic spectrum disorder (p < .001). It is reassuring that families with a child already affected by autism were not less likely to have younger children vaccinated. Numerous studies are evaluating measures to prevent the risk of infectious diseases. In low-income countries, Rotavirus disease occurs at a younger age than in high-income countries and in regions with a high burden of Rotavirus disease the success of Rotavirus vaccine is often suboptimal. Bines et al. [58] showed that coverage and efficacy of oral RV3-BB rotavirus vaccine derived from newborn with asymptomatic rotavirus infection are improved by a dose of rotavirus vaccine given at birth. In a phase 2a randomized, double-blind, three-arm, placebo-controlled trial, participants were randomized to oral RV3-BB rotavirus vaccine with the first dose given at 0–5 days after birth (neonatal schedule, 30 participants), to vaccine with the first dose given at about 8 weeks of age (infant schedule, 27 participants), or to placebo (32 participants). After randomization children received four oral doses of product at around 8 week intervals. The primary outcome was cumulative vaccine take (a serum immune response of anti-rotavirus IgA or serum neutralizing antibodies or detection of RV3-BB virus in stool after administration of vaccine or placebo) after three doses of RV3-BB. The RV3-BB rotavirus vaccine was found to be immunogenic with a positive cumulative vaccine take detected in 90 and 93 % of participants after three doses in a neonatal or an infant schedule, respectively, and well tolerated. Acute otitis media (AOM) in children can be managed without antibiotics. However, it is the most common reason for their use. Overusing antibiotic may increase serious side effects, antibiotic resistance and costs. Streptococcus salivarius 24SMB is able to produce bacteriocin-like substances with significant activity against AOM pathogens [59]. Marchisio et al. [60] investigated the efficacy of Streptococcus salivarius 24SMB nasal spray in preventing AOM in 100 otitis-prone children, aged 1–5 years, with recurrent AOM in a randomized, double-blind, placebo controlled trial. Children received intranasal S. salivarius 24SMB or placebo twice daily for 5 days each month for 3 consecutive months. When enrolled, the children were free of AOM. They were initially treated with amoxicillin-clavulanic acid for 10 days, to facilitate S. salivarius 24SMB colonization. During the study, when an AOM were diagnosed it was treated with amoxicillin plus clavulanic acid for 10 days. In the group treated with S. salivarius 24SMB, fewer children experienced any AOM in comparison with those in the placebo group (30.0 vs 14.9 %; p = 0.076). The number of children colonized by S. salivarius 24SMB who experienced any AOM was significantly lower compared with the children who were not colonized. The nasal administration of S. salivarius 24SMB was safe and well tolerated. In clinical practice, limitations of the study [61] include use of amoxicillin-clavulanic acid as first line antibiotic in AOM instead of amoxicillin and treatment of all enrolled children with an amoxicillin-clavulanic acid even if they were healthy at baseline. This can lead to an increasing antibiotic resistance and it is in contrast with a 2010 Italian pediatric guideline for the treatment of AOM [62] that recommended a “watchful waiting” approach for children with AOM to decrease the use of antibiotic. Unfortunately, watchful waiting strategy is still far from routine use and the antibiotic prescription rate continues to be high. Palma et al. [63] found no difference in the frequency of antibiotic prescriptions between before the publication of the guideline and after, both in a whole population of 4,573 children with AOM seen at the Pediatric Emergency Department (82 % versus 81 %) and in all age classes. The most frequently prescribed antibiotic was amoxicillin-clavulanic acid (51 %) and its prescription rate was similar before and after guideline publication. Neonatology Infants admitted to neonatal intensive care units (NICU), especially extremely low birth weight infants often receive an empirical antibiotic treatment, even if proven sepsis is diagnosed in a minority of cases [64]. Tzialla et al. [65] underline that unnecessary empirical administration of broad-spectrum antibiotics should be limited since it is associated with increasing multi-drug resistant bacterial infections, alteration of gut colonization which may lead to necrotizing enterocolitis (NEC) and increasing risk of invasive candidiasis. They also summarize current knowledge on the appropriate choice of antimicrobial agents and optimal duration of therapy in newborns with suspected or culture-proven sepsis. In the absence of randomized controlled trials that can definitely prove the best choice of antibiotics, an association of a penicillin or semisynthetic penicillin together with an aminoglycoside can be considered the best empirical regimen for early-onset sepsis [66]. Antistaphylococcal penicillin (oxacillin, flucloxacillin) plus an aminoglycoside should be the best option for late-onset sepsis (LOS); vancomycin should be given when infection by methicillin-resistant Staphylococcus aureus or coagulase-negative Staphylococcus is proven. Diagnosis of late-onset neonatal sepsis (LOS) can be difficult because clinical manifestations are not specific and none of the available laboratory tests can be considered an ideal marker. Poggi et al. [67] showed that presepsin (sCD14-ST) (P-SEP) may be an accurate marker for identifying LOS in preterm newborns. Presepsin is a soluble N-terminal fragment of the cluster of differentiation marker protein CD14, which is released into the circulation during monocyte activation upon the recognition of lipopolysaccharide from infectious agents. A prospective single-centre study on newborns <32 weeks’ gestational age who developed possible LOS was conducted. Blood samples for culture and count and for measuring P-SEP, procalcitonin, CRP were taken at enrolment (T0) and 1 (T1), 3 (T3), and 5 (T5) days after the first sample. P -SEP values were significantly higher in the LOS group (n = 21) than in the control group (n = 19) at any time. P-SEP achieved the best accuracy for prediction of probable sepsis at the cut-off of 885 ng/L with 94 sensitivity (95 CI 74–100) and 100 % specificity (95 % CI 84–100). Larger studies are required to confirm the role of P-SEP in early diagnosis of sepsis. NEC is a multifactorial disease associated with prematurity. Use of broad-spectrum antimicrobials and dysbiosis [68, 69], may play a role in NEC onset. Therefore, interest is growing in preventing NEC by given oral probiotics [70]. Along this line, in a meta-analysis Aceti et al. [71] selected twenty-six studies on the effect of probiotics for preventing NEC in preterm infants. The majority of studies had severe or moderate flaws. However, they found that probiotics may be effective in preterm and in very-low-birth-weight infants. Data were still insufficient in extremely-low-birth weight infants. A significant effect was found for Bifidobacteria and for probiotic mixtures. Well-designed studies are warranted to clarify the choice of the most effective strain, dose and duration of supplementation before recommending use of probiotics in clinical routine [70]. Other comorbidities of prematurity, such as delayed enteral feeding, low availability of human milk [72] and immunodeficiency [73] may favour NEC occurrence. Christensen et al. [74] found that neutropenic neonates have a significant higher risk for developing NEC. Neutropenia increased with prematurity and it was more common among small for gestational age neonates. Neonates with neutropenia were treated with either recombinant granulocyte colony-stimulating factor or intravenous immunoglobulin, but unfortunately, there was no reduction in LOS or NEC [74]. Respiratory distress syndrome at birth is one of the most frequent causes of admission to the NICU. More frequently due to pulmonary disease, and rarely to an extra pulmonary cause, such as a mediastinal mass. The first case of thymic haemorrhage was described by Ribet et al. in 1971 [75]. Gargano et al. [76] described a case of thymic haemorrhage with perinatal onset, associated with bilateral haemothorax and severe respiratory distress at birth. Thymic enlargement was evident after pleural evacuation and confirmed by chest conventional radiography, computed tomography (CT) and MRI sequences. The spontaneous resolution of the enlarged thymus suggested a thymic haemorrhage that did not require surgery. Thymic haemorrhage in the perinatal period is an exceedingly rare and severe condition that often leads to stillbirth or severe respiratory distress at birth. Spontaneous thymic haemorrhage should be considered in any neonate developing acute respiratory distress with widening of the mediastinum and pleural effusion on chest radiography. Neonatal hyperbilirubinemia is commonly due to an increase in unconjugated bilirubin that is usually benign and it resolves spontaneously. On the contrary, conjugated hyperbilirubinemia is infrequent and it should be promptly investigated. An evidence-based guideline for the management of new-borns with cholestasis [77] has been delivered this year. Early recognition of cholestasis helps to optimize the clinical management, to prevent clinical deterioration, to avoid premature, painful, expensive, and useless tests and to reduce underestimation and late referral. Many conditions can cause neonatal cholestasis and require medical or surgical treatment. The largest diagnostic group of patients includes those with biliary atresia. The clinical presentation of neonatal cholestasis may vary in relation to its aetiology. Nevertheless, the most common findings in an infant with cholestasis are prolonged jaundice, defined as jaundice lasting more than 14 days or recurring after the second week of life, acholic stools, dark yellow urine, and hepatomegaly. A practical approach for the management of the neonatal cholestasis in term and preterm infants has been provided [77]. Goetze et al. [78] recommended initial treatment and a clinically oriented overview of possible differential diagnoses. Massage therapy is a safe practice that can promote interaction between the mother and the infant, can improve weight gain, sleep patterns, growth, development, and autonomic nervous system functions, and can reduce rates of colic and infant mortality. Some clinical studies support the use of massage for reducing neonatal jaundice [79, 80], otherwise the apparent correlation has not yet been extensively examined among neonates with jaundice who are receiving phototherapy. Lin et al. [81] investigated the effects of infant massage on newborns with jaundice who are also receiving phototherapy. A total of 56 full-term neonates with jaundice, admitted for phototherapy at a regional teaching hospital, were included in the final study and randomly assigned to the control group (29 neonates; 16 males and 13 females) and the massage group (27 neonates; 11 males and 16 females). Authors showed that by day 3 of intervention, bilirubin levels were significantly lower in the massage group while stool frequency was significantly higher. Iron is an essential nutrient and plays a key role in many processes including human growth and development [82]. Low birth weight infants are particularly susceptible to develop iron deficiency anemia since they typically have small iron stores at birth and a greater need for iron due to the rapid increase in red cell mass [83, 84]. Nevertheless, it is currently unclear at what time iron supplementation in preterm very low birth weight infants should start. The report of Hong-Xing Jin et al. provides further insight into the issue of the timing of iron administration by evaluating early versus late iron supplementation in low birth weight infants through a meta-analysis of currently published studies [85]. Early supplementation ranged from as early as enteral feeding was tolerated to 3 weeks, and late supplementation ranged from 4 weeks to about 60 days. Early treatment was associated with significant smaller decrease in serum ferritin and haemoglobin levels (p < 0.001). In addition, the rate of blood transfusions was lower in early compared to late iron supplemented patients (p = 0.022). There was no difference between early and late supplementation in the number of patients who experienced NEC. Despite these results, authors recommended much caution when low birth weight infants are supplemented with iron as iron overload may have negative long-term effects on the neurodevelopment [86]. Neonates born with meconium stained amniotic fluid (MSAF) can develop feed intolerance in the first post-natal period of life. Gastric lavage is performed routinely in neonates with MSAF, but it is not a safe procedure as several complications have been described. The aim of Shah et al. [87] was to investigate the role of gastric lavage in vigorous late preterm and term newborns born with MSAF in a non-blinded randomized controlled trial conducted. No significant difference in the incidence of vomiting was found between 230 cases who received the gastric lavage, and 270 who did not. Feed intolerance was described in 51 neonates, with no significant difference in relation to the gestational age, gender, birth weight and modes of delivery. No complications of nasogastric tube insertion such as apnoea, bradycardia and local tissue trauma were observed in the gastric lavage group. These findings suggest that gastric lavage is not required in vigorous neonates born through MSAF to avoid feeding intolerance. Nevertheless, this topic is still debated. A systematic review of randomised controlled trials showed controversial evidences [88]. No association of feeding intolerance with gender or birth weight and gestation was found [89]. Nephrology Relapses and complications of nephrotic syndrome (NS), the most common manifestation of glomerular disease in childhood have been an active area of investigation. In children with steroid-sensitive nephrotic syndrome, relapses occur in most instances and 20 % to 60 % of patients develop steroid dependence. Uwaezuoke [90] reviewed triggers of relapses and their possible treatments. They documented that relapses may occur after viral infections, such as viral upper respiratory tract infections, urinary tract infections, diarrhea, or an atopic episode. This would support the hypothesis that relapses are caused by an ‘immune dysregulation’ and it is in agreement with findings of abnormalities of T cell subsets and/or T function. Rituximab a chimeric monoclonal antibody against CD 20 receptors on B cells may reduce proteinuria by inducing regulatory T cells. They also noted that relapses might be provoked by a systemic circulating factor or a primary defect of podocytes which might result in increased glomerular permeability. Corticosteroids and calcineurin inhibitors influence structure and function of podocytes. Finally, they reported that effective strategies for relapses in NS have been developed. In children, the risk for relapse has been shown to be reduced by giving prednisolone during upper respiratory infection. Zinc supplements prevent respiratory infections and reduce relapse rates in children. Ishikura et al. [91] in a long term follow-up study of 46 Japanese children with frequently relapsing NS, after an initial 2-year treatment of cyclosporine (CsA), showed that half of the patients still continued to relapse frequently or were on immunosuppressive agents at the last observation (mean age 18.7 years). They postulated that there may be an association between the experience of NS relapse during cyclosporine treatment and poorer outcome at the last observation. Fujinaga and Hirano [92] recruited 52 children with NS (median age 12.2 years). All patients had steroid-dependent NS (SDNS) or steroid-resistant NS (SRNS) and they were prescribed CsA for a median period of 39.5 months before being started on mycophenolate mofetil (MMF). During the follow-up period (median 6.1 years) after starting treatment with MMF, only 3 patients did not experience a relapse of NS, 14 were re-treated with CsA and 27 received rituximab infusions. At the last visit (median age 18.3 years), 19 patients (37 %) did not require any immunosuppressive agent, while 33 (63 %) continued to receive immunosuppressive agents. Univariate analysis revealed that the risk of persistent SDNS after MMF treatment, was positively associated with occurrence of SDNS during CsA treatment (24/33 vs. 3/19; p = 0.00019) and administration of rituximab during CsA treatment (12/33 vs. 0/19; p = 0.0017). Onset of relapse during initial CsA treatment predicts the poor clinical course of NS in the long run. The introduction of MMF following CsA treatment may not positively influence the long-term outcome in patients with SRNS and SDNS. McCaffrey et al. [93] reviewed the available evidence on complications of idiopathic NS and provided a guide for appropriate treatment. Infections should be treated promptly with broad-spectrum antibiotics in the nephrotic child. The role of antibiotic prophylaxis is still unclear, and data on the efficacy of the pneumococcal vaccine are lacking. Thromboembolic disease in NS may have potentially devastating effects, but whether primary prophylaxis is warranted is unknown. NS population are at risk for long-term cardiovascular disease and dyslipidemia, so novel management strategies, such as modifying sialylation of the circulating glycoprotein angiopoietin-like 4, have shown promising results. Neurology New insights into the pathophysiology of cerebral palsy, arterial ischemic stroke and acute encephalitis may drive advances in treatment. Cerebral palsy affects about 2 of every 1000 children. Preventing strategies are needed because of lack of effective treatments. Nelson et al. [94] analysed which factors operating before labor may increase the risk of cerebral palsy. Such factors include birth defects, defined as structural or functional defects present at birth, even if they are not recognized in the newborn period. Brain defects and congenital cardiac defects are common in children born at a gestational age of at least 35 weeks with cerebral palsy. Cerebral palsy is more likely when birth defects and fetal growth restriction are combined [95]. Cerebral palsy is also associated with low gestational age at birth, probably due to disorders that lead to malformation and delivery before term. On the other hand, malformations are more likely in at term infants with cerebral palsy. Other important risk factors for cerebral palsy are marked foetal restricted growth, especially in subjects with major birth defects, intrauterine infections (cytomegalovirus), and thromboembolism in brain, other organs and placenta. Thromboembolism may lead to malformations and perinatal stroke, the most common cause of cerebral palsy in children born at term. Finally, it has been found that a genetic component [96] is involved in many cases of cerebral palsy, especially when an older sibling is affected. A prevention programme should be based on early investigation of key prenatal risk factors for cerebral palsy, including birth defects, fetal growth restriction, and neonatal encephalopathy, that may act directly or may co-operate with other exposures or intrapartum events such asphyxia. In childhood, stroke is rare but the incidence of arterial ischemic stroke (AIS) in the neonate is similar to the incidence of large artery AIS in adults and is 17 times greater than the incidence of AIS in children. AIS is associated with a significant morbidity and mortality. In term neonates, AIS is the most common cause of cerebral palsy and is the second most common of seizures. Therefore, challenges presented by children with stroke have been the subject of considerable interest. Mechanisms, diagnosis and management of AIS in children have been summarized by Rosa et al. [97]. They documented that factors associated with AIS in childhood differ from those in adults. Focal cerebral arteriopathy caused by infections or genetically determined (Moyamoya Syndrome) seems to be the main cause. Other factors involved in the occurrence of AIS are vasculitis, arteriovenous malformation or dysplasia, hereditary coagulopathies, hematologic disorders, SCA, infarctions due to metabolic diseases, head or neck trauma, congenital and acquired heart disease. They pointed out that clinical presentation of stroke depends on the involved artery and patient’s age. Seizures and headache may develop irrespective of stroke subtype. Infants with stroke often present aspecific manifestations. Seizures and altered mental status may be the only symptoms. They emphasized that the cornerstone for diagnosing paediatric stroke is neuroimaging. In emergency settings, the diagnostic importance of non-contrast Computer Tomography has been underlined. However, MRI remains the “gold standard”. MRI Angiography should be useful to assess the involved territory. Vascular imaging of the carotid and vertebral arteries by duplex ultrasonography, Computer Tomography angiography, and catheter angiography should be considered. Laboratory tests for the investigation of diseases associated with AIS in infants and children should be performed. Understanding of the underlying etiology of AIS provides a basis for choosing the best therapeutic options. After excluding a haemorrhagic stroke, anticoagulant therapy is mandatory using aspirin or heparin. Thrombolysis may be considered. Further episodes of AIS may be prevented by Acetil-salycilic acid. Low molecular-weight heparin or warfarin are useful in children with cardioembolism, extracranial arterial dissection, cerebral venous sinus thrombosis and when aspirin treatment fails. Rosa et al. acknowledge that further studies are warranted to optimize antithrombotic treatment. Control of associated diseases, such as SCA or metabolic diseases is also recommended for secondary stroke prevention. Finally, rehabilitation can reduce long-term morbidity and improve quality of life. Encephalitis is a serious and disabling condition that can result from a large number of causes even if many cases remain undiagnosed. In a case series of acute encephalitis, Pillai et al. [98] provides insights into the relative frequency of causes. They observed that 49 (30 %) of 164 children had an infectious encephalitis, 13 (8 %) children had an infection associate encephalopathy. Pathogens detected in infectious encephalitis were enterovirus (12), mycoplasma pneumoniae (7), herpes simplex virus (5), and cytomegalovirus 2 %, in infection-associated encephalopathy, influenza virus and rotavirus. Immune-mediated/autoantibody-associated encephalitis was diagnosed in 56 (34 %) children. It was due to acute disseminated encephalomyelitis in 21 % of cases, N-methyl-D-aspartate receptor antibody encephalitis in 6 % of cases, voltage-gated potassium channel complex antibody encephalitis in 4 % of cases. In 46 (28 %) patients, the cause of encephalitis was unknown. Young children were more frequently affected with a median age at presentation of 5.5 years. The peak was during the winter season. Risk factors for poor outcomes were intensive care unit admission, diffuse restriction on MRI and status epilepticus. Nutrition This year saw the publication of the position statement of the Task Force on Breastfeeding of the Ministry of Health, the Italian Society of Neonatology, the Italian Society of Primary Pediatric Care and the Italian Society of Pediatric Gastroenterology and Nutrition [99]. They pointed out that breastfeeding is of benefit to the mother and the infant and should be promoted not only by neonatologists and pediatricians, but also by professionals, especially in maternity hospitals. They recommended exclusive breastfeeding for about 6 months of life. They emphasized that poor growth may suggest anticipating the introduction of weaning foods after 4 months of age. In agreement with those recommendation, there is evidence that starting breastfeeding within the first hour of life was associated with a reduced risk of neonatal mortality [100]. Moreover, exclusively breastfed newborns had a lower risk of mortality and infection-related deaths in the first month than partially breastfed neonates. Exclusively breastfed newborns also had a significant lower risk of sepsis, diarrhea and respiratory infections compared with those partially breastfed [100]. Khan et al. [100] concluded that substantial benefits in reducing neonatal mortality and morbidity can be achieved with effective promotion of early initiation of breastfeeding and exclusive breastfeeding during the first month of life. Giugliani et al. [101] reviewed studies that evaluated the effect of any type of intervention for promoting breastfeeding on child weight, length (or height) and weight/height (or BMI). Meta-analyses of studies reporting on mean weight, length, weight/length or BMI showed that the interventions had no impact on weight or length/height z scores and had a modest, but significant, reduction in body mass index/weight-for-height z scores, which was limited to studies from low- and high-incomes settings. An important heterogeneity among studies should be taken into account when interpreting the results. Authors stated that it is not possible to have a conclusive analysis of the impact of breastfeeding promotion interventions on child growth, although these seem to have little influence on child growth. Nevertheless, breastfeeding promotion must remain a priority, as many other benefits to the child and mother are already well established. Breast milk provides both protective and stimulatory immune signals, which may confer lower susceptibility to disorders such as allergic disease. Lodge et al. [102] systematically review the association between breastfeeding and childhood allergic disease. Longer duration of breastfeeding was associated with reduced risk of asthma for children (5–18 years), particularly in medium-/low-income countries, but this estimate had high heterogeneity and low quality. Exclusive breastfeeding for 3–4 months was associated with reduced risk of eczema ≤2 years (estimate principally from cross-sectional studies of low methodological quality). No association was found between breastfeeding and food allergy (estimate had high heterogeneity and low quality). The recommendations concerning complementary feeding from the Italian Society of Gastroenterology, Hepathology and Pediatric Nutrition and the Italian Society of Pediatric Allergy and Immunology Emilia Romagna [103] have been issued. Solid foods should be introduced between 4 and 6 months of age when the 6 months limit is not possible. An early exposure (before 4 months of age) or a delayed introduction of the main allergenic foods (after 6 months of age) has not been found to be useful to prevent allergy, even if infants were at risk of atopy. They noted that weaning while continuing breastfeeding might reduce the onset of allergies. The timing of introduction and amount of gluten has no effect on the onset of coeliac disease. The timing of introduction of complementary feeding is not related with the risk of obesity, T2DM and cardiovascular disease in later ages. They emphasized that infants should go on a nutritionally balanced diet, encouraging the daily consumption of fruit and vegetables. They recommended providing parents with the right information about daily intake of carbohydrates, lipids, proteins, liquids, fibers, vitamins and minerals. Baby led weaning and auto-weaning were discussed as a promising way for introducing complementary foods which respects the child’s self-regulatory capacity. They point out that baby led weaning is feasible when infants achieve postural stability to sit and to grasp objects. Zheng et al. [104] emphasized the risk of early weaning. They investigated the association between timing and types of complementary feeding and adiposity in children of 4–5 years of age enrolled in the Jiaxing birth cohort. Among 40510 children, 3.18 were overweight and 64.8 % were fed complementary food before 3 months of age. Early introduction of complementary foods before 3 months of age, was associated with greater BMI z-score (P-trend < .001) and higher risk of overweight (P-trend = .033). No significant association between timing of complementary feeding and obesity was observed. Fish liver oil was the major type of complementary food associated with adiposity when it was introduced before 3 months of age. Nutrition in early life is increasingly considered to be an important factor influencing later health. There is evidence that foods rich in n-3 polyunsaturated fatty acids have protective effects on oxidative damage and inflammation of the lung tissue, as in childhood wheezing and asthma. Lumia et al. [105] explored in a birth cohort study the association of food consumption in early life and the risk of atopic and non-atopic asthma. Among 182 children with asthma and 728 controls, a higher consumption of cow’s milk products was inversely associated with the risk of atopic asthma and higher consumption of breast milk and oats inversely with the risk of non-atopic asthma. Early consumption of fish was associated with a decreased risk of all asthma. This study indicated that dietary intake in early life combined with atopy history has an impact on the risk of developing asthma. A diet rich in fruits and vegetables supports healthy weight and weight loss in the context of a reduced-caloric diet. Moreover, it protects against chronic disease such as diabetes, stroke, cancer, and all-cause mortality. Herrick et al. [106] described youth fruit consumption, using data from 3129 participants, aged 2 to 19 years, in the US National Health and Nutrition Examination Survey. They found that 53 of total fruit intake was consumed as whole fruits and 34 % as fruit juices. Apples, apple juice, citrus juice, and bananas were responsible for almost half of total fruit consumption. There were differences by race and Hispanic origin in intake of citrus fruits, berries, melons, dried fruit, and citrus juices and other fruit juices. Children aged 6 to 11 years, consumed more apples and less bananas and other fruit juices compared with youth aged 2 to 5 years. Variations in the type of fruits consumed may be related to taste preference, repeated exposures to fruits, social experiences, and availability. Augusto et al. [107] underlined the importance of effective actions to promote the consumption of fruits and vegetables. They studied the relationship between frequency of fruit and vegetable consumption and nutritional deficiencies in 702 Brazilian schoolchildren (aged 4–10 years) in a cross-sectional population-based study. Only 5 % of children consumed fruits and vegetables ≥5 times/day. Overall, 6.3 of children were anaemic, 3.3 were stunted, 2.7 were obese and 33 % had multiple nutritional deficiencies. Consumption of vegetables ≤3 times/month and of fruits ≤3 times/week was associated with lower plasma concentrations of carotenoids and vitamin E. Nutritional deficiencies, including anemia, vitamin E insufficiency, vitamin D insufficiency and stunting, were more common in non-consumers of fruits and vegetables than in usual consumers (vegetables ≥1 time/week and fruits ≥4 times/week). Trends in childhood morbidity and mortality are strongly correlated with childhood infections and undernutrition. Most of the morbidity and mortality from diarrhea and pneumonia occur in children under 2 years of age, linking it closely to early nutrition and growth. Because of the complex interrelation between infections and nutrition conditions, Salam et al. [108] addressed several strategies to reduce major childhood infections and improve nutrition and growth and implications. These interventions include the following: exclusive breastfeeding up to 6 months of age, promotion of complementary feeding, pneumococcal, rotavirus and Haemophilus influenzae type b vaccine, community-based management of neonatal infections, case management of pneumonia infections, use of oral rehydration salt and zinc in diarrhea, vitamin A supplementation, antibiotics for dysentery, and Water Sanitation and Hygiene strategies. Utilization of preventive interventions will make an impact by reducing malnutrition and under-5 morbidity and mortality. White et al. [109] measured the prevalence of malnutrition, obesity and nutritional risk in 832 pediatric inpatients among multiple hospitals throughout Australia in a single day. The prevalence rates of malnourished, wasted, stunted, overweight and obese pediatric patients were 15, 13.8, 11.9, 8.8 and 9.9 %, respectively. Aboriginal and Torres Strait Islander patients were more likely to have lower height-for-age z-scores (P <0.01). However, BMI and weight-for-age z-scores were not significantly different. Children who were younger, from regional hospitals or with a primary diagnosis of cardiac disease or cystic fibrosis had significantly lower anthropometric z-scores (P = 0.05). Forty-four % of patients were identified as at high nutritional risk and requiring further nutritional assessment. They concluded that malnutrition and nutritional risk of Australian pediatric inpatients on a given day was much higher when compared with the healthy population. In contrast, the proportion of overweight and obese patients was less. Oncology Survival after childhood cancer is greatly enhanced in last decades. PanCare, an interdisciplinary Network of experts in the field, has been created to identify best practices of long-term care and address challenges in survivors, especially related to late effects of treatment [110]. Rhabdomyosarcoma (RMS), the most frequent soft tissue sarcoma in childhood, can be distinguished in embryonal RMS which occurs in 75 % of cases, with a higher incidence at age of 0–4 years and alveolar RMS found at every age with a worse prognosis. RMS has also been classified basing on molecular prognostic markers. PAX–FOXO1 fusion gene status was identified and compared to alveolar histology, whereas fusion gene–negative alveolar RMS patients were clinically similar to ERMS patients. Attempts to identify other prognostic gene signatures an independent 5-gene signature, MG5, showed a significant association with overall survival in the fusion gene–negative alveolar RMS and embryonal RMS patients. Hingorani et al. [111] studied whether expression of the MG5 metagene, measured using a technical platform that can be applied to routine pathology material, would correlate with outcome in a cohort of children with fusion gene–negative alveolar RMS. MG5 signature score showed a significant correlation with overall and failure-free survival. The ability of MG5 signature was confirmed to identify different risk groups within fusion gene–negative alveolar RMS from a different patient cohort composed of those with intermediate-risk disease. Ma et al. [112] explored the treatment results of childhood RMS and identified prognostic factors in a multicenter Chinese retrospective study. Medical records of 161 patients (median age 51 months) were analyzed. The genitourinary system was the most common primary site of tumor (43.5 %). The histological findings were: 80.7 embryonal, 11.9 alveolar and 3.1 % botryoid type. The 10-year event free survival rate was 53.4 ± 5.1 %, overall survival was 65.3 ± 6.3 %. Among variables, age (p = 0.028) and disease groups (p = 0.000) were significantly associated with overall survival. In conclusion, epidemiological characteristics of patients of this study were similar to worldwide data. Sangkhathat et al. [113] described current management of pediatric soft tissue sarcomas, based on multidisciplinary approach including surgery, chemotherapy and radiation therapy. Decision making in management protocol for each patient is determined on the risk induced by various clinical and pathological parameters. For cases of small resectable tumors in a favorable site, surgery provides the best choice of local control. Radiation therapy is added when surgery leaves residual disease or there is evidence of regional spread. Chemotherapy reduces risks of relapse and improves overall survival. The standard therapy for pediatric RMS is a multi-agent chemotherapy and surgery resection. This line aims to reduce the aggressiveness of the required surgery and frequently helps preserve organ function. Radiotherapy is often added. Proton beam therapy (PBT) is a type of radiotherapy with excellent dose localization. Takizawa et al. [114] showed that PBT can be useful in cases who cannot receive definitive photon radiotherapy. They performed, safely and effectively, fPBT in a 1-year-old girl with alveolar RMS with liver and cardiac invasion at the dose of 41.4 GyE. In this case, chemotherapy alone was not effective; surgery was not possible for broad tumour invasion of the surrounding diaphragm; photon radiotherapy could not be performed because the liver could not tolerate the treatment dose. In a retrospective study, Fukushima et al. [115] analyzed five children with genitourinary/pelvic rhabdomyosarcoma (GU/P-RMS) who had undergone multimodal therapy combined with PBT in a Japanese institution. All received neo-adjuvant chemotherapy and 3 underwent chemotherapy during PBT (Group Cx). All patients of Group Cx developed leukocytopenia but survived by their last hospital visit (after 36 months). They concluded that PBT was well tolerated and could be a plausible choice instead of photon therapy for GU/P-RMS children. Pulmonology Deletions of or mutations in genes encoding proteins important in surfactant production and function (SP-B, SP-C, and ABCA3), surfactant catabolism (GM-CSF receptor), or transcription factors important for surfactant production (TTF1) or lung development (Fox F1) may cause pediatric diffuse lung disease [116]. This is an heterogeneous group of uncommon disorders with impaired gas exchange and diffuse infiltrates at chest imaging [117]. Montella et al. [118] report on a 3-years-old boy with dry cough, progressive hypoxemia, dyspnea and bilateral ground glass opacities at chest high-resolution computed tomography (HRCT), in whom lung histology strongly suggested a surfactant disorder, but no variants conclusively classified as pathogenic in genes encoding surfactant proteins or transcription factors important for surfactant production were found. Pulse intravenous high-dose methylprednisolone, oral hydroxychloroquine and azithromycin were started as treatment, leading to gradual weaning from oxygen and to improvement of cough and dyspnea. Six months later, steroids and oxygen were stopped and a substantial improvement of exercise tolerance versus pre-treatment test was observed. Currently, the subject is 6 years-old, adherence to treatment is satisfactory, with no side effects. Repeated daily oxygen transcutaneous saturation showed values not lower than 98 % at room air, with median overnight values of 96 % (min 74 %, max 99 %), while chest HRCT was unmodified. The authors confirm that in children with suspected surfactant dysfunction, genetic testing is strongly recommended because it can provide a definitive diagnosis but when causative mutations are not found, lung biopsy with consistent histology may help physicians to address the definitive diagnosis. Respiratory Syncytial Virus (RSV) is by far the pathogen responsible for most of the airway infections in children under age 2 years. RSV bronchiolitis is a leading cause of severe respiratory disease that requires hospitalization and, in some cases, intensive care [119]. Quinonez et al. [120] pointed out important issues in the management of bronchiolitis, Pulse oximetry potentially overdiagnoses hypoxemia and increase hospital admission and length of stay [121]. Another point is that infants with bronchiolitis should be provided with antipyretics, intravenous hydration, oxygen when necessary, and salbutamol should be given to assess its efficacy. Finally, continuous treatment with epinephrine or saline may increase length of hospital stay and oxygen requirement when compared with “on-demand” treatments [122]. Infection can be consistently prevented by the application of environmental, hygienic and sanitary measures to minimize the viral diffusion [123]. Recommendations drafted by the Italian Society of Neonatology concerning prophylaxis of RSV infection with palivizumab, a humanized monoclonal antibody that blocks the viral replication during the epidemic season to children at high risk have been revised in 2015 through the update of epidemiological data [124]. Lanari et al. [125] evaluated the risk factors for hospitalization due to bronchiolitis during the first year of life in children born at different gestational ages in Italy. The study population consisted of 2314 newborns, of which 2210 (95.5 %) had a 1 year follow-up and were included in the analysis, while the remaining 120 cases (5.4 %) were hospitalized during the first year of life for bronchiolitis. Children born at 33–34 weeks gestational age had a higher hospitalization rate compared to other groups. Male gender, prenatal exposure to maternal smoking, neonatal surfactant therapy, having siblings aged less than 10 years, living in crowded conditions and being exposed to epidemic season during the first 3 months of life were identified as risk factors for high rate of hospitalization for bronchiolitis. This confirms that individual characteristics and exposure to environmental factors play an important role in determining the risk of severe infection and hospitalization, independently from preterm birth. The analysis of the weight of each risk factor allows defining accurately the risk for bronchiolitis hospitalization for any infant during the first year of life and letting authors to conclude that indications of palivizumab should be individually “tailored”. Conclusions This was an exciting year for pediatric research. This review provides novel information with regard to pathophysiology, clinical aspects, prevention and management of common diseases in childhood. Several Position statement of scientific societies on different issues including nutrition and hyperbilirubinemia in newborns were made available. New data on rare diseases, such as disorder of surfactant metabolism, and 48,XXXY/49,XXXXY mosaicism may improve their understanding. Overall, these advances can improve diagnosis and treatment of pediatric diseases and offer comprehensions on how we can develop patient care in the future. Abbreviations AISArterial Ischemic Stroke AOMAcute Otitis Media BMIBody Mass Index CDCeliac Disease CsACyclosporine CTComputed Tomography EIBExercise Induced Bronchoconstriction FeNOFractional Exhaled Nitric Oxide GDGraves’ Disease HRCTHigh-Resolution Computed Tomography HTHashimoto’s Thyroiditis IDInvasive Disease KSKlinefelter Syndrome LOSLate-Onset Neonatal Sepsis MMFMycophenolate Mofetil MRIMagnetic Resonance Imaging MSAFMeconium Stained Amniotic Fluid NECNecrotizing Enterocolitis NICUNeonatal Intensive Care Units NSNephrotic Syndrome PBTProton Beam Therapy P-SEPPresepsin PWSPrader Willi Syndrome rGHRecombinant Growth Hormone rIGF-1Insulin-Like Growth Factor-1 RMSRhabdomyosarcoma RSTSRubinstein-Taybi Syndrome RSVRespiratory Syncytial Virus SCASickle-Cell Anemia T2DMType 2 Diabetes Mellitus. Acknowledgements Not applicable. Funding No funding. Availability of data and materials The dataset supporting the conclusions of this article is included within the article. Authors’ contributions CC conceived the study, participated in its design, carried out the literature research and helped draft the manuscript. FSa conceived the study, participated in the design of the study, carried out the literature research and helped draft the manuscript. DD, CM, VM carried out the literature research and helped draft the manuscript. SB conceived the study, and participated in its design and coordination and helped draft 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. Du Toit G Roberts G Sayre PH Bahnson HT Radulovic S Santos AF Brough HA Phippard D Basting M Feeney M Turcanu V Sever ML Gomez Lorenzo M Plaut M Lack G Team LS Randomized trial of peanut consumption in infants at risk for peanut allergy N Engl J Med 2015 372 803 813 10.1056/NEJMoa1414850 25705822 2. Palmer DJ Metcalfe J Makrides M Gold MS Quinn P West CE Loh R Prescott SL Early regular egg exposure in infants with eczema: a randomized controlled trial J Allergy Clin Immunol 2013 132 387 392 10.1016/j.jaci.2013.05.002 23810152 3. Caffarelli C Coscia A Ridolo E Povesi Dascola C Gelmett C Raggi V Volta E Vanell M Dall’Aglio PP Parents’ estimate of food allergy prevalence and management in Italian school-aged children Pediatr Int 2011 53 505 510 10.1111/j.1442-200X.2010.03294.x 21105963 4. Fleischer DM Sicherer S Greenhawt M Campbell D Chan E Muraro A Halken S Katz Y Ebisawa M Eichenfield L Sampson H Lack G Du Toit G Roberts G Bahnson H Feeney M Hourihane J Spergel J Young M As’aad A Allen K Prescott S Kapur S Saito H Agache I Akdis CA Arshad H Beyer K Dubois A Eigenmann P Fernandez-Rivas M Grimshaw K Hoffman-Sommergruber K Host A Lau S O’Mahony L Mills C Papadopoulos N Venter C Agmon-Levin N Kessel A Antaya R Drolet B Rosenwasser L American Academy of Allergy A, Immunology, American Academy of Pediatrics ACoA, Asthma, Immunology ASoCI, Allergy, Canadian Society of A, Clinical I, European Academy of A, Clinical I, Israel Association of A, Clinical I, Japanese Society for A, Society for Pediatric D, World Allergy O Consensus communication on early peanut introduction and the prevention of peanut allergy in high-risk infants J Allergy Clin Immunol 2015 136 258 261 10.1016/j.jaci.2015.06.001 26100082 5. Ricci G Caffarelli C Early or not delayed complementary feeding?: This is the question J Allergy Clin Immunol 2016 137 334 335 10.1016/j.jaci.2015.09.038 26586038 6. Mondoulet L Dioszeghy V Thebault C Benhamou PH Dupont C Epicutaneous immunotherapy for food allergy as a novel pathway for oral tolerance induction Immunotherapy 2015 7 1293 1305 10.2217/imt.15.86 26584421 7. Leonardi S Pecoraro R Filippelli M Miraglia del Giudice M Marseglia G Salpietro C Arrigo T Stringari G Rico S La Rosa M Caffarelli C Allergic reactions to foods by inhalation in children Allergy Asthma Proc 2014 35 288 294 10.2500/aap.2014.35.3755 24992548 8. Povesi Dascola C Caffarelli C Exercise-induced anaphylaxis: a clinical view Ital J Pediatr 2012 38 43 10.1186/1824-7288-38-43 22980517 9. Caglayan Sozmen S Povesi Dascola C Gioia E Mastrorilli C Rizzuti L Caffarelli C Diagnostic accuracy of patch test in children with food allergy Pediatr Allergy Immunol 2015 26 416 422 10.1111/pai.12377 25808316 10. Fiocchi A Pecora V Petersson CJ Dahdah L Borres MP Amengual MJ Huss-Marp J Mazzina O Di Girolamo F Sensitization pattern to inhalant and food allergens in symptomatic children at first evaluation Ital J Pediatr 2015 41 96 10.1186/s13052-015-0204-9 26643320 11. Mastrorilli C Tripodi S Caffarelli C Perna S Di Rienzo-Businco A Sfika I Asero R Dondi A Bianchi A Povesi Dascola C Ricci G Cipriani F Maiello N Miraglia Del Giudice M Frediani T Frediani S Macri F Pistoletti C Dello Iacono I Patria MF Varin E Peroni D Comberiati P Chini L Moschese V Lucarelli S Bernardini R Pingitore G Pelosi U Olcese R Moretti M Cirisano A Faggian D Travaglini A Plebani M Verga MC Calvani M Giordani P Matricardi PM Italian Pediatric Allergy N Endotypes of pollen-food syndrome in children with seasonal allergic rhinoconjunctivitis: a molecular classification Allergy 2016 71 1181 1191 10.1111/all.12888 26999633 12. Asero R Tripodi S Dondi A Di Rienzo Businco A Sfika I Bianchi A Candelotti P Caffarelli C Povesi Dascola C Ricci G Calamelli E Maiello N Miraglia Del Giudice M Frediani T Frediani S Macri F Moretti M Dello Iacono I Patria MF Varin E Peroni D Comberiati P Chini L Moschese V Lucarelli S Bernardini R Pingitore G Pelosi U Tosca M Cirisano A Faggian D Plebani M Verga C Matricardi PM Italian Pediatric Allergy N Prevalence and clinical relevance of IgE sensitization to profilin in childhood: a multicenter study Int Arch Allergy Immunol 2015 168 25 31 10.1159/000441222 26528861 13. Corradi M Zinelli C Caffarelli C Exhaled breath biomarkers in asthmatic children Inflamm Allergy Drug Targets 2007 6 150 159 10.2174/187152807781696437 17897051 14. Lee YK Yang S Park J Kim H Hahn YS House dust mite-specific immunoglobulin E and longitudinal exhaled nitric oxide measurements in children with atopic asthma Korean J Pediatr 2015 58 89 95 10.3345/kjp.2015.58.3.89 25861331 15. Alvarez-Puebla MJ Olaguibel Rivera JM Almudevar E Echegoyen AA de Esteban CB Cambra K Cutoff point for exhaled nitric oxide corresponding to 3 % sputum eosinophils J Investig Allergol Clin Immunol 2015 25 107 111 25997303 16. Yang S Park J Lee YK Kim H Hahn YS Association of longitudinal fractional exhaled nitric oxide measurements with asthma control in atopic children Respir Med 2015 109 572 579 10.1016/j.rmed.2015.03.003 25840483 17. Caffarelli C Dascola CP Peroni D Rico S Stringari G Varini M Folesani G Corradi M Airway acidification in childhood asthma exacerbations Allergy Asthma Proc 2014 35 51 56 10.2500/aap.2014.35.3740 24801459 18. Caffarelli C Calcinai E Rinaldi L Povesi Dascola C Terracciano L Corradi M Hydrogen peroxide in exhaled breath condensate in asthmatic children during acute exacerbation and after treatment Respiration 2012 84 291 298 10.1159/000341969 23018317 19. Reis AC Alessandri AL Athayde RM Perez DA Vago JP Avila TV Ferreira TP de Arantes AC Coutinho Dde S Rachid MA Sousa LP Martins MA Menezes GB Rossi AG Teixeira MM Pinho V Induction of eosinophil apoptosis by hydrogen peroxide promotes the resolution of allergic inflammation Cell Death Dis 2015 6 e1632 10.1038/cddis.2014.580 25675292 20. Johansson H Norlander K Berglund L Janson C Malinovschi A Nordvall L Nordang L Emtner M Prevalence of exercise-induced bronchoconstriction and exercise-induced laryngeal obstruction in a general adolescent population Thorax 2015 70 57 63 10.1136/thoraxjnl-2014-205738 25380758 21. Caffarelli C Bacchini PL Gruppi L Bernasconi S Exercise-induced bronchoconstriction in children with atopic eczema Pediatr Allergy Immunol 2005 16 655 661 10.1111/j.1399-3038.2005.00327.x 16343087 22. Lobstein T Jackson-Leach R Moodie ML Hall KD Gortmaker SL Swinburn BA James WP Wang Y McPherson K Child and adolescent obesity: part of a bigger picture Lancet 2015 385 2510 2520 10.1016/S0140-6736(14)61746-3 25703114 23. Daniels LA Mallan KM Nicholson JM Thorpe K Nambiar S Mauch CE Magarey A An early feeding practices intervention for obesity prevention Pediatrics 2015 136 e40 49 10.1542/peds.2014-4108 26055848 24. Chang CJ Jian DY Lin MW Zhao JZ Ho LT Juan CC Evidence in obese children: contribution of hyperlipidemia, obesity-inflammation, and insulin sensitivity PLoS One 2015 10 e0125935 10.1371/journal.pone.0125935 26011530 25. Stagi S Lapi E Seminara S Pelosi P Del Greco P Capirchio L Strano M Giglio S Chiarelli F de Martino M Policaptil Gel retard significantly reduces body mass index and hyperinsulinism and may decrease the risk of type 2 diabetes mellitus (T2DM) in obese children and adolescents with family history of obesity and T2DM Ital J Pediatr 2015 41 10 10.1186/s13052-015-0109-7 25774705 26. Verduci E Lassandro C Giacchero R Miniello VL Banderali G Radaelli G Change in metabolic profile after 1-year nutritional-behavioral intervention in obese children Nutrients 2015 7 10089 10099 10.3390/nu7125520 26633492 27. Wang Q Liu DY Yang LQ Liu Y Chen XJ The epidemic characteristics of short stature in school students Ital J Pediatr 2015 41 99 10.1186/s13052-015-0207-6 26714873 28. Siklar Z Kocaay P Camtosun E Isakoca M Hacihamdioglu B Savas Erdeve S Berberoglu M The effect of recombinant growth hormone treatment in children with idiopathic short stature and Low insulin-like growth factor-1 levels J Clin Res Pediatr Endocrinol 2015 7 301 306 10.4274/jcrpe.2111 26777041 29. Valsesia A Chatelain P Stevens A Peterkova VA Belgorosky A Maghnie M Antoniazzi F Koledova E Wojcik J Farmer P Destenaves B Clayton P Group PI GH deficiency status combined with GH receptor polymorphism affects response to GH in children Eur J Endocrinol 2015 173 777 789 10.1530/EJE-15-0474 26340968 30. Butler MG Lee J Manzardo AM Gold JA Miller JL Kimonis V Driscoll DJ Growth charts for non-growth hormone treated Prader-Willi syndrome Pediatrics 2015 135 e126 135 10.1542/peds.2014-1711 25489013 31. Aversa T Valenzise M Salerno M Corrias A Iughetti L Radetti G De Luca F Wasniewska M Metamorphic thyroid autoimmunity in down syndrome: from Hashimoto’s thyroiditis to Graves’ disease and beyond Ital J Pediatr 2015 41 87 10.1186/s13052-015-0197-4 26558364 32. Shun CB Donaghue KC Phelan H Twigg SM Craig ME Thyroid autoimmunity in type 1 diabetes: systematic review and meta-analysis Diabet Med 2014 31 126 135 10.1111/dme.12318 24103027 33. Balsamo C Zucchini S Maltoni G Rollo A Martini AL Mazzanti L Pession A Cassio A Relationships between thyroid function and autoimmunity with metabolic derangement at the onset of type 1 diabetes: a cross-sectional and longitudinal study J Endocrinol Invest 2015 38 701 707 10.1007/s40618-015-0248-0 25722223 34. Sarno M Discepolo V Troncone R Auricchio R Risk factors for celiac disease Ital J Pediatr 2015 41 57 10.1186/s13052-015-0166-y 26268374 35. Ivarsson A Myleus A Norstrom F van der Pals M Rosen A Hogberg L Danielsson L Halvarsson B Hammarroth S Hernell O Karlsson E Stenhammar L Webb C Sandstrom O Carlsson A Prevalence of childhood celiac disease and changes in infant feeding Pediatrics 2013 131 e687 694 10.1542/peds.2012-1015 23420914 36. Guandalini S Polanco I Nonceliac gluten sensitivity or wheat intolerance syndrome? J Pediatr 2015 166 805 811 10.1016/j.jpeds.2014.12.039 25662287 37. Fasano A Sapone A Zevallos V Schuppan D Nonceliac gluten sensitivity Gastroenterology 2015 148 1195 1204 10.1053/j.gastro.2014.12.049 25583468 38. Caffarelli C Coscia A Baldi F Borghi A Capra L Cazzato S Migliozzi L Pecorari L Valenti A Cavagni G Characterization of irritable bowel syndrome and constipation in children with allergic diseases Eur J Pediatr 2007 166 1245 1252 10.1007/s00431-006-0410-y 17345097 39. Catassi C Elli L Bonaz B Bouma G Carroccio A Castillejo G Cellier C Cristofori F de Magistris L Dolinsek J Dieterich W Francavilla R Hadjivassiliou M Holtmeier W Korner U Leffler DA Lundin KE Mazzarella G Mulder CJ Pellegrini N Rostami K Sanders D Skodje GI Schuppan D Ullrich R Volta U Williams M Zevallos VF Zopf Y Fasano A Diagnosis of Non-Celiac Gluten Sensitivity (NCGS): The Salerno Experts’ Criteria Nutrients 2015 7 4966 4977 10.3390/nu7064966 26096570 40. Tartaglia N Ayari N Howell S D’Epagnier C Zeitler P 48, XXYY, 48, XXXY and 49, XXXXY syndromes: not just variants of Klinefelter syndrome Acta Paediatr 2011 100 851 860 10.1111/j.1651-2227.2011.02235.x 21342258 41. Milani D Bonarrigo F Avignone S Triulzi F Esposito S 48, XXXY/49, XXXXY mosaic: new neuroradiological features in an ultra-rare syndrome Ital J Pediatr 2015 41 50 10.1186/s13052-015-0156-0 26168786 42. Hoffman TL Vossough A Ficicioglu C Visootsak J Brain magnetic resonance imaging findings in 49, XXXXY syndrome Pediatr Neurol 2008 38 450 453 10.1016/j.pediatrneurol.2008.03.004 18486832 43. Tabarki B Shafi SA Adwani NA Shahwan SA Further magnetic resonance imaging (MRI) brain delineation of 49, XXXXY syndrome J Child Neurol 2012 27 650 653 10.1177/0883073811424797 22156792 44. Wallentin M Skakkebaek A Bojesen A Fedder J Laurberg P Ostergaard JR Hertz JM Pedersen AD Gravholt CH Klinefelter syndrome has increased brain responses to auditory stimuli and motor output, but not to visual stimuli or Stroop adaptation Neuroimage Clin 2016 11 239 251 10.1016/j.nicl.2016.02.002 26958463 45. Milani D Manzoni FM Pezzani L Ajmone P Gervasini C Menni F Esposito S Rubinstein-Taybi syndrome: clinical features, genetic basis, diagnosis, and management Ital J Pediatr 2015 41 4 10.1186/s13052-015-0110-1 25599811 46. Bentivegna A Milani D Gervasini C Castronovo P Mottadelli F Manzini S Colapietro P Giordano L Atzeri F Divizia MT Uzielli ML Neri G Bedeschi MF Faravelli F Selicorni A Larizza L Rubinstein-Taybi Syndrome: spectrum of CREBBP mutations in Italian patients BMC Med Genet 2006 7 77 10.1186/1471-2350-7-77 17052327 47. Bartholdi D Roelfsema JH Papadia F Breuning MH Niedrist D Hennekam RC Schinzel A Peters DJ Genetic heterogeneity in Rubinstein-Taybi syndrome: delineation of the phenotype of the first patients carrying mutations in EP300 J Med Genet 2007 44 327 333 10.1136/jmg.2006.046698 17220215 48. Spena S Milani D Rusconi D Negri G Colapietro P Elcioglu N Bedeschi F Pilotta A Spaccini L Ficcadenti A Magnani C Scarano G Selicorni A Larizza L Gervasini C Insights into genotype-phenotype correlations from CREBBP point mutation screening in a cohort of 46 Rubinstein-Taybi syndrome patients Clin Genet 2015 88 431 440 10.1111/cge.12537 25388907 49. Solomon BD Bodian DL Khromykh A Mora GG Lanpher BC Iyer RK Baveja R Vockley JG Niederhuber JE Expanding the phenotypic spectrum in EP300-related Rubinstein-Taybi syndrome Am J Med Genet A 2015 167A 1111 1116 10.1002/ajmg.a.36883 25712426 50. Bussel JB de Miguel PG Despotovic JM Grainger JD Sevilla J Blanchette VS Krishnamurti L Connor P David M Boayue KB Matthews DC Lambert MP Marcello LM Iyengar M Chan GW Chagin KD Theodore D Bailey CK Bakshi KK Eltrombopag for the treatment of children with persistent and chronic immune thrombocytopenia (PETIT): a randomised, multicentre, placebo-controlled study Lancet Haematol 2015 2 e315 325 10.1016/S2352-3026(15)00114-3 26688484 51. Grainger JD Locatelli F Chotsampancharoen T Donyush E Pongtanakul B Komvilaisak P Sosothikul D Drelichman G Sirachainan N Holzhauer S Lebedev V Lemons R Pospisilova D Ramenghi U Bussel JB Bakshi KK Iyengar M Chan GW Chagin KD Theodore D Marcello LM Bailey CK Eltrombopag for children with chronic immune thrombocytopenia (PETIT2): a randomised, multicentre, placebo-controlled trial Lancet 2015 386 1649 1658 10.1016/S0140-6736(15)61107-2 26231455 52. Neunert CE Grace RF Thrombopoietin-receptor agonists in children with immune thrombocytopenia Lancet 2015 386 1606 1609 10.1016/S0140-6736(15)61223-5 26231462 53. Mascarenhas MI Loureiro HC Ferreira T Dias A Sleep pathology characterization in sickle cell disease: case–control study Pediatr Pulmonol 2015 50 396 401 10.1002/ppul.23074 25045078 54. Eneh CI Okafor HU Ikefuna AN Uwaezuoke SN Nocturnal enuresis: prevalence and risk factors among school-aged children with sickle-cell anaemia in a South-east Nigerian city Ital J Pediatr 2015 41 66 10.1186/s13052-015-0176-9 26416787 55. MacNeil JR Bennett N Farley MM Harrison LH Lynfield R Nichols M Petit S Reingold A Schaffner W Thomas A Pondo T Mayer LW Clark TA Cohn AC Epidemiology of infant meningococcal disease in the United States, 2006–2012 Pediatrics 2015 135 e305 311 10.1542/peds.2014-2035 25583921 56. Azzari C Moriondo M Di Pietro P Di Bari C Resti M Mannelli F Esposito S Castelli-Gattinara G Campa A de Benedictis FM Bona G Comarella L Holl K Marchetti F The burden of bacteremia and invasive diseases in children aged less than 5 years with fever in Italy Ital J Pediatr 2015 41 92 10.1186/s13052-015-0189-4 26589787 57. Jain A Marshall J Buikema A Bancroft T Kelly JP Newschaffer CJ Autism occurrence by MMR vaccine status among US children with older siblings with and without autism JAMA 2015 313 1534 1540 10.1001/jama.2015.3077 25898051 58. Bines JE Danchin M Jackson P Handley A Watts E Lee KJ West A Cowley D Chen MY Barnes GL Justice F Buttery JP Carlin JB Bishop RF Taylor B Kirkwood CD Program RVRV Safety and immunogenicity of RV3-BB human neonatal rotavirus vaccine administered at birth or in infancy: a randomised, double-blind, placebo-controlled trial Lancet Infect Dis 2015 15 1389 1397 10.1016/S1473-3099(15)00227-3 26318715 59. Santagati M Scillato M Patane F Aiello C Stefani S Bacteriocin-producing oral streptococci and inhibition of respiratory pathogens FEMS Immunol Med Microbiol 2012 65 23 31 10.1111/j.1574-695X.2012.00928.x 22243526 60. Marchisio P Santagati M Scillato M Baggi E Fattizzo M Rosazza C Stefani S Esposito S Principi N Streptococcus salivarius 24SMB administered by nasal spray for the prevention of acute otitis media in otitis-prone children Eur J Clin Microbiol Infect Dis 2015 34 2377 2383 10.1007/s10096-015-2491-x 26385346 61. Di Mario S Basevi V D’Amico R Gagliotti C Gangemi M Marchetti F Moro ML Tamburlini G Streptococcus salivarius by nasal spray for recurrent otitis: how good is the evidence? Eur J Clin Microbiol Infect Dis 2016 35 1215 1216 10.1007/s10096-016-2648-2 27180246 62. Marchisio P Bellussi L Di Mauro G Doria M Felisati G Longhi R Novelli A Speciale A Mansi N Principi N Acute otitis media: from diagnosis to prevention. Summary of the Italian guideline Int J Pediatr Otorhinolaryngol 2010 74 1209 1216 10.1016/j.ijporl.2010.08.016 20843561 63. Palma S Rosafio C Del Giovane C Patianna VD Lucaccioni L Genovese E Bertolani P Iughetti L The impact of the Italian guidelines on antibiotic prescription practices for acute otitis media in a paediatric emergency setting Ital J Pediatr 2015 41 37 10.1186/s13052-015-0144-4 25948496 64. Stoll BJ Hansen N Fanaroff AA Wright LL Carlo WA Ehrenkranz RA Lemons JA Donovan EF Stark AR Tyson JE Oh W Bauer CR Korones SB Shankaran S Laptook AR Stevenson DK Papile LA Poole WK Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network Pediatrics 2002 110 285 291 10.1542/peds.110.2.285 12165580 65. Tzialla C Borghesi A Serra G Stronati M Corsello G Antimicrobial therapy in neonatal intensive care unit Ital J Pediatr 2015 41 27 10.1186/s13052-015-0117-7 25887621 66. Gray JW Ubhi H Milner P Antimicrobial treatment of serious gram-negative infections in newborns Curr Infect Dis Rep 2014 16 400 10.1007/s11908-014-0400-6 24535244 67. Poggi C Bianconi T Gozzini E Generoso M Dani C Presepsin for the detection of late-onset sepsis in preterm newborns Pediatrics 2015 135 68 75 10.1542/peds.2014-1755 25511124 68. Cassir N Benamar S Khalil JB Croce O Saint-Faust M Jacquot A Million M Azza S Armstrong N Henry M Jardot P Robert C Gire C Lagier JC Chabriere E Ghigo E Marchandin H Sartor C Boutte P Cambonie G Simeoni U Raoult D La Scola B Clostridium butyricum strains and dysbiosis linked to necrotizing enterocolitis in preterm neonates Clin Infect Dis 2015 61 1107 1115 10.1093/cid/civ468 26084844 69. McElroy SJ The role of bacteria in necrotizing enterocolitis: understanding the forest for the trees Neonatology 2015 108 196 197 10.1159/000437205 26278707 70. Caffarelli C Cardinale F Povesi-Dascola C Dodi I Mastrorilli V Ricci G Use of probiotics in pediatric infectious diseases Expert Rev Anti Infect Ther 2015 13 1517 1535 10.1586/14787210.2015.1096775 26496433 71. Aceti A Gori D Barone G Callegari ML Di Mauro A Fantini MP Indrio F Maggio L Meneghin F Morelli L Zuccotti G Corvaglia L Italian society of N, Probiotics for prevention of necrotizing enterocolitis in preterm infants: systematic review and meta-analysis Ital J Pediatr 2015 41 89 10.1186/s13052-015-0199-2 26567539 72. Alshaikh B Kostecky L Blachly N Yee W Effect of a quality improvement project to use exclusive Mother’s Own milk on rate of necrotizing enterocolitis in preterm infants Breastfeed Med 2015 10 355 361 10.1089/bfm.2015.0042 26230909 73. Maheshwari A Immunologic and hematological abnormalities in necrotizing enterocolitis Clin Perinatol 2015 42 567 585 10.1016/j.clp.2015.04.014 26250918 74. Christensen RD Yoder BA Baer VL Snow GL Butler A Early-onset neutropenia in small-for-gestational-Age infants Pediatrics 2015 136 e1259 1267 10.1542/peds.2015-1638 26459642 75. Ribet M Ponte C Lequien P Lacombe A Gosselin B Tumor of the thymus, hemothorax and respiratory distress in a newborn infant Presse Med 1971 79 1913 1914 5128491 76. Gargano G Paltrinieri AL Gallo C Di Pancrazio L Roversi MF Ferrari F Massive thymic hemorrhage and hemothorax occurring in utero Ital J Pediatr 2015 41 88 10.1186/s13052-015-0196-5 26568429 77. Dani C Pratesi S Raimondi F Romagnoli C Task Force for Hyperbilirubinemia of the Italian Society of N Italian guidelines for the management and treatment of neonatal cholestasis J Pediatr 2015 41 69 78. Gotze T Blessing H Grillhosl C Gerner P Hoerning A Neonatal cholestasis - differential diagnoses, current diagnostic procedures, and treatment Front Pediatr 2015 3 43 26137452 79. Basiri-Moghadam M Basiri-Moghadam K Kianmehr M Jani S The effect of massage on neonatal jaundice in stable preterm newborn infants: a randomized controlled trial J Pak Med Assoc 2015 65 602 606 26060154 80. Dalili H Sheikhi S Shariat M Haghnazarian E Effects of baby massage on neonatal jaundice in healthy Iranian infants: a pilot study Infant Behav Dev 2016 42 22 26 10.1016/j.infbeh.2015.10.009 26646074 81. Lin CH Yang HC Cheng CS Yen CE Effects of infant massage on jaundiced neonates undergoing phototherapy Ital J Pediatr 2015 41 94 10.1186/s13052-015-0202-y 26607061 82. Makrides M Anderson A Gibson RA Collins CT Improving the neurodevelopmental outcomes of low-birthweight infants Nestle Nutr Inst Workshop Ser 2013 74 211 221 10.1159/000348775 23887122 83. Girelli G Antoncecchi S Casadei AM Del Vecchio A Isernia P Motta M Regoli D Romagnoli C Tripodi G Velati C Recommendations for transfusion therapy in neonatology Blood Transfus 2015 13 484 497 26445308 84. Berglund SK Westrup B Domellof M Iron supplementation until 6 months protects marginally low-birth-weight infants from iron deficiency during their first year of life J Pediatr Gastroenterol Nutr 2015 60 390 395 10.1097/MPG.0000000000000633 25406528 85. Jin HX Wang RS Chen SJ Wang AP Liu XY Early and late Iron supplementation for low birth weight infants: a meta-analysis Ital J Pediatr 2015 41 16 10.1186/s13052-015-0121-y 25888053 86. Agostoni C Buonocore G Carnielli VP De Curtis M Darmaun D Decsi T Domellof M Embleton ND Fusch C Genzel-Boroviczeny O Goulet O Kalhan SC Kolacek S Koletzko B Lapillonne A Mihatsch W Moreno L Neu J Poindexter B Puntis J Putet G Rigo J Riskin A Salle B Sauer P Shamir R Szajewska H Thureen P Turck D van Goudoever JB Ziegler EE Nutrition ECo Enteral nutrient supply for preterm infants: commentary from the european society of paediatric gastroenterology, hepatology and nutrition committee on nutrition J Pediatr Gastroenterol Nutr 2010 50 85 91 10.1097/MPG.0b013e3181adaee0 19881390 87. Shah L Shah GS Singh RR Pokharel H Mishra OP Status of gastric lavage in neonates born with meconium stained amniotic fluid: a randomized controlled trial Ital J Pediatr 2015 41 85 10.1186/s13052-015-0194-7 26518433 88. Deshmukh M Balasubramanian H Rao S Patole S Effect of gastric lavage on feeding in neonates born through meconium-stained liquor: a systematic review Arch Dis Child Fetal Neonatal Ed 2015 100 F394 399 10.1136/archdischild-2015-308292 26040920 89. Ameta G Upadhyay A Gothwal S Singh K Dubey K Gupta A Role of gastric lavage in vigorous neonates born with meconium stained amniotic fluid Indian J Pediatr 2013 80 195 198 10.1007/s12098-012-0805-x 22990631 90. Uwaezuoke SN Steroid-sensitive nephrotic syndrome in children: triggers of relapse and evolving hypotheses on pathogenesis Ital J Pediatr 2015 41 19 10.1186/s13052-015-0123-9 25888239 91. Ishikura K Matsumoto S Sako M Tsuruga K Nakanishi K Kamei K Saito H Fujinaga S Hamasaki Y Chikamoto H Ohtsuka Y Komatsu Y Ohta T Nagai T Kaito H Kondo S Ikezumi Y Tanaka S Kaku Y Iijima K Japanese Society for Pediatric N, Japanese Society for Pediatric N Clinical practice guideline for pediatric idiopathic nephrotic syndrome 2013: medical therapy Clin Exp Nephrol 2015 19 6 33 10.1007/s10157-014-1030-x 25653046 92. Fujinaga S Hirano D Risk of persistent steroid dependency after switching from cyclosporine to mycophenolate mofetil in children with idiopathic nephrotic syndrome Pediatr Nephrol 2015 30 2051 2052 10.1007/s00467-015-3163-x 26194402 93. McCaffrey J, Lennon R, Webb NJ. The non-immunosuppressive management of childhood nephrotic syndrome. Pediatr Nephrol. 2015. 94. Nelson KB Blair E Prenatal factors in singletons with cerebral palsy born at or near term N Engl J Med 2015 373 946 953 10.1056/NEJMra1505261 26332549 95. Blair EM Nelson KB Fetal growth restriction and risk of cerebral palsy in singletons born after at least 35 weeks’ gestation Am J Obstet Gynecol 2015 212 520 10.1016/j.ajog.2014.10.1103 25448521 96. McMichael G Bainbridge MN Haan E Corbett M Gardner A Thompson S van Bon BW van Eyk CL Broadbent J Reynolds C O’Callaghan ME Nguyen LS Adelson DL Russo R Jhangiani S Doddapaneni H Muzny DM Gibbs RA Gecz J MacLennan AH Whole-exome sequencing points to considerable genetic heterogeneity of cerebral palsy Mol Psychiatry 2015 20 176 182 10.1038/mp.2014.189 25666757 97. Rosa M De Lucia S Rinaldi VE Le Gal J Desmarest M Veropalumbo C Romanello S Titomanlio L Paediatric arterial ischemic stroke: acute management, recent advances and remaining issues Ital J Pediatr 2015 41 95 10.1186/s13052-015-0174-y 26631262 98. Pillai SC Hacohen Y Tantsis E Prelog K Merheb V Kesson A Barnes E Gill D Webster R Menezes M Ardern-Holmes S Gupta S Procopis P Troedson C Antony J Ouvrier RA Polfrit Y Davies NW Waters P Lang B Lim MJ Brilot F Vincent A Dale RC Infectious and autoantibody-associated encephalitis: clinical features and long-term outcome Pediatrics 2015 135 e974 984 10.1542/peds.2014-2702 25802349 99. Davanzo R Romagnoli C Corsello G Position statement on breastfeeding from the Italian pediatric societies Ital J Pediatr 2015 41 80 10.1186/s13052-015-0191-x 26498033 100. Khan J Vesel L Bahl R Martines JC Timing of breastfeeding initiation and exclusivity of breastfeeding during the first month of life: effects on neonatal mortality and morbidity--a systematic review and meta-analysis Matern Child Health J 2015 19 468 479 10.1007/s10995-014-1526-8 24894730 101. Giugliani ER Horta BL Loret De Mola C Lisboa BO Victora CG Effect of breastfeeding promotion interventions on child growth: a systematic review and meta-analysis Acta Paediatr 2015 104 20 29 10.1111/apa.13160 26361071 102. Lodge CJ Tan DJ Lau MX Dai X Tham R Lowe AJ Bowatte G Allen KJ Dharmage SC Breastfeeding and asthma and allergies: a systematic review and meta-analysis Acta Paediatr 2015 104 38 53 10.1111/apa.13132 26192405 103. Alvisi P Brusa S Alboresi S Amarri S Bottau P Cavagni G Corradini B Landi L Loroni L Marani M Osti IM Povesi-Dascola C Caffarelli C Valeriani L Agostoni C Recommendations on complementary feeding for healthy, full-term infants Ital J Pediatr 2015 41 36 10.1186/s13052-015-0143-5 25928205 104. Zheng JS Liu H Zhao YM Li J Chen Y Zhu S Chen H Huang T Li D Complementary feeding and childhood adiposity in preschool-aged children in a large Chinese cohort J Pediatr 2015 166 326 331 10.1016/j.jpeds.2014.11.010 25491091 105. Lumia M Takkinen HM Luukkainen P Kaila M Lehtinen-Jacks S Nwaru BI Tuokkola J Niemela O Haapala AM Ilonen J Simell O Knip M Veijola R Virtanen SM Food consumption and risk of childhood asthma Pediatr Allergy Immunol 2015 26 789 796 10.1111/pai.12352 25693000 106. Herrick KA Rossen LM Nielsen SJ Branum AM Ogden CL Fruit consumption by youth in the United States Pediatrics 2015 136 664 671 10.1542/peds.2015-1709 26391940 107. Augusto RA Cobayashi F Cardoso MA Team AS Associations between low consumption of fruits and vegetables and nutritional deficiencies in Brazilian schoolchildren Public Health Nutr 2015 18 927 935 10.1017/S1368980014001244 24963861 108. Salam RA Das JK Bhutta ZA Current issues and priorities in childhood nutrition, growth, and infections J Nutr 2015 145 1116S 1122S 10.3945/jn.114.194720 25833888 109. White M Dennis N Ramsey R Barwick K Graham C Kane S Kepreotes H Queit L Sweeney A Winderlich J Wong See D Littlewood R Prevalence of malnutrition, obesity and nutritional risk of Australian paediatric inpatients: a national 1-day snapshot J Paediatr Child Health 2015 51 314 320 10.1111/jpc.12709 25123425 110. Hjorth L Haupt R Skinner R Grabow D Byrne J Karner S Levitt G Michel G van der Pal H Bardi E Beck JD de Vathaire F Essig S Frey E Garwicz S Hawkins M Jakab Z Jankovic M Kazanowska B Kepak T Kremer L Lackner H Sugden E Terenziani M Zaletel LZ Kaatsch P PanCare N Survivorship after childhood cancer: PanCare: a European network to promote optimal long-term care Eur J Cancer 2015 51 1203 1211 10.1016/j.ejca.2015.04.002 25958037 111. Hingorani P Missiaglia E Shipley J Anderson JR Triche TJ Delorenzi M Gastier-Foster J Wing M Hawkins DS Skapek SX Clinical application of prognostic gene expression signature in fusion gene-negative Rhabdomyosarcoma: a report from the Children’s oncology group Clin Cancer Res 2015 21 4733 4739 10.1158/1078-0432.CCR-14-3326 26473193 112. Ma X Huang D Zhao W Sun L Xiong H Zhang Y Jin M Zhang D Huang C Wang H Zhang W Sun N He L Tang J Clinical characteristics and prognosis of childhood rhabdomyosarcoma: a 10-year retrospective multicenter study Int J Clin Exp Med 2015 8 17196 17205 26770312 113. Sangkhathat S Current management of pediatric soft tissue sarcomas World J Clin Pediatr 2015 4 94 105 10.5409/wjcp.v4.i4.94 26566481 114. Takizawa D Oshiro Y Mizumoto M Fukushima H Fukushima T Sakurai H Proton beam therapy for a patient with large rhabdomyosarcoma of the body trunk Ital J Pediatr 2015 41 90 10.1186/s13052-015-0200-0 26573272 115. Fukushima H Fukushima T Sakai A Suzuki R Kobayashi C Oshiro Y Mizumoto M Hoshino N Gotoh C Urita Y Komuro H Kaneko M Sekido N Masumoto K Sakurai H Sumazaki R Tailor-made treatment combined with proton beam therapy for children with genitourinary/pelvic rhabdomyosarcoma Rep Pract Oncol Radiother 2015 20 217 222 10.1016/j.rpor.2014.12.003 25949226 116. Chen YJ Wambach JA DePass K Wegner DJ Chen SK Zhang QY Heins H Cole FS Hamvas A Population-based frequency of surfactant dysfunction mutations in a native Chinese cohort World J Pediatr 2016 12 190 195 10.1007/s12519-015-0047-x 26547207 117. Gower WA Nogee LM Surfactant dysfunction Paediatr Respir Rev 2011 12 223 229 10.1016/j.prrv.2011.01.005 22018035 118. Montella S Vece TJ Langston C Carrera P Nogee LM Hamvas A Manna A Cervasio M Santamaria F A disorder of surfactant metabolism without identified genetic mutations Ital J Pediatr 2015 41 93 10.1186/s13052-015-0198-3 26606984 119. Murray J Bottle A Sharland M Modi N Aylin P Majeed A Saxena S Medicines for Neonates Investigator G Risk factors for hospital admission with RSV bronchiolitis in England: a population-based birth cohort study PLoS One 2014 9 e89186 10.1371/journal.pone.0089186 24586581 120. Quinonez RA Schroeder AR Safely doing less and the new AAP bronchiolitis guideline Pediatrics 2015 135 793 795 10.1542/peds.2014-3703 25847807 121. Quinonez RA Garber MD Schroeder AR Alverson BK Nickel W Goldstein J Bennett JS Fine BR Hartzog TH McLean HS Mittal V Pappas RM Percelay JM Phillips SC Shen M Ralston SL Choosing wisely in pediatric hospital medicine: five opportunities for improved healthcare value J Hosp Med 2013 8 479 485 10.1002/jhm.2064 23955837 122. Skjerven HO Hunderi JO Brugmann-Pieper SK Brun AC Engen H Eskedal L Haavaldsen M Kvenshagen B Lunde J Rolfsjord LB Siva C Vikin T Mowinckel P Carlsen KH Lodrup Carlsen KC Racemic adrenaline and inhalation strategies in acute bronchiolitis N Engl J Med 2013 368 2286 2293 10.1056/NEJMoa1301839 23758233 123. Ricci V Delgado Nunes V Murphy MS Cunningham S Guideline Development G Technical T Bronchiolitis in children: summary of NICE guidance BMJ 2015 350 h2305 10.1136/bmj.h2305 26037525 124. Bollani L Baraldi E Chirico G Dotta A Lanari M Del Vecchio A Manzoni P Boldrini A Paolillo P Di Fabio S Orfeo L Stronati M Romagnoli C Italian Society of N Revised recommendations concerning palivizumab prophylaxis for respiratory syncytial virus (RSV) Ital J Pediatr 2015 41 97 10.1186/s13052-015-0203-x 26670908 125. Lanari M Prinelli F Adorni F Di Santo S Vandini S Silvestri M Musicco M Study Group of Italian Society of Neonatology on Risk Factors for RSVH Risk factors for bronchiolitis hospitalization during the first year of life in a multicenter Italian birth cohort Ital J Pediatr 2015 41 40 10.1186/s13052-015-0149-z 26006025

PMC005xxxxxx/PMC5002165.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 70110.1186/s12974-016-0701-9Research1-Oleyl-lysophosphatidic acid (LPA) promotes polarization of BV-2 and primary murine microglia towards an M1-like phenotype Plastira Ioanna ioanna.plastira@medunigraz.at 1Bernhart Eva eva.bernhart@medunigraz.at 1Goeritzer Madeleine madeleine.goeritzer@medunigraz.at 12Reicher Helga helga.reicher@medunigraz.at 1Kumble Vishwanath Bhat vishwanath.bhat-kumble@medunigraz.at 1Kogelnik Nora nora.kogelnik@medunigraz.at 1Wintersperger Andrea andrea.wintersperger@medunigraz.at 1Hammer Astrid astrid.hammer@medunigraz.at 3Schlager Stefanie stefanie.schlager@medunigraz.at 1Jandl Katharina katharina.jandl@medunigraz.at 4Heinemann Akos akos.heinemann@medunigraz.at 4Kratky Dagmar dagmar.kratky@medunigraz.at 12Malle Ernst ernst.malle@medunigraz.at 1Sattler Wolfgang +43-316-380-4188wolfgang.sattler@medunigraz.at 121 Institute of Molecular Biology and Biochemistry, Medical University of Graz, Harrachgasse 21, 8010 Graz, Austria 2 BioTechMed-Graz, Graz, Austria 3 Institute of Cell Biology, Histology and Embryology, Medical University of Graz, Graz, Austria 4 Institute of Experimental and Clinical Pharmacology, Medical University of Graz, Graz, Austria 26 8 2016 26 8 2016 2016 13 1 20521 7 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 Microglia, the immunocompetent cells of the CNS, rapidly respond to brain injury and disease by altering their morphology and phenotype to adopt an activated state. Microglia can exist broadly between two different states, namely the classical (M1) and the alternative (M2) phenotype. The first is characterized by the production of pro-inflammatory cytokines/chemokines and reactive oxygen and/or nitrogen species. In contrast, alternatively activated microglia are typified by an anti-inflammatory phenotype supporting wound healing and debris clearance. The objective of the present study was to determine the outcome of lysophosphatidic acid (LPA)-mediated signaling events on microglia polarization. Methods LPA receptor expression and cyto-/chemokine mRNA levels in BV-2 and primary murine microglia (PMM) were determined by qPCR. M1/M2 marker expression was analyzed by Western blotting, immunofluorescence microscopy, or flow cytometry. Cyto-/chemokine secretion was quantitated by ELISA. Results BV-2 cells express LPA receptor 2 (LPA2), 3, 5, and 6, whereas PMM express LPA1, 2, 4, 5, and 6. We show that LPA treatment of BV-2 and PMM leads to a shift towards a pro-inflammatory M1-like phenotype. LPA treatment increased CD40 and CD86 (M1 markers) and reduced CD206 (M2 marker) expression. LPA increased inducible nitric oxide synthase (iNOS) and COX-2 levels (both M1), while the M2 marker Arginase-1 was suppressed in BV-2 cells. Immunofluorescence studies (iNOS, COX-2, Arginase-1, and RELMα) extended these findings to PMM. Upregulation of M1 markers in BV-2 and PMM was accompanied by increased cyto-/chemokine transcription and secretion (IL-1β, TNFα, IL-6, CCL5, and CXCL2). The pharmacological LPA5 antagonist TCLPA5 blunted most of these pro-inflammatory responses. Conclusions LPA drives BV-2 and PMM towards a pro-inflammatory M1-like phenotype. Suppression by TCLPA5 indicates that the LPA/LPA5 signaling axis could represent a potential pharmacological target to interfere with microglia polarization in disease. Keywords IL-4IL-10LPSM1 and M2 polarizationNeuroinflammationhttp://dx.doi.org/http://dx.doi.org/10.13039/501100002428Austrian Science FundDK-MOLIN W1241http://dx.doi.org/http://dx.doi.org/10.13039/501100002428Austrian Science FundDK-MCD W1226BioTechMed Grazissue-copyright-statement© The Author(s) 2016 ==== Body Background Microglia are the resident immune cells of the brain and are endowed with specific receptor sets that are able to detect subtle alterations of the finely tuned micromilieu in the central nervous system (CNS) [1, 2]. Even in the resting state, the dynamic microglia processes scan the CNS environment and respond to danger signals [3]. Neuronal injury results in the release of ATP, neurotransmitters, growth factors and cytokines, or in changes of local ion homeostasis resulting in microglia activation [4]. Depending on the signal encountered, microglia can activate different repair programs that determine the severity of the response [4]. These responses include morphological transformation (increased size of cell bodies, thickening of proximal processes, decreased ramification of distal branches), proliferation, migration, phagocytosis, and the production of bioactive molecules [5]. These events are presumably the first steps that mobilize the cellular and molecular defense machinery in the CNS leading to subsequent microglia activation. Depending on the environmental milieu and stimulus encountered, this activation profile of microglia can reach from classically activated (M1) to alternatively activated (M2) cells [6]. Lysophosphatidic acid (LPA) is a mixture of saturated or unsaturated acyl/alkyl residues at the sn-1 or sn-2 position [7] that are present in biological fluids including cerebrospinal fluid (CSF) [8]. LPA is produced through different pathways by means of phospholipase A1- and A2-mediated hydrolysis of phosphatidic acid [8], de novo synthesis from glycerol-3-phosphate via acyltransferases [9], or autotaxin (ATX)-dependent cleavage of lysophosphatidylcholine [10]. The majority of circulating LPA is thought to originate from ATX [11]. In line, LPA levels in mice heterozygous for the ATX allele are reduced by 50 % [12–14]. Tissue distribution analyses revealed that ATX is expressed in murine and human brain [15]. Consistently, the brain contains significant levels of LPA [16], which increase in response to CNS injury [17–19]. In vitro and in vivo data demonstrated that LPA receptor-mediated signaling cascades play prominent roles in the CNS [20, 21]. The effects of LPA are mediated through the six currently recognized G protein-coupled receptors termed LPA receptors 1-6 (LPA1-6), which couple to one or more heterotrimeric G proteins [21, 22]. All members of the LPA receptor family are expressed in the CNS and regulate blood-brain barrier permeability [23], neuroprogenitor cell function [24], synaptic transmission [25], myelination [26], and brain immune responses [27–29]. Microglia express a range of LPA receptors that regulate cell morphology [30], membrane ruffling and hyperpolarization [31], metabolic changes [27], migration and ion mobilization [29], and growth factor production [31]. Recently, it was demonstrated that activation of LPA1 contributes to demyelination in spinal cord injury and that these effects are partly mediated by activated microglia [19]. In the peripheral immune system, LPA may adopt the role of sphingosine-1-phosphate during lymphocyte egress [32]. ATX is expressed in lymphoid organ high endothelial venules, where it generates LPA and promotes the entry of lymphocytes into lymphoid organs [33, 34]. It was suggested that local synthesis of LPA by ATX provides a chemotactic signal that facilitates T cell exit from the circulation into the lymph nodes [35]. In accordance, 1-bromo-3(S)-hydroxy-4-(palmitoyloxy)butyl]phosphonate (BrP-LPA), which antagonizes ATX and LPA receptor function, attenuated trafficking of lymphocytes into lymph nodes [36]. LPA potently affects migration, chemotaxis, proliferation, survival signaling, and interleukin-1β (IL-1β) secretion in different cell types of the immune system (reviewed in [35]). In monocyte-derived dendritic cells, LPA upregulates expression of CD86 (M1 marker), thereby enhancing T cell proliferation [37]. Unlike other CNS resident cells, microglia originate from hematopoietic stem cells in the yolk sac and act as primary cells responding to pathogens or injuries of the CNS. Similarly to macrophages, microglia polarize to an M1 phenotype (e.g. in response to lipopolysaccharide; LPS) in order to produce pro-inflammatory mediators [38]. In contrast, the anti-inflammatory cytokine IL-4 gives rise to the formation of a protective M2 microglia population in a murine stroke model [39]. Based on this evidence, it is conceivable that mediators controlling M1/M2 polarization of microglia could impact on promotion or resolution of neuroinflammation in the CNS. In earlier studies, we showed that LPA increases the expression of several glycolytic enzymes in microglia [27], indicating a shift towards a pro-inflammatory M1 polarization [40]. We suggest that LPA may act as a potent regulator of microglia signaling, thereby modulating microglia polarization. Methods Materials Cell culture medium RPMI1640 and DMEM (Dulbecco’s modified Eagle’s medium), fetal calf serum (FCS), antibiotics, and 0.25 % trypsin were from Invitrogen (Waltham, MA, USA). LPA (1-oleoyl-2-hydroxy-sn-glycero-3-phosphate) was from Sigma-Aldrich (St. Louis, MO, USA). Recombinant murine IL-4 and IL-10 were from Peprotech (Rocky Hill, NJ, USA) and LPS from E. coli (O111:B4) was from Sigma-Aldrich (St. Louis, MO, USA). Antibodies against cyclooxygenase 2 (COX-2) and arginase-1 (Arg-1; used only for Western blotting) were purchased from Cell Signaling (Beverly, MA, USA), inducible nitric oxide synthase (iNOS) antibody was from BD Biosciences (San Jose, CA, USA). For immunofluorescence, the COX-2 and Arg-1 antibodies were from Santa Cruz (Dallas, TX, USA), the antibodies against resistin-like alpha (RELMα, alternative name FIZZ-1) and iNOS were from Abcam (Cambridge, UK) and the CD11b antibody was from Novus Biologicals (Littleton, CO, USA). β-Tubulin monoclonal antibody, β-actin and tomato-lectin were from Sigma-Aldrich (St. Louis, MO, USA). Alexa Fluor® 488 Phalloidin was from Invitrogen (Waltham, MA, USA). Phycoerythrin (PE)-CD40, allophycocyanin (APC)-CD86, PE-CD206 antibodies and their isotype controls were from e-Bioscience (San Diego, CA, USA). 5-(3-Chloro-4-cyclohexylphenyl)-1-(3-methoxyphenyl)-1H-pyrazole-3-carboxylic acid (TCLPA5) was from Tocris Bioscience (Bristol, UK) and BrP-LPA was from Echelon Biosciences (Salt Lake City, UT, USA). tert-Butyl hydroperoxide was from Sigma-Aldrich. ELISA development kits were from Peprotech (Rocky Hill, NJ, USA). Primers were from QIAGEN or Invitrogen. Kits that were used for quantitative real-time PCR (qPCR) analysis were purchased from QIAGEN (Hilden, Germany) or Applied Biosystems (Foster City, CA, USA). BV-2 microglia The BV-2 murine microglial cell line was purchased from Banca Biologica e Cell Factory (Genova, Italy). Cells were grown and maintained in RPMI1640 medium supplemented with 10 % FCS, 1 % penicillin/streptomycin, and 1 % glutamine and cultured under standard conditions. Culture medium was changed to fresh medium every 2 or 3 days and when the cells reached confluence, they were splitted into new flasks or used immediately for the experiments. Primary microglia isolation and culture Primary murine microglia (PMM) were isolated and purified from cortices of neonatal (P0–P4) mice as previously described [41]. In brief, cortices were isolated from the whole brain, stripped from their meninges, and minced into small pieces. Glial cells were separated by trypsinization (0.1 % trypsin, 20 min, 37 °C, 5 % CO2), and the cell suspension was cultured in 75 cm2 tissue culture flasks precoated with 5 mg/ml poly-d-lysine in DMEM containing 15 % FCS, 1 % penicillin/streptomycin, and 1 % glutamine. After 2 to 3 days in culture, the medium was changed to DMEM/10 % FCS and cells were cultured for additional 10 to 14 days. Microglia were removed from the astrocytic monolayer by smacking the flasks 10-20 times and seeded onto poly-d-lysine-coated cell culture plates for further use. The purity of primary murine microglia was determined by immunocytochemistry with anti-CD11b or tomato-lectin staining and was always ≥95 %. qPCR analysis Total RNA was extracted from BV-2 or primary microglia cells using the RNeasy Mini or RNeasy Micro kit (QIAGEN, Hilden, Germany) and quantified using NanoDrop (Thermo Fisher Scientific, Waltham, MA, USA). RNA was reverse-transcribed by using the high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA) or by using the SuperScript® III reverse transcription kit (Invitrogen, Waltham, MA, USA). Quantitative real-time PCR (qPCR) was performed on an Applied Biosystems 7900HT Fast Real-Time PCR System using the QuantifastTM SYBR® Green PCR kit (QIAGEN, Hilden, Germany). Amplification of murine hypoxanthine-guanine phosphoribosyltransferase (HPRT) as housekeeping gene was performed on all samples as internal controls to account for variations in mRNA levels. Expression profiles and associated statistical parameters were analyzed by the 2-ddCt method [42]. Gene specific primers were purchased from QIAGEN (LPA receptors) and Invitrogen (cytokines and chemokines). Primer sequences are listed in Tables 1 and 2. In case of non-detects, we discriminate between undetermined values (that do not exceed the Ct threshold) and absent values (no reaction occurred) [43].Table 1 Primers (QIAGEN) used for qPCR analyses of LPA receptor expression Gene Cat. no. Product length (bp) Hprt QT00166768 168 Lpar1 QT00107709 94 Lpar2 QT00106008 94 Lpar3 QT00264320 99 Lpar4 QT00125888 96 Lpar5 QT00312571 100 Lpar6 QT00325668 118 Table 2 Primers (Invitrogen) used for expression analyses of cytokines and chemokines by qPCR Gene Forward primer (5′–3′) Reverse primer (5′–3′) Product length (bp) Ccl5 GCTGCTTTGCCTACCTCTCC TCGAGTGACAAACACGACTGC 104 Cxcl2 AGTGAACTGCGCTGTCAATG GCCCTTGAGAGTGGCTATGA 126 Il1β TGTGAAATGCCACCTTTTGA GGTCAAAGGTTTGGAAGCAG 94 Il6 TGATGCACTTGCAGAAAACA ACCAGAGGAAATTTTCAATAGGC 109 Tnfα CCACCACGCTCTTCTGTCTAC AGGGTCTGGGCCATAGAACT 103 LPA treatment For LPA treatment, BV-2 and PMM were plated in poly-d-lysine precoated plates (different sizes according to each experiment) and allowed to adhere for 2–3 days. Before treatments, cells were always incubated in serum-free DMEM overnight (o/n). The following day, medium was changed to serum-free DMEM containing 0.1 % BSA (control) or DMEM containing 0.1 % BSA and LPA (1 μM). BSA was used as LPA carrier. Aqueous LPA stock solutions (10 mM) were stored at −70 °C. Only freshly thawed stock solutions were used for the experiments. LPA solutions, culture medium, BSA, and PBS that were used during the experiments were tested for endotoxin content using the Limulus amebocyte lysate test. Endotoxin content was always <0.5 EU/ml. Pharmacological inhibition of LPA receptors We used BrP-LPA, a pan LPA receptor/ATX antagonist [44] and TCLPA5, a specific antagonist for LPA5 [45]. BrP-LPA was diluted in distilled water (stock concentration of 2 mM), aliquoted and kept at −20 °C. TCLPA5 was diluted in dimethylsulfoxide (DMSO) (stock concentration 100 mM) and kept at −20 °C. TCLPA5 solutions are stated to be stable for maximum 40 days. During the experiments, both antagonists were used at a final concentration of 5 μM. Cells were pretreated with the antagonists for 2 h before starting the experiment. Immunoblotting For Western blotting experiments, BV-2 cells were seeded onto 6-well plates at a density of 1 × 105 cells/well. Prior to experiments, cells were cultured in serum-free medium (o/n) and then incubated in serum-free medium containing 0.1 % BSA in the presence of 1 μM LPA or 1 μM LPA in the absence or presence of BrP-LPA (5 μM) or TCLPA5 (5 μM) for the indicated time periods. After removing the supernatant, cells were washed twice with ice-cold PBS and lysed in RIPA buffer (50 mM Tris-HCl pH 7.4, 1 % NP-40, 150 mM NaCl, 1 mM Na3VO4, 1 mM NaF, 1 mM EDTA) containing protease inhibitors (aprotinin, leupeptin, pepstatin: 1 μg/ml each), 10 μM PMSF, and phosphatase inhibitors (Thermo Scientific, Vienna, Austria). Cells were mechanically scraped off using a rubber scraper and centrifuged at 13,000 rpm for 10 min. Protein concentrations were determined using the BCA kit (Thermo Scientific) using BSA as a standard. One hundred micrograms of total cell protein was loaded per lane and separated by SDS-PAGE (10 %). After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes using electrophoretic transfer (Bio-Rad, Berkeley, CA, USA). Membranes were blocked with 5 % low-fat milk in TBST for 2 h at room temperature and incubated with the following primary antibodies: anti-iNOS (1:500), anti-Arg-1 (1:1000), and anti-COX-2 (1:1000). After removal of primary antibodies, the membranes were washed for 30 min in TBST and incubated for 2 h at room temperature with HRP-conjugated secondary antibodies (anti-rabbit 1:10,000; anti-mouse 1:5000). After washing with TBST for 1 h, immunoreactive bands were visualized using ECL or ECL plus reagents (Thermo Scientific) and detected with a chemiluminescence detection system (ChemiDoc Bio-Rad, Berkeley, CA, USA). Immunofluorescence To examine changes in cell morphology in response to LPA treatment, BV-2 and primary microglia were stained for β-tubulin or F-actin. After seeding, cells were serum-starved (untreated) or incubated in the presence of 1 μM LPA (24 h). Then, cells were washed with pre-warmed PBS, fixed with 4 % paraformaldehyde/PBS for 10 min, and permeabilized with 0.5 % TritonX/PBS for 10 min at room temperature. Following washing with PBS, cells were incubated with blocking buffer (Thermo Scientific, Waltham, MA, USA) for 1 h at 4 °C. Incubations with anti-β-tubulin (1:100) or phalloidin (1:50) were carried out overnight at 4 °C. Finally, cells stained for β-tubulin were incubated with cyanine (Cy)-3-labeled secondary antibody (1:200, 30 min, room temperature). Microglia nuclei were counterstained with Hoechst 33342 (Invitrogen, Waltham, MA, USA). All slides were washed three times with PBS and mounted using a mounting medium (Dako Austria GmbH, Vienna, Austria). To identify changes in M1 and M2 markers, BV-2 and PMM were seeded onto poly-d-lysine-coated chamber slides at a density of 4 × 104 and 1 × 105 cells, respectively, and incubated in serum-free medium in the absence or presence of LPA (1 μM) or LPA plus TCLPA5 (5 μM). LPS (20 ng/ml) and IL-4 (40 ng/ml) were used as positive controls for M1 and M2 polarization of PMM. Fixation, permeabilization, and blocking steps were performed as described above. Following blocking of unspecific binding, cells were incubated with FITC-conjugated tomato lectin (1:100) and antibodies against iNOS (1:20), Arg-1 (1:50), COX-2 (1:20), or RELMα (1:50), followed by incubation with labeled secondary antibodies. Slides were counterstained with Hoechst 33342. Confocal fluorescence microscopy imaging was performed using a Leitz/Leica TCSSP2 microscope (Leica Lasertechnik GmbH, Heidelberg, Germany). Quantification of fluorescence intensity was performed using ImageJ. At least 50 cells out of 3 different areas per chamber were measured. Flow cytometry Flow cytometry was used in order to assess the expression of CD40, CD86, and CD206 in microglia cells. BV-2 and PMM were seeded in triplicate onto 6-well and poly-d-lysine coated 24-well plates at a density of 1 × 105 and 1.5 × 105 cells per well, respectively. After 24 h serum starvation, cells were incubated in the presence of 1 μM LPA for 12, 24, and 48 h. BV-2 cells were also used to test effects of the inhibitors on surface marker expression. Serum-starved cells were incubated with vehicle controls, LPA, or LPA plus the antagonists for the abovementioned time periods. Cells were then collected, blocked using the Ultra V blocker (Thermo Scientific), and incubated with PE anti-CD40, APC anti-CD86, or PE anti-CD206 antibody (1:50). Finally, cells were fixed and measured using a Guava easyCyte 8 Millipore flow cytometer. ELISA IL-1β, tumor necrosis factor α (TNFα), interleukin-6 (IL-6), CCL5 (RANTES), and C-X-C motif chemokine 2 (CXCL2, alternative name MIP-2) concentrations in the cellular supernatant were quantitated using murine ELISA development kits (Peprotech, NJ, USA). Briefly, BV-2 and PMM were seeded in triplicate onto 12-well and poly-d-lysine coated 24-well plates at a density of 1 × 105 and 2.5 × 105 cells per well, respectively. After serum starvation (o/n), cells were incubated in serum-free medium, containing LPA in the absence or presence of BrP-LPA or TCLPA5 for the indicated times. For each time point, the supernatants were collected and kept at −70 °C until further use. The assays were performed according to manufacturer’s instructions. The standard curve for each ELISA was performed in triplicate. The concentrations of the cytokines and chemokines were determined using the external standard curve. NO detection iNOS activity was assessed indirectly by measuring the accumulated total nitrate levels in the supernatant after 12, 24, and 48 h in the case of BV-2 cells and 2, 8, 24, and 48 h in the case of primary microglia cells, using the total nitric oxide assay kit (ENZO Life Sciences, Switzerland). In this Griess assay, nitrate is reduced to nitrite by means of nitrate reductase. Fifty microliters of supernatant from each sample were processed according to manufacturer’s protocol. A standard curve was generated in the range between 0 and 100 μM using nitrate as standard. The total nitrate concentration per sample was determined using the external calibration curve. Measurement of carboxy-H2DCFDA oxidation Intracellular reactive oxygen species (ROS) levels were measured using the 2′,7′-dichlorofluorescin diacetate (DCFDA) cellular ROS detection kit (Abcam, Cambridge, UK). After internalization and subsequent hydrolysis, the redox indicator probe carboxy-H2DCFDA is converted to carboxy-H2DCF, which in the presence of oxidant species is oxidized to fluorescent carboxy-DCF [46]. BV-2 and PMM were seeded in black clear bottom 96-well plates at a density of 2.5 × 104 cells per well. Cells were allowed to adhere overnight and then incubated with 20 μM DCFDA for 40 min at 37 °C in the dark. The solution was removed and the cells were treated with vehicle control (DMSO), LPA, or LPA plus TCLPA5 for 0.5, 1, 3, and 6 h. Fluorescence intensity was measured with excitation and emission wavelengths of 485 and 535 nm, respectively. Statistical analysis All experiments were performed in triplicate and repeated at least three times. Data are presented as mean + SD unless otherwise stated. Statistical significance was determined by one-way ANOVA with Bonferroni correction using GraphPad 5.0 Prism. Values of p ≤ 0.05 were considered significantly different. Results LPA receptor expression in BV-2 and PMM was analyzed by qPCR. LPA2, LPA3, LPA5, and LPA6 were detected in BV-2 cells. LPA1 and 4 were undetectable (n.d.), whereas LPA3 was detected at very low copy numbers. In PMM LPA1, 2, 4, 5 and 6 were detected. LPA3 was undetectable (Fig. 1a). LPA6 expression was arbitrarily set to 1. To test whether LPA receptor expression changes upon cell activation, BV-2 and primary cells were incubated in the presence of LPA (1 μM). After 24 h, LPA receptor expression was analyzed by qPCR. These experiments (Fig. 1b) revealed that LPA is without effects on LPA receptor mRNA expression in BV-2 cells. LPA tended to decrease LPAR mRNA in PMM; however, these effects were statistically not significantly different from untreated cells.Fig. 1 LPA receptor expression in BV-2 cells and primary murine microglia (PMM). a Gene expression was monitored by qPCR and normalized to the housekeeping gene Hprt in BV-2 and PMM. Values are expressed as mean + SD (n = 6–9). LPA6 expression was arbitrarily set to 1. n.d. not detected. b LPA receptor expression after treatment with 1 μM LPA for 24 h. Results are expressed as mean + SD from three independent experiments Microglia are able to change their morphology in response to extracellular cues. Immunofluorescence studies revealed that LPA induced morphological changes of BV-2 (Fig. 2a) and PMM (Fig. 2b). F-actin staining (Fig. 2, upper panels) indicated that untreated cells were unipolar with one or more processes. After LPA treatment, the cells increased their surface area and acquired a flat morphology with more condensed actin labeling. Also the tubulin stains (Fig. 2, lower panels) indicated an increase in cell area in response to LPA treatment. Tubulin labeling changed from a diffuse pattern in untreated BV-2 cells to a more dense network in response to LPA. In primary cells, tubulin staining was also more condensed and occupied a larger area after LPA treatment.Fig. 2 LPA alters microglia morphology. Staining for F-actin and β-tubulin in a BV-2 cells and b primary murine microglia (PMM). Cells on chamber slides were serum starved (o/n) and incubated in the absence or presence of LPA (1 μM, 24 h). After washing and incubation with primary and secondary antibodies, cells were analyzed by confocal microscopy. Nuclei were stained with DAPI. Representative images reveal rearrangement of the cytoskeleton following LPA treatment. Scale bar = 20 μm. Results from one representative experiment (out of two) are shown Next, we evaluated the effects of LPA on BV-2 polarization by Western blot analyses to get an indication about time-dependent changes of M1/M2 marker expression. These analyses revealed that LPA increased the expression of iNOS and COX-2 (maximum induction at 24 h), but did not affect the M2 marker Arg-1 (Fig. 3a). As expected, stimulation of BV-2 cells with LPS (20 ng/ml) significantly increased iNOS and COX-2 levels. In contrast, IL-4 (a polarization signal towards M2) induced Arg-1 without affecting iNOS and COX-2 levels. IL-10 was without effect on iNOS and COX-2 levels but slightly decreased Arg-1 levels. The bar graphs in the right panel represent densitometric evaluation of the indicated protein bands.Fig. 3 LPA promotes classical activation of BV-2 and primary microglia. a Serum-starved BV-2 cells were treated with BSA (0.1 %; ‘c.’), LPA (1 μM), LPS (20 ng/ml), IL-4 (40 ng/ml), or IL-10 (40 ng/ml)], and cellular protein lysates were analysed by Western blotting. LPS, IL-4, and IL-10 were used to polarize cells to an M1- or M2-like phenotype, respectively. One representative plot for each protein and the densitometric analysis (mean + SD; normalized to actin) from four independent experiments is presented. Control = 0.1 % BSA. b Confocal immunofluorescence microscopy of PMM in the absence or presence of LPA (1 μM, 24 h). LPS (20 ng/ml) and IL-4 (40 ng/ml) were used to induce an M1- or M2-like phenotype, respectively. Cells were stained for CD11b (microglia marker) and COX-2 or Arg-1. Nuclei were counterstained with Hoechst. Scale bars = 20 μm. Results from one representative experiment (out of two) are shown. c PMM cultured on chamber slides were incubated in the absence or presence of LPA (1 μM) for 24 and 48 h. Cells were stained for specific inflammatory markers and nuclei were counterstained with Hoechst. The fluorescence intensity for each marker was quantitated with ImageJ. At least 50 cells out of 3 different areas per chamber were measured in two independent experiments. The results are presented as mean + SD (*p < 0.05, **p < 0.01, ***p < 0.001; one-way ANOVA with Bonferroni correction) Primary microglia were treated with LPA, LPS, or IL-4 and analyzed by confocal laser scanning microscopy for COX-2 and Arg-1 immunoreactivity. LPA treatment (1 μM, 24 h) led to increased COX-2 immunofluorescence that was mainly detected in cellular processes (Fig. 3b, upper panel). LPS induced a more rounded cell shape of PMM, and COX-2 staining was observed along the cell periphery. COX-2 expression was unaffected by IL-4. LPA (and LPS) treatment decreased Arg-1 staining, which was increased in response to IL-4 (Fig. 3b, lower panel). The bar graphs in the right panel show fluorescence intensities of micrographs displayed at the left. In time-dependent studies, LPA increased iNOS fluorescence intensity in PMM (Fig. 3c) by 1.6-fold (24 h). Using a double M1/M2 staining approach (COX-2 and RELMα), we detected 2.3-fold increased fluorescence intensity for COX-2 (at 48 h), while RELMα fluorescence was reduced by 22 % (48 h). Arg-1 expression was also reduced in response to LPA treatment (by 35 %). The right panel shows fluorescence intensities. We then analyzed surface marker expression by flow cytometry in LPA-treated BV-2 cells and PMM. LPA increased the percentage of CD40+ BV-2 cells from 7 to 20 % (Fig. 4a). LPS (20 ng/ml) increased the CD40+ population from 7 to 29 %. CD86 expression was increased approx. twofold in response to LPA (12 and 24 h), while CD206+ cells were reduced from 35 to 22 % (24 h). Primary microglia (Fig. 4b) showed a comparable though more pronounced response: LPA increased the percentage of CD40+ cells from 6 to 39 % (48 h), while LPS increased CD40+ cells to 67 %. The CD86+ cell population increased from 13 to 29 % (12 h) and then decreased to ≈12 % (48 h). The CD206+ population was reduced from 17 to 6 % (48 h).Fig. 4 LPA induces a pro-inflammatory microglia phenotype. a Serum-starved BV-2 cells (o/n) were treated with 1 μM LPA for the indicated time points. LPS (20 ng/ml) and IL-10 (40 ng/ml) were used as positive controls to induce M1- and M2-like phenotypes, respectively. Cells were stained with PE anti-CD40, APC anti-CD86 or PE anti-CD206 and analyzed using a FACSCalibur flow cytometer. Results (six separate experiments in triplicate) are expressed as mean + SD (*p < 0.05, **p < 0.01, *** p < 0.001; one-way ANOVA with Bonferroni correction). b Serum-starved PMM (o/n) were cultivated in the presence of LPA (1 μM) for the indicted times. LPS was used as a positive control. Cells were stained with PE-conjugated anti-CD40, APC-conjugated anti-CD86, or PE-conjugated anti-CD206 antibodies and analyzed using a Guava easyCyte 8 Millipore Flow Cytometer. Results from three individual preparations (measurements performed in duplicate) are shown as mean values + SD (*p < 0.05, *** p < 0.001 compared to untreated cells; one-way ANOVA with Bonferroni correction) Next, we determined the effect of LPA treatment on gene expression and secretion of selected pro-inflammatory cytokines and chemokines that are associated with an M1-like microglia phenotype [47]. qPCR analyses revealed that LPA increased transcription of Il1β, Tnfα, Il6, Ccl5, but not Cxcl2 in BV-2 cells (Fig. 5a, left panel). In PMM, LPA induced time-dependent transcription of Il1β, Tnfα, Il-6, Ccl5, and Cxcl2.Fig. 5 LPA induces expression and secretion of pro-inflammatory cytokines and chemokines. a BV-2 and PMM were cultured on 24-well plates and serum-starved (untreated) or incubated in the presence of 1 μM LPA for the indicated times. mRNA expression of different inflammatory cytokines and chemokines was monitored by qPCR and normalized to Hprt. Data are shown as mean + SD from three independent experiments performed in triplicate. Expression profiles were determined using the 2-ddCt method. b Serum-starved microglia cells were treated with LPA (1 μM) for the indicated times. Murine ELISA kits were used to quantitate the concentrations of IL-1β, TNFα, IL-6, CCL5 (RANTES), and CXCL2 (MIP-2) in the cellular supernatants. Results shown represent mean + SD from two independent experiments performed in triplicate. Data are expressed as mean values + SD (*p < 0.05; **p < 0.01; ***p < 0.001; one-way ANOVA with Bonferroni correction) In addition, we quantitated cytokine/chemokine concentrations in the cellular supernatants by ELISA. In both cell types, LPA augmented secretion of IL-1β, TNFα, IL-6, CCL5, and CXCL2 (Fig. 5b). In BV-2 cells, analytes were maximally induced at 2 or 8 h post activation. In PMM, maximal concentrations were observed between 24 (IL-1β, TNFα, IL-6, CCL5) and 48 h (CXCL2). To analyze the intracellular redox status, we measured DCF fluorescence. These experiments revealed a twofold increase of fluorescence 3 h post LPA addition in BV-2 cells (Fig. 6a, left panel). tert-Butyl hydroperoxide (tBHP; an inducer of intracellular ROS formation) was used as positive control (50 μM, 6 h). The DCF response was more pronounced in PMM and time-dependently increased (3.3-fold). In these cells, tBHP increased DCF fluorescence by 5.7-fold (Fig. 6a, right panel). Nitrate concentrations (surrogate markers for NO production via iNOS) in LPA-treated BV-2 and PMM increased by 1.3- and 1.1-fold, respectively (Fig. 6b).Fig. 6 LPA increases NO and ROS production in BV-2 and primary murine microglia. a The cellular redox status was determined using carboxy-H2DCFDA. Serum-starved BV-2 cells and PMM were incubated with carboxy-H2DCFDA and treated with LPA (1 μM), and the fluorescence intensity was quantitated. tert-Butyl hydroperoxide (TBHP) was used to induce intracellular ROS formation. Results (four independent experiments performed in triplicate) are expressed as mean values + SD. (**p ≤ 0.01; ***p < 0.001; one-way ANOVA with Bonferroni correction). b Serum-starved cells were treated with LPA (1 μM) for the indicated time periods. The production of NO was determined by measuring nitrate concentrations. Data (three independent experiments performed in triplicate) are presented as mean values + SD. (*p < 0.05; **p < 0.01; one-way ANOVA with Bonferroni correction) To get a more detailed picture which member(s) of the LPA receptor family is/are responsible for signal transmission in BV-2 and primary microglia, we used the pan LPA receptor inhibitor BrP-LPA and the LPA5 inhibitor TCLPA5. TCLPA5 was chosen since LPA5 was identified as a member of the microglia sensome [48]. To the best of our knowledge, no LPA6 inhibitor is currently commercially available. In BV-2 cells, BrP-LPA reduced LPA-induced COX2 expression (statistically not significant) and increased Arg-1 signals (Fig. 7a, upper panel). COX-2 activation by LPA was reduced by TCLPA5 at both time points analyzed (Fig. 7a; lower panel). During these experiments, LPA activation was also performed in the presence of DMSO to account for potential inadvertent effects mediated by the vehicle. Bar graphs at the right show densitometry of immunoreactive bands.Fig. 7 Inhibition of LPA5 suppresses the LPA-induced pro-inflammatory phenotype in BV-2 and primary murine microglia. a Serum-starved BV-2 cells were treated with LPA in the absence or presence of BrP-LPA (5 μM; upper panel) or TCLPA5 (5 μM; lower panel) added 2 h prior to LPA addition. COX-2 and Arg-1 response was monitored using Western blotting. One representative plot for each protein and the densitometric analysis (mean + SD) from four independent experiments is presented. (**p < 0.01; ***p < 0.001; # p < 0.05, inhibitor compared to LPA-treated cells; one-way ANOVA with Bonferroni correction). b PMM were incubated in the presence of vehicle (DMSO; 'untreated'), LPA (1 μM), vehicle (DMSO) plus LPA (1 µM), or TCLPA5 (5 μM in DMSO; added 2 h prior to LPA addition) plus LPA (1 µM) for 24 h. Cells were stained for iNOS, COX-2, Arg-1, or RELMα and visualized using confocal microscopy. Fluorescence intensity was quantitated with ImageJ. At least 50 cells out of 3 different areas per chamber were measured. Results (three independent experiments) are presented as mean + SD (*p < 0.05; ***p < 0.001; # p < 0.05 inhibitor compared to LPA-treated cells; one-way ANOVA with Bonferroni correction) To confirm the involvement of LPA5 in PMM, we performed immunofluorescence studies (Fig. 7b). These experiments revealed the expected induction of iNOS and COX-2 in response to LPA, while the M2 markers Arg-1 and RELMα were decreased. In response to TCLPA5, iNOS and COX-2 expression was significantly reduced with Arg-1 and RELMα being unaffected. The right panel shows fluorescence intensities of the corresponding micrographs. Surface marker expression analyses in LPA-stimulated BV-2 in the absence or presence of BrP-LPA and TCLPA5 is shown in Fig. 8. CD40 expression was lower in the presence of BrP-LPA (though not significant) whereas CD86 levels were significantly decreased at all time points analyzed (Fig. 8a). The percentage of CD206+ cells was unaffected by BrP-LPA. The presence of TCLPA5 during LPA activation significantly reduced the CD40 and CD86 positive cell populations to baseline levels at all time points analyzed while CD206 was unaffected (Fig. 8b).Fig. 8 LPA receptor antagonists attenuate M1 surface marker expression in BV-2 cells. Serum-starved (o/n) cells were cultivated in the presence of vehicle, LPA (1 μM), or LPA plus a BrP-LPA (5 μM) and b TCLPA5 (5 μM) for the indicated times. Inhibitors were added 2 h prior LPA addition. Cells were stained with PE-conjugated anti-CD40, APC-conjugated anti-CD86, or PE-conjugated anti-CD206 antibodies and analyzed using a Guava easyCyte 8 Millipore flow cytometer. Results from four individual experiments in triplicate are shown as mean values + SD. (*p < 0.05; **p < 0.01; ***p < 0.001 compared to untreated or DMSO-treated cells; # p < 0.05; ## p < 0.01; ### p < 0.001 inhibitor compared to LPA-treated cells; one-way ANOVA with Bonferroni correction) Cytokine secretion in response to LPA in BV-2 cells (Fig. 9a and PMM (Fig. 9b showed a general tendency to be reduced when TCLPA5 was present (Fig. 9). Secretion of all cyto-/chemokines was significantly reduced in BV-2 at one (TNFα, CCL5, and CXCL2) or two (IL-1β, IL-6) time points (Fig. 9a). In contrast, TCLPA5 inhibited secretion of IL-1β, TNFα, and IL-6 but was without effect on CCL5 and CXCL2 in PMM (Fig. 9b).Fig. 9 TCLPA5 inhibits the secretion of pro-inflammatory cytokines and chemokines. a BV-2 and b PMM were cultured on 24-well plates and serum-starved o/n. The supernatants were collected after incubation with vehicle, 1 μM LPA, or LPA plus TCLPA5 (5 μM) for the indicated times. The concentrations of IL-1β, TNFα, IL-6, CCL5 (RANTES), and CXCL2 (MIP-2) were quantitated by ELISA. Results shown represent mean + SD from two independent experiments performed in triplicate (*p < 0.05; **p < 0.01; ***p < 0.001 compared to vehicle control; # p < 0.05, ## p < 0.01; ### p < 0.001 TCLPA5 compared to LPA treated cells; one-way ANOVA with Bonferroni correction) Finally, we determined the effects of TCLPA5 on ROS and NO formation in both cell types. TCLPA5 significantly reduced LPA-mediated DCF fluorescence at 6 h in BV-2 and primary microglia (Fig. 10a). Nitrate concentrations were reduced at 48 (BV-2) and 24 h (PMM) by TCLPA5 (Fig. 10b).Fig. 10 TCLPA5 suppresses ROS and NO production. a The intracellular ROS levels generated by BV-2 and PMM were determined using carboxy-H2DCFDA. Serum-starved BV-2 and PMM were incubated with carboxy-H2DCFDA, treated with vehicle control, LPA (1 μM), or LPA plus TCLPA5 (5 μM) for the indicated time periods, and the fluorescence intensity was evaluated. Results (three independent experiments performed in triplicate) are presented as mean values + SD. (**p < 0.01; ***p < 0.001 compared to vehicle; # p < 0.05 compared to LPA treated cells; one-way ANOVA with Bonferroni correction). b Serum-starved BV-2 and PMM were incubated with vehicle (DMSO), LPA (1 μM), or LPA plus TCLPA5 (5 μM) for the indicated times, and the production of NO was determined by measuring the total nitrate concentration in the supernatants. Data (three independent experiments performed in triplicate) are presented as mean values + SD. (*p < 0.05; compared to untreated cells; # p < 0.05 compared to LPA treated cells; one-way ANOVA with Bonferroni correction) Discussion In the present study, we provide first evidence that LPA polarizes BV-2 and PMM towards an M1-like phenotype. Of note, the responses in BV-2 and primary microglia were qualitatively comparable, indicating that BV-2 cells represent a suitable prescreening model. Our findings might have bearings in neurological disease settings where LPA levels are increased, e.g., in spinal cord injury, traumatic brain injury, or multiple sclerosis [21]. LPA receptor profiling in the present study confirmed low expression of the classical LPA1-3 receptors [28] (and LPA4) but revealed high expression of LPA5 and LPA6. All of the LPA receptors are expressed in the developing brain and expression levels vary with developmental age [21]. LPA-mediated processes regulate proliferation, microtubule-dependent interkinetic nuclear migration, neurite retraction, cell survival, morphological changes, and cell migration [21]. Thus, in physiological conditions, LPA-mediated signaling contributes to normal development and function of the CNS. However, in response to injury, LPA levels can rise significantly in the brain and CSF [8, 17, 19, 49, 50]. LPA levels are elevated in human (0.05 controls vs. 0.27 μM post injury) and mouse (0.8 and 2 μM, prior vs. post injury) CSF in response to traumatic brain injury [51]. In this context, it is of importance that exogenous LPA can fuel endogenous LPA production via an LPA3-dependent pathway [52]. Overshooting LPA signaling has been linked to the development of fetal hydrocephalus in embryonic mice, a pathophysiological process that is ameliorated in LPA1/LPA2 double knockout animals [50]. LPA signaling is also involved in nerve injury-triggered pain responses [53], where LPA1 [54] and LPA5 [55] contribute via independent mechanisms. Findings that LPA5 is activated during nerve injury (but not under basal conditions) are consistent with the fact that LPA levels rise significantly in response to spinal cord injury [8, 19]. In the contused spinal cord, parenchymal LPA concentrations increase from 75 to 725 pmol/mg protein (naïve vs. 3 days post injury) and contribute to secondary injury manifested as demyelination [19]. Demyelination in the injured spinal cord was (at least in part) ascribed to LPA-activated microglia [19]. Lysophosphatidylcholine injected intrathecally is converted to LPA via ATX-mediated pathways, and an LPA3-dependent feed-forward loop induces further endogenous synthesis of LPA [52]. It was suggested that within this setting, microglia activation is responsible for de novo LPA synthesis and concomitant development of neuropathic pain [56]. Thus, findings of the present study that LPA induces an M1-like microglial phenotype are relevant to pathophysiology in the injured/diseased CNS. Here we show that LPA induces the expression of an M1 signature in BV-2 and primary microglia. In line with a previous study [47], LPS induced iNOS, COX-2, CD40, and CD86 expression. LPA treatment resulted in the upregulation of the M1 markers iNOS, COX-2, CD40, and CD86 and downregulation of the M2 markers CD206 (MRC1), Arg-1, and Relmα in BV-2 and primary microglia. The transcriptional programs that drive an LPA-mediated M1 signature are currently under investigation. Classical M1 marker expression was accompanied by increased cytokine/chemokine mRNA and protein levels as well as ROS and NO production in BV-2 and PMM. In PMM, LPS increased iNOS and COX-2 expression as well as IL-6 release [47] in a similar manner as observed here for LPA-treated BV-2 cells and PMM. Also, the temporal expression profiles are in agreement to what was reported for M1 marker expression in LPS-treated PMM: Gene expression of iNOS and COX-2 was significantly elevated between 4 and 72 h, and protein levels of IL-6 were significantly elevated over baseline up to 72 h [47]. The 2.5-fold increase in IL-6 secretion of primary microglia in response to LPA observed here is in a comparable range reported for LPA-stimulated fibroblast-like synoviocytes [57], which play an active role in synovial inflammation and damage via ATX/LPA-mediated pathways. iNOS is not expressed in the healthy brain, but expression is induced in response to inflammatory mediators like LPS or cytokines. Increased ROS and NO concentrations make it reasonable to assume that, in response to iNOS upregulation, excess NO reacts with NADPH oxidase-derived O2−. This reaction results in the formation of the highly neurotoxic peroxynitrite (ONOO−) in BV-2 microglia [58]. It is important to note that DCF (used during the present study to detect alterations in cellular redox balance) is not a species-specific probe but is, in addition to H2O2, also oxidized by hypochlorous acid (generated via the myeloperoxidase/H2O2/chloride system), other peroxidases, and ONOO− [59]. ONOO− was shown to induce mitochondrial dysfunction in neurons [60] to damage oligodendrocytes [61] and to compromise blood-brain barrier function [62]. Although our results suggest that LPA stimulation leads to ROS and NO production and is potentially neurotoxic, Awada and colleagues [63] have shown that overexpression of ATX in BV-2 microglia protects cells against H2O2-induced cell damage and oxidative stress. The same group [64] demonstrated ATX-mediated downregulation of cytokine production (mRNA and protein) in LPS-stimulated BV-2 cells. These seemingly contradictory results to the present study might be simply due to different incubation/culture conditions: In the ATX overexpression model [63], BV-2 cells are continuously exposed to LPA concentrations that are approx. fourfold elevated over the vector controls in contrast to the single addition used in the present study. Although we show that exposure to a single LPA bolus does not change LPA receptor expression in BV-2 cells and PMM, the situation might be different in ATX-overexpressing microglia. In terms of downregulated cytokine production in ATX-overexpressing BV-2 cells, Awada et al. [64] used LPS-stimulated cells while we studied effects of LPA on BV-2 and PMM that were exposed only to LPA (in the absence of a co-agonist). This is reminiscent of what was reported for peritoneal macrophages [65]: In that study, LPA induced IL-6 but not TNFα secretion in unstimulated macrophages while in LPS-stimulated cells, LPA downregulated TNFα but not IL-6 production. Thus, it appears that LPA-mediated effects depend on the cellular preactivation experience resulting in altered responsiveness upon rechallenge probably related to the intrinsic immune memory of microglia [66]. Time-dependent gene/protein expression of cytokines/chemokines have profiles specific for an M1 or M2 microglia phenotype [47]. We observed upregulated expression of IL-1β, IL-6, Tnfα, and the immunomodulatory chemokines CCL5 and CXCL2 in response to LPA treatment, findings reminiscent of what was reported for LPS-activated PMM [47]. In particular, increased IL-1β, IL-6, and TNFα concentrations were linked to a poor prognosis in infants suffering from ischemic encephalopathy [67]. In addition, IL-1β, IL-6, TNFα, and the chemokines CCL5 and CXCL2 (which were all upregulated by LPA in primary microglia) are implicated as modulators of the neuroinflammatory response during traumatic brain injury [68] where LPA levels are increased [51]. To get first indications about LPA receptors that are responsible for signal transmission in microglia, we have performed pharmacological inhibitor studies. BrP-LPA is a pan LPA receptor/ATX inhibitor [44] while TCLPA5 is a specific antagonist of LPA5 [45]. Of note, both inhibitors suppressed M1 marker expression in BV-2 and PMM and attenuated LPA-stimulated cytokine secretion. Since TCLPA5 blunted all pro-inflammatory signals, it is conceivable that LPA5 is a major player in LPA-dependent M1 polarization of microglia. Our findings are in line with LPA5-mediated signaling cascades in immunocompetent cells including a sensome function in microglia [48]. Using a novel chemical probe acting as specific antagonist for LPA5, it was shown that this receptor induces Ca2+ release from LPA-stimulated BV-2 cells in response to hexadecyl-LPA [69]. In line with results of the present study, Kozian and colleagues [69] demonstrated highest expression for LPA5 in BV-2 cells (LPA6 is not mentioned in this article) and reported a sub μM IC50 (730 nM) for this novel LPA5 antagonist for Ca2+ release in 16:0 alkyl-LPA stimulated BV-2 cells. Currently, no published data are available whether commercially available TCLPA5 or the LPA5 antagonist described in [69] crosses the blood-brain or blood-cerebrospinal fluid barriers to foster application in CNS disease models. In human mast cells, LPA5 is essential for conveying signals leading to MIP-1β (CCL4) generation and secretion underlining the importance of this signaling route as regulator of cyto- and/or chemokine production [70]. Based on the present results (graphically summarized in Fig. 11), we hypothesize that interference with the LPA/LPA5 signaling axis might provide an opportunity to pharmacologically shift microglia polarization. However, the therapeutic usefulness of such an approach has to be carefully evaluated in in vivo studies.Fig. 11 Graphical summary of findings obtained during the present study. LPA treatment induces a unique pro-inflammatory M1-like signature in BV-2 and PMM that is completely or partially reversed by pre-treatment with the LPA5 inhibitor TCLPA5. Data of the present study provide functional evidence for the role of LPA5 as member of the microglia sensome Conclusions We conclude that our data provide functional support for a sensome function of LPA5 in BV-2 and PMM. The present in vitro study indicates that interference with the LPA signaling axis either at the level of LPA synthesis (using, e.g., ATX inhibitors) or at the level of signal transmission (LPA receptor antagonists) could offer new means to modulate the microglia polarization status. Abbreviations APCAllophycocyanin Arg-1Arginase 1 ATXAutotaxin BrP-LPA1-Bromo-3(S)-hydroxy-4-(palmitoyloxy)butyl-phosphonate CXCL2C-X-C motif chemokine 2 CCL4C-C motif chemokine 4 CCL5C-C motif chemokine 5 CNSCentral nervous system COX-2Cyclooxygenase 2 DCFDA2′,7′-dichlorofluorescin diacetate DMSODimethylsulfoxide IL-1βInterleukin-1β IL-6Interleukin-6 iNOSInducible nitric oxide synthase LPALysophosphatidic acid LPA1-6LPA receptors 1-6 MRC1Macrophage mannose receptor 1 ONOO−Peroxynitrite o/nOvernight PEPhycoerythrin PMMPrimary murine microglia qPCRQuantitative real-time PCR RELMaResistin-like alpha ROSReactive oxygen species TBHPtert-Butylhydroperoxide TCLPA55-(3-Chloro-4-cyclohexylphenyl)-1-(3-methoxyphenyl)-1H-pyrazole-3-carboxylic acid TNFαTumor necrosis factor α Acknowledgements We express our thanks to Drs. Fritz Andrae and Andreas Artl (both piCHEM Graz; http://www.pichem.at) for the help with Limulus testing and to Anja Feiner for the expert technical support. Funding Financial support was provided by the Austrian Science Fund (FWF; DK-MOLIN W1241 and DK-MCD W1226), the Medical University of Graz (supporting the PhD programs DK-MOLIN and DK-MCD), and BioTechMed-Graz. Availability of data and materials There is no data, software, databases, and application/tool available apart from the data reported in the present study. All data is provided in the manuscript. Authors’ contributions IP, EB, AH, DK, MG, and WS conceived and designed the experiments. IP, EB, MG, HR, VBK, NK, AW, AH, SS, and KJ performed the experiments. IP, EB, MG, AW, SS, KJ, AH, DK, and WS analyzed the data. IP, EB, EM, and WS 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 and consent to participate Not applicable. ==== Refs References 1. Streit WJ Microglia as neuroprotective, immunocompetent cells of the CNS Glia 2002 40 2 133 9 10.1002/glia.10154 12379901 2. Kettenmann H Kirchhoff F Verkhratsky A Microglia: new roles for the synaptic stripper Neuron 2013 77 1 10 8 10.1016/j.neuron.2012.12.023 23312512 3. Nimmerjahn A Kirchhoff F Helmchen F Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo Science 2005 308 5726 1314 8 10.1126/science.1110647 15831717 4. Lucin KM Wyss-Coray T Immune activation in brain aging and neurodegeneration: too much or too little? Neuron 2009 64 1 110 22 10.1016/j.neuron.2009.08.039 19840553 5. Aldskogius H Regulation of microglia—potential new drug targets in the CNS Expert Opin Ther Targets 2001 5 6 655 68 10.1517/14728222.5.6.655 12540276 6. Cherry JD Olschowka JA O’Banion MK Neuroinflammation and M2 microglia: the good, the bad, and the inflamed J Neuroinflammation 2014 11 1 98 10.1186/1742-2094-11-98 24889886 7. Aoki J Inoue A Okudaira S Two pathways for lysophosphatidic acid production Biochim Biophys Acta 2008 1781 9 513 8 10.1016/j.bbalip.2008.06.005 18621144 8. Ma L Uchida H Nagai J Inoue M Aoki J Ueda H Evidence for de novo synthesis of lysophosphatidic acid in the spinal cord through phospholipase A2 and autotaxin in nerve injury-induced neuropathic pain J Pharmacol Exp Ther 2010 333 2 540 6 10.1124/jpet.109.164830 20123931 9. Shindou H Hishikawa D Harayama T Yuki K Shimizu T Recent progress on acyl CoA: lysophospholipid acyltransferase research J Lipid Res 2009 50 Suppl S46 51 18931347 10. Moolenaar WH Perrakis A Insights into autotaxin: how to produce and present a lipid mediator Nat Rev Mol Cell Biol 2011 12 10 674 9 10.1038/nrm3188 21915140 11. Umezu-Goto M Kishi Y Taira A Hama K Dohmae N Takio K Autotaxin has lysophospholipase D activity leading to tumor cell growth and motility by lysophosphatidic acid production J Cell Biol 2002 158 2 227 33 10.1083/jcb.200204026 12119361 12. Tanaka M Okudaira S Kishi Y Ohkawa R Iseki S Ota M Autotaxin stabilizes blood vessels and is required for embryonic vasculature by producing lysophosphatidic acid J Biol Chem 2006 281 35 25822 30 10.1074/jbc.M605142200 16829511 13. van Meeteren LA Ruurs P Stortelers C Bouwman P van Rooijen MA Pradere JP Autotaxin, a secreted lysophospholipase D, is essential for blood vessel formation during development Mol Cell Biol 2006 26 13 5015 22 10.1128/MCB.02419-05 16782887 14. Fotopoulou S Oikonomou N Grigorieva E Nikitopoulou I Paparountas T Thanassopoulou A ATX expression and LPA signalling are vital for the development of the nervous system Dev Biol 2010 339 2 451 64 10.1016/j.ydbio.2010.01.007 20079728 15. Giganti A Rodriguez M Fould B Moulharat N Coge F Chomarat P Murine and human autotaxin alpha, beta, and gamma isoforms: gene organization, tissue distribution, and biochemical characterization J Biol Chem 2008 283 12 7776 89 10.1074/jbc.M708705200 18175805 16. Triebl A Trotzmuller M Eberl A Hanel P Hartler J Kofeler HC Quantitation of phosphatidic acid and lysophosphatidic acid molecular species using hydrophilic interaction liquid chromatography coupled to electrospray ionization high resolution mass spectrometry J Chromatogr A 2014 1347 104 10 10.1016/j.chroma.2014.04.070 24813932 17. Tigyi G Hong L Yakubu M Parfenova H Shibata M Leffler CW Lysophosphatidic acid alters cerebrovascular reactivity in piglets Am J Phys 1995 268 5 Pt 2 H2048 55 18. Sun GY Lu FL Lin SE Ko MR Decapitation ischemia-induced release of free fatty acids in mouse brain. Relationship with diacylglycerols and lysophospholipids Mol Chem Neuropathol 1992 17 1 39 50 10.1007/BF03159980 1388450 19. Santos-Nogueira E Lopez-Serrano C Hernandez J Lago N Astudillo AM Balsinde J Activation of lysophosphatidic acid receptor type 1 contributes to pathophysiology of spinal cord injury J Neurosci 2015 35 28 10224 35 10.1523/JNEUROSCI.4703-14.2015 26180199 20. Choi JW Chun J Lysophospholipids and their receptors in the central nervous system Biochim Biophys Acta 2013 1831 1 20 32 10.1016/j.bbalip.2012.07.015 22884303 21. Yung YC Stoddard NC Mirendil H Chun J Lysophosphatidic acid signaling in the nervous system Neuron 2015 85 4 669 82 10.1016/j.neuron.2015.01.009 25695267 22. Kihara Y Maceyka M Spiegel S Chun J Lysophospholipid receptor nomenclature review: IUPHAR Review 8 Br J Pharmacol 2014 171 15 3575 94 10.1111/bph.12678 24602016 23. On NH Savant S Toews M Miller DW Rapid and reversible enhancement of blood-brain barrier permeability using lysophosphatidic acid J Cereb Blood Flow Metab 2013 33 12 1944 54 10.1038/jcbfm.2013.154 24045401 24. Hecht JH Weiner JA Post SR Chun J Ventricular zone gene-1 (vzg-1) encodes a lysophosphatidic acid receptor expressed in neurogenic regions of the developing cerebral cortex J Cell Biol 1996 135 4 1071 83 10.1083/jcb.135.4.1071 8922387 25. Trimbuch T Beed P Vogt J Schuchmann S Maier N Kintscher M Synaptic PRG-1 modulates excitatory transmission via lipid phosphate-mediated signaling Cell 2009 138 6 1222 35 10.1016/j.cell.2009.06.050 19766573 26. Weiner JA Hecht JH Chun J Lysophosphatidic acid receptor gene vzg-1/lpA1/edg-2 is expressed by mature oligodendrocytes during myelination in the postnatal murine brain J Comp Neurol 1998 398 4 587 98 10.1002/(SICI)1096-9861(19980907)398:4<587::AID-CNE10>3.0.CO;2-5 9717712 27. Bernhart E Kollroser M Rechberger G Reicher H Heinemann A Schratl P Lysophosphatidic acid receptor activation affects the C13NJ microglia cell line proteome leading to alterations in glycolysis, motility, and cytoskeletal architecture Proteomics 2010 10 1 141 58 10.1002/pmic.200900195 19899077 28. Moller T Contos JJ Musante DB Chun J Ransom BR Expression and function of lysophosphatidic acid receptors in cultured rodent microglial cells J Biol Chem 2001 276 28 25946 52 10.1074/jbc.M102691200 11340076 29. Schilling T Stock C Schwab A Eder C Functional importance of Ca2 + -activated K+ channels for lysophosphatidic acid-induced microglial migration Eur J Neurosci 2004 19 6 1469 74 10.1111/j.1460-9568.2004.03265.x 15066143 30. Muessel MJ Harry GJ Armstrong DL Storey NM SDF-1alpha and LPA modulate microglia potassium channels through rho gtpases to regulate cell morphology Glia 2013 61 10 1620 8 10.1002/glia.22543 23893870 31. Fujita R Ma Y Ueda H Lysophosphatidic acid-induced membrane ruffling and brain-derived neurotrophic factor gene expression are mediated by ATP release in primary microglia J Neurochem 2008 107 1 152 60 10.1111/j.1471-4159.2008.05599.x 18680554 32. Cyster JG Schwab SR Sphingosine-1-phosphate and lymphocyte egress from lymphoid organs Annu Rev Immunol 2012 30 69 94 10.1146/annurev-immunol-020711-075011 22149932 33. Kanda H Newton R Klein R Morita Y Gunn MD Rosen SD Autotaxin, an ectoenzyme that produces lysophosphatidic acid, promotes the entry of lymphocytes into secondary lymphoid organs Nat Immunol 2008 9 4 415 23 10.1038/ni1573 18327261 34. Nakasaki T Tanaka T Okudaira S Hirosawa M Umemoto E Otani K Involvement of the lysophosphatidic acid-generating enzyme autotaxin in lymphocyte-endothelial cell interactions Am J Pathol 2008 173 5 1566 76 10.2353/ajpath.2008.071153 18818380 35. Knowlden S Georas SN The autotaxin-LPA axis emerges as a novel regulator of lymphocyte homing and inflammation J Immunol 2014 192 3 851 7 10.4049/jimmunol.1302831 24443508 36. Bai Z Cai L Umemoto E Takeda A Tohya K Komai Y Constitutive lymphocyte transmigration across the basal lamina of high endothelial venules is regulated by the autotaxin/lysophosphatidic acid axis J Immunol 2013 190 5 2036 48 10.4049/jimmunol.1202025 23365076 37. Chen R Roman J Guo J West E McDyer J Williams MA Lysophosphatidic acid modulates the activation of human monocyte-derived dendritic cells Stem Cells Dev 2006 15 6 797 804 10.1089/scd.2006.15.797 17253943 38. Saijo K Glass CK Microglial cell origin and phenotypes in health and disease Nat Rev Immunol 2011 11 11 775 87 10.1038/nri3086 22025055 39. Zhao X Wang H Sun G Zhang J Edwards NJ Aronowski J Neuronal Interleukin-4 as a modulator of microglial pathways and ischemic brain damage J Neurosci 2015 35 32 11281 91 10.1523/JNEUROSCI.1685-15.2015 26269636 40. Orihuela R McPherson CA Harry GJ Microglial M1/M2 polarization and metabolic states Br J Pharmacol 2015 25800044 41. Deierborg T Preparation of primary microglia cultures from postnatal mouse and rat brains Methods Mol Biol 2013 1041 25 31 10.1007/978-1-62703-520-0_4 23813366 42. 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 4 402 8 10.1006/meth.2001.1262 11846609 43. McCall MN McMurray HR Land H Almudevar A On non-detects in qPCR data Bioinformatics 2014 30 16 2310 6 10.1093/bioinformatics/btu239 24764462 44. Yung YC Stoddard NC Chun J LPA receptor signaling: pharmacology, physiology, and pathophysiology J Lipid Res 2014 24643338 45. Kozian DH Evers A Florian P Wonerow P Joho S Nazare M Selective non-lipid modulator of LPA5 activity in human platelets Bioorg Med Chem Lett 2012 22 16 5239 43 10.1016/j.bmcl.2012.06.057 22801643 46. Halliwell B Whiteman M Measuring reactive species and oxidative damage in vivo and in cell culture: how should you do it and what do the results mean? Br J Pharmacol 2004 142 2 231 55 10.1038/sj.bjp.0705776 15155533 47. Chhor V Le Charpentier T Lebon S Ore MV Celador IL Josserand J Characterization of phenotype markers and neuronotoxic potential of polarised primary microglia in vitro Brain Behav Immun 2013 32 70 85 10.1016/j.bbi.2013.02.005 23454862 48. Hickman SE Kingery ND Ohsumi TK Borowsky ML Wang LC Means TK The microglial sensome revealed by direct RNA sequencing Nat Neurosci 2013 16 12 1896 905 10.1038/nn.3554 24162652 49. Savaskan NE Rocha L Kotter MR Baer A Lubec G van Meeteren LA Autotaxin (NPP-2) in the brain: cell type-specific expression and regulation during development and after neurotrauma Cell Mol Life Sci 2007 64 2 230 43 10.1007/s00018-006-6412-0 17192809 50. Yung YC Mutoh T Lin ME Noguchi K Rivera RR Choi JW Lysophosphatidic acid signaling may initiate fetal hydrocephalus Sci Transl Med 2011 3 99 99ra87 10.1126/scitranslmed.3002095 21900594 51. Crack PJ Zhang M Morganti-Kossmann MC Morris AJ Wojciak JM Fleming JK Anti-lysophosphatidic acid antibodies improve traumatic brain injury outcomes J Neuroinflammation 2014 11 37 10.1186/1742-2094-11-37 24576351 52. Ma L Uchida H Nagai J Inoue M Chun J Aoki J Lysophosphatidic acid-3 receptor-mediated feed-forward production of lysophosphatidic acid: an initiator of nerve injury-induced neuropathic pain Mol Pain 2009 5 64 10.1186/1744-8069-5-64 19912636 53. Ueda H Matsunaga H Olaposi OI Nagai J Lysophosphatidic acid: chemical signature of neuropathic pain Biochim Biophys Acta 2013 1831 1 61 73 10.1016/j.bbalip.2012.08.014 22960381 54. Inoue M Rashid MH Fujita R Contos JJ Chun J Ueda H Initiation of neuropathic pain requires lysophosphatidic acid receptor signaling Nat Med 2004 10 7 712 8 10.1038/nm1060 15195086 55. Lin ME Rivera RR Chun J Targeted deletion of LPA5 identifies novel roles for lysophosphatidic acid signaling in development of neuropathic pain J Biol Chem 2012 287 21 17608 17 10.1074/jbc.M111.330183 22461625 56. Ma L Nagai J Ueda H Microglial activation mediates de novo lysophosphatidic acid production in a model of neuropathic pain J Neurochem 2010 115 3 643 53 10.1111/j.1471-4159.2010.06955.x 20722972 57. Zhao C Fernandes MJ Prestwich GD Turgeon M Di Battista J Clair T Regulation of lysophosphatidic acid receptor expression and function in human synoviocytes: implications for rheumatoid arthritis? Mol Pharmacol 2008 73 2 587 600 10.1124/mol.107.038216 18006645 58. Kumar A Chen SH Kadiiska MB Hong JS Zielonka J Kalyanaraman B Inducible nitric oxide synthase is key to peroxynitrite-mediated, LPS-induced protein radical formation in murine microglial BV2 cells Free Radic Biol Med 2014 73 51 9 10.1016/j.freeradbiomed.2014.04.014 24746617 59. Tarpey MM Wink DA Grisham MB Methods for detection of reactive metabolites of oxygen and nitrogen: in vitro and in vivo considerations Am J Physiol Regul Integr Comp Physiol 2004 286 3 R431 44 10.1152/ajpregu.00361.2003 14761864 60. Ebadi M Sharma SK Peroxynitrite and mitochondrial dysfunction in the pathogenesis of Parkinson’s disease Antioxid Redox Signal 2003 5 3 319 35 10.1089/152308603322110896 12880486 61. Baud O Li J Zhang Y Neve RL Volpe JJ Rosenberg PA Nitric oxide-induced cell death in developing oligodendrocytes is associated with mitochondrial dysfunction and apoptosis-inducing factor translocation Eur J Neurosci 2004 20 7 1713 26 10.1111/j.1460-9568.2004.03616.x 15379992 62. Ding R Chen Y Yang S Deng X Fu Z Feng L Blood-brain barrier disruption induced by hemoglobin in vivo: Involvement of up-regulation of nitric oxide synthase and peroxynitrite formation Brain Res 2014 1571 25 38 10.1016/j.brainres.2014.04.042 24814387 63. Awada R Rondeau P Gres S Saulnier-Blache JS Lefebvre d’Hellencourt C Bourdon E Autotaxin protects microglial cells against oxidative stress Free Radic Biol Med 2012 52 2 516 26 10.1016/j.freeradbiomed.2011.11.014 22155714 64. Awada R Saulnier-Blache JS Gres S Bourdon E Rondeau P Parimisetty A Autotaxin downregulates LPS-induced microglia activation and pro-inflammatory cytokines production J Cell Biochem 2014 115 12 2123 32 10.1002/jcb.24889 25053164 65. Fan H Zingarelli B Harris V Tempel GE Halushka PV Cook JA Lysophosphatidic acid inhibits bacterial endotoxin-induced pro-inflammatory response: potential anti-inflammatory signaling pathways Mol Med 2008 14 7-8 422 8 10.2119/2007-00106.Fan 18431464 66. Kettenmann H Hanisch UK Noda M Verkhratsky A Physiology of microglia Physiol Rev 2011 91 2 461 553 10.1152/physrev.00011.2010 21527731 67. Aly H Khashaba MT El-Ayouty M El-Sayed O Hasanein BM IL-1beta, IL-6 and TNF-alpha and outcomes of neonatal hypoxic ischemic encephalopathy Brain Dev 2006 28 3 178 82 10.1016/j.braindev.2005.06.006 16181755 68. Gyoneva S Ransohoff RM Inflammatory reaction after traumatic brain injury: therapeutic potential of targeting cell-cell communication by chemokines Trends Pharmacol Sci 2015 36 7 471 80 10.1016/j.tips.2015.04.003 25979813 69. Kozian DH von Haeften E Joho S Czechtizky W Anumala UR Roux P Modulation of hexadecyl-LPA-mediated activation of mast cells and microglia by a chemical probe for LPA5 Chembiochem 2016 17 9 861 5 10.1002/cbic.201500559 26812365 70. Lundequist A Boyce JA LPA5 is abundantly expressed by human mast cells and important for lysophosphatidic acid induced MIP-1beta release PLoS One 2011 6 3 e18192 10.1371/journal.pone.0018192 21464938

PMC005xxxxxx/PMC5002166.txt

==== Front BMC Oral HealthBMC Oral HealthBMC Oral Health1472-6831BioMed Central London 27810.1186/s12903-016-0278-9Research ArticleShould oral anticoagulant therapy be continued during dental extraction? A meta-analysis Yang Shuo sophiasure@163.com 1Shi Quan shiquan3333@sina.cn 1Liu Jinglong summer_mimosa@hotmail.com 1Li Jinru 308928377@qq.com 2Xu Juan newxj@hotmail.com 11 Department of Stomatology, Chinese People’s Liberation Army General Hospital, 28 Fuxing Road, Beijing, 100853 China 2 Department of Stomatology, Chinese People’s Liberation Army 322 Hospital, 2 Yunzhong Road, Datong, 037000 China 26 8 2016 26 8 2016 2016 16 1 8122 6 2016 12 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 Oral anticoagulation therapy is widely used to reduce the risks of thromboembolism. However, the therapy increases the risk of hemorrhage during the surgical procedures. The aim of this meta-analysis was to evaluate the bleeding risk of patients continuing or discontinuing oral anticoagulant therapy while undergoing dental extractions. Methods Six electronic databases, including PubMed, Embase, Cochrane library, Web of Science, China Biology Medicine disc (CBM), and China National Knowledge Infrastructure (CNKI), were searched in March, 2016. Relevant articles were screened by two independent reviewers under our inclusion criteria. Quality was evaluated using the Cochrane Collaboration risk of bias tool. Meta-analyses were conducted with fixed and random effects models as appropriate. Results Six studies (with a total of 591 patients) were included in our meta-analysis. Our results showed that there was no significant difference in the bleeding risk between patients continuing or discontinuing oral anticoagulant therapy while undergoing dental extractions (risk ratio, 1.31; 95 % CI, 0.79, 2.14; P > 0.05). There was also no significant difference in bleeding risk 1 day (risk ratio, 0.91; 95 % CI, 0.35, 2.37; P > 0.05) and 7 days (risk ratio, 1.47; 95 % CI, 0.83, 2.59; P > 0.05) after the dental extraction. Conclusion Under current studies and evidence, it appears that patients continuing oral anticoagulant therapy do not have an increased risk of bleeding after dental extractions compared to patients who discontinue oral anticoagulant therapy. Keywords Oral anticoagulant therapyDental extractionBleedingMeta-analysisNational Natural Science Foundation of China emergency management project81541111Xu Juan Clinical scientific research fund from Wu Jieping Medical Foundation320.6750.15029Xu Juan issue-copyright-statement© The Author(s) 2016 ==== Body Background With the living standard gradually increasing and diet habits continuously changing, cardiovascular diseases, such as mechanical heart valve, atrial fibrillation and venous thromboembolism, have become more and more common. Millions of people receive oral anticoagulant therapy (OAT). Used to reduce thrombosis, oral anticoagulation therapy is one of the most effective prophylactic medications for preventing life-threatening events [1, 2]. Oral anticoagulant drugs mainly include heparin, warfarin and new oral anticoagulants (NOACs) [3]. Heparin should be administered intravenously, and its action is to interfere with thrombin-antithrombin pathways and reduce fibrin formation. Warfarin, which is derived from 4-hydroxycoumarin, is a competitive inhibitor of vitamin K. It is widely used all over the world [4]. New oral anticoagulants (NOACs) include factor Xa inhibitors, such as rivaroxaban, apixaban, and edoxaban, which mainly inhibit the factor Xa activity of the prothrombinase complex in the propagation phase, and factor IIa inhibitors such as dabigatran. Although their advantages included a rapid onset and a short half-life, NOACs are still not ready to completely replace conventional anticoagulants [3]. However, with the wide use of oral anticoagulant therapy (OAT), a major disadvantage of OAT that should receive more attention, is the increased risk of hemorrhage during surgical procedures. OATs are frequently prescribed to elder patients, who also have a higher demand for dental extraction caused by caries, periodontitis or other dental diseases. So, the problem of postoperative bleeding becomes a major concern for the dentist. Whether to continue or discontinue OAT has become a strong controversy. If OAT is continued, there is a high risk of bleeding, and if OAT is discontinued, the thromboembolic complications are potentially deadly. Discontinuing OAT 2 or 3 days before oral surgery has been a widely used strategy for managing patients on warfarin [5, 6]. However,the risk of postoperative bleeding apparently does not decrease, and the thromboembolic risk increases [7]. There have been documented incidences of thromboembolic events when warfarin was stopped prior to a dental procedure or a minor oral surgery [8, 9]. Garcia et al. [10] published the first important prospective study on the risk of thromboembolism with a short-term interruption of OAT by reporting that the incidence of thromboembolism within a 30-day follow-up period was 0.5 %. Though the incidence of thromboembolism is low, the consequence may be deadly. These thromboembolic events can have devastating clinical consequences, such as an embolic stroke, which can result in major disability or death, or myocardial ischaemia, which can increase risk of death two to four fold [11]. To solve this problem, some strategies have been developed in the last few decades. These strategies include reducing the dose of the anticoagulant drugs [12, 13] or bridging it with heparin [14, 15]. These two methods, however, do not completely eliminate the risk of thromboembolic events, such as stroke [9]. Currently, there are a lot of publications suggest that dental extractions may be carried out with no OAT interruption if local hemostasis is adequately maintained in the OAT patients [14, 16–20]. Nevertheless, there is still a lack of public recognition of the treatment guidelines for treating these patients. To address this controversy and to provide evidence-based recommendations, we have completed a meta-analysis. Articles on randomized controlled trials (RCTs) or controlled clinical trials (CCTs), including patients under OAT who undergo dental extractions were collected, and the postoperative bleeding was compared between patients who continued OAT and patients who discontinued OAT. Methods The methods for this review were based on the Cochrane Handbook for Systematic Reviews of Interventions. Throughout the whole process, the studies were assessed by 2 observers independently, and any disagreement was resolved by discussion. Database search A comprehensive search of the PubMed, EMBASE, Web of Science, Cochrane Library, China Biology Medicine disc (CBM), and China National Knowledge Infrastructure(CNKI) databases was conducted in March 2016. The search strategies for PubMed, EMBASE, Web of Science and Cochrane Library are shown in Table 1.Table 1 The key words, database and search result Database Key words Result Pubmed (oral anticoagulant therapy OR OAT OR anticoagulant*) AND (dental surgery OR dental extraction* OR tooth extraction*) 879 Embase (oral anticoagulant therapy OR OAT OR anticoagulant*) AND (dental surgery OR dental extraction* OR tooth extraction*) 634 Cochrane Library (oral anticoagulant therapy OR OAT OR anticoagulant*) AND (dental surgery OR dental extraction* OR tooth extraction*) 2 Web of Science (dental surgery OR dental extraction* OR tooth extraction*) AND (oral anticoagulant therapy OR OAT OR anticoagulant*) 664 Total 2179 Study selection Two reviewers independently evaluated all of the search results, and the inclusion criteria were as follows:Study design—studies were designed as randomized controlled trials (RCTs) or controlled clinical trials (CCTs) Participants—patients were receiving OAT and required dental extractions Comparators—the postoperative bleeding between patients who continue or discontinue OAT. Outcomes—postoperative bleeding (spontaneous bleeding, induced bleeding and minor bleeding) The exclusion criteria were as follows:In vitro study (laboratory studies or animal studies), case reports or letters. Studies on antiplatelet medications. Study outcomes were not clearly reported or the data could not be used for our meta-analysis. Data extraction The following parameters were extracted from each of the selected studies: the first author, country, year of publication, design type, patient characteristics, intervention method, and bleeding outcomes. The information is recorded in the table. Because there is no accepted standardized definition of bleeding outcomes for patients undergoing surgical procedures, we aimed to use the postoperative bleeding to summarize the bleeding outcome after thoroughly reading the included articles. The bleeding outcomes can be defined as two processes during the follow-up days. One is the patient’s own perceived bleeding, such as spontaneous bleeding that continues for more than 20 min [21] and oozing from the extraction site [7]. The other bleeding outcome is examined by doctors during the appointment, and the presence of a solid clot covering the extraction socket was considered as no bleeding, while the presence of a fresh clot that shed easily or oozing blood was considered to be positive bleeding [4]. Risk of bias evaluation The following seven items were taken into consideration: (1) allocation concealment, (2) random sequence generation, (3) blinding of participants and personnel, (4) blinding of outcome assessment, (5) incomplete outcome data, (6) selective reporting, and (7) other bias. The risk of bias for each item was judged as low risk, high risk, or unclear risk. The overall risk of bias for each study was evaluated by the following criteria: If the risk of bias was low for all the items, the study was of low risk. If one (or more than one) of the risk factors for bias was high for key items, the study was of high risk. If one (or more than one) of the risk factors for bias was unclear, the study had an unclear risk. Statistical analysis The bleeding outcomes of the included studies were combined, and a meta-analysis was performed using RevMan software (version 5.3). Subgroup analyses were performed once the included studies had the same evaluation intervals. The risk of bleeding for continuing or discontinuing OAT was expressed as a relative risk (RR) with an associated 95 % CI. In addition, a chi-square and I2 test were used to estimate the degree of heterogeneity, with values of 25, 50, and 75 % corresponding to the low, moderate, and high heterogeneity, respectively. Substantial heterogeneity was defined as a P value <0.05 and I2 >50 %. The fixed-effects model was applied when I2 <50 %, and the random-effects model was applied when I2 >50 %. Results Literature search The search process is shown in Fig. 1. There were 968 relevant studies identified during the database search, and after the exclusion of duplication as well as titles, abstracts, and full-text screening, finally, we included six studies [4, 7, 16, 21–23] in our meta-analysis. Five of the studies are in English, and the other one [22] is in Chinese.Fig. 1 Study flow diagram Data extraction The characteristics of the selected six studies are summarized in Tables 2 and 3. The studies were published between 1993 and 2010. Four of the six studies were RCT, and the other two were CCT. The total number of patients under OAT who received dental extractions was 581. The patients were divided into the following two groups: an OAT continued group and an OAT discontinued group. In the OAT continued group, patients did not stop their anticoagulant therapy when they needed dental extractions. In the OAT discontinued group, patients stopped their anticoagulant therapy 2 or 3 days before the dental extraction procedure and usually resumed the therapy on the same day. Four of the studies used warfarin as the anticoagulant drug, one used a vitamin K inhibitor and one had not report the anticoagulant. The details of the bleeding outcomes are shown in Table 3. The follow-up ranged from 1 to 7 days. The bleeding outcomes during patients’ review visits were recorded if the article mentioned.Table 2 Characteristics of included studies Study Country No.(T/C)a Design Anticoagulants Average age (mean ± SD or mean (range)) Borea (1993) [21] Italy 30(15/15) RCT NRb T: 62.7 ± 6.1 C: 61.1 ± 10 Gaspar (1997) [23] Israel 47(32/15) CCT Vitamin K inhibitor T: 61.1(34–85) C: 53.4(35–72) Devani (1998) [7] UK 65(33/32) CCT Warfarin T: 62.3(30–82) C: 61.3(32–81) Evans (2002) [16] UK 109(57/52) RCT Warfarin T: 67(36–92) C: 66 (30–93) Al-Mubarak (2007) [4] Saudi Arabia 214(110/114) RCT Warfarin T: 51.7 ± 14.7 53.1 ± 13.7 C: 52.3 ± 14.3 48.7 ± 13.1 Duan XQ (2010) [22] China 116(67/49) RCT warfarin T: 59.5 ± 12.6 C: 61.5 ± 11.3 a Ttreatment group, continue oral anticoagulant therapy group, C control group, discontinue oral anticoagulant therapy group b NRnot report Table 3 Characteristics of included studies Study Group INR (mean/range) Intervention method Number of extraction Follow up Bleeding outcomes Borea (1993) [21] T: continued OAT. C: discontinued OAT. T: 3–4.5 C: 1.5–2.5 T: Sutures and TA irrigation at surgery and MS for 7 day. C: Sutures and physiologic irrigation at surgery and MS for 7 days. Single dental 7 days Day 1: Spontaneous bleedings T: 1/15 C:2/15 Induced bleedings T:1/15 C:0/15 Day 2 to day 7: none Evans (2002) [16] T: continued OAT. C: discontinued OAT 2 days before extractions and resumed on the same day. T: 2.5(1.2–4.7) C: 1.6(1.2–2.3) All groups: Oxycellulose dressing and sutures for all patients. T: 2(1–7) C: 3(1–9) 7 days Postoperative bleeding T:15/57 C:7/52 Devani (1998) [7] T: continued OAT. C: discontinued OAT 2 days before extractions. T: 2.7(2.2–3.9) C: 1.6(1.2–2.1) All groups: Placing haemostatic pack and sutures. T: 2.1(1–9) C: 2(1–9) 5 days Minor bleeding: Day 1 T:0/33 C:0/32 Day 2 T:0/33 C:1/32 Day 3 T:1/33 C:0/32 Day 4 to day 5:none Gaspar (1997) [23] T: continued OAT C: discontinued OAT 3 days before and resumed the same day T: 2.5(1.9–3.5) C: 1.4(1.3–1.9) All groups: Sutures and TA irrigation at surgery and MS for 2 min.MS for 7 days postoperativedays NR 7 days Postoperative bleeding T:2/32 C:1/15 Al-Mubarak (2007) [4] T: continued OAT (group2 no sutures,group4 sutures). C: discontinued OAT 2 days before and resumed 12 h after dental extractions (group1 no sutures, group 3 sutures). T: 1.85(1.4–2.3) C:2.55(1.9–3.1) Local pressure (all) and sutures (group 3 and 4). All patients range 1–5, single extraction (63.3) two teeth (25 %), three teeth (7.5 %) four teeth (3.3 %) and five teeth (0.8 %). 7 days Postoperative bleeding T:8/110 C:7/114 Duan XQ (2010) [22] T: continued OAT. C: discontinued OAT 3 days before and resumed the next day. ALL: 1.80–2.67 All groups: Local pressure and sutures. T:1.5(1–5) C:1.3(1–5) 1 day Postoperative bleeding T:6/67 C:5/49 OAT oral anticoagulant therapy, T treatment group, C control group, INR international normalized ratio, MW mouthwash, TA tranexamic acid, NR not report Risk of bias evaluation The risk of bias summary is shown in Fig. 2. Of the six included studies, one [21] was judged to have a low risk of bias because all of the items had a low risk of bias. Five [4, 7, 16, 22, 23] of the studies were judged to have a high risk of bias because all of the studies failed to blind participants and personnel, and three [4, 22, 23] studies failed to describe the method of randomization and had no report of the allocation concealment.Fig. 2 Risk of bias of the studies. Summary of risk of bias for included studies. Green indicates a low risk of bias, yellow indicates an unclear risk of bias, and red indicates a high risk of bias Meta-analysis According to the postoperative bleeding outcomes, we performed a meta-analysis to compare the bleeding risk of the OAT continued and OAT discontinued groups. Six studies were included, with 314 subjects in the treated group and 277 in the control group. The incidence of postoperative bleeding was 10.8 % (34/314) in the OAT continued group and 8.30 % (23/277) in the OAT discontinued group. A fixed-effects model was applied because of the low heterogeneity across the studies. There was no significant difference in bleeding risk between the OAT continued group and OAT discontinued group (P = 0.29). The risk ratio was 1.31 and 95 % CI (0.79, 2.14) (Fig. 3).Fig. 3 Forest plot of the difference of postoperative bleeding between OAT continued or discontinued group Subgroup analyses were performed when three or more studies recorded the bleeding incidence using the same evaluation intervals. Two time points (1 and 7 days after the operation) fit the criteria. Three studies [7, 21, 22], with 115 subjects in the treated group and 96 in the control group were included in the 1 day subgroup. The bleeding occurred in 8 of 115 (6.9 %) patients in the treated group and in 7 of 96 (7.2 %) in the control group. A fixed-effects model was applied, and there was no significant difference between the two groups (P = 0.85). The risk ratio was 0.91 and 95 % CI (0.35, 2.37) (Fig. 4).Fig. 4 Forest plot of the difference of postoperative bleeding between OAT continued or discontinued group 1 day after the surgery Four studies [4, 16, 21, 23], with 214 subjects in the treated group and 196 in the control group were included in the 7 day subgroup. The bleeding occurred in 27 of 214 (12.6 %) patients in the treated group and in 17 of 196 (8.7 %) in the control group. A fixed-effects model was applied and there was no significant difference between the two groups (P = 0.19). The risk ratio was 1.47 and 95 % CI (0.83, 2.59) (Fig. 5).Fig. 5 Forest plot of the difference of postoperative bleeding between OAT continued or discontinued group 7 days after the surgery Discussion Numerous studies have reported that OAT reduces the risks of thromboembolism significantly [24, 25]. Other studies have documented that serious embolic complications, including death, are three times more likely to occur in patients whose anticoagulant therapy was interrupted, compared to patients with bleeding complications associated with anticoagulation [26]. For oral surgery procedures, no case of lethal postoperative bleeding in patients who continued OAT has been reported [16, 26], whereas several fatal thromboembolic events after stopping OAT for dental extractions have occurred [8, 27]. Nevertheless, there is still a lack of consensus regarding preoperative alteration of the anticoagulant regimen to prepare OAT patients for dental extractions. Summary of results The results of our meta-analysis revealed that patients continuing oral anticoagulant therapy do not have an increased risk of bleeding after dental extraction compared to patients who discontinue oral anticoagulant therapy. However, the follow-up dates were not the same between these studies. One of the studies [22] only reported the first day bleeding outcome, and one study [7] followed-up for 5 days with a detailed record of everyday situations. The other four studies [4, 16, 21, 23] followed up for 7 days after the dental extraction, but three studies only mentioned the total number of bleeding events and the remaining one recorded everyday situations. Subgroup analyses were performed since three of the included studies [7, 21, 22] recorded the bleeding outcome 1 day after surgery, and four included studies [4, 16, 21, 23] collected measurements for 7 days after the surgery. The results were consistent with the previous results. There was no significant difference in bleeding risk between the OAT continued group and the OAT discontinued group 1 and 7 days after the surgery. Of all the studies, none of these patients suffered serious bleeding, and the bleeding was easily controlled by patients at home or controlled with local measures during their visit to the hospital. No thromboembolic event was reported in these studies, whether OAT was continued or not. However, this outcome could be a result of the short follow-up period of these studies (from 1 to 7 days), which made it difficult assess the thromboembolism risk in patients who discontinued OAT. All of the 6 included studies came to the conclusion that OAT patients who do not discontinue the anticoagulant medication do not have a significantly higher risk of postoperative bleeding than OAT patients who stop the therapy. Some of the studies mentioned the importance of international normalized ratio (INR) and hemostatic procedures. INR has been used as a recommendation for monitoring patients’ oral anticoagulant therapy. For stroke prevention in atrial fibrillation patients, oral anticoagulation therapy that is dose adjusted to maintain an INR range of 2.0 to 3.0 is associated with a 64 % reduction in the risk of stroke compared to placebo [28]. And in patients suffering an acute venous thromboembolism (either deep vein thrombosis or pulmonary embolism), adjusted-dose OAT use significantly reduces the risk of recurrence of thrombotic events with a target INR range of 2.0–3.0. [29]. According to meta-analyses of atrial fibrillation or mixed populations assessing INR control and associated events [30], more than half of all thromboembolic events occurred when patients have an INR <2.0. Besides, the INR has a safe range before surgical procedures. This “safe range” is controversial because some experts recommended an INR of ≤3 [17, 31], whereas others suggest an INR of ≤4 [32, 33] as safe for dental extractions. An INR above 5 has been shown to be an unacceptable risk for postoperative bleeding [34]. However, none of the included patients had an INR of more than 5. Most of the patients had an INR no higher than 4, as shown in Table 3. Next to monitoring the INR, it is recommended to take a special care of patients with renal dysfunction. Given the renal excretion of drugs, renal dysfunction may result in a higher incidence of bleeding associated with oral anticoagulation [35, 36]. A variety of local hemostatic measures are used in oral surgeries, including sutures, local compression, adjuvants (such as fibrin and histoacryl glue), local antifibrinolytic solutions, collagen fleeces, acrylic splints, gelatin sponges and so on [37, 38]. Placing hemostatic, physiological or tranexamic acid irrigation and mouthwash were used in our included studies. However, one of the included studies [4] showed that suturing played no significant role in bleeding status. In their study, patients with sutures showed a higher incidence of bleeding than patients without sutures. This outcome might be caused by the mode of suturing that further traumatizes the soft tissue. Four of the included studies [7, 16, 21, 23] used local hemostatic measures, except for local compression and sutures. The effectiveness of these measures was worth studying. Limitations There were some limitations in the included studies. Only one of the six included studies was judged to have a low risk of bias, and the absence of randomization may predispose the analysis to a risk of selection bias and possible confounding effects [39]. Some studies were carried out with limited patient numbers [21, 23] and, as with all meta-analyses, there may be studies published in other languages or unpublished studies that we were not able to access. Additionally, the INR differed between different studies which influenced the bleeding outcomes. Patients with a low INR were less likely to suffer bleeding than patients with a higher INR. Other factors, such as gingival health, numbers and the complexity of the tooth extractions, the surgical skills, anesthesia type, material and suture technique, local hemostatic measures, and the use of anticoagulants differed from one study to another or were not reported. This heterogeneity limited the scientific evidence that could be obtained for this review. Conclusion Under current studies and evidence, the results of our meta-analysis revealed that patients continuing oral anticoagulant therapy do not increase the risk of bleeding after dental extractions compared to patients who discontinue oral anticoagulant therapy. However, with the limited number of included studies and the risk of bias, well-designed RCTs should be included with a larger sample size as well as specific inclusion and exclusion criteria. Moreover, it would be better if the bleeding outcomes had a reporting standard. This standardization would enhance the comparability of studies based on identical outcome measurements, which would help to establish guidelines for dentists who treat these patients with complex medical needs. Abbreviations NOACsNew oral anticoagulants OATOral anticoagulant therapy Acknowledgements This work was supported by a grant from the National Natural Science Foundation of China emergency management project (No. 81541111) and Clinical scientific research fund (No.320.6750.15029) from Wu Jieping Medical Foundation. Availability of data and material The authors declare that materials described in the manuscript, including all relevant raw data, will be freely available to any scientist wishing to use them for non-commercial purposes, without breaching participant confidentiality. We ensure that our datasets are either deposited in publicly available repositories (where available and appropriate) or presented in the main manuscript or additional supporting files, in machine-readable format (such as spreadsheets rather than PDFs) whenever possible. Authors’ contributions YS carried out the literature research and drafted the manuscript. YS, SQ, LJL and LJR gathered the information and performed the statistical analysis. XJ is the Corresponding author and she designed this meta-analysis, and helped to draft 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. Mok CK Boey J Wang R Chan TK Cheung KL Lee PK Chow J Ng RP Tse TF Warfarin versus dipyridamole-aspirin and pentoxifylline-aspirin for the prevention of prosthetic heart valve thromboembolism: a prospective randomized clinical trial Circulation 1985 72 1059 63 10.1161/01.CIR.72.5.1059 3899404 2. Stein PD Collins JJ Kantrowitz A Antithrombotic therapy in mechanical and biological prosthetic heart valves and saphenous vein bypass grafts Chest 1986 89 46S 53 10.1378/chest.89.2_Supplement.46S 3510830 3. CADTH Therapeutic Reviews Recommendations for antithrombotic agents for the prevention of stroke and systemic embolism in patients with atrial fibrillation 2013 Ottawa Canadian Agency for Drugs and Technologies in Health 4. Al-Mubarak S Al-Ali N Abou-Rass M Al-Sohail A Robert A Al-Zoman K Al-Suwyed A Ciancio S Evaluation of dental extractions, suturing and INR on postoperative bleeding of patients maintained on oral anticoagulant therapy Br Dent J 2007 203 E15 10.1038/bdj.2007.725 17694045 5. Mulligan R Weitzel KG Pretreatment management of the patient receiving anticoagulant drugs J Am Dent Assoc 1988 117 479 83 10.1016/S0002-8177(88)73021-4 2972766 6. Wahl MJ Demystifying medical complexities J Calif Dent Assoc 2000 28 510 8 11324133 7. Devani P Lavery KM Howell CJ Dental extractions in patients on warfarin: is alteration of anticoagulant regime necessary? Br J Oral Maxillofac Surg 1998 36 107 11 10.1016/S0266-4356(98)90177-2 9643595 8. Wahl MJ Dental surgery in anticoagulated patients Arch Intern Med 1998 158 1610 6 10.1001/archinte.158.15.1610 9701094 9. Dunn AS Turpie AG Perioperative management of patients receiving oral anticoagulants: a systematic review Arch Intern Med 2003 163 901 8 10.1001/archinte.163.8.901 12719198 10. Garcia DA Regan S Henault LE Upadhyay A Baker J Othman M Hylek EM Risk of thromboembolism with short-term interruption of warfarin therapy Arch Intern Med 2008 168 63 9 10.1001/archinternmed.2007.23 18195197 11. Kaplan RC Tirschwell DL Longstreth WJ Manolio TA Heckbert SR Lefkowitz D El-Saed A Psaty BM Vascular events, mortality, and preventive therapy following ischemic stroke in the elderly Neurology 2005 65 6 835 42 10.1212/01.wnl.0000176058.09848.bb 16186519 12. Johnson WT Leary JM Management of dental patients with bleeding disorders: review and update Oral Surg Oral Med Oral Pathol 1988 66 297 303 10.1016/0030-4220(88)90235-6 2971908 13. DeClerck D Vinckier F Vermylen J Influence of anticoagulation on blood loss following dental extractions J Dent Res 1992 71 387 90 10.1177/00220345920710020701 1532608 14. Souto JC Oliver A Zuazu-Jausoro I Vives A Fontcuberta J Oral surgery in anticoagulated patients without reducing the dose of oral anticoagulant: a prospective randomized study J Oral Maxillofac Surg 1996 54 27 32 10.1016/S0278-2391(96)90297-9 8530996 15. Johnson-Leong C Rada RE The use of low-molecular-weight heparins in outpatient oral surgery for patients receiving anticoagulation therapy J Am Dent Assoc 2002 133 1083 7 10.14219/jada.archive.2002.0333 12198988 16 Evans IL Sayers MS Gibbons AJ Price G Snooks H Sugar AW Can warfarin be continued during dental extraction? Results of a randomized controlled trial Br J Oral Maxillofac Surg 2002 40 248 52 10.1054/bjom.2001.0773 12054719 17 Blinder D Manor Y Martinowitz U Taicher S Hashomer T Dental extractions in patients maintained on continued oral anticoagulant: comparison of local hemostatic modalities Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1999 88 137 40 10.1016/S1079-2104(99)70106-X 10468454 18 Blinder D Manor Y Martinowitz U Taicher S Dental extractions in patients maintained on oral anticoagulant therapy: comparison of INR value with occurrence of postoperative bleeding Int J Oral Maxillofac Surg 2001 30 518 21 10.1054/ijom.2001.0172 11829234 19 Al-Belasy FA Amer MZ Hemostatic effect of n-butyl-2-cyanoacrylate (histoacryl) glue in warfarin-treated patients undergoing oral surgery J Oral Maxillofac Surg 2003 61 1405 9 10.1016/j.joms.2002.12.001 14663804 20 Carter G Goss A Lloyd J Tocchetti R Tranexamic acid mouthwash versus autologous fibrin glue in patients taking warfarin undergoing dental extractions: a randomized prospective clinical study J Oral Maxillofac Surg 2003 61 1432 5 10.1016/j.joms.2003.01.001 14663808 21 Borea G Montebugnoli L Capuzzi P Magelli C Tranexamic acid as a mouthwash in anticoagulant-treated patients undergoing oral surgery. An alternative method to discontinuing anticoagulant therapy Oral Surg Oral Med Oral Pathol 1993 75 29 31 10.1016/0030-4220(93)90401-O 8419869 22 Duan XQ Lv YL Zuo J Gong LN Wang J Meng X Clinical studies in patients with artificial heart valve replacement surgery tooth extraction Beijing J Stomatol. 2010 18 332 4 23 Gaspar R Brenner B Ardekian L Peled M Laufer D Use of tranexamic acid mouthwash to prevent postoperative bleeding in oral surgery patients on oral anticoagulant medication Quintessence Int 1997 28 375 9 9477900 24 Doraiswamy VA Slepian MJ Gesheff MG Tantry US Gurbel PA Potential role of oral anticoagulants in the treatment of patients with coronary artery disease: focus on dabigatran Expert Rev Cardiovasc Ther 2013 11 1259 67 10.1586/14779072.2013.827469 23968500 25 Agnelli G Becattini C Risk assessment for recurrence and optimal agents for extended treatment of venous thromboembolism Hematology Am Soc Hematol Educ Program 2013 2013 471 7 24319221 26 Wahl MJ Myths of dental surgery in patients receiving anticoagulant therapy J Am Dent Assoc 2000 131 77 81 10.14219/jada.archive.2000.0024 10649877 27 Yasaka M Naritomi H Minematsu K Ischemic stroke associated with brief cessation of warfarin Thromb Res 2006 118 290 3 10.1016/j.thromres.2005.08.009 16197984 28 Hart RG Pearce LA Aguilar MI Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation Ann Intern Med 2007 146 12 857 67 10.7326/0003-4819-146-12-200706190-00007 17577005 29 Kearon C Akl EA Ornelas J Blaivas A Jimenez D Bounameaux H Huisman M King CS Morris TA Sood N Antithrombotic Therapy for VTE Disease: CHEST Guideline and Expert Panel Report Chest 2016 149 2 315 52 10.1016/j.chest.2015.11.026 26867832 30 Erkens PM Ten CH Buller HR Prins MH Benchmark for time in therapeutic range in venous thromboembolism: a systematic review and meta-analysis Plos One 2012 7 9 e42269 10.1371/journal.pone.0042269 23049730 31 Campbell JH Alvarado F Murray RA Anticoagulation and minor oral surgery: should the anticoagulation regimen be altered? J Oral Maxillofac Surg 2000 58 131 5 10.1016/S0278-2391(00)90324-0 10670590 32 Cieslik-Bielewska A Pelc R Cieslik T Oral surgery procedures in patients on anticoagulants. Preliminary report Kardiol Pol 2005 63 137 40 16136412 33 Lippert S Gutschik E Views of cardiac-valve prosthesis patients and their dentists on anticoagulation therapy Scand J Dent Res 1994 102 3 168 71 8085123 34 Beirne OR Koehler JR Surgical management of patients on warfarin sodium J Oral Maxillofac Surg 1996 54 1115 8 10.1016/S0278-2391(96)90172-X 8811824 35 Piccini JP Stevens SR Chang Y Singer DE Lokhnygina Y Go AS Patel MR Mahaffey KW Halperin JL Breithardt G Renal dysfunction as a predictor of stroke and systemic embolism in patients with nonvalvular atrial fibrillation: validation of the R(2)CHADS(2) index in the ROCKET AF (Rivaroxaban Once-daily, oral, direct factor Xa inhibition Compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation) and ATRIA (AnTicoagulation and Risk factors In Atrial fibrillation) study cohorts Circulation 2013 127 2 224 32 10.1161/CIRCULATIONAHA.112.107128 23212720 36 Hart RG Eikelboom JW Ingram AJ Herzog CA Anticoagulants in atrial fibrillation patients with chronic kidney disease Nat Rev Nephrol 2012 8 10 569 78 10.1038/nrneph.2012.160 22825670 37 Rakocz M Mazar A Varon D Spierer S Blinder D Martinowitz U Dental extractions in patients with bleeding disorders. The use of fibrin glue Oral Surg Oral Med Oral Pathol 1993 75 280 2 10.1016/0030-4220(93)90135-Q 8469534 38 Bajkin BV Popovic SL Selakovic SD Randomized, prospective trial comparing bridging therapy using low-molecular-weight heparin with maintenance of oral anticoagulation during extraction of teeth J Oral Maxillofac Surg 2009 67 990 5 10.1016/j.joms.2008.12.027 19375008 39 Fleming PS Lynch CD Pandis N Randomized controlled trials in dentistry: common pitfalls and how to avoid them J Dent 2014 42 908 14 10.1016/j.jdent.2014.06.004 24930873

PMC005xxxxxx/PMC5002167.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 66110.1186/s12974-016-0661-0ResearchThe glucagon-like peptide-1 receptor agonist exendin-4 ameliorates warfarin-associated hemorrhagic transformation after cerebral ischemia Chen Fangzhe chen198761@sina.com 1Wang Weifeng weifengdangdao@163.com 2Ding Hongyan hyding2001@163.com 1Yang Qi qi_yang@hotmail.com 1Dong Qiang dong_qiang@fudan.edu.cn 1Cui Mei cuimei@fudan.edu.cn 11 Department of Neurology, Huashan Hospital, State Key Laboratory of Medical Neurobiology, Fudan University, No. 12 Middle Wulumuqi Road, Shanghai, 200040 China 2 The Department of Clinical Laboratory, Central Laboratory, Jing’an District Centre Hospital of Shanghai, Huashan Hospital Fudan University Jing’an Branch, No. 259 Xi Kang Road, Shanghai, 200040 China 26 8 2016 26 8 2016 2016 13 1 20429 12 2015 14 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 As the number of patients with cardioembolic ischemic stroke is predicted to be double by 2030, increased burden of warfarin-associated hemorrhagic transformation (HT) after cerebral ischemia is an expected consequence. However, thus far, no effective treatment strategy is available for HT prevention in routine clinical practice. While the glucagon-like peptide-1 receptor (GLP-1R) agonist exendin-4 (Ex-4) is known to protect against oxidative stress and neuronal cell death caused by ischemic brain damage, its effect on preventing warfarin-associated HT after cerebral ischemia is yet unknown. Therefore, we hypothesized that Ex-4 would stabilize the blood-brain barrier (BBB) and suppress neuroinflammation through PI3K-Akt-induced inhibition of glycogen synthase kinase-3β (GSK-3β) after warfarin-associated HT post-cerebral ischemia. Methods We used male C57BL/6 mice for all experiments. A 5-mg warfarin sodium tablet was dissolved in animals’ drinking water (effective warfarin uptake 0.04 mg (2 mg/kg) per mouse). The mice were fed for 0, 6, 12, and 24 h with ad libitum access to the treated water. To study the effects of Ex-4, temporary middle cerebral artery occlusion (MCAO) was performed. Then, either Ex-4 (10 mg/kg) or saline was injected through the tail vein, and in the Ex-4 + wortmannin group, PI3K inhibitor wortmannin was intravenously injected, after reperfusion. The infarct volume, neurological deficits, and integrity of the BBB were assessed 72 h post MCAO. One- or two-way ANOVA was used to test the difference between means followed by Newman–Keuls post hoc testing for pair-wise comparison. Results We observed that Ex-4 ameliorated warfarin-associated HT and preserved the integrity of the BBB after cerebral ischemia through the PI3K/Akt/GSK-3β pathway. Furthermore, Ex-4 suppressed oxidative DNA damage and lipid peroxidation, attenuated pro-inflammatory cytokine expression levels, and suppressed microglial activation and neutrophil infiltration in warfarin-associated HT post-cerebral ischemia. However, these effects were totally abolished in the mice treated with Ex-4 + the PI3K inhibitor—wortmannin. The PI3K/Akt-GSK-3β signaling pathway appeared to contribute to the protection afforded by Ex-4 in the warfarin-associated HT model. Conclusions GLP-1 administration could reduce warfarin-associated HT in mice. This beneficial effect of GLP-1 is associated with attenuating neuroinflammation and BBB disruption by inactivating GSK-3β through the PI3K/Akt pathway. Electronic supplementary material The online version of this article (doi:10.1186/s12974-016-0661-0) contains supplementary material, which is available to authorized users. Keywords Cerebral ischemiaExendin-4Hemorrhagic transformationBlood-brain barrierNeuroinflammationPI3K/Akt-GSK-3β signaling pathwayWarfarinhttp://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China81471173Cui Mei http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China81271295http://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China81571109issue-copyright-statement© The Author(s) 2016 ==== Body Background Globally, ischemic stroke is one of the leading causes of death and long-term disability [1]. The number of patients with cardioembolic ischemic stroke resulting from nonvalvular atrial fibrillation (AF), the major cause of cardioembolic ischemic stroke, is predicted to double by 2030 [2, 3]. Consequently, a growing burden of warfarin-associated hemorrhagic transformation (HT) after cerebral ischemia can be expected [4–6]. Early HT can occur as a complication of cardioembolic ischemic stroke [7]. Additionally, a higher rate of hematoma expansion and a worse clinical outcome have been reported in warfarin-associated HT patients [8–10]. However, no effective treatment strategy is available for prevention of HT in clinical practice. Experimental studies of cerebral ischemia have established increase in the permeability of the blood-brain barrier (BBB) after ischemia/reperfusion injury as one of the major causes of HT [11, 12]. The glucagon-like peptide-1 receptor (GLP-1R) agonist exendin-4 (Ex-4) is a long-acting analog of the endogenous insulinotropic peptide GLP-1. Both GLP-1 and Ex-4 have multiple physiologic functions, such as the induction of glucose-dependent insulin release, inhibition of glucagon secretion, stimulation of B cell replication, and antiapoptotic action [13]. Owing to their small molecule size, both GLP-1 and Ex-4 can diffuse across the BBB in the central nervous system and provide neuroprotection in cerebral ischemia [14, 15]. While it has been reported that Ex-4 can protect against oxidative products and neuronal cell death caused by ischemic brain damage, it is yet unknown whether Ex-4 is effective in preventing warfarin-associated HT after cerebral ischemia. Previous studies have shown that after a hemorrhagic stroke, cytotoxic events activate the ubiquitously expressed glycogen synthase kinase-3β (GSK-3β), which increases the expression of β-catenin [16, 17] and subsequently decreases the expressions of claudins [18]. There is substantial evidence that GSK-3β inhibition (tyrosine-216 dephosphorylation) reduces neuronal apoptosis [19–21] and attenuates neuroinflammation in neurodegenerative models [22–24]. Pharmacological stimulation of GLP-1R activates the phosphatidylinositol 3-kinase (PI3K)-Akt signaling pathway, and a number of studies have linked GSK-3β with the PI3K/Akt pathway, thereby showing that phosphorylated Akt inactivates GSK-3β via tyrosine-216 dephosphorylation. Herein, we hypothesized that Ex-4 would stabilize the BBB and suppress neuroinflammation through PI3K-Akt-induced inhibition of GSK-3β after warfarin-associated HT post-cerebral ischemia in mice. Methods Animals All experiments were conducted using male C57BL/6 mice (body weight 18–25 g) at a constant temperature and with a consistent light cycle (from 07:00 to 18:00) under normal diet. This study was carried out in accordance with the Guide for the National Science Council of the Republic of China. All animals were treated according to protocols approved by the Institutional Animal Care and Use Committee of Fudan University. A 5-mg warfarin sodium tablet (Coumadin™, Sigma-Aldrich, St. Louis, MO, USA) was dissolved in 375 mL of water. The C57 BL/6 mice were fed for 0, 6, 12, and 24 hours with ad libitum access to the treated water. Assuming a mouse body weight of 20 g and a water consumption rate of 15 mL/100 g per 24 h, this dosage corresponds to a warfarin uptake of 0.04 mg (2 mg/kg) per mouse over a 24-h period. Similar doses of warfarin have been previously used [25]. After 24 h, the warfarin was withdrawn and middle cerebral artery occlusion was performed (Additional file 1: Figure S1). For the international normalized ratio (INR) measurement, the mice were under deep anesthesia, a peritoneal midline incision was performed, and 0.6 mL blood was drawn from the inferior caval vein as previously described [26]. Blood was transferred to glass tubes (BD Vacutainer®) containing sodium citrate as the anticoagulant. Measurements of INR values and prothrombin time were performed in the Department of Central Laboratory, Jingan District Centre Hospital, Shanghai, China. Temporary middle cerebral artery occlusion and drug treatment Mice were anesthetized with ketamine/xylazine (65/6 mg/kg, i.p), and their body temperature was maintained at 37 °C by a heating pad and feedback control system (FHC, Bowdoin, ME, USA). A laser Doppler probe was fixed on the skull 5 mm lateral and 2 mm posterior to the bregma. A coated filament was placed on the right middle cerebral artery (MCA) with concurrent recording of laser Doppler cerebral blood flow to ensure that the cerebral blood flow decreased to below 25 % of the baseline. After 45 min, the filament was removed (Additional file 2: Figure S2). Either Ex-4 (10 mg/kg) or saline was injected through the tail vein immediately after reperfusion. In the Ex-4 + wortmannin group, we intravenously injected 15 μL/kg wortmannin (Sigma-Aldrich), a non-specific, covalent inhibitor of PI3K immediately after reperfusion. Assessment of infarct volume, neurological deficits, and blood-brain barrier All the mice were killed 72 h after temporary middle cerebral artery occlusion (MCAO), and brain tissues were incubated in 2,3,5-triphenyltetrazolium chloride (TTC) for 1 h. The infarct area in each slice was analyzed by a computerized image analysis system, and the infarct volume was calculated by multiplying the distance between sections [27]. Neurological score was determined 72 h after MCAO, according to the graded scoring system described previously by Li et al. [28]. Assessment of motor coordination deficits was performed on days 3 and 7 using the rota rod as previously described [29]. Investigators who performed MCAO models, evaluation of infarct volumes, neurological scales, and rota rod were blinded to all the experimental protocols and drug treatments. To measure BBB permeability, Evans blue (Sigma-Aldrich) was dissolved in saline (2 %) and injected into the right jugular vein 72 h after MCAO. The animals were then killed, and the brain hemispheres were homogenized in 3 mL of N,N-dimethylformamide (Sigma-Aldrich); incubated for 18 h at 55 °C; and centrifuged. The supernatants were subjected to spectrophotometry at 620 nm. Quantification of hemorrhagic transformation The hemoglobin content in brain tissue was quantified by spectrophotometric assay. The hemispheric brain tissue was homogenized with phosphate-buffered saline (PBS) and centrifuged at 13,000×g for 30 min. The hemoglobin-containing supernatant was collected, 80 μL of Drabkin reagent (Sigma) was added to 20-μL supernatant aliquots, and the sample was kept standing for 15 min at room temperature. The optical density in each group was measured at 540 nm, and hemorrhage volume was expressed in equivalent units by comparison with a reference curve generated using homologous blood. Western blotting Striatal brain tissues from the MCA were lysed with radioimmunoprecipitation assay buffer (RIPA) containing protease inhibitors (Sigma-Aldrich, St. Louis, MO, USA). Proteins were separated by SDS-PAGE and then transferred onto a nitrocellulose membrane. The membranes were incubated overnight at 4 °C with the following primary antibodies: anti-p-GSK-3β (Tyr216, 1:1000, Abcam Inc., Cambridge, MA); anti-GSK-3β (1:1000, Abcam); anti-β-actin (1:5000, Sigma-Aldrich); anti-p-β-catenin (Ser33/37/Thr41, 1:2000, Cell Signaling Technology Inc., Danvers, MA); anti-β-catenin (1:1000, Abcam), anti-claudin-3 (1:2000, Santa Cruz, CA); anti-claudin-5 (1:2000, Santa Cruz); anti-p-Akt (Ser473, 1:2000, Cell Signaling); anti-Akt (1:2000, Cell Signaling); anti-ICAM-1 (1:1000, Abcam); anti-VCAM-1 (1:1000, Abcam); anti-IKK-β (1:2000, Santa Cruz); anti-NF-kB (1:2000, Santa Cruz); anti-HHE (1:1,000, Abcam); anti-Iba1 (1:1,000, Abcam); and anti-myeloperoxidase (MPO) (1:2000, Santa Cruz). Secondary antibodies conjugated with horseradish peroxidase were used, and immunoreactivity was visualized by chemiluminescence (SuperSignal Ultra, Pierce, Rockford, IL, USA). Bands of interest were analyzed and quantified using Scion Image. siRNA-mediated GSK-3β gene knockdown The small interfering RNA (siRNA)-mediated GSK-3β gene knockdown was performed as previously described [30]. Briefly, two pairs of GSK-3β siRNAs (21500 R12-1717, R12-1719; Cell Signaling) with a total volume of 4 μL (2 μL each) were stereotaxically injected to the right lateral ventricle following coordinates relative to the bregma: AP = −0.4 mm, L = −1.0 mm, and H = − 2.0 mm (from the brain surface) 48 h prior to MCAO. Measurement of cytokine concentration Striatal brain tissues from the MCA were homogenized and collected by centrifugation at 15,000×g for 30 min at 4 °C and then stored at −70 °C until the assay was performed. The supernatant was assayed for tumor necrosis factor-α (TNF-α) and interleukin-1 beta (IL-1β) using enzyme-linked immunosorbent assays (ELISA; R&D Biosystems) as described previously [31]. Measurement of 8-OHdG formation in the brain Concentration of 8-hydroxy-2′-deoxyguanosine (8-OHdG) in brain DNA was measured by Piao et al.’s method [32], with slight modifications. Briefly, 200 mg of brain tissue was homogenized in 0.25 M sucrose solution. DNA was extracted from the homogenate under anaerobic conditions. The 8-OHdG content in the brain was measured by using an HPLC-ECD as previously described [33]. Each brain sample was examined in duplicate. Immunohistochemistry Seventy-two hours after MCAO, the mice were anesthetized and first perfused with saline followed by fixation with buffered paraformaldehyde (4 %). The brains were removed and post-fixed in 4 % paraformaldehyde; the paraformaldehyde was then removed and replaced with 30 % sucrose solution overnight. Then 15-μm coronal sections were obtained on a cryostat. The slices were blocked with PBS containing 5 % bovine serum albumin (BSA), 10 % goat serum, and 0.3 % Triton-X 100. Next, the slices were incubated with the primary antibodies anti-Iba1 (1:250, Abcam) and anti-TNF-α (1:100, Santa Cruz) overnight at 4 °C. Then Alexa Fluor 488 or 595 labeled secondary antibody (Molecular Probes Inc., Eugene, OR, USA) for 2 h at room temperature. The tissue sections were washed twice in PBS and then immersed in DAPI (Molecular Probes) solution (1:1000 dilution) for 10 min. The sections were finally rinsed in distilled water and fixed with a coverslip with anti-fade mounting medium. Assessment of microglia activation First, microglia activation were counted and morphologically characterized based on the following criteria. Cells with an oval cell body containing a small volume of cytoplasm and long, thin, delicate, and radially branched processes were classified as ramified microglia [34]. Activated microglia were defined as having an enlarged soma (width greater or equal to 30 μm) and a broad-flattened appearance with the common presence of several lamellapodia [35]. This morphological classification was then confirmed by using a methodology of semi-automatic image analysis to analyze the cell body to cell size ratio in Iba1-stained brain sections as described before [36] by ImageJ software. Statistical analysis All values are expressed as mean ± standard deviation (SD). Differences between means were analyzed using either one-way or two-way ANOVA followed by Newman–Keuls post hoc testing for pair-wise comparison using SigmaStat v 3.5®. A P value <0.05 was considered statistically significant. Results Exendin-4 ameliorated warfarin-associated HT after cerebral ischemia To examine the influence of warfarin on animal PT-INR values, the mice were killed at the indicated time points and the PT-INR values were measured. After warfarin administration, the PT-INR values increased in a time-dependent manner (Fig. 1a). After 24 h of warfarin administration, PT-INR values were elevated (mean = 3.85 ± 1.12; n = 6) and reached the therapeutic span used in humans. These results were consistent with those previously reported [26]. In view of these results, we decided to use 24 h as the warfarin administration time for all subsequent experiments.Fig. 1 Exendin-4 treatment reduced the stroke volume and improved neurological function after cerebral ischemia. a Prothrombin time-international normalized ratio values (PT-INR) in non-MCAO mice after 0, 6, 12, and 24 h of warfarin administration through drinking water. b Regional cerebral blood flow (rCBF) in both ischemic and reperfusion stages was recorded using laser Doppler cerebral blood flow. c Exendin-4 (Ex-4, 10 mg/kg) was injected through the tail vein immediately after reperfusion. The infarct volume was measured 72 h after middle cerebral artery occlusion (MCAO) using TTC straining. d Representative images of TTC straining showing the ischemic area and hemorrhage transformation. e, f Exendin-4 (Ex-4, 10 mg/kg) was injected through tail vein immediately after reperfusion. The neurological deficits were measured 72 h after MCAO, and assessment of motor function was analyzed on days 3 and 7 using rota rod after MCAO. Data are presented as mean ± SD and analyzed by two-way ANOVA. *P < 0.05 compared with the Ex-4(−) group, # P < 0.05 compared with the MCAO+/Ex-4(+) group MCAO induced a sharp drop of rCBF, leading to extensive infarction in the cerebral cortical and subcortical areas over a series of brain sections in the mice (Fig. 1b, d). Compared with the MCAO+/Ex-4 group, warfarin treatment did not increase the infarct size or neurological deficits. However, warfarin significantly exacerbated HT after cerebral ischemia. Ex-4 suppressed this exacerbation (Fig. 2a). Moreover, Ex-4 showed striking protective effects to reduce to infarct volume and improve neurological function in MCAO mice with or without warfarin treatment (Fig. 1c–f).Fig. 2 Exendin-4 treatment reduced warfarin-associated HT after cerebral ischemia. a Brain hemoglobin levels were evaluated at 72 h after middle cerebral artery occlusion (MCAO). Data are presented as mean ± SD and analyzed by two-way ANOVA. *P < 0.05 compared with the MCAO+/Ex-4(−) group, # P < 0.05 compared with the MCAO+/Ex-4(+) group. b Blood-brain barrier (BBB) integrity in MCAO mice were assessed after Evans blue staining. Data are presented as mean ± SD and analyzed by two-way ANOVA. *P < 0.05 compared with the Ex-4(−) group, # P < 0.05 compared with the MCAO+/Ex-4(+) group Exendin-4 preserves the BBB integrity in warfarin-associated HT after cerebral ischemia Functional barrier properties were evaluated using Evans blue assays, 72 h after surgery. Significantly more extravasated dye was measured in the ischemic hemispheres of mice subjected to warfarin treatment compared with the control group. Ex-4 preserved BBB integrity in the model of warfarin-associated HT after MCAO, which was associated with significantly reduced dye extravasation in Ex-4-treated animals (Fig. 2b). Claudin-3 and claudin-5 are transmembrane proteins essential for maintaining the diffusion barrier provided by tight junctions [37, 38]. Previous studies reported the regulatory role of activation of GSK-3β and β-catenin in claudin-3 and claudin-5 gene expression, respectively [39]. Western blot analyses of the ischemic brain were conducted at 72 h after MCAO. Changes in protein expression of phosphorylated and, therefore, activated GSK-3β (p-GSK-3β, Tyr216) were quantified as a ratio to total GSK-3β (Fig. 3a, b). GSK-3 phosphorylation was significantly increased in the warfarin-treated mice compared with the control group. However, Ex-4 significantly reduced the p-GSK-3β/GSK-3β ratio. Consistent with these results, increased phosphorylated β-catenin levels were also found in warfarin-treated mice compared with the control group (Fig. 3c, d). Ex-4 significantly reduced the p-β-catenin/β-catenin ratio. Tight junction protein expressions were also detected. As shown in Fig. 3e–h, claudin-3 and claudin-5 levels were reduced in the model of warfarin-associated HT after MCAO compared with MCAO mice. However, Ex-4 treatment significantly reversed the reduction.Fig. 3 Exendin-4 treatment preserved BBB integrity in warfarin-associated HT after cerebral ischemia. Western blot analysis of a, b p-GSK-3β/GSK-3β; c, d p-β-catenin/β-catenin; e, f claudin-3; and g, h claudin-5. Representative blots from six independent experiments with similar results are shown. (Sham: sham-operated group, Ctrl: MCAO group, W: warfarin-associated HT group, W + Ex-4: warfarin-associated HT pretreatment with Ex-4). Data are presented as mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 compared with warfarin-associated HT group Exendin-4 ameliorated warfarin-associated HT after cerebral ischemia through PI3K/Akt/GSK-3β pathway It has been reported that activated Akt (p-Akt) can inactivate GSK-3β and reduce the amount of GSK-3β available for phosphorylation (through the tyrosine-216 form) [40, 41]. The inactivation of GSK-3β, specifically through tyrosine-216 dephosphorylation, increased β-catenin, which is an important factor in maintaining BBB integrity [42, 43]. To examine if warfarin and Ex-4 could phosphorylate Akt after cerebral ischemia, the phosphorylation of Akt was examined 72 h after MCAO. After normalizing the values of the active p-Akt with the amount of total Akt (Akt) in each sample, we observed an increase in the Ex-4-treated mice compared to warfarin treatment alone. We indirectly studied the activation of Akt by measuring the phosphorylation of its downstream target GSK-3β in the same brain areas. As compared with the warfarin-treated MCAO mice, Ex-4 treatment significantly suppressed the phosphorylation of GSK-3β. These phosphorylation changes of Akt and GSK-3β were totally abolished when the mice were treated with Ex-4 in combination with the PI3K inhibitor—wortmannin (Fig. 4).Fig. 4 Effect of Ex-4 on PI3K/Akt/GSK-3β pathway in warfarin-associated HT after cerebral ischemia. Mice were intravenously treated either with Ex-4 (Ex-4, 10 mg/kg) or Ex-4 plus the PI3K inhibitor (wortmannin) (15 μL/kg) right after reperfusion. The expression levels of a, b p-Akt/Akt and c, d p-GSK-3β/GSK-3β were analyzed by immunoblotting. Representative blots from six independent experiments with similar results are shown. Data are presented as mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 vs. MCAO + warfarin group; **P < 0.05 vs. MCAO + warfarin + Ex-4 group These results showed that Ex-4 induced PI3K/Akt pathway activation and subsequent GSK-3β inactivation in the model of warfarin-associated HT after cerebral ischemia. Next, the role of PI3K/Akt/GSK-3β signaling pathway in the integrity of BBB was further investigated using the following antagonists: PI3K inhibitor wortmannin and GSK-3β siRNA in the model of warfarin-associated HT after cerebral ischemia. GSK-3β knockdown by siRNA significantly reduced the warfarin-associated HT. Ex-4 also reversed the warfarin-induced HT in ischemic mice. The mice receiving Ex-4 in combination with wortmannin, however, failed to show this protective effect (Fig. 5a).Fig. 5 Ex-4 alleviated warfarin-associated HT after cerebral ischemia through PI3K/Akt/GSK-3β pathway. Effects of GSK-3β siRNA and wortmannin on a brain hemoglobin level and b Evans blue extravasation were evaluated 72 h after MCAO. c, d The expression levels of p-β-catenin/β-catenin were also been detected by western blotting. Data are presented as mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 vs. MCAO + warfarin group; **P < 0.05 vs. MCAO + warfarin + Ex-4 group Exendin-4 preserves BBB integrity in warfarin-associated intracerebral hemorrhage after cerebral ischemia through PI3K/Akt/GSK-3β pathway To detect the role of the PI3K/Akt/GSK-3β signaling pathway in preventing BBB disruption in the Ex-4-treated mice, wortmannin and GSK-3β siRNA were used. In the model of warfarin-associated HT after cerebral ischemia, GSK-3β knockdown by siRNA significantly prevented the warfarin-induced BBB disruption. When administered alone, Ex-4 preserved BBB integrity after warfarin-associated HT. Mice receiving Ex-4 in combination with wortmannin failed to demonstrate reduced dye extravasation into the ischemic brain hemisphere (Fig. 5b). The effect of GSK-3β siRNA on the expression levels of the p-β-catenin/β-catenin ratio was also measured. GSK-3β knockdown by siRNA significantly reduced the expression of p-β-catenin. As shown in Fig. 5c, Ex-4 reduced GSK-3β activation, thereby stabilizing β-catenin. However, when the mice were treated with Ex-4 and wortmannin, this stabilization effect of Ex-4 was completely lost. The expression levels of tight junction proteins were also detected; warfarin-associated HT reduced claudin-3 and claudin-5 levels. However, Ex-4 treatment significantly increased their expression, and wortmannin reversed the initial increase of claudin-3 and claudin-5 by Ex-4 (Fig. 6a, b).Fig. 6 Ex-4 preserved the BBB integrity after warfarin-associated HT through PI3K/Akt/GSK-3β pathway. Mice were intravenously treated with either Ex-4 (Ex-4, 10 mg/kg) or Ex-4 plus wortmannin (15 μL/kg) right after reperfusion. The expression levels of a claudin-3, b claudin-5, c ICAM-1, and d VCAM-1 were analyzed by western blotting. Data are presented as mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 vs. MCAO + warfarin group; **P < 0.05 vs. MCAO + warfarin + Ex-4 group The PI3K/Akt pathway has been implicated in stabilization of the BBB through decreased expression of endothelial adherent proteins vascular cell adhesion molecule-1 (VCAM-1) and interstitial cell adhesion molecule-1 (ICAM-1) [44, 45]. Warfarin-associated HT significantly increased the expression of ICAM-1 and VCAM-1. Both adhesion molecules’ expressions were decreased by Ex-4 treatment, and wortmannin reversed the reduction of the adhesion molecules’ levels induced by Ex-4 (Fig. 6c, d). Exendin-4 suppresses oxidative DNA damage and lipid peroxidation in warfarin-associated HT after cerebral ischemia Next, we investigated whether Ex-4 can control oxidative stress in warfarin-associated HT using lipid peroxidation indicator (HHE) and DNA oxidative injure indicator (8-OHdG). 8-OHdG is a major form of oxidative DNA damage product, and 4-hydroxyhexenal (HHE) is one of the major lipid peroxidation products that are formed by n-3 polyunsaturated fatty acids in cells exposed to oxidative stress [46]. The expression levels of 8-OHdG and HHE were significantly increased in warfarin-associated HT brains compared to MCAO brains. The levels of these oxidative stress markers were significantly decreased in the Ex-4-treated group. When the mice were treated in combination with wortmannin, Ex-4 failed to suppress the expression levels of 8-OHdG and HHE (Fig. 7).Fig. 7 Ex-4 suppressed oxidative DNA damage and lipid peroxidation in warfarin-associated HT after cerebral ischemia. Mice were intravenously treated with either Ex-4 (Ex-4, 10 mg/kg) or Ex-4 plus wortmannin (15 μL/kg) right after reperfusion. The expression levels of a HHE and b 8-OHdG in the brain tissue were detected. Data are presented as mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 vs. MCAO + warfarin group; **P < 0.05 vs. MCAO + warfarin + Ex-4 group Exendin-4 attenuated pro-inflammatory cytokines in warfarin-associated HT after cerebral ischemia We additionally examined the role of Ex-4 in modulating neuroinflammation by measuring expression levels of several cytokines such as IKK-β, NF-kB, TNF-α, and IL-1β. The expression levels of IKK-β and NF-kB were significantly increased after warfarin-associated HT compared to MCAO alone, while Ex-4 treatment reduced the effect and wortmannin blocked the reduction induced by Ex-4 (Fig. 8a, b).Fig. 8 Ex-4 attenuated pro-inflammatory cytokines in warfarin-associated HT after cerebral ischemia. Mice were intravenously treated with either Ex-4 (Ex-4, 10 mg/kg) or Ex-4 plus wortmannin (15 μL/kg) right after reperfusion. The expression levels of a IKK-β and b NF-kB were analyzed by western blotting. The concentrations of pro-inflammatory cytokines c TNF-α and d IL-1β were detected by ELISA. Data are mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 vs. MCAO + warfarin group; **P < 0.05 vs. MCAO + warfarin + Ex-4 group The expression levels of TNF-α and IL-1β were evaluated by ELISA. Both these cytokines were upregulated in warfarin-associated HT mice, and Ex-4 blocked the increase concordantly with a similar pattern for IKK-β and NF-kB. The modulating effect of Ex-4 on the cytokines’ expression levels were reversed by co-treatment with wortmannin (Fig. 8c, d). Exendin-4 suppresses neuroinflammation in warfarin-associated HT after cerebral ischemia Consistent with the changes in cytokine levels, immunofluorescence analysis also showed that warfarin-associated HT robustly enhanced immunofluorescence intensity of Iba1 staining (a marker of microglia/macrophages) in the MCA area compared to the MCAO group (Fig. 9a). The quantification results showed Iba1-positive cells were significantly attenuated in the mice treated with the Ex-4 (Fig. 9b). Further, morphology analysis showed that the number of activated microglia was attenuated in the Ex-4-treated group (Fig. 9e–g). Consistent with these results, double immunofluorescent staining showed Iba1+/TNF-α + cells were elevated in the warfarin-associated HT group and Ex-4 significantly reduced the double positive cells (Additional file 3: Figure S3). Wortmannin blocked this function of Ex-4. Western blotting showed similar results with immunostaining (Fig. 9c). Taken together, these results suggest that the protection conferred by Ex-4 was likely mediated by the inhibition on warfarin-associated neuroinflammation after cerebral ischemia.Fig. 9 Ex-4 reduced Iba1+ microglial/macrophage cells and neutrophil infiltration in warfarin-associated HT after cerebral ischemia. Immunostaining of Iba1 was performed in the cortical and subcortical areas supplied by the middle cerebral artery. a Representative immunofluorescence images showed Iba1+ (green) and DAPI+ (blue) microglia/macrophages in the Ex-4-treated mice compared to the warfarin-associated HT group. Scale bar 100 μm. b Quantitative analysis of Iba1+ cells. The expression levels of Iba1 (c) and MPO (d) were detected by western blotting. e Representative images show microglial morphology in different groups. Scale bar 50 μm. f The number of activated microglia was expressed as a percentage of the total number of Iba1+ cells. g The cell body to cell size ratio of microglia provides additional information about microglial activation. Data are presented as mean ± SD from six independent experiments and analyzed by one-way ANOVA. *P < 0.05 vs. MCAO group; # P < 0.05 vs. MCAO + warfarin group; **P < 0.05 vs. MCAO + warfarin + Ex-4 group In addition to brain resident microglia, hematogenous leukocytes have been shown to play a pivotal role in post-stroke neuroinflammation. Among white blood cells, neutrophils have attracted much interest recently and have been intensively studied. The level of myeloperoxidase (MPO) was significantly increased in the warfarin-associated HT group compared to MCAO. Ex-4 treatment reversed the MPO level. The inhibition of Akt by wortmannin restored the MPO level back to that of the warfarin-associated HT group (Fig. 9d). Discussion Atrial fibrillation is a severe independent risk factor of stroke, its attributable risk increasing with age up to more than 20 % [47]. INR-driven oral anticoagulation with vitamin K antagonists to an INR of 2–3 reduces the risk of an ischemic stroke by over 60 % and has been the standard of stroke prevention in patients with AF [48]. However, anticoagulation therapy is closely related to HT after ischemia. In addition, cardioembolic stroke also carries with it an increased risk of HT [49]. The chief mechanism of HT is considered to be blood leakage due to disruption of the BBB. Our results showed that pretreatment with warfarin could significantly increase the INR level in a time-dependent manner and dramatically enhance Evans blue leakage provoked by MCAO. Although the infarct volume and neurological deficits were not significantly different between the groups with or without warfarin treatment, warfarin significantly promoted the HT after cerebral ischemia, which is consistent with the permeability measurement results. GLP-1 and long-acting Ex-4 induce numerous biological actions through the G protein-coupled GLP-1 receptor (GLP-1R). GLP-1R is reportedly expressed in a wide range of tissues, including the brain. Moreover, GLP-1R stimulation has shown neuroprotective actions in previous findings, thereby establishing that GLP-1R stimulation protects hippocampal neurons from amyloid-β peptide and glutamate-induced toxicity [50, 51]. As the GLP-1R agonist Ex-4 is permeable to the BBB with a relatively long half time, it has possible clinical applications. Several studies have shown that Ex-4 can protect against oxidative products and neuronal cell death caused by ischemic brain damage [15]. However, to the best of our knowledge, whether GLP-1R stimulation is associated with warfarin-associated HT has not yet been studied. Herein, we reported that Ex-4 prevented the exacerbation of HT caused by warfarin without affecting the infarct volume. The mechanism whereby Ex-4 prevented the exacerbation of HT might involve maintenance of the expression of tight junction proteins and suppress the neuroinflammation associated with warfarin treatment. The pathways that strengthen the antiapoptotic and neuroprotective effects of Ex-4 after cerebral ischemia mostly converge on activation of the transcription factor cAMP response element-binding protein (CREB) by phosphorylation. In the present study, the PI3K/Akt-GSK-3β signaling pathway appeared to contribute to the protection afforded by Ex-4 in the warfarin-associated HT model. PI3K/Akt plays a crucial role in the cell death/survival pathway through several different downstream targets including GSK-3β [52]. A temporal increase in phospho-Akt after cerebral ischemia has been reported, and GSK-3β dephosphorylation at tyrosine-216 is accelerated as a downstream target of Akt [53]. The inactivation of GSK-3β via tyrosine-216 dephosphorylation mediates neuronal survival after cerebral ischemia [43]. In addition, the inactivation of GSK-3β results in stabilization of β-catenin, a protein that plays a role in cell adhesion. As a result, free β-catenin is allowed to accumulate and be translocated to the nucleus, binding to the transcription factors to alter target gene expressions [54], such as those of tight junction proteins claudin-3 and claudin-5 [18, 39]. Furthermore, GSK-3β inactivation may also decrease NF-kB expression, thereby reducing neuroinflammation. In this study, Akt phosphorylation at Ser473 and GSK-3β dephosphorylation at tyr216 were increased in warfarin-associated HT after cerebral ischemia. Administration of Ex-4 substantially decreased HT and maintained the stability of BBB. The reduced dye extravasation and brain hemoglobin level were similar to that achieved by inhibition of GSK-3β. Evidence supporting enhanced BBB stabilization by Ex-4 including decreased adherens (VCAM-1 and ICAM-1) and increased tight junction (claudin-3 and claudin-5) proteins could be totally abolished by wortmannin, a specific PI3K inhibitor. These results suggest that warfarin-associated HT reduced the expression of tight junction proteins. This effect was prevented by treatment with Ex-4 through the PI3K/Akt-GSK-3β pathway. Furthermore, Ex-4 reduced the warfarin-induced hemorrhage volume via a protective effect on vascular endothelial cells. Inflammation has been recognized as a key contributor to the pathophysiology of cerebral ischemia [55]. Inflammation includes a series of cellular events such as infiltration of neutrophil cells and activation of microglia/macrophages and astrocytes [56]. We found that warfarin-associated HT significantly upregulated Iba1-positive cells. Microglia/macrophage activation, together with elevated expression of pro-inflammatory cytokines such as IKK-β, NF-kB, TNF-α, and IL-1β, demonstrated that the warfarin-associated HT induced a neuroinflammation after cerebral ischemia. It has also been reported that activated microglia/macrophages are major sources of metalloproteinase generation, which is closely associated with ischemia-induced cerebral hemorrhage and edema. NF-kB is a central mediator of these inflammatory processes. Recent evidence has shown that the PI3K/Akt signaling pathway may be an endogenous negative feedback regulator of NF-kB-mediated pro-inflammatory responses [57, 58]. Several pro-inflammatory NF-kB target genes including TNF-α and IL-1β could mediate the deleterious effects on neurons under ischemic conditions. In the present study, we showed that warfarin-induced HT markedly induced the activation of microglia/macrophages and consequently increased the production of pro-inflammatory cytokines and Ex-4 significantly inhibited the neuroinflammation induced by warfarin through the PI3K/Akt-GSK-3β pathway. Moreover, suppression of oxidative damage is also a key factor in neuroprotection. Using 8-OHdG and HHE as markers of oxidative stress, our study showed that Ex-4 reduced the warfarin-induced accumulation of oxidative DNA damage and lipid peroxidation after cerebral ischemia. Conclusions Our study results showed that administration of GLP-1 could reduce warfarin-associated HT in mice. This beneficial effect of GLP-1 was associated with attenuating neuroinflammation and BBB disruption by inactivating GSK-3β through the PI3K/Akt pathway. These findings have important clinical implications and would be particularly beneficial in those receiving anticoagulant therapy. Future clinical trials should focus on confirming the efficacy and safety of this therapy. Additional files Additional file 1: Figure S1. The PT-INR values after warfarin withdrawal. After warfarin withdrawal, INR values remained stable for the next 6 h and dropped to normal values after 24 h. Data are shown as mean ± SD. Additional file 2: Figure S2. The rCBF levels in the ischemia and reperfusion stages in MCAO mice. A coated filament was placed on the right middle cerebral artery (MCA) with concurrent recording of laser Doppler cerebral blood flow. In the ischemia stage, the rCBF decreased to <25 % of baseline. After 45 min, the filament was removed and the rCBF increased to 110 % of baseline. Additional file 3: Figure S3. Representative immunofluorescence images showed co-localization of Iba1 (green) and TNF-α (red) in microglia. Immunostaining of Iba1(green), TNF-α(red), and DAPI (blue) was performed in the cortical and subcortical areas supplied by the middle cerebral artery. (A) Representative immunofluorescence images showed the percentage of Iba1+/TNF-α + cells to total Iba1+ cells was increased after warfarin treatment. EX-4 treatment reduced the Iba1+/TNF-α + cells percentage, whereas wortmannin blocked this effect of EX-4. Scale bar 50 μm. (B) Quantitative analysis of Iba1 and TNF-α double positive cells/Iba1-positive cells. Abbreviations 8-OHdG8-hydroxy-2′-deoxyguanosine AFAtrial fibrillation BBBBlood-brain barrier Ex-4Exendin-4 GLP-1RGlucagon-like peptide-1 receptor GSK-3βGlycogen synthase kinase-3β HHE4-hydroxyhexenal HTHemorrhagic transformation ICAM-1Interstitial cell adhesion molecule-1 IL-1βInterleukin-1 beta INRInternational normalized ratio MCAOMiddle cerebral artery occlusion PBSPhosphate-buffered saline PI3KPhosphatidylinositol 3- kinase RIPARadioimmunoprecipitation assay buffer TNF-αTumor necrosis factor-α TTC2,3,5-triphenyltetrazolium chloride VCAM-1Vascular cell adhesion molecule-1 Acknowledgements We are grateful to Baoguo Xiao for his technical support and Min Guo for assisting in preparing this manuscript. Funding This study was supported by the National Natural Science Foundation of China 81471173 (to MC) and 81271295 and 81571109 (to QD). Authors’ contributions MC and QD designed the study. FC and WW performed the experiments, carried out the statistical analysis, and prepared the manuscript. They contributed equally to this work. HD and QY were involved in experiment performance and data collection. MC and QD were responsible for the supervision of the entire project and were involved in the study design, data interpretation, manuscript preparation, and funding. 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 The study was approved by the Ethics Committee of Fudan University, Shanghai, China. The approval number from IRB is “20150572A259.” ==== Refs References 1. Lopez AD Mathers CD Ezzati M Jamison DT Murray CJ Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data Lancet 2006 367 1747 57 10.1016/S0140-6736(06)68770-9 16731270 2. Go AS Hylek EM Phillips KA Chang Y Henault LE Selby JV Singer DE Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the anticoagulation and risk factors in atrial fibrillation (ATRIA) Study JAMA 2001 285 2370 5 10.1001/jama.285.18.2370 11343485 3. Go AS The epidemiology of atrial fibrillation in elderly persons: the tip of the iceberg Am J Geriatr Cardiol 2005 14 56 61 10.1111/j.1076-7460.2005.02278.x 15785146 4. Kucher N Castellanos LR Quiroz R Koo S Fanikos J Goldhaber SZ Time trends in warfarin-associated hemorrhage Am J Cardiol 2004 94 403 6 10.1016/j.amjcard.2004.04.050 15276120 5. Steiner T Rosand J Diringer M Intracerebral hemorrhage associated with oral anticoagulant therapy: current practices and unresolved questions Stroke 2006 37 256 62 10.1161/01.STR.0000196989.09900.f8 16339459 6. Flaherty ML Kissela B Woo D Kleindorfer D Alwell K Sekar P Moomaw CJ Haverbusch M Broderick JP The increasing incidence of anticoagulant-associated intracerebral hemorrhage Neurology 2007 68 116 21 10.1212/01.wnl.0000250340.05202.8b 17210891 7. Alexandrov AV Black SE Ehrlich LE Caldwell CB Norris JW Predictors of hemorrhagic transformation occurring spontaneously and on anticoagulants in patients with acute ischemic stroke Stroke 1997 28 1198 202 10.1161/01.STR.28.6.1198 9183351 8. Pfeilschifter W Spitzer D Czech-Zechmeister B Steinmetz H Foerch C Increased risk of hemorrhagic transformation in ischemic stroke occurring during warfarin anticoagulation: an experimental study in mice Stroke 2011 42 1116 21 10.1161/STROKEAHA.110.604652 21330626 9. Jickling GC Liu D Stamova B Ander BP Zhan X Lu A Sharp FR Hemorrhagic transformation after ischemic stroke in animals and humans J Cereb Blood Flow Metab 2014 34 185 99 10.1038/jcbfm.2013.203 24281743 10. Knight RA Barker PB Fagan SC Li Y Jacobs MA Welch KM Prediction of impending hemorrhagic transformation in ischemic stroke using magnetic resonance imaging in rats Stroke 1998 29 144 51 10.1161/01.STR.29.1.144 9445344 11. Ozkul-Wermester O Guegan-Massardier E Triquenot A Borden A Perot G Gerardin E Increased blood–brain barrier permeability on perfusion computed tomography predicts hemorrhagic transformation in acute ischemic stroke Eur Neurol 2014 72 45 53 10.1159/000358297 24853726 12. Khatri R McKinney AM Swenson B Janardhan V Blood–brain barrier, reperfusion injury, and hemorrhagic transformation in acute ischemic stroke Neurology 2012 79 S52 7 10.1212/WNL.0b013e3182697e70 23008413 13. Baggio LL Drucker DJ Biology of incretins: GLP-1 and GIP Gastroenterology 2007 132 2131 57 10.1053/j.gastro.2007.03.054 17498508 14. Zhang H Meng J Li X Zhou S Qu D Wang N Jia M Ma X Luo X Pro-GLP-1, a Pro-drug of GLP-1, is neuroprotective in cerebral ischemia Eur J Pharm Sci 2015 70 82 91 10.1016/j.ejps.2015.01.010 25640912 15. Teramoto S Miyamoto N Yatomi K Tanaka Y Oishi H Arai H Hattori N Urabe T Exendin-4, a glucagon-like peptide-1 receptor agonist, provides neuroprotection in mice transient focal cerebral ischemia J Cereb Blood Flow Metab 2011 31 1696 705 10.1038/jcbfm.2011.51 21487412 16. Guturi KK Mandal T Chatterjee A Sarkar M Bhattacharya S Chatterjee U Ghosh MK Mechanism of beta-catenin-mediated transcriptional regulation of epidermal growth factor receptor expression in glycogen synthase kinase 3 beta-inactivated prostate cancer cells J Biol Chem 2012 287 18287 96 10.1074/jbc.M111.324798 22493441 17. Wang WC Tsai JJ Kuo CY Chen HM Kao SH Non-proteolytic house dust mite allergen, Der p 2, upregulated expression of tight junction molecule claudin-2 associated with Akt/GSK-3beta/beta-catenin signaling pathway J Cell Biochem 2011 112 1544 51 10.1002/jcb.23067 21344483 18. Liebner S Corada M Bangsow T Babbage J Taddei A Czupalla CJ Reis M Felici A Wolburg H Fruttiger M Wnt/beta-catenin signaling controls development of the blood–brain barrier J Cell Biol 2008 183 409 17 10.1083/jcb.200806024 18955553 19. Krafft PR Altay O Rolland WB Duris K Lekic T Tang J Zhang JH alpha7 nicotinic acetylcholine receptor agonism confers neuroprotection through GSK-3beta inhibition in a mouse model of intracerebral hemorrhage Stroke 2012 43 844 50 10.1161/STROKEAHA.111.639989 22207510 20. Linseman DA Butts BD Precht TA Phelps RA Le SS Laessig TA Bouchard RJ Florez-McClure ML Heidenreich KA Glycogen synthase kinase-3beta phosphorylates Bax and promotes its mitochondrial localization during neuronal apoptosis J Neurosci 2004 24 9993 10002 10.1523/JNEUROSCI.2057-04.2004 15525785 21. Zhao R Zhang Z Song Y Wang D Qi J Wen S Implication of phosphatidylinositol-3 kinase/Akt/glycogen synthase kinase-3beta pathway in ginsenoside Rb1’s attenuation of beta-amyloid-induced neurotoxicity and tau phosphorylation J Ethnopharmacol 2011 133 1109 16 10.1016/j.jep.2010.11.054 21129477 22. Maixner DW Weng HR The Role of Glycogen Synthase Kinase 3 Beta in Neuroinflammation and Pain J Pharm Pharmacol (Los Angel) 2013 1 001 25309941 23. Morales-Garcia JA Susin C Alonso-Gil S Perez DI Palomo V Perez C Conde S Santos A Gil C Martinez A Perez-Castillo A Glycogen synthase kinase-3 inhibitors as potent therapeutic agents for the treatment of Parkinson disease ACS Chem Neurosci 2013 4 350 60 10.1021/cn300182g 23421686 24. Phiel CJ Wilson CA Lee VM Klein PS GSK-3alpha regulates production of Alzheimer’s disease amyloid-beta peptides Nature 2003 423 435 9 10.1038/nature01640 12761548 25. Foerch C Arai K Jin G Park KP Pallast S van Leyen K Lo EH Experimental model of warfarin-associated intracerebral hemorrhage Stroke 2008 39 3397 404 10.1161/STROKEAHA.108.517482 18772448 26. Kitashoji A Egashira Y Mishiro K Suzuki Y Ito H Tsuruma K Shimazawa M Hara H Cilostazol ameliorates warfarin-induced hemorrhagic transformation after cerebral ischemia in mice Stroke 2013 44 2862 8 10.1161/STROKEAHA.113.001183 23881959 27. Lin TN He YY Wu G Khan M Hsu CY Effect of brain edema on infarct volume in a focal cerebral ischemia model in rats Stroke 1993 24 117 21 10.1161/01.STR.24.1.117 8418534 28. Li X Blizzard KK Zeng Z DeVries AC Hurn PD McCullough LD Chronic behavioral testing after focal ischemia in the mouse: functional recovery and the effects of gender Exp Neurol 2004 187 94 104 10.1016/j.expneurol.2004.01.004 15081592 29. Doeppner TR Kaltwasser B Bahr M Hermann DM Effects of neural progenitor cells on post-stroke neurological impairment-a detailed and comprehensive analysis of behavioral tests Front Cell Neurosci 2014 8 338 25374509 30. Li L, McBride DW, Doycheva D, Dixon BJ, Krafft PR, Zhang JH, Tang J. G-CSF attenuates neuroinflammation and stabilizes the blood–brain barrier via the PI3K/Akt/GSK-3beta signaling pathway following neonatal hypoxia-ischemia in rats. Exp Neurol. 2015. 31. Souza DG Pinho V Pesquero JL Lomez ES Poole S Juliano L Correa A Jr de A Castro MS Teixeira MM de Castro MS Teixeira MM Role of the bradykinin B2 receptor for the local and systemic inflammatory response that follows severe reperfusion injury Br J Pharmacol 2003 139 129 39 10.1038/sj.bjp.0705200 12746231 32. Piao F Ma N Hiraku Y Murata M Oikawa S Cheng F Zhong L Yamauchi T Kawanishi S Yokoyama K Oxidative DNA damage in relation to neurotoxicity in the brain of mice exposed to arsenic at environmentally relevant levels J Occup Health 2005 47 445 9 10.1539/joh.47.445 16230839 33. Lemiere S Cossu-Leguille C Charissou AM Vasseur P DNA damage (comet assay) and 8-oxodGuo (HPLC-EC) in relation to oxidative stress in the freshwater bivalve Unio tumidus Biomarkers 2005 10 41 57 10.1080/13547500500038783 16097392 34. Soltys Z Ziaja M Pawlinski R Setkowicz Z Janeczko K Morphology of reactive microglia in the injured cerebral cortex. Fractal analysis and complementary quantitative methods J Neurosci Res 2001 63 90 7 10.1002/1097-4547(20010101)63:1<90::AID-JNR11>3.0.CO;2-9 11169618 35. Monif M Reid CA Powell KL Drummond KJ O’Brien TJ Williams DA Interleukin-1beta has trophic effects in microglia and its release is mediated by P2X7R pore J Neuroinflammation 2016 13 173 10.1186/s12974-016-0621-8 27364756 36. Hovens I Nyakas C Schoemaker R A novel method for evaluating microglial activation using ionized calcium-binding adaptor protein-1 staining: cell body to cell size ratio Neuroimmunol Neuroinflammation 2014 1 82 10.4103/2347-8659.139719 37. Rossa J Ploeger C Vorreiter F Saleh T Protze J Gunzel D Wolburg H Krause G Piontek J Claudin-3 and claudin-5 protein folding and assembly into the tight junction are controlled by non-conserved residues in the transmembrane 3 (TM3) and extracellular loop 2 (ECL2) segments J Biol Chem 2014 289 7641 53 10.1074/jbc.M113.531012 24478310 38. Jiao H Wang Z Liu Y Wang P Xue Y Specific role of tight junction proteins claudin-5, occludin, and ZO-1 of the blood–brain barrier in a focal cerebral ischemic insult J Mol Neurosci 2011 44 130 9 10.1007/s12031-011-9496-4 21318404 39. Taddei A Giampietro C Conti A Orsenigo F Breviario F Pirazzoli V Potente M Daly C Dimmeler S Dejana E Endothelial adherens junctions control tight junctions by VE-cadherin-mediated upregulation of claudin-5 Nat Cell Biol 2008 10 923 34 10.1038/ncb1752 18604199 40. Elyaman W Yardin C Hugon J Involvement of glycogen synthase kinase-3beta and tau phosphorylation in neuronal Golgi disassembly J Neurochem 2002 81 870 80 10.1046/j.1471-4159.2002.00838.x 12065646 41. Leis H Segrelles C Ruiz S Santos M Paramio JM Expression, localization, and activity of glycogen synthase kinase 3beta during mouse skin tumorigenesis Mol Carcinog 2002 35 180 5 10.1002/mc.10087 12489109 42. Ramirez SH Fan S Dykstra H Rom S Mercer A Reichenbach NL Gofman L Persidsky Y Inhibition of glycogen synthase kinase 3beta promotes tight junction stability in brain endothelial cells by half-life extension of occludin and claudin-5 PLoS One 2013 8 10.1371/journal.pone.0055972 23418486 43. Krafft PR Caner B Klebe D Rolland WB Tang J Zhang JH PHA-543613 preserves blood–brain barrier integrity after intracerebral hemorrhage in mice Stroke 2013 44 1743 7 10.1161/STROKEAHA.111.000427 23613493 44. Tsoyi K Jang HJ Nizamutdinova IT Park K Kim YM Kim HJ Seo HG Lee JH Chang KC PTEN differentially regulates expressions of ICAM-1 and VCAM-1 through PI3K/Akt/GSK-3beta/GATA-6 signaling pathways in TNF-alpha-activated human endothelial cells Atherosclerosis 2010 213 115 21 10.1016/j.atherosclerosis.2010.07.061 20864106 45. Tsoyi K Kim WS Kim YM Kim HJ Seo HG Lee JH Yun-Choi HS Chang KC Upregulation of PTEN by CKD712, a synthetic tetrahydroisoquinoline alkaloid, selectively inhibits lipopolysaccharide-induced VCAM-1 but not ICAM-1 expression in human endothelial cells Atherosclerosis 2009 207 412 9 10.1016/j.atherosclerosis.2009.05.012 19540498 46. Yamada S Funada T Shibata N Kobayashi M Kawai Y Tatsuda E Furuhata A Uchida K Protein-bound 4-hydroxy-2-hexenal as a marker of oxidized n-3 polyunsaturated fatty acids J Lipid Res 2004 45 626 34 10.1194/jlr.M300376-JLR200 14729859 47. Wolf PA Abbott RD Kannel WB Atrial fibrillation as an independent risk factor for stroke: the Framingham Study Stroke 1991 22 983 8 10.1161/01.STR.22.8.983 1866765 48. Hart RG Pearce LA Aguilar MI Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation Ann Intern Med 2007 146 857 67 10.7326/0003-4819-146-12-200706190-00007 17577005 49. Kelley RE Berger JR Alter M Kovacs AG Cerebral ischemia and atrial fibrillation: prospective study Neurology 1984 34 1285 91 10.1212/WNL.34.10.1285 6541300 50. Perry T Lahiri DK Sambamurti K Chen D Mattson MP Egan JM Greig NH Glucagon-like peptide-1 decreases endogenous amyloid-beta peptide (Abeta) levels and protects hippocampal neurons from death induced by Abeta and iron J Neurosci Res 2003 72 603 12 10.1002/jnr.10611 12749025 51. During MJ Cao L Zuzga DS Francis JS Fitzsimons HL Jiao X Bland RJ Klugmann M Banks WA Drucker DJ Haile CN Glucagon-like peptide-1 receptor is involved in learning and neuroprotection Nat Med 2003 9 1173 9 10.1038/nm919 12925848 52. Cross DA Alessi DR Cohen P Andjelkovich M Hemmings BA Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B Nature 1995 378 785 9 10.1038/378785a0 8524413 53. Noshita N Lewen A Sugawara T Chan PH Evidence of phosphorylation of Akt and neuronal survival after transient focal cerebral ischemia in mice J Cereb Blood Flow Metab 2001 21 1442 50 10.1097/00004647-200112000-00009 11740206 54. Lim JC Kania KD Wijesuriya H Chawla S Sethi JK Pulaski L Romero IA Couraud PO Weksler BB Hladky SB Barrand MA Activation of beta-catenin signalling by GSK-3 inhibition increases p-glycoprotein expression in brain endothelial cells J Neurochem 2008 106 1855 65 18624906 55. Iadecola C Anrather J The immunology of stroke: from mechanisms to translation Nat Med 2011 17 796 808 10.1038/nm.2399 21738161 56. Muir KW Tyrrell P Sattar N Warburton E Inflammation and ischaemic stroke Curr Opin Neurol 2007 20 334 42 10.1097/WCO.0b013e32813ba151 17495630 57. Fukao T Koyasu S PI3K and negative regulation of TLR signaling Trends Immunol 2003 24 358 63 10.1016/S1471-4906(03)00139-X 12860525 58. Broughton BR Reutens DC Sobey CG Apoptotic mechanisms after cerebral ischemia Stroke 2009 40 e331 9 10.1161/STROKEAHA.108.531632 19182083

PMC005xxxxxx/PMC5002168.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 69410.1186/s12974-016-0694-4Short Reportα-Synuclein deficiency promotes neuroinflammation by increasing Th1 cell-mediated immune responses Ettle Benjamin ben.ettle@uk-erlangen.de 1Kuhbandner Kristina kristina.kuhbandner@uk-erlangen.de 2Jörg Stefanie stefanie.joerg@uk-erlangen.de 2Hoffmann Alana alana.hoffmann@uk-erlangen.de 1Winkler Jürgen juergen.winkler@uk-erlangen.de 1Linker Ralf A. ralf.linker@uk-erlangen.de 21 Department of Molecular Neurology, University Hospital Erlangen, Friedrich-Alexander-University Erlangen-Nürnberg, Schwabachanlage 6, 91054 Erlangen, Germany 2 Department of Neurology, University Hospital Erlangen, Friedrich-Alexander-University Erlangen-Nürnberg, Schwabachanlage 6, 91054 Erlangen, Germany 26 8 2016 26 8 2016 2016 13 1 2015 5 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 Increased α-synuclein immunoreactivity has been associated with inflammatory activity in multiple sclerosis (MS) lesions, but the function of α-synuclein in neuroinflammation remains unknown. The aim of this study was to examine the role of α-synuclein in immunological processes in murine experimental autoimmune encephalomyelitis (EAE) as a model of MS. Findings We studied EAE in wildtype (aSyn+/+) and α-synuclein knockout (aSyn−/−) mice on a C57BL/6N background. In the spleen and spinal cord of aSyn+/+ mice, we observed a gradual reduction of α-synuclein expression during EAE, starting already in the pre-symptomatic disease phase. Compared to aSyn+/+ mice, aSyn−/− mice showed an earlier onset of symptoms but no differences in symptom severity at the peak of disease. Earlier symptom onset was accompanied by increased spinal cord infiltration of CD4+ T cells, predominantly of interferon-γ-producing T helper 1 (Th1) cells, and reduced infiltration of regulatory T cells, whereas antigen-presenting cells were unaltered. Pre-symptomatically, aSyn−/− mice exhibited hyperproliferative CD4+ splenocytes consistent with increased splenic interleukin-2 mRNA expression, resulting in increased numbers of Th1 cells in the spleen at the onset of symptoms. Conclusions Our findings indicate a functional role of α-synuclein in early EAE by increasing Th1 cell-mediated immune response. Keywords α-SynucleinExperimental autoimmune encephalomyelitisNeuroinflammationMultiple sclerosisInterdisciplinary Center for Clinical Research ErlangenE24E24Winkler Jürgen Linker Ralf A. German Parkinson Associationhttp://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftRTG 2162issue-copyright-statement© The Author(s) 2016 ==== Body Introduction Multiple sclerosis (MS) and its murine model experimental autoimmune encephalomyelitis (EAE) are characterized by a primarily T cell-mediated autoimmune attack against oligodendrocytes and myelin of the central nervous system (CNS), subsequently leading to axonal degeneration and neuronal loss. However, factors that drive and modify the immune response in MS and EAE are still incompletely defined. A candidate factor acting as modulator of inflammatory processes in MS and EAE is α-synuclein. Under physiological conditions, α-synuclein is predominantly expressed in neurons [1], but expression has also been observed in glia [1, 2] and hematopoietic cells such as T cells and monocytes [3, 4]. In human diseases, α-synuclein accumulates and forms insoluble fibrils as part of intraneuronal and intraoligodendroglial inclusions in chronic, age-related neurodegenerative diseases such as Parkinson’s disease (PD), dementia with Lewy bodies, and multiple system atrophy [5]. Importantly, peripheral and CNS inflammation aggravates degenerative features in these so-called synucleinopathies [6–8]. Given that increased α-synuclein immunoreactivity has been observed in microglia/macrophages within inflammatory-active MS lesions [9], α-synuclein may also be involved in the immune reaction that underlies degenerative changes in MS and EAE. To investigate this hypothesis, we employed EAE in wildtype (aSyn+/+) and α-synuclein knockout (aSyn−/−) mice, focusing on the pre-symptomatic and acute phase of EAE. Our data show an earlier onset of symptoms accompanied by increased T helper 1 (Th1) cell-mediated immune response in aSyn−/− mice, suggesting a role of α-synuclein for neuroinflammation in EAE and MS. Methods aSyn−/− mice [10] were maintained on a C57BL/6N background for more than 10 generations. All mice were kept under standard animal housing conditions with a 12-h day/night cycle and free access to food and water. For active EAE induction, aSyn+/+ (i.e., wildtype C57BL/6N) (Charles River) and aSyn−/− mice received 200 μg myelin oligodendrocyte glycoprotein (MOG)35–55 and 200 μg complete Freund’s adjuvant (CFA) subcutaneously. Pertussis toxin (200 ng) was applied intraperitoneally at the time of immunization and 48 h later. Mice were daily weighed and scored for clinical signs using a 10-point scale as described previously [11]. The spinal cord and spleen of aSyn−/− and aSyn+/+ mice were removed on day 10 or day 14 after EAE induction. Splenic single-cell suspensions were isolated by enzymatic degradation with DNaseI (10 mg/ml, Roche) and Liberase TL (1.67 Wünsch units/ml, Roche) or by mechanical dissociation [12]. CNS mononuclear cells were prepared by mechanical disruption and subsequent percoll gradient centrifugation as described [12]. Splenocytes and CNS mononuclear cells were analyzed by multicolor flow cytometry using the following fluorochrome-conjugated antibodies: anti-CD3 (17A2), anti-CD4 (RM4-5), anti-CD8a (53-6.7), anti-CD11b (M1/70), anti-CD11c (HL3), anti-CD44 (IM7), anti-CD16/32 (93), anti-CD25 (PC 61.5), anti-CD62L (MEL-14), anti-FoxP3 (FJK-16s), anti-IFN-γ (XMG1.2), and anti-IL-17a (eBio17B7). For intracellular staining, the FoxP3 staining kit (eBioscience) was used. Flow cytometry was performed on a FACS Canto II (BD), and results were analyzed with FlowJo software (Tree Star Inc.). For lymphocyte proliferation assay, splenic single-cell suspensions were isolated 10 days after EAE induction and stained with proliferation dye eFluor® 450 (eBioscience) according to the manufacturer’s instructions. 1 × 106 cells were seeded in 24-well plates and cultured with media alone or in the presence of 20 μg/ml MOG35–55 peptide or 1.25 μg/ml Concanavalin A as control. After 96 h, cells were harvested and processed for flow cytometry analysis. For in vitro T cell differentiation, splenic T cells were isolated by MACS and re-suspended in MACS buffer at 3 × 107 cells/ml. Sorted naive T cells (CD4+CD62L+CD44loCD25) were stimulated with 2 μg/ml anti-CD3 and 2 μg/ml anti-CD28. For Th1 cell differentiation, naive T cells were cultured for 4 days with interleukin (IL)-12 (20 ng/ml) and anti-IL-4 (10 mg/ml). For T helper 17 (Th17) cell differentiation, cells were cultured in the presence of IL-6 (40 ng/ml) and rhTGF-β1 (2 ng/ml) for 4 days. For Treg cell differentiation, only rhTGF-β1 (10 ng/ml) was added to culture media. To evaluate protein levels of α-synuclein, Western blot was performed as previously described [13]. Briefly, total protein was extracted from spinal cord tissue and MACS-purified CD11b+, CD11c+, and CD4+ splenocytes of aSyn+/+ and/or aSyn−/− mice either under native conditions or with active EAE prior to symptom onset (10 days after immunization) using radioimmunoprecipitation buffer. α-Synuclein protein was detected by using mouse anti-α-synuclein (Syn1; BD; 610786; 1:300). Mouse anti-β-actin (1:2000; ab8226) was obtained from Abcam. Donkey anti-mouse 488 (Dianova; 1:1000) was used as fluorescence-conjugated secondary antibody. Signal was detected with Fusion FX7 detection system (PeqLab), and densitometric quantifications were performed using the Bio1D software (Vilber Lourmat) by normalizing signals to β-actin. For real-time polymerase chain reaction (RT-PCR), total RNA was isolated from spleen and spinal cord tissue with the RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. Complementary DNA (cDNA) was prepared using the GoScript™ reverse transcription system (Promega). RT-PCR reactions were measured in triplicates either on a LightCycler 480 system (Roche) with the SSO Fast EvaGreen Supermix (Bio-Rad) [13] or on a qTower real-time PCR system (Analytik Jena) [12]. Relative quantification was performed using the ΔΔCT method after normalization to β-actin. Graphical and statistical analysis was performed with GraphPad Prism® 5 (GraphPad software). Data were analyzed either by ANOVA or Student’s t test. Comparison of survival curves (Gehan-Breslow-Wilcoxon) was performed to evaluate the percentage of disease-free mice. Data are presented as mean ± standard error of mean (SEM), and *p < 0.05, **p < 0.01, or ***p < 0.001 was considered to be statistically significant. Results We initially examined α-synuclein messenger RNA (mRNA) expression in the spinal cord and spleen in the pre-symptomatic and acute phase of EAE in aSyn+/+ (i.e., wildtype C57BL/6N) mice (Fig. 1a). Both in the spleen (Fig. 1b) and spinal cord (Fig. 1c), α-synuclein mRNA expression gradually declined during EAE as compared to native tissues. Reduced α-synuclein mRNA expression was already detected in the pre-symptomatic phase of EAE and reached its lowest expression (reduction by ~90 %) at the peak of disease. Matching the dynamic of α-synuclein gene expression, spinal cord α-synuclein protein levels were significantly reduced by 28 ± 4 % in the pre-symptomatic phase of EAE (Fig. 1d).Fig. 1 Reduced α-synuclein expression during EAE. a Clinical course of MOG35–55-induced EAE in C57BL/6N mice. Arrows indicate days prior to or post symptom onset on which α-synuclein mRNA expression was analyzed in the spleen (b) and spinal cord (c). b, c RT-PCR revealed a gradual downregulation of α-synuclein mRNA expression in the spleen and spinal cord (n = 4; *p < 0.05, **p < 0.01). d Western blot of spinal cord lysates from EAE mice 10 days after immunization (i.e., prior to symptom onset) demonstrates a significant reduction of α-synuclein (α-syn) protein as compared to the native spinal cord (n = 4; **p < 0.01) The pre-symptomatic and profound downregulation of α-synuclein expression in EAE suggests that α-synuclein may influence the immune response during the acute phase of EAE. To test this hypothesis, we induced EAE in aSyn−/− and aSyn+/+ mice and rated mice for clinical signs during EAE. Unchallenged, aSyn−/− do not display any obvious phenotype and show similar body weights as compared to age-matched aSyn+/+ mice. After the induction of EAE, disease onset was significantly earlier (1.2 ± 0.3 days) in aSyn−/− compared to aSyn+/+ mice as evidenced by lower number of disease-free mice (Fig. 2a) and a significantly more severe clinical course (Fig. 2b). At the peak of disease, all mice suffered from severe gait ataxia and no difference in clinical scores between aSyn−/− and aSyn+/+ mice was observed (Fig. 2b). There were no differences in the overall disease incidence and mortality between both groups.Fig. 2 Earlier symptom onset in aSyn−/− mice accompanied by increased spinal cord infiltration of Th1 cells during EAE. a Percentage of mice without symptoms during the onset phase of EAE is shown. The aSyn−/− cohort (dashed line) exhibited significantly less disease-free mice compared to the aSyn+/+ cohort (solid line) (n = 15/16; *p < 0.05). b aSyn−/− mice (white dots, dashed line) showed a significantly more severe clinical course during the initial phase of EAE than aSyn+/+ mice (black dots, solid line) (n = 15/16; **p < 0.01). Symptom severity at disease peak was not different between both groups. c Western blot analysis reveals the presence of α-synuclein (α-syn) in antigen-presenting cells (CD11b+ and CD11c+) and CD4+ T cells isolated from the spleen of aSyn+/+ mice. Spinal cord (SC) lysates of aSyn+/+ and aSyn−/− mice were loaded as positive and negative controls for α-synuclein immunoreactivity. d Representative contour plots of ex vivo flow cytometry to determine frequencies of IL-17A+ and/or IFN-γ+ cells within the spinal cord of aSyn+/+ (upper panel) and aSyn−/− (lower panel) mice 2 days after symptom onset of EAE are depicted. e In the spinal cord of aSyn−/− mice (gray bars), there were significantly higher frequencies of IFN-γ+ Th1 and IFN-γ+/IL-17A+ Th1/Th17 cells as compared to aSyn+/+ mice (black bars) (n = 4; *p < 0.05, **p < 0.01). f Frequency of regulatory T cells was lower in aSyn−/− mice (n = 4; **p < 0.01). g There was no difference in the frequency of antigen-presenting cells (n = 4; p > 0.05) Symptom onset and severity in EAE are tightly linked to Th1 and/or Th17 cell infiltration into the spinal cord [14]. Moreover, the physiological expression of α-synuclein within antigen-presenting cells (CD11b+ or CD11c+) and CD4+ T cells (Fig. 2c) in wildtype mice suggests that α-synuclein deficiency directly impacts immune responses during EAE in aSyn−/− mice. Thus, to dissect the immunological mechanism responsible for earlier symptom onset in aSyn−/− mice, we next performed ex vivo flow cytometry to determine frequencies of CD4+ T cells within the spinal cord (Fig. 2d). Total CD4+ cell frequency was twofold but not significantly increased in aSyn−/− spinal cords (aSyn−/− 32.4 ± 8.9, aSyn+/+ 14.6 ± 5.1, p > 0.05). However, aSyn−/− mice exhibited a significantly increased frequency of interferon gamma (IFN-γ)-producing Th1 cells compared to aSyn+/+ mice, whereas IL-17A-producing Th17 cells were not significantly different (Fig. 2e). Notably, T cells that produced both IFN-γ and IL-17A were also significantly increased in aSyn−/− mice (Fig. 2e). Further characterization of infiltrating immune cells revealed a reduced frequency of regulatory T cells in aSyn−/− compared to aSyn+/+ mice (Fig. 2f), but neither alterations of CD11b+/CD11c− nor of CD11b−/CD11c+ antigen-presenting cells were detected (Fig. 2g). In a next step, we investigated primary immune responses in the periphery by phenotyping splenic CD4+ cells prior to (i.e., day 10 post immunization) and 2 days after the onset of symptoms (Fig. 3). Overall, we observed an increased frequency of CD4+ T cells in aSyn−/− as compared to aSyn+/+ mice. Matching the immunological profile in the spinal cord, a higher frequency of IFN-γ-producing Th1 cells was detected in the spleen of aSyn−/− compared to aSyn+/+ mice, particularly 2 days after the occurrence of the first symptoms (Fig. 3a). In line, IL-17A-producing Th17 cells were altered neither prior to nor 2 days after the onset of symptoms (Fig. 3b). Furthermore, T cells producing both IFN-γ and IL-17A were also significantly increased in the spleen of aSyn−/− mice but only after symptom onset (Fig. 3c). There were no differences in the frequency of regulatory T cells (Fig. 3d). In parallel, α-synuclein deficiency resulted in an increase of activated splenic effector T cells prior to symptom onset with significantly increased CD4+/CD44high (aSyn−/− 52.13 ± 0.82, aSyn+/+ 48.27 ± 1.27; p < 0.05) and CD4+/CD25high cell frequency (aSyn−/− 55.03 ± 1.03, aSyn+/+ 51.60 ± 0.99; p < 0.05). An in vitro T cell differentiation assay with splenic T cells revealed no difference regarding the differentiation into T helper cell subtypes (i.e., Th1, Th17, regulatory T cells) between aSyn−/− mice and controls (Fig. 3e).Fig. 3 Increased splenic Th1 cell frequency in aSyn−/− mice during the onset phase of EAE. Ex vivo flow cytometry revealed a higher frequency of splenic IFN-γ+ Th1 cells 2 days after disease onset in aSyn−/− mice as compared to aSyn+/+ mice (n = 4; *p < 0.05) (a) but no differences in the frequency of IL-17A+ Th17 cells (n = 4; p > 0.05) (b). The spleen of aSyn−/− mice also exhibited higher frequencies of IFN-γ+/IL-17A+ Th1/Th17 cells (n = 4; ***p < 0.01) (c), whereas regulatory T cells remained unaltered (n = 4; p > 0.05) (d). e In vitro differentiation under Th1-, Th17-, or regulatory T cell-polarizing conditions was similar for naive CD4+ T cells derived from aSyn−/− and aSyn+/+ mice (n = 3; p > 0.05) In line with the increased frequency of splenic CD4+ T cells, CD4+ splenocytes isolated from aSyn−/− mice showed enhanced proliferation upon ex vivo re-stimulation with MOG35–55 peptide as compared to aSyn+/+ splenocytes (Fig. 4a). Consisting with this hyperproliferative phenotype of aSyn−/− splenocytes, the levels of IL-2 mRNA were significantly increased in the spleen of aSyn−/− mice in EAE prior to symptom onset (Fig. 4b), whereas IL-4 expression was unaltered (aSyn−/− 1.20 ± 0.37, aSyn+/+ 1.44 ± 0,41; p > 0.05).Fig. 4 Increased proliferation of α-synuclein-deficient splenocytes associated with increased splenic IL-2 mRNA expression in aSyn−/− mice prior to onset of EAE. a Proliferation of CD4+ cells was measured by flow cytometry in splenocyte cultures after ex vivo recall with MOG35–55 (splenocytes harvested on day 10 post immunization in aSyn−/− and aSyn+/+ mice), revealing significantly increased proliferation of aSyn−/−-derived splenocytes (n = 4; *p < 0.05). b RT-PCR was performed on cDNA derived from total RNA that was isolated from the spleen of aSyn−/− and aSyn+/+ mice 10 days after induction of EAE. IL-2 mRNA expression was significantly increased in the spleen of aSyn−/− mice (n = 4; *p < 0.05) Discussion Our study reveals that α-synuclein expression gradually declines during EAE of C57BL/6N mice, beginning already in the pre-symptomatic phase. Furthermore, EAE in aSyn−/− mice is characterized by earlier symptom onset and enhanced T cell—in particular Th1 cell—infiltration into the spinal cord, consisting of higher splenic T cell frequencies and increased T cell proliferation. Our data on the role of α-synuclein during the acute phase of EAE extends previous descriptive work on α-synuclein in MS and EAE. In demyelinated MS lesions with active inflammation, but not in chronic inactive lesions, increased α-synuclein immunoreactivity has been observed in microglia/macrophages as well as in neurons and oligodendrocytes [9]. Interestingly, elevated α-synuclein expression in neurons and glia during MOG-induced EAE in rats has been correlated with degree of inflammation and demyelination [15]. In a small study investigating cerebrospinal fluid, α-synuclein concentrations were significantly higher in patients with inflammatory demyelinating disease than in patients suffering from PD or controls [16]. The phenotypical characterization of T cells within the spinal cord at the onset of EAE showed increased Th1 cell infiltration in aSyn−/− mice. Matching this finding, the spleen of aSyn−/− mice also exhibited higher frequencies of Th1 cells. Furthermore, increased spinal cord Th1 cell infiltration was accompanied by lower frequencies of anti-inflammatory neuroprotective regulatory T cells, whereas Th17 and antigen-presenting cells remained unaffected. These data suggest that α-synuclein deficiency promotes Th1 cell activation and negatively impacts regulatory T cells, thereby exacerbating Th1-mediated immune response in EAE. In agreement with this interpretation, a functional role of α-synuclein for T cell activation has previously been suggested as splenic CD4+ cells isolated from α-synuclein-deficient mice showed increased activity [17]. Moreover, increased CNS regulatory T cell frequency has been observed upon immunization with nitrated α-synuclein protein in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) model of PD [18]. Results from splenocyte recall assays on day 10 after EAE induction imply that enhanced T cell proliferation underlies expansion of Th1 cells in aSyn−/− mice. In good agreement with this finding, we observed increased mRNA expression of IL-2 as central activator of T cell proliferation [19] in the spleen of aSyn−/− mice on day 10 after EAE induction, whereas mRNA expression of IL-4 as key cytokine during Th2 cell-mediated immune response was unaltered. These data suggest that α-synuclein deficiency leads to increased IL-2 production, thereby expanding the population of Th1 cells. In line with this interpretation, immunization with nitrated α-synuclein protein in the MPTP model of PD elicits decreased Th1 cell-mediated inflammation and reduction of IL-2 levels [18]. Overall, our data show that endogenous α-synuclein plays a functional role for immunological processes during early EAE as a new regulator of Th1 responses in neuroinflammation. Hence, our study identifies α-synuclein as an interesting new target for modulating the immune response in MS. Abbreviations aSyn+/+Wildtype C57BL/6N aSyn−/−α-Synuclein knockout CFAComplete Freund’s adjuvant CNSCentral nervous system EAEExperimental autoimmune encephalomyelitis IFN-γInterferon gamma ILInterleukin MOGMyelin oligodendrocyte glycoprotein MPTP1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine MSMultiple sclerosis PDParkinson’s disease RT-PCRReal-time polymerase chain reaction Th1T helper 1 Th17T helper 17 Acknowledgements RAL holds an endowed professorship supported by Novartis. Funding This work was supported by the Interdisciplinary Center for Clinical Research Erlangen (IZKF; E18 and E24). BE is supported by a research grant of the German Parkinson Association. KK (MSc) and AH (MSc) are graduate students of the Research Training Group 2162 (“Neurodevelopment and Vulnerability of the Central Nervous System”) funded by the Deutsche Forschungsgemeinschaft (DFG). Availability of data and materials The datasets supporting the conclusions of this article are included within the article. Authors’ contributions BE, SJ, JW, and RAL conceived the project. KK, BE, SJ, and AH performed experiments. BE and KK 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 All animal experiments were performed in accordance with the German laws for animal protection and were approved by the local ethics committees at the Friedrich-Alexander-University Erlangen-Nuremberg. ==== Refs References 1. Mori F Tanji K Yoshimoto M Takahashi H Wakabayashi K Demonstration of alpha-synuclein immunoreactivity in neuronal and glial cytoplasm in normal human brain tissue using proteinase K and formic acid pretreatment Exp Neurol 2002 176 1 98 104 10.1006/exnr.2002.7929 12093086 2. Djelloul M Holmqvist S Boza-Serrano A Azevedo C Yeung MS Goldwurm S Alpha-synuclein expression in the oligodendrocyte lineage: an in vitro and in vivo study using rodent and human models Stem Cell Reports 2015 5 2 174 84 10.1016/j.stemcr.2015.07.002 26235891 3. Barbour R Kling K Anderson JP Banducci K Cole T Diep L Red blood cells are the major source of alpha-synuclein in blood Neurodegener Dis 2008 5 2 55 9 10.1159/000112832 18182779 4. Shin EC Cho SE Lee DK Hur MW Paik SR Park JH Expression patterns of alpha-synuclein in human hematopoietic cells and in Drosophila at different developmental stages Mol Cells 2000 10 1 65 70 10.1007/s10059-000-0065-x 10774749 5. Spillantini MG Crowther RA Jakes R Cairns NJ Lantos PL Goedert M Filamentous alpha-synuclein inclusions link multiple system atrophy with Parkinson’s disease and dementia with Lewy bodies Neurosci Lett 1998 251 3 205 8 10.1016/S0304-3940(98)00504-7 9726379 6. Grozdanov V Bliederhaeuser C Ruf WP Roth V Fundel-Clemens K Zondler L Inflammatory dysregulation of blood monocytes in Parkinson’s disease patients Acta Neuropathol 2014 128 5 651 63 10.1007/s00401-014-1345-4 25284487 7. Lim S Chun Y Sung Lee J Lee SJ Neuroinflammation in synucleinopathies Brain Pathol 2016 26 3 404 9 10.1111/bpa.12371 26940152 8. Sommer A Fadler T Dorfmeister E Hoffmann AC Xiang W Winner B Infiltrating T lymphocytes reduce myeloid phagocytosis activity in synucleinopathy model J Neuroinflammation 2016 13 1 174 10.1186/s12974-016-0632-5 27364890 9. Lu JQ Fan Y Mitha AP Bell R Metz L Moore GR Association of alpha-synuclein immunoreactivity with inflammatory activity in multiple sclerosis lesions J Neuropathol Exp Neurol 2009 68 2 179 89 10.1097/NEN.0b013e318196e905 19151622 10. Abeliovich A Schmitz Y Farinas I Choi-Lundberg D Ho WH Castillo PE Mice lacking alpha-synuclein display functional deficits in the nigrostriatal dopamine system Neuron 2000 25 1 239 52 10.1016/S0896-6273(00)80886-7 10707987 11. Linker RA Maurer M Gaupp S Martini R Holtmann B Giess R CNTF is a major protective factor in demyelinating CNS disease: a neurotrophic cytokine as modulator in neuroinflammation Nat Med 2002 8 6 620 4 10.1038/nm0602-620 12042814 12. Haghikia A Jorg S Duscha A Berg J Manzel A Waschbisch A Dietary fatty acids directly impact central nervous system autoimmunity via the small intestine Immunity 2015 43 4 817 29 10.1016/j.immuni.2015.09.007 26488817 13. Ettle B Reiprich S Deusser J Schlachetzki JC Xiang W Prots I Intracellular alpha-synuclein affects early maturation of primary oligodendrocyte progenitor cells Mol Cell Neurosci 2014 62 68 78 10.1016/j.mcn.2014.06.012 25019582 14. Fletcher JM Lalor SJ Sweeney CM Tubridy N Mills KH T cells in multiple sclerosis and experimental autoimmune encephalomyelitis Clin Exp Immunol 2010 162 1 1 11 10.1111/j.1365-2249.2010.04143.x 20682002 15. Papadopoulos D Ewans L Pham-Dinh D Knott J Reynolds R Upregulation of alpha-synuclein in neurons and glia in inflammatory demyelinating disease Mol Cell Neurosci 2006 31 4 597 612 10.1016/j.mcn.2006.01.007 16503161 16. Wang H Wang K Xu W Wang C Qiu W Zhong X Cerebrospinal fluid alpha-synuclein levels are elevated in multiple sclerosis and neuromyelitis optica patients during replase J Neurochem 2012 122 1 19 23 10.1111/j.1471-4159.2012.07749.x 22469018 17. Shameli A Xiao W Zheng Y Shyu S Sumodi J Meyerson HJ A critical role for alpha-synuclein in development and function of T lymphocytes Immunobiology 2016 221 2 333 40 10.1016/j.imbio.2015.10.002 26517968 18. Reynolds AD Stone DK Hutter JA Benner EJ Mosley RL Gendelman HE Regulatory T cells attenuate Th17 cell-mediated nigrostriatal dopaminergic neurodegeneration in a model of Parkinson’s disease J Immunol 2010 184 5 2261 71 10.4049/jimmunol.0901852 20118279 19. Cornely R Pollock AH Rentero C Norris SE Alvarez-Guaita A Grewal T Annexin A6 regulates interleukin-2-mediated T-cell proliferation Immunol Cell Biol 2016 94 6 543 53 10.1038/icb.2016.15 26853809

PMC005xxxxxx/PMC5002169.txt

==== Front Crit CareCritical Care1364-85351466-609XBioMed Central London 143010.1186/s13054-016-1430-2LetterIs first-line antimicrobial therapy still adequate to treat MRSA in the ICU? A report from a highly endemic country Bassetti Matteo + 39 0432 559353mattba@tin.it 14Righi Elda 1Peghin Maddalena 1Carnelutti Alessia 1Ansaldi Filippo 2Trucchi Cecilia 2Alicino Cristiano 2Tricarichi Enrico Maria 3Del Giacomo Paola 3Tumbarello Mario 31 Infectious Diseases Division, Santa Maria Misericordia University Hospital, Udine, Italy 2 IRCCS AOU San Martino IST, Department of Health Sciences, University of Genoa, Genoa, Italy 3 Institute of Infectious Diseases, Catholic University of the Sacred Heart, Rome, Italy 4 Clinica Malattie Infettive, Azienda Ospedaliero-Universitaria “Santa Maria della Misericordia”, Piazzale S. Maria della Misericordia, n. 15, 33100 Udine, Italy 27 8 2016 27 8 2016 2016 20 1 246© 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 Methicillin-resistant Staphylococcus aureus (MRSA) infections cause great concern in intensive care units (ICUs) [1]. Although strict infection control protocols have reduced staphylococcal colonization, the ICU still represents a reservoir for MRSA infections, playing a role in their circulation to multiple wards and hospitals [2–4]. In critically ill patients, lack of adequate treatment may lead to increased mortality [1]. For this reason, broad-spectrum antimicrobial therapy is often justified among critically ill patients. We retrospectively analyzed the characteristics of S. aureus bloodstream infections (SA-BSI) from two Italian University hospitals during 2010–2014. A total of 17/337 (5 %) were ICU patients; of these, 16 (94 %) had MRSA-BSI compared with 36 % (116/320) from other wards (P < 0.001). Lower adequate first-line therapy (defined as therapy administered within 48 h of the positive blood culture and effective against a susceptible pathogen) and infectious diseases (ID) specialist consultation were documented in ICU versus non-ICU patients (18 versus 60 % and 53 versus 24 %, P < 0.001 and P = 0.02, respectively). When only MRSA infections were considered, adequate therapy and ID consultation remained more common in non-ICU patients (Table 1; difference not significant). Inadequate therapy for non-ICU and ICU patients was mainly associated with beta-lactam use (62/104 versus 10/12, respectively, P = 0.19). Patients with MRSA infections in the ICU displayed a lower Charlson score, longer hospitalizations, higher rates of nosocomial infections, endocarditis, and central venous catheter (CVC) and urinary catheter placement. Source control, including CVC removal, was significantly higher in ICU versus non-ICU patients (Table 1). A stepwise logistic regression analysis identified ICU stay (odds ratio (OR) 19.5, 95 % confidence interval (CI) 3.4–384.2, P < 0.001), presence of intravascular devices other than CVCs for over 72 h (OR 3.5, 95 % CI 1–13.2, P = 0.04), and pulmonary source of infection (OR 3.2, 95 % CI 1.2–9.4, P = 0.02) as factors associated with MRSA-BSI. Overall crude 7- and 30-day mortality was similar for MRSA- (Table 1) and SA-BSI (13 versus 18 % in non-ICU and 25 versus 29 % in ICU patients, respectively). Multivariate analysis identified as independent factors for 7-day mortality among patients with MRSA an inadequate targeted treatment (OR 0.19, 95 % CI 0.04–0.86, P = 0.03), absence of ID consultation (OR 0.17, 95 % CI 0.04–0.6, P = 0.004), and occurrence of endocarditis (OR 4.8, 95 % CI 1.4–17.5, P = 0.01) or septic shock (OR 15.9, 95 % CI 4.6– 66.7, P < 0.001). High Charlson score (OR 1.25, 95 % CI 1.1–1.5, P = 0.004) and septic shock (OR 2.8, 95 % CI 1.0–7.9, P = 0.04) were significantly associated with 30-day mortality.Table 1 Characteristics of MRSA bloodstream infections in patients hospitalized in the ICU compared with other wards Characteristic Non-ICU (n = 116) ICU (n = 16) P value Age, years (median, IQR) 70.5 (56–77) 59 (56–68) 0.07 Males (%) 82/116 (70.7) 10/16 (62.5) 0.50 Charlson score (median, IQR) 6 (3–7) 2 (1–4) <0.001 CVC (>72 h) (%) 58/116 (50) 12/16 (75) 0.05 Other intravascular devices (>72 h) (%) 8/116 (6.9) 2/16 (12.5) 0.35 Urinary catheter (>72 h) (%) 40/116 (34.5) 14/16 (87.5) <0.001 Antimicrobial therapy (<30 days) (%) 50/116 (43.1) 10/16 (62.5) 0.14 Source of infection  Unknown 46/116 (39.6) 0/16 (0) 0.002  CVC 12/116 (10.3) 3/16 (18.8) 0.53  Pulmonary 12/116 (10.3) 1/16 (6.3) 1.00  Endocarditis 15/116 (12.9) 8/16 (50) 0.001  Skin and soft tissue 12/116 (10.3) 3/16 (18.7) 0.53  Other 19/116 (4.3) 1/16 (6.3) 1.00 CVC removal (%) 42/58 (72.4) 12/12 (100) 0.06 Source control (%) 22/70 (31.4) 12/16 (75) 0.003 Hospitalization, days (median, IQR) 30 (18–44) 56 (25–116) 0.01 Acquisition (%)  Community acquired 2/116 (1.7) 0/16 (0) 1.00  Health-care associated 40/116 (34.5) 0/16 (0) <0.001  Hospital-acquired 74/116 (63.8) 16/16 (100) 0.003 Septic shock 22/116 (19) 1/16 (6.3) 0.37 Infectious disease consultation (%) 53/116 (45.7) 4/16 (25) 0.11 Empirical antimicrobial therapy (%) 110/116 (94.8) 16/16 (100) 1.00  Daptomycin 7/104 (6.7) 0/12 (0) 0.35  Glycopeptides 24/104 (23.1) 1/12 (8.3) 0.46 Therapy duration, days (median, IQR) 16 (7-22) 18 (14-27) 0.31 Adequate initial therapy (%) 30/116 (25.9) 2/16 (12.5) 0.39 7-day mortality (%) 27/116 (23.3) 3/16 (18.8) 0.69 30-day mortality 42/116 (36.2) 5/16 (31.3) 0.7 Values are expressed as percentage and median (25th and 75th percentile) CVC central venous catheter, IQR interquartile range The prevalence of MRSA varies widely by geographic region [5]. In our report, overall MRSA rates were comparable to those reported in Italy (33.6 %) by the European Antimicrobial Resistance Surveillance (EARS) in 2014 [5]. Our data highlight that inadequate MRSA first-line therapy can occur in clinical settings known to be at high risk for multi-drug resistant (MDR) infections. Low ID involvement and a priority towards multidrug-resistant Gram-negatives are possible reasons for a reduced first-line use of anti-MRSA compounds in the ICU. Although ICU patients displayed higher rates of inadequate first-line therapy and risk factors associated with increased mortality (e.g., reduced ID consultation and endocarditis) [6], overall mortality was comparable between groups. This may be related to a limited number of patients with septic shock in the ICU group; furthermore, low Charlson scores and higher source control in the ICU group compared with the non-ICU group may have contributed to achieve positive outcomes. In conclusion, our study draws attention to an alarming proportion of first-line inadequate therapy among patients with SA-BSI in the ICU. In this setting, the use of protocols including anti-MRSA agents in patients at risk for Staphylococcus aureus bacteremia (SAB) should be recommended, and clinicians must retain a high level of suspicion for MRSA infections in order to select an appropriate early antimicrobial treatment and ultimately reduce mortality. ==== Refs References 1. Klevens RM Morrison MA Nadle J Invasive methicillin-resistant Staphylococcus aureus infections in the United States JAMA. 2007 298 1763 1771 10.1001/jama.298.15.1763 17940231 2. Mitchell BG Collignon PJ McCann R A major reduction in hospital-onset Staphylococcus aureus bacteremia in Australia-12 years of progress: an observational study Clin Infect Dis 2014 59 7 969 975 10.1093/cid/ciu508 24973314 3. Stenehjem E Stafford C Rimland D Reduction of methicillin-resistant Staphylococcus aureus infection among veterans in Atlanta Infect Control Hosp Epidemiol 2013 34 1 62 68 10.1086/668776 23221194 4. Hoefnagels-Schuermans A Borremans A Peetermans W Origin and transmission of methicillin resistant Staphylococcus aureus in an endemic situation: differences between geriatric and intensive care patients J Hosp Infect. 1997 36 209 222 10.1016/S0195-6701(97)90196-1 9253702 5. European Centre for Disease Prevention and Control (ECDC). Antimicrobial resistance surveillance in Europe. 2014. http://ecdc.europa.eu/en/publications/Publications/antimicrobial-resistance-europe-2014.pdf. Accessed 31 July 2016. 6. Vogel M Schmitz RP Hagel S Infectious disease consultation for Staphylococcus aureus bacteremia--a systematic review and meta-analysis J Infect 2016 72 1 19 28 10.1016/j.jinf.2015.09.037 26453841

PMC005xxxxxx/PMC5002170.txt

==== Front J Exp Clin Cancer ResJ. Exp. Clin. Cancer ResJournal of Experimental & Clinical Cancer Research : CR0392-90781756-9966BioMed Central London 40310.1186/s13046-016-0403-2ResearchMicroarray-based identification of genes associated with cancer progression and prognosis in hepatocellular carcinoma Yin Fuqiang yinfq@mail2.sysu.edu.cn 12Shu Lipei shulipei888@163.com 3Liu Xia realliuxia@sina.com 4Li Ting 516225191@qq.com 1Peng Tao pengtaocn@hotmail.com 3Nan Yueli 331578797@qq.com 5Li Shu 369363170@qq.com 5Zeng Xiaoyun zxyxjw@21cn.com 25Qiu Xiaoqiang xqqiu9999@sina.com 51 Medical Scientific Research Centre, Guangxi Medical University, 22 Shuangyong Rd, Nanning, Guangxi 530021 People’s Republic of China 2 Key Laboratory of High-Incidence-Tumor Prevention and Treatment (Guangxi Medical University), Ministry of Education, 22 Shuangyong Rd, Nanning, Guangxi 530021 People’s Republic of China 3 The First Affiliated Hospital, Guangxi Medical University, 22 Shuangyong Rd, Nanning, Guangxi 530021 People’s Republic of China 4 Centre for Translational Medicine, Guangxi Medical University, 22 Shuangyong Rd, Nanning, Guangxi 530021 People’s Republic of China 5 School of Public Health, Guangxi Medical University, 22 Shuangyong Rd, Nanning, Guangxi 530021 People’s Republic of China 27 8 2016 27 8 2016 2016 35 1 12719 6 2016 9 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 Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths. The average survival and 5-year survival rates of HCC patients still remains poor. Thus, there is an urgent need to better understand the mechanisms of cancer progression in HCC and to identify useful biomarkers to predict prognosis. Methods Public data portals including Oncomine, The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) profiles were used to retrieve the HCC-related microarrays and to identify potential genes contributed to cancer progression. Bioinformatics analyses including pathway enrichment, protein/gene interaction and text mining were used to explain the potential roles of the identified genes in HCC. Quantitative real-time polymerase chain reaction analysis and Western blotting were used to measure the expression of the targets. The data were analysed by SPSS 20.0 software. Results We identified 80 genes that were significantly dysregulated in HCC according to four independent microarrays covering 386 cases of HCC and 327 normal liver tissues. Twenty genes were consistently and stably dysregulated in the four microarrays by at least 2-fold and detection of gene expression by RT-qPCR and western blotting showed consistent expression profiles in 11 HCC tissues compared with corresponding paracancerous tissues. Eleven of these 20 genes were associated with disease-free survival (DFS) or overall survival (OS) in a cohort of 157 HCC patients, and eight genes were associated with tumour pathologic PT, tumour stage or vital status. Potential roles of those 20 genes in regulation of HCC progression were predicted, primarily in association with metastasis. INTS8 was specifically correlated with most clinical characteristics including DFS, OS, stage, metastasis, invasiveness, diagnosis, and age. Conclusion The significantly dysregulated genes identified in this study were associated with cancer progression and prognosis in HCC, and might be potential therapeutic targets for HCC treatment or potential biomarkers for diagnosis and prognosis. Electronic supplementary material The online version of this article (doi:10.1186/s13046-016-0403-2) contains supplementary material, which is available to authorized users. Keywords Hepatocellular carcinomaMicroarrayProgressionPrognosishttp://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China81360448Zeng Xiaoyun http://dx.doi.org/10.13039/501100004607Natural Science Foundation of Guangxi Province2014GXNSFAA118139Zeng Xiaoyun Key Laboratory of High-Incidence-Tumor Prevention and Treatment (Guangxi Medical University), Ministry of EducationGK2015-ZZ03 and GK2014-ZZ03Zeng Xiaoyun Guangxi Outstanding Teachers Training Project for Collegesissue-copyright-statement© The Author(s) 2016 ==== Body Background Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths [1]. There are 750,000 new cases of HCC and nearly 700,000 deaths each year, making this a particularly lethal form of cancer [2]. Over the past decade major progress has been made in our understanding of the risk factors and molecular pathways driving liver carcinogenesis, and these advances have led to substantial opportunities for HCC prevention, surveillance, early diagnosis, prediction of prognosis, and therapy [1]. However, the average survival of HCC patients is normally between 6 and 20 months [3], and long-term prognosis is poor with reported 5-year survival rates ranging from 17 to 53 % [4]. Thus, there is an urgent need to better understand the mechanism of cancer progression and development in HCC and to identify useful biomarkers for diagnosis and prognosis. High-throughput profiling technologies such as microarrays and, more recently, next-generation sequencing have become invaluable tools for biomedical research, and large amounts of data generated by those tools, including mRNA expression, DNA methylation, and microRNA expression, are collected in public archives such as the major public projects The Cancer Genome Atlas (TCGA) [5] and the International Cancer Genome Consortium [6], and the most prominent primary data archives, ArrayExpress [7], Gene Expression Omnibus (GEO) [8], Oncomine [9] and the databases of the International Nucleotide Sequence Database Collaboration [10]. The wide range of those databases, the various ways in which publicly archived gene expression data are being used in support of new studies, and reuse of these public data can be very powerful [11]. In particular, reusing of the data has the potential to predict treatment response and disease progression and was advantageous to develop precision therapies [12]. For example, based on data retrieved from Oncomine, TCGA, and GEO, Liu et al. identified several genes associated with ovarian cancer progression [13] and drug resistance [14]. In a similar manner, we identified that upregulation of E2F transcription factor 3 is associated with poor prognosis in HCC [15]. In the present study, using data of mRNA expression, DNA methylation, and clinical data retrieved from Oncomine, GEO, and the TCGA cohort, we identified a group of genes associated with cancer progression and prognosis in HCC. Methods Samples All patients who underwent curative hepatectomy for primary HCC at the First Affiliated Hospital of Guangxi Medical University between March 2015 and September 2015 were eligible for inclusion in this study. Total of 11 HCCs and the matched paracancerous tissues were collected during surgery and stored in a liquid nitrogen tank until use for mRNA isolation and protein extraction. The study was endorsed by the Ethics Committee of Guangxi Medical University and was performed according to the Declaration of Helsinki, 2013 edition. All patients received an explanation of the aims of the study and signed informed consent. mRNA isolation and quantitative real-time polymerase chain reaction (RT-qPCR) analysis Total RNA from 11 HCC and their matched paracancerous tissues was isolated using a miRNeasy Mini Kit (Qiagen, Hilden, Germany). RNA was quantified by spectrophotometry on a NanoDrop 2000 (Thermo Scientific, DE, USA). A total of 2 μg RNA was subjected to cDNA synthesis using the miScript II RT Kit (Qiagen, Hilden, Germany). RT-qPCR was performed with the QuantiFast SYBR Green PCR Kit (Qiagen, Hilden, Germany). Data were collected with the StepOnePlus Real-Time PCR System (ABI, CA, USA) according to the manufacturer’s instructions. The gene expression was compared in each HCC sample and the matched paracancerous tissue, and then the homogeneity of variance in all samples was analysed using the t-test. The RT-qPCR gene-specific primers were as follows: TBCE: forward primer, 5′-AGGCCAACAGATGTTCTCCAG-3′, reverse primer, 5′-CAGGGGGTTTCTTAGGCAGG-3′; INTS8: forward primer, 5′-AACTGAGAGTTCTACTGCTGGA-3′, reverse primer, 5′-GCTGCGCCCAAATCATAGC-3′; VIPR1: forward primer, 5′-TGCTGGGACACCATCAACTC-3′, reverse primer, 5′-TTGTCCGGAAAGAAGGCGAA-3′; CLEC4M: forward primer, 5′-TACTTCATGTCTAACTCCCAGCG-3′, reverse primer, 5′-GCTCCTCAGCAGTTTTGATTACG-3′; MARCO: forward primer, 5′-GGGGACACAGGACTTCAAGG-3′, reverse primer, 5′-CCCTGTTCTCCCTTCACACC-3′; DNASE1L3: forward primer, 5′-AGCCCTTTGTGGTCTGGTTC-3′, reverse primer, 5′-CGTCCGTGTAGACCTCAACC-3′; CRHBP: forward primer, 5′-AAATCCTCAGCAGGTTGCGA-3′, reverse primer, 5′-AAGGCGTCATCTTGGAAGGG-3′; FCN2: forward primer, 5′-CTGCAAGGACCTGCTAGACC-3′, reverse primer, 5′-TGTCATTCCCCAGCCAGAAC-3′; GAPDH (used as the control): forward primer, 5′-GAAGGTGAAGGTCGGAGT-3′, reverse primer, 5′-GAAGATGGTGATGGGATTT-3′. Protein extraction and western blotting Total protein was extracted from HCC and paracancerous tissues with RIPA lysis buffer (Solarbio, Beijing, China) and proteinconcentration was determined using an Enhanced BCA Protein Quantification Kit (KeyGEN BioTECH, Jiangsu, China). Then the samples were separated by Novex NuPAGE SDS-PAGE Gel System (Thermo Fisher Scientific, MA, USA) and were transferred to the PVDF membrane using the Bio-Rad Criterion System (Bio-Rad, CA, USA). Membranes were blocked with 8 % non-fat dry milk in PBS containing 0.1 % Tween-20 (0.1 % TBST, pH7.4) for 1 h. Membranes were incubated with antibodies specific for human INTS8 (rabbit polyclonal antibody, 1:750 dilutions; Proteintech, Hubei, China) and GAPDH (rabbit polyclonal antibody, 1:1,000 dilution; Boster, Hubei, China) overnight at 4 °C. After 3 washings with 0.1 % TBST for 5 min, horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (1:5,000 dilution; Bioss, Beijing, China) were applied, followed by washings with 0.1 % TBST for 5 min each at room temperature (RT). The bound immunocomplexes were detected using ECL+ reagent (GE Healthcare Bio-Sciences, NJ, USA) with a FluorChem M system (Proteinsimple, CA, USA). Gene expression profiles The genes significantly dysregulated in HCC were identified based on the 4 microarrays, Chen Liver microarray (104 HCCs vs. 76 liver tissues), Roessler Liver microarray (22 HCCs vs. 21 liver tissues), Roessler Liver 2 microarray (225 HCCs vs. 220 liver tissues) and Wurmbach Liver microarray (35 HCCs vs. 10 liver tissues), which are all deposited in Oncomine database (https://www.oncomine.org/resource/login.html) [9]. The 4 microarrays together covering total of 386 cases of HCCs and 327 cases of normal liver tissues. The rank for a gene is the median rank for that gene across each of the analyses. DNA methylation, mRNA expression, and clinical data of 379 HCC patients in a TCGA cohort were retrieved from cBioPortal for Cancer Genomics (http://cbioportal.org) [16, 17], but only 157 samples with matched gene expression data, prognosis data and most of the other clinical data were used to analyze the clinical importance of the target genes. mRNA expression data associated with HCC metastasis were retrieved from microarray GDS3091 [18] and GDS274 [19], which were deposited in the GEO profiles databases (http://www.ncbi.nlm.nih.gov/geoprofiles/) [8]. Bioinformatics analyses Enrichment of the biological process and cellular component of a group of genes was determined using the DAVID online tool (http://david.abcc.ncifcrf.gov/) [20, 21]. Protein/gene-protein/gene interaction analysis was performed using the GeneMANIA online tool (http://www.genemania.org/) [22, 23]. Function prediction based on text mining was performed using the Coremine Medical online database (http://www.coremine.com/medical/) [24]. Data analysis The data were analysed by SPSS 20.0 software. The mRNA expression of a gene is presented as the mean ± SD. Homogeneity of variance was analysed using the t-test. Expression values of a gene were dichotomised into high and low expression using the median as a cutoff for analysis of clinical importance in a TCGA cohort, as described in a previous study [25]. The probability of survival and its significance was calculated using the Kaplan-Meier method and log-rank test, respectively. A Cox proportional hazard model was performed for multivariate analysis of prognosis. The correlation between gene expression and clinicopathologic characteristics was evaluated by Pearson’s χ2 test (two-sided). The correlation between DNA methylation and gene expression was analysed using bivariate correlations. P values < 0.05 were considered to indicate statistically significant differences. Results Retrieval of significantly dysregulated genes in HCC Four independent microarrays deposited in the Oncomine database were selected to identify genes associated with cancer development and progression in HCC. These microarrays were Chen Liver Statistics covering 104 cases of HCC and 76 cases of liver tissue, Roessler Liver Statistics covering 22 cases of HCC and 21 cases of liver tissue, Roessler Liver 2 Statistics covering 225 cases of HCC and 220 cases of liver tissue, and Wurmbach Liver Statistics covering 35 cases of HCC and 10 cases of liver tissues. Based on analysis of these four independent microarrays, 40 genes that were significantly upregulated (P < 1.36E-10) and 40 genes that were significantly downregulated (P < 1.31E-10) in HCC were retrieved (Fig. 1). Analysis of the 80 genes by the DAVID online tool indicated that cell cycle was the top biological process, covering 17 genes, and microtubule cytoskeleton was the top cellular component, covering 14 genes (Additional file 1: Table S1).Fig. 1 The 80 genes that were significantly dysregulated in hepatocellular carcinomas according to four independent microarrays retrieved from the Oncomine database. a The top 40 genes that were significantly upregulated in four microarrays. b The top 40 genes that were significantly downregulated in four microarrays. The four microarrays cover a total of 386 cases of hepatocellular carcinomas and 327 cases of normal liver tissue: (1) Chen Liver Statistics, 104 cases of hepatocellular carcinoma and 76 cases of liver tissue; (2) Roessler Liver Statistics, 22 cases of hepatocellular carcinoma and 21 cases of liver tissue; (3) Roessler Liver 2 Statistics, 225 cases of hepatocellular carcinoma and 220 cases of liver tissue; (4) Wurmbach Liver Statistics, 35 cases of hepatocellular carcinoma and 10 cases of liver tissue. The rank for a gene is the median rank for that gene across each of the analyses. The P value given for a gene is for the median-ranked analysis. The genes labelled in red and in blue were significantly and consistently up- and downregulated in the four microarrays, respectively Among the 80 genes that were dysregulated in HCCs according to four independent microarrays covering a total of 386 cases of HCC and 327 cases of normal liver tissues, nine genes (CAP2, PTTG1, TOP2A, GMNN, GPC3, UBE2C, UBAP2L, TBCE, and INTS8) were consistently and stably upregulated and 18 genes (CXCL14, VIPR1, CLEC4M, MARCO, CLEC1B, NAT2, FCN2, EGR1, DNASE1L3, MT1F, CRHBP, LCAT, PAMR1, ACSM3, MT1G, MT1X, SRPX, and MT1H) were consistently and stably downregulated in HCC, by least 2-fold (Fig. 1; Table 1). Among the above 27 genes, seven genes—CAP2, GMNN, PTTG1, TBCE, TOP2A, UBE2C, and FCN2—encode proteins associated with cell cycle and microtubule cytoskeleton (Additional file 1: Table S1). Protein/gene-protein/gene interaction analysis was performed to further explain the interrelationships of these genes in HCC. As shown in Additional file 2: Figure S2, the 27 proteins/genes directly/indirectly interacted with each other via co-localisation, genetic interactions, shared common pathways, and protein domains, and, in particular, co-expression, and 10 of them—VIPR1, DNASE1L3, SRPX, MT1H, CXCL14, CLEC4M, CRHBP, GPC3, NAT2, and MARCO—interacted with at least 14 other genes, more than half of all the genes in the interaction network (Additional file 2: Figure S2). Moreover, these genes were also those that were dysregulated at least 4-fold in HCC (Table 1).Table 1 Genes that were stably and consistently dysregulated in 386 cases of hepatocellular carcinoma compared with 327 cases of normal liver tissues according to four independent microarrays retrieved from the Oncomine database, and their associations with hepatocellular carcinoma Gene Independent microarray data (Fold change) No. of articlesa Associations with hepatocellular carcinoma Direction of regulation Chen Liver Roessler Liver Roessler Liver 2 Wurmbach Liver TBCE b Up 2.125 2.403 2.822 2.419 - - INTS8 b Up 2.393 3.102 2.340 2.115 - - UBAP2L b Up 2.108 2.959 2.819 2.742 - - GMNN b Up 3.362 7.340 4.696 3.394 1 Potential oncogene [38] UBE2C Up 4.733 3.661 3.422 5.113 4 Cancer progression and poor prognosis [39] PTTG1 Up 4.688 4.741 5.773 10.622 9 Angiogenesis, progression, and poor prognosis [40, 41], therapeutic target [42] CAP2 Up 3.526 4.254 5.790 8.569 10 Multistage hepatocarcinogenesis [43], early detection [44] TOP2A Up 2.663 11.236 8.292 13.321 11 Early age onset, shorter patient survival and chemoresistance [45] GPC3 Up 16.826 26.693 28.236 76.162 199 Diagnosis [29], cell proliferation and invasion [28]; prediction of recurrence [30] VIPR1 b Down 9.979 5.310 7.202 4.855 - - CLEC4M b Down 28.107 9.276 4.361 36.431 - - MARCO b Down 11.333 6.107 3.984 20.154 - - DNASE1L3 b Down 8.386 12.378 7.653 10.303 - - PAMR1 b Down 2.726 2.381 2.473 2.917 - - ACSM3 b Down 2.902 6.135 4.836 11.262 - - CLEC1B b Down 6.600 6.605 4.748 36.770 1 Downregulated in a cohort of 65 pairs of human HCCs [46] MT1F b Down 14.107 18.140 15.749 9.680 1 Inhibition of cancer growth [47] CRHBP b Down 16.565 7.020 4.822 46.837 1 Downregulated in a cohort of 65 pairs of human HCCs [46] LCAT b Down 4.917 8.507 8.064 7.435 1 LCAT activity correlated with serum albumin and serum bilirubin level [48] MT1X b Down 10.812 11.558 8.227 6.903 1 HCC-related [49] SRPX b Down 4.929 5.104 5.879 7.202 1 Proliferation, migration and invasiveness [50] MT1H b Down 13.846 9.037 8.473 7.723 1 Potential tumour suppressor [51] FCN2 b Down 10.881 9.089 6.299 44.688 2 HBV- and HCV-related HCC [52], FCN2 haplotypes associate with HCC [53] CXCL14 b Down 12.903 9.667 10.940 13.977 4 Potential diagnostic marker [54]; rs2237062 polymorphism influences HBV-related HCC progression [52, 55]; potential tumour suppressor [56] MT1G b Down 13.065 11.134 11.160 11.187 4 Tumour suppressor gene [51, 57], biomarker [58] EGR1 Down 3.541 10.547 6.769 9.241 12 Critical for hepatocarcinogenesis [59] NAT2 Down 8.024 16.088 13.999 36.890 14 NAT2 polymorphism is risk factor for developing HCC [60], NAT2 activity is critical in smoking-related hepatocarcinogenesis [61] aNo. of articles was based on a search in the PubMed database bpoorly studied genes in HCC Measurement of gene expression at mRNA and protein level Among the 27 genes, the associations of seven with HCC are relatively well studied and described in published papers. However, the relationship of the remaining 20 genes with HCC was poorly understood, and these genes were selected for further analyses (Table 1). The expression of eight genes that were randomly selected from the 20 genes was measured by RT-qPCR in 11 tissues of HCC patients compared with matched paracancerous tissues. As shown in Fig. 2a, the expression of TBCE and INTS8 was increased, whereas that of VIPR1, CLEC4M, MARCO, DNASE1L3, CRHBP, and FCN2 was decreased in HCC tissues, although the changes in TBCE and VIPR1 expression were not statistically significant. Compared with the average expression in paracancerous tissues, the expression of INTS8 in HCC was upregulated with 2.06-fold and the expression of CLEC4M, MARCO, DNASE1L3, CRHBP, and FCN2 was downregulated with 3.83-, 5.70-, 5.63-, 3.87-, and 8.94-fold, respectively. All results of gene expression determined by RT-qPCR were completely consistent with their expression identified by the four independent microarrays (Fig. 1; Table 1). Furthermore, a significant increase at the protein level of INTS8 was observed in HCC tissues compared with corresponding paracancerous tissues (Fig. 2b), which was consistent with its expression at the mRNA level.Fig. 2 Measurement of gene expression at mRNA and protein level. a mRNA expression of genes in 11 tissues of HCC patients compared with matched paracancerous tissue. * P < 0.05; ** P < 0.01. b Protein expression of INTS8 in four tissues of HCC patients compared with expression in corresponding paracancerous tissues. The intensity of protein bands was measured by Image J software.. T, HCC tissue; P, paracancerous tissue Analysis of clinical importance The clinical importance in HCC of the 20 selected genes (Table 1) was evaluated on the basis of TCGA clinical data. A total of 379 HCC patient samples with clinical data in a cohort of TCGA were retrieved. Among these, 157 samples with mRNA expression values were selected for analysis of the relationship between genes and clinical characteristics. The expression values of a gene were categorised as high or low according to the median value in accordance with a previous study [25]. A total of 11 genes were associated with DFS and/or OS (Table 2); among those, low expression of ACSM3 and CXCL14 was associated with poor DFS, and low expression of CRHBP, DNASE1L3, FCN2, MT1X, and VIPR1 was associated with poor OS (Fig. 3, Table 2). Four genes were associated with both DFS and OS: high expression of INTS8 in HCC patients, and low expression of LCAT, MARCO, and PAMR1, was associated with poor DFS and OS (Fig. 4, Table 2). To elucidate whether any of the above genes was an independent factor for predicting patient survival, we performed multivariate analyses of tumour stage, tumour pathologic PT, tumour residual, tumour status, vital status, age, gender, and the 11 genes by a Cox proportional hazards model (Table 3). We found that stage (P = 0.050), tumour status (P = 0.001), DNASE1L3 expression (P = 0.042), and INTS8 expression (P = 0.023) were independent risk prognostic factors for OS in HCC patients, although no gene was found to be an independent prognostic factor for DFS (data not shown).Table 2 The associations of 11 genes with disease-free survival (DFS) and/or overall survival (OS) of patients with hepatocellular carcinoma in a TCGA cohort, analysed using Kaplan-Meier survival plots DFS (Median) OS (Median) 95 % Confidence Interval 95 % Confidence Interval Estimate Std. Error Lower Boundary Upper Boundary Estimate Std. Error Lower Boundary Upper Boundary ACSM3 H 24.800 7.587 9.930 39.670 L 14.400 2.391 9.714 19.086 Overall 19.300 3.428 12.581 26.019 CXCL14 H 29.300 10.355 9.004 49.596 L 16.400 2.222 12.045 20.755 Overall 19.300 3.428 12.581 26.019 INTS8 H 14.400 2.408 9.681 19.119 21.700 5.048 11.805 31.595 L 27.200 4.266 18.839 35.561 53.300 10.531 32.659 73.941 Overall 19.300 3.428 12.581 26.019 37.800 8.792 20.568 55.032 LCAT H 29.300 5.287 18.937 39.663 55.600 13.029 30.063 81.137 L 14.800 2.129 10.628 18.972 21.700 5.133 11.640 31.760 Overall 19.300 3.428 12.581 26.019 37.800 8.792 20.568 55.032 MARCO H 24.800 6.094 12.856 36.744 53.300 16.525 20.911 85.689 L 15.600 1.710 12.248 18.952 23.300 5.664 12.199 34.401 Overall 19.300 3.428 12.581 26.019 37.800 8.792 20.568 55.032 PAMR1 H 29.300 7.881 13.853 44.747 69.500 7.445 54.908 84.092 L 16.400 3.616 9.312 23.488 21.100 1.762 17.647 24.553 Overall 19.300 3.428 12.581 26.019 37.800 8.792 20.568 55.032 CRHBP H 55.600 13.080 29.964 81.236 L 27.500 7.523 12.754 42.246 Overall 37.800 8.792 20.568 55.032 DNASE1L3 H 55.600 6.310 43.232 67.968 L 23.300 5.103 13.298 33.302 Overall 37.800 8.792 20.568 55.032 FCN2 H 53.300 12.677 28.453 78.147 L 30.600 10.341 10.331 50.869 Overall 37.800 8.792 20.568 55.032 MT1X H 58.800 12.301 34.690 82.910 L 23.300 5.997 11.546 35.054 Overall 37.800 8.792 20.568 55.032 VIPR1 H 51.300 7.615 36.374 66.226 L 20.600 5.643 9.540 31.660 Overall 37.800 8.792 20.568 55.032 The gene expression and survival data of 157 HCC patients in a TCGA cohort were used for the analysis. Expression values of a gene were dichotomised into high and low expression using the median as a cutoff H high expression, L low expression Fig. 3 Association of seven genes (ACSM3, CXCL14, CRHBP, DNASE1L3, FCN2, MT1X, and VIPR1) with DFS or OS, analysed using Kaplan-Meier survival plots. The survival data of 157 HCC patients in a TCGA cohort were used for the analysis. Expression values of a gene were dichotomised into high expression (blue line) and low expression (green line) using the median as a cutoff Fig. 4 Association of INTS8, LCAT, MARCO, and PAMR1 with DFS and OS, analysed using Kaplan-Meier survival plots. The survival data of 157 HCC patients in a TCGA cohort were used for the analysis. Expression values of a gene were dichotomised into high expression (blue line) and low expression (green line) using the median as a cutoff Table 3 Multivariate analysis of prognosis of 157 HCC patients in a TCGA cohort using Cox proportional hazard model Factors B SE Wald df Sig. Exp(B) 95.0 % CI Lower Upper CRHBP -.267 1.272 .044 1 .834 .766 .063 9.268 DNASE1L3 -.969 .476 4.140 1 .042 .379 .149 .965 FCN2 .517 .896 .333 1 .564 1.676 .290 9.704 INTS8 .204 .090 5.175 1 .023 1.227 1.029 1.463 LCAT .030 .194 .024 1 .877 1.031 .704 1.509 MARCO -.070 .859 .007 1 .935 .932 .173 5.020 MT1X .788 1.051 .561 1 .454 2.198 .280 17.256 PAMR1 -.158 .236 .448 1 .503 .854 .538 1.355 VIPR1 .194 .287 .459 1 .498 1.215 .692 2.131 Stage (I/II–III) .901 .463 3.784 1 .050 2.462 .993 6.101 PT (1–2/3–4) -.223 .426 .273 1 .601 .800 .347 1.844 Residual (R0/R1–2) -.175 .586 .089 1 .765 .839 .266 2.649 Tumour status (free/with) 1.300 .404 10.359 1 .001 3.669 1.662 8.097 Vital status (dead/alive) −13.599 55.193 .061 1 .805 .000 .000 1.2e + 41 Age .000 .014 .001 1 .981 1.000 .974 1.028 Gender -.276 .312 .781 1 .377 .759 .412 1.399 PT AJCC Tumour Pathologic PT Six genes were associated with tumour pathologic PT and tumour stage (Table 4); among these, high expression of INTS8 and UBAP2L, and low expression of ACSM3, FCN2, LCAT, and MT1G, was significantly associated with metastatic tumour and late stage (P ≤ 0.05). In particular, UBAP2L was markedly and highly expressed in T2 tumours (72.5 % vs. 27.5 %) and LCAT was lowly expressed in T2 tumours (30.0 % vs. 70.0 %) and highly expressed in T1 tumours (72.6 % vs. 27.4 %). In addition, LCAT was highly expressed in stage I tumours (71.2 % vs. 28.8 %).Table 4 Associations of genes expression with AJCC tumour pathologic PT, tumour stage, age and gender in 157 patients with hepatocellular carcinoma Factors No. of patients ACSM3 FCN2 INTS8 LCAT MT1G UBAP2L High Low High Low High Low High Low High Low High Low PT 157 P = 0.037 P = 0.026 P = 0.046 P = 0.000 P = 0.016 P = 0.004  T1 62 (39.5 %) 39 (62.9 %) 23 (37.1 %) 39 (62.9 %) 23 (37.1 %) 23 (37.1 %) 39 (62.9 %) 45 (72.6 %) 17 (27.4 %) 39 (62.9 %) 23 (37.1 %) 22 (35.5 %) 40 (64.5 %)  T2 40 (25.5 %) 20 (50.0 %) 20 (50.0 %) 13 (32.5 %) 27 (67.5 %) 24 (60.0 %) 16 (40.0 %) 12 (30.0 %) 28 (70.0 %) 13 (32.5 %) 27 (67.5 %) 29 (72.5 %) 11 (27.5 %)  T3 46 (29.3 %) 16 (34.8 %) 30 (65.2 %) 22 (47.8 %) 24 52.2 %) 28 (60.9 %) 18 (39.1 %) 18 (39.1 %) 28 (60.9 %) 21 (45.7 %) 25 54.3 %) 23 (50.0 %) 23 (50.0 %)  T4 9 (5.7 %) 4 (44.4 %) 5 (55.6 %) 4 (44.4 %) 5 (55.6 %) 4 (44.4 %) 5 (55.6 %) 4 (44.4 %) 5 (55.6 %) 6 (66.7 %) 3 (33.3 %) 5 (55.6 %) 4 (44.4 %) Stage 143 P = 0.016 P = 0.032 P = 0.026 P = 0.000 P = 0.037 P = 0.009  I 59 (41.3 %) 36 (61.0 %) 23 (39.0 %) 36 (61.0 %) 23 (39.0 %) 23 (39.0 %) 36 (61.0 %) 42 (71.2 %) 17 (28.8 %) 37 (62.7 %) 22 (37.3 %) 22 (37.3 %) 37 (62.7 %)  II 36 (25.2 %) 19 (52.8 %) 17 (47.2 %) 12 (33.3 %) 24 (66.7 %) 22 (61.1 %) 14 (38.9 %) 12 (33.3 %) 24 (66.7 %) 13 (36.1 %) 23 (63.9 %) 25 (69.4 %) 11 (30.6 %)  III 48 (33.6 %) 16 (33.3 %) 32 66.7 %) 25 (52.1 %) 23 (47.9 %) 30 (62.5 %) 18 (37.5 %) 18 (37.5 %) 30 (62.5 %) 23 (47.9 %) 25 (52.1 %) 25 (52.1 %) 23 (47.9 %) CXCL14 GMNN INTS8 MT1F MT1G SPRX High Low High Low High Low High Low High Low High Low Agea 157 P = 0.031 P = 0.031 P = 0.005 P = 0.013 P = 0.031 P = 0.031   < 65 80 (51.0 %) 47 (58.8 %) 33 (41.3 %) 47 (58.8 %) 33 (41.3 %) 49 (61.2 %) 31 (38.8 %) 47 (58.8 %) 32 (40.0 %) 47 (58.8 %) 33 (41.3 %) 47 (58.8 %) 33 (41.3 %)   ≥ 65 77 (49.0 %) 32 (41.6 %) 45 (58.4 %) 32 (41.6 %) 45 (58.4 %) 30 (39.0 %) 47 (61.0 %) 32 (41.6 %) 46 (59.7 %) 32 (41.6 %) 45 (58.4 %) 32 (41.6 %) 45 (58.4 %) CLEC1B CRHBP FCN2 MT1G TBCE High Low High Low High Low High Low High Low Gender 157 P = 0.003 P = 0.019 P = 0.043 P = 0.003 P = 0.019  Female 62 (39.5 %) 22 (35.5 %) 40 (64.5 %) 24 (38.7 %) 38 (61.3 %) 25 (61.0 %) 37 (39.0 %) 22 (35.5 %) 40 (64.5 %) 24 (38.7 %) 38 (61.3 %)  Male 95 (60.5 %) 57 (60.0 %) 38 (40.0 %) 55 (57.9 %) 40 (42.1 %) 54 (56.8 %) 41 (43.2 %) 57 (60.0 %) 38 (40.0 %) 55 (57.9 %) 40 (42.1 %) aAge was dichotomised into < 65 and ≥ 65 using the median as a cutoff. PT, AJCC Tumour Pathologic PT. Expression values of a gene were dichotomised into high and low expression using the median as a cutoff. P value determined using Pearson’s χ2 test (2-sided) Ten genes were associated with age and gender. As shown in Table 4, we found that six genes—CXCL14, GMNN, INTS8, MT1F, MT1G, and SPRX—were expressed at low levels in HCC patients aged ≥ 65 years. Expression of five genes was related to the gender of HCC patients. Except for FCN2, which is lowly expressed in male HCC patients, the other four genes, CLEC1B, CRHBP, MT1G, and TBCE, were all lowly expressed in female HCC patients. In addition, PAMR1 and MT1X were closely related to the vital status; both showed low expression in 60.3 % (38/63) of HCC patients with dead status, compared with high expression in 57.4 % (54/94) of patients with alive status (P = 0.022). Potential roles of the genes in HCC progression The potential roles of the 20 genes in HCC were predicted on the basis of Coremine Medical mining. As shown in Fig. 5, the associations of the genes with diagnosis, prognosis, drug resistance, recurrence, metastasis, and invasiveness of HCC was comprehensively analysed. The results indicated that, with the exception of PAMR1, the other 19 genes were all associated with at least one factor contributing to cancer progression, and many of the genes, for example GMNN, CXCL14, MT1G, MT1X, SPRX, and VIPR1, were closely associated with almost all of the factors included in this analysis. Most of the genes were extensively associated with several factors. For example, 15 genes (including INTS8, LCAT, MARCO, and DANSE1L3) were associated with diagnosis, 14 genes (including INTS8, MARCO, CRHBP, and VIPR1) were associated with metastasis, and 13 genes (including LCAT, MARCO, FCN2, and CXCL14) were associated with prognosis.Fig. 5 Association of the genes with HCC characteristics was determined by text mining using Coremine Medical and probabilistic scoring (P < 0.05). HCC: hepatocellular carcinoma, DR: drug resistance Based on the gene expression in two independent GEO microarrays corresponding to HCC metastasis, the association of the genes CLEC4M, CRHBP, MARCO, MT1X, SRPX, UBAP2L, and VIPR1 with metastasis was further analysed; unfortunately, data for the other genes were unavailable. The expression of CRHBP, LCAT, and SPRX was significantly dysregulated in nine HCCs with venous metastasis compared with 11 HCC without (Fig. 6a). Genes VIPR1, LCAT, BAP2L, CLEC4M, CRHBP, and SRPX were significantly dysregulated in 32 HCCs with portal vein tumour thrombus metastasis and 33 HCCs with intrahepatic spread metastasis compared with 22 HCCs with no metastasis (Fig. 6b&c). In particular, LCAT was highly expressed in HCC patients with venous metastasis and patients with portal vein tumour thrombus metastasis, and SRPX was lowly expressed in HCC patients with venous metastasis and patients with intrahepatic spread metastasis (Fig. 6).Fig. 6 mRNA expression of the genes in HCC patients with and without metastasis according to microarray data retrieved from the GEO online database. a Microarray data GDS3091 [18] cover nine HCCs with venous metastasis and 11 without as controls. b, c Microarray data GDS274 [19] cover 32 HCCs with portal vein tumour thrombus metastasis, 33 with intrahepatic spread metastasis, and 22 HCCs with no metastasis as controls. *, P < 0.05; **, P < 0.01 Correlation of DNA methylation with mRNA expression of the target genes DNA methylation and mRNA expression data from 379 HCC patients in a TCGA cohort were retrieved and the correlations between them were analysed using bivariate correlations. Among the 20 genes that are poorly studied in HCC (Table 1), DNA methylation data of CLEC1B and SRPX were not available. DNA methylation was negatively correlated with the mRNA expression for eight genes, ACSM3, INTS8, LCAT, MT1X, CRHBP, MARCO, PAMR1, and VIPR1. In particular, high methylation of the first four genes was significantly correlated with lower mRNA expression (Fig. 7), indicating that the expression of these genes in HCC might be regulated by DNA methylation.Fig. 7 DNA methylation of four genes was significantly and negatively correlated with their mRNA expression. Data for gene expression and DNA methylation in 379 HCCs were retrieved from a TCGA cohort. The correlation between DNA methylation and gene expression was analysed using bivariate correlations Discussion Cancer is frequently considered to be a disease of the cell cycle because alterations in different families of cell cycle regulators cooperate in tumour development. Molecular analysis of human tumours has shown that cell cycle regulators are frequently mutated in human neoplasms, underscoring the importance of maintaining cell cycle commitment in the prevention of human cancer [26]. Abnormal expression of cell cycle controllers, particularly G1/S-phase transition, is often implicated in the pathogenesis of most human cancers, including HCC. For example, vaccinia-related kinase 1 promotes HCC by controlling the levels of cell cycle regulators associated with G1/S transition [27]. In this study, 80 genes that were significantly dysregulated in HCC were identified based on four independent microarrays covering a total of 386 cases of hepatocellular carcinoma and 327 cases of normal liver tissues (Fig. 1), and biological process annotation of these genes revealed that 17 of these genes were implicated in cell cycle functions (Additional file 1: Table S1). These results suggested that these genes might contribute to cancer progression and development in HCC at least in part through regulation of the cell cycle. Twenty-seven genes were further identified to be consistently dysregulated in all four microarrays by at least 2-fold (Table 1). The expression of eight of these genes (TBCE, INTS8, VIPR1, CLEC4M, MARCO, DNASE1L3, CRHBP, and FCN2) was confirmed in 11 tissues of HCC patients compared with matched paracancerous tissues by RT-qPCR (Fig. 2a). Seven of the 27 genes (UBE2C, PTTG1, CAP2, TOP2A, GPC3, EGR1, and NAT2) have been well studied in HCC (Table 1). For example, GPC3 plays critical roles in cell proliferation and invasion through the induction of apoptosis [28] and is a biomarker for diagnosis [29] and recurrence [30]. Protein/gene-protein/gene interaction analyses indicated that these 27 proteins/genes strongly interacted with each other, and 10 of them interacted with at least half of all the genes (Additional file 2: Figure S2). Moreover, six of these genes were related to the cell cycle in HCC (Additional file 1: Table S1). Together, these results indicate that the genes identified in this study might play crucial roles in HCC progression, probably functioning as a group. Biomarkers not only have prognostic implications, but are also helpful for measurement of treatment responses and surveillance for tumour recurrence and for guiding clinical decisions [31]. Thus, prognostic biomarkers for HCC patients are necessary and crucial, and there is an ongoing search for predictive biomarkers. In this study, a group of genes associated with DFS and OS (Table 2) were identified in 157 HCC patients. Among these genes, low expression of ACSM3 and CXCL14 was associated with poor DFS, low expression of CRHBP, DNASE1L3, FCN2, MT1X, and VIPR1 was associated with poor OS (Fig. 3, Table 2), high expression of INTS8 was associated with poor DFS and OS, and low expression of LCAT, MARCO, and PAMR1 was associated with poor DFS and OS (Fig. 4, Table 2). Furthermore, DNASE1L3 and INTS8 were identified as independent risk prognostic factors for OS (Table 3). There are few reports of the association of these genes with prognosis in HCC or in other cancers. Previous studies indicate that downregulation of CXCL14 is associated with prognosis in gastric cancer patients [32], MT1X may aid in the prognostic discrimination of oral squamous cell carcinoma cases [33], and MARCO expression is associated with breast cancer survival and risk of recurrence [34]. Twenty genes that have been less studied in HCC (Table 1) were further evaluated to predict their potential roles in HCC progression. Coremine medical mining suggested that most of those genes were associated with diagnosis, prognosis, drug resistance, recurrence, metastasis, and invasiveness. In particular, 13, 14, and 15 genes were potentially associated with prognosis, metastasis, and diagnosis in HCC, respectively (Fig. 5). The association of these genes with prognosis appears to have clinical importance, as 11 genes were shown to be associated with DFS or/and OS (Table 2, Fig. 3 & 4). The role of these genes in metastasis was further confirmed by gene expression analysis, which showed that five genes were significantly dysregulated in HCC with venous metastasis, portal vein tumour thrombus metastasis, or intrahepatic spread metastasis, compared with the appropriate controls. Specifically, LCAT was highly expressed in HCC patients with venous metastasis and patients with portal vein tumour thrombus metastasis, and SRPX was lowly expressed in HCC patients with venous metastasis and patients with intrahepatic spread metastasis (Fig. 6), suggesting that these two genes might be closely related to HCC metastasis. There are few studies on LCAT and SRPX in cancer metastasis, with only one reported that SRPX is upregulated in gastric cancer cells after depletion of TWIST, which promoted the epithelial-mesenchymal transition that occurs during the initial steps of tumour metastasis [35]. INTS8 encodes a subunit of the integrator complex that is involved in the cleavage of small nuclear RNAs, and its association with cancer is poorly understood. Limited studies indicate that INTS8 contains mutations in peripheral T cell lymphoma compared with non-malignant samples from 12 patients [36], and a combination of INTS8 with SULF1, ATP6V1C1, and GPR172A can be used to discriminate gastric carcinomas from adjacent noncancerous tissues [37]. In this study, we found that, potentially regulated by demethylation (Fig. 7), INTS8 was significantly and consistently upregulated at least 2.115-fold in HCC according to four independent microarrays (Fig. 1; Table 1) and that INTS8 mRNA was upregulated 2.06-fold on average in 11 tissues of HCC patients compared with corresponding paracancerous tissues, with a similar expression profile at the protein level (Fig. 2). Based on the clinical importance analysis of 157 HCC patients in a TCGA cohort, we found that high expression of INTS8 was associated with poor DFS and OS (Fig. 4, Table 2), and was an independent risk prognostic factor for OS (Table 3). Moreover, high expression of INTS8 was associated with metastatic tumours and late stage (Table 4), and with younger HCC patients (<65 years old) (Table 4). In addition, text mining indicated that INTS8 was closely related with metastasis, invasiveness, and diagnosis (Fig. 5). The above results strongly indicate that this gene is indeed upregulated in HCC, where it might play crucial roles in HCC cancer progression and development, and is a potential biomarker for diagnosis and, in particular, prognosis. Conclusion In summary, by means of data retrieved from six independent microarrays, RT-qPCR and western blotting detection in 11 pairs of tissues, clinical importance analyses in a cohort of 157 patients, and bioinformatics analyses including biological process annotation, protein interaction and text mining, we have identified a group of genes that are significantly dysregulated in HCC and might be associated with cancer progression, development, and, in particular, prognosis. These genes could be potential therapeutic targets for HCC treatment, and might be useful biomarkers for diagnosis and prognosis. Additional files Additional file 1: Table S1. Biological process and cellular component annotation of the 80 genes associated with HCC development and progression by DAVID online tool. (PDF 167 kb) Additional file 2: Figure S2. Protein/gene-protein/gene interaction network of the 27 genes that were stably and consistently dysregulated in 386 cases of hepatocellular carcinoma compared with 327 cases of normal liver tissue according to the four independent microarrays retrieved from the Oncomine database. (PDF 306 kb) Fuqiang Yin, Lipei Shu and Xia Liu are co-first authors. Acknowledgements None. Funding The present study was supported by the National Natural Science Foundation of China (grant nos. 81360448), the Natural Science Foundation of Guangxi (nos. 2014GXNSFAA118139), Key Laboratory of High-Incidence-Tumor Prevention and Treatment (Guangxi Medical University), Ministry of Education (nos. GK2015-ZZ03 and GK2014-ZZ03) and Guangxi Outstanding Teachers Training Project for Colleges. Availability of data and materials None. Authors’ contributions FY and XL performed most analysis. FY wrote the manuscript. LS provided the clinical samples. TL performed the RT-qPCR and western blotting. TP helped collect samples and revise manuscript. YN and SL performed mRNA and protein isolation. XZ and XQ designed the study and helped draft manuscript. All authors 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 None. Ethics approval and consent to participate The study was endorsed by the Ethics Committee of Guangxi Medical University and was performed according to the Declaration of Helsinki, 2013 edition. All patients received an explanation of the aims of the study and signed informed consent. We are free to use ovarian cancer data in TCGA by meeting its freedom-to-publish criteria: A marker paper has been published on that tumour type. ==== Refs References 1. Yang JD Roberts LR Hepatocellular carcinoma: A global view Nat Rev Gastroenterol Hepatol 2010 7 448 58 10.1038/nrgastro.2010.100 20628345 2. Miki D Ochi H Hayes CN Aikata H Chayama K Hepatocellular carcinoma: towards personalized medicine Cancer Sci 2012 103 846 50 10.1111/j.1349-7006.2012.02242.x 22339805 3. Byam J Renz J Millis JM Liver transplantation for hepatocellular carcinoma Hepatobiliary Surg Nutri 2013 2 22 30 4. Singhal A Jayaraman M Dhanasekaran DN Kohli V Molecular and serum markers in hepatocellular carcinoma: predictive tools for prognosis and recurrence Crit Rev Oncol Hematol 2012 82 116 40 10.1016/j.critrevonc.2011.05.005 21680198 5. Cancer Genome Atlas Research N Weinstein JN Collisson EA Mills GB Shaw KR Ozenberger BA The Cancer Genome Atlas Pan-Cancer analysis project Nat Genet 2013 45 1113 20 10.1038/ng.2764 24071849 6. International Cancer Genome C Hudson TJ Anderson W Artez A Barker AD Bell C International network of cancer genome projects Nature 2010 464 993 8 10.1038/nature08987 20393554 7. Rustici G Kolesnikov N Brandizi M Burdett T Dylag M Emam I ArrayExpress update--trends in database growth and links to data analysis tools Nucleic Acids Res 2013 41 D987 90 10.1093/nar/gks1174 23193272 8. 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 10.1093/nar/gks1193 23193258 9. Rhodes DR Yu J Shanker K Deshpande N Varambally R Ghosh D ONCOMINE: a cancer microarray database and integrated data-mining platform Neoplasia 2004 6 1 6 10.1016/S1476-5586(04)80047-2 15068665 10. Coordinators NR Database resources of the National Center for Biotechnology Information Nucleic Acids Res 2015 43 D6 17 10.1093/nar/gku1130 25398906 11. Rung J Brazma A Reuse of public genome-wide gene expression data Nat Rev Genet 2013 14 89 99 10.1038/nrg3394 23269463 12. Kannan L, Ramos M, Re A, El-Hachem N, Safikhani Z, Gendoo DM, et al. Public data and open source tools for multi-assay genomic investigation of disease. Briefings in bioinformatics. 2016;17:603–15. 13. Liu X Gao Y Zhao B Li X Lu Y Zhang J Discovery of microarray-identified genes associated with ovarian cancer progression Int J Oncol 2015 46 2467 78 25891226 14. Liu X Gao Y Lu Y Zhang J Li L Yin F Downregulation of NEK11 is associated with drug resistance in ovarian cancer Int J Oncol 2014 45 1266 74 24969318 15. Zeng X Yin F Liu X Xu J Xu Y Huang J Upregulation of E2F transcription factor 3 is associated with poor prognosis in hepatocellular carcinoma Oncol Rep 2014 31 1139 46 24402082 16. Gao J Aksoy BA Dogrusoz U Dresdner G Gross B Sumer SO Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal Sci Signal 2013 6 11 10.1126/scisignal.2004088 17. Cerami E Gao J Dogrusoz U Gross BE Sumer SO Aksoy BA The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data Can Dis 2012 2 401 4 10.1158/2159-8290.CD-12-0095 18. Budhu A Forgues M Ye QH Jia HL He P Zanetti KA Prediction of venous metastases, recurrence, and prognosis in hepatocellular carcinoma based on a unique immune response signature of the liver microenvironment Cancer Cell 2006 10 99 111 10.1016/j.ccr.2006.06.016 16904609 19. Ye QH Qin LX Forgues M He P Kim JW Peng AC Predicting hepatitis B virus-positive metastatic hepatocellular carcinomas using gene expression profiling and supervised machine learning Nat Med 2003 9 416 23 10.1038/nm843 12640447 20. da Huang W Sherman BT Lempicki RA Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources Nat Protoc 2009 4 44 57 10.1038/nprot.2008.211 19131956 21. da Huang W Sherman BT Lempicki RA Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists Nucleic Acids Res 2009 37 1 13 10.1093/nar/gkn923 19033363 22. Mostafavi S Ray D Warde-Farley D Grouios C Morris Q GeneMANIA: a real-time multiple association network integration algorithm for predicting gene function Genome Biol 2008 9 Suppl 1 S4 10.1186/gb-2008-9-s1-s4 18613948 23. Zuberi K Franz M Rodriguez H Montojo J Lopes CT Bader GD GeneMANIA prediction server 2013 update Nucleic Acids Res 2013 41 W115 22 10.1093/nar/gkt533 23794635 24. de Leeuw N Dijkhuizen T Hehir-Kwa JY Carter NP Feuk L Firth HV Diagnostic interpretation of array data using public databases and internet sources Hum Mutat 2012 33 930 40 10.1002/humu.22049 26285306 25. Hedditch EL, Gao B, Russell AJ, Lu Y, Emmanuel C, Beesley J, et al. ABCA transporter gene expression and poor outcome in epithelial ovarian cancer. J Natl’ Cancer Institute. 2014; 106 26. D’Andrilli G Kumar C Scambia G Giordano A Cell cycle genes in ovarian cancer: steps toward earlier diagnosis and novel therapies Clin Cancer Res 2004 10 8132 41 10.1158/1078-0432.CCR-04-0886 15623586 27. Lee N Kwon JH Kim YB Kim SH Park SJ Xu W Vaccinia-related kinase 1 promotes hepatocellular carcinoma by controlling the levels of cell cycle regulators associated with G1/S transition Oncotarget 2015 6 30130 48 26375549 28. Pan Z Chen C Long H Lei C Tang G Li L Overexpression of GPC3 inhibits hepatocellular carcinoma cell proliferation and invasion through induction of apoptosis Mol Med Rep 2013 7 969 74 23338845 29. Yu JP, Xu XG, Ma RJ, Qin SN, Wang CR, Wang XB, et al. Development of a Clinical Chemiluminescent Immunoassay for Serum GPC3 and Simultaneous Measurements Alone With AFP and CK19 in Diagnosis of Hepatocellular Carcinoma. J Clin Lab Anal. 2015;29:85–93. 30. Wang Y Shen Z Zhu Z Han R Huai M Clinical values of AFP, GPC3 mRNA in peripheral blood for prediction of hepatocellular carcinoma recurrence following OLT: AFP, GPC3 mRNA for prediction of HCC Hepat Mon 2011 11 195 9 22087143 31. Wong KF Xu Z Chen J Lee NP Luk JM Circulating markers for prognosis of hepatocellular carcinoma Expert Opinion Med Diagnos 2013 7 319 29 10.1517/17530059.2013.795146 32. Hu C Lin F Zhu G Xue X Ding Y Zhao Z Abnormal hypermethylation of promoter region downregulates chemokine CXC ligand 14 expression in gastric cancer Int J Oncol 2013 43 1487 94 23982764 33. Brazao-Silva MT Rodrigues MF Eisenberg AL Dias FL de Castro LM Nunes FD Faria PR Cardoso SV Loyola AM de Sousa SCOM Metallothionein gene expression is altered in oral cancer and may predict metastasis and patient outcomes Histopathology 2015 67 358 67 10.1111/his.12660 25640883 34. Bergamaschi A Tagliabue E Sorlie T Naume B Triulzi T Orlandi R Russnes HG Nesland JM Tammi R Auvinen P Kosma V-M Ménard S Børresen-Dale A-L Extracellular matrix signature identifies breast cancer subgroups with different clinical outcome J Pathol 2008 214 357 67 10.1002/path.2278 18044827 35. Feng MY Wang K Shi QT Yu XW Geng JS Gene expression profiling in TWIST-depleted gastric cancer cells Anat Rec 2009 292 262 70 10.1002/ar.20802 36. Simpson HM Khan RZ Song C Sharma D Sadashivaiah K Furusawa A Liu X Nagaraj S Sengamalay N Sadzewicz L Luke J Concurrent Mutations in ATM and Genes Associated with Common gamma Chain Signaling in Peripheral T Cell Lymphoma PLoS One 2015 10 e0141906 10.1371/journal.pone.0141906 26536348 37. Cheng L Zhang Q Yang S Yang Y Zhang W Gao H Deng X Zhang Q A 4-gene panel as a marker at chromosome 8q in Asian gastric cancer patients Genomics 2013 102 323 30 10.1016/j.ygeno.2013.05.004 23722107 38. Kim HE Kim DG Lee KJ Son JG Song MY Park YM Frequent amplification of CENPF, GMNN and CDK13 genes in hepatocellular carcinomas PLoS One 2012 7 e43223 10.1371/journal.pone.0043223 22912832 39. Ieta K Ojima E Tanaka F Nakamura Y Haraguchi N Mimori K Identification of overexpressed genes in hepatocellular carcinoma, with special reference to ubiquitin-conjugating enzyme E2C gene expression Int J Cancer 2007 121 33 8 10.1002/ijc.22605 17354233 40. Fujii T Nomoto S Koshikawa K Yatabe Y Teshigawara O Mori T Overexpression of pituitary tumor transforming gene 1 in HCC is associated with angiogenesis and poor prognosis Hepatology 2006 43 1267 75 10.1002/hep.21181 16628605 41. Su MC Hsu HC Liu YJ Jeng YM Overexpression of pituitary tumor-transforming gene-1 in hepatocellular carcinoma Hepatogastroenterology 2006 53 262 5 16608036 42. Liang M Chen X Liu W Li S Li C Jiang L Role of the pituitary tumor transforming gene 1 in the progression of hepatocellular carcinoma Cancer Biol Ther 2011 11 337 45 10.4161/cbt.11.3.14102 21099350 43. Shibata R Mori T Du W Chuma M Gotoh M Shimazu M Overexpression of cyclase-associated protein 2 in multistage hepatocarcinogenesis Clin Cancer Res 2006 12 5363 8 10.1158/1078-0432.CCR-05-2245 17000669 44. Sakamoto M Mori T Masugi Y Effendi K Rie I Du W Candidate molecular markers for histological diagnosis of early hepatocellular carcinoma Intervirology 2008 51 Suppl 1 42 5 10.1159/000122603 18544947 45. Wong N Yeo W Wong WL Wong NL Chan KY Mo FK TOP2A overexpression in hepatocellular carcinoma correlates with early age onset, shorter patients survival and chemoresistance Int J Cancer 2009 124 644 52 10.1002/ijc.23968 19003983 46. Ho DW Kai AK Ng IO TCGA whole-transcriptome sequencing data reveals significantly dysregulated genes and signaling pathways in hepatocellular carcinoma Frontiers Med 2015 9 322 30 10.1007/s11684-015-0408-9 47. Lu DD Chen YC Zhang XR Cao XR Jiang HY Yao L The relationship between metallothionein-1F (MT1F) gene and hepatocellular carcinoma Yale J Biol Med 2003 76 55 62 15369632 48. Tahara D Nakanishi T Akazawa S Yamaguchi Y Yamamoto H Akashi M Lecithin-cholesterol acyltransferase and lipid transfer protein activities in liver disease Metabolism 1993 42 19 23 10.1016/0026-0495(93)90166-L 8446043 49. Gui T Dong X Li R Li Y Wang Z Identification of hepatocellular carcinoma-related genes with a machine learning and network analysis J Computational Biol 2015 22 63 71 10.1089/cmb.2014.0122 50. Lin ZY Chuang WL Genes responsible for the characteristics of primary cultured invasive phenotype hepatocellular carcinoma cells Biomed Pharmacother 2012 66 454 8 10.1016/j.biopha.2012.04.001 22681909 51. Udali S Guarini P Ruzzenente A Ferrarini A Guglielmi A Lotto V DNA methylation and gene expression profiles show novel regulatory pathways in hepatocellular carcinoma Clin Epigen 2015 7 43 10.1186/s13148-015-0077-1 52. Zhou X Zhu HQ Lu J Regulation of gene expression in HBV- and HCV-related hepatocellular carcinoma: integrated GWRS and GWGS analyses Int J Clin Experimental Med 2014 7 4038 50 53. Hoang TV Toan NL le Song H Ouf EA Bock CT Kremsner PG Ficolin-2 levels and FCN2 haplotypes influence hepatitis B infection outcome in Vietnamese patients PLoS One 2011 6 e28113 10.1371/journal.pone.0028113 22140517 54. Sun H Chua MS Yang D Tsalenko A Peter BJ So S Antibody Arrays Identify Potential Diagnostic Markers of Hepatocellular Carcinoma Biomark Insights 2008 3 1 18 19578489 55. Gu X Wang H Wang A Dou T Qi P Ji Q An intronic polymorphism rs2237062 in the CXCL14 gene influences HBV-related HCC progression in Chinese population Mol Biol Rep 2012 39 797 803 10.1007/s11033-011-0801-7 21556757 56. Wang W Huang P Zhang L Wei J Xie Q Sun Q Antitumor efficacy of C-X-C motif chemokine ligand 14 in hepatocellular carcinoma in vitro and in vivo Cancer Sci 2013 104 1523 31 10.1111/cas.12279 24033560 57. Kanda M Nomoto S Okamura Y Nishikawa Y Sugimoto H Kanazumi N Detection of metallothionein 1G as a methylated tumor suppressor gene in human hepatocellular carcinoma using a novel method of double combination array analysis Int J Oncol 2009 35 477 83 10.3892/ijo_00000448 19639168 58. Ji XF Fan YC Gao S Yang Y Zhang JJ Wang K MT1M and MT1G promoter methylation as biomarkers for hepatocellular carcinoma World JGastroenterol 2014 20 4723 9 10.3748/wjg.v20.i16.4723 24782625 59. Lu D Han C Wu T Microsomal prostaglandin E synthase-1 promotes hepatocarcinogenesis through activation of a novel EGR1/beta-catenin signaling axis Oncogene 2012 31 842 57 10.1038/onc.2011.287 21743491 60. Agundez JA Olivera M Ladero JM Rodriguez-Lescure A Ledesma MC Diaz-Rubio M Increased risk for hepatocellular carcinoma in NAT2-slow acetylators and CYP2D6-rapid metabolizers Pharmacogenetics 1996 6 501 12 10.1097/00008571-199612000-00003 9014199 61. Farker K Schotte U Scheele J Hoffmann A Impact of N-acetyltransferase polymorphism (NAT2) in hepatocellular carcinoma (HCC)--an investigation in a department of surgical medicine Exp Toxicol Pathol 2003 54 387 91 10.1078/0940-2993-00275 12877350

PMC005xxxxxx/PMC5002171.txt

==== Front Lipids Health DisLipids Health DisLipids in Health and Disease1476-511XBioMed Central London 30410.1186/s12944-016-0304-6ResearchEvaluating oxidative stress, serological- and haematological status of dogs suffering from osteoarthritis, after supplementing their diet with fish or corn oil Barrouin-Melo Stella Maria barrouin@ufba.br 12Anturaniemi Johanna johanna.roine@helsinki.fi 1Sankari Satu satu.sankari@helsinki.fi 1Griinari Mikko mikko.griinari@clanet.fi 3Atroshi Faik faik.atroshi@helsinki.fi 45Ounjaijean Sakaewan sakaewan@gmail.com 5Hielm-Björkman Anna Katrina +358-(0)44-3270462anna.hielm-bjorkman@helsinki.fi 11 Department of Equine and Small Animal Medicine, Faculty of Veterinary Medicine, University of Helsinki, P.O. Box 57, 00014 Helsinki, Finland 2 Department of Anatomy, Pathology and Clinics, School of Veterinary Medicine and Zootechny, Federal University of Bahia, Av. Adhemar de Barros, 500, CEP: 40170-110 Salvador, Bahia Brazil 3 Clanet Oy, Lotankatu 1, 02680 Espoo, Finland 4 Department of Pharmacology and Toxicology, Faculty of Veterinary Medicine, University of Helsinki, P.O. Box 57, 00014 Helsinki, Finland 5 Rinnekoti Research Centre, Nousumäki 2, 02980 Espoo, Finland 26 8 2016 26 8 2016 2016 15 1 13910 5 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 Oxidative stress plays an important role in the pathogenesis of disease, and the antioxidant physiological effect of omega-3 from fish oil may lead to improvement of canine spontaneous osteoarthritis (OA). Methods In this prospective randomized, controlled, double-blinded study, we assessed haematological and biochemical parameters in dogs with OA following supplementation with either a concentrated omega-3 deep sea fish oil product or corn oil. Blood samples from 77 client-owned dogs diagnosed as having OA were taken before (baseline) and 16 weeks after having orally ingested 0.2 ml/Kg bodyweight/day of deep sea fish oil or corn oil. Circulating malondialdehyde (MDA), glutathione (GSH), non-transferrin bound iron (NTBI), free carnitine (Free-Car), 8-hydroxy-2-deoxyguanosine (8-OH-dG), and serum fatty acids, haemograms and serum biochemistry were evaluated. Differences within and between groups from baseline to end, were analysed using repeated samples T-test or Wilcoxon rank test and independent samples T-test or a Mann-Whitney test. Results Supplementation with fish oil resulted in a significant reduction from day 0 to day 112 in MDA (from 3.41 ± 1.34 to 2.43 ± 0.92 μmol/L; P < 0.001) and an elevation in Free-Car (from 18.18 ± 9.78 to 21.19 ± 9.58 μmol/L; P = 0.004) concentrations, whereas dogs receiving corn oil presented a reduction in MDA (from 3.41 ± 1.34 to 2.41 ± 1.01 μmol/L; P = 0.001) and NTBI (from −1.25 ± 2.17 to −2.31 ± 1.64 μmol/L; P = 0.002). Both groups showed increased (albeit not significantly) GSH and 8-OH-dG blood values. Dogs supplemented with fish oil had a significant reduction in the proportions of monocytes (from 3.84 ± 2.50 to 1.77 ± 1.92 %; P = 0.030) and basophils (from 1.47 ± 1.22 to 0.62 ± 0.62 %; P = 0.012), whereas a significant reduction in platelets counts (from 316.13 ± 93.83 to 288.41 ± 101.68 × 109/L; P = 0.029), and an elevation in glucose (from 5.18 ± 0.37 to 5.32 ± 0.47 mmol/L; P = 0.041) and cholesterol (from 7.13 ± 1.62 to 7.73 ± 2.03 mmol/L; P = 0.011) measurements were observed in dogs receiving corn oil. Conclusions In canine OA, supplementation with deep sea fish oil improved diverse markers of oxidative status in the dogs studied. As corn oil also contributed to the reduction in certain oxidative markers, albeit to a lesser degree, there was no clear difference between the two oil groups. No clinical, haematological or biochemical evidence of side effects emerged related to supplementation of either oil. Although a shift in blood fatty acid values was apparent due to the type of nutraceutical product given to the dogs, corn oil seems not to be a good placebo. Keywords OAOsteoarthritisDogNatural modelMDAGSHNTBI8-OH-dGOmega-3Omega-6Fatty acidCAPESProc. no. BEX 2686/13-9Barrouin-Melo Stella Maria Helvi KnuuttilaNutraceuticoilsMila Laboratoriesissue-copyright-statement© The Author(s) 2016 ==== Body Background Alongside the importance of knowledge of canine diseases in veterinary medicine, pet dogs have increasingly been used as translational models for human diseases such as osteoarthritis (OA). The three joints predominantly affected in canine OA are the hips, elbows and knees [1]. OA commonly occurs as a consequence of hip and elbow dysplasia [2]. In these cases, even though the genetic background might be crucial for the development of joint mechanical instability, years of genetic selection by dog breeders has not reduced the incidence of the disease [3]. Factors such as nutritional imbalance [4, 5], chronic inflammation [6, 7], ageing [8, 9] and obesity [10, 11] are linked to the development of OA in dogs and in humans. These factors are also associated with oxidative stress [12]. Structural, cellular and molecular alterations due to proteoglycan breakdown and inflammation triggered by e.g. IL-1 are present in the evolution of OA [13]. Osteoarthritis is a condition that causes pain, inflammation and stiffness in many joints. Non-steroidal anti-inflammatory drugs (NSAIDs) and disease-modifying anti-rheumatic drugs (DMARDs) are available today for the treatment of inflammatory disorders; however, these drugs have side effects. NSAIDs react by blocking the activity of cyclooxygenases (COXs) during inflammation [14]. COX is a bifunctional enzyme exhibiting both COX and peroxidase activities [15]. While the COX component converts arachidonic acid (AA) to a hydroperoxy-endoperoxide, the peroxidase component reduces the endoperoxide to the corresponding alcohol, the precursor of prostanoids, such as thromboxanes and prostaglandins [16]. Many promising new treatment approaches for OA are available. The search for more natural anti-inflammatory agents that can selectively block the activity of COX-2 during inflammation is ongoing, with enhanced therapeutic effect and little or no side effects even with prolonged usage high on the wish list. Scientists have provided new evidence that fish oil supplementation decreases the formation of pro-inflammatory prostanoids, which, when produced in excess, increase inflammation in various tissues and organs [17, 18]. Fish oil is rich in docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), n-3 fatty acids that act as competitive substrates for the enzymes of AA metabolism and its products [19]. Reactive oxygen species (ROS), associated with pro-inflammatory cytokines and prostaglandins, have been shown to play a deleterious role in the course of joint diseases, leading to cartilage damage [20, 21] and progressive chronic inflammation [22, 23]. Comparable changes in the synovia immune cellularity have been described in dogs as those occurring in humans [24, 25]. An increased level of oxidation in the synovial fluid of osteoarthritic joints has been observed [26, 27], and increased activity of antioxidant enzymes has been associated with decreased viscosity of the synovial fluid [28]. Also, markers of oxidative stress and levels of antioxidant enzymes change in the blood of human patients and dogs with OA. These include serum catalase (CAT), superoxide dismutase (SOD), glutathione (GSH) and malondialdehyde (MDA) [29, 30]. Lipid metabolism in tissues affected by OA includes peroxidation reactions that link inflammation, oxidative stress and cartilage/bone tissue damage [28]. Lipid peroxidation may be involved in the clinical consequences of pain and dysfunction in the joint [31, 32]. Thus, dietary lipid composition is an important issue. Fish oil supplements, rich in omega-3 polyunsaturated fatty acids (PUFAs), e.g. EPA, have been claimed to be beneficial in the treatment of diseases such as rheumatoid arthritis [33]. After fish oil supplementation, the omega-3 s EPA and DHA increased in both plasma and neutrophil membranes of human patients exhibiting clinical improvement of OA [34]. Supplementation of n-3 PUFA in cell culture studies as well as in animal in vivo studies showed that cell proliferation was inhibited by production of lipid peroxides [35, 36]. Omega-3-related inhibition of cell proliferation via lipid peroxidation is helpful in conditions such as cancer, although lipid peroxidation and formation of ROS are thought to produce tissue damage [28]. Under in vitro conditions using endothelial cells, omega-3 supplementation resulted in an indirect antioxidant effect, characterized by reduced formation of lipid peroxidation products associated with direct superoxide scavenging, thus reducing inflammation, as compared with supplementing with other long-chain PUFAs [37]. Evidence suggests that omega-3 supplementation benefits human patients [38–40] and dogs [41–44] suffering from OA. Earlier in vitro studies have reported some beneficial effects of omega-3 fatty acids on cartilage cell inflammation and metabolism [45, 46] that are consistent with clinical observations. In dogs, the quality and mechanisms underlying nutraceutical therapy with omega-3 oils are just beginning to be investigated. This paper reports our findings on the circulating levels of fatty acids, oxidative stress markers and antioxidant molecules as well as blood counts and serum biochemistry as outcome measures in a controlled double-blind clinical trial designed to compare the effects of omega-3 fatty acids and corn oil in dogs suffering from OA. Methods Data on the clinical parameters have been published earlier [43], with a detailed description of the study protocol, the inclusion criteria for dogs, the diet and the supplements and the blood sampling protocol. The materials and methods section here contains only a brief description of the dogs and study design. Dogs and ethics Briefly, 77 dogs met the inclusion criteria of age of at least one year, a body weight exceeding 18 kg, predetermined clinical signs of OA, a Helsinki chronic pain index (HCPI) [47] of over 6 and a radiographic diagnosis of hip, knee or elbow OA. Exclusion criteria were neurological deficits, articular infection, recent trauma, pregnancy, lactation or the owner would not comply with changing the dogs’ diet or with giving the supplement on a daily basis. The dogs lived with their owners during the entire study period, and generally had low physical activity due to their OA condition. The owners of the dogs were instructed not to give non-steroidal anti-inflammatory drugs (NSAIDs) or other analgesics for the two weeks preceding the baseline visit if possible, and not to give medication containing Na-pentosan polysulphate for 30 days. They were also required to sign an informed consent form acknowledging that they could leave the study at any time, without giving a reason. The study protocol was approved by the Committee of Ethics of the University of Helsinki and was conducted at the Small Animal Hospital of the same University. Study design and test products The study was run as a double-blind, randomized clinical trial with a treatment group (an omega-3 supplement comprising a concentrated oil product from deep sea fish) and a negative control group (corn oil), using the CONSORT guidelines [48]. All dogs were evaluated by radiographs, submitted to blood sampling, and assigned to groups using a stratified computer-generated four-block random list [43] using the HCPI pain severity level and the diet as strata. The groups were statistically equal in total number of dogs, gender (also whether castrated, sterilized or intact), mean body weight, mean age and mean disease (OA) duration in months at baseline [43]. All eligible dogs were started on a diet of commercial dry food, which was either a basic wheat/beef (Royal Canin® Croc) or a rice/lamb sensitive formula containing no wheat, soy, beef, preservatives or artificial colorants (Jahti & Vahti®). Both foods had similar nutritional composition with no fish oil-derived ω-3 fatty acids and a ω-6/ω-3 ratio of 11:1. Neither food contained ingredients with a known positive effect on OA. At baseline, the dogs were evaluated clinically, as described elsewhere [43], and had their blood samples collected. The dogs started the supplement trial by receiving a pharmaceutically cleaned dense deep sea fish oil with added vitamin E (Doils® Joints, Nutraceuticoils, Belgium) or corn oil (control) daily at the dose of 0.2 ml/Kg of body weight (BW) per os. Fish oil used in this study was concentrated by molecular distillation with a total omega-3 content of 63.6 % (Table 1). The fish oil contained 450 mg of EPA, 100 mg of DHA and 27 mg of eicosatetraenoic acid (ETA) per ml, with a total ω-3 content of 63.6 % of the total fatty acid content (Table 1). The same manufacturer supplied the corn oil with fish smell in identical containers to be used as a control. The corn oil contained only 1 % of ω-3, mainly α-linolenic acid (ALA), and no EPA, DHA or docosapentaenoic acid (DPA) [43]. To avoid oxidation of the fatty acids all dog owners were instructed to keep the oil supplement with the cap tightly shut in a dark refrigerator at 4-8 °C. At the end of the trial, at week 16, the dogs were again clinically evaluated [43] and blood samples were taken after an overnight fast. All outcome evaluators (veterinarians and owners) and technical assistants were blinded (Fig. 1).Table 1 Composition of fish oil and corn oil products used as nutraceutical supplements for the study dogs suffering from osteoarthritis Component Fish oil g/100 g Corn oil g/100 g Analyzed fatty acids  Palmitic acid (PA), C16:0a 0.6 10.2  Stearic acid (SA), C18:0a 4.6 1.8  Oleic acid (OA), C18:1 ω-9b 9.7 28.6  Linoleic acid (LA), C18:2 ω-6c 1.3 51.8  Gamma-Linolenic acid (GLA), C18:3 ω-6c 0.7 0.1  Eicosadienoic acid (EDA), C20:2 ω-6c 0.3 0  Dihomo-gamma-linolenic acid (DHGLA), C20:3 ω-6c 0.4 0  Arachidonic acid (AA), C20:4 ω-6c 2.1 0  Palmitoleic acid (PO), C16:1 ω-7b 0.2 0.1  α-linolenic acid (ALA), C18:3 ω-3c 0.6 1.0  Stearidonic acid (SDA), C18:4 ω-3c 2.4 0  Eicosatetraenoic acid (ETA), C20:4 ω-3c 1.8 0  Eicosapentaenoic acid (EPA), C20:5 ω-3c 43.9 0  Docosapentaenoic acid (DPA), C22:5 ω-3c 2.2 0  Docosahexaenoic acid (DHA), C22:6 ω-3c 11.4 0  Other fatty acids combined 17.8 6.4 %  ω-3 fatty acids 63.6 1.0  ω-6 fatty acids 4.6 51.8  Ratio ω-6 / ω-3 0.07 52.2  Total SAFAa 6.4 12.8  Total MUFAb 16.4 29.5  Total PUFAc 69.5 53.2 a SAFA saturated fatty acid, b MUFA monounsaturated fatty acid, c PUFA polyunsaturated fatty acid; Source: Hielm-Björman et al. (2012) [43], with a few additions Fig. 1 Flow chart showing the design and inclusion/exclusion criteria of the prospective study of fish oil as a nutraceutical intervention in dogs suffering from osteoarthritis (OA). Corn oil was compared with fish oil as placebo. *Information on the dogs’ clinical follow up and pain outcome are described by Hielm-Björkman et al. [43] Fatty acid analysis The fatty acid composition of the oil products used was analysed by an international food testing laboratory using a modified AOCS Ce 1c-89 method (Eurofins, Raisio, Finland). The fatty acid profile of the serum total lipids was analysed according to Seppänen-Laakso et al. [49]. Briefly, the serum lipids were first extracted with chloroform:dimethyl ether (2:1), then centrifuged at 3500 rpm. The lower chloroform layer was separated and evaporated overnight. Petroleum ether (0.5 ml) was added. The methyl esters were formed by adding 1 ml of 0.5 M Na-methylat and keeping the substance in a 50 °C water bath for 10 min. The samples were neutralized with 1 ml of NaHSO4, and the fatty acid methyl esters were extracted by adding 0.5 ml of petroleum ether. The samples were shaken for 10 min and centrifuged at 3500 rpm. The petroleum ether layer was analysed by DANI GC 1000 gas chromatography with NB-351 0.32 mm × 25 m capillary column by HNU-Nordion Ltd. (Helsinki, Finland). The column oven was programmed from 120 °C to 236 °C at 5.9 °C/min. Identification was based on retention times. Dietary fatty acid content was used to evaluate the adherence of each owner to providing the dogs’ respective supplement of fish oil or corn oil. Blood counts and biochemical analysis Venous blood samples from overnight fasting dogs were taken at all three visits from the cephalic vein prior to any clinical tests or sedation. Whole-blood EDTA samples were immediately used for blood counts. To obtain plasma, part of the blood samples were taken in polypropylene tubes containing EDTA or Li-heparin and centrifuged at 1000 × g for 15 min. For serum, the blood was allowed to clot at room temperature for 30 min and then centrifuged at 1000 × g for 15 min. All serum and plasma samples were stored at −80 °C until analysed. Complete blood counts were performed with an automated haematology multiparameter analyser adjusted for animal-cell counting (Cell-Dyn 3700 System, ABBOTT Diagnostics Division, ABBOTT Park, IL, USA). Blood smears were stained with May-Grünwald-Giemsa, and manual leucocyte differentials were determined by counting 200 cells. The activity of serum alanine aminotransferase (ALT, EC 2.6.1.2 was measured according to the recommendations of the International Federation of Clinical Chemistry [50] and serum alkaline phosphatase (ALP, EC 3.1.3.1) activity according to the recommendations of the Scandinavian Society for Clinical Chemistry and Clinical Physiology [51]. Spectrophotometric methods were used for the determination of serum albumin, cholesterol, creatinine, glucose, total protein, triglycerides and urea. Analyses were performed by using a clinical chemistry analyser (Konelab 30i, Thermo Fisher Scientific, Vantaa, Finland). Lipid peroxidation was determined by HPLC-based TBARS modified assay [52] and reported as MDA-equivalent. In brief, the malondialdehyde-thiobarbituric acid complex formed under high temperature (100 °C) and acidic conditions were injected onto the Spherisorb ODS2 (C18 column); 5 μm, 250 × 3.2 mm with guard column (C8) and eluted with 65 %, 50 mM potassium phosphate buffer pH 7.0 and 35 % methanol at a flow rate of 1.0 ml/min and detected at a wavelength of 532 nm. Blood reduced GSH was determined using Ellman’s reagent [53] and served/applied as a marker of the antioxidant capacity of the organism. Briefly, an aliquot (0.2 ml) of whole blood was combined with 10 % sulphosalicylic acid. After centrifugation at 12,000 rpm for 2 min in a Eppendorf centrifuge at 4 °C, the supernatant was analysed spectrophotometrically at 412 nm, with 5,5′-dithiobis (2-nitro benzoic acid) in 0.1 M phosphate buffer, pH 8.0, for non-protein thiols. Concentrations of 8-hydroxy-2-deoxy Guanosine (8-OH-dG) were measured in serum samples by an 8-hydroxy-2-deoxy Guanosine EIA kit (Cayman Chemical, Ann Arbor, MI, USA) to evaluate oxidative DNA damage and used as an indicator of oxidative stress [54]. Serum free iron (NTBI) concentration was measured by NTA-chelation/HPLC-based assay principally described elsewhere [54, 55] and used as a surrogate marker of iron metabolism and oxidation in the body. Free carnitine (Free-Car) was measured using a Konelab 20XTi analyser (ThermoFisher Scientific, Vantaa, Finland) and used as a marker of antioxidant capacity of the body. Serum samples were pipetted to tubes (Centrifree Amicon, Gloucestershire, UK), which were centrifuged at 2000 g for one hour. Clear supernatants were used in assays of free carnitine. For determination of total carnitine, this supernatant was alkalized with 1 M NaOH. The incubation time at 60 °C was one hour. After cooling the sample was acidified with 2 M HCl. This sample was used for analysis of total carnitine. Carnitine reacted with acetyl-CoA to form acetylcarnitine and CoA. Carnitine acetyltransferase was the catalyser in this reaction. Free CoA reacted then with DTNB (5, 5-dithiobis (2-nitro benzoicacid)). Forming a complex, the thiophenolate ion was then measured at 405 nm. The detection limit was 2.0 μmol/L. Statistical analysis All 77 intended-to-treat dogs were analysed. To calculate differences within groups from baseline to end, a repeated samples T-test and Wilcoxon rank test were used if the variables were normally distributed or not, respectively. To test differences between groups at the different evaluation times, an independent samples T-test or a Mann-Whitney test was used, using the same criteria of normality. Results Blood concentrations of fatty acids Dogs were supplemented for 16 weeks with fish oil (N = at start 39/at end 35) or corn oil (N = 38/36). No significant differences between groups in terms of dogs per group, gender, age, body weight or disease severity/body distribution were observed [43]. As expected, serum concentrations of different fatty acids reflected the fatty acid composition in the nutraceutical products. All in all 69.5 % of the total fat content of the fish oil comprised PUFAs, whereof 63.6 % omega-3 and 4.6 % omega-6 fatty acids. The fat content of the corn oil had 52.8 % PUFAs, 1.0 % omega-3 and 51.8 % omega-6 fatty acids. The omega-6/omega-3 ratio of the fish oil and the corn oil was 0.07 and 51.8, respectively [43]. At the beginning of the trial, all dogs randomly distributed in either group had statistically similar blood concentrations of each evaluated fatty acid (Table 2). The fatty acid blood concentrations at the end of the trial showed that the dogs from each group had received the intended nutraceutical product. The dogs that had received fish oil exhibited a significant rise in blood levels of the omega-3 fatty acids EPA, DPA and DHA (all with P < 0.001) and a reduction in the concentrations of omega-6 gamma-linolenic acid (GLA), AA and linoleic acid (LA) (all with P < 0.001) and palmitoleic acid (PO) (P = 0.004) at 16 weeks. A significant difference was also found in the blood concentrations of the same fatty acids between dogs from the fish and corn oil groups at the end of the trial, with dogs in the former group having higher EPA, DPA and DHA concentrations and lower GLA, AA and LA concentrations (all with p ≤ 0.001).Table 2 Blood concentrations of fatty acids at baseline and at 16 weeks in dogs receiving fish oil or corn oil supplementation Variable Fish oil Corn oil Between groups Baseline End of trial P-value/direction Baseline End of trial P-value/direction Baseline P-value End P-value Fatty acids, %  PA, C16:0 15.44 ± 1.38 14.99 ± 1.88 0.191 14.99 ± 1.31 14.60 ± 1.28 0.002 ↓ 0.183 0.336  SA, C18:0 21.42 ± 1.49 21.38 ± 1.65 0.891 21.87 ± 1.62 21.85 ± 1.67 0.919 0.246 0.252  OLA, C18:1 ω-9 9.57 ± 1.43 9.43 ± 2.41 0.758 9.13 ± 1.23 8.90 ± 1.14 0.243 0.180 0.255  PO, C16:1 ω7 0.99 ± 0.35 0.85 ± 0.30 0.004 ↓ 0.91 ± 0.28 0.87 ± 0.25 0.407 0.339 0.739  LA, C18:2 ω-6 22.97 ± 2.36 20.91 ± 2.26 <0.001 ↓ 23.17 ± 2.60 24.13 ± 2.79 0.007 ↑ 0.743 <0.001 b  GLA, C18:3 ω-6 0.19 ± 0.07 0.13 ± 0.05 <0.001 ↓ 0.17 ± 0.07 0.20 ± 0.10 0.018 ↑ 0.322 0.001 b  EDA, C20:2 ω6 0.32 ± 0.09 0.31 ± 0.09 0.545 0.37 ± 0.08 0.36 ± 0.08 0.625 0.065 0.044 b  DHGLA, C20:3 ω-6 0.88 ± 0.19 0.81 ± 0.27 0.111 0.95 ± 0.24 0.89 ± .25 0.011 ↓ 0.246 0.223  AA, C20:4 ω-6 19.48 ± 2.26 14.2 ± 2.06 <0.001 ↓ 19.48 ± 2.59 20.15 ± 2.51 0.030 ↑ 0.995 <0.001 b  ALA, C18:3 ω-3 0.33 ± 0.11 0.31 ± 0.08 0.353 0.33 ± 0.14 0.34 ± 0.28 0.919 0.971 0.632  EPA, C20:5 ω-3 0.98 ± 0.48 6.81 ± 2.96 <0.001 ↑ 1.04 ± 0.44 0.70 ± 0.31 <0.001 ↓ 0.633 <0.001 a  DPA, C22:5 ω-3 2.07 ± 0.45 3.11 ± 0.99 <0.001 ↑ 2.16 ± 0.62 1.96 ± 0.57 0.004 ↓ 0.529 <0.001 a  DHA, C22:6 ω-3 1.52 ± 0.59 3.00 ± 1.04 <0.001 ↑ 1.56 ± 0.71 1.30 ± .50 0.006 ↓ 0.790 <0.001 a P-values for changes within and between groups; significant (P < 0.05) values bolded. aHigher values in the fish oil group; bHigher values in the corn oil group. See Table 1 for fatty acid abbreviations. ↑ Value increased significantly; ↓ Value decreased significantly Conversely, dogs in the corn oil group showed a significant elevation in the blood concentrations of the omega-6 fatty acids GLA (P = 0.018), AA (P = 0.030) and LA (P = 0.007) and a decrease in omega-3 EPA (P < 0.001), DPA (P = 0.004) and DHA (P = 0.006) (Table 2). Reductions in Palmitic acid (PA) (P = 0.002) and omega-6 dihomogammalinolenic acid (DHGLA) (P = 0.011) serum concentrations were also observed within the corn oil group, but no significant difference was seen in the comparison between supplement groups at the end of the trial. The corn oil group showed higher concentrations of omega-6 eicosadienoic acid EDA (P = 0.044) at 16 weeks than the dogs that had received fish oil. No side effects were reported or seen in either group. Blood cell counts and biochemical analytes At baseline, all blood parameters were statistically similar among all dogs (Table 3). The haemograms revealed some significant changes within and between the groups of dogs receiving fish or corn oil from the beginning to the end of the trial, although most values remained within the reference ranges (Table 3). The Mean Corpuscular Haemoglobin (MCH) and the Mean Corpuscular Haemoglobin Concentration (MCHC) values were slightly below the normal ranges and increased in dogs from both groups (fish oil: P = 0.003 and P < 0.001, corn oil: P < 0.001 and P < 0.001, respectively) during the 16-week trial. A significant reduction in the Mean Corpuscular Volume (MCV) (P = 0.010) was found only within the fish oil group. No significant difference was found in the comparison between groups.Table 3 Biomarkers for lipid metabolism, oxidation and inflammation; blood counts and biochemistry per group at start (baseline) and end (16 weeks) of fish oil or corn oil trial Variable Fish oil Corn oil Between Groups Baseline End of trial P-value/direction Baseline End of trial P-value/direction Baseline P-value End P-value Oxidation marker  MDA (μmol/l) 3.41 ± 1.34 2.43 ± 0.92 <0.001 ↓ 3.41 ± 1.34 2.41 ± 1.01 0.001 ↓ 0.957 0.926  GSH (mmol/l) 1.97 ± 0.22 2.02 ± 0.17 0.200 2.01 ± 0.21 2.07 ± 0.22 0.184 0.645 0.296  8-OH-dG (pg/ml) 1.54 ± 0.60 1.64 ± 0.74 0.315 1.68 ± 0.60 1.75 ± 0.79 0.410 0.340 0.545  NTBI (μmol/l) −1.36 ± 4.11 −1.92 ± 2.58 0.248 −1.25 ± 2.17 −2.31 ± 1.64 0.002↓ 0.950 0.445  FCar (μmol/l) 18.18 ± 9.78 21.19 ± 9.58 0.004 ↑ 20.53 ± 9.78 22.23 ± 10.12 0.121 0.222 0.657 Haemograms (Unit) RV = Reference value  Haematocrit % RV = 38–57 50.15 ± 4.24 47.55 ± 8.95 0.096 49.22 ± 5.15 48.38 ± 5.88 0.237 0.729 0.881  Haemoglobin (g/l) RV = 140–203 164.23 ± 14.00 162.57 ± 15.19 0.453 160.41 ± 16.71 160.94 ± 19.23 0.816 0.616 0.409  MCV (fl) RV = 67–80 69.17 ± 2.15 68.73 ± 2.23 0.010 ↓ 69.90 ± 3.11 69.76 ± 2.54 0.636 0.339 0.180  MCH (pg) RV = 24–29 22.64 ± 0.77 22.88 ± 0.87 0.003 ↑ 22.76 ± 0.95 23.20 ± 0.97 <0.001 ↑ 0.572 0.494  MCHC (g/l) RV = 345–367 327.33 ± 3.96 332.80 ± 5.93 <0.001 ↑ 325.81 ± 6.18 332.56 ± 6.79 <0.001 ↑ 0.379 0.364  Platelet (n × 109/l) RV = 102–395 304.76 ± 88.37 309.83 ± 98.56 0.692 316.13 ± 93.83 288.41 ± 101.68 0.029 ↓ 0.991 0.513  RBC (n × 1012/l) RV = 5.3–8.0 7.25 ± 0.61 7.11 ± 0.74 0.198 7.04 ± .76 6.94 ± 0.84 0.319 0.474 0.291  WBC (n × 109/l) RV = 5.4–17.4 8.37 ± 2.27 8.20 ± 2.36 0.564 8.45 ± 3.22 8.51 ± 3.31 0.913 0.712 0.975  Neutrophil % RV = 40–80 68.97 ± 10.52 66.46 ± 14.40 0.350 69.89 ± 11.30 70.56 ± 8.44 0.644 0.971 0.195  Lymphocyte % RV = 10–36 18.10 ± 8.61 19.68 ± 7.74 0.109 16.11 ± 7.20 16.82 ± 5.88 0.482 0.504 0.166  Monocyte % RV = 0–13 3.84 ± 2.50 1.77 ± 1.92 <0.001 ↓ 4.11 ± 3.02 3.06 ± 2.70 0.050 ↓ 0.340 0.030 a  Eosinophil % RV = 2–10 7.59 ± 3.53 9.01 ± 4.79 0.069 8.54 ± 5.33 8.42 ± 6.00 0.927 0.605 0.551  Basophil % RV = 0–1 1.47 ± 1.22 0.62 ± 0.62 0.001 1.33 ± 1.08 1.13 ± 0.96 0.396 0.882 0.012 a Blood biochemistry  Urea (mmol/l) RV = 2.4–8.8 5.23 ± 1.61 5.48 ± 1.31 0.308 5.12 ± 1.37 5.25 ± 1.80 0.695 0.430 0.536  Creatinine (μmol/l) RV = 57–116 86.85 ± 9.36 90.65 ± 11.73 0.008 ↑ 88.62 ± 13.34 93.51 ± 14.85 0.001 ↑ 0.372 0.311  ALP (U/l) RV = 33–215 127.83 ± 143.75 146.85 ± 170.32 0.052 147.32 ± 107.88 153.70 ± 116.36 0.283 0.598 0.843  ALT (U/l) RV = 18–77 52.35 ± 24.24 77.22 ± 136.25 0.279 58.60 ± 26.96 58.12 ± 28.77 0.888 0.452 0.414  Albumin (g/l) RV = 30–41 33.68 ± 2.02 33.77 ± 2.57 0.790 32.80 ± 2.39 33.24 ± 2.51 0.160 0.129 0.385  Total protein (g/l) RV = 58–77 63.53 ± 4.07 64.42 ± 4.34 0.158 63.97 ± 3.18 64.50 ± 3.59 0.469 0.990 0.932  Glucose (mmol/l) RV = 4.0–6.4 5.22 ± 0.38 5.34 ± 0.43 0.171 5.18 ± 0.37 5.32 ± 0.47 0.041 ↑ 0.645 0.854  Cholesterol (mmol/l) RV = 3.7–9.8 6.51 ± 1.90 6.94 ± 2.40 0.109 7.13 ± 1.62 7.73 ± 2.03 0.011 ↑ 0.463 0.135  Triglycerides (mmol/l) RV = 0.5–1.1 0.79 ± 0.37 0.71 ± 0.46 0.406 0.80 ± 0.31 0.81 ± 0.36 0.854 0.299 0.335 MCV Mean Corpuscular Volume, MCH Mean Corpuscular Haemoglobin, MCHC Mean Corpuscular Haemoglobin Concentration, RDW Red Cell Distribution Width, RBC Red Blood Cell Count, WBC White Blood Cell Count, ALP Alkaline Phosphatase, ALT Alanine Aminotransferase. P-values for changes within and between groups; significant (P < 0.05) values bolded. aLower counts in the Fish oil group; bHigher values in the Corn oil group. RV Reference value. ↑ Value increased significantly; ↓ Value decreased significantly A significant reduction in monocyte counts in dogs from both the fish oil (P < 0.001) and corn oil (P = 0.050) groups between baseline and 16 weeks was observed. Despite the reduction in monocyte percentage in both groups, there was also a significant difference between the groups (P = 0.030), so that the decrease was greater in the fish oil group. Basophil counts, obtained by manual counting, were reduced significantly within the group of dogs receiving fish oil (P = 0.001) from the beginning to the end of the study, but not within the corn oil group. The comparison between groups also revealed that the basophil counts were significantly lower in the fish oil group (P = 0.012) than in the corn oil group at the end of the trial. The platelet counts were significantly reduced within the corn oil group (P = 0.029) from the beginning to the end of the study. By contrast, in the fish oil group the platelets were increased, albeit not significantly. The creatinine values increased in the serum of dogs from both groups (fish oil: P = 0.008, corn oil: P = 0.001) from the beginning to the end of the study. The final mean creatinine value was higher in the corn oil group, although the difference was not significant. The group of dogs that received corn oil presented a significant elevation in serum values of glucose (P = 0.041) and cholesterol (P = 0.011), whereas the increases in the fish oil group were not significant. At the end of the trial, no statistical difference emerged between groups for either glucose or cholesterol serum values. Oxidative stress markers The biomarkers that were chosen to reflect oxidative stress (MDA, 8-OH-dG and NTBI) and antioxidant capacity of the body (GSH and Free-Car) were statistically similar in the two dog groups at the beginning of the trial (Table 3). MDA values exhibited a significant reduction in dogs within both groups (fish oil: P < 0.001, corn oil: P = 0.001) from the beginning to the end of the trial. The comparison between groups did not reveal a statistical difference at the end of the study. Serum values for 8-OH-dG exhibited a discrete and non-significant elevation within both groups. When these values were compared between groups, the mean value for the corn oil group was slightly higher than for the fish oil group, although the difference did not reach significance. Serum free iron, in the form of non-transferrin bound iron (NTBI), resulted in negative values. Concentrations were higher among dogs receiving corn oil than among those receiving fish oil. The values for serum NTBI exhibited a decrease in both groups, but the reduction was significant only in the corn oil group (P = 0.002). No significant difference was found between groups at the end of the trial. Blood GSH (mg/L) mean values exhibited minor (non-significant) elevations within both groups from the beginning to the end of the trial. The comparison between groups also did not reveal significant differences at the end of the trial. Serum Free-Car values increased in both dog groups from the beginning to the end of the trial, but the difference was significant only in dogs within the fish oil group (P = 0.004). No difference was found between groups at any time. Discussion Dietary supplementation of dogs suffering from OA with fish oil and corn oil clearly resulted in changes in blood concentrations of fatty acids, reflecting the fatty acid composition of the supplements. In a previous study, authors have reported a similar effect on blood profile of fatty acids in healthy dogs undergoing dietary trials with standard fish and corn oils [56]. Thus, the usage of oils for nutraceutical purposes, by the oral route, will promptly promote the bioavailability of fatty acids to the target tissues via the bloodstream. Both oils used in the trials were well tolerated and did not produce any adverse effects in the dogs over the 16 weeks; neither in clinical follow-up, haemograms or serum biochemistry for liver and renal function. So far, the best clinical tools to definitively diagnose, classify and define disease progression or prognosis of OA in dogs are radiography, pain intensity measurements and functional assessments [57]. Nevertheless, such methods are sensitive only when the disease reaches an advanced stage, often with irreversible lesions in cartilage and bone [58]. A number of potential biomarkers have been studied to aid diagnosis, particularly in the early phase, and consensus exists that combined molecules could be applied for the assessment of disease outcome after the use of disease-modifying drugs [58]. A lipidomic approach has suggested that lipid metabolism can be altered during the course of OA in people, and an association between AA release from phospholipids due to an increased activity of phospholipase A2 (PLA2) and inflammatory pain has been described [59]. At least one clinical study has shown a relation between higher blood concentrations of AA, higher blood concentrations of omega-6 PUFAs and synovitis in patients with OA [39]. In this study in dogs, supplement-related changes in the serum fatty acid profile also influenced the associated metabolites. Fish oil supplementation induced a reduction in the serum concentration of LA and therefore also in the metabolites GLA, DHGLA and AA. Although the fish oil preparation used in our study contained 2 % AA, there was a pronounced decrease in serum AA concentration in the fish oil-supplemented dogs. A decrease in the AA concentration is a result of competition between AA and long-chain omega-3 PUFA for incorporation into circulating blood and tissue phospholipids [60]. Similar changes over a long period would be expected also in the phospholipids of other cells’ membranes. Long-chain PUFA omega-3 fatty acids (EPA, DPA and DHA) increased among serum total lipids, consistent with their concentrations in the supplemented fish oil. Supplementation with corn oil increased the LA content among serum total lipids and produced small but consistent increases in LA’s metabolites (EDA, GLA, DHGLA and AA). Although there was a significant difference in omega-6 EDA between the two groups at the end of the trial, this seems to be mostly due to the near-difference between the two groups already at baseline. EDA was also shown to be able to modulate the metabolism of PUFAs and alter the responsiveness of macrophages toward inflammatory stimulation [61]. However, the amelioration of functional parameters in both groups of dogs seen in the clinical data published earlier [43] suggests that not only the fish oil but also the corn oil produces some beneficial effect on the pathological condition. Despite not always reaching statistical significance, the global results of clinical and functional tests [43], together with the laboratory data obtained in this study, indicate a more consistent improvement in the dogs receiving fish oil than in the ones receiving corn oil. But, no substantial global difference between the two groups was seen. Because EPA metabolism in cell membranes results in less pro-inflammatory eicosanoids, namely 3-series prostaglandins and 5-series leukotrienes, than the respective 2-series and 4-series derived from AA metabolism, these omega-3 fatty acids have been considered anti-inflammatory [18]. The anti-inflammatory properties attributed to omega-3 PUFAs that are potentially responsible for healing articular structures, cells, and molecules whose metabolism is affected by OA include reduction of proteoglycan-degrading enzymes, COX-2, and inflammatory cytokines IL-1 and TNF-α, as shown in vitro [45]. There is a need to prove these effects in clinical studies with fish oil products used as nutraceutical compounds to treat OA of companion animals [62, 63] and humans [5, 39]. In the dogs whose samples were analyzed here, the evaluated clinical outcome measures reflecting joint function and pain improved in the group receiving fish oil, whereas only pain improved in the corn oil group [43]. Changes in the diagnostic markers of oxidation (MDA, NTBI and 8-OH-dG) indicate a reduced concentration of oxidation in the fish oil group and suggest a possible effect of EPA and DHA on oxidative damage in tissue lipids. EPA and DHA might have acted as antioxidants or boosted the endogenous antioxidant mechanisms. The exact mechanism of action by which fish oil can regulate the level of plasma and tissue diagnostic marker enzymes in OA condition is unclear. Despite scant knowledge based on in vivo models, it is believed that dietary omega-3 PUFA stabilizes cell membranes by modulating the lipid environment [64], thereby making it less susceptible to damage caused by inflammatory agents. Concerning the oxidative stress molecules involved in OA, our findings of decreased MDA serum concentrations in all dogs may indicate that a similar degree of protection against lipid peroxidation was conferred by both fish and corn oil. Osteoarthritis is believed to be caused by mechanical stress on the joint leading to low grade inflammatory processes [7]. ROS are key components of many normal physiological processes that, at moderate levels, act as indispensable second messengers. However, higher than normal intracellular ROS concentrations can still overpower the homeostatic proteins and cause oxidative damage to the cell and contribute to the onset and progression of OA by inducing chondrocyte death and matrix degradation [65]. It has been shown that MDA, as a marker of lipid peroxidation, is increased in inflamed cartilage tissue and synovial fluid [23, 28, 66] as well as in blood of humans [29, 67] and dogs [32] suffering from OA. In vitro assays have demonstrated cartilage degradation mediated by lipid peroxidation [8]. On the other hand, the possibility of a heightened risk of lipid peroxidation after using fish oil has also been emphasized in a recent review [18], but was not confirmed here. Another study has demonstrated that fish oil decreased MDA levels in healthy dogs [68]. The slight elevation in values of 8-OH-dG that we found in both fish and corn oil groups was non-significant and fell below the reported plasma values for healthy dogs [69]. It indicates that during the study period the dogs did not exhibit evidence of DNA damage detected by measuring 8-OH-dG either before or after receiving fish or corn oil. In humans, elevations of 8-OH-dG have been reported in diverse conditions, including rheumatoid arthritis [70]. However, some authors consider 8-OH-dG measurement from the blood to be a questionable marker of oxidative stress, emphasizing the need for more research [71]. We suggest that the same is true for veterinary medicine, given the scarcity of reports on 8-OH-dG measurements in dogs. In any case, the results described herein contribute to the research, confirming that supplementation with fish oil, and even with corn oil, is significantly associated with decreased concentrations of lipid peroxidation products, namely MDA, in the plasma of OA dogs. Moreover, no evidence for lipid peroxidation emerged with the combined assays used in this study. Other authors have demonstrated that fish oil, which is rich in omega-3 PUFA, decreases the production of inflammatory series of prostaglandins and cytokines [72]. Our results indicate reduced oxidative stress markers in OA dogs supplemented with dietary oils, perhaps involving an equivalent mechanism, boosting the antioxidant defence in vivo. Thus, it can be assumed that dietary fish oil might enhance the antioxidant capability by inhibiting the formation of pro-inflammatory mediators or by upregulating the activities of other enzymes directly or indirectly associated with the antioxidant system. The reduction of serum free iron that we found in the corn oil group (significant reduction) and in the fish oil group (non-significant) suggests at least a less favourable metabolic environment for oxidative damage. Presence and accumulation of iron has been reported to facilitate the formation of ROS, such as the hydroxyl radical, which is the most important radical in 8-OHdG production [54, 73]. The forms of free iron include non-transferrin bound iron (NTBI) [74]. An excess of intercellular free iron causes cell damage by catalysing the production of the hydroxyl radical through the Fenton reaction [75, 76]. Thus, excessive free iron could be a reliable parameter of iron toxicity [75, 77]. Evidence has suggested that oxygen free radicals, especially ones promoted by free iron, play an important role in the development of diseases [75, 77]. Free iron can damage tissues by catalysing the conversion of hydrogen peroxide to free radical ions that attack lipids, proteins and DNA [75, 78]. Fish and corn oil, as seen in the present study, may have played a protective role against increased free iron, possibly through iron binding and export, thus preventing iron-induced toxicity via the Fenton reaction [76]. The results obtained for blood measurements of GSH and Free-Car indicate that both oils were capable of maintaining the integrity of the antioxidant defence system. Moreover, fish oil supplementation induced a significant rise in blood concentrations of Free-Car. L-carnitine protects cells against oxidative damage since it acts as a free radical scavenger and has an important function in lipid metabolism by transporting fatty acids across the inner mitochondrial membrane [79]. Abundant evidence indicates the importance of GSH in maintaining the antioxidant homeostasis in cartilage and during articular pathological conditions [23, 27, 29, 67, 80–83]. Glutathione is a tripeptide (y-Glutamyl-cysteinyl-glycine) thiol present in virtually all animal cells – in normal conditions, both the reduced (GSH) and oxidized (GSSG) forms of glutathione remain in a balanced state [53]. GSH acts as a free radical scavenger and neutralizes superoxides, peroxide radicals and singlet oxygen by donating hydrogen atoms [84]. Antioxidant enzymes like glutathione S-transferase (GST) and glutathione peroxidase (GPX) utilize reduced forms of GSH to carry out their scavenging and detoxification activities, and at the end of the process, an oxidized form of GSSG is the final product [85]. Kumar and Das [86] reported a higher concentration of glutathione in rats with Freund’s adjuvant (CFA)-induced inflammation that were treated with fish oil. Arab et al. [87] noted that DHA enhanced the cellular GSH concentration by elevating the activities of gamma-glutamyl-cysteinyl ligase and glutathione reductase enzymes. PUFAs were reported to enhance the activity of glutathione reductase [88] and glucose-6-phosphate dehydrogenase enzymes [89], which in turn can restore GSH at a faster rate. So far, a number of clinical studies have shown systemic evidence of a reduced antioxidant defence system, including reduced blood concentrations of GSH, in the course of inflammatory articular diseases in humans [90–93]. Although the blood values for GSH were within the reference values for dogs [94] before the trial in dogs with OA, we observed elevated values after 16 weeks of oil supplementation. Thus, we attribute these elevations to the supplementation with either fish oil or corn oil. The increase in GSH may be beneficial since it was concurrent with a reduction of MDA in both groups of dogs. In fact, it has been shown in dogs that GSH has an important role in adequate lipid peroxide detoxification during illness [94]. In the literature on canine OA, however, the only report on blood markers of oxidative stress associated with the disease was in a study using an experimental Pond-Nuki model [30]. As far as we know, no studies exist on oxidative and antioxidant parameters in naturally occurring canine OA, apart from the results reported here. Moreover, our study shows inedited results of the effect of supplementation with fish and corn oil on the oxidative/antioxidant balance in natural occurring canine OA. Nonetheless, further research should include analyses of serum isoprostane and prostaglandin as markers of oxidation and inflammation. In our study, limited biological samples were available. The results of blood biochemistry, all within the normal range, support the conclusion that 16-week supplementation with fish or corn oil was not associated with toxicity or metabolic imbalance in dogs. Although hyperglycaemia was listed as a potential adverse effect of fish oil supplementation of dogs [18], we found no evidence of such an effect. In fact, the glucose concentrations were significantly higher within the corn oil group after 16 weeks of supplementation, although no significant difference was seen between the two groups. Moreover, all values remained within the reference range. Similarly, the cholesterol concentrations were significantly higher in the dogs after receiving corn oil, but not fish oil, but at the end of the 16-week trial they remained within normal reference values. These are in tune with a recent study where fish oil supplementation increased plasma triglycerides and ghrelin but did not appear to affect protein metabolism or postprandial glycaemia in adult lean dogs, whereas an increase in cholesterol concentration could be seen in the control diet group [95]. Concerning haematology tests, most values stayed within references for the dogs in our study. Administration of fish or corn oil during 16 weeks caused no change in blood haemoglobin concentration, although significant elevations of MCH and MCHC were seen in both groups from baseline to the end of the trial. The elevation of haemoglobin concentration within cells in both groups of dogs may be attributed to the fact that the mean values were at the lower end of the reference range at the beginning of the study. Thus, the oils seem to have enhanced haemoglobin synthesis towards normality during the 16-week study. There are no similar or comparable reports in dogs in the literature. In humans, there are reports of haemoglobin reduction in healthy individuals [96] and no change in patients with risk of anaemia after supplementation with fish oil [97, 98]. Likewise, we observed a significant reduction of the MCV values within the fish oil group, but comparable changes in dogs were not found elsewhere; also, no statistical difference was observed between groups. Elevations in packed cell volume after ingestion of fish oil for six weeks have been reported in healthy human beings [96], and no difference were found in patients with risk of anaemia [97]. The leucocyte counts of dogs in the present study revealed a significant reduction in the number of monocytes and basophils in the group receiving fish oil, from baseline to the end of the trial. These changes were also significant in the comparison with the dogs receiving corn oil over the same period. However, all counts, including those for other white cells, remained within reference values for dogs. Changes in eicosanoid expression towards the generation of leukotriene B5 induced by EPA ingestion have been speculated not to promote differences in circulating leucocyte counts in humans [99], although at least one study has shown increases in total leucocyte and monocyte counts after fish oil supplement [96]. It has also been shown that omega-3 fatty acids are able to interfere with leucocyte chemotaxis, adhesion molecule expression and leucocyte-endothelial adhesive interactions [60], the effect of which, attributed to the decrease of AA content of cells involved in immune responses, is associated with clinical improvement of inflammatory diseases including rheumatoid arthritis [100]. Even though a higher ratio of leukotriene B5/B4 has been reported in dogs supplemented with fish oil than in dogs given corn oil [17], no information on leucocyte counts in studies with fish oil supplementation of healthy dogs or dogs with OA was earlier available in the literature. A clinical study in dogs has, however, shown that the concentration of leukotriene B4 in osteoarthritic hip joint capsules was higher than in clinically normal ones [101]. Still, extensive and recent reviews on the therapeutic use of fish oil for animal diseases, including OA, do not mention its impact on haemograms [62]. The platelet counts in the present study showed that supplementation with corn oil resulted in lowered circulating numbers from the beginning to the end of the 16-week trial and that fish oil ingestion was associated with a non-significant elevation of platelets during the same period. Nevertheless, the values remained within the normal ranges for dogs in both groups, and the comparison between groups at the end of the trial showed negligible and non-significant differences. No bleeding events occurred during the clinical follow-up of the dogs. Although altered platelet function has been considered a potential risk of fish oil supplementation in dogs, a recent review concluded that there is no evidence of a harmful effect on platelet function or counts or clinical consequences due to fish oil or its fatty acids [18]. From our observations of both dog groups, we can hypothesize that the oils, particularly the fish oil given to the dogs as a supplement seemed to have promoted an abundance of beneficial polyunsaturated DHA and EPA, thus altering the availability of certain fatty acids in the body. At the cell membrane level, these fatty acids could have facilitated the repair of mitochondrial and other membranes, which could have been damaged by ROS due to inflammation in OA. As depicted by the significantly elevated concentrations of carnitine and reduced concentrations of MDA within the group of dogs that received fish oil, these fatty acids may also have promoted a better state of redox regulation. Clinically, the favourable shift in the membrane composition of lipids would be related to the reduction of pain and lameness indicators described in the dogs [43]. As OA often has a neuropathic aspect when chronic, the ROS contribution might be even more pronounced clinically at later stages: Neurons are especially sensitive to ROS since neurons have greater energy demands to function as compared to glial and other cells in the central nervous system [102, 103]. Also lipid peroxidation products may contribute to neuropathic pain in OA as they have been shown to contribute to neuropathic pain in chronic spinal cord injury animals [104]. However, the mechanism that ROS and lipid peroxidation play in chronic neuropathic pain is not well understood. When the organism has more beneficial fatty acids it becomes less prone to peroxidation. For therapeutic use, a recent review has discussed and clarified the basis for what the authors called Lipid Replacement Therapy, in which membrane lipids can be replaced by, among others, oral supplements, promoting therapeutic effects on different diseases [105]. As well, Scicchitano et al. (2012) have reviewed and proposed an important role for nutraceuticals, including fish oils, in the control of lipid metabolism thus improving the overall burden of oxidation of lipids [106]. In the present study, fish oil supplementation seems to have promoted such effects, and corn oil to a lesser extent [43]. Considering that OA is a chronic disease whose clinical presentation takes longer than the actual onset of pathological symptoms, our results suggest that a longer follow-up would maybe have allowed clearer differences to emerge. In the future, the extent to which fish oil modifies the redox balance in healthy dogs and dogs with OA should be evaluated. Future fish oil supplementation study groups should also be stratified according to if they have a neuropathic aspect of pain. Further, one should consider another control substance than corn oil. Conclusion In global terms, there was no clear difference between dogs fed fish oil or corn oil. The oxidative status markers indicated a decrease in oxidative stress (MDA and iron) and an increase in antioxidant capacity (Free-Car and GSH) both in dogs receiving fish oil and corn oil. The elevation of the antioxidant capacity and the decrease of inflammatory monocytes and basophils were, however, more significant in the fish oil group and are unprecedented results reported in dogs with OA. Despite the fact that values remained within the normal range, cholesterol and glucose increased significantly only in the corn oil group. Abbreviations 8-OH-dG8-hydroxy-2-deoxyguanosine AAArachidonic acid ALAα-linolenic acid ALPAlkaline phosphatase ALTAlanine aminotransferase ANOVAAnalysis of variance BWBody weight CATCatalase CFAComplete Freund’s adjuvant CONSORTConsolidated standards of reporting trials COXCyclooxygenase DHADocosahexaenoic acid DHGLADihomo-gamma-linolenic acid DMARDDisease-modifying anti-rheumatic drug DPADocosapentaenoic acid EPAEicosapentaenoic acid ETAEicosatetraenoic acid FAFatty acid FFAFree fatty acid Free-CarFree carnitine GLAGamma-linolenic acid GPXGlutathione peroxidase GSHReduced glutathione GSSGOxidized glutathione GSTGlutathione S-transferase HCPIHelsinki chronic pain index HPLCHigh-performance liquid chromatography IL-1Interleukin-1 LALinoleic acid MCHMean Corpuscular Haemoglobin MCHCMean Corpuscular Haemoglobin Concentration MCVMean Corpuscular Volume MDAMalondialdehyde MMPMatrix metalloproteinase NONitric oxide NSAIDNon-steroidal anti-inflammatory drug NTBINon-transferrin bound iron OAOsteoarthritis OLAOleic acid PAPalmitic acid PUFApolyunsaturated fatty acid PVFPeak vertical force RARheumatoid arthritis ROSReactive oxygen species RVReference values SDStandard deviation SDAStearidonic acid SODSuperoxide dismutase TBARSThiobarbituric acid reactive substance TIMPTissue inhibitor metalloproteinase TNF-αTransforming growth factor alpha TPTotal protein W0Baseline Wxx weeks after baseline W-xx weeks before baseline ωOmega Acknowledgements We thank the Helvi Knuuttila Foundation for funding study materials, evaluation costs at the animal hospital, and analyses; Mila Laboratories (Helsinki, Finland) for enabling analyses at a reduced price in their facilities; Nutraceuticoils (Belgium) for paying for the dogs’ radiographs and blood samples for inclusion, the serum fatty acid analysis before and after the trial, the dogs’ foods, the supplements to be tested and their identical controls. We also thank the CAPES Foundation, Ministry of Education of Brazil, for the postdoctoral scholarship to SMBM (Proc. no. BEX 2686/13-9). Availability of data and materials The datasets generated during and analysed during the current study are not publicly available at this time due to being part of a PhD work that still is underway. They are, however, available from the corresponding author on request. Authors’ contributions SMBM participated in data analyses and interpretation and wrote the manuscript. JA participated in evaluations and sample collections, input all data into the database and helped to draft the manuscript. SS helped with planning the laboratory methods and interpreting the laboratory data. MG planned and analysed the fatty acid data. FA planned, organized and analysed the results of the oxidative stress laboratory tests. SO carried out the oxidative stress laboratory assays. AKHB conceived the study, coordinated its planning, organized the finances and the research agreement, participated in most of the data collection, performed the statistical analyses and helped to draft the manuscript. All authors read and approved the final manuscript. Authors’ information SMBM carried out her postdoctoral research at the Department of Equine and Small Animal Clinical Medicine, University of Helsinki, Finland; she is an associate professor and a researcher in small animal internal medicine at the Department of Anatomy, Pathology and Clinics, School of Veterinary Medicine and Zootechny, Federal University of Bahia, Brazil. JA is conducting her PhD on the relationships between gene expression, disease and nutrition. SS is a professor in veterinary clinical chemistry and the head of the clinical laboratories of the Department of Equine and Small Animal Medicine, University of Helsinki, Finland. MG is a specialist in nutraceutical oils and has a PhD in Animal Sciences. At the time of the study he was working as a visiting professor at the Swedish University of Agricultural Sciences, Department of animal nutrition and management. Nowadays he has his own firm that produces and sells oil supplements for dogs and cats. FA is a researcher on free radicals and inflammation, antioxidants and DNA repair. SO is a researcher on iron, iron overload, iron chelating agents and lipid peroxidation. AKHB is an assistant and adjunct professor and a principal investigator in small animal surgery and the head of the Pain and Rehabilitation Clinic at the Department of Equine and Small Animal Clinical Medicine, University of Helsinki, Finland, where she studies non-pharmaceutical treatments for canine OA, including food and nutraceuticals. Competing interests The authors declare that they have no competing interests. The study protocol was written at and accepted by the Department of Equine and Small Animal Medicine, University of Helsinki, Finland. The corresponding author/principal investigator (AKHB) proposed the study to the fish oil manufacturing company Nutriceuticoils and was not remunerated for the work. The execution, ordering, payment and distribution of the project were taken care of by the University and the principal investigator in Finland. However, Nutriceuticoils paid for the dogs’ radiographs and blood samples for inclusion, the serum fatty acid analysis before and after the trial, the dogs’ foods, the supplements to be tested and their identical controls. Nutriceuticoils had no input in study design, data collection/analysis/interpretation, and the decision to write and submit the paper for publication was the sole responsibility of the corresponding author. No medical writer was involved in drafting this paper. ==== Refs References 1. Johnston SA Osteoarthritis. Joint anatomy, physiology, and pathobiology Vet Clin North Am Small Anim Pract 1997 27 699 723 10.1016/S0195-5616(97)50076-3 9243777 2. Innes J. Canine osteoarthritis - initiating factors. In Hill’s European symposium on osteoarthritis and joint health. Genova; ©2005 Hill’s Pet Nutrition, Inc; 2005. p. 6–13. 3. Leppänen M Saloniemi H Controlling canine hip dysplasia in Finland Prev Vet Med 1999 42 121 131 10.1016/S0167-5877(99)00059-8 10551430 4. Lauten SD Nutritional risks to large-breed dogs: from weaning to the geriatric years Vet Clin North Am Small Anim Pract 2006 36 1345 1359 10.1016/j.cvsm.2006.09.003 17085239 5. Lopez HL Nutritional interventions to prevent and treat osteoarthritis. Part I: focus on fatty acids and macronutrients PM R 2012 4 Suppl 5 S145 154 10.1016/j.pmrj.2012.02.022 22632694 6. Sutton S Clutterbuck A Harris P Gent T Freeman S Foster N Barrett-Jolley R Mobasheri A The contribution of the synovium, synovial derived inflammatory cytokines and neuropeptides to the pathogenesis of osteoarthritis Vet J 2009 179 10 24 10.1016/j.tvjl.2007.08.013 17911037 7. Berenbaum F Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!) Osteoarthritis Cartilage 2013 21 16 21 10.1016/j.joca.2012.11.012 23194896 8. Tiku ML Shah R Allison GT Evidence linking chondrocyte lipid peroxidation to cartilage matrix protein degradation. Possible role in cartilage aging and the pathogenesis of osteoarthritis J Biol Chem 2000 275 20069 20076 10.1074/jbc.M907604199 10867027 9. Lawler DF Larson BT Ballam JM Smith GK Biery DN Evans RH Greeley EH Segre M Stowe HD Kealy RD Diet restriction and ageing in the dog: major observations over two decades Br J Nutr 2008 99 793 805 10.1017/S0007114507871686 18062831 10. Fantuzzi G Adipose tissue, adipokines, and inflammation J Allergy Clin Immunol 2005 115 911 919 10.1016/j.jaci.2005.02.023 15867843 11. Laflamme DP Companion Animals Symposium: Obesity in dogs and cats: What is wrong with being fat? J Anim Sci 2012 90 1653 1662 10.2527/jas.2011-4571 21984724 12. Rajappa M Tagirasa R Nandeesha H Hamide A Sundar I Hariharan P Ananthanarayanan PH Vengattaraman A Thiyagarajan D Harichandrakumar KT Synergy of iron, high sensitivity C-reactive protein and ceruloplasmin with oxidative stress in non-diabetic normo-tensive South Indian obese men Diabetol Metab Syndr 2013 7 214 217 10.1016/j.dsx.2013.10.014 13. Goldring SR Goldring MB The role of cytokines in cartilage matrix degeneration in osteoarthritis Clin Orthop Rel Res 2004 427 27 36 10.1097/01.blo.0000144854.66565.8f 14. Smith CJ Zhang Y Koboldt CM Muhammad J Zweifel BS Shaffer A Talley JJ Masferrer JL Seibert K Isakson PC Pharmacological analysis of cyclooxygenase-1 in inflammation Proc Natl Acad Sci U S A 1998 95 13313 13318 10.1073/pnas.95.22.13313 9789085 15. Blobaum AL Marnett LJ Structural and functional basis of cyclooxygenase inhibition J Med Chem 2007 50 1425 1441 10.1021/jm0613166 17341061 16. Nugteren DH Hazelhof E Isolation and properties of intermediates in prostaglandin biosynthesis Biochim Biophys Acta 1973 326 448 461 10.1016/0005-2760(73)90145-8 4776443 17. Hall JA Henry LR Jha S Skinner MM Jewell DE Wander RC Dietary (n-3) fatty acids alter plasma fatty acids and leukotriene B synthesis by stimulated neutrophils from healthy geriatric Beagles Prostaglandins Leukot Essent Fatty Acids 2005 73 335 341 10.1016/j.plefa.2005.07.006 16146686 18. Lenox CE Bauer JE Potential adverse effects of omega-3 fatty acids in dogs and cats J Vet Intern Med 2013 27 217 226 10.1111/jvim.12033 23323770 19. Schmitz G Ecker J The opposing effects of n-3 and n-6 fatty acids Prog Lipid Res 2008 47 147 155 10.1016/j.plipres.2007.12.004 18198131 20. Henrotin YE Bruckner P Pujol J-PL The role of reactive oxygen species in homeostasis and degradation of cartilage Osteoarthritis Cartilage 2003 11 747 755 10.1016/S1063-4584(03)00150-X 13129694 21. Afonso V Champy R Mitrovic D Collin P Lomri A Reactive oxygen species and superoxide dismutases: Role in joint diseases Joint Bone Spine 2007 74 324 329 10.1016/j.jbspin.2007.02.002 17590367 22. Doom M Bruin T Rooster H Bree H Cox E Immunopathological mechanisms in dogs with rupture of the cranial cruciate ligament Vet Immunol Immunopathol 2008 125 143 161 10.1016/j.vetimm.2008.05.023 18621423 23. Dycus DL Au AY Grzanna MW Wardlaw JL Frondoza CG Modulation of inflammation and oxidative stress in canine chondrocytes Am J Vet Res 2013 74 983 989 10.2460/ajvr.74.7.983 23802669 24. Hewicker-Trautwein M Cartera SD Bennett D Kellya DF Immunocytochemical demonstration of lymphocyte subsets and MHC class II antigen expression in synovial membranes from dogs with rheumatoid arthritis and degenerative joint disease Vet Immunol Immunopathol 1999 67 341 357 10.1016/S0165-2427(98)00229-3 10206202 25. Muir P Kelly JL MarvelS J Heinrich DA Schaefer SL Manley PA Tewari K Singh A Suresh M Hao Z Plisch E Lymphocyte populations in joint tissues from dogs with inflammatory stifle arthritis and associated degenerative cranial cruciate ligament rupture Vet Surg 2011 40 753 761 10.1111/j.1532-950X.2011.00867.x 21770988 26. Yudoh K Trieu N Nakamura H Hongo-Masuko K Kato T Nishioka K Potential involvement of oxidative stress in cartilage senescence and development of osteoarthritis: oxidative stress induces chondrocyte telomere instability and down regulation of chondrocyte function Arthritis Res Ther 2005 7 380 391 10.1186/ar1499 27. Regan EA Bowlerand RP Crapo JD Joint fluid antioxidants are decreased in osteoarthritic joints compared to joints with macroscopically intact cartilage and sub-acute injury Osteoarthritis Cartilage 2008 16 515 521 10.1016/j.joca.2007.09.001 18203633 28. Ostalowska A Birkner E Wiecha M Kasperczyk S Kasperczyk A Kapolka D Zon-Giebel A Lipid peroxidation and antioxidant enzymes in synovial fluid of patients with primary and secondary osteoarthritis of the knee joint Osteoarthritis Cartilage 2006 14 139 145 10.1016/j.joca.2005.08.009 16289733 29. Sarban S Kocyigit A Yazar M Isikan UE Plasma total antioxidant capacity, lipid peroxidation, and erythrocyte antioxidant enzyme activities in patients with rheumatoid arthritis and osteoarthritis Clin Biochem 2005 38 981 986 10.1016/j.clinbiochem.2005.08.003 16150434 30. Goranov NV Serum markers of lipid peroxidation, antioxidant enzymatic defense, and collagen degradation in an experimental (Pond-Nuki) canine model of osteoarthritis Vet Clin Pathol 2007 36 192 195 10.1111/j.1939-165X.2007.tb00208.x 17523095 31. Brown DC Boston RC Coyne JC Farrar JT Development and psychometric testing of an instrument to measure chronic pain in dogs with osteoarthritis Am J Vet Res 2007 68 631 637 10.2460/ajvr.68.6.631 17542696 32. Sellam J Berenbaum F The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis Nat Rev Rheumatol 2010 6 625 635 10.1038/nrrheum.2010.159 20924410 33. Darlington LG Stone TW Antioxidants and fatty acids in the amelioration of rheumatoid arthritis and related disorders Br J Nutr 2001 85 251 269 10.1079/BJN2000239 11299072 34. van der Tempel H Tulleken JE Limburg PC Muskiet FA van Rijswijk MH Effects of fish oil supplementation in rheumatoid arthritis Ann Rheum Dis 1990 49 76 80 10.1136/ard.49.2.76 2138449 35. Ruggiero C Lattanzio F Lauretani F Gasperini B Andres-Lacueva C Cherubini A Omega-3 polyunsaturated fatty acids and immune-mediated diseases: inflammatory bowel disease and rheumatoid arthritis Curr Pharm Des 2009 15 4135 4148 10.2174/138161209789909746 20041815 36. Wooki K McMurray DN Chapkin RS n-3 Polyunsaturated fatty acids - physiological relevance of dose Prostaglandins Leukot Essent Fatty Acids 2010 82 155 158 10.1016/j.plefa.2010.02.028 20188532 37. Richard D Kefi K Barbe U Bausero P Visioli F Polyunsaturated fatty acids as antioxidants Pharmacol Res 2008 57 451 455 10.1016/j.phrs.2008.05.002 18583147 38. Gruenwald J Petzold E Busch R Petzold HP Graubaum HJ Effect of glucosamine sulfate with or without omega-3 fatty acids in patients with osteoarthritis Adv Ther 2009 26 858 871 10.1007/s12325-009-0060-3 19756416 39. Baker KR Matthan NR Lichtenstein AH Niu J Guermazi A Roemer F Grainger A Nevitt MC Clancy M Lewis CE Torner JC Felson DT Association of plasma n-6 and n-3 polyunsaturated fatty acids with synovitis in the knee: the MOST study Osteoarthritis Cartilage 2012 20 382 387 10.1016/j.joca.2012.01.021 22353693 40. Zawadzki M Janosch C Szechinski J Perna canaliculus lipid complex PCSO-524™ demonstrated pain relief for osteoarthritis patients benchmarked against fish oil, a randomized trial, without placebo control Mar Drugs 2013 11 1920 1935 10.3390/md11061920 23739042 41. Roush JK Cross AR Renberg WC Dodd CE Sixby KA Fritsch DA Allen TA Jewell DE Richardson DC Leventhal PS Hahn KA Evaluation of the effects of dietary supplementation with fish oil omega-3 fatty acids on weight bearing in dogs with osteoarthritis J Am Vet Med Assoc 2010 236 67 73 10.2460/javma.236.1.67 20043801 42. Fritsch DA Allen TA Dodd CE Jewell DE Sixby KA Leventhal PS Brejda J Hahn KAJ A multicenter study of the effect of dietary supplementation with fish oil omega-3 fatty acids on carprofen dosage in dogs with osteoarthritis Am Vet Med Assoc 2010 236 535 539 10.2460/javma.236.5.535 43. Hielm-Björkman A Roine J Elo K Lappalainen A Junnila J Laitinen-Vapaavuori O An un-commissioned randomized, placebo-controlled double-blind study to test the effect of deep sea fish oil as a pain reliever for dogs suffering from canine OA BMC Vet Res 2012 8 157 10.1186/1746-6148-8-157 22950577 44. Rialland P Bichot S Lussier B Moreau M Beaudry F del Castillo JR Gauvin D Troncy E Effect of a diet enriched with green-lipped mussel on pain behavior and functioning in dogs with clinical osteoarthritis Can J Vet Res 2013 77 66 74 23814358 45. Curtis CL Hughes CE Flannery CR Little CB Harwood JL Caterson B N-3 fatty acids specifically modulate catabolic factors involved in articular cartilage degradation J Biol Chem 2000 275 721 724 10.1074/jbc.275.2.721 10625599 46. Tattersall AL Wilkins RJ Effects of hexosamines and omega-3/omega-6 fatty acids on pH regulation by interleukin 1-treated isolated bovine articular chondrocytes Pflugers Arch 2008 456 501 506 10.1007/s00424-007-0425-x 18204856 47. Hielm-Björkman A Rita H Tulamo R-M Psychometric testing of the Helsinki chronic pain index by completion of a questionnaire in Finnish by owners of dogs with chronic signs of pain caused by osteoarthritis Am J Vet Res 2009 70 727 734 10.2460/ajvr.70.6.727 19496661 48. Schulz KF Altman DG Moher D CONSORT Group CONSORT 2010 Statement: updating guidelines for reporting parallel group randomized trials Br Med J 2010 340 c332 10.1136/bmj.c332 20332509 49. Seppänen-Laakso T Laakso I Hiltunen R Simultaneous analysis of bound and free fatty acids in human plasma by quantitative gas chromatography Acta Pharmaceutica Fennica 1990 99 109 117 50. IFCC 2002/5 IFCC Primary Reference Procedures for the Measurement of Catalytic Activity Concentrations of Enzymes at 37 °C Part 4. Reference Procedure for the Measurement of Catalytic Concentration of Alanine Aminotransferase Clin Chem Lab Med 2002 40 718 724 12241021 51. Scandinavian Society for Clinical Chemistry and Clinical Physiology Recommended methods for the determination of four enzymes in blood Scand J Clin Lab Invest 1974 33 291 306 10.3109/00365517409082499 4369123 52. Halliwell B Chirico S Lipid peroxidation: its mechanism, measurement, and significance Am J Clin Nutr 1993 57 715 724 53. Atroshi F Sandholm M Red blood cell glutathione (GSH) as a marker of milk production in Finn sheep Res Vet Sci 1982 33 256 259 7146633 54. Ounjaijean S Westermarck T Partinen M Plonka-Poltorak E Kaipainen P Kaski M Fucharoen S Srichairatanakool S Atroshi F Increase in non-transferrin bound iron and the oxidative stress status in epilepsy patients treated using valproic acid monotherapy Int J Clin Pharmacol Ther 2011 49 268 276 10.5414/CP201466 21429441 55. Singh S Hider RC Porter JB A direct method for quantification of non-transferrin-bound iron Anal Biochem 1990 186 320 323 10.1016/0003-2697(90)90088-Q 2363505 56. Hall JA Picton RA Skinner MM Jewell DE Wander RC The (n-3) fatty acid dose, independent of the (n-6) to (n-3) fatty acid ratio, affects the plasma fatty acid profile of normal dogs J Nutr 2006 136 2338 2344 16920851 57. Schulz KF Fossum TW Diseases of the joint Small animal surgery 2007 3 St. Louis Mosby Elsevier 1143 1315 58. Lotz M Martel-Pelletier J Christiansen C Brandi M-L Bruyère O Chapurlat R Collette J Cooper C Giacovelli G Kanis JA Karsdal MA Kraus V Lems WF Meulenbelt I Pelletier J-P Raynauld J-P Reiter-Niesert S Rizzoli R Sandell LJ Van Spil WE Reginster J-Y Value of biomarkers in osteoarthritis: current status and perspectives Ann Rheum Dis 2013 72 1756 1763 10.1136/annrheumdis-2013-203726 23897772 59. Castro-Perez JM Kamphorst J Degroot J Lafeber F Goshawk J Yu K Shockcor JP Vreeken RJ Hankemeier T Comprehensive LC-MS E lipidomic analysis using a shotgun approach and its application to biomarker detection and identification in osteoarthritis patients J Proteome Res 2010 9 2377 2389 10.1021/pr901094j 20355720 60. Calder PC Omega-3 polyunsaturated fatty acids and inflammatory processes: nutrition or pharmacology? Br J Clin Pharmacol 2013 75 645 662 22765297 61. Huang YS Huang WC Li CW Chuang LT Eicosadienoic acid differentially modulates production of pro-inflammatory modulators in murine macrophages Mol Cell Biochem 2011 358 85 94 10.1007/s11010-011-0924-0 21688154 62. Vandeweerd JM Coisnon C Clegg P Cambier C Pierson A Hontoir F Saegerman C Gustin P Buczinski S Systematic review of efficacy of nutraceuticals to alleviate clinical signs of osteoarthritis J Vet Intern Med 2012 26 448 456 10.1111/j.1939-1676.2012.00901.x 22404506 63. Bauer JE Therapeutic use of fish oils in companion animals J Am Vet Med Assoc 2011 239 1441 1451 10.2460/javma.239.11.1441 22087720 64. Kim YR Lee YH Park KA Kim JO Ha YL A simple method for the preparation of highly pure conjugated linoleic acid (CLA) synthesized from safflower seed oil J Food Sci Nutr 2000 5 10 14 65. Russell EG Cotter TG New insight into the role of Reactive Oxygen Species (ROS) in cellular signal-transduction processes Int Rev Cell Mol Biol 2015 319 221 54 10.1016/bs.ircmb.2015.07.004 26404470 66. Watari T Naito K Sakamoto K Kurosawa H Nagaoka I Kaneko K Evaluation of the effect of oxidative stress on articular cartilage in spontaneously osteoarthritic STR/OrtCrlj mice by measuring the biomarkers for oxidative stress and type II collagen degradation/synthesis Exp Ther Med 2011 2 245 250 22977492 67. Suantawee T Tantavisut S Adisakwattana S Tanavalee A Yuktanandana P Anomasiri W Deepaisarnsakul B Honsawek S Oxidative stress, vitamin E, and antioxidant capacity in knee osteoarthritis J Clin Diagn Res 2013 7 1855 1859 24179881 68. Kearns RJ Hayek MG Turek JJ Meydani M Burr JR Greene RJ Marshall CA Adams SM Borgert RC Reinhart GA Effect of age, breed and dietary omega-6 (n-6): omega-3 (n-3) fatty acid ratio on immune function, eicosanoid production, and lipid peroxidation in young and aged dogs Vet Immunol Immunopathol 1999 69 165 83 10.1016/S0165-2427(99)00052-5 10507303 69. Baskin CR Hinchcliff KW DiSilvestro RA Reinhart GA Hayek MG Chew BP Burr JR Swenson RA Effects of dietary antioxidant supplementation on oxidative damage and resistance to oxidative damage during prolonged exercise in sled dogs Am J Vet Res 2000 61 886 891 10.2460/ajvr.2000.61.886 10951977 70. Jacob KD Noren Hooten N Trzeciak AR Evans MK Markers of oxidant stress that are clinically relevant in aging and age-related disease Mech Ageing Dev 2013 134 139 157 10.1016/j.mad.2013.02.008 23428415 71. Poulsen HE Nadal LL Broedbaek K Nielsen PE Weimann A Detection and interpretation of 8-oxodG and 8-oxoGua in urine, plasma and cerebrospinal fluid Biochim Biophys Acta 1840 2014 801 808 72. Appel MJ Woutersen RA Effects of dietary fish oil (MaxEPA) on N-nitrosobis(2-oxopropyl)amine (BOP)-induced pancreatic carcinogenesis in hamsters Cancer Lett 1995 94 179 189 10.1016/0304-3835(95)03848-Q 7634246 73. Kato J Kobune M Nakamura T Kuroiwa G Takada K Takimoto R Sato Y Fujikawa K Takahashi M Takayama T Ikeda T Niitsu Y Normalization of elevated hepatic 8-hydroxy-2′-deoxyguanosine levels in chronic hepatitis C patients by phlebotomy and low iron diet Cancer Res 2001 61 8697 8702 11751387 74. Hershko C Grady RW Cerami A Mechanism of iron chelation in the hypertransfused rat: definition of two alternative pathways of iron mobilization J Lab Clin Med 1978 92 144 151 681807 75. Imlay JA Chin SM Linn S Toxic DNA damage by hydrogen peroxide through the Fenton reaction in vivo and in vitro Science 1988 240 640 642 10.1126/science.2834821 2834821 76. Bhubhanil S Chamsing J Sittipo P Chaoprasid P Sukchawalit R Mongkolsuk S Roles of Agrobacterium tumefaciens membrane-bound ferritin (MbfA) in iron transport and resistance to iron under acidic conditions Microbiology 2014 160 863 871 10.1099/mic.0.076802-0 24600024 77. Tang Y Li Y Yu H Gao C Liu L Xing M Liu L Yao P Quercetin attenuates chronic ethanol hepatotoxicity: implication of free iron uptake and release Food Chem Toxicol 2014 67 131 138 10.1016/j.fct.2014.02.022 24569067 78. Cao D Zheng J Xian LF Tang GM Sun XJ Xu WG Tao HY Liu K Liu WW Role of iron in lung injury-induced by hyperoxia Undersea Hyperb Med 2014 41 27 31 24649714 79. Ribas GS Vargas CR Wajner M L-carnitine supplementation as a potential antioxidant therapy for inherited neurometabolic disorders Gene 2014 533 469 476 10.1016/j.gene.2013.10.017 24148561 80. Studer RK Nitric oxide decreases IGF-1 receptor function in vitro; glutathione depletion enhances this effect in vivo Osteoarthritis Cartilage 2004 12 863 869 10.1016/j.joca.2004.07.005 15501401 81. Surapaneni KM Venkataramana G Status of lipid peroxidation, glutathione, ascorbic acid, vitamin E and antioxidant enzymes in patients with osteoarthritis Indian J Med Sci 2007 61 9 14 10.4103/0019-5359.29592 17197733 82. Mendis E Kim MM Rajapakse N Kim SK Sulfated glucosamine inhibits oxidation of biomolecules in cells via a mechanism involving intracellular free radical scavenging Eur J Pharmacol 2008 579 74 85 10.1016/j.ejphar.2007.10.027 18036590 83. Vaillancourt F Fahmi H Shi Q Lavigne P Ranger P Fernandes JC Benderdour M 4-Hydroxynonenal induces apoptosis in human osteoarthritic chondrocytes: the protective role of glutathione-S-transferase Arthritis Res Ther 2008 10 R107 10.1186/ar2503 18782442 84. Meister A Selective modification of glutathione metabolism Science 1983 220 472 477 10.1126/science.6836290 6836290 85. Meister A Function of glutathione in kidney via the gamma-glutamyl cycle Med Clin North Am 1975 59 649 66 10.1016/S0025-7125(16)32015-6 236410 86. Kumar GS Das UN Free radical-dependent suppression of growth of mouse myeloma cells by alpha-linolenic and eicosapentaenoic acids in vitro Cancer Lett 1995 92 27 38 10.1016/0304-3835(95)03753-J 7757958 87. Arab K Rossary A Flourié F Tourneur Y Steghens JP Docosahexaenoic acid enhances the antioxidant response of human fibroblasts by upregulating gamma-glutamyl-cysteinyl ligase and glutathione reductase Br J Nutr 2006 95 18 26 10.1079/BJN20051626 16441913 88. Hsu CS Shen WW Hsueh YM Yeh SL Soy isoflavone supplementation in postmenopausal women. Effects on plasma lipids, antioxidant enzyme activities and bone density J Reprod Med 2001 46 221 226 11304862 89. Yilmaz HR Songur A Ozyurt B Zararsiz I Sarsilmaz M The effects of n-3 polyunsaturated fatty acids by gavage on some metabolic enzymes of rat liver Prostaglandins Leukot Essent Fatty Acids 2004 71 131 135 10.1016/j.plefa.2004.03.002 15207530 90. Maneesh M Jayalekshmi H Suma T Chatterjee S Chakrabarti A Singh TA Evidence for oxidative stress in osteoarthritis Ind J Clin Biochem 2005 20 129 130 10.1007/BF02893057 91. Fernandez-Moreno M Soto-Hermida A Pertega S Oreiro N Fernandez-Lopez C Rego-Perez I Blanco FJ Mitochondrial DNA (mtDNA) haplogroups and serum levels of anti-oxidant enzymes in patients with osteoarthritis BMC Musculoskelet Disord 2011 12 264 10.1186/1471-2474-12-264 22108257 92. Kotani K Sakane N Kamimoto M Taniguchi N Levels of reactive oxygen metabolites in patients with knee osteoarthritis Australas J Ageing 2011 30 231 233 10.1111/j.1741-6612.2010.00500.x 22176570 93. Bhattacharya I Saxena R Gupta V Efficacy of vitamin E in knee osteoarthritis management of North Indian geriatric population Ther Adv Musculoskelet Dis 2012 4 11 19 10.1177/1759720X11424458 22870491 94. Viviano KR VanderWielen B Effect of N-Acetylcysteine Supplementation on Intracellular Glutathione, Urine Isoprostanes, Clinical Score, and Survival in Hospitalized Ill Dogs J Vet Intern Med 2013 27 250 258 10.1111/jvim.12048 23458734 95. de Godoy MR Conway CE Mcleod KR Harmon DL Influence of feeding a fish oil-containing diet to young, lean, adult dogs: effects on lipid metabolites, postprandial glycaemia and body weight Arch Anim Nutr 2015 69 499 514 10.1080/1745039X.2015.1100866 26490201 96. Sanders TA Hinds A The influence of a fish oil high in docosahexaenoic acid on plasma lipoprotein and vitamin E concentrations and haemostatic function in healthy male volunteers Br J Nutr 1992 68 163 173 10.1079/BJN19920074 1390601 97. Kooshki A Taleban FA Tabibi H Hedayati M Effects of omega-3 fatty acids on serum lipids, lipoprotein (a), and hematologic factors in hemodialysis patients Ren Fail 2011 33 892 898 10.3109/0886022X.2011.605536 21859401 98. Gharekhani A Khatami M-R Dashti-Khavidaki S Razeghi E Abdollahi A Hashemi-Nazari S-S Mansournia M-A Potential effects of omega-3 fatty acids on anemia and inflammatory markers in maintenance hemodialysis patients DARU J Pharm Sci 2014 22 11 10.1186/2008-2231-22-11 99. Payan DG Wong MY Chernov-Rogan T Valone FH Pickett WC Blake VA Gold WM Goetzl EJ Alterations in human leukocyte function induced by ingestion of eicosapentaenoic acid J Clin Immunol 1986 6 402 410 10.1007/BF00915380 3021806 100. Miles EA Calder PC Influence of marine n-3 polyunsaturated fatty acids on immune function and a systematic review of their effects on clinical outcomes in rheumatoid arthritis Br J Nutr 2012 107 Suppl 2 S171 184 10.1017/S0007114512001560 22591891 101. Lascelles BDX King S Roe S Marcellin-Little DJ Jones S Expression and activity of COX-1 and 2 and 5-LOX in joint tissues from dogs with naturally occurring coxofemoral joint osteoarthritis J Orthop Res 2009 27 1204 1208 10.1002/jor.20864 19274754 102. Bell KF Insight into a neuron’s preferential susceptibility to oxidative stress Biochem Soc Trans 2013 41 6 1541 5 10.1042/BST20130245 24256251 103. Hassler SN Johnson KM Hulsebosch CE Reactive oxygen species and lipid peroxidation inhibitors reduce mechanical sensitivity in a chronic neuropathic pain model of spinal cord injury in rats J Neurochem 2014 131 4 413 7 10.1111/jnc.12830 25051888 104. Gwak YS Hassler SN Hulsebosch CE Reactive oxygen species contribute to neuropathic pain and locomotor dysfunction via activation of CamKII in remote segments following spinal cord contusion injury in rats Pain 2013 154 9 1699 708 10.1016/j.pain.2013.05.018 23707296 105. Nicolson GL Ash ME Lipid Replacement Therapy: A natural medicine approach to replacing damaged lipids in cellular membranes and organelles and restoring function Biochim Biophys Acta 1838 2013 1657 1679 106. Scicchitano P Cameli M Maiello M Modesti PA Muiesan ML Novo S Palmiero P Saba PS Pedrinelli R Ciccone MM Nutraceuticals and dyslipidaemia: beyond the common therapeutics J Funct Foods 2014 6 11 32 10.1016/j.jff.2013.12.006

PMC005xxxxxx/PMC5002172.txt

==== Front BMC Pregnancy ChildbirthBMC Pregnancy ChildbirthBMC Pregnancy and Childbirth1471-2393BioMed Central London 100010.1186/s12884-016-1000-2Research ArticleUterine salvage management for atonic postpartum hemorrhage using “modified lynch suture” El-Sokkary M. mohammedelsokkary1@yahoo.com Wahba K. Karimwahbaobgyn@yahoo.com El-Shahawy Y. ayman_gamal007@yahoo.com Dept. of Obstetrics and Gynecology (OBGY), Ain Shams University Maternity Hospital, Cairo, Egypt 27 8 2016 27 8 2016 2016 16 1 25121 11 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 To assess the effectiveness of the new modified technique in order to control bleeding in women presenting with atonic, flabby uterus compared to the most commonly described technique of classic B-Lynch suture. Method This study included 160 women of uncontrolled atonic postpartum hemorrhage delivered by cesarean section at Ain Shams University Maternity Hospital between January 2013 and October 2015. Participants were randomly assigned following simple randomization procedures (computerized random numbers) and divided into two groups. Group, I (80 patients) operated upon by the modified (new technique) stitch while group II (80 patients) operated upon by the classic technique. The ultimate goals were to stop blood loss after placement of the sutures and avoid life-saving hysterectomy thus preserving the life and fertility of the patient. Results The modified new technique was done in 80 patients with atonic postpartum hemorrhage and it was found to be superior to the classic technique with a success rate 95 % (4 cases needed hysterectomy as a lifesaving measure) compared to 85 % with the classic technique (in 12 cases, a life-saving hysterectomy was done). Conclusions This technique can replace the classic B-lynch in flabby unresponsive atonic uteri as it has 8 shaped placement of the stitch which causes more firm compression on the uterus and simultaneous bilateral uterine artery ligation. This technique was proved valuable and successful in many patients who suffer from uncontrolled massive postpartum hemorrhage (PPH). Keywords Postpartum hemorrhageAtonyB-LynchUterine salvagesissue-copyright-statement© The Author(s) 2016 ==== Body Background Primary postpartum hemorrhage (PPH) refers to the bleeding from the genital tract that is more than 500 ml after vaginal delivery or more than 1000 ml following cesarean section during the first 24 h following delivery of the fetus [1]. However, intraoperative blood loss estimation is inaccurate in most cases; therefore, the American College of Obstetrics and Gynecology has defined it as a drop of hematocrit value of more than 10 % from the predelivery value [2]. PPH accounts for most cases of maternal morbidity and mortality in developing countries [3]. The causes of PPH include trauma, retained placenta, abnormal coagulation that may be congenital or acquired, and atonic uterus, which is one of the preventable causes of primary PPH and accounts for more than 80 % of cases of primary PPH [4]. In March 1997, Lynch [5] published his brace suture for controlling PPH when other primary measures failed. The procedure was simple and effective with the primary goal to compress the uterus without occluding the uterine arteries or uterine cavity [5]. Since this publication, more than 10 variants of uterine compressing sutured have been reported [6–10]. If this method fails, the next step will be vascular ligation (uterine, ovarian and hypogastric) or hysterectomy as a last resort [5]. In the current literature, a new modified technique of classic B-Lynch has been reported that has more effective compression on the uterus with added hemostatic effect of uterine artery ligation. Method This prospective clinical trial was performed in the Department of Obstetrics and Gynecology, Ain Shams University Maternity Hospital, Cairo, Egypt, between January 2013 and October 2015. A total of 160 women who suffered from PPH during lower segment cesarean section (LSCS) and were not responding to uterotonics were operated upon by the modified technique or the classic one after counseling of the patient and a written consent given by the patient. Patient groups 160 women with PPH refractory to uterotonics were divided into:Group I (modified technique): 80 patients were operated upon by the new technique Group II (classic technique): 80 cases were operated upon by the classical technique of B-Lynch. Patient selection Patients under general anesthesia with atonic postpartum hemorrhage refractory to uterine massage, ecbolics (oxytocin 30 units IM, methergine 0.25 mg, misoprostol 1000 μg rectal) and bimanual compression. If these measures failed, the modified or classic technique was done randomly. 5 min later, the patient is observed abdominally and vaginally for bleeding. If bleeding had stopped, closure of the abdomen with intraperitoneal drain is done, but if bleeding still continued, this represents a failure of the compression sutures thus, internal iliac artery ligation or hysterectomy will be done. Exclusion criteria Patients with traumatic PPH, DIC, bleeding diathesis, retained parts of the placenta or cases with uterine anomalies. Outcome Successful: if the bleeding stopped and no need for hysterectomy Unsuccessful: if the bleeding continued and there is a need for hysterectomy Randomization For allocation of the participants, a computer-generated list of random numbers was used and was kept in Ain Shams Maternity Hospital computer and with research supervisors. Participants were randomly assigned for following simple randomization procedures (computerized random numbers) to 2 treatment groups modified Lynch and classical Lynch groups. Group assignments were allocated according to a computer-generated randomized series, were kept in sealed envelopes. Postoperative results Vital data, urine output, follow up of any vaginal bleeding, the output of the intraperitoneal drains, hemoglobin and hematocrit concentration. Blood transfusion was given according to patient preoperative hemoglobin and amount of blood loss intraoperative. In stable cases, observations were made for 48 h, then discharging the patient with discharge card, including all the operative details to be rechecked after one week for any abnormality (wound gaping, deep venous thrombosis, puerperal sepsis, uterine wall necrosis, vesicovaginal fistula). Steps of the new modified Lynch technique (See fig. 1) Exteriorize the uterus, removal of sutures of lower segment cesarean section. The assistant stretched up the uterus, the 1st stitch is placed 2 cm below the lower segment incision and 2 cm medial to the lateral angle to come on the same side on the upper flap then cross on the fundus to the contralateral side above the uterosacral then to the other uterosacral then to the contralateral flap in a figure of eight fashion then after tightening of this suture the needle is passed through avascular area in the broad ligament to the back while the tape is passed through a window on the opposite side made by an artery forceps to become on the posterior aspect of the uterus. Tightening of the transverse suture is done.Fig. 1 Steps of the new modified Lynch technique Advantages of the modified technique over the classical technique The 8 shaped ligatures appear to be more hemostatic and compressing the uterus. Also, the transverse limb ligates the uterine artery with more compression on the lower segment, so it is more effective in case of placenta previa or bleeding from the lower segment in general. Results A total of 160 women was recruited in the current study. The clinic-epidemiological data of women under the study were analyzed, in group I, the mean age was 29.6 ± 4.5 years, the mean parity was 1.55 ± 1.35, the mean weight was 76 ± 13.13151 kg, the gestational age was 39.1 ± 1.1 weeks and the neonatal birth weight was 3.49 ± 0.365 kg compared to group II in whom the mean age was 29.3 ± 5.1 years, the mean parity was 1.6 ± 1.21, the mean weight was 79 ± 12.325 kg, the gestational age was 38.7 ± 1.8 weeks and the neonatal birth weight was 3.49 ± 0.365 kg with no significant difference between the two groups (P < 0.001). The modified new technique was done in 80 patients with atonic postpartum hemorrhage refractory to usual measures and it was found to be superior to the classic technique with a success rate 95 % (4 cases needed hysterectomy as a lifesaving measure) compared to 85 % with the classic technique (in 12 cases, a life-saving hysterectomy was done) as shown in Table 1.Table 1 Shows the clinic-epidemiological, operative and postoperative data of women under the study Parameter G1 G2 P-value Age (year) Mean ± SD 29.6 ± 4.5 29.3 ± 5.1 0.694 Parity 1.55 ± 1.35 1.61 ± 1.21 0.431 Weight 76 ± 13.151 79 ± 12.325 0.139 Gestational age 39.1 ± 1.1 38.7 ± 1.8 0.092 Neonatal birth weight 3.496 ± 0.365 3.397 ± 0.389 0.099 Postoperative blood loss 324 ± 105 568 ± 209 0.0001 (HS) Need for blood transfusion 2.8 ± 0.5 4.2 ± 0.8 0.0001 (HS) Need for hysterectomy 4 12 0.0001 (HS) Hospital stay 2.8 ± 1.5 2.7 ± 1.6 0.684 Duration of procedure 6 ± 0.95 3 ± 1.3 0.0001 (HS) Pre-operative Hb% 9.8 ± 1.2 10.1 ± 1.3 0.131 Post-operative Hb% 8.7 ± 1.1 8.9 ± 1.8 0.398 Bleeding from multiple bites 7 13 0.152 Hematometra 1 5 0.212 Wound hematoma 4 4 1.00 Wound infection 4 4 1.00 Fever 5 8 0.563 Table (1) shows that there was no significant difference between the two groups as regards the clinic-epidemiological characters of the study groups and the postoperative data, but the modified new technique was found to be superior to the classic technique with a success rate 95 % compared to 85 % with the classic technique The two groups were compared as regards the complications of the conservative operative intervention, as regards the bleeding from multiple bites; it was clear that it was lower in group I when compared to group II which appears due to the hemostatic added effect of uterine artery ligation but the difference was not statistically significant, the same with the other postoperative parameters as development of hematometra, wound infection or post-operative fever, there was no significant difference between the two groups (Table 1). Discussion Postpartum hemorrhage is a potentially life-threatening complication of fetal delivery [11]. It may occur after vaginal delivery (4 %) or caesarean births (6 %) [12, 13]. The most important step in the management of PPH is to identify and correct the underlying cause [14]. Most of the cases of PPH can be controlled by traditional treatment modalities like uterotonics, uterine massage, bimanual compression and balloon tamponade [14]. Uncontrolled PPH is usually managed by different uterine suture techniques (modified B-Lynch, and square suture) or with stepwise devascularization surgical procedures. These techniques have reported variable outcomes and many of the patients finally require emergency hysterectomy [5, 7, 14]. A review of peripartum or cesarean hysterectomy reported an average mortality rate of 4.8 % [15]. This does not mean that hysterectomy caused this high maternal mortality, but critically ill situations requiring this surgery may eventually cause it. Indeed, cesarean hysterectomy is one of the most difficult obstetric surgeries and is always challenging [16]. The surgical method of controlling uterine bleeding by inserting B-Lynch suture has been developed to reduce the incidence of emergency hysterectomy and to preserve fertility in these patients. Because of simplicity of application and less time taken to put the modified B-Lynch stitch, it should be the preferred choice [6]. Various parameters in the current study are compared as follows In Khatoon et al. study [17], B-Lynch stitch was applied in 9 cases, i.e. 60 % after vaginal delivery and on 6 cases i.e. 40 % during cesarean section. In the current study the new technique of the B-Lynch stitch was taken on 160 cases during cesarean section, but actually, in our institute the new technique is now being performed on cases also after vaginal delivery and on the closed uterus to avoid the additional bleeding on incision of the uterus but this is still under trial. In a prospective study of Hackenthal et al. [18], Hb difference is 3gm% after using the modified B-Lynch technique and in the current study, the mean intraoperative blood loss was 568 ± 209 ml after classical technique and 324 ± 105 ml after the new technique with a highly significant difference between the two groups. After the blood transfusion according to the clinical condition, there was no significant difference between the two groups as regards the postoperative hemoglobin. In Koh et al. study [19], 4 patients required more than 3 units of blood transfusion and 2 patients did not require any blood transfusion while in the current study the mean units of blood required was 4.2 ± 0.8 in the group who was operated upon by the classical technique and was 2.8 ± 0.5 in the group who was operated upon by the new technique with highly significant difference between the two groups, reflecting the effectiveness of new modified B-Lynch stitch in the control of atonic postpartum hemorrhage. In a study of Anamika et al. [20], time was taken to put stitch was 11 to 20 min in 35 patients and less than 10 min in 3 patients, more than 20 min in 5 patients. In Our study, time taken to put stitch of the classic technique was 3 ± 1.3 while the time taken to put stitches of the new technique was 6 ± 0.95 with highly significant difference between the two groups, but this factor may be changed with more training and familiarity with the new technique and must be kept in mind the effectiveness of the new method. Study conducted by Hackenthal et al. [18] and Anamika et al. [20] had a success rate of 100 %, thus proving that this technique was highly effective and 1 patient died on the 21st postpartum day due to Acute Respiratory Distress Syndrome and Septicaemia In Our study success rate was 85 % with the classic technique and 95 % with the new technique with highly significant difference between the two groups with no mortalities.. In a prospective study conducted by Ghodake et al. [21], 31 patients underwent B-Lynch stitch, out of which 5 patients had a post-operative fever, 3 patients had surgical wound gaping. In our study, 8 patients had a post-operative pyrexia in the group operated upon by the classic technique and 5 patients in the group operated upon by the new technique with a highly significant difference between the two groups. But as regards wound infection and gaping there was no significant difference between the two groups. In Our study, there were no major complications. Conclusion The new technique of the B-Lynch is highly effective in controlling an atonic postpartum hemorrhage so we suggest strongly this technique as an alternative safe option to stop an atonic postpartum hemorrhage. There was no adverse effect on the fertility potential for the observed 2 years; however, a long-term follow-up is required to comment on its actual rate. All pertinent study information was explained to them and they were informed that rejection or withdrawal from the study will not affect any medical service provided. A summarized study information sheet was shown to all cases before obtaining their verbal agreement. Finally, an informed verbal consent was obtained and witnessed by the attending nurse. A log book was created including the participant’s study number and the date of consent. The IRB waived the requirement for taking a written consent as the research had minimal risk of harm to subjects and involved no risky procedures for which written consent is required. Acknowledgements None. Funding None declared. Availability of data and materials Data are available from Ain Shams University Maternity Hospital, for researchers who meet the criteria for access to confidential data. Data is restricted to protect participant privacy. Any interested researcher can obtain a minimal data-set after contacting one of the main authors. Assistant professor Mohammed Elsokkary can be contacted at Mohammedelsokkary1@yahoo.com. Authors’ contributions Assistant professor ME conceived the idea of the study, collected data, drafted the manuscript and gave the final approval for the manuscript to be published. Assistant professor KW participated in data analysis and interpretation and drafted the manuscript. Lecturer YE designed the protocol, carried out the statistical analysis and shared in writing the manuscript draft. All authors reviewed and approved the final manuscript. All authors read and approved the final manuscript. Competing interests The authors have declared that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of Ain Shms Mater University. ==== Refs References 1. Arulkumaran S Tanizian O the management of postpartum hemorrhage Arulkumaran’s: the management of labor 2006 2 208 228 2. Roman AS Rebarber A Seven ways to control postpartum hemorrhage Contemp Ob/Gyn 2003 3 34 53 3. World Health Organization Maternal Mortality: A Global Factbook 1991 Geneva WHO 4. Condous GS Arulkumaran S Obstetric Hemorrhage Williams Obstetrics 2005 22 New York McGraw-Hill Medical Publication division 823 824 5. B-Lynch C Coker A Lawal AH Abu J Cowen MJ The B-Lynch surgical technique for the control of massive postpartum hemorrhage: an alternative to hysterectomy? Five cases reported Br J ObstetGynaecol 1997 104 3 372 375 10.1111/j.1471-0528.1997.tb11471.x 6. Mallappa Saroja CS Nankani A El-Hamamy E Uterine compression sutures, an update: review of efficacy, safety and complications of B-Lynch suture and other uterine compression techniques for postpartum hemorrhage Arch. Gynecol. Obstet. 2009 281 4 581 588 10.1007/s00404-009-1249-z 19834719 7. Hwu YM Chen CP Chen HS Su TH Parallel vertical compression sutures: a technique to control bleeding from placenta praevia or accreta during caesarean section BJOG 2005 112 10 1420 1423 10.1111/j.1471-0528.2005.00666.x 16167948 8. Dawlatly B Wong I Khan K Agnihotri S Using the cervix to stop bleeding in a woman with placenta accreta: a case report BJOG 2007 114 4 502 504 10.1111/j.1471-0528.2006.01189.x 17378822 9. Schnarwyler B Passweg D Von Castelrg B Successful treatment of drug refractory uterine atony by fundus compression sutures [in German] Geburt shelf Frauenheilkd 1996 56 3 151 153 10.1055/s-2007-1022282 10. Cho JH Jun HS Lee CN Hemostatic suturing technique for uterine bleeding during cesarean delivery Obstet. Gynecol. 2000 96 1 129 131 10928901 11. Lewis G the Confidential Enquiry into Maternal and ChildHealth (CEMACH) Saving Mothers’ Lives: Reviewing Maternal Deaths to Make Motherhood Safer—2003–2005 The Seventh Report on the Confidential Enquiries into Maternal Deaths in the United Kingdom 2007 London The Confidential Enquiry into Maternal and Child Health (CEMACH) 12. Combs CA Murphy EL Laros RK Jr Factors associated with postpartum hemorrhage with vaginal birth Obstet. Gynecol. 1991 77 1 69 76 1984230 13. Prasertcharoensuk W Swadpanich U Lumbiganon P Accuracy of the blood loss estimation in the third stage of labor Int. J. Gynecol. Obstet. 2000 71 1 69 70 10.1016/S0020-7292(00)00294-0 14. Weisbrod AB Sheppard FR Chernofsky MR Emergent management of postpartum hemorrhage for the general and acute care surgeon World Journal of Emergency Surgery 2009 4 1 43 10.1186/1749-7922-4-43 19939251 15. Machado LS Emergency peripartum hysterectomy: Incidence indications, risk factors and outcome N. Am. J. Med. Sci. 2011 3 358 361 10.4297/najms.2011.358 22171242 16. Matsubara S Kuwata T Usui R Watanabe T Izumi A Important surgical measures and techniques at cesarean hysterectomy for placenta previa accreta Acta Obstet. Gynecol. Scand. 2013 92 372 377 10.1111/aogs.12074 23323568 17. Khatoon A Syeda Fariha H Junaid A B-Lynch brace suture for the treatment of major primary postpartum hemorrhage: An experience at Abbasi Shahi Hospital Karachi Mc 2011 17 3 36 38 18. Hackethal A Bruggmann D Oehmke F Tinneberg HR Uterine compression sutures in primary PPH after Caesarean section: Fertility Preservation with a simple and effective technique Human Reprod 2008 23 1 74 79 10.1093/humrep/dem364 19. Koh E Devendra K Tan L B-Lynch suture for the treatment of uterine atony Singapore Meds 2009 50 7 693 697 20. Anamika Majumdar A Mallick K Vasava B Desai KT Dalal M A descriptive study on Hayman suture technique to control postpartum hemorrhage: Sri Lanka J Obstet Gynecol 2012 34 79 83 21. Ghodake VB Pandit SN Umbardand SM Role of modified B-Lynch suture in modern day management of atonic PPH Bombay Hospital Journal 2008 50 2 205 210

PMC005xxxxxx/PMC5002173.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 131210.1186/s12906-016-1312-9Research ArticleA preliminary randomised controlled study of short-term Antrodia cinnamomea treatment combined with chemotherapy for patients with advanced cancer Tsai Ming-Yen +886-7-731 7123-2332missuriae0116@gmail.com 124Hung Yu-Chiang 13Chen Yen-Hao 4Chen Yung-Hsiang 2Huang Yu-Chuen 5Kao Chao-Wei 1Su Yu-Li 4Chiu Hsien-Hsueh Elley 3Rau Kun-Ming +886-7-731 7123-2332kmrau58@cgmh.org.tw 41 Department of Chinese Medicine, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, 83301 Taiwan 2 Graduate Institute of Integrated Medicine, College of Chinese Medicine, China Medical University, Taichung, 40402 Taiwan 3 School of Chinese Medicine for Post Baccalaureate, I-Shou University, Kaohsiung, 83301 Taiwan 4 Department of Internal Medicine, Division of Oncology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, 83301 Taiwan 5 Department of Medical Research, China Medical University Hospital and School of Chinese Medicine, China Medical University, Taichung, 40402 Taiwan 26 8 2016 26 8 2016 2016 16 1 32211 11 2015 23 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 Antrodia cinnamomea (AC) is a popular medicinal mushroom in Taiwan that has been widely used for treatment of various cancers. Few clinical studies have reported its application and efficiency in therapeutic chemotherapy strategies. We performed a double-blind, randomized clinical study to investigate whether AC given for 30 days had acceptable safety and efficacy in advanced cancer patients receiving chemotherapy. Methods Patients with advanced and/or metastatic adenocarcinoma, performance status (PS) 0–2, and adequate organ function who had previously been treated with standard chemotherapy were randomly assigned to receive routine chemotherapy regimens with AC (20 ml twice daily) orally for 30 days or placebo. The primary endpoint was 6-month overall survival (OS); the secondary endpoints were disease control rate (DCR), quality of life (QoL), adverse event (AE), and biochemical features within 30 days of treatment. Results From August 2010 to July 2012, 37 subjects with gastric, lung, liver, breast, and colorectal cancer (17 in the AC group, 20 in the placebo group) were enrolled in the study. Disease progression was the primary cause of death in 4 (33.3 %) AC and 8 (66.7 %) placebo recipients. Mean OSs were 5.4 months for the AC group and 5.0 months for the placebo group (p = 0.340), and the DCRs were 41.2 and 55 %, respectively (p = 0.33). Most hematologic, liver, or kidney functions did not differ significantly between the two groups, but platelet counts were lower in the AC group than in the placebo group (p = 0.02). QoL assessments were similar in the two groups, except that the AC group showed significant improvements in quality of sleep (p = 0.04). Conclusions Although we found a lower mortality rate and longer mean OS in the AC group than in the control group, A. cinnamomea combined with chemotherapy was not shown to improve the outcome of advanced cancer patients, possibly due to the small sample size. In fact, the combination may present a potential risk of lowered platelet counts. Adequately powered clinical trials will be necessary to address this question. Trial registration ClinicalTrials.gov NCT01287286. Keywords Antrodia cinnamomeaChemotherapyAdvanced and metastatic cancerOutcomeissue-copyright-statement© The Author(s) 2016 ==== Body Background Cancer is a serious health problem in Asia, and it is now the leading cause of death in Asian Pacific countries such as China, Japan, Korea, and Taiwan. The most commonly diagnosed cancers in Eastern Asia, in descending order, are lung, breast, colon, and rectum, and the leading causes of cancer deaths are lung, stomach, and liver cancer [1]. Garcia et al. reported that in the year 2007, an estimated 3.3 million people were diagnosed with cancer and 2.3 million people died from cancer in Eastern Asia. Their report also predicted that by 2020, an estimated 7.1 million new cases of cancer will be diagnosed each year if existing prevention and management strategies remain unchanged [2]. The major conventional therapies for cancer are surgery, chemotherapy, and radiotherapy. However, these therapies have numerous limitations: (1) Most cancer patients are diagnosed too late to undergo surgery, despite advances in biomarker and radiographic research; (2) although chemotherapy and radiotherapy are effective against cancer, they also have serious side effects and complications (e.g., fatigue, pain, diarrhea, nausea, vomiting, and skin reactions); (3) most advanced cancers have a survival rate of less than 5 years, and recurrence or metastasis is quite common even after surgical resection; and (4) since some cancers are relatively chemo- or radio-resistant, systemic cytotoxic chemotherapy and radiotherapy are minimally effective at improving patient survival [3, 4]. For these reasons, and in spite of the uncertainty of its benefits due the lack of well-controlled scientific studies, cancer patients have increasingly turned to complementary and alternative medicine (CAM) for treatment [5]. A recent study showed that in Taiwan, up to 50.4 % of adults and 84.6 % of females use some form of CAM to treat cancer [6]. Antrodia cinnamomea (AC), a well-known medical mushroom used in Taiwan, has several pharmacological functions, including antioxidant, anti-itching and hepatoprotective effects, as demonstrated by experimental studies [7, 8]. AC extract contains a complex mixture of bioactive ingredients, including triterpenoids, steroids, polysaccharides, and phenyl and biphenyl compounds [9]. It has recently become popular as a potential complementary and alternative therapeutic agent for the treatment of various cancers [10]. Previous studies revealed that maleimide derivative isolated from AC attenuates the migration and invasion of breast cancer cells [11]. Investigations by Hsu et al. have shown that AC fruiting body extract exhibits a significant cytotoxic effect against hepatoma cell line Hep G2 and PLC/PRF/5 cells [12]. Furthermore, AC extract has adjuvant antiproliferative effects on hepatoma cells in vitro and in vivo when combined with anti-tumor agents through the inhibition of Multi-Drug Resistance (MDR) gene expressions and the pathway of COX-2-dependent inhibition of phospho-AKT (p-AKT) [13]. However, the clinical effects of AC on cancer have not yet been demonstrated in human trials. To increase understanding of the effects of A. cinnamomea on cancer patients, we designed the preliminary study and worked with oncologists and a manufacturer to settle some of the controversy surrounding AC treatment. In addition, we also evaluated the safety and the alleviation of symptoms as a result of treatment with AC in combination with chemotherapy for patients in advanced stages. Methods Study design and patient selection This prospective randomized, double-blind study was undertaken in the Medical Oncology Department of Kaohsiung Chang Gung Medical Hospital (KCGMH) between August 2010 and July 2012. The protocol was approved by the KCGMH Ethics Committee (IRB-98-3904A3) and registered at clinicaltrials.gov under the identification code NCT01287286. All participants provided written informed consent to participate in the study. Patients aged ≥18 years were eligible if they had advanced or recurrent, untreated, stage III-IV adenocarcinomas that were histologically or cytologically confirmed; had previously received standard chemotherapy regimens; and had an Eastern Co-operative Oncology Group (ECOG) performance status (PS) of 0–2. Patients were excluded if they were pregnant or lactating; had used any other investigational agent within the 28 days prior to registration; had a severe current illness; had an abnormal blood cell count, liver function test, or creatinine clearance; or had known hypersensitivity to any formulas of AC in the market. Treatment Eligible patients were randomly assigned to either the AC group or the placebo group. The participants were given 20-ml oral formulations twice a day for 30 days. The A. cinnamomea used in the study comprised mainly 2100 mg polysaccharides, 172 mg triterpenoids, and 2687.5 mg γ-aminobutyric acid, and its Taiwan Symbol of National Quality (SNQ) approval number was A00851. The aqueous extract from mycelium of AC displayed cytotoxic activity with an IC50 value > 400 μg/ml in murine RAW 264.7 cells. In addition, the freeze-dried state of AC mycelia was administered for 28 days and 90 days respectively to evaluate the oral toxicity in Sprague-Dawley (SD) rats [14, 15]. Both studies showed the no-observed-adverse-effect-level (NOAEL) and are safe to use as dietary supplements or nutraceuticals. Strains of AC were provided by the Research and Development Center (New Bellus Enterprises Co., Ltd, Tainan, Taiwan). Extracts were prepared as previously described by Cheng et al [16]. Briefly, AC was first cultured with Malt Extract Agar (MEA) I broth containing 1.5 % agar, 1.5 % malt extract, 2 % glucose, and 0.1 % peptone at 30 °C for 7–10 days. AC mycelia were inoculated into MEA I broth and incubated at 30 °C with shaking for 8–10 days. After sterilization at 121 °C for 30 min, fermented extracts were centrifuged at 3000 g for 10 min and then permeated through a 3000 MW ultrafiltration membrane. The placebo was made of malt, sucrose, and xanthan gum to achieve comparable color, outlook, and taste to the AC. Both the AC and placebo solutions were produced by Amon Biotech Co., Ltd (Lot no. AC9905245TA, Expiry Date/ EXP.: June 2010/ June 2013). Quality control was ensured. The AC and placebo solutions were packed in sealed opaque aluminum packages; only the treatment codes and administration instructions were printed on the outsides of the package to ensure successful blinding of the patients. Eligible patients with lung cancer were scheduled to receive platinum-based chemotherapy with cisplatin dosed at 60 − 75 mg/m2 or a carboplatin area under the curve (AUC5). The following regimens were sequentially adopted for this study at the discretion of the investigator in accordance with the National Comprehensive Cancer Network (NCCN) guidelines for patients with breast, stomach, liver, and colorectal cancers: adriamycin-based, taxane- or vinorelbine-based, 5-fluorouracil (5-FU)-based and oxaliplatin-based chemotherapy (Table 1) [17]. After the first dose of standard chemotherapy, the patients began treatment with a 30-day supply of the agent and were asked to return the containers at the next scheduled visit (on days 7, 15, and 30). Counting the unused boxes served as a measure for compliance with treatment. Patients were withdrawn from the trial if they failed to appear for the next two scheduled visits or if they had not used at least 80 % of the supplied treatment at 75 % of the treatment intervals.Table 1 Chemotherapeutic protocols for different advanced cancer types according to NCCN guidelines and their adverse events Cancer type Chemotherapeutic regimens Common adverse events Lung Platinum-based Nausea, vomiting, nephrotoxicity, bone marrow suppression Breast Adriamycin-based or taxane-based Bone marrow suppression, nausea, vomiting, alopecia Stomach Oxaliplatin/5-fluouracil Nausea, vomiting, mucositis, diarrhea, bone marrow suppression Liver Adriamycin and Cisplatin Nausea, vomiting, bone marrow suppression, nephrotoxicity Colorectal Oxaliplatin-based or Irinotecan-based Nausea, vomiting, diarrhea, bone marrow suppression Efficacy and safety assessment The primary endpoint of this analysis was overall survival (OS). OS was calculated from the date of randomization until the date of death or the last 6-month follow-up. The secondary endpoints were the disease control rate (DCR) (complete response (CR), partial response (PR), and stable disease (SD)), as determined by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Tumor assessments using computed tomography scans were performed before A. cinnamomea treatment and 30 days thereafter. Other secondary outcome measures included Quality of life (QoL), adverse event (AE), and clinical and laboratory assessments conducted at baseline and at 30 days after administration. QoL was measured using the European Organization for Research and Treatment-Quality of Life Questionnaire (EORTC-QLQ-C30) and the Multidimensional Fatigue Inventory (MFI-20). The incidence and severity of AEs were graded according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE) version 3.0. Statistical analyses The baseline characteristics and clinical data of the patients treated with chemotherapy plus AC or placebo are presented as number and mean ± standard deviation (SD). Overall survival (OS) for both groups was measured from randomization until death from any cause. Patients that were alive but lost to follow-up were excluded at the last follow-up date. OS was illustrated by means of Kaplan-Meier curves, and log-rank tests were used to compare survival between the two groups. The Chi-square test was used to compare the treatment response in the two groups, separated as DCR vs. progressive disease (PD). The data on the QoL, clinical, and laboratory assessments for the pre- and post-tests within the AC group and the placebo group were assessed by Wilcoxon signed-rank test. Differences in efficacy between the two groups were determined by Mann-Whitney U-test. All statistical analyses were performed with SPSS version 18 for Windows. A p value < 0.05 was considered statistically significant. Result Between August 2010 and July 2012, a total of 37 patients were recruited into the study and considered for the full analysis (Fig. 1). Of the 37, 17 received chemotherapy with A. cinnamomea, and 20, chemotherapy with placebo. Patient demographics and disease characteristics are presented in Table 2.Fig. 1 Flow chart Table 2 Baseline characteristics of patients treated with chemotherapy plus A. cinnamomea (AC) or placebo Chemotherapy + AC (n = 17) Chemotherapy + placebo (n = 20) Sex (male/female) 10/7 8/12 Age (Mean ± SD) 56.71 ± 9.61 53.10 ± 10.86 Body weight in kg (Mean ± SD) 61.98 ± 9.09 60.36 ± 11.11 Months of prior treatment (Mean ± SD) 9.80 ± 3.46 8.55 ± 4.03 Tumor location (TNM stage, AJCC, 2002)  Lung (stage IV) 1 7  Breast (stage IV) 4 4  Stomach (stage III-IV) 5 5  Liver (stage III-IV) 3 1  Colorectal (stage IV) 4 3 At the 30-day analysis, patients had experienced 2.35 and 1.75 chemotherapy cycles schedule while on AC and placebo treatment, respectively. Four cases of disease progression (23.5 %) had occurred in the AC group and 4 (20 %) in the placebo group, and no deaths or withdrawals were recorded. By the cut-off date at 6 months of treatment, a total of 12 patients had died (32.4 %), 25 patients were surviving (67.6 %), and 4 patients in the placebo group were lost to follow-up. Disease progression was the primary cause of death in 4 (33.3 %) AC and 8 (66.7 %) placebo recipients. There was no statistically significant difference in survival between the treatment groups; mean OSs were 5.4 months (AC group) and 5.0 months (placebo group; p = 0.340; Fig. 2). As shown in Table 3, 29 of the 37 patients (13 patients in the AC group and 16 patients in the placebo group) had measurable lesions and could be evaluated for response based on the RECIST criteria. No complete response was observed in either treatment group. Three patients in the AC group achieved PR, four patients were SD, and the remaining six patients were PD. In the placebo group, three patients achieved PR, eight patients were SD, and the remaining five patients were PD. The DCRs were 41.2 and 55 % in the AC and placebo groups, respectively. There were no significant differences in the objective DCR vs. PD in the two groups (p = 0.33).Fig. 2 Kaplan-Meier curves of 6-month survival for chemotherapy plus A. cinnamomea (AC) or placebo Table 3 Best overall response and disease control rate per RECIST (Full Analysis Population) Chemotherapy + AC (n = 17) Chemotherapy + placebo (n = 20) p-value Best overall response, No. (%)a Complete response (CR) 0 0 Partial response (PR) 3 (17.6) 3 (15) Stable disease (SD) 4 (23.5) 8 (40) Progressive disease (PD) 6 (35.3) 5 (25) Unknown 4 (23.5) 4 (20) Disease control rate (CR + PR + SD) 7 (41.2) 11 (55) 0.33 aProportion of patients whose best overall response is complete response, partial response, or stable disease according to RECIST (Response Evaluation Criteria in Solid Tumors) All patients completed the two quality of life forms (Table 4). Most of the EORTC-QLQ-C30 assessments did not show significant differences between the treatment groups with regard to the mean scores from baseline to 30 days, but quality of sleep improved significantly more in the AC group than in the placebo group (p = 0.04). There was no difference in fatigue levels on the general subscale of the MFI-20 between the two groups.Table 4 Quality of life of cancer patients treated with chemotherapy plus A. cinnamomea (AC) or placebo Chemotherapy + AC (n = 17) p-valuea Chemotherapy + placebo (n = 20) p-valuea p-valueb Pre-test Post-test Pre-test Post-test EORTCQLQ-C30  General condition 49.51 ± 11.20 55.55 ± 28.63 0.08 53.75 ± 26.27 53.07 ± 23.10 0.28 0.76  Fatigue 26.79 ± 27.23 34.72 ± 36.03 0.18 28.33 ± 18.19 41.52 ± 28.16 0.06 0.32  Nausea and vomiting 22.54 ± 34.32 10 ± 18.68 0.08 9.16 ± 18.31 16.66 ± 26.64 0.24 0.51  Pain 14.70 ± 21.95 25.55 ± 30.12 0.04 19.16 ± 23.74 21.92 ± 28.89 0.38 0.77  Shortness of breath 9.80 ± 19.59 11.11 ± 20.57 0.31 13.33 ± 19.94 19.29 ± 25.61 0.15 0.30  Sleep disturbance 47.05 ± 42.58 40.00 ± 36.07 0.37 46.66 ± 38.08 61.40 ± 35.59 0.01 0.04  Lack of appetite 27.45 ± 35.81 22.22 ± 29.99 0.44 18.33 ± 27.51 31.57 ± 34.19 0.03 0.44  Constipation 15.68 ± 20.80 17.77 ± 27.79 0.28 21.66 ± 31.11 8.77 ± 18.73 0.02 0.29  Diarrhea 5.88 ± 13.09 6.66 ± 13.80 0.12 1.66 ± 7.45 10.52 ± 24.97 0.21 0.86 MFI - 20  General Fatigue 10.41 ± 1.50 10.6 ± 2.26 0.35 10.55 ± 1.84 10.15 ± 2.03 0.21 0.41  Physical Fatigue 12.23 ± 0.90 12.13 ± 1.41 0.33 11.95 ± 0.82 12.25 ± 1.11 0.24 0.78  Reduced Activity 11.23 ± 1.60 11.66 ± 0.90 0.47 11.20 ± 1.54 11.15 ± 1.26 0.43 0.12  Reduced Motivation 12.76 ± 1.82 13.00 ± 1.41 0.15 12.55 ± 1.23 12.50 ± 1.39 0.33 0.37  Mental Fatigue 11.17 ± 1.07 11.33 ± 1.11 0.20 10.9 ± 1.77 11.95 ± 1.35 0.01 0.28 aWilcoxon signed-rank test (Within group) bMann-Whitney U test (Differences in efficacy between two groups) All laboratory parameters were comparable in the AC group and the placebo group as compared with baseline, and similar trends between both groups were observed throughout the study period (Table 5). After 30 days of treatment, the AC group showed a significant decrease in platelet count (p < 0.01). We also detected a significant difference in the decreases in the platelet counts in the two groups (p = 0.02). Other tested biomarkers did not demonstrate significant changes in the two groups.Table 5 Body weight, hematological and non-hematological data follow-up for patients taking chemotherapy plus A. cinnamomea (AC) or placebo Chemotherapy + AC (n = 17) p-valuea Chemotherapy + placebo (n = 20) p-valuea p-valueb Pre-test Post-test Pre-test Post-test Hematological  WBC (103/ul) 6.1 ± 2.64 11.55 ± 1.30 0.38 5.75 ± 3.23 11.47 ± 1.48 0.34 0.10  Hemoglobin (g/dl) 11.57 ± 1.17 6.17 ± 2.77 0.43 11.75 ± 2.03 5.69 ± 2.16 0.77 0.79  Platelet (103/ul) 179.11 ± 65.11 141.71 ± 63.47 <.001 201 ± 68.97 196.55 ± 74.47 0.33 0.02  Non-hematological Bilirubin (mg/dl) 0.64 ± 0.17 0.79 ± 0.78 0.48 0.49 ± 0.19 0.62 ± 0.28 0.12 0.62  AST (u/l) 37.88 ± 22.01 39.62 ± 26.38 0.33 27.75 ± 11.68 26.50 ± 9.73 0.30 0.07  ALT (u/l) 29.70 ± 17.50 28.70 ± 15.46 0.14 20.33 ± 10.28 29.89 ± 9.46 0.42 0.09  Creatinine (mg/dl) 0.82 ± 0.29 0.86 ± 0.30 0.11 0.72 ± .17 0.78 ± 0.17 0.10 0.73  Body weight (kg) 61.98 ± 9.09 61.24 ± 9.13 0.21 60.36 ± 11.11 59.45 ± 11.36 0.13 0.53 aWilcoxon signed-rank test (Within group) bMann-Whitney U test (Differences in efficacy between two groups Various treatment-related adverse events (any grade) due to the chemotherapy plus either A. cinnamomea or placebo in this study are reported in Table 6. Gastrointestinal complaints were the most common AEs reported during the study. Most of the gastrointestinal complaints were of grade 1–2 in the AC group. Additionally, one patient receiving placebo reported grade 3 diarrhea (increasing to ≥7 stools per day). Four patients suffered severe abdominal pain in the placebo group due to disease progression. The grade 3–4 AEs in other patients in both groups were related to disease progression, except that an 81 y/o patient with gastric cancer in the AC group developed upper gastrointestinal (GI) bleeding after 10 doses of AC.Table 6 Common adverse events (AEs) of 37 cancer patients treated with chemotherapy plus either A. cinnamomea or placebo Chemotherapy + AC (n = 17) Chemotherapy + placebo (n = 20) Grade 1–2, N (%) Grade 3–4, N (%) Grade 1–2, N (%) Grade 3–4, N (%) Abdominal pain 5 (29.4) 1 (5) 4 (20) Diarrhea 3 (17.6) 1 (5) Dry mouth 1 (5.9) Blurred vision 1 (5) Dyspnea 2 (11.8) Intracranial hemorrhage 1 (5.9) Upper gastrointestinal bleeding 1 (5.9) 1 (5) Infection 1 (5.9) 2 (10) Discussion This double-blind randomized clinical trial examined the impact of 30-day A. cinnamomea treatment on the efficacy and safety associated with chemotherapy in patients with advanced cancers. Despite the strong scientific rationale in the biotechnology industry [12, 18, 19], the combination of AC and standard chemotherapy regimens failed to improve the survival and response rates of the AC group relative to those of the placebo group in patients with advanced cancers. A recent preclinical study demonstrated the antitumor effect of AC on non-small cell lung cancer cells both in vitro and in vivo with bioluminescence imaging and indicated that the inhibition of tumor proliferation may be related to induced apoptosis [20]. Moreover, AC has been shown to enhance the cytotoxic effect of chemotherapy on ovarian cancer and hepatoma cells [13, 21]. These results have led to AC being used widely in certain populations and kept its price high in Taiwan. However, the sweet success of herbal or folk medicine agents in animal studies and the bitter aftertaste in clinical studies have recently been a major topic of discussion in research [22]. Known experimental factors, including the animal model, target selection, recommended dose, and differences in comorbidities, influence all treatments from cellular therapy to new drug application [23]. Despite the nonexistence of a clear anti-cancer effect found in this trial, given the analyses of OS and DCR of AC treatment, a potential finding is that AC does not affect the progression of cancer when combined with chemotherapy. Advanced or metastatic cancer is predictably associated with challenges and burdens that may lead to symptoms of fatigue, toxicities of cancer therapy, and compromised patient QoL and mortality [24, 25]. The production of reactive oxygen species (ROS) is an important contributor to these symptoms and metabolic abnormalities [26]. Although AC is used for protection against oxidative damage and elimination of fatigue [27, 28], this study did not find a significant benefit on fatigue or QoL-related outcomes in patients treated with AC. Only for quality of sleep was AC more effective than the placebo. It is known that sleep disturbances affect 30–50 % of cancer patients and contribute additional risks of depression, fatigue, heightened pain, and decrement of survival rates [29, 30]. Evidence is growing that disruptions in biological rhythmicity are also relevant to cancer. The mitotic properties of cancerous cells themselves, the treatments of cancer, the time of day of treatment administration, and possibly the quality of life of cancer patients all affect outcomes [31]. A recent study reported that poor sleep quality may be associated with increases in oxidative stress indicators and decreases in free radical scavenging anti-oxidants [32]. However, few studies to date have investigated the relationship between AC and the improvement of quality of sleep. This finding in our study, that AC may improve the quality of sleep, is worthy of further exploration, but the evidence does not support a strong recommendation. Further research is needed in order to examine the potential impacts of the integrative model of care on specific outcomes, such as chemotherapy-induced and cancer-related fatigue, as well as other QoL-related concerns. In the minds of many patients, A. cinnamomea is a natural component with high antioxidant and polysaccharide contents. The lack of toxicity reported in previous studies might indicate that is safe for use as a functional food ingredient [33, 34]. However, the adverse events observed in patients treated with AC in our study were generally consistent with its known adverse event profile. A study published in China concluded that the property and flavor of AC are cold and bitter [35]. According to Traditional Chinese Medicine (TCM) theory, the barrier function of gastric mucosa is inhibited after the intake of an excessive amount of bitter and cold medicine, possibly leading to damage to the stomach [36]. If these medicines are taken for an extended period or in large doses, the decreases in prostaglandin E2 and increases in products of lipid peroxidation in gastric mucosa will produce oxidative stress and its companions, such as inflammation. Our findings showed AC to cause mild to moderate irritation in the GI tract at intake, especially in debilitated patients. The incidence of GI reactions, including abdominal pain and diarrhea, was much greater in the AC group (47.1 %) than in the placebo group. In clinical practice, the AC-associated GI reactions can be alleviated by keeping pieces of ginger (Zingiber officinale Roscoe) in the mouth after AC intake, for this warm and acrid medicine can restrain the cold and bitterness, according to the TCM compatibility of herbal medicines [37]. Recent studies have demonstrated that ginger may provide a protective effect against drug-induced gastric injury and have a potential ulcer-preventive ability [38, 39]. However, because the precise mechanism of the interaction of AC and ginger has yet to be elucidated, we cannot exclude the possibility that an antagonistic effect of the combined agents may play some role. The majorities of enrolled patients in both groups had advanced adenocarcinoma and were receiving similar chemotherapy regimens. The laboratory parameters regarding hematological and non-hematological safety were comparable in both treatment groups. There was a significant decrease in platelet counts within 30 days of adjuvant AC treatment. Our data also indicated a large difference in the decline in platelet levels between the two groups in patients with lung cancer (83.5 % vs. −5.0 %) and gastric cancer (22.4 % vs. −4.9 %). However, the relationship between the decrease in platelet counts and adjuvant AC treatment was not well established in our study. Thrombocytopenia in itself is sometimes just a paper laboratory value, and the toxicity does not affect patient safety or outcome. However, a platelet count below 10 × 109/L (and maybe < 5 × 109/L) often leads to an increased risk of bleeding in cancer patients and can accelerate morbidity [40, 41]. Patients exposed to chemotherapeutic agents, such as cyclophosphamide (CTX), methotrexate (MTX), 5-FU, cytarabine (Ara-C), and etoposide (VP-16), are at higher risk of experiencing thrombocytopenia due to their effects on the production of blood cell progenitors [42]. In our study, few patients received the above chemotherapeutic agents, except 5-FU for liver or colorectal cancer. The extent of thrombocytopenia was analyzed in both groups, and patients with liver or colorectal cancer exhibited no significant differences from the others. Though this observation spanned only a short period, the finding may imply an association between AC and its platelet-lowering effect in patients with lung or gastric cancer. Most research has focused on the antiplatelet activity of AC as a potential therapeutic agent for treating thromboembolic disorders [43, 44]. However, no definitive proof demonstrates that AC interferes with the processes of platelet production, destruction, and pooling. We are fully aware that our sample size was limited and that the data analysis in our study had its restrictions. However, the harmful effect of AC in the treatment of cancer is unknown to the public, especially in the case of advanced cancer patients, so the findings of this study could provide some preliminary data. The small sample size and the recruitment of patients with different types of cancer are the most important problem of our study. For this reason, the study may have had insufficient power to detect significant differences for anti-cancer effects. We know every cancer has its own characteristics and natural course, but it is important to manage the cancer-related symptoms and adverse events of chemotherapy, especially in the advanced stage. Although disappointing results could be anticipated for cancer patients in advanced stage even with the use of plant extract. We still can observe some phenomena, including safety and physical symptoms, after AC treatment. The results should be informative for health care professionals and our patients. This clinical study had several limitations. First, this plant extract has multiple potent mechanisms and different activity, but this study was assayed only some test and, in fact is a marginal experiments. Second, the treatment lasted only 30 days, so related changes in prognostic indicators could not be demonstrated. Longer or continued AC treatment may be required for clinically significant changes in patient well-being to be observed. Third, due to the restrictions on sample collection from advanced cancer patients, the study neither evaluated the relationship between a target cancer and AC usage nor elucidated the effects of AC treatment. Finally, AC is not yet regulated by the Food and Drug Administration. Some studies previously also reported that different extracts from A. camphorata showed cytotoxic activity, so it is need to determine cytotoxic dose and IC50 value before being a complementary therapy [45]. Further investigation using a more sample, distinct cancer in early stage, long-term treatment, and rigorous design will be needed to determine the efficacy of A. cinnamomea. Conclusions A. cinnamomea is a popular folk medicine and has attracted great attention due to its reputation for anti-cancer activity against several types of cancer, but little information is available on its clinical application. This study is the first to report the therapeutic effects of this medical fungus in advanced cancer patients receiving standard chemotherapy. The mean 6-month survival rate was not significantly different between groups (5.4 months vs. 5 months, AC versus placebo, respectively). The only significant difference between groups was the platelet counts, which were lower in the AC group, and sleep quality, which was significantly improved in the AC group. Overall, AC combined with chemotherapy did not improve the outcomes in the advanced cancer patients. Abbreviations 5-FU5-fluorouracil ACAntrodia cinnamomea AEAdverse event Ara-CCytarabine CAMComplementary and alternative medicine CRComplete response CTCAENational Cancer Institute’s Common Terminology Criteria for Adverse Events CTXCyclophosphamide DCRDisease control rate ECOGEastern co-operative oncology group EORTC-QLQ-C30European organization for research and treatment-quality of life questionnaire GIGastrointestinal KCGMHKaohsiung Chang Gung Medical Hospital MDRMulti-drug resistance MFIMultidimensional fatigue inventory MTXMethotrexate NCCNNational Comprehensive Cancer Network NOAELNo-observed-adverse-effect-level OSOverall survival p-AKTphospho-AKT PRPartial response PSPerformance status QoLQuality of life RECISTResponse evaluation criteria in solid tumors ROSReactive oxygen species SDSprague-dawley SDStandard deviation SDStable disease SNQTaiwan Symbol of National Quality TCMTraditional chinese medicine VP-16Etoposide Acknowledgements This work was supported by the Chang Gung Memorial Hospital with XMRPG890251. M.Y. Tsai and H.H.E. Chiu contributed equally to this work. The study drug Antrodia cinnamomea was provided by Amon Biotech Co., Ltd. The authors thank all the physicians and nurses who take care of the patients in the Department of Oncology at the KCGMH. Funding This study was not funded by any organization. Availability of data and materials Related data has been provided in the manuscript in the form of figures and tables. Authors’ contributions MYT and YHC participated in assignment at KCGMH, collected data, interpreted the final results, conducted the study, and wrote the manuscript. MYT and YC Huang were involved in statistical analysis. HMR and YLS participated in clinical evaluation and enrollment. CWK and YHC took part in data collection. HHEC and YC Hung were responsible for study design 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 Ethical approval was received from the Institutional Review Board of Chang Gung Medical Foundation (protocol number: ID 98-3904A3). All participants provided informed written consent. ==== Refs References 1. Magrath I Litvak J Cancer in developing countries: opportunity and challenge J Natl Cancer Inst 1993 85 862 74 10.1093/jnci/85.11.862 8492315 2. Garcia M Jemal A Ward EM Center MM Hao Y Siegel RL Global cancer facts & figures 2007 2007 Atlanta American Cancer Society 3. Urruticoechea A Alemany R Balart J Villanueva A Viñals F Capellá G Recent advances in cancer therapy: an overview Curr Pharm Des 2010 16 3 10 10.2174/138161210789941847 20214614 4. U.S. National Institutes of Health National Cancer Institute. Chemotherapy Side Effects Fact Sheets. http://www.cancer.gov/cancertopics/coping/physicaleffects/chemo-side-effects. Accessed 29 Apr 2015. 5. Li X Yang G Li X Zhang Y Yang J Chang J Traditional Chinese medicine in cancer care: a review of controlled clinical studies published in Chinese PLoS One 2013 8 e60338 10.1371/journal.pone.0060338 23560092 6. Lee YW Chen TL Shih YR Tsai CL Chang CC Liang HH Adjunctive traditional Chinese medicine therapy improves survival in patients with advanced breast cancer: a population-based study Cancer 2014 120 1338 44 10.1002/cncr.28579 24496917 7. Lee TH Lee CK Tsou WL Liu SY Kuo MT Wen WC A new cytotoxic agent from solid-state fermented mycelium of Antrodia camphorata Planta Med 2007 73 1412 5 10.1055/s-2007-990232 17932820 8. Ao ZH Xu ZH Lu ZM Xu HY Zhang XM Dou WF Niuchangchih (Antrodia camphorata) and its potential in treating liver diseases J Ethnopharmacol 2009 121 194 212 10.1016/j.jep.2008.10.039 19061947 9. Hsieh YH Chu FH Wang YS Chien SC Chang ST Shaw JF Antrocamphin A, an anti-inflammatory principal from the fruiting body of Taiwan ofungus camphoratus, and its mechanisms J Agric Food Chem 2010 58 3153 8 10.1021/jf903638p 20128588 10. Patel S Goyal A Recent developments in mushrooms as anti-cancer therapeutics: a review. 3 Biotech 2012 2 1 15 11. Kumar KJ Vani MG Chueh PJ Mau JL Wang SY Antrodin C inhibits epithelial-to-mesenchymal transition and metastasis of breast cancer cells via suppression of Smad2/3 and β-catenin signaling pathways PLoS One 2015 10 e0117111 10.1371/journal.pone.0117111 25658913 12. Hsu YL Kuo YC Kuo PL Ng LT Kuo YH Lin CC Apoptotic effects of extract from Antrodia camphorata fruiting bodies in human hepatocellular carcinoma cell lines Cancer Lett 2005 221 77 89 10.1016/j.canlet.2004.08.012 15797630 13. Chang CY Huang ZN Yu HH Chang LH Li SL Chen YP The adjuvant effects of Antrodia camphorata extracts combined with anti-tumor agents on multidrug resistant human hepatoma cells J Ethnopharmacol 2008 118 387 95 10.1016/j.jep.2008.05.001 18571350 14. Chiu YJ Nam MK Tsai YT Huang CC Subacute oral toxicity assessment in rats for freeze-dried Antrodia cinnamomea mycelium culture “antro-Gem” J Testing Qual Assur 2013 2 145 53 15. Huang CC Nam MK Tsai YT Lan A Evaluation of the sub-chronic toxicity and teratogenicity of Antrodia cinnamomea mycelia Life Sci J 2014 11 1090 8 16. Cheng PC Huang CC Chiang PF Lin CN Li LL Lee TW Radioprotective effects of Antrodia cinnamomea are enhanced on immune cells and inhibited on cancer cells Int J Radiat Biol 2014 90 841 52 10.3109/09553002.2014.911989 24708166 17. National Comprehensive Cancer Network, Inc. [US]. http://www.nccn.org/professionals/physician_gls/f_guidelines.asp. Accessed 15 Oct 2015. 18. Liu YW Lu KH Ho CT Sheen LY Protective effects of Antrodia cinnamomea against liver injury J Tradit Complement Med 2012 2 284 94 10.1016/S2225-4110(16)30114-6 24716143 19. Song TY Hsu SL Yen GC Induction of apoptosis in human hepatoma cells by mycelia of Antrodia camphorata in submerged culture J Ethnopharmacol 2005 100 158 67 10.1016/j.jep.2005.02.043 15949907 20. Chiou JF Wu AT Wang WT Kuo TH Gelovani JG Lin IH A preclinical evaluation of Antrodia camphorata alcohol extracts in the treatment of Non-small cell lung cancer using Non-invasive molecular imaging Evid Based Complement Alternat Med 2011 2011 914561 21785640 21. Liu FS Yang PY Hu DN Huang YW Chen MJ Antrodia camphorata induces apoptosis and enhances the cytotoxic effect of paclitaxel in human ovarian cancer cells Int J Gynecol Cancer 2011 21 1172 9 21897275 22. Meng QX Roubin RH Hanrahan JR Ethnopharmacological and bioactivity guided investigation of five TCM anticancer herbs J Ethnopharmacol 2013 148 229 38 10.1016/j.jep.2013.04.014 23623820 23. Ouseph S Tappitake D Armant M Wesselschmidt R Derecho I Draxler R Cellular therapies clinical research roadmap: lessons learned on how to move a cellular therapy into a clinical trial Cytotherapy 2015 17 339 43 10.1016/j.jcyt.2014.10.008 25484311 24. Cleeland CS Sloan JA Cella D Chen C Dueck AC Janjan NA CPRO (assessing the symptoms of cancer using patient-reported outcomes) multisymptom task force. Recommendations for including multiple symptoms as endpoints in cancer clinical trials: a report from the ASCPRO (assessing the symptoms of cancer using patient-reported outcomes) multisymptom task force Cancer 2013 119 411 20 10.1002/cncr.27744 22930243 25. Fan G Filipczak L Chow E Symptom clusters in cancer patients: a review of the literature Curr Oncol 2007 14 173 9 10.3747/co.2007.145 17938700 26. Gramignano G Lusso MR Madeddu C Massa E Serpe R Deiana L Efficacy of l-carnitine administration on fatigue, nutritional status, oxidative stress, and related quality of life in 12 advanced cancer patients undergoing anticancer therapy Nutrition 2006 22 136 45 10.1016/j.nut.2005.06.003 16459226 27. Huang CC Hsu MC Huang WC Yang HR Hou CC Triterpenoid-rich extract from Antrodia camphorata improves physical fatigue and exercise performance in mice Evid Based Complement Alternat Med 2012 2012 364741 22829854 28. Hseu YC Chang WC Hseu YT Lee CY Yech YJ Chen PC Protection of oxidative damage by aqueous extract from Antrodia camphorata mycelia in normal human erythrocytes Life Sci 2002 71 469 82 10.1016/S0024-3205(02)01686-7 12044846 29. Kripke DF Garfinkel L Wingard DL Klauber MR Marler MR Mortality associated with sleep duration and insomnia Arch Gen Psychiatry 2002 59 131 6 10.1001/archpsyc.59.2.131 11825133 30. Irwin MR Depression and insomnia in cancer: prevalence, risk factors, and effects on cancer outcomes Curr Psychiatry Rep 2013 15 404 10.1007/s11920-013-0404-1 24078066 31. Miller AH Ancoli-Israel S Bower JE Capuron L Irwin MR Neuroendocrine-immune mechanisms of behavioral comorbidities in patients with cancer J Clin Oncol 2008 26 971 82 10.1200/JCO.2007.10.7805 18281672 32. Gulec M Ozkol H Selvi Y Tuluce Y Aydin A Besiroglu L Oxidative stress in patients with primary insomnia Prog Neuropsychopharmacol Biol Psychiatry 2012 37 247 51 10.1016/j.pnpbp.2012.02.011 22401887 33. Chen TI Chen CW Lin TW Wang DS Chen CC Developmental toxicity assessment of medicinal mushroom Antrodia cinnamomea T.T. Chang et W.N. Chou (higher Basidiomycetes) submerged culture mycelium in rats Int J Med Mushrooms 2011 13 505 11 10.1615/IntJMedMushr.v13.i6.20 22181838 34. Chang JB Wu MF Lu HF Chou J Au MK Liao NC Toxicological evaluation of Antrodia cinnamomea in BALB/c mice In Vivo 2013 27 739 45 24292577 35. Su ZC Ruan SB Investigating the Xingwei, Guijing and efficacy of Taiwan Antrodia camphorate Clin J Chinese Med 2013 20 26 30 36. Qiu SH Sun BQ Li L Experimental study on the influence of common-used bitter and cold medicine to the barrier effect of gastric mucosa Chinese J Med Guide 2007 9 140 2 37. Pang B Yu XT Zhou Q Zhao TY Wang H Gu CJ Effect of Rhizoma coptidis (Huang Lian) on treating diabetes Mellitus Evid Based Complement Alternat Med 2015 2015 921416 26508987 38. Wang Z Hasegawa J Wang X Matsuda A Tokuda T Miura N Protective effects of ginger against aspirin-induced gastric ulcers in rats Yonago Acta Med 2011 54 11 9 24031124 39. Nanjundaiah SM Annaiah HN Dharmesh SM Gastroprotective effect of ginger rhizome (Zingiber officinale) extract: role of gallic acid and cinnamic acid in H(+), K(+)-ATPase/H. pylori inhibition and anti-oxidative mechanism Evid Based Complement Alternat Med 2011 2011 249487 10.1093/ecam/nep060 19570992 40. Stanworth SJ Hyde C Heddle N Rebulla P Brunskill S Murphy MF Prophylactic platelet transfusion for haemorrhage after chemotherapy and stem cell transplantation Cochrane Database Syst Rev 2004 4 CD004269 15495093 41. Lopes RD Ohman EM Granger CB Honeycutt EF Anstrom KJ Berger PB Six-month follow-up of patients with in-hospital thrombocytopenia during heparin-based anticoagulation (from the complications after thrombocytopenia caused by heparin [CATCH] registry) Am J Cardiol 2009 104 1285 91 10.1016/j.amjcard.2009.06.045 19840578 42. Kuhn JG Chemotherapy-associated hematopoietic toxicity Am J Health Syst Pharm 2002 59 Suppl 4 4 7 43. Lu WJ Lin SC Lan CC Lee TY Hsia CH Huang YK Effect of Antrodia camphorata on inflammatory arterial thrombosis-mediated platelet activation: the pivotal role of protein kinase C ScientificWorldJournal 2014 2014 745802 25541625 44. Tsai TC Tung YT Kuo YH Liao JW Tsai HC Chong KY Anti-inflammatory effects of Antrodia camphorata, a herbal medicine, in a mouse skin ischemia model J Ethnopharmacol 2015 159 113 21 10.1016/j.jep.2014.11.015 25449461 45. Geethangili M Tzeng YM Review of pharmacological effects of antrodia camphorata and its bioactive compounds Evid Based Complement Alternat Med 2011 2011 212641 10.1093/ecam/nep108 19687189

PMC005xxxxxx/PMC5002174.txt

==== Front BMC BioinformaticsBMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 118010.1186/s12859-016-1180-9Methodology ArticleStatistical inference for time course RNA-Seq data using a negative binomial mixed-effect model Sun Xiaoxiao xiaosun@uga.edu 1Dalpiaz David dalpiaz2@illinois.edu 2Wu Di dwu@fas.harvard.edu 3S. Liu Jun jliu@stat.harvard.edu 3Zhong Wenxuan wenxuan@uga.edu 1Ma Ping pingma@uga.edu 11 Department of Statistics, University of Georgia, 101 Cedar Street, Athens, 30602 USA 2 Department of Statistics, University of Illinois at Urbana-Champaign, 725 South Wright Street, Champaign, 61820 USA 3 Department of Statistics, Harvard University, One Oxford Street, Cambridge, 02138 USA 26 8 2016 26 8 2016 2016 17 1 3243 3 2016 11 8 2016 © The Author(s) 2016 Open Access This 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 Accurate identification of differentially expressed (DE) genes in time course RNA-Seq data is crucial for understanding the dynamics of transcriptional regulatory network. However, most of the available methods treat gene expressions at different time points as replicates and test the significance of the mean expression difference between treatments or conditions irrespective of time. They thus fail to identify many DE genes with different profiles across time. In this article, we propose a negative binomial mixed-effect model (NBMM) to identify DE genes in time course RNA-Seq data. In the NBMM, mean gene expression is characterized by a fixed effect, and time dependency is described by random effects. The NBMM is very flexible and can be fitted to both unreplicated and replicated time course RNA-Seq data via a penalized likelihood method. By comparing gene expression profiles over time, we further classify the DE genes into two subtypes to enhance the understanding of expression dynamics. A significance test for detecting DE genes is derived using a Kullback-Leibler distance ratio. Additionally, a significance test for gene sets is developed using a gene set score. Results Simulation analysis shows that the NBMM outperforms currently available methods for detecting DE genes and gene sets. Moreover, our real data analysis of fruit fly developmental time course RNA-Seq data demonstrates the NBMM identifies biologically relevant genes which are well justified by gene ontology analysis. Conclusions The proposed method is powerful and efficient to detect biologically relevant DE genes and gene sets in time course RNA-Seq data. Electronic supplementary material The online version of this article (doi:10.1186/s12859-016-1180-9) contains supplementary material, which is available to authorized users. Keywords Differentially expressed geneGene set enrichmentAnalysis of varianceSmoothing splinePenalized likelihoodhttp://dx.doi.org/10.13039/100000001National Science FoundationDMS-1440037Ma Ping http://dx.doi.org/10.13039/100000001National Science FoundationDMS-1438957Ma Ping http://dx.doi.org/10.13039/100000001National Science FoundationDMS-1440038Zhong Wenxuan http://dx.doi.org/10.13039/100000002National Institutes of Health1R01GM113242-01Zhong Wenxuan http://dx.doi.org/10.13039/100000002National Institutes of Health1R01GM122080-01Ma Ping issue-copyright-statement© The Author(s) 2016 ==== Body Background RNA-sequencing (RNA-Seq) technology has become a preferred choice for studying transcriptomes [1, 2]. Compared to microarray, RNA-Seq provides a single nucleotide level measurement of mRNA expression levels. It offers the chance to detect novel transcripts by obtaining tens of millions of short reads. When mapped to the genome or reference transcripts, RNA-Seq data are summarized by a number of read counts. The huge number of read counts enables researchers to quantify transcriptomes in ultra-high resolution [3, 4]. To study the dynamics of genome-wide mRNA expression levels during a biological process, e.g., development, researchers often conduct time course RNA-Seq experiments. As in static RNA-Seq experiments (RNA-Seq taken irrespective of time), identifying differentially expressed (DE) genes across different treatments or conditions is still a key task in time course RNA-Seq experiments. Inferring DE genes in time course RNA-Seq experiments has a number of interesting challenges. First, the DE genes in time course data are those with different gene expression profiles along the time across treatments or conditions. However, most of the available methods treat expressions of a gene at different time points as replicates and test the significance of the mean expression difference between treatments or conditions irrespective of time, e.g., edgeR [5] and DESeq [6]. They thus fail to identify many DE genes with different profiles across time. Second, some methods have been developed recently to identify the DE genes with different expression profiles over time. A recent work by Oh et al. [7] models time dependency using a hidden Markov model. Such a model requires the Markov property. In particular, the Markov property states that the conditional dependency of prior information from all time can be simplified to the conditional dependency of prior information of k time points (kth order Markov chain). It is still unclear whether such Markov property holds for general time course RNA-Seq data. Finally, both edgeR and DESeq use the total read counts of each gene and model the variation of the read counts across the replicates at gene level. When RNA-Seq experiments do not have replicates or the number of replicates is small, the statistical significance tests in edgeR and DESeq have small degrees of freedom and may result in a high false discovery rate (FDR). To surmount these challenges, we develop a novel statistical method to identify DE genes in this article. The input of our method is the read counts at the exon level for each gene at each time point. The read counts of genes at the exon level across different time points are modeled by a negative binomial mixed-effect model (NBMM). In this model, the mean gene expression profiles over time across treatments are modeled by a nonparametric bivariate function of time and treatments, while the time dependency is characterized by a parametric random effect. The nonparametric bivariate function has great flexibility in modeling different expression profiles over possibly non-equally spaced time points across treatments and conditions. The parametric random effects are used to define a variety of time dependency correlation structures. The model is fitted by a penalized likelihood method. In order to identify DE genes unique to time course experiments, we define two types of DE genes in time course RNA-Seq experiments: nonparallel differentially expressed (NPDE) genes with nonparallel expression profiles over time across treatments, see Fig. 1, and parallel differentially expressed (PDE) genes with parallel expression profiles over time across treatments, see Fig. 2. PDE genes are those consistently up-regulated or down-regulated over time across treatments, whereas NPDE genes are those that have significant expression profile changes over time across treatments. Compared with PDE genes, in many scientific investigations, NPDE genes are of primary interest. Focused study of the NPDE genes may provide more information on how the cell responds differently to different stimulus or treatments. Moreover, time course RNA-Seq experiments are commonly used in case-control studies and in clinical trials. In such experiments, mRNA samples are taken from a small number of subjects over time in the treatment group and from another small number of subjects in the control group. Because each group only consists of a small number of subjects, one subject with high baseline gene expression can cause a high average gene expression for the whole group. Thus, there are many PDE genes between treatments, but they are biologically irrelevant [8]. To distinguish the two types of DE genes, we decompose the nonparametric bivariate function in our model into the main effects of time and treatment separately, as well as their interaction through a functional ANOVA decomposition. The identification of DE genes is equivalent to testing significance of treatment-time interactions in the functional ANOVA decomposition. We fit this model to the exon level read counts data using penalized maximum likelihood. The tuning parameter is selected by cross-validation [9]. Fig. 1 NPDE gene. Gene ss (FlyBase ID: FBgn0003513) was identified as non-parallel differentially expressed with p value=0.00. Different exons are represented by curves with varying colors. This gene participates in antennal development, antennal morphogenesis, and imaginal disc-derived leg segmentation. Read counts on the y-axis are the average counts (The total read counts on each exon divided by the length of exon). The left panel and right panel represent the early and late embryonic developmental stages respectively Fig. 2 PDE gene. Gene I d g f2 (FlyBase ID: FBgn0020415) was identified as parallel differentially expressed with p value=0.00. Different exons are represented by curves with varying colors. This gene participates in imaginal disc development. Read counts on the y-axis are the average counts (The total read counts on each exon divided by the length of exon). The left panel and right panel represent the early and late embryonic developmental stages respectively Methods Nonparametric model and penalized likelihood method We first provide a short review of nonparametric model and penalized likelihood method. Consider the nonparametric model for data points (ti,yi), 1 yi=η(ti)+εi,i=1,⋯,T, where η is the mean function and random noise εi are independently Gaussian distributed. When η is assumed to be of form η(ti)=tiβ, which is linear in β, one has a standard linear model. The disadvantage of linear model is illustrated by a toy example, where we generated 100 data points, faded circles in Fig. 3, from a nonlinear function. The linear model fit, the dashed straight line in Fig. 3, does not provide a good fit. Since linear model is too restrictive to model nonlinear function, we allow η to vary in a high-dimensional functional space, leading to diverse nonparametric estimators. Fig. 3 Curve fitting examples. The curve fitted using penalized likelihood is in the solid line and the linear fit is in dashed line, with the interpolation fit superimposed in faded line and the data in circles An approach to the estimation of η is via the minimization of negative log likelihood, 2 ∑i=1Tyi−η(ti)2. Without any constraint, the minimizer η^ in (2) simply interpolates the data and has no predicting power, see the faded line in Fig. 3. To avoid this problem, one uses penalized likelihood to get a smoothing estimator of η via minimization of 3 ∑i=1Tyi−η(ti)2+λ∫η′′(t)2dt, where η′′ is the second derivative and characterizes the smoothness of η,λ is a smoothing parameter, which controls the trade-off between the lack of fit of the nonparametric model and the roughness of η. To select the proper λ, researchers often use generalized cross validation [10]. An adequate fit by a proper selected λ is illustrated by the solid curve in Fig. 3. Negative binomial mixed-effect model In time course RNA-Seq experiments, the short read counts cannot be adequately modeled by independent Gaussian distribution. We extend the aforementioned modeling strategy to develop a negative binomial mixed-effect model (NBMM) for modeling time course RNA-Seq data. The model specification Suppose the time course RNA-Seq experiments are conducted across G conditions/treatments. For each gene, the mapped read counts on exon k at time ti in condition/treatment g, denoted by Yigk, are assumed to follow a negative binomial distribution (NegBin), 4 Yigk∼NegBin(ν,p(ti,g,k)), where the negative binomial distribution has the probability distribution, 5 P(Yigk=y)=Γ(ν+y)y!Γ(ν)p(ti,g,k)ν(1−p(ti,g,k))y, where ν is a nuisance parameter, which is the number of reads that cannot be mapped to the reference genome, and 1−p(ti,g, k) is the probability that a read is mapped to exon k in condition g at time ti,g=1,⋯,G, i=1,⋯,ng,k=1,⋯,K. In this setting, ng is the number of time points in the gth condition, and K is the number of exons. In most cases, we only have two treatments: case and control or mutant and wild type (G=2). To model the time trend and capture the time dependence, we use a nonparametric mixed-effect model with logit link ([11], p.199) 6 log{p(ti,g,k)/(1−p(ti,g,k))}=log(βti,g)+η(ti,g)+zkbk, where βti,g is the effective library size, used in edgeR [12], of the tith time point, mean expression η is assumed to be a smooth function of time t for each treatment g, zk is the length of the kth exon, bk represents the exon specific random effect to model the intra-exon variation with bk∼N(0,σ2), and the random effect variance σ2 is to be estimated from the data. The log(βti,g) term provides a convenient device to normalize the reads to a common scale. In model (6), the bivariate function η is decomposed as 7 η(t,g)=η∅+η1(t)+η2(g)+η1,2(t,g), where η∅ is the baseline expression irrespective of time and treatment, η1(t) is the time effect at time t, η2(g) is the treatment effect of the gth condition, and η1,2(t, g) is the interaction between time and treatment effects. The time and treatment effects are defined as the deviation from the baseline expression, and, therefore, ∫0Tη1(t)dt=0 and ∑g=1Gη2(g)=0. Analogously, the time-treatment interaction is defined as ∫0Tη1,2(t,g)dt=0 for all g, and ∑g=1Gη1,2(t,g)=0 for all t. This decomposition is referred to as the functional ANOVA decomposition [11, 13]. If the time-treatment interaction term η1,2(t, g) is significant, we have η(t, g1)−η(t, g2)=η2(g1)−η2(g2)+η1,2(t, g1)−η1,2(t, g2) for every t. In the right hand side, the first two terms are constants and the remaining terms vary with t. When the time-treatment interaction η1,2(t, g) is not significant in (7), the model reduces to 8 η(t,g)=η∅+η1(t)+η2(g), which produces the parallel population mean time course profiles for different treatment conditions, i.e., η(t, g1)−η(t, g2)=η2(g1)−η2(g2) for each t, where the right hand side of the equation is a constant which does not vary with t. To distinguish the expression profiles, we define the genes with significant time-treatment interaction term in (7), i.e., η1,2(t, g)≠0, as non-parallel differentially expressed (NPDE) genes. If genes have a significant main effect in treatment g but no time-treatment interaction in (7), i.e., η2(g)≠0 and η1,2(t, g)=0, we define those as parallel differentially expressed (PDE) genes [8]. Estimation By (5), one has a minus log likelihood 9 ∑k=1K∑g=1G∑i=1ng(ν+Yigk)log1+elog{p(ti,g,k)/(1−p(ti,g,k))}−νlogp(ti,g,k)/(1−p(ti,g,k)). Substituting (6) into (9), we get the minus log likelihood of Y conditioning on random effects b, where Y=(Y111,⋯,YnG,G,K)T, and b=(b1,⋯,bK)T. Therefore, the (Henderson) likelihood [14] of (Y,b) is 10 log(fy|b(Y|b)fb(b))∝∑k=1K∑g=1G∑i=1ng(ν+Yigk)log1+elog(βti)+η(ti,g)+zkbk−νlog(βti)+η(ti,g)+zkbk+∑k=1Kbk2/σ2. In (10), the fy|b denotes the conditional distribution (negative binomial) of Y given b, and fb denotes the distribution (normal) of b. In the end, we derive a penalized (Henderson) likelihood ([9], p.486) as 11 ∑k=1K∑g=1G∑i=1ng(ν+Yigk)log1+elog(βti)+η(ti,g)+zkbk−νlog(βti)+η(ti,g)+zkbk+∑k=1Kbk2/σ2+NλJ(η), where N=∑k=1K∑g=1Gng, the quadratic functional J(η) quantifies the smoothness of η, and the smoothing parameter λ controls the trade-off between the goodness-of-fit and the smoothness of η. The minimization of (11) is performed in a reproducing kernel Hilbert space ℋ⊆{η:J(η)<∞}, in which J(η) is a square semi-norm [13]. For model (6) with functional ANOVA (7), we employ the following quadratic penalty, which produces a cubic spline estimate, 12 J(η)=θ1−1∫0Td2η1/dt22dt+θ1,2−1∫0T∑g=1Gd2η1,2/dt22dt, where θ1 and θ1,2 are extra smoothing parameters that adjust the relative penalties on the roughness of different components. See detailed examples in Sect. 2.4 of [11]. For model (6) with functional ANOVA (8), we use penalty 13 J(η)=∫0Td2η1/dt22dt. To perform the penalized likelihood estimation of (11), we implement two nested iterative loops [9]. Fixing the smoothing parameter, the inner loop minimizes (11), and the outer loop estimates the smoothing parameters and variance of random effects via the minimization of certain cross-validation score, see [9] for details. For fixed smoothing parameter λ, (11) can be minimized through Newton iteration. Write 14 ligk(ζigk)=(ν+Yigk)log(1+eζigk)−νζigk, where ζigk=log(βti)+η(ti,g)+zkbk. The quadratic approximation of ligk(ζigk) at ζ~igk is 15 ligk(ζigk)≈ligk(ζ~igk)+μ~igk(ζigk−ζ~igk)+ω~igk(ζigk−ζ~igk)2/2=ω~igk(Y~igk−ζigk)2/2+Eigk, where Y~igk=ζ~igk−μ~igk/ω~igk and Eigk is independent of ζigk; μ~igk=(ν+Yigk)p~(ti,g,k)−ν and ω~igk=ν(1−p~(ti,g,k)). The Newton iteration can thus be performed via iterated weighted least squares, 16 ∑k=1K∑g=1G∑i=1ngω~igk(Y~igk−log(βti)+η(ti,g)+zkbk)2+∑k=1Kbk2/σ2+NλJ(η). Since ν is unknown, we estimate it from data. We apply the log operation to (5), and drop the terms that do not involve ν to get the individual objective function. Then the joint objective function is the sum of minus individual objective functions, 17 1N∑k=1K∑g=1G∑i=1nglog(Γ(ν))−logΓ(ν+Yigk)−νlog(p(ti,g,k)), where Γ is the gamma function. Given (Yigk,p(ti,g, k)), one estimates ν via the minimization of (17). We iterate between the estimations of η and ν in (11) and (17) [11]. Significance testing for individual gene Once the model (6) is fitted to the exon level read counts data, we identify NPDE and PDE genes by testing the significance of the interaction and main effects in (7). To identify NPDE genes, we test the significance of the time-treatment interaction in (7), which is, 18 H0:η1,2(t,g)=0;H1:η1,2(t,g)≠0. To derive the needed test statistic, we first define the Kullback-Leibler distance [11] 19 KL(η,η^)=1N∑k=1K∑g=1G∑i=1ngνp(ti,g,k)log1−p(ti,g,k)1−p^(ti,g,k)+ν(η(ti,g)−η^(ti,g)). Then, we use the following Kullback-Leibler distance ratio (KLR) [15] as our test statistic 20 KLR=KL(η^F,η^R)KL(η^F,ηC), where η^F stands for a full model estimate given that H1 is true in the ANOVA decomposition (7), and η^R represents a reduced model estimate under the hypothesis that H0 is true in (7). Analogously, we define ηC as a constant function. For genes that are not considered as NPDE by the preceding test, we further investigate whether they are PDE or not. In model (6) with functional ANOVA (8), we are interested in testing 21 H0:η2(g)=0;H1:η2(g)≠0. In testing for PDE genes, the full model estimate η^F does not include a time-treatment interaction, and η^R only has an overall mean and time effect in (8). The p values for identifying NPDE and PDE genes are calculated through a permutation procedure. First, we compute a Kullback-Leibler distance ratio KLR for a gene. Second, the time labels for the gene are shuffled, and we recompute the statistic for the shuffled gene. We repeat the second step B times to obtain KLR1∗,⋯KLRB∗. In the end, the p value for the gene is given by, 22 #KLRi∗>KLR,i=1,⋯,B/B, where #{·} represents the cardinality of the set, i.e., the number of permuted KLR∗s which is larger than the KLR. Gene set significance testing In many studies, researchers are not only interested in identifying individual DE genes, but also in finding DE gene sets. A gene set may be defined by known biological information, for instance, a group of genes within the same biological pathway. Since genes within the same gene set are closely related, we increase statistical power of significance tests by borrowing information across genes. In addition, we obtain more robust results from gene sets than from individual genes. Subramanian et al. [16] proposed an approach named Gene Set Enrichment Analysis (GSEA), which tested the significance of pre-defined gene sets through a Kolmogorov-Smirnov like test. Efron and Tibshirani [17] proposed gene set analysis (GSA), which was shown to make a significant improvement over GSEA. Following the ideas from GSEA and GSA, we test for significant NPDE gene sets via the following steps. Initially, pre-defined gene sets S1,S2, …, SP are collected. Then, we compute the Kullback-Leibler distance ratio KLR based on (20) for all genes. For each gene set, Sk, we calculate a gene set score, Rk, defined as the average of the Kullback-Leibler distance ratios in (20), 23 Rk=∑i∈SkKLRi/#{Sk}, where #{Sk} is the number of genes in gene set Sk. The gene set score Rk defines an enrichment test statistic, with a larger value of Rk suggesting a greater enrichment of NPDE genes. The PDE gene sets can be tested in the same way. To test the significance of the gene set, a threshold is needed. The following permutation procedure is used to determine the threshold, and gene sets with values of Rk above the threshold are declared significant. In particular, we shuffle the time label for each gene and recompute the statistic for each permuted gene. We utilize formula (23) to calculate the permuted gene set scores R1∗,⋯,RB∗, where B is permutation times. In the end, we calculate the p value of the kth gene set, given by, 24 #Ri∗>Rk,i=1,⋯,B/B. Results Simulation study We evaluated the performance of the proposed method by carrying out extensive analysis on simulated datasets. Datasets were generated from both the NBMM model and an RNA-Seq simulator. All p values were adjusted by Benjamini and Hochberg (BH) method for multiple testing corrections [18]. Single gene simulation We simulated exon level read counts according to Eqs. (4), (5) and (6). The effective library sizes of all time points were estimated by edgeR. We have three settings in this section. For each setting, bk∼N(0,1),k=1,2,3, accounts for variation of different exons, z1=0.1,z2=0.25 and z3=0.4 and ν is set to be 1000 for all those settings. Each exon was simulated with both single replicate and three replicates. First setting: linear pattern. In the first setting, we generated exon level read counts of DE genes, see the top panel in the Fig. 4, using the following function, 25 η(ti,g)=C((0.9−2ti)I[g=2]+ti), Fig. 4 Simulated read counts. Simulated read counts generated from a negative binomial distribution. Samples of DE genes in the first, second and third setting are shown in the top, middle and bottom panel respectively. Different exons are represented by curves with varying colors where ti=i/10,i=1,2⋯,8,g=1,2, and C=2 is a scale factor, I[g=2] is an indicator function which equals one when g=2 and zero otherwise. Second setting: exponential pattern In the second setting, we simulated exon level read counts of DE genes, see the middle panel in the Fig. 4, using the following smooth function, 26 η(ti,g)=exp104F111F26+102F13F29+Cg, where F1=(0.9−2ti)I[g=2]+ti,F2=0.1I[g=2]+I[g=1]+(1−2I[g=1])ti, and C1=C2=1. The constants Cg,g=1,2, define fixed reference expression levels for different conditions. Third setting: cyclic pattern In the third setting, exon level read counts of DE genes, see the bottom panel in Fig. 4, were generated using the following smooth function, 27 η(ti,g)=sin(2.5π((0.9−2ti)I[g=2]+ti)+2. There were two scenarios in each setting. In the first scenario, we simulated time course exon level read counts of 50 genes. Half of the genes were DE genes generated by the above mean functions, and the remaining genes were generated as non-differentially expressed (NDE) genes by using the same mean function for different conditions. In the second scenario, 25 DE genes had the same profiles as those in the first scenario and 225 NDE genes were modeled as flat profiles. We compared the NBMM with three methods, maSigPro [19], DyNB [20] and edgeR. The former two methods are designed for time course data. Analysis followed the steps described in the R package documentation and unless stated otherwise default parameters were used. Table 1 summarizes the performance of each method. The FDR was calculated as the number of false positives divided by the number of identified DE genes, and the False Non-Discovery Rate (FNR) as the number of false negatives divided by the number of genes which were not identified as DE genes. DyNB was only applied to the simulated data set of the first scenario in each setting due to its extensive computational cost, see Table 2. In the third setting, the DyNB failed to report the results for the data set with one replicate. In addition, edgeR was not recommended for single replicate data sets and, therefore, not used in each single replicate dataset [19]. Table 1 The FDR and FNR of all methods for detecting DE genes in simulation studies. If the method failed to report any significant genes, the FDR was NA and FNR was 0.50 for scenario 1 and 0.09 for scenario 2 Setting 1 Setting 2 Setting 3 FDR FNR FDR FNR FDR FNR NBMM Scenario 1 1 Rep 0.00 0.00 0.00 0.17 0.00 0.14 3 Rep 0.00 0.00 0.00 0.00 0.00 0.14 Scenario 2 1 Rep 0.07 0.00 0.00 0.02 0.21 0.02 3 Rep 0.16 0.00 0.15 0.01 0.09 0.02 maSigPro Scenario 1 1 Rep 0.11 0.00 0.00 0.00 NA 0.50 3 Rep 0.00 0.07 0.00 0.04 NA 0.50 Scenario 2 1 Rep 0.00 0.00 0.00 0.00 NA 0.09 3 Rep 0.00 0.01 0.00 0.01 NA 0.09 DyNB Scenario 1 1 Rep NA 0.50 0.00 0.36 NA 0.50 3 Rep 0.54 0.54 0.32 0.32 0.43 0.20 Scenario 2 1 Rep NA 3 Rep NA edgeR Scenario 1 Rep 1 NA 3 Rep 0.50 NA 0.50 NA 0.50 NA Scenario 2 1 Rep NA 3 Rep 0.88 0.00 0.00 0.00 0.86 0.00 The best result in each scenario is shown in boldface Table 2 The running CPU time (seconds) for all methods in simulation studies Setting 1 Setting 2 Setting 3 NBMM 1 Rep 7.133 6.182 7.261 3 Rep 6.240 6.271 7.000 maSigPro 1 Rep 0.215 0.025 0.200 3 Rep 0.235 0.091 0.236 DyNB 1 Rep 31944.470 NA 42513.210 3 Rep 36228.200 36335.970 40412.250 edgeR 1 Rep 0.004 0.001 0.001 3 Rep 0.001 0.001 0.001 The performance of edgeR, DyNB and maSigPro in terms of FDR and FNR was not as good as that of NBMM in the first scenario. This is expected since edgeR is not designed for time course data and the accuracy of detecting DE genes is affected by the estimated effective library size. When the NDE genes do not show flat profiles, the prediction performance of edgeR and maSigPro relying on TMM normalization [12] will be impaired. maSigPro had a better performance compared with NBMM method in the second scenario in linear and exponential settings. However, our method performed much better than other methods in more complicated patterns, such as a cyclic pattern. For this pattern, other methods either failed to detect any DE genes or identified almost all the genes as DE genes. In particular, in the first setting, the proposed NBMM method identified all DE genes. In the third setting, our approach identified about 88 % of DE genes with FDR 0.00 in the first scenario, whereas the maSigPro failed to detect any DE genes. In summary, as the pattern of the mean function moves away from linear to nonlinear, the advantage of the NBMM over other methods is getting more significant in detecting DE genes. The NBMM took 7 s (CPU time) to process 50 genes with three replicates. Running CPU time for other settings are shown in Table 2. In summary, edgeR is not designed for time course RNA-Seq data, and, therefore, their performance is not as good as that of the NBMM and maSigPro in most settings. The maSigPro is applicable to time course RNA-Seq data and has a good performance in the roughly linear pattern. Its performance in the highly nonlinear pattern is not as good as the NBMM. Simulation using RNA-Seq simulator An RNA-Seq simulator, polyester [21], was applied to simulate RNA-Seq experiments. The simulator takes a set of annotated transcripts as input and produces files containing simulated RNA-Seq reads after simulating the steps of an RNA-Seq experiment. The reference genome used in the simulation was from Drosophila melanogaster. Tophat [22], samtools [23] and DEXSeq [24] were utilized to estimate the read counts data from the simulated fasta files. Analysis followed the steps described in the documentations and unless stated otherwise, default parameters were used. We simulated the data of 7763 transcripts. By directly specifying the number of reads in each transcript, we simulated two expression patterns, linear expression pattern in (28) and nonlinear expression pattern in (29). In each pattern, 125 DE genes were created. 28 vti,g=r((5−ti)I[g=2]+ti), where r is the reference expression level defined in (30) and ti=1+3(i−1)/7. 29 vti,g=r(sin(2.5π((0.9−2ti)I[g=2]+ti)+2). The reference expression level is 30 r=20ι/υ, where ι is the length of transcript and υ=100 is the length of short reads. The expression values for NDE genes in all time points are defined in (30). Removing genes with zero expression values over all time points, we came down with a data set including 4526 genes, among which 219 genes were DE genes. We applied NBMM, maSigPro and edgeR to the dataset and results were summarized in Table 3. NBMM and maSigPro detected all DE genes with linear change pattern, however, NBMM identified 40 DE genes with nonlinear pattern whereas maSigPro found no genes with this pattern. As we can see in Table 3, the FDR and FNR of NBMM are lower than those of maSigPro. edgeR identified almost all the genes as DE genes and resulted in a higher FDR in Table 3. Table 3 The FDR and FNR of all methods for detecting DE genes in simulation using polyester FDR FNR NBMM 0.621 0.018 maSigPro 0.737 0.028 edgeR 0.925 0.00 The best result in each scenario is shown in boldface Gene sets simulation In this study, we simulated 30 gene sets, each with ten genes. All 100 genes in the first ten gene sets were NPDE genes generated by the first setting in (25). The rest of the gene sets were NDE genes with the same mean function for two conditions. We chose ν=1000,C=2 and calculated the gene set scores and p values for the simulated data. The R package GSA developed in [17] was used to detect DE genes enriched gene sets. In GSA package, we set method=“mean”, minsize=10, resp.type=“two class unpaired” and other parameters as default. The p values for all 30 gene sets calculated by NBMM and GSA are plotted in Fig. 5. The NBMM method detected all NPDE genes enriched gene sets, whereas the GSA method did not identify any significant gene sets. Fig. 5 A comparison between result of NBMM and that of GSA. The p values of the proposed method are shown as pink cycles. The p values from GSA are shown as blue circles. The x-axis represents the gene set index, and the first 10 gene sets are the NPDE gene enriched gene sets Real data analysis Study of the development of Drosophila melanogaster (fruit fly) is important since this biological process shares many common features among different organisms. Graveley et al. [25] reported a time course RNA-Seq experiment of Drosophila melanogaster embryogenesis. The dataset included 12 embryonic samples collected at 2-hour intervals for 24 h. Each sample was collected at different stages of development. Sequencing was performed using the Illumina Genome Analyzer II platform. Reads of length 75 were uniquely aligned to the Drosophila melanogaster r5 genome using Bowtie [26]. Since in the first six time points, fruit flies were in the cleavage and gastrulation processes, whereas in the remaining six time points, they were in the process of differentiation [27], we divided the 12 time points into two developmental stages: early and late embryonic developmental stages. After data screening [5], the dataset used in our analysis consists of 1900 genes with different numbers of exons. Among these 1900 genes, 161 genes are related to embryo development (GO: 0009790) [28]. We aim to identify DE genes between the two developmental stages and find the significant pathways. Single gene testing The NBMM model was fitted gene-by-gene and the KLRs were calculated. The permutation procedure was used to obtain the p value for each individual gene. After multiple testing corrections, our method identified 192 NPDE genes and 751 PDE genes at a significance level of 0.05. We conducted functional annotation clustering for these genes using DAVID [29]. For NPDE genes, eight annotation clusters with enrichment scores above 2.0 were found. Seven of them are related to embryo development. For PDE genes, ten annotation clusters with enrichment scores above 2.0 were found. These clusters are associated with the regulation of RNA splicing, mitosis, and development related pathways. Moreover, edgeR was applied to this dataset and 518 DE genes were found. There were 292 genes in common between the edgeR and proposed approach, see Fig. 6. Therefore, 651 DE genes were specifically found by NBMM and 226 DE genes were identified exclusively by edgeR. Among 161 genes in embryo development (GO: 0009790), 86 genes were identified by NBMM method, whereas edgeR detected 39 genes. For genes exclusively selected by edgeR, only two clusters with enrichment scores above 2.0 were found. These clusters are associated with certain catabolic processes. However, there were 11 clusters with enrichment scores above 2.0 for DE genes exclusively identified by the NBMM method. The biological processes associated with the clusters are the regulation of mRNA processing, mitosis, nuclear division, determination of anterior/posterior axis, embryo, and neuroblast differentiation, etc. Fig. 6 A comparison between the result of NBMM and that of edgeR. The Venn diagram between the sets of DE genes identified by NBMM and edgeR In addition, we compared the NBMM with maSigPro, which detected 1012 DE genes. There were 588 genes in common between these two models, see Fig. 7. The NBMM specifically found 355 DE genes and 424 DE genes were identified exclusively by maSigPro. The annotation clustering was applied to these specifically identified DE genes. For genes exclusively selected by maSigPro, five clusters with enrichment scores above 2.0 were found. These clusters are associated with neuron projection morphogenesis, regulation of nuclear mRNA splicing and stem cell maintenance, etc. There were three clusters with enrichment scores above 2.0 for DE genes exclusively identified by the NBMM. The biological processes associated with the clusters are the mitosis, embryonic hindgut morphogenesis, gut development, etc. For the detailed functional annotation clustering, see the Additional files 1-6. Fig. 7 A comparison of the result of NBMM and that of maSigPro. The Venn diagram between the sets of DE genes identified by NBMM and maSigPro Gene sets testing The pathway gene sets of the fruit fly were compiled using the Bioconductor package “org.Dm.eg.db”. The Entrez Gene identifier (version in Nov 2012) in each gene ontology term of org.Dm.egGO2ALLEGS was converted to official gene symbols using the org.Dm.egSYMBOL. We selected the gene sets with 15 to 30 genes and at least five of the 1900 genes were in the gene sets. We performed 100 permutations and chose pathways at the significance level of 0.05. Among 340 tested gene sets, 22 NPDE gene sets were selected by the NBMM, and 18 significant gene sets were selected by the GSA. Among 22 NPDE gene sets, eight gene sets are involved in the cell differentiation and cell development, see Table 4. The 18 significant gene sets detected by the GSA are the induction of apoptosis, chromosome localization, establishment of chromosome localization, cytoskeletal anchoring at plasma membrane, sarcomere organization, etc. These 18 gene sets are not associated with embryonic pathways. For the detailed information, see the Additional file 7. This shows that gene sets detected by the NBMM are more biologically relevant to development. Table 4 The significant pathways identified by the NBMM gene set analysis of the fruit fly data Pathway name p value Segment polarity determination 0.00 Salivary gland boundary specification 0.00 Glial cell differentiation 0.00 Glial cell development 0.00 Axon choice point recognition 0.00 Epithelial cell differentiation 0.00 Regulation of tube length, open tracheal system 0.00 Establishment of blood-brain barrier 0.00 Discussion Time course RNA-Seq data provide valuable insights into biological development and identifying biologically relevant DE genes is a key issue. We classify DE genes into two types: NPDE and PDE genes. Compared with PDE genes, NPDE genes are more likely to be biologically relevant. Therefore, focused study of the NPDE genes may provide more information on the underlying biological mechanisms. In this article, we proposed a statistical method, NBMM, for identifying DE genes in time course RNA-Seq experiments. Compared to other available methods, such as edgeR, the NBMM models time dependency and exon variation using a mixed-effect model. Moreover, the proposed NBMM method outperforms other approaches designed for time course RNA-Seq data in terms of DE genes detection accuracy, such as maSigPro and DyNB. The advantage of the NBMM over other competing methods is significant when they are applied to single replicate time course RNA-Seq data. Furthermore, gene sets significance test is shown to effectively detect DE gene sets. The NBMM method is applied to gene expression data on a gene-by-gene basis. Thus, parallel computing can be employed for testing the significance of multiple genes simultaneously. We implemented a parallel computing option in our timeSeq package to speed up the computing process. Conclusions In this paper, we developed a negative binomial mixed-effect model (NBMM) to detect the differentially expressed (DE) genes in time course RNA-Seq data. We showed that our approach outperforms other currently available methods in both synthetic and real data. The timeSeq, an open source software package, is freely available from CRAN. Additional files Additional file 1 Functional Annotation Clustering for NPDE genes. (XLSX 12 kb) Additional file 2 Functional Annotation Clustering for PDE genes. (XLSX 14 kb) Additional file 3 Functional Annotation Clustering for genes specifically found by NBMM in comparison with edgeR. (XLSX 14 kb) Additional file 4 Functional Annotation Clustering for genes specifically found by edgeR in comparison with NBMM. (XLSX 10 kb) Additional file 5 Functional Annotation Clustering for genes specifically found by NBMM in comparison with maSigPro. (XLSX 10 kb) Additional file 6 Functional Annotation Clustering for genes specifically found by maSigPro in comparison with NBMM. (XLSX 11 kb) Additional file 7 Significant gene sets detected by the GSA. (XLSX 44 kb) Acknowledgements We are grateful to Fan Gao for assistance in developing the R package and Patrick Kriebel for critical reading of this manuscript. Funding This work was funded in part by NSF DMS-1440037, 1438957, 1440038 and NIH 1R01GM113242-01, 1R01GM122080-01. This work used the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by National Science Foundation grant number ACI-1053575. Authors’ contributions XS and PM designed research; XS performed research; XS, DD, and DW analyzed data; XS and WZ designed simulation; and XS, JL, WZ, and PM 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 and consent to participate Not applicable. ==== Refs References 1 Salzman J Jiang H Wong WH Statistical modeling of RNA-seq data Stat Sci. 2011 26 1 62 83 10.1214/10-STS343 2 Mortazavi A Williams BA McCue K Schaeffer L Wold B Mapping and quantifying mammalian transcriptomes by RNA-seq Nature Methods 2008 5 7 621 8 10.1038/nmeth.1226 18516045 3 Nagalakshmi U Wang Z Waern K Shou C Raha D Gerstein M Snyder M The transcriptional landscape of the yeast genome defined by RNA sequencing Science 2008 320 5881 1344 1349 10.1126/science.1158441 18451266 4 Wilhelm BT Marguerat S Watt S Schubert F Wood V Goodhead I Penkett CJ Rogers J Bähler J Dynamic repertoire of a eukaryotic transcriptome surveyed at single-nucleotide resolution Nature 2008 453 7199 1239 1243 10.1038/nature07002 18488015 5 Robinson MD McCarthy DJ Smyth GK edgeR: a Bioconductor package for differential expression analysis of digital gene expression data Bioinforma 2010 26 1 139 40 10.1093/bioinformatics/btp616 6 Anders S Huber W Differential expression analysis for sequence count data Genome Biology 2010 11 106 10.1186/gb-2010-11-10-r106 20236492 7 Oh S Song S Grabowski G Zhao H Noonan JP Time series expression analyses using RNA-seq: a statistical approach BioMed Res Int. 2013 2013 1 16 8 Ma P Zhong W Liu JS Identifying differentially expressed genes in time course microarray data Stat Biosci 2009 1 2 144 59 10.1007/s12561-009-9014-1 9 Gu C Ma P Generalized nonparametric mixed-effect models: computation and smoothing parameter selection J Comput Graph Stat 2005 14 2 485 504 10.1198/106186005X47651 10 Craven P Wahba G Smoothing noisy data with spline functions Numerische Mathematik 1978 31 4 377 403 10.1007/BF01404567 11 Gu C Smoothing Spline ANOVA Models 2013 New York Springer 12 Robinson M Oshlack A A scaling normalization method for differential expression analysis of RNA-seq data Genome Biol 2010 11 3 25 10.1186/gb-2010-11-3-r25 13 Wahba G Spline Models for Observational Data 1990 Philadelphia Society for Industrial and Applied Mathematics 14 Robinson GK That BLUP is a good thing: The estimation of random effects Stat Sci. 1991 6 15 32 10.1214/ss/1177011926 15 Gu C Model diagnostics for smoothing spline ANOVA models Canadian J Stat 2004 32 4 347 58 10.2307/3316020 16 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 2005 102 43 15545 15550 10.1073/pnas.0506580102 16199517 17 Efron B Tibshirani R On testing the significance of sets of genes Annals Appl Stat 2007 1 1 107 29 10.1214/07-AOAS101 18 Benjamini Y Hochberg Y Controlling the false discovery rate: a practical and powerful approach to multiple testing Royal Stat Soc Ser B (Methodological), J 1995 57 289 300 19 Nueda MJ Tarazona S Conesa A Next maSigPro: updating maSigPro bioconductor package for RNA-seq time series Bioinformatics 2014 30 18 2598 602 10.1093/bioinformatics/btu333 24894503 20 Äijö T Butty V Chen Z Salo V Tripathi S Burge CB Lahesmaa R Lähdesmäki H Methods for time series analysis of RNA-seq data with application to human Th17 cell differentiation Bioinformatics 2014 30 12 113 20 10.1093/bioinformatics/btu274 21 Frazee AC Jaffe AE Langmead B Leek JT Polyester: simulating RNA-seq datasets with differential transcript expression Bioinformatics 2015 31 17 2778 784 10.1093/bioinformatics/btv272 25926345 22 Trapnell C Roberts A Goff L Pertea G Kim D Kelley DR Pimentel H Salzberg SL Rinn JL Pachter L Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks Nature Protocols 2012 7 3 562 78 10.1038/nprot.2012.016 22383036 23 Li H Handsaker B Wysoker A Fennell T Ruan J Homer N Marth G Abecasis G Durbin R The sequence alignment/map format and samtools Bioinformatics 2009 25 16 2078 079 10.1093/bioinformatics/btp352 19505943 24 Anders S Reyes A Huber W Detecting differential usage of exons from RNA-seq data Genome Res 2012 22 10 2008 017 10.1101/gr.133744.111 22722343 25 Graveley BR Brooks AN Carlson JW Duff MO Landolin JM Yang L Artieri CG van Baren MJ Boley N Booth BW The developmental transcriptome of Drosophila melanogaster Nature 2011 471 7339 473 9 10.1038/nature09715 21179090 26 Langmead B Trapnell C Pop M Salzberg S Ultrafast and memory-efficient alignment of short DNA sequences to the human genome Genome Biol 2009 10 3 25 10.1186/gb-2009-10-3-r25 27 Campos-Ortega JA, Hartenstein V. The Embryonic Development of Drosophila Melanogaster. Berlin: Springer. 28 Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Gene ontology: tool for the unification of biology Nature Genetics 2000 25 1 25 9 10.1038/75556 10802651 29 Huang DW Sherman BT Lempicki RA Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists Nucleic Acids Res 2009 37 1 1 13 10.1093/nar/gkn923 19033363

PMC005xxxxxx/PMC5002175.txt

==== Front BMC GenomicsBMC GenomicsBMC Genomics1471-2164BioMed Central London 303210.1186/s12864-016-3032-4Research ArticleSmall RNA and degradome profiling reveals miRNA regulation in the seed germination of ancient eudicot Nelumbo nucifera Hu Jihong hujh2010@163.com 12Jin Jing jinjing1130@whu.edu.cn 1Qian Qian qqian@whu.edu.cn 1Huang Keke kekehuang@whu.edu.cn 1Ding Yi 86 27 68754319yiding@whu.edu.cn 11 State Key Laboratory of Hybrid Rice, College of Life Sciences, Wuhan University, Wuhan, 430072 China 2 State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences, Kunming, 650223 China 26 8 2016 26 8 2016 2016 17 1 68412 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 MicroRNAs (miRNAs) play important roles in plant growth and development. MiRNAs and their targets have been widely studied in model plants, but limited knowledge is available concerning this small RNA population and their targets in sacred lotus (Nelumbo nucifera Gaertn.). Results In this study, a total of 145 known miRNAs belonging to 47 families and 78 novel miRNAs were identified during seed germination using high-throughput small RNA sequencing. Furthermore, some miRNA families which have not yet been reported in monocot or eudicot species were detected in N. nucifera, indicating that these miRNAs was divergence from monocots and core eudicots during evolution. Using degradome sequencing, 2580 targets were detected for all the miRNAs. GO (Gene Ontology) and KEGG pathway analyses showed that many target genes enriched in “regulation of transcription” and involved in “carbohydrate”, “amino acid and energy metabolism”. Nine miRNAs and three corresponding targets of them were further validated by quantitative RT-PCR. Conclusions The results present here suggested that many miRNAs were involved in the regulation of seed germination of sacred lotus, providing a foundation for future studies of sacred lotus seed longevity. Comparative analysis of miRNAs from different plants also provided insight into the evolutionary gains and losses of miRNAs in plants. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-3032-4) contains supplementary material, which is available to authorized users. Keywords miRNATarget genesDegradome sequencingQuantitative qRT-PCRNelumbo nuciferaNational Natural Science Foundation of China (CN)31271310Ding Yi Opening Foundation of State Key Laboratory of Genetic Resources and EvolutionGREKF14-02Hu Jihong http://dx.doi.org/10.13039/501100002858China Postdoctoral Science Foundation2014M562053Hu Jihong issue-copyright-statement© The Author(s) 2016 ==== Body Background MicroRNAs (miRNAs) are approximately 21 nucleotides endogenous non-coding RNAs that play important roles in the regulation of gene expression at the post-transcriptional level, resulting in the cleavage or translation repression of target mRNAs [1, 2]. In plants, miRNA-guided gene regulation is involved in multiple developmental processes, such as floral development [3], male and female sterility [4, 5], leaf growth [6], and RNA metabolism [7]. Most plant miRNAs have perfect or near-perfect sequences complementary with their target mRNA to regulate gene expression at the post-transcriptional level [8]. Thus, identifying miRNAs and their targets, and elucidating their regulatory mechanisms are critical to understand the development processes. Recently, degradome sequencing has been successfully used to analyse the miRNA targets. Combined with small RNA sequencing, the pace of miRNA and their targets identification have been greatly improved in plants [9]. Other studies have also suggested that whole-scale miRNA identification and targets analysis might reveal regulatory networks in plant growth and development [10]. Lately, degradome sequencing has been used to find target genes on Arabidopsis, rice, cotton and so on [11–13]. However, the miRNA targets in sacred lotus (Nelumbo nucifera Gaertn.) have only been predicted by bioinformatics method, and little information has been experimentally validated [14, 15]. Sacred lotus is a perennial aquatic plant with ornamental, edible, medicinal value and phylogenetic importance [16]. In Asia, sacred lotus has been cultivated for more than 7000 years due to its beautiful flowers, edible rhizomes and seeds [17, 18]. Recently, the genome sequences of the sacred lotus variety ‘China Antique’ have been sequenced, which is available for further identifying miRNAs in N. nucifera [19, 20]. As an ancient basal eudicots, sacred lotus seed is one of the oldest directly dated seeds for ~1300 year-old. Previous studies reported that 16 old fruit of ‘China Antique’ were collected from the same lakebed and 84 % of them can be germinated in ~3 days [21, 22]. The proteins which were heat stable and ROS removal were reported to be important for seed longevity [21–23]. ROS adversely affects cellular proteins and enzymes and renders them inactive. Some thermal proteins, including superoxidae dismutase (SOD) and HSP were demonstrated to reduce deleterious ROS accumulation for improving the seed longevity [22, 23]. Because of the sacred lotus seeds remain viable after hundreds of years, resistance and damage repair in seeds must be extraordinarily effective [24]. Therefore, sacred lotus seeds could be an excellent model for studying the adaptive mechanisms of seeds [25]. Seed germination is the first step for plant growth and plays a pivotal role in seedling establishment. It is a complex physiological and biochemical process and previous studies suggested a role for miRNAs in seed germination [26–29]. Using small RNA sequencing, the miRNA expression patterns in rice embryos at 0, 12 and 24 h after imbibition showed that osa-miR159f, osa-miR166l and osa-319b were predominantly expressed in rice seeds [27]. In soybean, gma-miR1530 and gma-miR1536 have been implicated in the conversion between photosynthesis and lipid accumulation by regulating their target genes of transketolase and carboxylase, respcetively [28]. In maize, 115 know miRNAs were identified in the seeds at 24 h after imbibition, and the regulation of the genes targeted by these miRNAs was involved in the early stage of seed germination [29]. However, little miRNAs were known about the seed germination of sacred lotus. Therefore, identification of novel miRNAs and elucidation of their functions in the seed germination will help us to understand the long living of sacred lotus seed. Furthermore, as a basal eudicot, identification of the sacred lotus miRNA and comparison with other angiosperms will aid the study of conservation and evolution of miRNAs in plants. To comprehensively investigate small RNAs and their targets and provide some information for further understanding the miRNA-mediated regulation network in the seed germination of sacred lotus, five small RNA libraries and a degradome library were constructed in this study. The profiling of the miRNAs and their target genes provides some insights into the regulatory pathways governing gene expression during the seed germination, and the data presented here will lay a foundation for future studies of sacred lotus seed longevity. Results Overview of small RNA library sequencing To elucidate the regulatory roles of miRNAs during the seed germination of sacred lotus, five small RNA libraries were constructed from germinating seeds at 0 h, 12 h, 24 h, 36 h and 72 h after imbibition and then sequenced by a Hiseq2,500 platform. After filtering the low-quality reads, including reads <18 nt in length, 5′ adaptor contaminates, reads with polyA, and reads without 3′adaptors, ranging from 10,245,283 clean reads (at 72 h) to 14,412,672 clean reads (at 0 h) were obtained. More than 60 % of the clean reads were successfully mapped to the sacred lotus reference genome (Table 1). Moreover, 3,347,363, 4,084,098, 3,575,988, 4,002,600 and 3,497,201 clean reads were unique mapped to the genome. Other unannotated unique sequences were used for novel miRNAs prediction and further analysis (Table 1).Table 1 Data set summary of sequencing of the five small RNA and one degradome libraries 0 h 12 h 24 h 36 h 72 h Small RNA data Clean reads 14,412,672 14,045,040 10,736,476 12,983,765 10,245,283 Mapped reads 9,362,670(64.96 %) 8,479,342(60.37 %) 6,700,319(62.41 %) 8,378,691(64.53 %) 6,474,276(63.19 %) Unique mapped 3,347,363(100) 4,084,098(100) 3,575,988(100) 4,002,600(100) 3,497,201(100) Match miRNA 3,347(0.10 %) 5,309 (0.13 %) 4,648(0.13 %) 4,803(0.12 %) 3,847(0.11 %) Unannotation 3,047,439(91.04 %) 3,725,514(91.02 %) 3,248,070(90.83 %) 3,633,160(90.77 %) 3,152,027(90.13 %) Total Unique Degradome data Clean Reads 13,852,945(100) 4,362,200(100) Mapped reads 13,787,820(99.53 %) 4,333,854(99.35 %) Transcript Mapped 7,659,047(55.29 %) 2,018,752(46.28 %) Unannotation 6,193,898(44.71 %) 2,343448(53.72 %) In all of the five libraries, the small RNA (sRNA) sequences were ranged from 18 to 28 nt in length, with the majority were 19 to 24 nt. And the 24 nt sRNAs were the most abundant within the total sRNA, which approximately accounted from 20 % to 35 % in all of the five sRNA libraries, followed by 21 nt sRNAs, the typical length of canonical miRNAs (Additional file 1: Figure S1). The common and specific of total and unique small RNA sequences were compared between the adjacent libraries (0 vs12h, 12 h vs 24 h, 24 h vs 36 h and 36 h vs 72 h), showing that more than 75 % of the total sRNAs were shared (Additional file 1: Figure S2). However, only a small fraction (approximately 12.5 % ~ 14 %) of unique sRNA shared in adjacent libraries, suggesting that these small RNAs were low in abundance but highly in diverse. Identification of miRNAs and compared with other plants Conserved miRNA families have been identified in many plant species and have important functions in plant development [29]. In the present study, 145 known miRNAs (belonging to 47 families) were identified (Table 2 and Additional file 2: Table S1). The unannotated small RNA tags were used to predict novel miRNAs and their secondary structures of the precursors were further detected by MFOLD or RNAfold to remove the false positive (Additional file 1: Figure S3). Finally, 78 novel miRNAs were identified in sacred lotus (Additional file 1: Figure S3, Additional file 2: Table S2). The largest miRNA family size identified was Nnu-miR396 comprising 15 members, following by Nnu-miR169 and Nnu-miR393 possessed 12 and 8 members, respectively. Other miRNA families such as Nnu-miR397, Nnu-miR403, Nnu-miR529 and Nnu-miR827 possessed only one member detected in sacred lotus during seed germination of sacred lotus (Table 2 and Additional file 2: Table S1).Table 2 Comparion of the miRNA families identified in Arabidopsis, rice, Amborella and sacred lotus (Nelumbo nucifera) Family Arabidopsis Rice Amborella Sacred lotus miR156 10 12 4 5 miR157 4 0 0 4 miR159 3 6 1 3 miR160 3 6 1 3 miR162 2 2 0 3 miR164 3 6 2 4 miR165 2 0 0 2 miR166 7 13 4 4 miR167 4 10 1 4 miR168 2 2 1 3 miR169 14 18 3 12 miR170 1 0 0 3 miR171 3 9 3 7 miR172 5 4 1 7 miR319 3 2 5 4 miR390 2 1 1 2 miR393 2 2 1 8 miR394 2 1 1 3 miR395 6 25 1 2 miR396 2 8 5 15 miR397 2 2 2 1 miR398 3 2 1 3 miR399 6 11 0 6 miR403 1 0 0 1 miR408 1 1 0 6 miR529 0 2 0 1 miR530 0 1 0 2 miR535 0 1 1 1 miR827 1 1 0 1 miR828 1 0 1 1 miR837 1 0 0 1 miR1432 0 1 0 1 miR2111 2 0 1 2 miR2118 0 17 0 2 miR2275 0 4 0 2 miR2950 0 0 1 2 A comparison of the identified known miRNA families from sacred lotus with other three different plants (Arabidopsis, rice and Amborella) showed that the number of miRNA families widely varied (Table 2) [30, 31]. In sacred lotus, most of the families were represented by from 1 to 8 variants, except for miR396 (15) and miR169 (12). More than 10 variants of miR169 family were also detcted in Arabidopsis and rice (Table 2). When comparing 25 miRNA families of sacred lotus miRNAs to those in 10 other angiosperm species, five families (miR165, miR403, miR828, miR837 and miR2111) were only present in eudicot species. And other two families (miR1432 and miR2275) were found in basal angiosperm and monocots but not detected in core eudicot species (Fig. 1). MiR529 was identified in basal eudicots (Eschscholzia californica and N. nucifera) and most of the monocots, but no homologues were found in core eudicots (Fig. 1).Fig. 1 miRNA distribution in different plant species. The miRNA data were obtained from the miRbase 20 and the present study To validate the predicted novel miRNAs, stem-loop RT-PCR was performed for six novel miRNAs (novel_miR_2, novel_miR_13, novel_miR_29, novel_miR_32, novel_miR_36 and novel_miR_66) with relatively high abundance at all the five stages of seed germination (Fig. 2, Additional file 1: Figure S3 and Additional file 2: Table S2). As a result, all the selected six novel miRNAs were found to be expressed in the seed germination (Fig. 2b).Fig. 2 Potential novel miRNAs identified in this study. a The secondary structure of novel_ miR_36 precursor was predicted by MFOLD. Mature miRNA is highlighted in red. b Stem-loop RT-PCR analysis of the identified novel miRNAs. Six novel miRNAs (novel_miR_2, novel_miR_13, novel_miR_29, novel_miR_32, novel_miR_36, novel_miR_66) were confirmed via stem-loop RT-PCR. The sizes of the obtained PCR products were approximately 60 bp. Marker indicates a 50 bp DNA Ladder Marker Differentially expressed miRNA at different stages of seed germination To compare the different miRNA expression profiles during the seed germination of sacred lotus, the differential expression analysis of the miRNAs was performed between the four stages with 0 h (CK) of germination, based on the normalized read count for each identified miRNA (Additional file 2: Tables S3 and S4). And a total of 107 differentially expressed (DE) miRNAs were found at the four stages during sacred lotus seed germination. Then cluster analysis was conducted to further elucidate the expression patterns of all known miRNAs, based on the criteria (at least one comparison has a fold change value ≥ 2.0 or ≤ −2.0 with a p-value < 0.05) (Fig. 3a). Compared with the 0 h of seed germination, many DE miRNAs were detected in the other four stages (Additional file 2: Table S4). Furthermore, eight DE miRNAs were differentially expressed in all the four RNA libraries, while 26 DE miRNAs were expressed in at least two of the four small RNA libraries, and 17 miRNAs were identified only once in the four samples (Fig. 3b). These results indicated that a larger number of miRNAs were differentially expressed during all five stages of seed germination. Compared with germinating seed at 0 h, a total of 35 DE miRNAs were up-regulated, while 16 miRNAs were down-regulated (Table 3). For instance, Nnu-miR156c-5p and Nnu-miR159a were significantly expressed at 36 h (Table 3). Nnu-miR165a-5p, Nnu-miR166c, Nnu-miR166d and Nnu-miR2111b showed the highest expression levels at 24 h, while Nnu-miR396b-3p showed maximum expression level at 0 h (Table 3). Most of the novel miRNAs showed low expression levels and some of them were only found at one stage during the seed germination. However, novel_mir_13, novel_mir_33 and novel_mir_39 were highly expressed during all the five stages (Additional file 2: Table S4) and stem-loop RT-PCR also confirmed the expression of novel_mir_13 during sacred lotus seed germination (Fig. 2b).Fig. 3 Small RNAs and their expression patterns at each of the five stages during the seed germination of sacred lotus. a Heatmap for the clustering analysis of differentially expressed known miRNAs. The bar represents the scale of the expression levels of the miRNAs. b Venn diagram of the common and specific known miRNAs at four different stages (12 h, 24 h, 36 h and 72 h) compared with 0 h Table 3 Differential expressed known miRNAs during the sacred lotus seed germination Change miRNA 0 h-std 12 h-std 24 h-std 36 h-std 72 h-std Significance Up-regulated Nnu-miR156c-5p 6.68 10.04 10.59 12.26 8.62 * Nnu-miR157a 56.90 100.44 144.83 169.62 69.79 ** Nnu-miR159a 164.04 434.63 1008.33 1065.40 334.58 ** Nnu-miR160b 6.99 10.04 16.09 13.28 11.49 ** Nnu-miR162a-3p 78.52 132.40 218.52 125.68 167.91 ** Nnu-miR165a-5p 8.27 25.26 45.31 32.02 20.12 ** Nnu-miR165a-3p 5100.3 10745.91 18344.31 10534.78 10529.28 ** Nnu-miR166b-5p 0.64 1.52 6.78 6.47 6.57 ** Nnu-miR166c 5106.39 10766.00 18370.99 10560.33 10551.04 ** Nnu-miR166d 2117.83 5231.12 9055.48 5330.06 4818.39 ** Nnu-miR167a-3p 0 0.3 1.27 1.70 0.41 * Nnu-miR2111b 2.86 8.83 8.89 6.47 7.39 ** Nnu-miR319b 103.63 150.05 214.29 149.52 114.13 ** Nnu-miR319d 736.57 1670.66 2534.59 1737.40 1320.68 ** Down-regulated Nnu-miR1030 4.45 3.04 1.69 1.70 1.23 * Nnu-miR170a 33.06 15.52 29.22 29.63 12.73 ** Nnu-miR171g 33.38 16.44 30.07 29.97 12.73 ** Nnu-miR2111a-5p 137.01 98.01 99.10 122.62 55.83 ** Nnu-miR393a-5p 57.86 39.87 42.77 37.13 28.74 ** Nnu-miR393c 57.22 38.65 42.35 36.78 28.33 ** Nnu-miR394a 61.67 19.18 28.37 30.99 22.58 ** Nnu-miR394b 103.32 44.13 52.09 53.13 32.43 ** Nnu-miR396b-3p 8.27 1.83 3.39 1.70 1.23 ** Nnu-miR408a 13.35 6.09 8.05 12.26 5.34 ** Nnu-miR408c 29.56 17.04 20.75 24.18 16.01 ** Nnu-miR530a 4.45 3.04 1.69 1.70 1.23 * "std" represents normalized expression level (Normalised expression, RPM) of miRNAs "*" and "**" represnent significant differentially expression between 0 h and other hours, *p < 0.05, **p < 0.01 Degradome sequencing analysis In this study, a degradome library was constructed from the germinating seed of sacred lotus, followed by degradome sequencing. After discarding the low quality sequences, a total of 13,852,945 clean reads with 4,362,200 unique clean reads was remained. And 99.53 % (13,787,820) of the total sequences, corresponding to representing 99.35 % (4,333,854) of the unique sequences, were mapped to the reference genome to identify the fragments of degraded mRNAs. Cleavage of target mRNAs of the miRNAs were identified by degradome sequencing (Fig. 4 and Additional file 1: Figure S4). Generally, the target genes of kown miRNAs were mostly transcription factors [32], and similar results were also obtained for the novel miRNAs (Additional file 2: Tables S5 and S6). For example, both Nnu-miR159b and Nnu-miR319c targeted mRNA encoding the MYB transcription factor. Furthermore, Nnu-miR156a, Nnu-miR160a, Nnu-miR160a-5p and Nnu-miR169a targeted mRNA encoding squamosa promoter-binding-like protein 17, auxin response factor 18, NAC transcription factor and nuclear transcription factor Y subunit A-10, respectively (Fig. 4, Additional file 1: Figure S4 and Additional file 2: Table S5). The targets of novel miRNAs were also identified by degradome sequencing and many of them were transcription factors, including GAMYB (novel_mir_13), MYB (novel_mir_32), NAC (novel_mir_36) and TCP (novel_mir_27) (Additional file 2: Table S6).Fig. 4 Target plots (t-plots) of six identified known miRNA targets using degradome sequencing. The red lines indicate signatures consistent with miRNA-directed cleavage. a Nnu-miR156a slicing NNU_02903-RA at nt 1107(SPL17: Squamosa promoter-binding-like protein 17). b Nnu-miR159b slicing NNU_10896-RA at nt 1284 (GAM1: Transcription factor GAMYB). c Nnu-miR160a slicing NNU_12564-RA at nt 1549 (ARF18: Auxin response factor 18). d Nnu-miR169a slicing NNU_09768-RA at nt 1642 (NFYA10: Nuclear transcription factor Y subunit A-10). e Nnu-miR319c slicing NNU_00026-RA at nt 2081(TCP2: Transcription factor TCP2). f Nnu-miR393a-5p slicing NNU_01291-RA at nt 1689 (TIR1: Protein TRANSPORT INHIBITOR RESPONSE 1). "alignment score" is the score for mismatch. Score = 0 represents perfect match and G:U = 0.5 Functional annotation of the target genes To further perceive the putative function of the predicted target genes, Gene Ontology (GO) analysis was performed. And twenty-five different biological processes, fifteen different cellular components and ten different molecular functions were predicted (Additional file 1: Figure S5 and Additional file 2: Table S7). The most significantly enriched GO terms were involved in the three main categories, “protein phosphorylation”, “nucleus”, and “ATP binding”, followed by “regulation of transcription”, “membrane” and “DNA binding” (Fig. 5a and Additional file 1: Figure S5). These results suggested that regulation of transcription played a vital role in the process of the seed germination of sacred lotus. Moreover, all the targets could be used for KOG classifications which were classified into 23 groups (Additional file 1: Figure S6). The most dominative group was the “general function prediction only”, followed by the “transcription” (Additional file 1: Figure S6).Fig. 5 GO analysis and KEGG pathways of the target genes. a GO analysis of the target genes for differentially expressed miRNAs. The colouring of the p-values indicates the significance of the rich factor. The circle indicates the target genes that are involved, and the size is proportional to the gene numbers. The Y-axis represents the name of enrichment GO terms. The X-axis represents rich factor. The "Rich factor" was calculated by the number of genes mapped to the GO term divided by the number of all genes in the input list. b KEGG pathways of the target genes for differentially expressed miRNAs In addition, KEGG analysis was further conducted to elucidate the biological pathways of the DE miRNA target genes, which were classified into five groups including “organismal system”, “metabolism”, “genetic information processing”, “environment information processes” and “cellular processes” (Fig. 5b and Additional file 2: Table S8). The categories of “carbohydrate metabolism”, “amino acid metabolism”, “energy metabolism” and “biosynthesis of other secondary metabolism” were the most enriched pathways, which were all in the group of “metabolism”, suggesting that the miRNA regulating the expression of metabolic genes played crucial roles in the seed germination. Quantitative RT-PCR validation of miRNAs and their target genes In this study, real time quantitative RT-PCR (qRT-PCR) was performed to confirm the expression levels of nine significantly DE miRNAs and three respective predicted target genes, obtained from the high-throughput sequencing (Fig. 6 and Additional file 2: Table S5). Three miRNAs (Nnu-miR393a-5p, Nnu-miR396b-5p and Nnu-miR397) showed the similar expression pattern in terms of read abundance during the five stages of seed germination (Fig. 6 and Additional file 2: Table S3). For instance, Nnu-miR156a showed the highest expression levels at 0 h, and then gradually declined from 12 h to 72 h (Fig. 6g). Moreover, the expression levels of Nnu-miR393a-5p, Nnu-miR394a, Nnu-miR396b-5p, Nnu-miR397 and Nnu-miR2111a were up-regulated and maximized at 24 h and then dramatically decreased to relatively low levels at 36 h and 72 h (Fig. 6). Some of the miRNAs showed the minor discrepancies between sequencing data and the qPCR results, which partially due to the differences in the sensitivity, specificity and algorithm between the two techniques.Fig. 6 Validation of differentially expressed miRNAs and some of their targets during the seed germination in sacred lotus by qRT-PCR a Nnu-miR157a, b Nnu-miR168a-5p, c Nnu-miR393A-5p, d Nnu-miR396b-5p, e Nnu-miR397, f Nnu-miR2111a,  g Nnu-miR156a, h Nnu-miR319c, i Nnu-miR394a,  j Target of Nnu-miR156a (NNU_02036-RA, SPL16), k Target of Nnu-miR319c (NNU_00026-RA, TCP2), l Target of Nnu-miR394a (NNU_22183-RA, FBX) The expression levels of some of the predicted target genes (SPL16, TCP2 and FBX) were selected for validation using qRT-PCR, and multiple of target genes showed contrasting expression patterns with their corresponding miRNAs (Fig. 6 and Additional file 2: Table S5). For example, TCP2 (NNU_00026-RA) was the predicted target gene of Nnu-miR319c which expression level was sharply declined to the lowest level at 24 h, then gradually increased and remained at a relatively high level at 72 h. On the contrary, the target (TCP2) of the Nnu-miR319c was increased at the 24 h and then gradually declined at 72 h (Fig. 6h and k). Moreover, Nnu-miR156a also showed a negative relationship with its target SPL16 (Fig. 6g and j). The results suggested that some miRNAs might be invovled in the regulation of seed germination of sacred lotus [33]. However, the expression of FBX showed partially negative relationship with Nnu-miR394a and both of them showed high expression levels at 24 h (Fig. 6i and l). Thus, it is likely that the inverse relationship between miRNAs and their targets might be limited during specific stages of germination, because miRNAs are not the only regulatory factors affecting their targets [34]. Since miRNAs have been shown to also act by translation inhibition in plants, the fact that the RNA expression level of a target is not negatively correlated with miRNA expression level does not mean that the target is not real. Putative microRNA editing in sacred lotus The miRNA editing patterns showed a significant discrepancy among the germinating seeds at 0 h, 12 h, 24 h, 36 h and 72 h (Fig. 7a,b). For example, the most dominant nucleotide substitution type were at 12 h and 24 h with T to C (more than 35 % in both libraries); while the most common nucleotide substitutions were at 0 h and 36 h with A to G and C to T (Fig. 7b). At the nucleotide positions of 4 and 11, the miRNA editing events showed the specifically lowest number (less than 5 number of editing) (Fig. 7a, b and Additional file 2: Table S9). These results were consistent with previous reported miRNA editing in rice grain filling [35].Fig. 7 Analysis of miRNA editing events. a Summary of the nucleotide substitutions positions among miRNAs. b Summary of the nucleotide substitution types among miRNAs. c Validation of the editing sites in miRNAs obtained from deep sequencing by Sanger sequencing. The edited sites are highlighted with black frames. The upper is sequences of genomic DNA (gDNA), and the lower is sequences of cDNA To validate the occurrence of miRNA editing events in sacred lotus, two editing types (T to A, and U to C) were examined in Nnu-miR160a and Nnu-miR319c, respectively (Fig. 7c). Using PCR amplification, precursor miRNA sequences from genomic DNA and mature miRNA sequences from cDNA were cloned and sequenced, respectively. Comparison analysis of the sequences further confirmed that the miRNA editing events really existed in sacred lotus (Fig. 7c). Discussion As non-coding RNAs, miRNAs play key roles in plant growth and development. Many studies of miRNAs have been reported in model plants such as rice, Arabidopsis and maize and so on. Sacred lotus is a basal eudicot with ornamental, agricultural, medicinal importance. Previous studies on miRNAs in sacred lotus were only in leaves or flowers and for bioinformatics searching without experimentally validation [14, 15]. Phased small interfering RNAs (phasiRNAs) which regulated by miRNAs were also identified in Chinese scared lotus by bioinformatics analysis [36]. However, the number of miRNAs identified in sacred lotus is very few and little is known about the miRNAs involved in the seed germination of N. nucifera. In this study, 148 known miRNAs (belonging to 47 families) and 78 novel miRNAs were identified in sacred lotus via deep sequencing and the results were validated by experimental approaches. Most of the sacred lotus miRNAs identified in the study were present in monocots and at least one of the core eudicot species, indicating their existence in the ancestor of the two large angiosperm clades. Eleven miRNA families were conserved among monocots, basal angiosperm, basal eudicots and core eudicots, suggesting that these miRNAs were involved in important gene regulation pathways in plants (Fig. 1). However, some miRNAs might become lost or evolve within some angiosperm lingeages [30]. MiR165 and miR828 were found in Arbidopsis and two basal eudicot species, but not reported in monocots, while miR529 was present in monocots but not found in the core eudicot species (Fig. 1). That may be because the last common ancestor of monocots and eudicots likely had miR529, with the gene appearing near the speciation of that last common ancestor, but the gene was lot in the ancestor of core eudicots [30]. In addition, most of the target mRNAs of miR165 and miR166 as well as miR156 and miR529 were the same, respectively [33, 34]. The loss of miR165 or miR529 in monocot or core eudicot species indicated the distinct evolution of these species. Moreover, miR1432 and miR2275 were only present in monocots and N. nucifera, suggesting that they might be ancient miRNAs in angiosperm. Further investigation of these miRNAs and their targets would give some insights into the functional divergence and evolution in angiosperm. Using small RNA sequencing, many novel miRNAs with low abundance are discovered from various developmental stages. In this study, 78 novel miRNAs were identified and some of them were validated in the sacred lotus during seed germination (Fig. 2). The expression profiles of these novel potential miRNAs during the different stages might provide valuable information for seed germination and seed longevity in sacred lotus. In addition, a large number of miRNA editing events were found in the seed germination of sacred lotus. Analysis of the nucleotide positions of editing sites showed that the 4 and 11 sites had less editing frequency and positions of 6 and 7 sites had more editing frequency (Fig. 7a). The results were in accordance with rice grains, while in rice young panicles, the two sites had more editing events [5, 35]. These results indicated that miRNA editing was diversity in different tissues and species. The computational analysis using the cleaveland pipeline revealed a total of 2580 targets for these known and novel miRNAs by degradome sequencing in the present study. Our results showed that most miRNA families had multiple target genes, while several miRNAs from different miRNA families shared some target genes. GO analysis showed that these targets mainly involved in “protein phosphorylation” and “regulation of transcription” (Additional file 1: Figure S5). For KEGG pathway, “carbohydrate metabolism” was enriched in the germinating seeds of sacred lotus (Fig. 5b), indicating that many carbohydrate metabolism related genes were regulated by miRNAs during the seed germination in sacred lotus. In many plant species, plant hormones including Gibberellins (GAs) and ABA, can promote and inhibit seed germination. And some studies also indicated that cross talk between auxin and GA or ethylene might affect seed germination [29, 37]. In Arabidopsis thaliana, overexpression of MIR160 (35S:MIR160) resulted in reduced ABA sensitivity during germination. Transcriptome analysis of germinating ARF10 and the transgenic plants expressing ARF10 with silent mutations (mARF10) revealed that many ABA-responsive genes were overexpressed in germinating mARF10 seeds [38]. Furthermore, miR160 has been shown to regulate the expression of ARF17 which is another auxin response transcription factor [39]. In this study, the Nnu-miR160 also targeted the ARF17 (NNU_16091-RA) and the expression of the three miRNAs of the miR160 families was up-regulated at 24 h during the seed germination (Additional file 2: Tables S3 and S5). These results indicated that overexpression of miR160 might promote seed germinating by decreasing the ABA levels in the seed germination of sacred lotus. Auxin signal pathways are also regulated by miR393 which targets the auxin receptor TIR1 in Arabidopsis [40]. Two TIR1 gene homologues (NNU_01291-RA and NNU_06904-RA) were also detected for the targets of Nnu-miR393 in our degradome dataset (Additional file 1: Figure S4 and Additional file 2: Table S5). The expression level of Nnu-miR393a-5p was also validated by qRT-PCR (Fig. 6c). The down-regulation of some components in auxin signal transduction by miRNAs might be a regulatory step to decrease ABA sensitivity in mature seeds and to switch to the germination mode [29]. MiR159 has been reported to target GAMYB transcription factors which interact with GA-response elements [4]. In Arabidopsis seeds, miR159 regulates the abundance of MYB33 and MYB101 mRNAs during germination in response to ABA. These two MYB transcription factors were positive regulators of ABA responses and overexpressed miR159 would render plants hyposensitive to ABA [41, 42]. In addition, the two target genes of Nnu-miR159 also participate in aleurone vacuolation which was in GA-induced pathways during seed germination. In this study, Nnu-miR159 targeted three GAMYB genes and their expression were highly up-regulated at 36 h during seed germination (Additional file 2: Table S3). The Nnu-miR156 family members targeted 7 Squamosa Promoter Binding proteins like (SPL) plant specific transcription factors in the present study (Additional file 1: Figure S4 and Additional file 2: Table S5). The expression levels of Nnu-miR156a and its target gene SPL16 (NNU_02036-RA) were also validated (Fig. 6g and j). In plants, SPLs were reported to be involved in diverse developmental processes including leaf development, phase change and flowering in plants [43]. Overexpression of miR156 in Arabidopsis and rice suppressed the expression of SPL genes and reduced apical dominance, delayed flowering time and affected the juvenile to adult transition [43–45]. MiR156 was also reported to be involved in seed germination in soybean and Lotus japonicas [46, 47]. However, unlike the maize [43], the expression level of miR156 in sacred lotus was down-regulated after 72 h of the seed germination (Additional file 2: Table S2). That may be a factor for the long living of sacred lotus seed and viable for nearly thousands years. Nuclear factor Y (NF-Y) was documented to affect the seed development and was targeted by miR169 in many plants. In the degradome dataset in the present study, NFYA2 (NNU_25577-RA), NFYA3 (NNU_11871-RA) and NFYA10 (NNU_09768-RA) were the target genes of miR169 in sacred lotus (Additional file 1: Figure S4 and Additional file 2: Table S5). In Arabidopsis, overexpression of the target of miR169 (NF-YA5) resulted in ABA hypersensitivity during seed germination [48]. Thus, up-regulation of most of the miR169 genes in sacred lotus might reduce the ABA responsiveness during seed germination (Additional file 2: Table S2). In plants, seed germination is promoted by appropriate environmental conditions including light, temperature, and nutrient availability [49]. Reactive oxygen species (ROS) are largetly generated during the seed germination and reducing the ROS and minimizing cell damage might affect the seed germination. MiR408 which targets L-ascorbate oxidase (NNU_10795-RA) might scavenge the oxidative species produced during stress. In the present study, all the miRNAs of Nnu-miR408 family were significant differentially expressed, indicating their roles in seed germination (Table 3 and Additional file 2: Table S3). PPR proteins that regulate gene expression in mitochondria and chloroplasts were also regulated by some miRNAs, including Nnu-miR168, Nun-miR2673, novel_mir_7 and novel_mir_40. These miRNAs might affect the pathways of photosynthesis and lipid accumulation during the seed germination [50]. Taken together, we proposed potential roles for miRNAs playing during the early stages of seed germination in the sacred lotus (Fig. 8). These miRNA mediated the regulation of gene expression in many processes, including morphological changes, developmental process, metabolism and responsive to stress process (Fig. 8). These results indicated that diverse and complex miRNAs were involved in the seed germination. Further functional research and target analyses of the conserved and novel miRNAs could provide additional clues to the different regulation of gene expression during seed germination. The results in this study will lay a foundation for further investigating of long living of sacred lotus.Fig. 8 The potential regulatory network for miRNAs in the seed germination of sacred lotus Conclusions In this study, combined with small RNA sequencing and degradome sequencing, 145 known miRNAs and 78 novel miRNAs with 2580 targets were identified during seed germination of sacred lotus. GO and KEGG pathway analyses showed that many target genes enriched in regulation and metabolism. Compared with other angiosperm species, some miRNA families were present or absent in monocots and core eudicots, indicating the functional divergence and evolution in angiosperm. These results suggested that many miRNAs were involved in the regulation of seed germination of sacred lotus and provided insights into the evolutionary gains and losses of miRNAs in plants. Methods Plant materials Sacred lotus (N. nucifera Gaertn.) seeds used here are maintained in our laboratory. The mature seeds of 3 years old were germinated in distilled water under 16-h light photoperiod at 28 °C in a phytotron. The seeds at 0, 12, 24, 36 and 72 h after imbibition was collected and immediately frozen in liquid nitrogen, then stored in -80 °C until further use. RNA isolation and small RNA library construction Total RNA was extracted from germinating seeds at 0, 12, 24, 36 and 72 h after imbibition using TRIZOL reagent (TianGen, China) according to the manufacturer’s instructions. At least five germinating seeds were mixed for each sample. RNA-free Dnase I (Promega) was then used to remove DNA contamination from total RNA for 30 min at 37 °C. RNA quantity was determined by a Qubit Fluometer. The RNA purity and integrity was detected by Agilent 2100 Bioanalyzer. Small RNAs of 18 ~ 30 nt were purified from the 5 μg of total RNA by 15 % polyacrylamide gel electrophoresis (PAGE) as described previously [51]. A pair of Solexa adaptors was ligated to the 5′ and 3′ termini of the small RNAs. And then reverse transcription was performed by Super-Script II Reverse Transcriptase (Invitrogen) and amplified using 15 PCR cycles to produce sequencing library. The final quality of the cDNA libraries were examined with the Agilent 2100 Bioanalyzer. The small RNA fragments were finally sequenced by Hiseq2, 500 platforms. Bioinformatics analysis of the miRNAs The raw reads from the Illumina sequencing were first filtered to remove low-quality reads which including reads in length <18 nt, 5′ adaptor contaminates, reads with polyA, and reads without 3′ adaptor. And then adaptor sequences were trimmed to get clean reads. All unique sequences were aligned against the sacred lotus genome (http://lotus-db.wbgcas.cn/) to map the sequences using SOAP [52, 53]. Subsequently, the clean sequences were used to search the RepeatMasker and Rfam database to annotate repeats, rRNA, tRNA, scRNA, snoRNA and snRNA. After removing the sequences belonging to rRNA, tRNA, scRNA, snoRNA, snRNA and repeat sequence tags, the remaining sequences were used to BLAST against miRbase 20.0 (http://www.mirbase.org) to identify conserved miRNA [54]. The remaining unannotated small RNA tags were used to predict novel miRNAs using the miRNA prediction algorithm MIREAP (http://sourceforge.net/projects/mireap). This program was utilized to identify novel miRNA candidates by exploring the secondary structures, Dicer cleavage sites and the minimum free energy (MFE) of the unannotated small RNA tags that could be mapped to the genome. The parameters for identifying potential miRNAs or pre-miRNAs as described previously [5, 33]. And the secondary structures of potential precursor miRNAs were constructed by MFOLD 3.2 [55]. To identify the expression patterns of miRNAs among the five different stages, the read count of each identified miRNA was normalized to reads per million (RPM) using the following formula: normalized expression = count of miRNA/total count of clean reads × 10-6. If one miRNA have no reads in a sample, the expression was arbitrarily set to 0.01 for further calculation [56]. Fold change of log2 scale value and P value were calculated using the formula as described [57]. The expression of |fold change| >2 and P value <0.01 was as significant differential expression. All the significantly differential expression profiles were clustered into different groups by Cluster 3.0 with default parameters. Degradome library construction and data analysis The degradome library was constructed as described previously [13]. Poly (A) RNA was extracted from total RNA of mixture of the five samples (1 μg of total RNA for each) using the Oligotex kit (Qiagen). With 5′-monophosphates, polyadenylated transcripts were ligated to RNA adapters which consisted of a Mme I recognition site at its 3′ end. Then, first-strand cDNA was reverse-transcribed using oligo d (T) and amplified for five PCR cycles. The PCR product was purified and digested with Mme I and was then ligated to a 3′ double DNA adapter. After amplifying 18 PCR cycles, the PCR product was gel-purified for deep sequencing. The sliced miRNA targets were identified according to the CleaveLand pipeline [13]. The potential targets were as followed: alignments with scores less than four (G: U pairs scored 0.5), and no mismatches were found at the site between the 10th and 11th nucleotides of the miRNAs. The target genes were annotated using the information from Lotus-DB (http://lotus-db.wbgcas.cn/) [47]. The functional classification and pathways based on the three database: KOG (Clusters of Orthologous Groups of proteins), KO (KEGG Ortholog database) and GO (Gene Ontology), using BLAST with a cutoff E-value of 10−5. GO enrichment analysis was implemented by the TopGO R package and GO terms with corrected FDR < 0.05 were considered significantly enrichment [58]. Identification of potential miRNA editing sites The miRNA editing sites located on the known miRNA sequences were identified according the method as described previously [35]. The RNA editing level was calculated by the ratio of the reads support the mismatch in the site to the total reads detected on this site. The common and specific miRNAs with RNA editing during the different time points were also analyzed. Genomic DNA and total RNA were isolated from germinating seeds of sacred lotus to validate miRNA editing. Precursor miRNA sequences were amplified from DNA, and the corresponding mature miRNA sequences were also amplified from cDNA which was reverse-transcribed using stem-loop RT-PCR. Then the PCR products were purified, cloned and sequenced with a minimum of six clones for DNA products as well as at least twelve clones for miRNA PCR products. All primers used in this study are listed in Additional file 2: Table S10. Detection of potential miRNA using stem-loop RT-PCR and qRT-PCR Total RNA was extracted from sacred lotus germinating seeds at different stages. Then, RNase-free DNase I (Promega) was used to remove the DNA contamination by incubating at 37 °C for 30 min. For each miRNA, approximately 2 μg of total RNA was reverse-transcribed using miRNA-specific stem-loop primers in a 20 μl of reaction volume using a Fermentas RevertAid First Strand cDNA Synthesis Kit (Fermentas, USA) [59]. The reactions were incubated for 30 min at 16 °C, followed by pulsed RT of 60 cycles at 30 °C for 30s, 42 °C for 30s and 50 °C for 1 s and finally the reactions were terminated at 70 °C for 5 min [59]. The cDNA for the miRNA target was generated using 2 μg of total RNA and OligodT18 primer with the RevertAid First Strand cDNA Synthesis Kit (Fermentas, USA). The novel miRNAs were cloned using specific primers and the PCR products were detected by gel electrophoresis. All the primers were listed in Additional file 2: Table S10. Real time qRT-PCR analysis of the miRNA and their targets was performed using the FastStart Universal SYBR Green Master Mix (Roche) on the StepOne plus PCR platform (Applied Biosystems). The qRT-PCR reactions were conducted with the following protocol: 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 56 °C for 30s and 72 °C for 15 s. Scared lotus NnEF1a (GI: 226897264) was used as an endogenous control. To avoid non-specific amplification, melting curve was carried out for each PCR product. The expression level of the miRNAs and their targets in different samples were calculated by comparative 2-△△CT method [60]. Availability of supporting data All the sequencing data were deposited in the NCBI Short Read Archive (SRA) database under the accession number SRP070743. Additional files Additional file 1: Figure S1. Size distribution of small RNAs in all of the five libraries from germinating seeds at 0 h, 12 h, 24 h, 36 h and 72 h. Figure S2. Venn chart for total sRNA (A) and unique sRNAs (B) between the adjacent pairwise libraries. Figure S3. The secondary structures of novel miRNA precursors in sacred lotus. The mature miRNAs are in red and miRNA* in blue. ("." represent baase mismatches, "("represent base matches). Figure S4. Target plats (T-plots) of identified known and novel miRNA targets using degradome sequencing. The red lines indicate signatures consistent with miRNA-directed cleavage. Figure S5. GO classification of target transcripts for all identified miRNAs in sacred lotus. Figure S6. KOG function classification of the miRNA targets in sacred lotus. (PDF 3161 kb) Additional file 2: Table S1. Detailed information of the known miRNAs identified in sacred lotus during seed germination. Table S2. Detailed information of the novel miRNAs identified from sacred lotus during seed germination. Table S3. The expression level of all known miRNAs of the germinating seeds at 0 h, 12 h, 24 h, 36 h and 72 h. Table S4. The significantly differential expression level of the known miRNAs of the germinating seeds at 12 h, 24 h, 36 h and 72 h in comparison with 0 h. Table S5. The targets of known miRNA identified in sacred lotus during the seed germination. Table S6. The targets of novel miRNA predicted in sacred lotus during the seed germination. Table S7. GO enrichment analysis for all the target genes. Table S8. KEGG pathway enrichment analysis for the targets of differentially expressed miRNAs. Table S9. The common and specific miRNA editing in sacred lotus during the seed germination. Table S10. Primers of miRNAs and targets in sacred lotus for qRT-PCR and miRNA editing in this study. (XLSX 203 kb) Abbreviations ABAAbscisic acid ARFAuxin response factor DEDifferentially expressed FBXF-Box GAGibberellins GOGene Ontology HSPHeat shock protein KEGGKyoto Encyclopedia of Genes and Genomes KOGEukaryotic Orthologous Groups MFEMinimum free energy miRNAMicroRNA NACNAM, ATAF, and CUC NF-YNuclear factor Y PAGEPolyacrylamide gel electrophoresis PCDProgrammed cell death PPRPentatricopeptide repeat qRT-PCRQuantitative reverse transcription polymerase chain reaction ROSReactive oxygen species RPMReads per million SOAPShort oligonucleotide alignment program SPLSQUAMOSA promoter-binding protein TCPTEOSINTE BRANCHED/CYCLOIDEA/PCF Funding This research is financially supported by National Natural Science Foundation of China (31271310), Opening Foundation of State Key Laboratory of Genetic Resources and Evolution (GREKF14-02) and the China Postdoctoral Science Foundation (2014 M562053). Availability of data and materials All the sequencing data were deposited in the NCBI Short Read Archive (SRA) database under the accession number SRP070743. And the relevant supporting data can be found within the article and its additional files. Authors’ contributions YD and JHH conceived and designed the experiments. JHH, JJ and QQ performed the experiments. JHH and KKH analyzed the data. YD supervised the progress of the project. JHH, JJ and YD wrote the paper. All authors have read and approved this 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. Bartel DP MicroRNAs: Genomics, Biogenesis, Mechanism, and Function Cell 2004 116 281 297 10.1016/S0092-8674(04)00045-5 14744438 2. Voinnet O Origin, biogenesis, and activity of plant microRNAs Cell 2009 136 669 87 10.1016/j.cell.2009.01.046 19239888 3. Chen XM A microRNA as a translational repressor of APETALA2 in Arabidopsis flower development Science 2004 303 2022 5 10.1126/science.1088060 12893888 4. Millar AA Gubler F The Arabidopsis GAMYB-like genes, MYB33 and MYB65, are microRNA-regulated genes that redundantly facilitate anther development Plant Cell 2005 17 705 21 10.1105/tpc.104.027920 15722475 5. Yan J Zhang HY Zheng YZ Ding Y Comparative expression profiling of miRNAs between the cytoplasmic male sterile line Meixiang A and its maintainer line Meixiang B during rice anther development Planta 2015 241 109 23 10.1007/s00425-014-2167-2 25228384 6. Chuck G Cigan AM Saeteurn K Hake S The heterochronic maize mutant Corngrass1 results from overexpression of a tandem microRNA Nat Genet 2007 39 544 9 10.1038/ng2001 17369828 7. Vaucheret H Vazquez F Crete P Bartel DP The action of ARGONAUTE1 in the miRNA pathway and its regulation by the miRNA pathway are crucial for plant development Gene Dev 2004 18 1187 97 10.1101/gad.1201404 15131082 8. Hu JH Sun LL Zhu ZX Zheng Y Xiong W Ding Y Characterization of conserved microRNAs from five different cucurbit species using computational and experimental analysis Biochimie 2014 102 137 44 10.1016/j.biochi.2014.03.002 24657600 9. Gutierrez L Mongelard G Flokova K Pacurar DI Novak O Staswick P Auxin controls Arabidopsis adventitious root initiation by regulating jasmonic acid homeostasis Plant Cell 2012 24 2515 27 10.1105/tpc.112.099119 22730403 10. Vidal EA Araus V Lu C Parry G Green PJ Coruzzi GM Nitrate-responsive miR393/AFB3 regulatory module controls root system architecture in Arabidopsis thaliana Proc Natl Acad Sci U S A 2010 107 4477 82 10.1073/pnas.0909571107 20142497 11. German MA Pillay M Jeong DH Hetawal A Luo SJ Janardhanan P Global identification of microRNA-target RNA pairs by parallel analysis of RNA ends Nat Biotechnol 2008 26 941 6 10.1038/nbt1417 18542052 12. Zhou M Gu LF Li PC Song XW Wei LY Degradome sequencing reveals endogenous small RNA targets in rice (Oryza sativa L. ssp. indica) Front Biol 2010 5 67 90 10.1007/s11515-010-0007-8 13. Liu N Tu LL Tang WX Gao WH Lindsey K Zhang XL Small RNA and degradome profiling reveals a role for miRNAs and their targets in the developing fibers of Gossypium barbadense Plant J 2014 80 331 44 10.1111/tpj.12636 25131375 14. Zheng Y Jagadeeswaran G Gowdu K Wang N Li SH Ming R Genome-wide analysis of microRNAs in sacred lotus, Nelumbo nucifera (Gaertn) Trop Plant Biol 2013 6 117 130 10.1007/s12042-013-9127-z 15. Pan L Wang XL Jin J Yu XL Hu JH Bioinformatic identification and expression analysis of Nelumbo nucifera microRNA and their targets Appl Plant Sci 2015 3 9 1500046 16. Wu ZH Gui ST Quan ZW Pan L Wang SZ Ke WD A precise chloroplast genome of Nelumbo nucifera (Nelumbonaceae) evaluated with Sanger, Illumina MiSeq, and PacBio RS II sequencing platforms: insight into the plastid evolution of basal eudicots BMC Plant Biol 2014 14 289 10.1186/s12870-014-0289-0 25407166 17. Pan L Quan ZW Hu JH Wang GY Liu SN He Y Genetic diversity and differentiation of lotus (Nelumbo nucifera ) accessions assessed by simple sequence repeats Ann Appl Biol 2011 159 428 441 10.1111/j.1744-7348.2011.00509.x 18. Yang M Xu L Liu YL Yang PF RNA-Seq Uncovers SNPs and Alternative Splicing Events in Asian Lotus (Nelumbo nucifera ) PLoS One 2015 10 e0125702 10.1371/journal.pone.0125702 25928215 19. Wang Y Fan GY Liu YM Sun FM Shi CC Liu X The sacred lotus genome provides insights into the evolution of flowering plants Plant J 2013 76 557 567 10.1111/tpj.12313 23952714 20. Ming R VanBuren R Liu YL Yang M Han YP Li LT Genome of the long-living sacred lotus (Nelumbo nucifera Gaertn.) Genome Biol 2013 14 R41 10.1186/gb-2013-14-5-r41 23663246 21. Shen-Miller J Turek J William SJ Tholandi M Yang M Centuries-Old Viable Fruit of Sacred Lotus Nelumbo nucifera Gaertn var. China Antique Trop Plant Biol 2013 6 53 68 10.1007/s12042-013-9125-1 22. Shen-Miller J Berger R Exceptional seed longevity and robust growth: ancient Sacred Lotus from China Am J Bot 1995 82 1367 1380 10.2307/2445863 23. Kaur H Petla BP Kamble NU Singh A Rao V Salvi P Ghosh S Majee M Differentially expressed seed aging responsive heat shock protein OsHSP18.2 implicates in seed vigor, longevity and improves germination and seedling establishment under biotic stress Front Plant Sci 2015 6 713 26442027 24. Shen-Miller J Schopf JW Harbottle G Cao RJ Ouyang S Zhou KS Long-living lotus: germination and soil {gamma}-irradiation of centuries-old fruits, and cultivation, growth, and phenotypic abnormalities of offspring Am J Bot 2002 89 236 47 10.3732/ajb.89.2.236 21669732 25. Chu P Chen HH Zhou YL Li Y Ding Y Jiang LW Proteomic and functional analyses of Nelumbo nucifera annexins involved in seed thermotolerance and germination vigor Planta 2012 235 1271 88 10.1007/s00425-011-1573-y 22167260 26. Han C Yang PF Studies on the molecular mechanisms of seed germination Proteomics 2015 15 1671 1679 10.1002/pmic.201400375 25597791 27. He DL Wang Q Wang K Yang PF Genome-Wide Dissection of the MicroRNA Expression Profile in Rice Embryo during Early Stages of Seed Germination PLoS One 2015 10 12 e0145424 10.1371/journal.pone.0145424 26681181 28. Song QX Liu YF Hu XY Zhang WK Ma B Chen SY Identification of miRNAs and their target genes in developing soybean seeds by deep sequencing BMC Plant Biol 2011 11 5 10.1186/1471-2229-11-5 21219599 29. Wang L Liu H Li D Chen H Identification and characterization of maize microRNAs involved in the very early stage of seed germination BMC Genomics 2011 12 154 10.1186/1471-2164-12-154 21414237 30. Barakat A Wall K Leebens‐Mack J Wang JYJ Carlson JE dePamphilis CW Large‐scale identification of microRNAs from a basal eudicot (Eschscholzia californica ) and conservation in flowering plants Plant J 2007 51 6 991 1003 10.1111/j.1365-313X.2007.03197.x 17635767 31. Albert VA Barbazuk WB Der JP Leebensmack J Ma H Palmer JD The Amborella genome and the evolution of flowering plants Science 2013 342 6165 1241089 10.1126/science.1241089 24357323 32. Jones-Rhoades MW Bartel DP Bartel B MicroRNAs and their regulatory roles in plants Annu Rev Plant Biol 2006 57 19 53 10.1146/annurev.arplant.57.032905.105218 16669754 33. Bi F Meng X Ma C Yi G Identification of miRNAs involved in fruit ripening in Cavendish bananas by deep sequencing BMC Genomics 2015 16 776 10.1186/s12864-015-1995-1 26462563 34. Li T Ma L Geng YK Hao CY Chen XH Zhang XY Small RNA and degradome sequencing reveal complex roles of miRNAs and their targets in developing wheat grains PLoS One 2015 10 e0139658 10.1371/journal.pone.0139658 26426440 35. Yi R Zhu ZX Hu JH Qian Q Dai JC Ding Y Identification and expression analysis of microRNAs at the grain filling stage in rice (Oryza sativa L.) via deep sequencing PLoS One 2013 8 e57863 10.1371/journal.pone.0057863 23469249 36. Zheng Y Wang SP Sunkar R Genome-wide discovery and analysis of phased small interfering RNAs in Chinese sacred lotus PLoS One 2014 9 12 e113790 10.1371/journal.pone.0113790 25469507 37. Das SS Karmakar P Nandi AK Sanan-Mishra N Small RNA mediated regulation of seed germination Front Plant Sci 2015 6 828 26528301 38. Liu PP Montgomery TA Fahlgren N Kasschau KD Repression of AUXIN RESPONSE FACTOR10 by microRNA160 is critical for seed germination and post-germination stages Plant J 2007 52 133 46 10.1111/j.1365-313X.2007.03218.x 17672844 39. Mallory AC Bartel DP Bartel B MicroRNA-directed regulation of Arabidopsis AUXIN RESPONSE FACTOR17 is essential for proper development and modulates expression of early auxin response genes Plant Cell 2005 17 1360 75 10.1105/tpc.105.031716 15829600 40. Navarro L Dunoyer P Jay F Arnold B Dharmasiri N Estelle M A plant miRNA contributes to antibacterial resistance by repressing auxin signaling Science 2006 312 436 439 10.1126/science.1126088 16627744 41. Reyes JL Chua NH ABA induction of miR159 controls transcript levels of two MYB factors during Arabidopsis seed germination Plant J 2007 49 592 606 10.1111/j.1365-313X.2006.02980.x 17217461 42. Alonso-Peral MM Li JY Li YJ Allen RS Schnippenkoetter W Ohms S The microRNA159-regulated GAMYB-like genes inhibit growth and promote programmed cell death in Arabidopsis Plant Physiol 2010 154 757 771 10.1104/pp.110.160630 20699403 43. Li DT Wang LW Liu X Cui DZ Chen TT Zhang H Deep sequencing of maize small RNAs reveals a diverse set of microRNA in dry and imbibed seeds PLoS One 2013 8 e55107 10.1371/journal.pone.0055107 23359822 44. Xie KB Wu CQ Xiong L Genomic organization, differential expression, and interaction of SQUAMOSA promoter-binding-like transcription factors and microRNA156 in rice Plant Physiol 2006 142 280 293 10.1104/pp.106.084475 16861571 45. Wang JW Czech B Weigel D miR156-regulated SPL transcription factors define an endogenous flowering pathway in Arabidopsis thaliana Cell 2009 138 4 738 749 10.1016/j.cell.2009.06.014 19703399 46. Shamimuzzaman M Vodkin L Identification of soybean seed developmental stage-specific and tissue-specific miRNA targets by degradome sequencing BMC Genomics 2012 13 310 10.1186/1471-2164-13-310 22799740 47. Hu JH Zhang HY Ding Y Identification of conserved microRNAs and their targets in the model legume Lotus japonicus J Biotech 2013 164 520 524 10.1016/j.jbiotec.2013.01.026 48. Mu J Tan H Hong S Liang Y Zuo J Arabidopsis transcription factor genes NF-YA1, 5, 6, and 9 play redundant roles in male gametogenesis, embryogenesis, and seed development Mol Plant 2013 6 188 201 10.1093/mp/sss061 22933713 49. Penfield S Josse E Kannangara R Gilday AD Halliday KJ Graham IA Cold and light control seed germination through the bHLH transcription factor SPATULA Curr Biol 2005 15 1998 2006 10.1016/j.cub.2005.11.010 16303558 50. Wang K, Deng J, Damaris RN, Yang M, Xu LM, Yang PF. LOTUS-DB: an integrative and interactive database for Nelumbo nucifera study. Database (Oxford). 2015;2015,bav023. 51. Hafner M Landgraf P Ludwig J Rice A Ojo T Lin C Holoch D Lim C Tushcl T Identification of microRNAs and other small regulatory RNAs using cDNA library sequecing Methods 2008 44 3 12 10.1016/j.ymeth.2007.09.009 18158127 52. Li R Li Y Kristiansen K Wang J SOAP: short oligonucleotide alignment program Bioinformatics 2008 24 713 4 10.1093/bioinformatics/btn025 18227114 53. Kozomara A Griffiths-Jones S miRBase: integrating microRNA annotation and deep-sequencing data Nucleic Acids Res 2011 39 D152 7 10.1093/nar/gkq1027 21037258 54. Zuker M Mfold web server for nucleic acid folding and hybridization prediction Nucleic Acids Res 2003 31 3406 15 10.1093/nar/gkg595 12824337 55. Chen L Wang T Zhao M Tian Q Zhang WH Identification of aluminum-responsive microRNAs in Medicago truncatula by genome-wide high-throughput sequencing Planta 2012 235 375 86 10.1007/s00425-011-1514-9 21909758 56. Audic S Claverie JM The significance of digital gene expression profiles Genome Res 1997 7 986 95 9331369 57. Addo-Quaye C Eshoo TW Bartel DP Axtell MJ Endogenous siRNA and miRNA targets identified by sequencing of the Arabidopsis degradome Curr Biol 2008 18 758 62 10.1016/j.cub.2008.04.042 18472421 58. Alexa A Rahnenfuhrer J Lengauer T Improved scoring of functional groups from gene expression data by decorrelating GO graph structure Bioinformatics 2006 22 13 1600 07 10.1093/bioinformatics/btl140 16606683 59. Varkonyi-Gasic E Wu R Wood M Walton EF Hellens RP Protocol: a highly sensitive RT-PCR method for detection and quantification of microRNAs Plant Methods 2007 3 12 10.1186/1746-4811-3-12 17931426 60. 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 8 10.1006/meth.2001.1262 11846609

PMC005xxxxxx/PMC5002176.txt

==== Front Chem Cent JChem Cent JChemistry Central Journal1752-153XSpringer International Publishing Cham 19810.1186/s13065-016-0198-4Research ArticleNew method for sequestration of silver nanoparticles in aqueous media: in route toward municipal wastewater Roy Marie-Laine marie-laine_roy@hotmail.com 1Gagnon Christian christian.gagnon@ec.gc.ca 2http://orcid.org/0000-0001-6463-0413Gagnon Jonathan 1-418-723-1986jonathan_gagnon@uqar.ca 11 Département de Biologie, chimie et géographie, Université du Québec à Rimouski, 300 allée des Ursulines, Rimouski, QC G5L 3A1 Canada 2 Centre Saint-Laurent, Environment Canada, 105 McGill st., 7th floor, Montreal, QC H2Y 2E7 Canada 26 8 2016 26 8 2016 2016 10 1 546 1 2016 4 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 Nanomaterials are widely used in industry for their specific properties. Silver nanoparticles (Ag NPs) are largely used in several consumer products notably for their antibacterial properties and will likely be found in wastewater, then in the receiving environment. The development of a product capable to sequestrate those released contaminants is needed. Under environmental conditions, the biopolymer chitosan can generally coordinate the cationic metals. Ag NPs present unique properties due to their high surface/mass ratio which are promising for their sequestration. Results The immobilization of chitosan on functionalized silica assisted by microwaves gives a sequestering agent of silver while being easily recoverable. The IR spectrum confirmed the immobilization of N,N–dimethylchitosan (DMC) on silica core. The immobilized DMC gave a percentage of sequestration of Ag NPs (120 µg L−1) in nanopure water of 84.2 % in 4 h. The sequestration efficiency was largely dependent of temperature. By addition of hydrosulfide ions, the percentage of sequestration increased up to 100 %. The immobilized DMC recovered a large portion of silver regardless the speciation (Ag NP or Ag+). In wastewater, the immobilized DMC sequestered less Ag NPs (51.7 % in 97 % wastewater). The presence of anionic (sodium dodecyl sulfate and sodium N–lauroylsarcosinate) and non-ionic surfactants (cetyl alcohol) increased the hydrophobicity of Ag NPs and decreased the percentage of sequestration. Conclusions The immobilized DMC is a promising tool for sequestrating such emerging pollutant at low concentrations in a large volume of sample that would allow the characterization of concentrated Ag NPs by transmission electron microscopy. The efficiency of the support to sequestrate would likely be influenced by the chemical environment of the Ag NP solution. Electronic supplementary material The online version of this article (doi:10.1186/s13065-016-0198-4) contains supplementary material, which is available to authorized users. Keywords Ag NPSupported polysaccharideSilicaRemovalWastewaterSilver sulfideOrganic matterhttp://dx.doi.org/10.13039/501100000038Natural Sciences and Engineering Research Council of Canadaissue-copyright-statement© The Author(s) 2016 ==== Body Background Nanomaterials are widely used in industry for their specific properties. A nanoparticle is defined as a particle possessing at least two dimensions measuring between 1 and 100 nm [1, 2]. In recent years, silver nanoparticles (Ag NPs) have been widely studied since they have a high surface/mass ratio that confers a higher reactivity. They are used in catalysis and for their antimicrobial properties in many areas of applications including consumer products and textiles [1–4]. In 2012, approximately 55 tons of Ag NPs were produced and used [5]. The majority of Ag NPs in consumer products will be likely found in municipal wastewater treatment plants and exposure to aquatic organisms could result in different toxicological effects [6]. The development of new sequestration techniques is therefore important tools for their removal [2]. Chitosan represents a rare example of cationic biopolymer that is mainly extracted from crustacean exoskeletons. This aminopolysaccharide is known as coagulant and flocculent [7] and for its capacity to bind transition metals. The alcohol and amino groups in raw chitosan allow the chelation of transition metals. At neutral pH, cationic metals are coordinated by unbounded electrons of nitrogen atoms [4, 8]. Applications of chitosan are limited by its insolubility in aqueous solutions and organic solvents. The protonation of amino groups lead to the solubilization of chitosan in diluted acid conditions. However, its sorption capacity [4] and utilization in wastewater treatment [9] are limited. Ag NP recovery methods have been developed including cloud point extraction with Triton X-114 [10] and activated carbon [11]. These methods work for high concentrations of Ag NPs only. Silica is a widely used support for chromatography and for supported reagents and catalysts [12]. Silica with silanol groups on the surface and a large surface area allow coupling with many molecules including polymers [8, 12]. The immobilization of polymers on silica can be used for a variety of applications such as biosensors and drug delivery, for instance [13]. The use of polymers in the catalytic reactions of chemicals and biological processes is growing. Supported polymers offer opportunities in the production of chemical and new intermediates [14]. Supported polymers are been used in various combinatorial chemicals, in the research for new drugs, in the oil refinery and in catalysis and biosynthesis [14, 15]. Supported polysaccharides allow the formation of support with high surface for sorption where some polysaccharides are used to immobilize various molecules such as enzymes. Some studies have been realized to immobilize chitosan on a support made of silica gel [16]. Immobilized chitosan can bound copper ions [9] or acted as affinity support for the adsorption of proteins [17]. These syntheses imply more than three steps that necessitate several days and require the removal of starting compounds. Moreover, concentrations of heavy metal ions were as high as the order of milligram per liter. The microwave-assisted heating is a technique with many advantages including the ability to accelerate chemical reactions and to achieve higher heating rates and better reaction yields [18, 19]. Therein, we report the preparation of immobilized chitosan derivative on modified silica and the assessment of potential sequestration of Ag NPs in municipal wastewater. In this work, the removal capacity of this sequestration was then studied against two other silver species (ionic silver and Ag2S NPs). Results and discussion Formation of immobilized N,N–dimethylchitosan (DMC) on modified silica The immobilization of DMC on the modified silica is summarized in Scheme 1. The alkyl halide of modified silica reacts with tertiary amine of DMC in a one-step process using microwave. Some tertiary amine groups are converted into quaternary ammonium allowing to chemically bound DMC onto silica propyl bromide. The reaction between DMC and modified silica lead to an insoluble product even in the protonated form whereas the free protonated DMC is soluble under acidic conditions. The immobilized DMC was washed with a solution of acetic acid to remove unbounded DMC. The supported DMC was then characterized by IR and Raman spectroscopy (see Additional file 1: Figures S1 and S2).Scheme 1 Immobilization of DMC on modified silica In Fig. 1, the IR spectrum of silica propyl bromide shows a broad Si–O stretching at 1086 cm−1. The IR spectrum of DMC shows bands at 3423, 2869, 1586, 1455, 1364 and 1019 cm−1 representing OH stretching, CH vibrations, CH2 deformation, CH3 deformations, C-N stretching and C-O stretching, respectively (Additional file 1: Figure S1). The IR spectra of immobilized DMC after washing and those of free DMC are similar but the relative intensity of bands is different. The intensity of OH, C–O, C–N stretching are higher for immobilized DMC whereas the CH2 deformation band of immobilized polymer is lower. Considering that unbounded DMC was washed out, these bands indicate that DMC was fixed on the modified silica. These differences in IR spectroscopy indicate that the polymer is immobilized on silica and its surface is covered by DMC.Fig. 1 Infrared spectra in the 800–1700 cm−1 region of A immobilized DMC after washing; B DMC; C silica propyl bromide According to the literature [20], the C–Br stretching of bromoalkane compounds absorb in Raman at 645–635 and 565–555 cm−1 (general stretching zone). In Fig. 2, the Raman spectrum of the silica propyl bromide shows C–Br stretching at 634 and 562 cm−1. These vibrational bands disappeared after the immobilization of DMC. The disappearance of these bands in immobilized DMC spectrum indicates that DMC was bound to silica. The comparison of Raman spectra of immobilized DMC, free DMC and modified silica shows a new vibrational band at 853 cm−1 for immobilized DMC.Fig. 2 Raman spectra of A immobilized DMC after washing; B DMC; C silica propyl bromide With the DMC/silica ratio used during the reaction, the nitrogen/carbon ratio of supports was quite constant within a variation of 5 %, a small decrease is observed for polymer/silica ratio of 2–5 (Additional file 1: Figure S3). The nitrogen percentage increases until a DMC/silica ratio of 1 and after is relatively constant as well demonstrating that immobilization of DMC on silica is saturated. Sequestration of silver nanoparticles It is possible to qualitatively verify the sequestration of a solution of Ag NPs (120 µg L−1) by comparing UV–visible spectra before and after sequestration (Additional file 1: Figure S4). The intensity of the absorption band of citrate-coated Ag NPs at 400 nm decreases after sequestration that was attributed to the reduction of Ag NP concentration. The ICP-MS analyses of the supernatant and immobilized DMC were carried out to verify the mass balance of silver content. Different sequestration parameters were evaluated whose influence of sequestration such as time, temperature and different forms of silver that can be found in the waters. These results are presented in Table 1. Table 1 (lines 1–3) shows the percentage of sequestration after 0.5, 2 and 4 h of Ag NPs in nanopure water. During the first 30 min, the support sequestrated a large proportion (59.9 %) of Ag NPs. After that the percentage of sequestration increased with time, but more slightly between 2 and 4 h to reach around 80 %. For the lower amount of ionic silver (1.34 mg L−1; line 6), the immobilized DMC recovered totally the metal. At higher concentration (4.25 mg L−1; line 7), the immobilized DMC sequestrated a lower proportion of ionic silver (84.2 %) since there must be probable saturation of the immobilized DMC. The maximum sorption capacity of the immobilized DMC at those concentrations was 10.1 µg g−1 for Ag NPs and 0.36 mg/g−1 for ionic silver (Ag+). The unbounded electron of nitrogen atoms would be available for the coordination of Ag+. Thus, the immobilized DMC sequesters silver despite its form. Ag NPs and ionic silver (AgNO3) are mostly recovered. A support composed of positively charged quaternary trimethylated amines (TMC) was also used to verify if it would be more selective for Ag NPs. The ionic silver in presence of immobilized TMC (DQ of 47.6 %) was sequestrated at 28.0 % (line 8) and 23.0 % of Ag NPs for immobilized TMC (line 9). The decrease of sequestration would be explained by the steric hindrance around the cationic charge of the polymer.Table 1 Percentage of sequestration of Ag NPs, Ag+ and Ag2S NPs by immobilized DMC at different conditions Line Time of sequestration (h) Parameters Percentage of sequestration (%)a 1 0.5 25 °C 59.9 ± 4.1 2 2 25 °C 77.7 ± 3.9 3 4 25 °C 84.2 ± 5.8 4 4 2 °C 3.5 ± 0.2 5 4 40 °C 26.9 ± 0.5 6 4 1.34 mg L−1 of Ag+ 100.0 ± 0.0 7 4 4.25 mg L−1 of Ag+ 84.2 ± 4.8 8 4 4.25 mg L−1 of Ag+; with immobilized TMC 28.0 ± 8.4 9 4 With immobilized TMC 23.0 ± 1.4 10 4 +16.0 mg L−1 of Ag2S 24.1 ± 9.1 Ag NPs (120 µg L−1) were added by default excepted in cases where the source of silver is mentioned. Ag+ was added as silver nitrate aAverage ± SD In an environment with high concentrations of sulfur like municipal wastewater, Ag NPs can also be transformed into Ag2S [21]. The Ag2S nanoparticles of size of 77.1 ± 56.8 nm were synthesized from l-cysteine and silver nitrate. The immobilized DMC sequestrated 24.1 % of Ag2S NPs (line 10) corresponding to a sorption capacity of 0.39 mg g−1. The zeta potential was used to quantify the nanoparticle charge and provide information on electrostatic interactions (Table 2). The zeta potential of the Ag NPs was −7.4 mV (line 11) while the zeta potential of Ag2S NPs was −6.1 mV. With a zeta potential being less negative, Ag2S NPs would be more difficultly adsorbed on the immobilized DMC (line 8).Table 2 Zeta potential of Ag NPs (120 µg L−1) by addition of electrolytes Line Parameters Zeta potential of Ag NPs (mV)a 11 Nanopure water −7.4 ± 1.2 12 20 mg L−1 NaSH −49.3 ± 0.8 13 35.2 mg L−1 SDS −9.0 ± 1.3 14 20 mg L−1 NaSH; 8.8 mg L−1 SDS −58.2 ± 3.1 aAverage ± SD Table 3 shows the percentage of sequestration of Ag NPs, by the immobilized DMC, increases with addition of hydrosulfide. The hydrosulfide concentrations correspond to the minimum amounts of sulfur found in wastewater according to Hurse and Abeydeera [22]. Hydrosulfide ions can strongly coordinate silver because they modify the electronic environment and creates strong covalent bonds [23]. By coordinating the surface of Ag NPs, the particle becomes strongly negative. Indeed, the zeta potential of Ag NPs was −7.4 mV (line 11) while the zeta potential of Ag NPs with hydrosulfide ions was −49.3 mV (line 12). This strong negative charge promotes electrostatic interactions with the cationic immobilized DMC.Table 3 Percentage of sequestration of Ag NPs (120 µg L−1) by immobilized DMC with addition of NaSH Line NaSH concentration (mg L−1) Molar ratio NaSH/Ag NP Percentage of sequestration (%)a 15 0 0 84.2 ± 4.8 16 20 0.2 99.5 ± 0.2 17 100 1.0 100.0 ± 0.0 18 200 2.0 99.6 ± 0.6 aAverage ± SD Sodium dodecyl sulfate (SDS) and cetyl alcohol are surfactants commonly used in consumer products, which are found in municipal wastewater. Surfactants could affect Ag NPs properties and their interactions with immobilized DMC [24]. SDS concentrations used in experiments were the upper and lower limits found in wastewater influents in the USA according to Knepper and coworkers [25], whereas the cetyl alcohol concentration is limited by the solubility. In the presence of SDS, an anionic surfactant (Table 4, lines 19–21), the percentage of sequestration of Ag NPs decreases to around 20 %. In the presence of sodium N-lauroylsarcosinate (SLS), another anionic surfactant (line 22), the percentage of sequestration decreases to 2.7 %. The zeta potential of Ag NPs in water was −7.4 mV (line 11) while the zeta potential with addition of SDS was −9.0 mV (line 13). The charge on the surface does not change within precision. Surfactants, due to their partial charge (SLS -1/2 and SDS -1/3 per oxygen atom), would replace the citrate ion and would increase the hydrophobicity of Ag NPs. The partial charge of SLS being greater than SDS would coordinate more Ag NPs and replace more the citrate ion, hence the lower sequestration by addition of SLS. Highly hydrophobic species could reduce sequestration. In the presence of cetyl alcohol, a non-ionic surfactant (lines 25–26), the percentage of sequestration became at around 4 %. The same reduction due to hydrophobicity occurs with cetyl alcohol. Ag NP behavior in wastewater would be changed. In the presence of both SDS and sulfide, the DMC sequestered 6 % of Ag NP (line 23). In this case, the zeta potential was −58.2 mV (line 14). The particles are strongly negative as well as being very hydrophobic that prevents sequestration by DMC.Table 4 Percentage of sequestration of Ag NPs by immobilized DMC at differents conditions after 4 h Line Parameters Percentage of sequestration (%)a 19 3.6 µg L−1 SDS 27.0 ± 1.4 20 8.8 mg L−1 SDS 21.6 ± 7.9 21 35.2 mg L−1 SDS 23.5 ± 3.3 22 11.8 mg L−1 SLS 2.7 ± 0.6 23 8.8 mg L−1 SDS; 20 mg L−1 NaSH 6.0 ± 0.8 24 20 mg L−1 NaSH; 8.8 mg L−1 SDS 6.4 ± 0.6 25 0.335 µg L−1 cetyl alcohol 2.4 ± 0.5 26 1.34 µg L−1 cetyl alcohol 5.9 ± 0.1 The concentration of Ag NPs was 120 μg L-1 aAverage ± SD The sequestration percentage reached very high values as high as 99–100 % (Table 3, lines 16–18) for solution containing sodium hydrosulfide and decreased to 90.7 % (line 27) by addition of 10 % municipal wastewater. Municipal wastewater contains compounds like sulfur and organic matter leading to a decrease of sequestration (lines 27–29). A solution composed of 50 % wastewater gave a sequestration of 84.6 % (line 28) while a solution of 97 % wastewater had a percentage of sequestration of 51.7 % by immobilized DMC (line 29). In the absence of suspended matter—municipal wastewater previously filtered through GF/F 0.7 μm—the percentage of sequestration was 27.4 % (line 30). The organic matter is known to form complexes with silver [23]. The presence of humic substances stabilizes Ag NPs by covering them that reduced agglomeration or sedimentation [26]. The lower electrostatic charge would decrease interaction with the cationic immobilized DMC as observed in the presence of organic matter (Table 5).Table 5 Percentage of sequestration of Ag NPs (120 µg L−1) by immobilized DMC in wastewater after 4 h Line Composition of aqueous solutions Percentage of sequestration (%)a 27 10 % wastewater 90.7 ± 0.2 28 50 % wastewater 84.6 ± 7.6 29 97 % wastewater 51.7 ± 6.7 30 97 % wastewater filtered GF/F 0.7 μm 27.4 ± 5.8 aAverage ± SD Characterization of immobilized DMC after sequestration of Ag NP Figure 3 shows the infrared spectra of immobilized DMC before and after sequestration of citrate coated Ag NPs. Figure 3a (after sequestration) shows a band at 1558 cm−1 associated to asymmetric carboxylate stretching band of citrate carbonyl on Ag NPs [27] while Fig. 3b (before sequestration) does not have any band in this region. The C = O stretching in Fig. 3a (after sequestration) indicates that Ag NPs were sequestered by the immobilized DMC.Fig. 3 Infrared spectra of immobilized DMC after (A) and before (B) Ag NP sequestration SEM allows visualizing certain characteristics like the size and morphology. Figure 4 shows SEM image (Fig. 4a) of the immobilized DMC after Ag NP sequestration in water. There are no observable structural differences in SEM between the silica (not shown) and immobilized DMC. Thus, the silica would have a homogeneous covering of DMC, which is coherent with the IR spectrum (Fig. 1). SEM images show that the immobilized DMC is porous. In Fig. 4b, the black dots on TEM image represent Ag NPs of 20 nm size while the light gray shape without distinct outline would be organic matter (DMC or citrate).Fig. 4 a Scanning electron microscope (SEM) image of immobilized DMC and b transmission electron microscope (TEM) image of Ag NPs after sequestration The average diameter of Ag NPs and their size distribution can be determined by TEM. In the stock solution, citrate-coated Ag NPs do not agglomerate (Fig. 5a), Ag NPs are monodisperse with an average diameter of 22.1 nm (Fig. 6a). Adding NaSH, a part of Ag NPs agglomerates while the other part remains in monomeric form (Fig. 5c, d) with an average diameter of 20.4 nm (Fig. 6c). After sequestration in nanopure water (Fig. 4b) or NaSH solution (Fig. 5b), Ag NPs appeared with defined sizes without agglomeration. However they are polydispersed with sizes of 15, 22–29, 44 and 59–88 nm, resulting in an average diameter of 39.8 nm (Fig. 6b). After sequestration in NaSH solution, the range was mainly between 20 and 24 nm and 42–44 nm, with an average diameter of 35.9 nm (Fig. 6d). During sequestration, the DMC counterion (acetate ion) could exchange with the citrate ion. Thus, Ag NPs would be less stable and will agglomerate.Fig. 5 TEM images of a Ag NPs citrate; b Ag NPs citrate with NaSH in molar ratio NaSH/Ag NP 1:5 after sequestration; c, d Ag NP citrate with NaSH in molar ratio NaSH/Ag NP 1:5 Fig. 6 Particle size distribution (n = 100) from TEM images of a Ag NPs before sequestration; b Ag NPs after sequestration; c Ag NPs with NaSH before sequestration; d Ag NPs with NaSH after sequestration Effect of temperature on sequestration By varying the temperature during sequestration, it was possible to determine the activation energy from the Arrhenius relationship. A plot of 1/T according to the natural logarithm of the first order rate constant is performed. The slope of the line corresponds to the activation energy divided by the gas constant (8.314 J K−1 mol−1). The activation energy was 803 J mol−1. The immobilized DMC sequesters 3.5 % at 275 K, 84.2 % at 298 K and 26.9 % at 313 K (Table 1, lines 3–5). Sequestration was largely affected by temperature where the best sequestration was obtained at 25 °C. At 2 °C, the activation energy is not completely attained and a low amount of Ag NP is sequestered by the immobilized DMC. This energy is achieved at room temperature. Thus, environmental samples could be easily handled. At 40 °C, the activation energy is reached and environmental temperature increases the molecular motion. The increase of temperature in the reaction medium would result in competitive reactions explaining the low percentage of sequestration. When modifying the order of addition between NaSH and SDS, the percentage of sequestration of Ag NPs is similar, 6.0 % when SDS (Table 4, line 23) is added first compared to 6.4 % when NaSH (line 24) is added first. These observations indicate that the process is reversible and that there is competition between anions. Experimental General information SiliaBond® propyl bromide (particle size 40–63 µm, loading 1.69 mmol/g, specific surface area 470–530 m2 g−1), chitosan (viscosity <20 mPa s (cP), degree of deacetylation >95 % from shrimp exoskeletons, Pandalus borealis) and silver nanoparticles (20 nm, 0.02 mg/mL) coated with citrate were purchased respectively from Silicycle (Quebec), Primex (Iceland) and TedPella (USA). All other reagents were bought from Aldrich except sodium N-lauroyl sarcosinate (ICN biomedicals). Sodium dodecyl sulfate (SDS) was reagent plus grade. Concentrated nitric acid (≥69 % v/v) and hydrogen peroxide (≥30 % v/v) were ultrapure grade whereas other reagents were ACS grade. N,N–dimethylchitosan (DMC) was synthesized according to literature [28]. Nanopure water was obtained from a Barnstead nanopure infinity ultrapure water system. Ionic silver comes from AgNO3. All materials were washed with nitric acid and rinsed with nanopure water before use. Municipal wastewater was collected on June 18, 2013 as a 24 h-composite sample from aerated lagoons at Rimouski-Est station (Quebec, Canada). The sample was stored at −20 °C. Municipal wastewater had 0.61 g L−1 of total matter and 0.46 g L−1 of dissolved matter. The microwave heating was realized with a Mars microwave system from CEM Corporation using MarsXpress™ close-vessels. Infrared and Raman spectra were recorded on a Thermo scientific Nicolet iS10 spectrometer with Smart Omni transmission in KBr pellets and on a Thermo scientific DXR Raman Microscope directly on solid, respectively. Elemental analyses were determined using analyzer Costech instruments elemental combustion system 4100. NMR spectra were performed using an Avance III HD 600 MHz NMR from Bruker by NanoQAM (Université du Québec à Montréal). UV–visible spectra were recorded on a Cary 100 Bio UV–visible spectrophotometer from Varian. ICP–MS measurements were achieved on an Agilent 7500c spectrometer octopole reaction system using argon plasma at 7000 K, autosampler ASX-520 Cetac and software ChemStation v.3.04. Analyses from MP-AES were achieved on an Agilent Technologies 4200 MP-AES with a nitrogen generator, autosampler ASX-520 Cetac and MP Expert software version 1.5.0.6545. Zeta potentials were measured by Malvern zetasizer nano ZS with Malvern Zetasizer software version 7.11. Solutions were placed in disposable capillary cells (DTS1070) of Malvern which were washed with nanopure water, nitric acid 10 % v/v, nanopure water and ethanol. A single measurement with zetasizer had 100 runs in manual mode, the zetasizer took three measurements with a delay of 45 s. Zeta potentials of Ag NPs were measured in nanopure water excepted when presence of salts is mentioned. Transmission electron microscopy (TEM) was recorded on a Delong Instruments model LVEM5. Before TEM analyses, dried supports were ground in an agate mortar and then suspended in dry ethanol. A few drops of solution were placed on a copper grid of 400 mesh covered with a hexagonal carbon film provided by Ted Pella Inc. (Redding, CA). SEM microscope was a JEOL JSM-6460 LV scanning electron microscope. Dried supports were placed on a carbon tape and placed on the sample holder. The uncertainty of zeta potential measurements was estimated using the standard deviation between three data collections. The uncertainty on percentage of sequestration comes from the standard deviation between two independent sequestrations. General procedure for the preparation of immobilized DMC on modified silica (example for polymer/silica ratio 1:1) A suspension containing DMC (0.30 g), sodium carbonate (0.90 g) and SiliaBond® propyl bromide (0.32 g) was prepared in a mixture of methanol/water (8 mL, 1:9 v/v). In MarsXpress™ close-vessels, the suspension was heated by microwave at 100 °C during 5 min and the temperature was maintained at 100 °C during 15 min using a maximum power of 1600 W. The solution was allowed to reach room temperature (rt). The solid was filtered and suspended in a 1 % (v/v) aqueous acetic acid solution (50 mL) during 15–30 min. The solid was filtered and dried at normal atmosphere. A white solid was obtained (0.33 g). The solid was ground to a size of 250 μm. IR υ (cm−1) 3430 (OH), 2900 (CH), 1558 (CH2 def), 1462 (CH3 def), 1380–1265 (C–N), 1110–1090 (C–O pyranosyl). Raman υ (cm−1) 634, 562 (C–Br). General procedure for the N-methylation of immobilized DMC on modified silica (polymer/silica ratio 1:1) A suspension containing immobilized DMC (0.30 g), sodium carbonate (0.90 g) in a mixture of methanol/water (8 mL, 1:9 v/v) and iodomethane (3 mL). In MarsXpress™ close-vessels, the suspension was heated by microwave at 100 °C during 5 min and the temperature was maintained at 100 °C during 15 min using a maximum power of 1600 W. The solution was allowed to reach rt. The solid was then filtered and dried under normal atmosphere. The yield is quantitative. Degree of quaternization (DQ) of TMC in the protonated form was obtained by comparing the integrals of N(CH3)3+ (3.3 ppm), N(CH3)2 (3.0 ppm) and CH3CO (2.1 ppm) peaks from the 1H NMR spectrum in D2O. DQ of TMC was 47.6 % from the following equation. DQ =N(CH3)3+/9(N(CH3)3+/9)+ (N(CH3)2/6)+ (CH3CO/3)×100% A suspension containing TMC (0.30 g), sodium carbonate (0.90 g) and SiliaBond® propyl bromide (0.32 g) in a mixture of methanol/water (8 mL, 1:9 v/v). In MarsXpress™ close-vessels, the suspension was heated by microwave at 100 °C during 5 min and the temperature was maintained at 100 °C during 15 min using a maximum power of 1600 W. The solution was allowed to reach rt. The solid was then filtered and dried under normal atmosphere. The solid was ground to a size of 250 μm. Procedure for formation of Ag2S nanoparticles The synthesis method of Ag2S nanoparticles was adapted from Xiang and coworkers [29] and Brelle and coworkers [30]. Silver nitrate (68 µmol, 11.5 mg) was added to a stirred solution of l-cysteine (68 µmol, 8.2 mg) in 10 mL ethanol. After 15 min, the solution was transferred into a 15 mL Teflon tube. The tube was placed in 120 mL high pressure reactor from Parr Instrument. The reactor was heated at 180 °C during 10 h after that it was allowed to reach rt. The resulting precipitate was centrifuged at 3000 rpm during 10 min and washed using nanopure water and absolute ethanol several times. The dark precipitate was dried at 60 °C during 6 h. The black, dried precipitate was then put into a tube with ethanol and placed in a bath sonicator for 5 min. The solution was decanted for 1 h. The suspension was recovered and evaporated. The Ag2S NP mean size of 77.1 ± 56.8 nm was determined by TEM. General procedure of sequestration of silver nanoparticles Ag NP solution was prepared by dilution (factor 33×) of the commercial stock solution. In a Falcon tube (15 mL), the Ag NP solution (10 mL) and immobilized DMC on silica (0.100 g) were stirred with a magnetic bar during 4 h. The sequestrations were carried out in duplicate. The suspension was then centrifuged at 1000×g for 5 min. The supernatant was first collected and the residual solid was filtered (Whatman cellulose filter papers grade 2) and dried under normal atmosphere. The samples were placed in the dark at 4 °C until further analysis. IR of immobilized DMC after sequestration υ (cm−1) 3430 (OH), 2900 (CH), 1651 (C = O in COOH), 1557 (CH2 def), 1110–1090 (C–O pyranosyl). Sample preparation prior to ICP–MS and MP-AES analyses Dried support (0.100 g), concentrated nitric acid (6 mL) and hydrogen peroxide (1 mL) were mixed in open flasks until the complete gas evolution during at least 2 h. The flasks were closed and heated to 70 °C in a hot bath for an additional 2 h. The supernatant (2 mL) was digested in the same way than for the support except that 4 mL of nitric acid was used. The samples were stored in dark at 4 °C until ICP-MS or MP-AES analyses. All analyses were performed with the ICP-MS except analyses with TMC immobilized, Ag2S, Ag NPs at 2 and 40 °C, NaSH with SDS and SLS that have been made by MP-AES. The detection of 107Ag was used to measure the total silver in ICP–MS. The limit of detection for silver by ICP–MS was 0.04 µg L−1. Conclusions The synthesis of immobilized DMC on silica meets some principles of green chemistry [31, 32], for example, by using renewable products and microwave-assisted heating. Ag NPs are retained in the immobilized DMC by electrostatic interactions. When Ag NPs are highly negatively charged, for example by addition of NaSH, interactions are stronger and thus the sorption efficiency increases. Sequestrations are also highly dependent of hydrophobicity of Ag NPs coming from surfactants or organic matter that would decrease the electrostatic interactions with Ag NPs and lower the sorption efficiency. In a more complex environment, many factors likely influence the interactions between Ag NPs and immobilized DMC resulting in lower sorption capacity. Immobilized polysaccharides as chitosan derivatives could serve as a promising approach for retrieving or removing emerging pollutants and heavy metals due to the chelation of the nitrogen atom. The immobilized DMC can be used on large volume samples and at low metallic pollutant concentrations. Sequestration optimizations should be carried out to increase selectivity and sensitivity of the method for potential number of uses. Additional file 10.1186/s13065-016-0198-4 Infrared and Raman spectra of immobilized DMC, variation of N/C ratio, UV-visible spectra of Ag NP solutions. Authors’ contributions MLR carried out the syntheses and the sequestration experiments and drafted the manuscript. CG and JG conceived of the study, and participated in its design and coordination and helped interpreting data and writing up the manuscript. All authors read and approved the final manuscript. Acknowledgements We thank the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Chemical Management Plan (CMP) of Environment Canada for their financial support. We also acknowledge professor Émilien Pelletier (TEM), Mr. Mathieu Babin (ICP-MS) and Mr. Claude Belzile (SEM) for their supports. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Farkas J Peter H Christian P Gallego-Urrea JA Hassellöv M Tuoriniemi J Gustafsson S Olsson E Hylland K Thomas KV Characterization of the effluent from a nanosilver producing washing machine Environ Int 2011 37 1057 1062 10.1016/j.envint.2011.03.006 21470683 2. Wijnhoven SWP Peijnenburg WJGM Herberts CA Haggens WI Oomen AG Heugens EHW Roszek B Bisschops J Gosens I Van De Meent D Dekkers S De Jong WH Van Zijverden M Sips AJA Geertsma RE Nano-silver—a review of available data and knowledge gaps in human and environmental risk assessment Nanotoxicology 2009 3 109 138 10.1080/17435390902725914 3. Savage N Diallo MS Nanomaterials and water purification: opportunities and challenges J Nanopart Res 2005 7 331 342 10.1007/s11051-005-7523-5 4. Guidal E Heterogenous catalysis on chitosan-based materials: a review Prog Polym Sci 2005 30 71 109 10.1016/j.progpolymsci.2004.12.001 5. Piccino F Gottschalk F Seeger S Nowack B Industrial production quantities and uses of ten engineered nanomaterials in Europe and the world J Nanopart Res 2012 14 1109 1120 10.1007/s11051-012-1109-9 6. Benn TM Westerhoff P Nanoparticle silver released into water from commercially available sock fabrics Environ Sci Technol 2008 42 4133 4139 10.1021/es7032718 18589977 7. Renault F Sancey B Charles J Morin-Crini N Badot PM Winterton P Crini G Chitosan flocculation of cardboard-mill secondary biological wastewater Chem Eng J 2009 155 775 783 10.1016/j.cej.2009.09.023 8. Liu XD Tokura S Haruki M Nishi N Sakairi N Surface modification of nonporous glass beads with chitosan and their adsorption property for transition metal ions Carbohyd Polym 2002 49 103 108 10.1016/S0144-8617(01)00308-3 9. Li F Du P Chen W Zhang S Preparation of silica-supported porous sorbent for heavy metal ions removal in wastewater treatment by organic-inorganic hybridization combined with sucrose and polyethylene glycol imprinting Anal Chim Acta 2007 585 211 218 10.1016/j.aca.2006.12.047 17386667 10. Chao J Liu J Yu S Feng Y Tan Z Liu R Yin Y Speciation analysis of silver nanoparticles and silver ions in antibacterial products and environmental waters via cloud point extraction-based separation Anal Chem 2011 83 6875 6882 10.1021/ac201086a 21797201 11. Gicheva G Yordanov G Removal of citrate-coated silver nanoparticles from aqueous dispersions by using activated carbon Coll Surf A 2013 431 51 59 10.1016/j.colsurfa.2013.04.039 12. Yin R Otterbrite RM Siddiqui JA Grafting of poly(acrylic acid) onto nonporous glass bead surfaces Polym Adv Technol 1997 8 761 766 10.1002/(SICI)1099-1581(199712)8:12<761::AID-PAT714>3.0.CO;2-6 13. Ehlert N Müller PP Stieve M Behrens P Immobilization of alkaline phosphate on modified silica coatings Micropor Mesopor Mat 2010 131 51 57 10.1016/j.micromeso.2009.11.040 14. Buchmeiser MR Polymeric materials in organic synthesis and catalysis 2003 Weinheim Wiley 15. De Jong KP Synthesis of supported catalysts Curr Opin Solid State Mater Sci 1999 4 55 62 10.1016/S1359-0286(99)80012-6 16. Kucherov AV Kramareva NV Finashina ED Koklin AE Kustov LM Heterogenized redox catalyst on the basis of the chitosan matrix 1. Copper complexes J Mol Cata A Chem 2003 198 377 389 10.1016/S1381-1169(03)00002-5 17. Shi QH Tian Y Dong XY Bai S Sun Y Chitosan-coated silica beads as immobilized metal affinity support for protein adsorption Biochem Eng J 2003 16 317 322 10.1016/S1369-703X(03)00095-0 18. Fukaya N Yamashita H Haga H Tsuchimoto T Onozawa S Sakakura T Yasuda H Microwave-assisted organic functionalization of silica surfaces: effect of selectively heating silylating agents J Organomet Chem 2011 696 825 828 10.1016/j.jorganchem.2010.10.008 19. Gabriel C Gabriel S Grant EH Halstead BSJ Mingos MP Dielectric parameters relevant to microwave dielectric heating Chem Soc Rev 1998 27 213 224 10.1039/a827213z 20. Socrates G Infrared and Raman characteristic group frequencies: tables and charts 2011 3e England Wiley 21. Liu JL Pennel KG Hurt RH Kinetics and mechanisms of nanosilver oxysulfidation Environ Sci Technol 2011 45 7345 7353 10.1021/es201539s 21770469 22. Hurse TJ Abeydeera WPP Quantification of sulfur and sulfur-containing compounds in wastewaters by means of combination of liquid chromatographic methods J Chromatogr A 2002 942 201 210 10.1016/S0021-9673(01)01379-6 11826872 23. Smith DS Bell RA Kramer JR Metal speciation in natural waters with emphasis on reduced sulfur group as strong metal binding sites Comp Biochem Physiol C 2002 133 65 74 24. Eriksson E Auffarth K Henze M Ledin A Characteristics of grey wastewater Urban Water 2002 4 85 104 10.1016/S1462-0758(01)00064-4 25. Knepper TP Barcelo D De Voogt P Volume XL: analysis and fate of surfactants in the aquatic environment 2003 Boston Elsevier Science 26. Cumberland S Lead JR Particle size distributions of silver nanoparticles at environmentally relevant conditions J Chromatogr A 2009 1216 9099 9105 10.1016/j.chroma.2009.07.021 19647834 27. Wulandari P Nagahiro T Fukada N Kimura Y Niwano M Tamada K Characterization of citrates on gold and silver nanoparticles J Coll Interf Sci 2015 438 244 248 10.1016/j.jcis.2014.09.078 28. Verheul RJ Amidi M Van der Wal S Van Riey E Jiskoot W Hennink WE Synthesis, characterization and in vitro biological properties of O-methyl free N ,N,N -trimethylated chitosan Biomaterials 2008 29 3642 3649 10.1016/j.biomaterials.2008.05.026 18556059 29. Xiang J Cao H Wu Q Zhang S Zhang X Watt AAR l -cysteine-assisted synthesis and optical properties of Ag2 S nanospheres J Phys Chem C 2008 112 3580 3584 10.1021/jp710597j 30. Brelle MC Zhang JZ Nguyen L Mehra RK Synthesis and ultrafast study of cysteine-and glutathione-capped Ag2 S semiconductor colloidal nanoparticles J Phys Chem A 1999 103 10194 10201 10.1021/jp991999j 31. Galuszkaa A Migaszewshi Z Namiesʹnik J The 12 principles of green analytical chemistry and the SIGNIFICANCE mnemonic of green analytical practices review Trend Anal Chem 2013 50 78 84 10.1016/j.trac.2013.04.010 32. Anastas P Warner JC Green chemistry: theory and practice 2000 New York Oxford University Press

PMC005xxxxxx/PMC5002177.txt

==== Front BMC Infect DisBMC Infect. DisBMC Infectious Diseases1471-2334BioMed Central London 169410.1186/s12879-016-1694-9Research ArticleEvaluation of SD Bioline HIV/syphilis Duo rapid test kits in Nepal Shakya Geeta geeta.nphl@gmail.com 1Singh Dipendra Raman dipendra_raman@gmail.com 2Ojha Hemanta Chandra drojhas@hotmail.com 2Ojha Chet Raj crojha@yahoo.com 1Mishra Shravan Kumar shravanmishra11@yahoo.com 1Malla Karishma rupkari@yahoo.com 3Chaudhary Puspa deopushpa@gmail.com 3Regmi Kiran regmikiran@gmail.com 41 National Public Health Laboratory, Teku, Kathmandu, Nepal 2 National Centre for AIDS and STD Control, Teku, Kathmandu, Nepal 3 Paropakar Maternity Hospital, Kathmandu, Nepal 4 Family Health Division, Teku, Kathmandu, Nepal 26 8 2016 26 8 2016 2016 16 1 45018 6 2015 5 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 Accurate and prompt diagnosis of HIV and syphilis simultaneously has reinforcing effect on their control program because of their prevalent co-infection. Availability of a simple user-friendly two-pronged and affordable detection tools brings down the cost of health care. They are important in the antenatal clinics, with added opportunity for intervention and prevention of mother to child transmission. In cooperation with rapid test kit manufacturers, SD Bioline, NPHL and NCASC, an evaluation of commercially available HIV/syphilis Duo rapid test kit (SD Bioline) to assess its performance and operational characteristics was done in the present study. Method A prospective laboratory-based cross sectional study was conducted at a large Women’s Hospital. Ten thousand pregnant women, visiting the Hospital for antenatal care or for delivery, were enrolled in study. Tests were performed by the SD Bioline HIV/Syphilis Duo kit as well as national algorithm for HIV and syphilis diagnosis which were considered gold standard. Sensitivity, Specificity, positive predictive value and negative predictive value along with kappa coefficient were calculated for the kit under evaluation. Result The sensitivity, specificity, Negative predictive value and Positive predictive value of the kit for HIV diagnosis were 100 % (95 % CI 83.18–100 %, 99.96–100 %, 83.18–100 %, and 99.96–100 %, respectively). Kappa value was found to be 1.0. Out of total cases, results of 9985 (99.85 %) cases were concordant with National algorithm for syphilis diagnosis. Thirteen (0.13 %) cases were found false positive while two were false negative. The sensitivity of the kit for syphilis diagnosis was found to be 95.45 % (95 % CI 84.86–98.74 %) and specificity was 99.87 % (95 % CI; 99.78–99.92 %). Positive predictive value was 76.36 % (95 % CI; 63.65–85.63 %) and Negative predictive value was 99.89 % (95 % CI; 99.39–99.99 %). Kappa value was found to be 0.85. Conclusion The performance characteristics of SD Bioline HIV/Syphilis duo kit were found almost concordant with the kits being used for HIV and Syphilis diagnosis separately. Its implementation in antenatal clinics/VCTs could be an added opportunity for simultaneous diagnosis of HIV and syphilis. Keywords Duo testSensitivitySpecificityKappa coefficientSD Bioline, South Koreaissue-copyright-statement© The Author(s) 2016 ==== Body Background Accurate and prompt diagnosis of HIV and syphilis is critical to disease control especially during the antenatal period with an additional opportunity for treatment and prevention of adverse effects of these disorders on the offspring. Review of literature shows that co-infection with syphilis can increase the transmission of HIV by both increasing viral shedding through open ulcers [1, 2] and by increasing patient viral load [3, 4]. Simple user-friendly two-pronged and affordable detection tools could save lives and reduce overall costs borne by patients and health systems. Rapid, accessible tests for simultaneous screening for HIV as well as syphilis are available on the market, but reliable information about their efficiency and dependability in quality or performance is coming up only gradually. Reliable evaluation reports of commercially available rapid diagnostic tests for tuberculosis [5] has been published but for syphilis and HIV large sized study reports of dual test kits are yet awaited. Screening for HIV and syphilis is a basic routine test being an inseparable part of any antenatal check-up protocol and is also endorsed in the American Congress of Obstetricians and Gynecologists (ACOG) Guidelines for Perinatal Care [6]. Simultaneous screening for both these disorders would be cost-effective as well as time-saving. The current situation in Nepal entails the practice of testing separately for syphilis by screening by RPR test and confirmation by TPHA method and testing for HIV by following the National algorithm for HIV rapid diagnosis [7]. So, a rapid dual testing kit would meet the long felt need in the field as well as in the hospital settings. Hence, in cooperation with rapid test kit manufacturers, SD Bioline/NPHL, we undertook an evaluation of commercially available HIV/syphilis Duo rapid test kit (SD Bioline) to assess its performance, reproducibility and operational characteristics and to identify its utility as a reliable screening tool. Methods The study was carried out with the cases enrolled at the antenatal clinic of Paropkar Maternity and Women’s Hospital (PMWH), Kathmandu, Nepal with the test-confirmation and backup-reference laboratory testing supported by the National Public Health Laboratory, Teku, Kathmandu, Nepal during the study period between July, 2014 and December, 2014. Paropkar Maternity and Women’s Hospital is the largest maternity and gynecological hospital of the country serving population of the local embankment area communities as well as a large number of referral cases from all over the country. The National Public Health Laboratory is the apex laboratory institution of the country serving as a general and reference laboratory especially for investigations of public health importance. All women attending the antenatal clinic of PMWH were enrolled in the study. This included all cases on their first visit to the ANC clinic as well as those on subsequent visits or for delivery, without being previously screened for HIV and/or syphilis. Women who were diagnostically confirmed cases or those already on therapeutic intervention with presumed diagnosis fell into the exclusion criteria. Informed written consent were taken in local language from all participants and ethical approval obtained from the ethical committee of the Nepal Health Research Council (NHRC). Sample collection: Two types of samples were collected depending upon suitability for the tests and patient convenience: [1] blood collection by venipuncture for serum collection and [2] capillary blood collection by finger-prick method. All procedures for specimen collection were performed by laboratory personnels taking all biosafety precautions. Test kit: SD BIOLINE HIV/syphilis Duo test kits (Standard Diagnostic Inc., Korea) were obtained for the evaluation study. Procedural steps (Fig. 1) Fig. 1 Procedural flow-chart All the collected samples were tested using SD Bioline HIV/Syphilis duo test kit and after interpreting the test results as per manufacturer’s guideline entered into the records. In addition, all samples were simultaneously tested by using national HIV testing algorithm [7] with a combination of Determine HIV-1/2/O assay (Abbott Laboratories, Abbott Park, IL), Uni-Gold Recombinant HIV-1/2 (Trinity Biotech, Bray, Ireland) HIV 1/2 Stat-Pak Ultra Fast (Chembio Diagnostic Systems, Medford, NY). Testing for syphilis was done, first by RPR test by using IMMUTREP RPR test kit (Omega Diagnostics Scotland UK) and all test positive samples were re-confirmed by TPHA method. The used SD Bioline HIV/Syphilis duo test kits were preserved for dispatching to SD Bioline Company, South Korea. Statistical analysis was carried out by using “openepi” online software (http://openepi.com/Diagnostic Test/DiagnosticTest.htm) for calculating sensitivity, specificity and predictive values. Kappa values was calculated according to the following formula: k=P0‐Pe1‐Pe Where, P0 was observed agreement and Pe was expected agreement [8]. Result Total number of samples collected were 10,000, out of which 9,453 (94.53 %) were serum samples and 547 (5.47 %) were finger-prick whole blood samples as shown in Table 1. Of the women attending the ante-natal clinic 57.71 % were already in advanced pregnancy and near to delivery.Table 1 Sample size and sample characteristics (n = 10,000) Sample collection method: Number of samples (%) Venepuncture blood samples for serum collection 9,453 (94.53 %) Finger-prick whole blood samples 547 (5.47 %) Total number of samples collected by both methods (9,453 + 547) = 10,000 Table 2 Assay results of the SD BIOLINE HIV/syphilis Duo test for dual detection of HIV and syphilis, Kathmandu, Nepal, 2014 RDT Infection RDT Result Reference methoda Positive Negative Total Using SD Bioline HIV/Syphilis duo test kit Syphilis Positive 42 13 55 Negative 2 9943 9945 Total 10,000 HIV Positive 19 0 19 Negative 0 9981 9981 Total 10,000 RDT rapid diagnostic test aTested by using national HIV testing algorithm [7] with a combination of Determine HIV-1/2/O assay (Abbott Laboratories, Abbott Park, IL), Uni-Gold Recombinant HIV-1/2 (Trinity Biotech, Bray, Ireland) HIV 1/2 Stat-Pak Ultra Fast (Chembio Diagnostic Systems, Medford, NY). Testing for syphilis was done, first by RPR test by using IMMUTREP RPR test kit (Omega Diagnostics Scotland UK) and all test positive samples were re-confirmed by TPHA method. (Thirteen (0.13 %) cases in syphilis testing were found false positive while two were false negative and in case of HIV all of 19 cases tested positive using SD Bioline HIV/Syphilis duo test kit also tested positive by the reference method.) Testing for HIV For HIV diagnosis, the sensitivity, specificity, negative predictive value and positive predictive value of the test kit were 100 % (95 % CI 83.18–100 %, 99.96–100 %, 83.18–100 %, and 99.96–100 %, respectively) and Kappa value was found to be 1.0. Information about the number of sample tested and result obtained by given methods are provided in Table 2. Testing for syphilis In the diagnosis of syphilis the test results of 9985 (99.85 %) samples using the test kit matched with that of the Reference method. Thirteen (0.13 %) cases were false positive and two were false negative. The sensitivity of the test kit for syphilis diagnosis was 95.45 % (95 % CI: 84.86–98.74 %) and specificity was 99.87 % (95 % CI: 99.78–99.92 %). Positive predictive value was 76.36 % (95 % CI; 63.65–85.63 %) and Negative predictive value was 99.89 % (95 % CI: 99.39–99.99 %) and Kappa value was found to be 0.85. Of all samples tested for syphilis by the Reference method (RPR method) using IMMUTREP RPR test kits, 44 (0.44 %) tested positive and of those tested using SD Bioline HIV/Syphilis duo test kit 55 tests were positive. Only 13 (23 %) tests were positive for syphilis in the tests using HIV/Syphilis duo test kit and 2 (4 %) tested positive for syphilis using IMMUTREP RPR test kit in the whole series. Discussion An evaluation of the diagnostic performance of the SD BIOLINE HIV/syphilis Duo test which is a two-pronged screening kit for simultaneous detection of HIV and syphilis, has been undertaken in the present study. The sensitivity and specificity of the kit in detecting HIV infection were found to be remarkably high at 100 %. These findings reiterate the results of earlier studies done in other countries [9, 10]. Comparable results have also been reported from the Southern Ethiopia, 2014 [11] study using the same test kit (shown in Table 3).Table 3 Comparison of results of Southern Ethiopia and Kathmandu, Nepal studies RDT Southern Ethiopia, 2014a Kathmandu, Nepal, 2014b SD HIV/syphilis Duo Infection % Sensitivity 95 % CI HIV 100 (98.5 to 100) 100 % (95 % CI: 83.18–100 %) Syphilis 97.6 (92.4 to 99.6) 95.45 (95 % CI: 84.86–98.74 %) % Specificity % CI HIV 99.5 (97.6 to 100) 100 % (95 % CI: 99.96–100 %) Syphilis 96 (90.6 to 98.7) 99.87 (95 % CI: 99.78–99.92 %) % PPV 95 % CI HIV 99.5 (97.6 to 100) 100 % (95 % CI: 83.18–100 %) Syphilis 95.4 (89.3 to 98.5) 76.36 (95 % CI: 63.65–85.63 %) % NPV 95 % CI HIV 100 (98.5 to 100) 100 % (95 % CI: 99.96–100 %) Syphilis 98 (93.4 to 99.7) 99.89 (95 % CI; 99.39–99.99 %) aSyphilis reference method, TPHA; HIV reference method, KHB/STAT-PAK/Unigold algorithm + ELISA. KHB, Shenghai Kehua Bioengineering; NPV, negative predictive value; PPV, positive predictive value; RDT, rapid diagnostic test. Ref: [11] bReference method: Tested by using national HIV testing algorithm [7] with a combination of Determine HIV-1/2/O assay (Abbott Laboratories, Abbott Park, IL), Uni-Gold Recombinant HIV-1/2 (Trinity Biotech, Bray, Ireland) HIV 1/2 Stat-Pak Ultra Fast (Chembio Diagnostic Systems, Medford, NY). Testing for syphilis was done, first by RPR test by using IMMUTREP RPR test kit (Omega Diagnostics Scotland UK) A rapid and sensitive tool for HIV testing has been a long felt need among health workers in the fields and the hospitals. Moreover, a two-pronged test kit designed for simultaneous HIV and syphilis detection has been a pressing need in the settings of the antenatal clinics and as a screening test for prospective mothers. SD BIOLINE HIV/syphilis Duo test has displayed many merits needed for such a test. The consequences of false negative test results in HIV testing of pregnant mothers have negative impact on efforts for the prevention of mother to child transmission of HIV infection (PMTCT). So, the results of the present evaluation as a reliable screening test for HIV without false negative test result need to be interpreted from these perspectives also. However, the SD BIOLINE HIV/syphilis Duo test’s usefulness for screening HIV-infected cases missed by other tests must be emphasized. Good screening tests do not replace the confirmatory reference tests but by their performances the consequences of false negative tests could be prevented to a great extent. In the present study, for the diagnosis of syphilis The SD BIOLINE HIV/syphilis Duo test demonstrated a sensitivity of 95.45 % (95 % CI: 84.86–98.74 %) and specificity was 99.87 % (95 % CI; 99.78–99.92 %). The positive predictive value was 76.36 % (95 % CI; 63.65–85.63 %) and negative predictive value was 99.89 % (95 % CI; 99.39–99.99 %) and Kappa value was found to be 0.85. But these results have to be interpreted in the light of the fact that the predictive value of a test depends on the prevalence of a particular infection, and that the test may also demonstrate different performance elsewhere in lower prevalence regions. However, within the constraints of limited availability of diagnostic options in many resource-limited countries the SD BIOLINE HIV/syphilis Duo test can be a strategic tool for cost effective intervention in syphilis control through antenatal screening and treatment of positive cases [12]. Simple and easy method of dual screening of HIV and syphilis as observed in case of SD BIOLINE HIV/syphilis Duo test enhances the accessibility of quality screening service for most communities of the developing countries. Besides, this presents an opportunity to be automatically tested for syphilis for all those who get HIV testing without having to wait for the results of a second test. This may be an important step towards global elimination of HIV and syphilis in the pregnant women and other vulnerable risk groups. Both HIV and syphilis being sexually transmitted diseases share common mode of transmission because of which simultaneous screening for both reinforces the control strategy of one by the other. By using this dual test kit as a point of care testing great impact can be expected in the control of both HIV and syphilis in the economically disadvantaged communities as point of care tests are strategic tools for the control and management of STIs [13]. Conclusion The performance characteristics of the SD Bioline HIV/Syphilis Duo test kit were found almost concordant with the reference kits currently being used for HIV and Syphilis diagnosis in Nepal. The performance of the kit was observed better for HIV diagnosis than for diagnosis of syphilis. Use of SD Bioline HIV/Syphilis Duo test in antenatal clinics/HTCs would meet the long felt need for a good screening tool for their day to day use with opportunity for simultaneous diagnosis of HIV and syphilis. Abbreviations AIDSAcquired immune deficiency syndrome ARTAnti-retroviral therapy DoHSDepartment of health services HIVHuman immunodeficiency virus HPHealth post HTCHIV testing and counseling NCASCNational center for AIDS and STD control NPHLNational public health laboratory PHCPrimary health care PMTCTPrevention of mother to child transmission POCTPoint of care testing RPRRapid plasma reagin WHOWorld health organization Acknowledgement We acknowledge Paropakar Maternity Hospital Management for allowing us to carry out the study in the hospital. Our thanks are due to laboratory staff of this hospital, for helping us in specimen collection, at field site. We also acknowledge National Center for AIDS and STD Control (NCASC) and Family Health Division (FHD) for their support. We appreciate the contribution made by National Public Health Laboratory (NPHL) staff for confirmation of positive cases and other reference laboratory supports. We are also thankful to Professor K. N. Shakya of Kathmandu Medical College, Kathmandu for reviewing and making important improvements in the manuscript. Last but not the least, we would like to acknowledge SD Bioline Company for giving such an opportunity to evaluate the kit in Nepal and WHO Nepal for their technical support. Funding The study was done in cooperation with NPHL and NCASC and the commercially available HIV/syphilis Duo rapid test kit (SD Bioline) was provided by rapid test kit manufacturers, SD Bioline for the present study. Availability of data and materials Approval for collection of data and materials was obtained from the Paropakar Maternity Hospital Management which is duly acknowledged. Authors’ contributions GS (Director, NPHL), principal investigator, designed, facilitated and supervised the study at each step from inception to manuscript writing and submission for publication. Any inquiry and communication should be addressed to her. Email: <geeta.nphl@gmail.com>. DRS (Epidemiologist), HCO (Epidemiologist)- Patient Counseling, Data collection, manuscript drafting. KM (Pathologist), SKM, (laboratory personnel)- Supervision in laboratory testing. CRO- Manuscript drafting. PC (gynecologist), KR (public health)- Manuscript Editing. 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 Informed written consent were taken in local language from all participants and ethical approval obtained from the ethical committee of the Nepal Health Research Council (NHRC). ==== Refs References 1. Mertz KJ Trees D Levine WC Lewis JS Litchfield B Pettus KS Morse SA St Louis ME Weiss JB Schwebke J Dickes J Kee R Reynolds J Hutcheson D Green D Dyer I Richwald GA Novotny J Weisfuse I Goldberg M O’Donnell JA Knaup R Etiology of genital ulcers and prevalence of human immunodeficiency virus coinfection in 10 US cities. The Genital Ulcer Disease Surveillance Group J Infect Dis 1998 178 1795 1798 10.1086/314502 9815237 2. Mertz KJ Weiss JB Webb RM Levine WC Lewis JS Orle KA Totten PA Overbaugh J Morse SA Currier MM Fishbein M St Louis ME An investigation of genital ulcers in Jackson, Mississippi, with use of a multiplex polymerase chain reaction assay: high prevalence of chancroid and human immunodeficiency virus infection J Infect Dis 1998 178 1060 1066 10.1086/515664 9806035 3. Buchacz K Patel P Taylor M Kerndt PR Byers RH Holmberg SD Klausner JD Syphilis increases HIV viral load and decreases CD4 cell counts in HIV-infected patients with new syphilis infections AIDS 2004 18 2075 2079 10.1097/00002030-200410210-00012 15577629 4. Kofoed K Gerstoft J Mathiesen LR Benfield T Syphilis and human immunodeficiency virus (HIV)-1 coinfection: influence on CD4 T-cell count, HIV-1 viral load, and treatment response Sex Transm Dis 2006 33 143 148 10.1097/01.olq.0000187262.56820.c0 16505739 5. World Health Organization on behalf of the Special Programme for Research and Training in Tropical Diseases Laboratory-based evaluation of 19 commercially available rapid diagnostic tests for tuberculosis 2008 6. American Congress of Obstetricians and Gynecologists (ACOG) Guidelines for Perinatal Care, Seventh Edition October 2012 is available at http://www.acog.org/resources_and_publications/. Accessed Oct 2012. 7. National Public Health Laboratory, 2011. National HIV testing algorithm. available at http://www.nphl.gov.np/index.php?obj=content&id=205. 8. Viera AJ Garrett JM Understanding interobserver agreement: the kappa statistic Fam Med 2005 37 5 360 3 15883903 9. Bristow CC Adu-Sarkodie Y Ondondo O Multisite laboratory evaluation of a dual human immunodeficiency virus (HIV)/syphilis point-of-care rapid test for simultaneous detection of HIV and syphilis infection OFID 2014 1 ofu015 25734088 10. Omoding D Katawera V Siedner M Evaluation of the SD bioline HIV/syphilis Duo assay at a rural health center in Southwestern Uganda BMC Res Notes 2014 7 746 10.1186/1756-0500-7-746 25339379 11. Shimelis T Tadesse E The diagnostic performance evaluation of the SD BIOLINE HIV/syphilis Duo rapid test in southern Ethiopia: a cross-sectional study BMJ Open 2015 5 e007371 10.1136/bmjopen-2014-007371 25908677 12. Connor N Roberts J Nicoll A Strategic options for antenatal screening for syphilis in the United Kingdom: a cost effectiveness analysis J Med Screen 2000 7 1 7 13 10.1136/jms.7.1.7 10807140 13. Saloojee H Velaphi S Goga Y Afadapa N Steen R Lincetto O The prevention and management of congenital syphilis: an overview and recommendations Bull World Health Organ 2004 82 6 424 30 15356934

PMC005xxxxxx/PMC5002178.txt

==== Front Acta Neuropathol CommunActa Neuropathol CommunActa Neuropathologica Communications2051-5960BioMed Central London 36310.1186/s40478-016-0363-yResearchRevisiting rodent models: Octodon degus as Alzheimer’s disease model? Steffen Johannes 1Krohn Markus 2Paarmann Kristin 125Schwitlick Christina 12Brüning Thomas 2Marreiros Rita 3Müller-Schiffmann Andreas 3Korth Carsten 3Braun Katharina 4Pahnke Jens +47 230 71466jens.pahnke@medisin.uio.nohttp://www.pahnkelab.eu 12561 University of Lübeck (UzL), LIED, Lübeck, Germany 2 Department of Neuro-/Pathology, University of Oslo (UiO) & Oslo University Hospital (OUS), Oslo, Norway 3 Department of Neuropathology, Heinrich Heine University Düsseldorf (HHU), Düsseldorf, Germany 4 Department of Zoology/Developmental Neurobiology, Otto von Guericke University Magdeburg (OvGU), Magdeburg, Germany 5 Leibniz Institute for Plant Biochemistry (IPB), Halle, Germany 6 Department of Pathology (PAT), Translational Neurodegeneration Research and Neuropathology Lab, University of Oslo, Postboks 4950, Nydalen, 0424 Oslo Norway 26 8 2016 26 8 2016 2016 4 1 9111 8 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.Alzheimer’s disease primarily occurs as sporadic disease and is accompanied with vast socio-economic problems. The mandatory basic research relies on robust and reliable disease models to overcome increasing incidence and emerging social challenges. Rodent models are most efficient, versatile, and predominantly used in research. However, only highly artificial and mostly genetically modified models are available. As these ‘engineered’ models reproduce only isolated features, researchers demand more suitable models of sporadic neurodegenerative diseases. One very promising animal model was the South American rodent Octodon degus, which was repeatedly described as natural ‘sporadic Alzheimer’s disease model’ with ‘Alzheimer’s disease-like neuropathology’. To unveil advantages over the ‘artificial’ mouse models, we re-evaluated the age-dependent, neurohistological changes in young and aged Octodon degus (1 to 5-years-old) bred in a wild-type colony in Germany. In our hands, extensive neuropathological analyses of young and aged animals revealed normal age-related cortical changes without obvious signs for extensive degeneration as seen in patients with dementia. Neither significant neuronal loss nor enhanced microglial activation were observed in aged animals. Silver impregnation methods, conventional, and immunohistological stains as well as biochemical fractionations revealed neither amyloid accumulation nor tangle formation. Phosphoepitope-specific antibodies against tau species displayed similar intraneuronal reactivity in both, young and aged Octodon degus. In contrast to previous results, our study suggests that Octodon degus born and bred in captivity do not inevitably develop cortical amyloidosis, tangle formation or neuronal loss as seen in Alzheimer’s disease patients or transgenic disease models. Keywords Neurodegenerative diseasesNeuropathologyRodentiaAmyloid beta-PeptidesTau proteinsAlzheimer’s diseaseAnimal modelsOctodon degushttp://dx.doi.org/10.13039/501100002347Bundesministerium für Bildung und ForschungUBICATRANS-GENBraun Katharina http://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftDFG930/9DFG930/12Pahnke Jens http://dx.doi.org/10.13039/501100005416Norges ForskningsrådNFR247179NFR260786Pahnke Jens VIAA LatviaNFI/R/2014/023Pahnke Jens http://dx.doi.org/10.13039/501100001664Leibniz-GemeinschaftSAW-2015-IPB-2Pahnke Jens http://dx.doi.org/10.13039/501100006095Helse Sør-Øst RHF2016062Pahnke Jens http://dx.doi.org/10.13039/501100001736German-Israeli Foundation for Scientific Research and DevelopmentEU FP7MC-ITN IN-SENS #607616)Korth Carsten issue-copyright-statement© The Author(s) 2016 ==== Body Introduction Senile plaques, a hallmark of Alzheimer’s disease (AD), were long suggested to initiate the destructive cascade to progressive neuronal dysfunction and death. Nowadays small, soluble oligomers of β-amyloid (Aβ) are deemed the primary toxic species [1]. These oligomers disrupt a variety of receptors [2], increase membrane permeability [2] and are suspected to induce hyperphosphorylation and aggregation of tau [3]. Physiologically, Aβ is largely eliminated from the brain by LRP1 [4] and several ABC transporters (reviewed in [5, 6]). The vast majority of cases occur sporadically and a large series of risk factors have been identified, including age, type 2 diabetes, high blood pressure, and various genetic factors like specific alleles of apolipoprotein E (APOE) [7, 8]. A small proportion of AD cases involve genetic variations which entail alterations in amount, ratio or amino acid sequence of Aβ [9]. However, these rare inherited forms are the fundament of both, disease models and our current understanding of AD. Due to a lack of alternatives, the main focus lies on the usage of these genetically manipulated research animals (thus non-sporadic AD models), which restricts the progress of research and limits the scope of detailed analyses. To successfully combat the sporadic form of AD, models that develop the disease on a more ‘natural’ basis would certainly help to understand the underlying mechanisms which are essential for developing advanced and efficient therapy options. The South American rodent Octodon degus (degu) may be a promising candidate for physiologically modelling sporadic AD, as it was reported to develop the ‘full range of AD-like pathologies’ [10] without any genetic manipulation. While wild degus have only a limited life expectancy (mean: <1 year; common max: 3–4 years), captivity lowers mortality and increases mean life span to 5–8 years [11] (Ebensperger L.A. & Hayes L.D., unpublished data). This captivity-dependent, aged phenotype combined with the highly homologous Aβ sequence [12], differing only in one amino acid from human Aβ (see Fig. 1b), might be the main reasons for their vulnerability. Thus, during the past years wild-caught and captive born degus were used in AD research [12–15] and have been referenced in numerous review publications [10, 16–22]. In this course, prominent intra- and extracellular Aβ deposits were reported in cortical and hippocampal areas of aged animals (>3 years) [12, 19]. Furthermore, APP and Aβ positive axonal bulbs were observed in hippocampal white matter tracts of old animals (6 years) which preceded cerebral amyloid angiopathy [13]. Biochemical analysis of aged degus (5-years-old) indicated a correlation between Aβ*56 oligomers and tau phosphorylation on the one hand and decreased synaptic plasticity and impaired memory performance on the other hand [14]. However, the most recent study, analysing young (1 to 3 years old) and old animals (4 to 6 years old), found AD associated autophagy markers LC3 and p62 unchanged [15]. Furthermore, GFAP (glial fibrillary acidic protein), CD11b expression, oxidative stress and apoptosis markers were unaffected in cortices but elevated in hippocampi of old animals. Cortical and hippocampal levels of AD-linked IL-6 seemed increased in old degus [15].Fig. 1 Comparison of human and rodent Aβ sequences. Part a shows a simplified phylogenetic tree of rodents. In b, human and rodent sequences of Aβ42 are compared. The Aβ sequence of guinea pigs (green) equals the human sequence. Chinchillas, degus and naked mole rats (NMRs) share the same sequence (yellow) with one variation at position 13 as compared to humans. The lesser Egyptian jerboa has an additional difference at position 10 (orange). A larger group of rodents, including mice and rats, show 3 sequential differences at 5, 10 and 13 (red). This sequence is often erroneously referred to as “rodent Aβ”, as it is the most frequent sequence in rodents. The uniformity of human and degu tau at the analysed phosphorylation sites is shown in c The aim of the present study was to critically re-evaluate the suitability of degus as ‘natural’ AD model. To characterize amyloidosis and tau deposition, different cortical and hippocampal regions of young (1-year-old) and aged (5-years-old) wild-type, colony-bred degus were screened for neurodegenerative changes. Materials and methods Drugs and chemicals used in the study were purchased from Carl Roth, Karlsruhe, Germany; except for those specifically mentioned. Animals Wild-type degus used in the present study were bred at the Institute of Biology, Otto-von-Guericke-University Magdeburg. Degus were housed in a climate-controlled environment (22 °C, 55 ± 5 % humidity) on a 12 h light/dark cycle with an enriched environment. Social enrichment was provided by housing in same sex groups of up to 4 animals. Sensory, motor and cognitive enrichment was provided by housing the animals in large cages (510x420x680 mm3, EBECO, Castrop Rauxel, Germany), equipped with a drinking bottle, burrows for hiding, a running wheel (Europet-Bernina International, Iserlohn, Germany) for physical exercise and material for nest building. Nutritional enrichment was provided by providing dry bread, fresh vegetables and fresh tree branches in addition to the regular pellet food (rat diet pellets and cereals; ssniff Spezialdiäten, Soest, Germany) ad libitum. A first cohort of degus was immunohistochemically analysed (J.P.) in two groups: young (12 months) and aged (60 months), both sexes, at least two animals per sex and age group. A second cohort from the same colony was independently analysed in another neuropathological laboratory (C.K.) and used for quantitative and western blot analyses: seven animals (age in months: 3, 24, 25, 56, 56, 65, and 65) Tg-mice, harbouring mutant human amyloid precursor protein (KM670/671NL) and presenilin 1 (L166P) both driven by murine Thy1.2-promotor [23], were kindly provided by the University of Tübingen, Germany. All mice were housed in a climate-controlled (22 °C) environment on a 12 h light/dark cycle in same sex groups of up to 4 animals and free access to food and water. All experiments were conducted in accordance to the EU and state law of Saxony-Anhalt and approved by the local animal ethics committee. Sequences & alignments Protein sequences were gathered using NCBI Protein database (http://www.ncbi.nlm.nih.gov/protein; for accession numbers see Table 1) and protein alignments were performed using BLASTP 2.2.30+ [24, 25].Table 1 Accession numbers of featured proteins from the NCBI Protein database (http://www.ncbi.nlm.nih.gov/protein/) Species Accession number Amyloid-beta A4 protein isoform a precursor  Homo sapiens NP_000475.1  Cavia porcellus XP_003467233.1  Chinchilla lanigera XP_005375649.1  Octodon degus XP_004627753.1  Heterocephalus glaber XP_004898345.1  Jaculus jaculus XP_004654437.1  Mus musculus NP_001185752.1  Rattus norvegicus NP_062161.1  Cricetulus griseus ERE75573.1  Mesocricetus auratus XP_005073973.1  Peromyscus maniculatus XP_006988006.1  Microtus ochrogaster XP_005345348.1 Microtubule-associated protein tau  Homo sapiens NP_001116538.2  Octodon degus XP_004630049 Immunohistochemistry Animals were sacrificed by cervical dislocation and immediately perfused with 50 mL PBS followed by 50 mL PFA for fixation. Paraffin-embedded, 4 μm-thick coronal sections were deparaffinised and stained using Haematoxylin and Eosin. Immunohistochemical analyses were performed using Bond-Max™ (Leica Microsystems, Wetzlar, Germany) automated staining system as described previously [5, 26–28]. Epitope retrieval was carried out as follows: 5 min in 95 % (v/v) formic acid for 6F3D, 4G8 and 6E10; 20 min in EDTA buffer pH 9.0 for IBA1 and AT180; 10 min enzymatic digestion (Bond Enzyme Pretreatment Kit, Leica Biosystems Nussloch, Nussloch, Germany) for GFAP or 20 min in citric acid buffer pH 6.0 for NeuN, AT8 and AT100. Antibodies against ionized calcium-binding adapter molecule 1 (IBA1 1:1000, 019-19741, Wako Chemicals, Neuss, Germany), glial fibrillary acid protein (GFAP, 1:500, Z033401, Dako Deutschland, Hamburg, Germany), neuronal nuclear antigen (NeuN, 1:500, MAB377, Merck Chemicals, Darmstadt, Germany) phosphorylated tau (AT8, 1:50, MN1020; AT100, 1:500, MN1060; AT180, 1:50, MN1040, Thermo Fisher Scientific, Waltham, MA, USA); β-amyloid (6E10, 1:100, SIG-39320, Covance Research Products, Denver. PA, USA; 4G8, 1:2000, SIG-39220, Covance Research Products, Denver. PA, USA; 6F3D, 1:100, M0872, Dako Deutschland, Hamburg, Germany) were used according to manufactures instructions. Slides were developed using Bond™ Polymer Refine Detection kit (Leica Microsystems, Wetzlar, Germany). For Campbell-Switzer staining, paraffin-embedded, 4 μm-thick coronal sections were deparaffinised, stirred for 5 min in ammonium hydroxide and washed twice in distilled water for 1 min. Sections were incubated for 40 min in Silver-Pyridine-Carbonate solution (14 % (v/v) pyridine, 0.49 % (w/v) silver nitrate, 0.37 % (w/v) potassium carbonate) followed by 3 min in citric acid and washed in acetate buffer (33.6 mM sodium acetate, 14.4 mM acetic acid, pH 4.99). Sections were developed in developer solution (236 mM sodium carbonate, 12.5 mM ammonium nitrate, 5.9 mM silver nitrate, 1.7 mM tungstosilicic acid, 0.87 mM formaldehyde) under a light source for about 6 min. Sections were washed three times in acetate buffer and once in distilled water, each for 30 s. Sections were finally placed in 0.5 % (w/v) sodium thiosulfate solution for 45 s, washed twice in distilled water and subsequently mounted. Slides were digitized using MIRAX MIDI Scanner (Carl Zeiss MicroImaging, Jena, Germany). The scanned slides were processed using the free Pannoramic Viewer software (3DHISTECH, Budapest, Hungary), and analysed under blinded conditions using either AxioVision (Carl Zeiss Imaging Solutions; Munich, Germany) or ImageJ (Wayne Rasband, National Institutes of Health, Bathesda, MD, USA) and the ITCN plugin (Thomas Kuo and Jiyun Byun, University of California, CA, USA). Four-step fractionation and quantification of Aβ Animals were sacrificed by cervical dislocation and transcardially perfused with PBS. The brain was immediately snap-frozen in liquid nitrogen and stored at −80 °C. Fractionation of brain tissue was performed by preparative ultracentrifugation as described previously [29]. In brief, brain tissue was homogenated in 9 volumes of TBS buffer (150 mM sodium chloride, 50 mM Tris, pH 7, supplemented with protease inhibitor, Complete EDTA-free, Roche, Basel, Switzerland) and subsequently centrifuged (100,000 g, 4 °C, 1 h). The supernatant (TBS-fraction, soluble Aβ) was harvested, pellet was sonicated in 100 μl TBS/ 1 % (v/v) Triton X-100 and centrifuged again (100,000 g, 4 °C, 1 h). The supernatant (TX-100-fraction, detergent soluble Aβ) was harvested, the pellet was sonicated in TBS/ 2 % (w/v) SDS and centrifuged (100,000 g, room temperature, 1 h). Supernatant (SDS-fraction, protein bound Aβ) was harvested and the remaining pellet was finally resolved in 70 % formic acid (FA-fraction, insoluble Aβ). For quantification, SDS-fraction was diluted 20-fold in TBS and FA-fraction was neutralized with 19 volumes of 1 M Tris (pH 11). Quantification was performed using the Human/Rat β-amyloid (40) or (42) ELISA Kit (Wako Chemicals, Neuss, Germany) according to the manufacturer's instructions, which uses BNT77 (epitope: amino acids 11-16) and BA27 (Aβ40-specific) or BC05 (Aβ42-specific), respectively [30]. Tau preparation Brain tissue was mixed with 9 volumes of TBS buffer (supplemented with protease and phosphatase inhibitor, Sigma-Aldrich, St. Louis, MO, USA), homogenized (total fraction) and centrifuged (130,000 g, 4 °C, 20 min). The pellet was resuspended in RIPA buffer (150 mM sodium chloride, 50 mM Tris, 0.5 % (w/v) deoxycholic acid, 1 % (v/v) Triton X-100, 0.5 % (w/v) SDS, 25 mM EDTA, pH 8 supplemented with protease and phosphatase inhibitor) and centrifuged again (130,000 g, 4 °C, 20 min). Supernatant was discarded; pellet was resuspended in 70 % formic acid and centrifuged (130,000 g, 4 °C, 20 min). Supernatant (FA-fraction) was neutralized with two volumes of neutralization buffer (5 M sodium hydroxide, 0.5 M Tris, 0.25 M monosodium phosphate). Proteins were precipitated using 2,2,2-trichloroacetic acid (25 % (v/v) TCA) for 30 min at 4 °C and subsequently separated by centrifugation (22,000 g, 4 °C, 15 min). Resulting pellet (insoluble fraction) was washed two times with cold acetone and subsequently air-dried. Western blot Total and insoluble fractions were separated on a reducing 4–12 % Bis-Tris gel (Thermo Fisher Scientific, Waltham, MA, USA) and subsequently transferred to nitrocellulose membrane. Phosphorylated tau was detected using AT8 (1:1,000, MN1020, Thermo Fisher Scientific, Waltham, MA, USA) and total tau levels using HT7 antibody (1:1,000, MN1000, Thermo Fisher Scientific, Waltham, MA, USA), and protein expression was normalized with internal control anti-actin (1:10,000, Sigma-Aldrich, St. Louis, MO, USA). Statistics For all quantifications, animals of either sex were used (n ≥ 2 per sex; n ≥ 4 per group), statistical significance (p ≤ 0.05) was determined using unpaired t-tests with Welch’s correction (Prism 6, GraphPad Software La Jolla, CA, USA). Results Routine histological Haematoxylin and Eosin (H&E) stains of young (1 year) and old (5 years) wild-type degus revealed normal age-related changes in the examined brain regions, but no obvious signs for specific lesions, neurodegeneration, or neuronal displacement (Fig. 2a, b). Neuronal marker NeuN unveiled no generalized loss of cortical neurons between both time points (Fig. 2c, d, m). Microglia-specific stains of ionized calcium-binding adapter molecule 1 (IBA1), which can unravel early signs of pathology shown by localised microglia accumulation, displayed homogenously distributed populations of resting microglia in young and aged animals (Fig. 2e, f). Individual differences in young and aged degus were apparent, but the cortical coverage was not significantly different between the two groups (Fig. 2n). Cortical clustering of microglia as seen in pre-depositing APP-transgenic mice, pinpointing towards starting Aβ deposition, was not detected. Cortical astroglia (GFAP-positive) were nearly exclusively located in cortical layer 1 and around blood vessels (Fig. 2g, h, o), without any age-dependent changes in spatial distribution or intensity.Fig. 2 Immunohistochemical analysis of young and aged degus. H&E stain revealed normal age-related changes but no signs for lesions, neurodegeneration, or displacement in young (1-year-old, a) and aged (5-years-old, b) animals. Density of cortical neurons (NeuN-satin) remained virtually unchanged in aged degus (d), compared to young (c). IBA1-stain (e, f) revealed homologues populations of resting microglia cells (young, e; aged, f)). GFAP Immunoreactivity was slightly decreased in aged animals (h), but spatial distribution (layer 1, surrounding vessels) was similar (young: g; aged: h). Campbell-Switzer stain unveiled neither extracellular plaques nor tangles (i, j). Thioflavin T likewise indicated no amyloid plaques (young, k; aged, l). Semi-automatically determination of neurons (m) as well as microglial cells (n) and astrocytes (o) in cortices revealed no significant changes in aged animals. Scale bars = 100 μm. Data is presented as mean ± SEM (n ≥ 4) To evaluate the extent of any cortical amyloidosis and neuronal destruction, a modified Campbell-Switzer stain was applied, but no extracellular deposits were exposed (Fig. 2i, j). This result was supported by thioflavin T stains which revealed no specific cortical fluorescence as well (Fig. 2k, l). To examine potentially undiscovered amyloid deposits, we performed more sensitive immunohistochemical stains by employing commonly used antibodies against different Aβ epitopes (clones 6F3D, 4G8, 6E10). The epitope of clone 6E10 is located N-terminal of the H13R substitution (within amino acids 3-8). Besides high unspecific background staining, limited intracellular immunoreactivity was detected in all cortices of young and aged degus. However, extracellular deposits (e.g. plaques) could not be traced in any of the examined brain regions (Fig. 3a, b). The epitope of anti-Aβ antibody clone 6F3D (amino acids 8-17) includes the H13R substitution and showed neither intra- nor extracellular immunoreactivity in young or aged animals (Fig. 3c, d). Likewise, no aggregates could be detected using the 4G8 antibody (Fig. 3e, f) with an epitope C-terminally of the H13R substitution (amino acids: 18-22). The lack of age-dependent, immunohistological changes in degus was independently confirmed by a second neuropathological laboratory (C.K.) in an additional study (1 to 5 year old degus from the same colony, anti-amyloid stains using clones 6E10 and IC16; data not shown).Fig. 3 β-amyloid pathology in young and aged degus. Aβ staining using 6E10-antibody (a, b) resulted in unspecific background signals accompanied by spot-like intracellular immunoreactivity. Although visible in both, young (a) and aged (b), intensity generally tends to be elevated in aged animals. In contrast, no immunoreactivity was detected in young (c, e) or aged (d, f) degus using 6F3D (c, d), or 4G8 (e, f). Scale bar = 100 μm Levels of cortical and hippocampal Aβ40 and Aβ42 in young and aged degus were quantitatively measured using immunoassays (Fig. 4), revealing very low levels of soluble and membrane-bound Aβ and low levels of protein-bound and insoluble Aβ. The levels of insoluble Aβ did not age-dependently change. However, the concentration of insoluble Aβ was generally several magnitudes below those of established AD mouse models, and even lower than those of wild-type naked mole rats and guinea pigs (Table 2). The comparison of wild-type degus, wild-type and transgenic mice demonstrates that aged wild-type degus very closely resemble histological parameters of wild-type mice in terms of Aβ deposition and unspecific activation of microglial and astroglial cells (Fig. 5).Fig. 4 Aβ levels in fractionated brain tissue of young and aged degus. Levels of Aβ40 and Aβ42 were measured in fractionated cortical (black) and hippocampal (grey) tissue of young (a, c) and aged (b, d) degus using immunoassays. a, b In both groups, Aβ40 was rarely present in soluble (TBS) and membrane-bound (TX-100) forms. The highest amounts were protein-bound (SDS) and smaller proportions were insoluble (formic acid; FA). Overall, young and aged animals demonstrated very similar levels Aβ40. c, d Young and aged degus showed low levels of Aβ42 in soluble (TBS) and membrane-bound (TX-100) fractions and higher levels in protein-bound (SDS) and insoluble (FA) fractions in both, cortex and hippocampus. Aβ42 levels were likewise not crucially changed in aged animals. Data is presented as mean ± SEM (young: 3, 24 months; aged: 56, 56, 65, 65 months) Table 2 Levels of insoluble Aβ in transgenic AD models and wild-type rodents Model Age (months) Aβ40 (pg/mg) Aβ42 (pg/mg) Reference AD models APP23 12 3098 746 [54] APPLondon 24 1300 3300 [55] Tg2576 2–3 700 2100 [56] >20 39,900 40,900 [57] APP/PS1 5 26,200 49,400 [58] 8 166,000 113,500 Wild-type naked mole rats 2–9 years 37 60 [34] Guinea pig adult 79 18 [59] Fig. 5 Comparison of neuropathological changes in wild-type degus, wild-type and transgenic mice. While aged wild-type degus (a, 5 years old) and mice (b, 200-days-old) exhibit no sign of β-amyloid deposition, APP/PS1 transgenic mice present with obvious cortical amyloidosis (c, 150-days-old). Compared to wild-type degus (d, g) and mice (e, h), transgenic mice manifest with pronounced micro- (f) and astrogliosis (i). Scale bars = 200 μm Phosphorylated tau is the second protein accumulating during disease progression and another histopathological hallmark of AD [31]. We utilized antibodies against different epitopes of phosphorylated tau to screen for neurofibrillary tangles. Although sequences of human and degu tau vary, the analysed phosphorylation sites (Ser202/Thr205, Thr212/Ser214 and Thr231) are identical (Fig. 1c). AT8 (Ser202/Thr205) labelled cortical neurons in young and aged animals (Fig. 6a, b) and AT100 (Thr212/Ser214) showed nuclear-localized reactivity (Fig. 6c, d). AT100 epitope is known for nuclear co-localization and considered not tau-specific, as it appears likewise in tau knockout mice [32]. AT180 (Thr231) equally stained cortical neurons of young and aged degus (Fig. 6e, f). In sum, the detected tau did morphologically not correspond to neurofibrillary tangles and showed no age-dependent intensification.Fig. 6 Tau pathology in young and aged degus. Phosphoepitope-specific antibodies were used to examine involvement of tau. AT8 revealed intracellular immunoreactivity in young and aged degus (a, b). AT100 was unspecific and appeared almost entirely in the nucleus (c, d). AT180-labelled cortical neurons in young and aged animals (e, f). Scale bar = 100 μm Moreover, the independent quantification of tau revealed neither elevated levels of total tau, nor an increase in phosphorylated or insoluble fraction (Fig. 7).Fig. 7 Total and insoluble levels of cortical tau in young and aged degus. Biochemical analysis of total and insoluble tau in cortices of young and aged degus revealed high individual variance, but no significant age-related changes in levels of neither total (HT7) nor phosphorylated (AT8) tau. β-actin served as loading control. Age of degus (left to right in months): 3, 24, 25, 56, 56, 65, 65 Discussion In the near future, aging societies will be particularly challenged by age-related diseases demanding intensive care. Exceptional research endeavours are necessary to face these approaching challenges. Thus, reliable and accurate animal models are one of the major prerequisites in basic research. Although transgenic rodent models of AD are constantly refined, no model yet mimics all pathological features of this complex disease. More accurate models maintaining the beneficial characteristics of rodents would lead to a better understanding and more expedient therapeutic approaches. Degus were described as a promising natural model of Alzheimer’s disease during the last years by a Chilean group. However, in our studies we were unable to detect any systematic occurrence of the typical histopathological hallmarks of AD in relation to age. Haematoxylin and Eosin as well as NeuN stains showed not more than slight differences between young and aged degus, rather indicative for the natural aging process than pathological neurodegeneration. In contrast to previous results from Inestrosa et al. [15], we could not detect elevated GFAP expression in old degus compared to young animals. The lack of signs for microglial and astrocytic inflammation further reiterates the absence of a pathological degeneration in the examined brains of the old degus. Aβ pathology The sparse intracellular reactivity of anti-Aβ clone 6E10, which was lacking in clone 4G8, clone 6F3D, Campbell-Switzer and thioflavin T stains, most likely indicates an unspecific reaction [33]. In contrast, no sign of any extracellular deposition of Aβ was detected in aged animals by any of the applied staining methods (Fig. 3). Quantitative measurements underpinned the absence of considerable amounts of insoluble Aβ (Fig. 4) and revealed Aβ-levels that are in the same range as those in wild-type mice [29] and below those of wild-type naked mole rats [34]. Consistent with results of van Groen et al. no significant neuronal loss was found in the brain of 5-years-old degus [13]. These findings are in sharp contrast to observations in brains of degus obtained from their natural habitat, in which prominent intra- and extracellular Aβ deposits in cortices and hippocampi of aged animals (>3 years) were reported [12, 19]. These differences may, at least in part, be caused by different rearing conditions (laboratory housing versus natural wildlife conditions) and it has to be considered that in their wildlife habitat the animals are exposed to stress, may suffer from hypertension, viral infections and diabetes, i.e. known risk factors contributing to the aetiology of AD and the early development of AD-type neuropathology. Furthermore, the single amino-acid-difference between degus and humans at position 13 (histidine to arginine) affects a histidine residue (His13) which is crucial for aggregation and toxicity of Aβ. His13 is involved in early N-terminal β-sheet formation [35] and a substitution lowers aggregation propensity [36], neuronal binding [37], and cytotoxicity [36]. Moreover, His13 is involved in the coordination of metal ions [38] and methylation or substitution by arginine, as seen in degus, lowers the affinity for metal ions and thus depletes aggregation [38–40] and attenuates toxicity [41, 42] of Aβ. Two other species which are related to degus share a similar Aβ sequence, but, despite higher life expectancies, lack the neuropathological features as reported for degus. Naked mole rats (Heterocephalus glaber) have the identical Aβ sequence (see Fig. 1) and an exceptional lifespan of more than 30 years [34]. Although young naked mole rats naturally exhibit pronounced oxidative stress [43] and Aβ levels similar to 3xTg-AD mice [34, 43], they do not develop amyloid plaques with age [34]. Furthermore, naked mole rats even present with high levels of phosphorylated tau without any tangle formation [44]. In Guinea pigs (Cavia porcellus), with a human identical Aβ sequence (see Fig. 1) and a lifespan similar to degus (average 5–7 [45]), dense amyloid deposits do not occur [45], despite similar APP processing [46, 47] and high β–secretase activity [47]. Tau pathology The additional screening for tau deposition, the second aggregating protein in AD, revealed similar intracellular reactivity in young and aged degus using phosphoepitope-specific antibodies AT8 and AT180. AT100 staining showed the previously described, unspecific nuclear localization [32]. Biochemical analysis did not reveal an age-dependent increase of total, insoluble or phosphorylated tau (Fig. 7). Some variability observed in the levels of total tau or insoluble tau could hint subsets with different aggregation propensities but the very same animals did not exhibit tau pathology in IHC, and larger number of animals would be needed to identify the existence of such subsets. Hence, no evidence for a pathological deposition of tau could be detected in the examined animals. Methodological considerations The animals used in a variety of studies were collected from different sources [13, 48, 49], including animals caught in the wild [12, 50], the latter does usually not allow a precise age determination and thus hinders precisely controlled analyses. However, standardised housing conditions as used for the degus examined in the present study seem to prevent the development of ‘AD-like’ pathology described for wild-caught animals. Nonetheless, it would be fascinating to decipher the factors inducing the histopathological and biochemical changes in degus previously described [12, 13, 15]. Moreover, not only the particular species or the specific amino acid sequence seems influence the extent of amyloid deposition, but the genetic background likewise enfolds a strong effect [26]. As degus are not yet an established research model, they lack a defined, stable and characterized inbred genetic background. An interesting approach of separating old degus with severely aberrant behaviour as disease model takes high individual variability into account and indicates that ‘AD-like’ pathology might not necessarily develop in old degus. However, even in this selected subgroup with increased levels of inflammatory and oxidative stress markers, no correlation between altered behaviour and specific neuropathological symptoms could be established [50]. The stated impairments of spatial memory and cognition in aged degus [14] may therefore be just a part of the normal aging process, since physiologic aging is linked to significant impairments in memory [51], cognition [52] and hippocampal long-term potentiation [53] in mice as well. Thus, symptoms of normal aging may not be misinterpreted to model AD. Conclusion Octodon degus was re-evaluated in the context of existing rodent AD models and human AD pathology. Performing immunohistological and molecular analyses of young and aged animals, we were able to show exclusively normal age-related cortical changes without indications for extensive degeneration as seen in patients with dementia and transgenic AD mice. Neither significant neuronal loss nor enhanced microglial activation were observed in aged animals of our degu population. Furthermore, no amyloid accumulation or tangle formation as seen in sporadic Alzheimer’s disease patients could be determined. Phosphoepitope-specific antibodies against tau species displayed similar intracellular neuronal reactivity in both, young and aged degus. Moreover, we highlighted some previous results, which stand in contrast to the assumption of degus as natural AD model and seem to be thus far neglected. Bearing that in mind, assessment of degus as AD models should be meticulously done and receive particular attention. Currently, it is not clear if unnoticed environmental or rearing factors might play a role in triggering AD-like neuropathology. Therefore, we conclude that presently, the rodent Octodon degus is neither a superior model which is more suitable than other frequently used rodent models nor a ‘natural’ model of Alzheimer’s disease in general. Acknowledgements The work was financed by the following grants to J.P. (Deutsche Forschungsgemeinschaft DFG PA930/9, DFG PA930/12; VIAA Latvia NFI/R/2014/023; Leibniz Society SAW-2015-IPB-2; HelseSØ No:2016062; Norsk forskningsrådet: NFR251290, NFR246392, NFR248772, JPND NeuroGEM NFR247179, JPND PROP-AD NFR260786), K.B. (German-Israeli Foundation for Scientific Research and Development (GIF), BMBF Grants UBICA, TRANS-GEN). and C.K. (BMBF #01GQ1422A EU-FP7 MC-ITN IN-SENS #607616). Authors’ contributions JS, MK, KP, CS conducted investigations, wrote manuscript; TB conducted investigations; RM, AMS, CK: conducted experiments, wrote manuscript; KB provided O. degu cohort, wrote manuscript; JP planned study, conducted investigations, wrote manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Glabe CG Common mechanisms of amyloid oligomer pathogenesis in degenerative disease Neurobiol Aging 2006 27 4 570 575 10.1016/j.neurobiolaging.2005.04.017 16481071 2. Kayed R Lasagna-Reeves CA Molecular mechanisms of amyloid oligomers toxicity J Alzheimers Dis 2013 33 Suppl 1 S67 S78 22531422 3. Choi SH Kim YH Hebisch M Sliwinski C Lee S D'Avanzo C Chen H Hooli B Asselin C Muffat J Klee JB Zhang C Wainger BJ Peitz M Kovacs DM Woolf CJ Wagner SL Tanzi RE Kim DY A three-dimensional human neural cell culture model of Alzheimer's disease Nature 2014 515 7526 274 278 10.1038/nature13800 25307057 4. Pflanzner T Janko MC Andre-Dohmen B Reuss S Weggen S Roebroek AJ Kuhlmann CR Pietrzik CU LRP1 mediates bidirectional transcytosis of amyloid-beta across the blood-brain barrier Neurobiol Aging 2011 32 12 2323 10.1016/j.neurobiolaging.2010.05.025 20630619 5. Krohn M Lange C Hofrichter J Scheffler K Stenzel J Steffen J Schumacher T Bruning T Plath AS Alfen F Schmidt A Winter F Rateitschak K Wree A Gsponer J Walker LC Pahnke J Cerebral amyloid-beta proteostasis is regulated by the membrane transport protein ABCC1 in mice J Clin Invest 2011 121 10 3924 3931 10.1172/JCI57867 21881209 6. Pahnke J Frohlich C Paarmann K Krohn M Bogdanovic N Arsland D Winblad B Cerebral ABC transporter-common mechanisms may modulate neurodegenerative diseases and depression in elderly subjects Arch Med Res 2014 45 8 738 743 10.1016/j.arcmed.2014.10.010 25446622 7. Reitz C Mayeux R Alzheimer disease: epidemiology, diagnostic criteria, risk factors and biomarkers Biochem Pharmacol 2014 88 4 640 651 10.1016/j.bcp.2013.12.024 24398425 8. Hollingworth P Harold D Sims R Gerrish A Lambert JC Carrasquillo MM Abraham R Hamshere ML Pahwa JS Moskvina V Dowzell K Jones N Stretton A Thomas C Richards A Ivanov D Widdowson C Chapman J Lovestone S Powell J Proitsi P Lupton MK Brayne C Rubinsztein DC Gill M Lawlor B Lynch A Brown KS Passmore PA Craig D McGuinness B Todd S Holmes C Mann D Smith AD Beaumont H Warden D Wilcock G Love S Kehoe PG Hooper NM Vardy ER Hardy J Mead S Fox NC Rossor M Collinge J Maier W Jessen F Ruther E Schurmann B Heun R Kolsch H van den Bussche H Heuser I Kornhuber J Wiltfang J Dichgans M Frolich L Hampel H Gallacher J Hull M Rujescu D Giegling I Goate AM Kauwe JS Cruchaga C Nowotny P Morris JC Mayo K Sleegers K Bettens K Engelborghs S De Deyn PP Van Broeckhoven C Livingston G Bass NJ Gurling H McQuillin A Gwilliam R Deloukas P Al-Chalabi A Shaw CE Tsolaki M Singleton AB Guerreiro R Muhleisen TW Nothen MM Moebus S Jockel KH Klopp N Wichmann HE Pankratz VS Sando SB Aasly JO Barcikowska M Wszolek ZK Dickson DW Graff-Radford NR Petersen RC Alzheimer's Disease Neuroimaging I van Duijn CM Breteler MM Ikram MA DeStefano AL Fitzpatrick AL Lopez O Launer LJ Seshadri S Consortium C Berr C Campion D Epelbaum J Dartigues JF Tzourio C Alperovitch A Lathrop M Consortium E Feulner TM Friedrich P Riehle C Krawczak M Schreiber S Mayhaus M Nicolhaus S Wagenpfeil S Steinberg S Stefansson H Stefansson K Snaedal J Bjornsson S Jonsson PV Chouraki V Genier-Boley B Hiltunen M Soininen H Combarros O Zelenika D Delepine M Bullido MJ Pasquier F Mateo I Frank-Garcia A Porcellini E Hanon O Coto E Alvarez V Bosco P Siciliano G Mancuso M Panza F Solfrizzi V Nacmias B Sorbi S Bossu P Piccardi P Arosio B Annoni G Seripa D Pilotto A Scarpini E Galimberti D Brice A Hannequin D Licastro F Jones L Holmans PA Jonsson T Riemenschneider M Morgan K Younkin SG Owen MJ O'Donovan M Amouyel P Williams J Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated with Alzheimer's disease Nat Genet 2011 43 5 429 435 10.1038/ng.803 21460840 9. Schellenberg GD Montine TJ The genetics and neuropathology of Alzheimer's disease Acta Neuropathol 2012 124 3 305 323 10.1007/s00401-012-0996-2 22618995 10. Braidy N Munoz P Palacios AG Castellano-Gonzalez G Inestrosa NC Chung RS Sachdev P Guillemin GJ Recent rodent models for Alzheimer's disease: clinical implications and basic research J Neural Transm (Vienna) 2012 119 2 173 195 10.1007/s00702-011-0731-5 22086139 11. Ebensperger LA Tapia D Ramirez-Estrada J Leon C Soto-Gamboa M Hayes LD Fecal cortisol levels predict breeding but not survival of females in the short-lived rodent, Octodon degus Gen Comp Endocrinol 2013 186 164 171 10.1016/j.ygcen.2013.02.044 23524002 12. Inestrosa NC Reyes AE Chacon MA Cerpa W Villalon A Montiel J Merabachvili G Aldunate R Bozinovic F Aboitiz F Human-like rodent amyloid-beta-peptide determines Alzheimer pathology in aged wild-type Octodon degu Neurobiol Aging 2005 26 7 1023 1028 10.1016/j.neurobiolaging.2004.09.016 15748782 13. van Groen T Kadish I Popovic N Popovic M Caballero-Bleda M Bano-Otalora B Vivanco P Rol MA Madrid JA Age-related brain pathology in Octodon degu: blood vessel, white matter and Alzheimer-like pathology Neurobiol Aging 2011 32 9 1651 1661 10.1016/j.neurobiolaging.2009.10.008 19910078 14. Ardiles AO Tapia-Rojas CC Mandal M Alexandre F Kirkwood A Inestrosa NC Palacios AG Postsynaptic dysfunction is associated with spatial and object recognition memory loss in a natural model of Alzheimer's disease Proc Natl Acad Sci U S A 2012 109 34 13835 13840 10.1073/pnas.1201209109 22869717 15. Inestrosa NC Rios JA Cisternas P Tapia-Rojas C Rivera DS Braidy N Zolezzi JM Godoy JA Carvajal FJ Ardiles AO Bozinovic F Palacios AG Sachdev PS Age Progression of Neuropathological Markers in the Brain of the Chilean Rodent Octodon degus, a Natural Model of Alzheimer's Disease Brain Pathol 2015 25 6 679 691 10.1111/bpa.12226 25351914 16. Ardiles AO Ewer J Acosta ML Kirkwood A Martinez AD Ebensperger LA Bozinovic F Lee TM Palacios AG Octodon degus (Molina 1782): a model in comparative biology and biomedicine Cold Spring Harb Protoc 2013 2013 4 312 318 10.1101/pdb.emo071357 23547147 17. Tarragon E Lopez D Estrada C Ana GC Schenker E Pifferi F Bordet R Richardson JC Herrero MT Octodon degus: a model for the cognitive impairment associated with Alzheimer's disease CNS Neurosci Ther 2013 19 9 643 648 10.1111/cns.12125 23710760 18. Iqbal K Bolognin S Wang X Basurto-Islas G Blanchard J Tung YC Animal models of the sporadic form of Alzheimer's disease: focus on the disease and not just the lesions J Alzheimers Dis 2013 37 3 469 474 23948903 19. Rios JA Cisternas P Arrese M Barja S Inestrosa NC Is Alzheimer's disease related to metabolic syndrome? A Wnt signaling conundrum Prog Neurobiol 2014 121 125 146 10.1016/j.pneurobio.2014.07.004 25084549 20. Castro-Fuentes R Socas-Perez R Octodon degus: a strong attractor for Alzheimer research Basic Clin Neurosci 2013 4 1 91 96 25337333 21. Braidy N Poljak A Jayasena T Mansour H Inestrosa NC Sachdev PS Accelerating Alzheimer's research through 'natural' animal models Curr Opin Psychiatry 2015 28 2 155 164 25602247 22. Rivera DS Inestrosa NC Bozinovic F On cognitive ecology and the environmental factors that promote Alzheimer disease: lessons from Octodon degus (Rodentia: Octodontidae) Biol Res 2016 49 10 10.1186/s40659-016-0074-7 26897365 23. Radde R Bolmont T Kaeser SA Coomaraswamy J Lindau D Stoltze L Calhoun ME Jaggi F Wolburg H Gengler S Haass C Ghetti B Czech C Holscher C Mathews PM Jucker M Abeta42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology EMBO Rep 2006 7 9 940 946 10.1038/sj.embor.7400784 16906128 24. Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 17 3389 3402 10.1093/nar/25.17.3389 9254694 25. Altschul SF Wootton JC Gertz EM Agarwala R Morgulis A Schaffer AA Yu YK Protein database searches using compositionally adjusted substitution matrices FEBS J 2005 272 20 5101 5109 10.1111/j.1742-4658.2005.04945.x 16218944 26. Frohlich C Paarmann K Steffen J Stenzel J Krohn M Heinze HJ Pahnke J Genomic background-related activation of microglia and reduced beta-amyloidosis in a mouse model of Alzheimer's disease Eur J Microbiol Immunol (Bp) 2013 3 1 21 27 10.1556/EuJMI.3.2013.1.3 23814667 27. Schumacher T Krohn M Hofrichter J Lange C Stenzel J Steffen J Dunkelmann T Paarmann K Frohlich C Uecker A Plath AS Sommer A Bruning T Heinze HJ Pahnke J ABC transporters B1, C1 and G2 differentially regulate neuroregeneration in mice PLoS One 2012 7 4 e35613 10.1371/journal.pone.0035613 22545122 28. Frohlich C Zschiebsch K Groger V Paarmann K Steffen J Thurm C Schropp EM Bruning T Gellerich F Radloff M Schwabe R Lachmann I Krohn M Ibrahim S Pahnke J Activation of Mitochondrial Complex II-Dependent Respiration Is Beneficial for alpha-Synucleinopathies Mol Neurobiol 2016 53 7 4728 4744 10.1007/s12035-015-9399-4 26319560 29. Muller-Schiffmann A Herring A Abdel-Hafiz L Chepkova AN Schable S Wedel D Horn AH Sticht H de Souza Silva MA Gottmann K Sergeeva OA Huston JP Keyvani K Korth C Amyloid-beta dimers in the absence of plaque pathology impair learning and synaptic plasticity Brain 2016 139 Pt 2 509 525 10.1093/brain/awv355 26657517 30. Oshima N Morishima-Kawashima M Yamaguchi H Yoshimura M Sugihara S Khan K Games D Schenk D Ihara Y Accumulation of amyloid beta-protein in the low-density membrane domain accurately reflects the extent of beta-amyloid deposition in the brain Am J Pathol 2001 158 6 2209 2218 10.1016/S0002-9440(10)64693-7 11395399 31. Grundke-Iqbal I Iqbal K Quinlan M Tung YC Zaidi MS Wisniewski HM Microtubule-associated protein tau. A component of Alzheimer paired helical filaments J Biol Chem 1986 261 13 6084 6089 3084478 32. Whiteman IT Minamide LS de Goh L Bamburg JR Goldsbury C Rapid changes in phospho-MAP/tau epitopes during neuronal stress: cofilin-actin rods primarily recruit microtubule binding domain epitopes PLoS One 2011 6 6 e20878 10.1371/journal.pone.0020878 21738590 33. Aho L Pikkarainen M Hiltunen M Leinonen V Alafuzoff I Immunohistochemical visualization of amyloid-beta protein precursor and amyloid-beta in extra- and intracellular compartments in the human brain J Alzheimers Dis 2010 20 4 1015 1028 20413866 34. Edrey YH Medina DX Gaczynska M Osmulski PA Oddo S Caccamo A Buffenstein R Amyloid beta and the longest-lived rodent: the naked mole-rat as a model for natural protection from Alzheimer's disease Neurobiol Aging 2013 34 10 2352 2360 10.1016/j.neurobiolaging.2013.03.032 23618870 35. Dong X Chen W Mousseau N Derreumaux P Energy landscapes of the monomer and dimer of the Alzheimer's peptide Abeta(1-28) J Chem Phys 2008 128 12 125108 10.1063/1.2890033 18376983 36. Dai X Sun Y Gao Z Jiang Z Copper enhances amyloid-beta peptide neurotoxicity and non beta-aggregation: a series of experiments conducted upon copper-bound and copper-free amyloid-beta peptide J Mol Neurosci 2010 41 1 66 73 10.1007/s12031-009-9282-8 19685013 37. Poduslo JF Howell KG Olson NC Ramirez-Alvarado M Kandimalla KK Alzheimer's disease amyloid beta-protein mutations and deletions that define neuronal binding/internalization as early stage nonfibrillar/fibrillar aggregates and late stage fibrils Biochemistry 2012 51 19 3993 4003 10.1021/bi300275g 22545812 38. Liu ST Howlett G Barrow CJ Histidine-13 is a crucial residue in the zinc ion-induced aggregation of the A beta peptide of Alzheimer's disease Biochemistry 1999 38 29 9373 9378 10.1021/bi990205o 10413512 39. Huang J Yao Y Lin J Ye YH Sun WY Tang Dagger WX The solution structure of rat Abeta-(1-28) and its interaction with zinc ion: insights into the scarcity of amyloid deposition in aged rat brain J Biol Inorg Chem 2004 9 5 627 635 10.1007/s00775-004-0556-x 15160315 40. Miura T Suzuki K Kohata N Takeuchi H Metal binding modes of Alzheimer's amyloid beta-peptide in insoluble aggregates and soluble complexes Biochemistry 2000 39 23 7024 7031 10.1021/bi0002479 10841784 41. Smith DP Smith DG Curtain CC Boas JF Pilbrow JR Ciccotosto GD Lau TL Tew DJ Perez K Wade JD Bush AI Drew SC Separovic F Masters CL Cappai R Barnham KJ Copper-mediated amyloid-beta toxicity is associated with an intermolecular histidine bridge J Biol Chem 2006 281 22 15145 15154 10.1074/jbc.M600417200 16595673 42. Tickler AK Smith DG Ciccotosto GD Tew DJ Curtain CC Carrington D Masters CL Bush AI Cherny RA Cappai R Wade JD Barnham KJ Methylation of the imidazole side chains of the Alzheimer disease amyloid-beta peptide results in abolition of superoxide dismutase-like structures and inhibition of neurotoxicity J Biol Chem 2005 280 14 13355 13363 10.1074/jbc.M414178200 15668252 43. Edrey YH Oddo S Cornelius C Caccamo A Calabrese V Buffenstein R Oxidative damage and amyloid-beta metabolism in brain regions of the longest-lived rodents J Neurosci Res 2014 92 2 195 205 10.1002/jnr.23320 24273049 44. Orr ME Garbarino VR Salinas A Buffenstein R Sustained high levels of neuroprotective, high molecular weight, phosphorylated tau in the longest-lived rodent Neurobiol Aging 2015 36 3 1496 1504 10.1016/j.neurobiolaging.2014.12.004 25576082 45. Bates K Vink R Martins R Harvey A Aging, cortical injury and Alzheimer's disease-like pathology in the guinea pig brain Neurobiol Aging 2014 35 6 1345 1351 10.1016/j.neurobiolaging.2013.11.020 24360504 46. Beck M Bruckner MK Holzer M Kaap S Pannicke T Arendt T Bigl V Guinea-pig primary cell cultures provide a model to study expression and amyloidogenic processing of endogenous amyloid precursor protein Neuroscience 2000 95 1 243 254 10.1016/S0306-4522(99)00390-5 10619481 47. Beck M Bigl V Rossner S Guinea pigs as a nontransgenic model for APP processing in vitro and in vivo Neurochem Res 2003 28 3-4 637 644 10.1023/A:1022850113083 12675155 48. Palacios-Munoz A Escobar MJ Vielma A Araya J Astudillo A Valdivia G Garcia IE Hurtado J Schmachtenberg O Martinez AD Palacios AG Role of connexin channels in the retinal light response of a diurnal rodent Front Cell Neurosci 2014 8 249 25202238 49. Kumazawa-Manita N Katayama M Hashikawa T Iriki A Three-dimensional reconstruction of brain structures of the rodent Octodon degus: a brain atlas constructed by combining histological and magnetic resonance images Exp Brain Res 2013 231 1 65 74 10.1007/s00221-013-3667-1 23995563 50. Deacon RM Altimiras FJ Bazan-Leon EA Pyarasani RD Nachtigall FM Santos LS Tsolaki AG Pednekar L Kishore U Biekofsky RR Vasquez RA Cogram P Natural AD-Like Neuropathology in Octodon degus: Impaired Burrowing and Neuroinflammation Curr Alzheimer Res 2015 12 4 314 322 10.2174/1567205012666150324181652 25817252 51. Li Q Zhao HF Zhang ZF Liu ZG Pei XR Wang JB Cai MY Li Y Long-term administration of green tea catechins prevents age-related spatial learning and memory decline in C57BL/6 J mice by regulating hippocampal cyclic amp-response element binding protein signaling cascade Neuroscience 2009 159 4 1208 1215 10.1016/j.neuroscience.2009.02.008 19409206 52. Palmeri A Privitera L Giunta S Loreto C Puzzo D Inhibition of phosphodiesterase-5 rescues age-related impairment of synaptic plasticity and memory Behav Brain Res 2013 240 11 20 10.1016/j.bbr.2012.10.060 23174209 53. Puzzo D Bizzoca A Loreto C Guida CA Gulisano W Frasca G Bellomo M Castorina S Gennarini G Palmeri A Role of F3/contactin expression profile in synaptic plasticity and memory in aged mice Neurobiol Aging 2015 36 4 1702 1715 10.1016/j.neurobiolaging.2015.01.004 25659859 54. He P Cheng X Staufenbiel M Li R Shen Y Long-term treatment of thalidomide ameliorates amyloid-like pathology through inhibition of beta-secretase in a mouse model of Alzheimer's disease PLoS One 2013 8 2 e55091 10.1371/journal.pone.0055091 23405115 55. Pype S Moechars D Dillen L Mercken M Characterization of amyloid beta peptides from brain extracts of transgenic mice overexpressing the London mutant of human amyloid precursor protein J Neurochem 2003 84 3 602 609 10.1046/j.1471-4159.2003.01556.x 12558980 56. Toda T Noda Y Ito G Maeda M Shimizu T Presenilin-2 mutation causes early amyloid accumulation and memory impairment in a transgenic mouse model of Alzheimer's disease J Biomed Biotechnol 2011 2011 617974 10.1155/2011/617974 21234330 57. Parkinson J Ploeger B Appelkvist P Bogstedt A Dillner Bergstedt K Eketjall S Visser SA Modeling of age-dependent amyloid accumulation and gamma-secretase inhibition of soluble and insoluble Abeta in a transgenic mouse model of amyloid deposition Pharmacol Res Perspect 2013 1 2 e00012 10.1002/prp2.12 25505567 58. Krauthausen M Kummer MP Zimmermann J Reyes-Irisarri E Terwel D Bulic B Heneka MT Muller M CXCR3 promotes plaque formation and behavioral deficits in an Alzheimer's disease model J Clin Invest 2015 125 1 365 378 10.1172/JCI66771 25500888 59. Beach TG Kuo YM Schwab C Walker DG Roher AE Reduction of cortical amyloid beta levels in guinea pig brain after systemic administration of physostigmine Neurosci Lett 2001 310 1 21 24 10.1016/S0304-3940(01)02076-6 11524148

PMC005xxxxxx/PMC5002179.txt

==== Front BMC MicrobiolBMC MicrobiolBMC Microbiology1471-2180BioMed Central London 81310.1186/s12866-016-0813-5Research ArticlePyrosequencing analysis revealed complex endogenetic microorganism community from natural DongChong XiaCao and its microhabitat Xia Fei xiafei1177@163.com 12Liu Yan liuyan8468@sjtu.edu.cn 1Guo Meng-Yuan 971276109@qq.com 1Shen Guang-Rong sgrong@sjtu.edu.cn 1Lin Juan +86-21-51630536linjuan@fudan.edu.cn 2Zhou Xuan-Wei +86-21-34205778xuanweizhou@sjtu.edu.cn 11 Key Laboratory of Urban Agriculture (South) Ministry of Agriculture, and Engineering Research Center of Cell & Therapeutic Antibody (Ministry of Education), and School of Agriculture and Biology, Shanghai Jiao Tong University, No. 1-411# Agriculture and Biology Building, 800 Dongchuan Road, Shanghai, 200240 People’s Republic of China 2 Department of Ecology and Evolutionary Biology, School of Life Sciences, Fudan University, E-401-8#, Life Science Building, 2005 Songhu Road, Shanghai, 200438 People’s Republic of China 26 8 2016 26 8 2016 2016 16 1 1963 5 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 Ophiocordyceps sinensis (DongChong XiaCao (DCXC) in Chinese), a fungal parasite of caterpillars, is a traditional Chinese medicine. Bioactive components isolated from natural DCXC possess a wide range of pharmacological actions. Many efforts have been directed towards isolating the fungi based on culture-dependent methods for investigation of fungal diversity in order to determine the anamorph of natural DCXC and find new medicinal fungi resources, and the results have been varied. Results In the present study, a total of 44,588 bacterial and 51,584 fungal sequences corresponding to 11,694 and 9297 putative operational taxonomic units (OTU) were respectively identified by a Roche/454-based, high throughput sequence analysis of 16S rRNA genes and ITS regions. The main bacterial groups were Proteobacteria, Acidobacteria, Bacteroidetes, Actinobacteria and Firmicutes, while the Ascomycota, Basidiomycota and Zygomycota were the main fungal phyla. Proteobacteria presented 68.4, 49.5, 38.9 and 35.6 % of all bacteria in the sclerotia, stromata, external mycelial cortices and soil, respectively. As the main fungi phyla, Ascomycota presented 21.0, 45.6 26.4 and 59.3 % in the sclerotia, stromata, external mycelial cortices and soil, respectively. Bacterial and fungal communities were more diverse in the environmental sample than in the natural DCXC sample. Microbial communities were obviously distinct in each sample. Several novel unclassifiable bacterial (10.41 %) and fungal (37.92 %) species were also detected. Conclusions This study revealed an abundant endogenetic fungal and bacterial resources and a variety of genetic information in natural DCXC by high-throughput 454 sequencing technology. Microorganism that had been discovered in natural DCXC will provide sources for screening the new bioactive metabolites and its biotechnological application. Electronic supplementary material The online version of this article (doi:10.1186/s12866-016-0813-5) contains supplementary material, which is available to authorized users. Keywords Ophiocordyceps sinensisPyrosequencingMicroorganism communityCommunity diversityMicrohabitatsTibet Shenglong Industry Co., Ltd2013310031001210issue-copyright-statement© The Author(s) 2016 ==== Body Background Natural Ophiocordyceps sinensis (syn. Cordyceps sinensis), known as DongChong XiaCao (DCXC, short for “natural DCXC” in following text) in Chinese and as the Chinese caterpillar fungus in western countries, is traditionally considered a parasitic complex of a fungus (Hirsutella sinensis) and caterpillars that belong to Thitarodes, Hepialidae and Lepidoptera [1]. It is endemic to alpine habitats on the Tibetan Plateau, located predominantly in Tibet and Tibetan autonomous prefectures of neighboring provinces and the high Himalayas [2]. Being described as an organism complex rather than a simple O. sinensis organism, the terms caterpillar fungus and natural DCXC are an entomophagous flask fungus in the new family Ophiocrdycipitaceae (Prenemycetes, Ascomycota) [3]. Isolated O. sinensis strain and DCXC have been used for academic research and processing products, however their names in concrete practices were often confused. Experimental methods in biochemistry have proved that DCXC consists of active constituents such as cordycepin, cordycepic acid, polysaccharides, nucleosides, ergosterol, peptides, aminophenol and trace elements [4]. The modern medical researches prove that natural DCXC and its active components possess a wide range of pharmacological actions, such as anti-inflammatory [5], antioxidant, antitumor, antihyperglycemic, antiapoptosis, immunomodulatory, nephroprotective and hepatoprotective [6]. Its immunoregulatory function plays an important role in anti-tumor effects, organ transplantation and the prevention of kidney, liver and heart disease [7]. Therefore, these myco-medicinal products, coupled with their associated mycelial cultures, are popular items in the traditional medicine market [8]. High commercial value and habitat degradation of natural DCXC has led to overexploitation, which has endangered the species in recent years. The number of natural DCXC populations is extremely limited, and its residential ecological environments are also potentially threatened [9]. To protect DCXC resources and its habitats, an effective approach is to study and develop substitutes represented by endogenetic microorganisms isolated from natural DCXC, based on consideration of its specific bioactive components. On the other hand, artificial cultivation of DCXC or artificial assistance of its growth may provide an alternative to protect this valuable resource [1]. In the past 30 years, scientists have been devoted to determine the anamorph of DCXC, and the investigation and isolation of the endogenetic fungi from natural DCXC [10–15]. Meanwhile, genetic diversity of natural DCXC has been estimated from a limited number of individuals or populations using various molecular methods [16–19]. Currently, the microorganism community composition, structure and functional activity of natural DCXC remain unclear. With conserved and hypervariable sequence characteristics, the 16S rRNA genes and the internal transcribed spacer (ITS) regions were commonly used as the ideal marker sequence in the analysis of microorganism community diversity [20]. The 16S rRNA gene and ITS regions sequences permit the identification and prediction of phylogenetic relationships of prokaryotes and eukaryota [21, 22]. High-throughput sequencing, in particular pyrosequencing [23], has been applied widely to studies of microorganism community composition, structure and its ecology [24, 25]. Using this method, some trace microbes have been detected, and there have been new discoveries in the field of microbial population diversity in traditional Chinese medicinal materials (TCMM) [26], agriculture [27], marine ecology [28] and soil ecology [29]. According to traditional Chinese medicine theory, the geoherbalism formation of TCMM is based on many factors, among which the microenvironment of rhizosphere soil plays an important role in the formation of secondary metabolism in medicinal plants. Therefore, based on the diversity of microbial communities by the clone library approach [30], the natural DCXC couple with its microhabitat soil was divided into four parts: stroma, larva, external mycelial cortices and soil, to in each sample using high throughput pyrosequencing platform (Roche/454 GS-FLX Titanium System) and quantitative real-time PCR (qPCR) techniques. The aim of the study was to detect the microbial diversity and community structure of natural DCXC and its microhabitats. In subsequent, we analyzed the relationship between the microorganism community composition and structure in various specimens with regard to the medicinal components, fungal parasitisation of the larva, and growth and development. It lays a foundation for exploring the genetic diversity and functioning of DCXC microbiota as a model for further biotechnological developments. Results Community diversity High quality sequences of 44,658 16S rRNA genes and 51,584 ITS reads were obtained with at least 9042 bacterial 16S rRNA gene and 8585 fungal ITS reads in each sample. The results indicated that a large number of bacterial and fungal microorganisms inhabiting the natural DCXC (sclerotia and stromata) (Fig. 1g, h) and its microhabitat (external mycelial cortices and soil adhering to the surface of the membrane covered DCXC) (Fig. 1i, f). Although we got nearly ten thousandsequences per sample for bacteria and fungus community, the slope of the rarefaction curve was still not flat at different similarity cutoff values, indicating that there were still microorganisms that was not detected (Additional file 1: Figure S1). The diversity and richness of bacterial and the fungal community of different natural DCXC samples and its microhabitat were represented by the indices (Table 1). The OTU (Operational taxonomic units) numbers and Shannon’s index of the soil sample were higher than the other samples, both in the bacterial and fungal communities (Table 1), which indicated a higher diversity of the bacterial and fungal community in the soil adhering to natural DCXC. This was also demonstrated by the curve of OTU Rank abundance (Additional file 2: Figure S2) and Simpson’s index (Table 1). Interestingly, the OTU number and the Shannon’s index showed that the bacterial community diversity was highest in the soil, followed by external mycelial cortices and stromata, and fewest in sclerotia at each similarity level. For the fungal community, the OTU number revealed the same trend as the bacteria; however, this trend in Shannon’s index was disturbed for external mycelial cortices and stromata.Fig. 1 Collection and isolation of DCXC and its microhabitat samples. a The sample was collected from Ya-an ethnic township, Baqing country of Nagqu Prefecture; b The samples for this study were collected in three different populations at least 50–100 m apart at 4520 m above sea level; the stars indicated the governments of Biru, Nierong, Sog and Baqing county; the triangles indicated the sampling locations in the Ya-an ethnic township, Baqing County; c habitat of natural DCXC; d-i after collection of growing DCXC (d and e). The collected samples were divided into DCXC and its growth microhabitats. Microhabitat samples included soil adhering to the surface of the membrane (f), and external mycelial cortices (membrane covered around the larva) (i). The DCXC was divided into sclerotia (g) and stroma (h) Table 1 Summary of the diversity indices of different samples from DongChong XiaCao (DCXC; Ophiocordyceps sinensis) and its environmental samples Bacteria Fungi Similarity Sample names OTU Numbers Chao1 a Simpson Shannon OTU Numbers Chao1 Simpson Shannon 97 % sclerotia b 1982 3050.38 0.0069 6.23 1709 3638.45 0.0162 5.81 stromata 2914 5409.71 0.0069 6.95 2109 4894.79 0.0120 5.68 external mycelial cortices 4270 9127.36 0.0010 7.90 2450 6611.38 0.0200 5.60 Soil 5809 13855.43 0.0003 8.31 3204 9554.88 0.0071 6.84 95 % sclerotia 1447 2085.29 0.0097 5.75 1372 2855.18 0.0182 5.53 stromata 2254 3711.34 0.0081 6.45 1639 3681.45 0.0154 5.35 external mycelial cortices 3505 6743.64 0.0016 7.57 1998 4892.88 0.0210 5.40 Soil 4912 10329.53 0.0005 8.05 2722 7445.69 0.0084 6.55 90 % sclerotia 897 1188.80 0.0182 5.09 956 1926.36 0.0223 5.07 stromata 1341 1984.05 0.0139 5.72 1039 1934.23 0.0181 4.92 external mycelial cortices 2228 3577.16 0.0027 6.88 1391 2861.02 0.0253 5.03 Soil 3216 5724.56 0.0012 7.38 2081 4787.90 0.0100 6.17 a To calculate all the indices, the sequences number in each samples were normalized to the same, with 9042 bacterial 16S rRNA gene sequences and 8585 fungal ITS sequences b Abbreviation of the samples name: stroma (ZiZuo in Pinyin simplified as stromata), membrane covered around larva (JunPi in Pinyin simplified as external mycelial cortices), larva (ChongTi in Pinyin simplified as sclerotia) and the soil adhering to the surface of the membrane covered DCXC. The detailed image is shown in Fig. 1 Similarly, richness was highest in the soil sample, followed by external mycelial cortices and stromata. The least rich bacterial and fungal communities were in the sclerotia sample, which was in accordance with the Chao1 index (Table 1). The richness of microorganism community indicated that the soil and the external mycelial cortices in contact with the soil promoted microorganism growth. Community composition According to the taxonomy results of 16S rRNA gene and ITS sequences obtained by a localblast search against the Silva and UNITE databases, the microbial compositions in the different samples were analyzed (Fig. 2). The bacterial community composition was varied in different samples of natural DCXC and the microhabitat samples. Proteobacteria was the main group in different samples and presented 68.4, 49.5, 38.9 and 35.6 % of the total bacteria detected in sclerotia, stromata, external mycelial cortices and soil, respectively (Fig. 2a). The proportion of Acidobacteria in the sclerotia was rather small, representing only 2.5 %. In contrast, Acidobacteria in the other three samples were higher: 8.4 % in stromata, 19.0 % in external mycelial cortices and 17.9 % in the soil. Bacteroidetes was another notably abundant bacterial group, with the lowest proportion, 10.2 %, in sclerotia and the highest, 26.7 %, in stromata. In addition, Actinobacteria, Planctomycetes and Verrucomicrobia bacteria, as well as Firmicutes, were also detected in the samples, although at small fractions. Finally, we also detected notable numbers of unclassified bacteria, especially in the external mycelial cortices and soil samples, representing 15.7 and 16.5 % of the total sequences in these samples, respectively, indicating a higher diversity of unknown microorganisms in these two samples.Fig. 2 Bacterial and fungal community patterns of different samples. a bacterial community at the Phylum level; b fungal community at the Phylum level. The percentages on the vertical axis of the graph represent the proportion of each microbe group in the total obtained sequences for each sample For the fungal community, we detected three main phyla in the samples from DCXC: Ascomycota, Basidiomycota and Zygomycota (Fig. 2b). Ascomycota was the most dominant group in the samples, representing 21.0 % in sclerotia, 45.6 % in stromata, 26.4 % in external mycelial cortices and 59.3 % in the soil. Basidiomycota, as the second most abundant fungal community group, was present at its lowest level, 11.0 %, in the sclerotia and was highest in the soil at 20.1 %. While in the stromata and external mycelial cortices, the portions of Basidiomycota fungi were 19.8 and 19.6 %, respectively. Zygomycota was present in low numbers in the sclerotia, external mycelial cortices and soil samples, but was rather higher, accounting for 13.9 % of the total detected fungi, in the stromata. The limitation of the resolution of the ITS sequence meant that there was a notable fraction of the ITS sequences in each sample that could not be classified exactly, especially in the sclerotia and external mycelial cortices, with 65.9 and 51.3 % of undefined sequences in these two samples, respectively. These results suggested an undiscovered abundant fungal community resource in the samples from natural DCXC. The compositions of bacterial and fungal community were further analyzed at the genus level. The huge undiscovered microorganism resource in the natural DCXC and the resolution limit of the rRNA gene and the ITS region sequence, resulted in a large proportion of unclassified sequences at the genus level. For the bacterial community, undefined sequences accounted for 29.06 in the sclerotia, 54.63 in the stromata, 78.54 in the external mycelial cortices and 80.73 in the soil. These unclassified sequences suggested a huge bio-resource that may be related to the producing of medical materials in natural DCXC. Meanwhile, the most abundant genera detected in the sclerotia included Pseudomonas and Rhodoferax, with 9.24 and 9.11 % of the total sequences, respectively. While other genera, such as Pedobacter (3.42 %), Sphingomonas (2.29 %), Collimonas (2.78 %) were also detected. The main genera detected in the stromata were Pedobacter (8.05 %), Rhodoferax (6.91 %), Variovorax (4.41 %) and Mucilaginibacter (3.96 %). For the external mycelial cortices sample, the Thermomonas (3.43 %) and Gemmatimonas (2.13 %) were the main genera detected. In addition, Ferruginibacter (3.07 %), Gemmatimonas (2.22 %) and Terrimonas (1.93 %) were the main genera in the soil sample. Detailed information on the bacterial genera and their proportions detected in each sample is shown in Additional file 3: Table S1. The undefined fungal ITS sequences also accounted for a significant percentage of the total sequences in the sclerotia (73.64 %), stromata (85.52 %), external mycelial cortices (87.55 %) and soil (67.32 %). The Entoloma were the main genus detected in the sclerotia (2.62 %) and external mycelial cortices (3.47 %) samples. It was also distributed in the soil sample, but the percentage was only 0.37 %. Exophiala (2.90 %), Cladophialophora (1.77 %) and Phaeomollisia (1.45 %), were the main genera detected in the sample stromata. Tetracladium (1.38 %), were the main genus in the external mycelial cortices sample. Finally, for the soil sample, Cladophialophora (3.30 %), Verrucaria (3.10 %), Sebacina (2.96 %) and Gyoerffyella (2.33 %) were the main genera detected. For more information of the fungal community genera and their proportions detected in each sample are shown in Additional file 4: Table S2. Community similarities in different samples The microorganism community similarity in different samples of natural DCXC was determined by weighted principal coordinates analysis (PCoA). The bacterial community in the soil sample appeared more similar to the bacterial community in the external mycelial cortices. These two samples clustered separately from the other two samples, sclerotia and stromata (Fig. 3a). This was also demonstrated by the unweighted pair group method with arithmetic averages (UPGMA) trees constructed from thetaYC distances, in which the bacterial communities of soil and external mycelial cortices were in the same group, while the samples stromata and sclerotia were in another group (Additional file 5: Figure S3-A). In addition, a Venn diagram showed the shared OTUs with 97 % similarity of the bacterial community in each samples from DCXC (Additional file 6: Figure S4-A). Soil and external mycelial cortices samples shared 1054 OTUs; meanwhile, that soil and stromata samples shared 877 OTUs, and there were only 406 OTUs shared by soil and sclerotia. This suggested that the bacterial communities in soil and external mycelial cortices were more similar, which was constant to the results of PCoA and the UPGMA clustering analysis. The defined bacteria at the genus level in the soil sample mainly were Ferruginibacter and Gemmatimonas, while the Thermomonas and Pseudomonas were the main genera in the sample external mycelial cortices and stromata, which was constant with the composition of the bacterial community at the genus level (Additional file 3: Table S1).Fig. 3 Principal coordinates analysis of bacterial and fungal communities in different samples of natural DCXC. Percentages on the axes of the graph represent the explained variance of total variance. The OTU data matrix used in the analyses was clustered at the 97 % similarity and the principal coordinates analysis was based on the jclass model. OTUs shown in the profiles were relatively more abundant ones. The circle in color indicated the four samples distributed in the ordination. The circle in gray indicated the OTUs that areas symbolize the abundant, and the different circles indicated the number of sequences contained in each OTU Fungal communities in the sclerotia and external mycelial cortices samples were more similar, with Entoloma being abundant in these two samples. Meanwhile, fungal communities in samples of soil and stromata were only distantly related to those in the sclerotia and external mycelial cortices (Fig. 3b). UPGMA trees constructed from a ThetaYC distances matrix showed that the fungal community structure in external mycelial cortices and sclerotia was more similar to each other than they were to those of the sample stromata and soil (Additional file 5: Figure S3-B). In addition, a Venn diagram (Additional file 6: Figure S4-B) showed that the shared OTUs with 97 % similarity of the fungal community in each samples from DCXC and its microhabitats, which indicated that external mycelial cortices and sclerotia shared 406 OTUs. Meanwhile, 344 OTUs were shared by external mycelial cortices and stromata, and 250 OTUs were shared by external mycelial cortices and soil. This demonstrated that a similar fungal community structure in the external mycelial cortices and sclerotia samples. Soil and stromata samples contained rather different fungal communities compared to the external mycelial cortices and sclerotia samples, suggesting a high diversity of fungi in the different samples of DCXC and its microhabitats. For the defined fungal genera in the samples, the levels of Entoloma provided a partial explanation for the distribution of sclerotia and external mycelial cortices, while other genera did not show a tendency to be present in a particular sample internal microenvironment, indicating that these fungi were disparate in the different part of DCXC. Abundance of total bacteria and fungi The abundance of total bacterial and fungal communities was detected qPCR (quantitative real-time PCR) based on the SYBR Green method. The qPCR results indicated that the soil sample contained the most abundant bacterial and fungal community compared with the other samples from natural DCXC. The abundance of bacterial community were higher than the fungal community in natural DCXC samples (Fig. 4). Bacterial community abundance was 1.18 ± 0.01 × 106 copies per gram dry materials in the soil sample, which was significant higher than in in the sclerotia, stromata and external mycelial cortices samples (p < 0.05). With the fungal abundance was 3.80 ± 0.15 × 107 copies per gram dry material, abundance in the soil sample was significant higher than in in the other samples (p < 0.05). The results of a T-test indicated that the abundances of total bacterial and fungal communities were significantly higher in the soil sample than in the other samples (p < 0.05), and the abundance of the total fungal community was significant higher than the total bacterial community abundance in each sample (p < 0.05).Fig. 4 Abundance of the total bacterial and fungal community in different samples isolated from natural DCXC. The error bars indicate SDs (n = 3), with some error bars smaller than the symbol. Samples containing a different letter indicate that there were significant differences (p < 0.05). The unit of abundance was the log value of the gene copies per gram of dry materials. The moisture content of each sample was measured by gravimetric analysis after oven-drying (105 °C) of triplicate samples Discussion Natural DCXC is an expensive, endangered, traditional medicinal drug, which comprises many fungi and a larva. Since the 1980s, the separation and identification of the anamorph of DCXC have been the main focus in this field. More than 30 fungal strains belonging to 13 genera, which are involved in anamorphic types of the DCXC, have been reported in China. In recent years, the relationship between DCXC and its microhabitats has received more attention. Zhang et al. [31] investigated the mycobiota of natural DCXC using a traditional culture-dependent method. In total, 572 fungal cultures were isolated from different parts (including stromata, sclerotia, and external mycelial cortices) of natural DCXC. For the growth microhabitat of DCXC, Li et al. [32] analyzed soil fungal community structures in propagated DCXC by DGGE, and found that the ITS2 sequences of fungi in soil samples were highly similar to those of Inocybe (Fr.) Fr., Tricholoma (Fr.) Staude, Entorrhiza C.A. Weber, unconfirmed Ascomycota, and soil fungi in GenBank; the similarity of the soil fungal population structure was only 19.4 % ~ 50.1 %. Based on the above-mentioned analysis, at the technical level, the traditional culture-dependent method was used for primary separation of fungi from the DCXC. Modern bio-techniques were used for the identification of the DCXC anamorph in the isolates, rather than to investigate the microorganism community [33–35]. Many years of research have proved that H. sinensis is the anamorph of natural DCXC collected from different geographical regions [10, 36]. However, H. sinensis and O. sinensis belong to different stages of the life cycle of the same organism [12, 37]. O. sinensis pathogenicity has clearly been shown, but only by a few researchers. The aims of the previous studies were to determine the anamorph of DCXC that could produce similar metabolites to DCXC. Currently, H. sinensis mycelia have been applied successfully for in large-scale fermentation to produce various drugs and health products [1], which has led to limited protection of natural DCXC resources. Although O. sinensis is defined to be only one species, natural DCXC and its microhabitats contain a variety of microorganism [10], and form a complex host-microbiota-environment microecosystem. In this study, the pyrosequencing analysis reveals an unexpectedly high diversity of the microorganism community in the complex microecosystem. Analysis of composition and structure of microorganism communities benefits to unsolved problems in DCXC study, for example, how fungus infected the larvae, what was the relationship (symbiosis, parasitism or pythogenesis and etc.) of endophytic fungus and its host organism interaction. However, in this microecosystem, the relationships among microorganisms inhabiting in DCXC need to be further explored. The influence of these microorganisms to the reproductive and individual development of DCXC will also need further investigation. In addition, numerous of unidentified sequences were detected in the natural DCXC samples and its micro-environments. There are two possible reasons for explanation of this result. Firstly, the research targeting the micro-ecology in Tibetan Plateau was few all over the world. Secondly, scientists for the study of DCXC mainly came from China, and focused their interesting on isolation and identification of the fungus of the asexual generations (anamorph), and fermented these fungi by using artificial methods, etc. So there was a little piece of information on endogenetic microorganism from natural DCX in the database. The bacterial population of DCXC was only one report available [30], and the effects of bacteria on the growth process and medicinal composition of natural DCXC are unclear. Based on the results analysis of this study, the main groups of bacteria in the different samples were Proteobacteria, Acidobacteria, Bacteroidetes, Actinobacteria and Firmicutes. Among the bacterial communities, the Proteobacteria is a major group (phylum) of bacteria, which include a wide variety of pathogens. The results suggest that bacteria may be relative to pathopoiesis of the larva. Acidobacteria is widely distributed in nature and plays important roles in various ecosystems [38]. Seven genera of Acidobacteria were isolated and reported, and confirmed by Bergey’s manual of systematic bacteriology, including Acidobacterium, Geothrix, Holophaga and the recently published Edaphobacter aggregans, Edaphobacter modestus, Chloracidobacterium thermophilum and Terriglobus roseus [39]. After the confirmation of the symbiotic microbes in the larva guts, genomic analysis will be possible to reveal the molecular foundations of the relationships between the insect and its microbiome [40]. Furtherly, the roles of the bacterial communities in growth and development, and in the formation of bioactive ingredients of DCXC need further exploration. It is need further for study that endogenetic bacteria play roles in larva pathopoiesis, secondary metabolite formation and fruit body development. Although, pooling the samples from different sampling sites together only was a preliminary approach in study design in the early high-throughput sequencing, which might resulted in the extreme limitations of data, the results of current study also have certain values for the further study. Study design will be improved in the following research of comparing the endogenetic microorganism community in natural DCXC from different areas. Pharmaceutical substances from natural DCXC are a complex product, just as the DCXC itself is a natural complex of larva and fungal, as well as bacterial communities. Hence, we have good reason to believe that the complicated communities of endogenetic microorganisms in DCXC are related to the secretion of the pharmaceutical substances, and are involved in the growth, development and metabolic process of DCXC. In other words, bioactive ingredients and the infection process of fungus to larva should be related to the endogenetic microorganism community structure of DCXC. In studies of the fungal community structure [15, 41], it was demonstrated that the culture-dependent method only identified limited amounts of microbes [42]. The limitation of the approach resulted few kinds of fungi being found in DCXC. Many studies have indicated a diverse microorganism community in natural DCXC using culture-dependent and independent methods [14, 31]. Obviously, the culture - dependent methods are not sufficient to discover the endogenetic complete community resource, especially the bacterial community. Surprisingly, the DCXC microbiota has so far been poorly explored and exploited for screening bioactive component proposes. In this study, the large proportion of unidentified sequences was found, and the function of these sequences was still unclear. However, illustration of the endogenetic microorganism community structure in DCXC using metagenome data would be beneficial to clarify the relationship between the DCXC and its endogenetic microorganisms, metagenome analysis provides new avenues for the study of the DCXC microorganism community and will reveal the molecular foundations of the relationships between the larva and its microbiome. The unidentified sequences, as a new potential microbial resource, will lay a foundation in the discovery of new drugs and drug leads, and the potential biotechnological application of DCXC microbiota. Conclusions This study revealed an abundant endogenetic fungal and bacterial resources and a variety of genetic information in natural DCXC by high-throughput 454 sequencing technology. In the natural DCXC and its microhabitat, the main bacterial groups were Proteobacteria, Acidobacteria, Bacteroidetes, Actinobacteria and Firmicutes, while the Ascomycota, Basidiomycota and Zygomycota were the main fungal phyla. Proteobacteria presented 68.4, 49.5, 38.9and 35.6 % of all bacteria in the sclerotia, stromata, external mycelial cortices and soil, respectively. And as the main fungi phyla, Ascomycota presented 21.0, 45.6, 26.4and 59.3 % in the sclerotia, stromata, external mycelial cortices and soil, respectively. The bacterial and fungal communities were obviously distinct in each sample and the microorganism communities were more diverse in the environmental sample than in the natural DCXC sample. Microorganisms that had been discovered in natural DCXC will provide sources for screening the new bioactive metabolites and its biotechnological application. Methods Sample collection and preparation Natural DCXC in the Nagqu district of Tibet is mainly distributed in six counties: Baqing, Sog, Biru, Jiali, Nagqu and Nie Rong. In particular, with high quality and yield, natural DCXC was distributed in Baqing, Sog, and Biru. The average density of natural DCXC in these counties is 0.42 fruiting body per square meter, with the highest density being four fruiting bodie sper square meter. In current study, about 30 natural DCXC in total were collected with digging them up from the meadow soil of three sampling areas in the Ya-an ethnic township, Baqing County of the South of Nagqu Prefecture of Tibet Autonomous Region (Fig. 1a). In each sampling area, 5–10 natural DCXC were collected from three locations which departed 50–100 m (Fig. 1b). The samples were kept in aseptic valve bags and transported in an ice box to laboratory of Plant Biotechnology R&D Center of Shanghai Jiao Tong University, Shanghai, China. To get rid of the microorganisms inhabiting the surface of natural DCXC, the samples were treated with different disinfectants, including 75 % (v/v) ethanol, 2.5 % sodium hypochlorite, 0.01 % mercuric chloride individually, and then washed with sterilized water three times [30]. Subsequently, all the natural DCXC samples collected from three sampling locations were homogenized and divided into stroma, sclerotia and its microhabitat samples, including external mycelial cortices (membrane covering the larvae) and the soil adhering to the surface of the membrane covering DCXC, to investigate the microorganism community (Fig. 1d-i). The average moisture content of the samples were determined by drying to constant weight. The fresh samples were frozen at −20 °C until DNA was extracted. The same sections of natural DCXC samples from 3 different sampling location were pooled together to extract the genomic DNA. Genome DNA extraction, PCR amplification and pyrosequencing Natural DCXC samples were ground with liquid nitrogen to increase the fungus DNA extraction efficiency. About 0.5 g ground samples were used to extract total genomic DNA with a PowerSoil™ soil DNA Isolation Kits (Mo Bio Laboratories, Solana Beach, CA, USA) according to the manufacturer’ s protocol. The V1-V3 region of the 16S rRNA gene was amplified using the primer set 8F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 536R (5′-GWATTACCGCGGCKGCTG-3′) [43] for bacterial community analysis. For fungal community analysis, the ITS sequence, including the partial 18S rRNA gene, was amplified with primer set ITS 1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS 4R (5′-TCCTCCGCTTATTGATATGC-3′) [44–46]. For 454-pyrosequencing, a barcode with eight nucleotides was contained in the forward primer, and the requisite adapters A and B were added for the forward and reverse primers, respectively. The PCR amplification was carried out in a total volume of 25 μL PCR mixture containing 12.5 μL Ex Taq DNA polymerase (Takara, Japan), 1 μL bovine serum albumin (25 mg•ml-1), 1 μM of each primer, 1 μL template and 8.5 μL of ultrapure water. The PCR protocol consisted of an initial denaturation step at 95 °C for 5 min; 30 cycles at 94 °C for 30 s and annealing at 55 °C (for the bacterial community study) or 56 °C (for the fungal community study) for 45 s; and extension at 72 °C for 1 min (5 min at 72 °C for the last cycle) [30]. Three replicates of the amplifications were pooled together to minimize the PCR bias. PCR products were detected by electrophoresis through 1.2 % agarose gels and appropriately sized fragments were purified with DNA Gel Extraction Kit (Axygen, Union City, CA, USA). Purified PCR amplicons were sequencing using the Roche-454 GS FLX system, according to the instructions of the Shanghai Personal Biotechnology Co. Ltd. Quantitative real-time PCR Quantitative real-time PCR (qPCR) was used to determine the abundance of bacterial and fungal communities. Universal primer set 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 536R [43] were employed to quantify members of total bacteria. Total fungal community abundance was quantified using the same ITS primers used for the pyrosequencing. Standard curves were constructed by serial dilution of linearized plasmids containing the corresponding gene fragment. Gene quantification was performed on a Mx3000P real-time PCR system (Stratagene, La Jolla, CA, USA) with SYBR® Premix ExTaq™ (Takara, Tokyo, Japan). The total reaction volume of 20 μL contained 10 μL SYBR® Premix ExTaq), 0.2 μL bovine serum albumin (25 mg · mL−1), 1 μM of each primer, 2 μL template DNA and 6.6 μL of ultrapure water. The thermal programs for assaying total bacterial and fungal community abundance were as described in the pyrosequencing PCR section. The qPCR results were analyzed using MxPro qPCR software version 3.0 (Stratagene, La Jolla, CA, USA). Standard curves had R2 >0.99 and the efficiencies for PCR reactions were 95 and 85 % for total bacteria and fungi community, respectively. The copy numbers of the genes were calculated to the unit of “copies per gram dry material”. The Analysis of Variance (ANOVA) and pairwised analysis was conducted with R [47] by commands of “aov” and “pairwise.t.test”, respectively. The p values indicated the difference of copy numbers of ITS and 16S rRNA genes were corrected with “holm” [48] method by default of “pairwise.t.test” command and the p values < 0.05 were considered significant. Analysis of pyrosequencing data The average length of the raw pyrosequencing data was 550 bp in ITS and 450 bp in 16S rRNA genes sequences. These raw data were processed with the Mothur software package, version 1.27.0 [49]. Briefly, after trimming off the barcode, primers and filtering out the sequences with a quality score of any base below 25 with the “trim.seqs” command. The sequences were also trimmed if the fragment length beyond of 450 bp to 600 bp and 400 bp to 550 bp for ITS sequence and 16S rRNA genes, respectively. After potential chimeric sequences were checked with the chimera.slayer command, 44,658 16S rRNA genes and 51,584 ITS high quality sequences were obtained. The pyrosequencing reads of 16S rRNA gene were aligned against SILVA [50] with the command of “align.seqs”. After the distance matrix was calculated with the “dist.seqs” command, the qualified sequences were clustered into OTUs at 97, 95 and 90 % similarity with the “cluster” command, respectively. The OTUs of ITS sequences were picked with USEARCH61 [51] without alignment. After the bacterial 16S rRNA gene and ITS sequences in each sample were normalized to the minimum of 9042 and 8585, respectively, diversity indices (Shannon-Weiner), richness estimator (Chao1) and the evenness index (ACE) were calculated using the “summary.seqs” command for each sample. Taxonomy information was assigned to 16S rRNA gene and ITS sequences using localblast program against SILVA [50] and UNITE [52] databases, respectively with the “classify.seqs” command. Finally, weighted-Principal coordinates analysis was performed using QIIME version 1.7 [53]. Additional files Additional file 1: Figure S1. Rarefaction curves of 16S rRNA genes and ITS sequences. (a), (c) and (e) are the rarefaction curves of the 16S rRNA genes at cut off values of 97, 95 and 95 % similarity. While (b), (d) and (f) are the rarefaction curves for ITS sequences at cut off values of 97, 95 and 95 % similarity. In each panel, the x-axis represents the number of sequences and the y-axis represents the number of operational taxonomic units (OTUs) determined at a particular cut off value. (TIF 396 kb) Additional file 2: Figure S2. Operational taxonomic units (OTUs) rank curves of 16S rRNA gene (a) and ITS sequences (b) at a 97 % cut off value. In each figure, the x-axis represents the OTUs listed as the abundance in descending order and the y-axis represents the relative abundance of each OTU. The OTU rank curve could explain the diversity and the evenness of the organism community. The width of the curve reveals the diversity of the organism community along the x-axis; a wider curve indicates more diversity. Evenness of the organism community is suggested by the pattern of the curve. A flatter curve suggests a more even organism community. (TIF 216 kb) Additional file 3: Table S1. The classification of the bacteria detected in each sample of natural DCXC at the genus level. CT, larva; ZZ, stroma; JP, the membrane of DCXC; soil, bacteria isolated from the surface of natural DCXC. The percentages following the genus names are the proportions of each microbe group in the total obtained sequences of each sample. (DOCX 44 kb) Additional file 4: Table S2. The classification of the fungi detected in each sample of natural DCXC at the genus level. CT, larva; ZZ, stroma; JP, the membrane of DCXC; soil, bacteria isolated from the surface of natural DCXC. The percentages following the genus names are the proportion of each microbe group in the total obtained sequences of each sample. (DOCX 31 kb) Additional file 5: Figure S3. Unweighted Pair Group Method with Arithmetic Averages (UPGMA) dendrogram constructed from ThetaYC distances of bacterial communities in each sample (A) and fungal community (B). The values were shown above the lines indicate differences between the organism community structure in each sample. (TIF 106 kb) Additional file 6: Figure S4. Venn diagram of the shared operational taxonomic units (OTUs) of the bacterial (a) and fungal (b) community in each sample from DCXC. The OTUs were generated with a 97 % similarity cut off value. Different colors show the different samples and the numbers in the across area are the number of OTUs shared by different samples. (TIF 434 kb) Abbreviations DCXCNatural Ophiocordyceps sinensis (syn. Cordyceps sinensis) ITSInternal transcribed spacer TCMMTraditional Chinese medicinal materials qPCRQuantitative real-time PCR OTUOperational taxonomic units PCoAPrincipal coordinates analysis UPGMAUnweighted pair group method with arithmetic averages PCRPolymerase chain reaction The author would like to thank Director Buyong Jiacuo and Xin Wang for his helping to collect the natural O. sinenesis samples. Metagenome sequencing was performed in Shanghai Personal Biotechnology Co., Ltd. Funding This work was financially supported by Tibet Shenglong Industry Co., Ltd (No: 2013310031001210), the Social Science and Technology Development Project of Dongguan, China (No. 2013108101055). Availability of data and materials The datasets generated during and analysed during the current study are available in the NCBI Short Read Archive repository, under accession No. SRP052214, [http://www.ncbi.nlm.nih.gov/sra/?term=SRP052214]. Authors’ contributions Dr. JL and X-WZ managed the project and were final approval of the version to be published; X-WZ designed the experiments and prepared the samples; FX and JL led the data analysis and drafted the work, FX, YL and XZ wrote the manuscript. G-RS and M-YG led the data analysis of geographic information and preparation of Fig. 1. All authors reviewed the manuscript. We agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. 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. Zhou XW Li LJ Tian EW Advances in research of the artificial cultivation of Ophiocordyceps sinensis in China Crit Rev Biotechnol 2014 34 233 43 10.3109/07388551.2013.791245 23875938 2. Chen J Lee S Cao Y Peng Y Winkler D Yang D Ethnomycological use of medicinal chinese caterpillar fungus, Ophiocordyceps sinensis (Berk.) GH Sung et al. (ascomycetes) in Northern Yunnan Province, SW China Int J Med Mushrooms 2010 12 427 34 10.1615/IntJMedMushr.v12.i4.100 3. Sung GH Hywel-Jones NL Sung JM Luangsa-ard JJ Shrestha B Spatafora JW Phylogenetic classification of Cordyceps and the clavicipitaceous fungi Stud Mycol 2007 57 5 59 10.3114/sim.2007.57.01 18490993 4. Zhou X Gong Z Su Y Lin J Tang K Cordyceps fungi: natural products, pharmacological functions and developmental products J Pharm Pharmacol 2009 61 279 91 10.1211/jpp.61.03.0002 19222900 5. Yang ML Kuo PC Hwang TL Wu TS Anti-inflammatory principles from Cordyceps sinensis J Nat Prod 2011 74 1996 2000 10.1021/np100902f 21848266 6. Yue K Ye M Zhou Z Sun W Lin X The genus Cordyceps: a chemical and pharmacological review J Pharm Pharmacol 2013 65 474 93 10.1111/j.2042-7158.2012.01601.x 23488776 7. Shashidhar M Giridhar P Sankar KU Manohar B Bioactive principles from Cordyceps sinensis : A potent food supplement-A review J Funct Foods 2013 5 1013 30 10.1016/j.jff.2013.04.018 8. Lo HC Hsieh C Lin FY Hsu TH A systematic review of the mysterious caterpillar fungus Ophiocordyceps sinensis in Dong-ChongXiaCao (冬蟲夏草 Dōng Chóng Xià Cǎo) and related bioactive ingredients J Tradit Complement Med 2013 3 16 23 10.1016/S2225-4110(16)30164-X 24716152 9. Quan QM Wang QX Zhou XL Li S Yang XL Zhu YG Cheng Z Comparative phylogenetic relationships and genetic structure of the caterpillar fungus Ophiocordyceps sinensis and its host insects inferred from multiple gene sequences J Microbiol 2014 52 99 105 10.1007/s12275-014-3391-y 24500473 10. Jiang Y Yao Y Anamorphic fungi related to Cordyceos sinensis Mycosystema 2003 22 161 76 11. Jiang Y Yao Y Current understanding of molecular systematics of Cordyceps J Fungal Res 2004 2 58 67 12. Liu X Guo Y Yu Y Zeng W Isolation and identification of the anamorphic state of Cordyceps sinensis (Berk.) Sacc Acta Mycol Sin 1989 8 35 40 13. Mo M Chi S Zhang K Microcycle conidiation of Cordycps sinensis and anamorph isolation Mycosystema 2000 20 482 5 14. Zhang Y Zhang S Wang M Bai F Liu X High diversity of the fungal community structure in naturally-occurring Ophiocordyceps sinensis PLoS One 2010 5 e15570 10.1371/journal.pone.0015570 21179540 15. Zhang Y Li E Wang C Li Y Liu X Ophiocordyceps sinensis , the flagship fungus of China: terminology, life strategy and ecology Mycology 2012 3 2 10 16. Zhang Y Xu L Zhang S Liu X An Z Wang M Guo Y Genetic diversity of Ophiocordyceps sinensis , a medicinal fungus endemic to the Tibetan Plateau: implications for its evolution and conservation BMC Evol Biol 2009 9 290 10.1186/1471-2148-9-290 20003548 17. Hao J Cheng Z Liang H Yang X Li S Zhou T Zhang W Chen J Genetic differentiation and distributing pattern of Cordyceps sinensis in China revealed by rDNA ITS sequences Chin Tradit Herb Drugs 2009 40 112 6 18. Liang HH Cheng Z Yang XL Li S Ding ZQ Zhou TS Zhang WJ Chen JK Genetic diversity and structure of Cordyceps sinensis populations from extensive geographical regions in China as revealed by inter-simple sequence repeat markers J Microbiol 2008 46 549 56 10.1007/s12275-008-0107-1 18974957 19. Chen Y Zhang YP Yang Y Yang D Genetic diversity and taxonomic implication of Cordyceps sinensis as revealed by RAPD markers Biochem Genet 1999 37 201 13 10.1023/A:1018738706123 10544805 20. Sun X Gao Y Yang Y Recent advancement in microbial environmental research using metagenomics tools Biodivers Sci 2013 21 393 400 21. Kirk JL Beaudette LA Hart M Moutoglis P Klironomos JN Lee H Trevors JT Methods of studying soil microbial diversity J Microbiol Methods 2004 58 169 88 10.1016/j.mimet.2004.04.006 15234515 22. Pace NR A molecular view of microbial diversity and the biosphere Science 1997 276 734 40 10.1126/science.276.5313.734 9115194 23. Metzker ML Sequencing technologies-the next generation Nat Rev Genet 2010 11 31 46 10.1038/nrg2626 19997069 24. Buee M Reich M Murat C Morin E Nilsson RH Uroz S Martin F 454 Pyrosequencing analyses of forest soils reveal an unexpectedly high fungal diversity New Phytol 2009 184 449 56 10.1111/j.1469-8137.2009.03003.x 19703112 25. Roesch LF Fulthorpe RR Riva A Casella G Hadwin AK Kent AD Daroub SH Camargo FA Farmerie WG Triplett EW Pyrosequencing enumerates and contrasts soil microbial diversity ISME J 2007 1 283 90 18043639 26. Chen XC Song JY Dong LL Yao H Chen SL Han JP Study on metagenomics and genuine traditional Chinese medicinal materials Chin Tradit Herb Drugs 2012 43 2315 20 27. Melcher U Verma R Schneider WL Metagenomic search strategies for interactions among plants and multiple microbes Front Plant Sci 2014 5 268 10.3389/fpls.2014.00268 24966863 28. Massana R Eukaryotic picoplankton in surface oceans Annu Rev Microbiol 2011 65 91 110 10.1146/annurev-micro-090110-102903 21639789 29. Becher D Bernhardt J Fuchs S Riedel K Metaproteomics to unravel major microbial players in leaf litter and soil environments: challenges and perspectives Proteomics 2013 13 2895 909 23894095 30. Xia F Liu Y Shen GR Guo LX Zhou XW Investigation and analysis of microbiological communities in natural Ophiocordyceps sinensis Can J Microbiol 2014 3 104 11 25578897 31. Zhang Y Sun B Zhang S Wang M Liu X Gong W Mycobiotal investigation of natural Ophiocordyceps sinensis based on culture-dependent investigation Mycosystema 2010 29 518 27 32. Li T Song B Wang CQ Li TH Analysis on soil fungal community structures in propagation of Cordyceps sinensis by denaturing gradient gel electrophoresis Chin Tradit Herb Drugs 2013 44 477 81 33. Chen YQ Ning W Hui Z Liang HQ Differentiation of medicinal Cordyceps species by rDNA ITS sequence analysis Planta Med 2002 68 635 9 10.1055/s-2002-32892 12142999 34. Liu ZY Yao YJ Liang ZQ Liu AY Pegler DN Chase MW Molecular evidence for the anamorph-teleomorph connection in Cordyceps sinensis Mycol Res 2001 105 827 32 10.1017/S095375620100377X 35. Shrestha B Zhang W Zhang Y Liu X What is the Chinese caterpillar fungus Ophiocordyceps sinensis (Ophiocordycipitaceae)? Mycology 2010 1 228 36 10.1080/21501203.2010.536791 36. Chen YQ Hu B Xu F Zhang W Zhou H Qu LH Genetic variation of Cordyceps sinensis , a fruit-body-producing entomopathogenic species from different geographical regions in China FEMS Microbiol Lett 2004 230 153 8 10.1016/S0378-1097(03)00889-9 14734179 37. Xiao W Yang JL Zhu P Cheng K He H Zhu H Wang Q Non-support of species complex hypothesis of Cordyceps sinensis by targeted rDNA-ITS sequence analysis Mycosystema 2009 28 724 30 38. Zhou J Xia B Huang H Palumbo AV Tiedje JM Microbial diversity and heterogeneity in sandy subsurface soils Appl Environ Microbiol 2004 70 1723 34 10.1128/AEM.70.3.1723-1734.2004 15006798 39. Janssen PH Identifying the dominant soil bacterial taxa in libraries of 16S rRNA and 16S rRNA genes Appl Environ Microbiol 2006 72 1719 28 10.1128/AEM.72.3.1719-1728.2006 16517615 40. Dillon RJ Dillon VM The gut bacteria of insects: nonpathogenic interactions Annu Rev Entomol 2004 49 71 92 10.1146/annurev.ento.49.061802.123416 14651457 41. Zhu JS Yao Y Chen W Zheng T Lu J Guo Y Molecular co-existence of Paecitomyces hepiali and Hirsutella sinensis in caterpillar and fruiting bodies of Cordyceps sinensis FASEB J 2007 21 A1079 10.1096/fj.07-8386com 42. Daniel R The soil metagenome-a rich resource for the discovery of novel natural products Curr Opin Biotechnol 2004 15 199 204 10.1016/j.copbio.2004.04.005 15193327 43. Lane DJ Pace B Olsen GJ Stahl DA Sogin ML Pace NR Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses Proc Natl Acad Sci U S A 1985 82 6955 9 10.1073/pnas.82.20.6955 2413450 44. Larena I Salazar O González V Julián MC Rubio V Design of a primer for ribosomal DNA internal transcribed spacer with enhanced specificity for ascomycetes J Biotechnol 1999 75 187 94 10.1016/S0168-1656(99)00154-6 10553657 45. Manter DK Vivanco JM Use of the ITS primers, ITS1F and ITS4, to characterize fungal abundance and diversity in mixed-template samples by qPCR and length heterogeneity analysis J Microbiol Methods 2007 71 7 14 10.1016/j.mimet.2007.06.016 17683818 46. Nilsson RH Kristiansson E Ryberg M Hallenberg N Larsson KH Intraspecific ITS variability in the kingdom Fungi as expressed in the international sequence databases and its implications for molecular species identification Evol Bioinforma 2008 4 193 201 47. RcoreTeam R: A language and environment for statistical computing 2013 Vienna R Foundation for Statistical Computing 48. Holm S A simple sequentially rejective multiple test procedure Scand J Stat. 1979 6 65 70 49. Schloss PD Westcott SL Ryabin T Hall JR Hartmann M Hollister EB Lesniewski RA Oakley BB Parks DH Robinson CJ Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities Appl Environ Microbiol 2009 75 7537 41 10.1128/AEM.01541-09 19801464 50. Pruesse E Quast C Knittel K Fuchs BM Ludwig W Peplies J Glöckner FO SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB Nucleic Acids Res 2007 35 7188 96 10.1093/nar/gkm864 17947321 51. Edgar RC Search and clustering orders of magnitude faster than BLAST Bioinformatics 2010 26 2460 1 10.1093/bioinformatics/btq461 20709691 52. Kõljalg U Larsson KH Abarenkov K Nilsson RH Alexander IJ Eberhardt U Erland S Høiland K Kjøller R Larsson E UNITE: a database providing web-based methods for the molecular identification of ectomycorrhizal fungi New Phytol 2005 166 1063 8 10.1111/j.1469-8137.2005.01376.x 15869663 53. Caporaso JG Kuczynski J Stombaugh J Bittinger K Bushman FD Costello EK Fierer N Pena AG Goodrich JK Gordon JI QIIME allows analysis of high-throughput community sequencing data Nat Methods 2010 7 335 6 10.1038/nmeth.f.303 20383131

PMC005xxxxxx/PMC5002180.txt

==== Front World J Surg OncolWorld J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 97810.1186/s12957-016-0978-2ResearchAssociation between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility: a meta-analysis based on 38 case-control studies with 40,985 subjects Zhang Jinghong zhangjinghong_zjh@163.com Zhang Lijun zhanglijun_zlj16@163.com Li Guangming +86 13501122244liguangming_lgm16@163.com Department of General Surgery, Beijing TongRen Hospital, No.1 Jia Dong Jiao Min Xiang, Dongcheng District, Beijing, 100730 China 27 8 2016 27 8 2016 2016 14 1 2305 5 2016 13 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 Studies investigating the association between the methylenetetrahydrofolate reductase (MTHFR) gene 1298A>C polymorphism and the risk of breast cancer have reported inconsistent results. So, we performed this updated meta-analysis and tried to give a more precise estimation of association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility. Methods Relevant studies published before 1 January 2016 were identified by searching PubMed and EMBASE. The strength of relationship between the MTHFR gene 1298A>C polymorphism and breast cancer susceptibility was assessed using odds ratio (OR) and corresponding 95 % confidence interval (95 % CI). The meta-analysis was performed using Stata 11.0 software. Results A total number of 38 case-control studies including 18,686 cases and 22,299 controls were identified. No association was found in five genetic models (dominant model: OR = 0.99, 95 % CI 0.99–1.00, P = 0.218; recessive model: OR = 1.00, 95 % CI 0.97–1.02, P = 0.880; homozygote genetic model: OR = 0.99, 95 % CI 0.98–1.01, P = 0.390; heterozygote genetic model: OR = 0.99, 95 % CI 0.97–1.00, P = 0.138; and allele contrast genetic model: OR = 0.99, 95 % CI 0.98–1.01) for MTHFR gene 1298 A>C polymorphism and breast cancer susceptibility. In the subgroup analysis stratified by source of control, decreased risk of breast cancer was found in studies with hospital-based controls in dominant model (OR = 0.98, 95 % CI 0.96–1.00, P = 0.037). Conclusions Our meta-analysis suggested that there is no significant association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility for overall population. Keywords MTHFR gene 1298A>C polymorphismBreast cancerGene polymorphismMeta-analysisOne-carbon metabolismVariantissue-copyright-statement© The Author(s) 2016 ==== Body Background Breast cancer is the most frequently diagnosed cancer among women, which contributed to 25 % of all cancer cases in women worldwide, and it is the leading cause of female cancer-related death [1]. In UK, 48,034 women were diagnosed as breast cancer holders in 2008, and in USA, more than 2.8 million women suffered from breast cancer in 2015 [2, 3]. In China, breast cancer mortality have also raised quickly in recent years, from 3.53/100,000 in 1990–1992 to 4.25 in 2012 [4]. The high morbidity and mortality of the disease lead to increasing global public health burden gradually. It is widely accepted that several factors, such as hormonal, environmental, and genetic factors as well as their interactions contribute to the onset of breast cancer [5, 6]. In 1993, mutations in breast cancer (BRCA1) gene were suggested to be linked with high incidence of breast cancer in some families [7]. Since then, many susceptible genes involved in initiation and evolution of breast cancer have been researched, and one of them, the methylenetetrahydrofolate reductase (MTHFR) gene has been widely studied. The MTHFR locus locates on chromosome 1 at the end of short arm (1p36.6), which encodes enzymes relevant to folates metabolism. The enzyme encoded by MTHFR gene takes part in the irreversible conversion of 5,10-metylenetetrahydrofolate to 5-methyltetrahydrofolate, which plays a crucial role in homocysteine remethylation to methionine [8]. Previous studies have indicated that functional single nucleotide polymorphisms (SNPs) of MTHFR gene participate in the folate-metabolizing genetic pathway and are fundamental during the synthesis, repair, and methylation process of DNA, RNA, and protein, which may affect folate and vitamin B12 level [9, 10]. Of these SNPs, 1298A>C polymorphism is caused by A to C transition in exon 7 and results in alanine in substitution of glutamine at codon 429 of the protein [11]. Subjects with mutated MTHFR 1298A>C genetic polymorphisms have higher plasma level of homocysteine [12] and may be more susceptible to different kinds of cancers, including breast cancer. Many studies have investigated the association between MTHFR gene 1298A>C polymorphism and breast cancer risk. However, the results are inconsistent, with some studies found significant association [13, 14], while others were not [15, 16]. Although previous meta-analysis has tried to clarify the association [17], recently, several new case-control studies have been published [18–20]. In order to avoid the limitations of single case-control studies and provide renewed evidence, we performed this updated meta-analysis and tried to give a more precise and comprehensive estimation of association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility. Methods Data sources Two databases were electronically searched, including PubMed and EMBASE, to retrieve studies analyzing the association between breast cancer susceptibility and MTHFR gene 1298A>C polymorphism until January 1, 2016. Searching terms were “breast cancer” or “breast neoplasm”, in combination with “methylenetetrahydrofolate reductase” or “MTHFR” or “MTHFR A1298C” or “MTHFR 1298A>C” or “rs1801131” or “Glu429Ala”, and in combination with, “polymorphism” or “variant” or “genotype” or “allele”. We also hand-checked the reference lists of all the included studies to make sure no study was missed. Two researches conducted the searches independently. If several publications carried out among same patients and controls, we only included one study with the most complete data. Inclusion criteria We first performed initial screening of titles and abstract. A second round screening was based on full-text reviews. Studies were considered eligible if they met the following criteria: (1) it was a case-control study in design; (2) it evaluated the MTHFR gene 1298 A>C polymorphism and breast cancer susceptibility; (3) breast cancer was pathologically confirmed for all of the patients; (4) sample sizes and individual genotype frequencies in cases and controls were available; and (5) cases and controls should be matched. Exclusion criteria Researches were excluded if they met any one of the following criteria: (1) data came from reviews or abstracts; (2) genotype and allele frequencies were both unavailable; (3) subjects with other malignant tumor were included in controls; (4) repeatedly published literature; (5) not breast cancer susceptibility outcome; and (6) controls were chosen from women with a family history of breast cancer or with other kinds of malignant tumors. Data extraction and quality assessment Two reviewers independently searched and selected literature, and then, extracted relevant data according to a data extraction form. Disagreements were solved by discussion until consensus was made. The extracted data including the first author, year of publication, country of origin, ethnicity of the study population, source of control, sample size, the genotype and allele frequencies of the MTHFR gene 1298A>C polymorphism, and information of Hardy-Weinberg equilibrium (HWE) in control groups. Different ethnicity descents were categorized as Caucasian, Asian, African, and if studies were with more than one ethnicity, they were categorized as mixed ethnicity. For each included study, the quality assessment was conducted according to the STrengthening the REporting of Genetic Association (STREGA) studies). If the study met all or most of the criteria in this approach, it would be classified as “++” or “high quality”. For study in which some of the criteria were fulfilled and the others were not likely to change the results and conclusions, it would be graded as “+” or “moderate quality”. For studies fulfilled few or no criteria and the results were thought to be with non-ignorable bias, it would be classified as “−” or “low quality” [21]. Statistical analysis Data analysis was conducted using STATA 11.0 software (Stata Statistical software, College Station, TX, USA, www.stata.com). Odds ratio (OR) and its corresponding 95 % confidence intervals (95 % CI) were used to evaluate the strength of association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility. Heterogeneity among included studies was tested using chi-square-based Q test and I2 test. Phet < 0.05 or I2 > 50 % were considered as statistically significant for heterogeneity. The Mantel-Haenszel method was used for fix-effect model if no heterogeneity was found. Otherwise, the DerSimonian-Laird random-effect model was used. Fix-effect model considers that across all studies, the genetic factors have similar effects on genetic disorder susceptibility and the observed differences among studies are caused just by chance [22]. Random-effect model considers that different studies may have substantial diversity, and it calculates within- as well as between-study difference [23]. Five comparison genetic models were used to assess the association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility. We assessed the dominant model (AA + AC vs. CC), recessive model (AA vs. AC + CC), allele contrast genetic model (A vs. C), the heterozygote comparison (AC vs. CC), and the homozygote comparison (AA vs. CC). P < 0.05 showed the statistical significance. HWE was tested for included studies if no relevant information was provided in original research. Sensitivity analyses were conducted by omitting individual studies sequentially. Moreover, we performed subgroup analysis stratified by ethnicity, source of control, and deviation from HWE. Publication bias was quantitatively assessed by Egger’s linear regression test [24] and visual inspection of Begg’s funnel plots. Results Literature search We initially identified 373 potentially relevant studies from searching the two databases and the reference lists of relevant studies. Firstly, we eliminated duplications, and after this procedure, 248 studies were retained. After reading the titles and abstracts, we excluded 193 studies. Among them, 89 were not case-control studies, 91 were irrelevant to MTHFR polymorphism or breast cancer susceptibility, and 13 were reviews or meta-analysis. Then, we read the full texts of the 55 retained articles and 17 were excluded. Of them, 11 was irrelevant to 1298A>C polymorphism, four focused on breast cancer mortality, one conducted among the same patients and controls with another study, but provided less completed data, and for one study, the controls were chosen from BRCA1 carriers. We finally identified 38 case-control studies eligible for the meta-analysis [13–16, 18–20, 25–55], including 18,686 cases and 22,299 controls. A flow chart of data selection was presented in Fig. 1.Fig. 1 Flow chart of data selection Main characteristics of included studies Table 1 presents the main characteristics and genotype frequencies of the included studies. Of the 38 studies, 15 studies were carried out among Asians, 13 among Caucasians, and 10 among mixed populations. All studies included were case-control studies in design, and all patients with breast cancer fulfilled the pathological diagnosis. The number ranged from 35 to 1986 for cases, and 33 to 2414 for controls. In 21 studies, controls were normal healthy people randomly recruited from general population, and in 15 studies, controls were recruited from hospital among women with benign disease or through women going to hospital for routine physical examines, but in the two studies, we were unable to find out the source of controls. In most of the included studies, controls were matched with cases in ethnicity and age. In quality assessment, 17 studies included were categorized as “high quality,” and 21 as “moderate quality” (Table 1). In eight studies, the genotype distributions in control groups were deviated from HWE (Table 1).Table 1 The main characteristics of studies included in this meta-analysis and the distribution of MTHFR gene 1298A>C genotypes and alleles among cases and controls First author Year Ethnicity Source of controls Cases Controls Cases Controls Deviation from HWE Quality grade AA AC CC AA AC CC A C A C Aram 2012 Caucasian HB 35 55 20 30 75 5 125 95 135 85 Yes + Awwad 2015 Asian PB 68 61 17 58 64 13 197 95 180 90 No ++ Carvalho Barbosa Rde 2012 Mixed PB 68 80 17 72 84 9 216 114 228 102 Yes + Chen 2005 Mixed PB 558 417 87 536 457 110 1533 591 1529 677 No ++ Cheng 2008 Asian HB 207 125 19 310 207 17 539 163 827 241 Yes + Chou 2006 Asian HB 104 30 8 172 95 18 238 46 439 131 No + Ergul 2003 Caucasian HB 50 48 20 90 85 18 148 88 265 121 No + Ericson 2009 Caucasian PB 242 242 57 487 480 105 726 356 1454 690 No ++ Forsti 2004 Caucasian NA 94 102 27 133 127 38 290 156 393 203 No + Gao 2009 Asian PB 446 165 9 425 188 11 1057 183 1038 210 No ++ He 2014 Asian HB 138 132 40 173 155 53 408 212 501 261 No + Hosseini 2011 Caucasian HB 36 96 162 60 135 105 168 420 255 345 No + Inoue 2008 Asian PB 225 139 16 387 234 41 589 171 1008 316 No ++ Justenhoven 2005 Caucasian PB 273 256 53 295 266 73 802 362 856 412 No ++ Kakkoura 2015 Mixed PB 138 465 468 150 500 484 741 1401 800 1468 No ++ Kotsopoulos 2008 Caucasian HB 466 390 85 398 309 73 1322 560 1105 455 No + Lajin 2012 Caucasian HB 44 52 23 65 48 13 140 98 178 74 No + Le Marchand 2004 Mixed PB 741 372 77 1493 801 120 1854 526 3787 1041 No ++ Lissowska 2007 Caucasian PB 892 874 220 1086 941 251 2658 1314 3113 1443 Yes + Liu 2013 Asian HB 206 176 53 214 172 49 588 282 600 270 No + Lopez-Cortes 2015 Mix PB 110 3 1 191 3 1 223 5 385 5 Yes + Lu 2015 Asian HB 369 172 19 352 185 23 910 210 889 231 No + Ma 2009 Mixed HB 269 168 21 279 157 22 706 210 715 201 No + Mir 2008 Asian NA 15 19 1 11 22 0 49 21 44 22 Yes + Ozen 2013 Mix PB 17 29 5 71 35 0 63 39 177 35 Yes + Papandreou 2012 Caucasian HB 129 135 36 136 116 31 393 207 388 178 No + Platek 2009 Mix PB 443 402 83 842 758 181 1288 568 2442 1120 No ++ Qi 2004 Asian PB 155 58 4 144 71 3 368 66 359 77 No ++ Sangrajrang 2010 Asian HB 302 223 38 258 206 23 827 299 722 252 Yes + Sharp 2002 Caucasian PB 27 25 3 24 25 11 79 31 73 47 No ++ Shrubsole 2004 Asian PB 768 311 42 824 344 40 1847 395 1992 424 No ++ Stevens 2007 Mixed PB 224 228 42 252 201 40 676 312 705 281 No ++ Vainer 2010 Caucasian HB 398 353 80 379 330 76 1149 513 1088 482 No + Weiwei 2014 Asian HB 135 129 32 151 130 25 399 193 432 180 No + Wu 2012 Asian PB 37 32 6 42 28 5 106 44 112 38 No ++ Xu 2007 Mixed PB 558 417 87 536 457 110 1533 591 1529 677 No ++ Zhang 2015 Asian PB 98 87 31 105 84 27 283 149 294 138 No ++ Ziva Cerne 2011 Caucasian PB 258 219 47 131 117 21 735 313 379 159 No ++ PB population-based study, HB hospital-based study, NA not available, HWE Hardy-Weinberg equilibrium Quantitative data analysis Association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility The results of the five genetic models testing MTHFR gene 1298A>C polymorphism and breast cancer susceptibility are presented in Table 2. In the dominant model (AA + AC vs. CC), P value for heterogeneity was 0.000, and I2 was 50.5 %, indicating significant heterogeneity among studies. Thus, random-effect model was used. The overall effect Z value was 1.12 (P = 0.218) and OR was 0.99 (95 % CI 0.99–1.00), suggesting that no association was found in the dominant model. The Egger’s linear regression test indicated that there was some evidence of publication bias in this model (Egger, P = 0.01). Other four genetic models were also performed (Table 2), but no association was found. In subgroup analyses stratified by source of control, a significant decrease in breast cancer susceptibility was found in hospital-based controls in dominant model (OR = 0.98, 95 % CI 0.96–1.00, P = 0.037), but not in allele contrast genetic model (OR = 0.97, 95 % CI 0.94–1.00, P = 0.092) (Table 3). Moreover, the results showed that in subgroups of Asians and population-based studies, the heterogeneity among studies was significantly reduced. Figure 2 shows the forest plot of the dominant model testing the association between MTHFR 1298A>C polymorphism and breast cancer risk, stratified by ethnicity. Figure 3 shows the forest plot of the dominant model testing the association between MTHFR 1298A>C polymorphism and breast cancer risk, stratified by source of control.Table 2 Summary of different genetic model comparison results of MTHFR gene 1298A>C polymorphism Genetic model OR (95 % CI) Z P value I 2 % P het Effect model Egger’s test t value P value AA + AC vs. CC 0.99 (0.99–1.00) 1.23 0.218 50.5 0.000 R −2.72 0.010 AA vs. AC + CC 1.00 (0.97–1.02) 0.15 0.880 35.9 0.016 R −1.45 0.155 AA vs. CC 0.99 (0.98–1.01) 0.86 0.390 43.8 0.002 R −2.75 0.014 AC vs. CC 0.99 (0.97–1.00) 1.48 0.138 41.2 0.005 R −2.55 0.015 A vs. C 0.99 (0.98–1.01) 0.92 0.360 55.5 0.000 R −2.27 0.029 OR odds ratio, CI confidence interval, R random-effect model, P het P value for heterogeneity P < 0.05 stands for statistical significance Table 3 Results of subgroup analyses of MTHFR gene 1298A>C polymorphism Stratified by Comparison Number of datasets Dominant genetic model Allele contrast OR (95 % CI) P value OR (95 % CI) P value Ethnicity Asian 15 1.00 (0.99–1.00) 0.506 1.01 (0.99–1.02) 0.249 Caucasian 13 0.98 (0.95–1.01) 0.129 0.97 (0.93–1.00) 0.059 Mixed 10 0.99 (0.99–1.00) 0.852 0.99 (0.98–1.01) 0.660 Source of control PB 21 1.00 (0.99–1.01) 0.931 1.00 (0.99–1.02) 0.830 HB 15 0.98 (0.96–1.00) 0.037 0.97 (0.94–1.00) 0.092 NA 2 0.99 (0.94–1.04) 0.793 0.99 (0.91–1.08) 0.892 Deviation from HWE Yes 8 0.98 (0.95–1.00) 0.019 0.98 (0.95–1.01) 0.102 No 30 1.00 (0.99–1.01) 0.909 1.00 (0.98–1.01) 0.801 PB population-based study, HB hospital-based study, NA not available Fig. 2 Shows the forest plot of the dominant model testing the association between MTHFR 1298A>C polymorphism and breast cancer risk, stratified by ethnicity Fig. 3 Shows the forest plot of the dominant model testing the association between MTHFR 1298A>C polymorphism and breast cancer risk, stratified by source of control Sensitivity analysis and publication bias Sensitivity analyses were conducted by omitting each dataset sequentially, and the result did not change under any genetic model. Sensitivity analysis suggested that for all of the five genetic comparisons of MTHFR gene 1298A>C polymorphism and breast cancer susceptibility, the results were statistically robust. Visual inspection of Begg’s funnel plots identified the substantial asymmetry for dominant model, the allele contrast genetic model, the heterozygote comparison, and the homozygote comparison. The Egger’s linear regression test also indicated the similar results (P < 0.05 for all models tested except the recessive genetic model) (Table 2). Figure 4 shows the Begg’s funnel plot under dominant model of MTHFR 1298A>C polymorphism.Fig. 4 Begg’s funnel plot under dominant model of MTHFR 1298A>C polymorphism Discussion MTHFR is an essential gene in the one-carbon metabolism pathway. During the past few years, many meta-analyses assessing the association between MTHFR gene polymorphism and cancer risks have been published, including liver cancer, ovary cancer, lung cancer, gastric cancer, pancreatic cancer, cervical cancer, and esophageal cancer [56–60]. Genetic variation in enzymes and proteins involved in folate metabolism is also a rational candidate for studying the genetic of breast cancer. Therefore, the interest in MTHFR gene 1298A>C polymorphism and breast cancer susceptibility has existed for a long time. In 2002, Sharp et al. for the first time published a case-control study estimating the association between MTHFR gene 1298A>C polymorphism and breast cancer risk. Their result suggested that risk was significantly lower for the 1298CC genotype compared to AA genotype (OR = 0.24, 95 % CI 0.06–0.97) [49]. However, after that, a number of subsequent studies were conducted and their results were inconsistent, with some studies showed significant associations while others were not. The inconsistency may be caused by several reasons. First of all, although in vitro, the variant genotype is associated with a substantial decrease in enzymatic activity [11], this functional polymorphism may be an important but not the exclusive influencing factor in etiology and pathogenesis of breast cancer. Special lifestyle and environmental factors, such as tea drinking [61], dietary intake of folate, vitamin B6 and B12 [62], physical activities [63], long-term oral contraceptive use [64], and hormone replacement therapy use [65], are possibly confounding factors taking part in the disease etiology. Moreover, differences in patient choosing criteria, ethnicity, sample size, and sources of control could contribute to inconsistency. Hence, it is necessary to conduct a meta-analysis providing quantitative approach for pooling the results of all studies with the same purpose and explaining the overall estimation as well as the diversity. Our study has important strengths. All original studies used a case-control study design, which is a useful tool to identify gene and disease associations. However, individual genotype case-control studies could not be based on a large number of subjects or contain patients in different ethnicities, and thus has insufficient statistical power. Our meta-analysis based on case-control studies involving 40,985 subjects brings to light that there is no significant association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility for overall population, with ORs from 0.99 to 1.00 and narrow 95 % CIs for all of the five genetic models. Moreover, in our study, no association was found in different ethnicities or in population-based studies, which thereby strengthened this association. As shown in our meta-analysis, studies with hospital-based design or controls deviated from HWE had a weak, but statistical significant decreased association with breast cancer in dominant model. However, in these two kinds of studies, the controls may not represent the whole population and thereby, the results from them should be interpreted with caution. Overall, our meta-analysis based on 38 case-control studies provided reliable and comprehensive estimations. The association in the five genetic models sustained unchanged in the sensitivity analysis, which further confirmed the results of main analysis. It is also important to mention the heterogeneity existed in this study. For all genetic models in the main analysis, P value for heterogeneity was less than 0.05, indicating significant heterogeneity among the included studies. Finding the potential sources of heterogeneity is an important part of meta-analysis, which can greatly influence the results of the research. To detect the possible source of heterogeneity in our meta-analysis, we performed the subgroup analysis stratified by ethnicity, source of control, and deviation from HWE. When stratified by ethnicity and source of control, the heterogeneity was significantly decreased in Asian and population-based subgroups. Therefore, the different ethnicity and source of control may contribute to the overall heterogeneity. However, heterogeneity still existed in Caucasian, mixed ethnicity, and hospital-based control subgroups, suggesting that ethnicity and source of control did not fully explain the heterogeneity among studies. Further studies may try to explore in interactions between different factors and to minimize the heterogeneity in subgroups. Several previous meta-analyses have been published to analyze the association between MTHFR gene polymorphisms and breast cancer susceptibility, and the majority of them concerned on 677C>T polymorphism [66–69]. Two studies published in 2014 have detected the 1298A>C polymorphism [17, 70]. The main result of our study was consistent with the previous meta-analyses. Comparing with these two studies, our study has some important improvements. In 2014–2015, some new studies were published and they were included in our meta-analysis. Through strict methodological process, we provided a more comprehensive view of included studies. The abovementioned meta-analyses only stratified by ethnicity to test if there existed differences in variant ethnicities. In present study, we also conducted subgroup study stratified by source of control and deviation from HWE in control group, to analyze if there were differences between subgroups. We should also pay attention to the several limitations in our study, which may affect the result. Firstly, we only included published studies meeting our inclusion criteria from two databases, similar studies in other databases and unpublished researches may have been missed, and this is also the main reason for the publication bias we found in four of the five genetic models. Secondly, the control groups in some of the included studies were deviated from HWE, which may fail to represent the whole population and have some effects on the overall estimation. Thirdly, although the results from subgroup and sensitivity analyses were quite similar to the main analysis, significant heterogeneity was detected in all five genetic models of MTHFR gene 1298A>C polymorphism and breast cancer susceptibility. Different characteristics in study population and study design may contribute to the heterogeneity. Considering that meta-analysis is a kind of retrospective research and may easily be affected by methodological deficiencies of the included studies, we developed a detailed protocol before conducting this analysis, to ensure the quality of our research. Conclusions From the combination results of currently included studies, our meta-analysis suggested that there is no significant association between MTHFR gene 1298A>C polymorphism and breast cancer susceptibility for overall population. Abbreviations MTHFRMethylenetetrahydrofolate reductase OROdds ratio SNPsSingle nucleotide polymorphisms HWEHardy-Weinberg equilibrium Acknowledgements None. Funding None. Availability of data and materials This research is a meta-analysis, and all data and materials are available in database of PubMed (http://www.ncbi.nlm.nih.gov/pubmed/) and EMBASE (http://www.embase.com/). Authors’ contributions JHZ and GML wrote the paper. JHZ and JLZ analyzed the data. GML organized the whole work. All authors read and approved the final manuscript. Competing interests All 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. Torre LA Bray F Siegel RL Ferlay J Lortet-Tieulent J Jemal A Global cancer statistics, 2012 CA Cancer J Clin 2015 65 2 87 108 10.3322/caac.21262 25651787 2. CancerStats Incidence 2008-UK. Available from: http://info.cancerresearchuk.org/prod_consump/groups/cr_common/@nre/@sta/documents/generalcontent/cr_072111.pdf Accessed date 2 Feb 2016. 3. McGuire A Brown JA Malone C McLaughlin R Kerin MJ Effects of age on the detection and management of breast cancer Cancers 2015 7 2 908 29 10.3390/cancers7020815 26010605 4. Yearbook of Chinese health statistics, 2013. Available from: http://www.nhfpc.gov.cn/htmlfiles/zwgkzt/ptjnj/year2013/index2013.html Accessed date 2 Feb 2016. 5. Hankinson SE Colditz GA Willett WC Towards an integrated model for breast cancer etiology: the lifelong interplay of genes, lifestyle, and hormones Breast Cancer Research: BCR 2004 6 5 213 8 10.1186/bcr921 15318928 6. Dumitrescu RG Cotarla I Understanding breast cancer risk—where do we stand in 2005? J Cell Mol Med 2005 9 1 208 21 10.1111/j.1582-4934.2005.tb00350.x 15784178 7. Weber BL Abel KJ Couch FJ Merajver SD Chandrasekharappa SC Castilla L Progress toward isolation of a breast cancer susceptibility gene, BRCA1 Cold Spring Harb Symp Quant Biol 1994 59 531 6 10.1101/SQB.1994.059.01.059 7587109 8. Xu X Gammon MD Zeisel SH Lee YL Wetmur JG Teitelbaum SL Choline metabolism and risk of breast cancer in a population-based study FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology 2008 22 6 2045 52 10.1096/fj.07-101279 18230680 9. Wei LK Sutherland H Au A Camilleri E Haupt LM Gan SH A potential epigenetic marker mediating serum folate and vitamin B12 levels contributes to the risk of ischemic stroke BioMed Research International 2015 2015 167976 25705649 10. Xu X Chen J One-carbon metabolism and breast cancer: an epidemiological perspective Journal of Genetics and Genomics = Yi chuan xue bao 2009 36 4 203 14 10.1016/S1673-8527(08)60108-3 19376481 11. Weisberg I Tran P Christensen B Sibani S Rozen R A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity Mol Genet Metab 1998 64 3 169 72 10.1006/mgme.1998.2714 9719624 12. Lee YS Han DH Jeon CM Lyoo IK Na C Chae SL Serum homocysteine, folate level and methylenetetrahydrofolate reductase 677, 1298 gene polymorphism in Korean schizophrenic patients Neuroreport 2006 17 7 743 6 10.1097/01.wnr.0000215777.99473.52 16641680 13. Ergul E Sazci A Utkan Z Canturk NZ Polymorphisms in the MTHFR gene are associated with breast cancer Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine 2003 24 6 286 90 10.1159/000076460 15004488 14. Stevens VL McCullough ML Pavluck AL Talbot JT Feigelson HS Thun MJ Association of polymorphisms in one-carbon metabolism genes and postmenopausal breast cancer incidence Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology 2007 16 6 1140 7 10.1158/1055-9965.EPI-06-1037 15. Ericson U Sonestedt E Ivarsson MI Gullberg B Carlson J Olsson H Folate intake, methylenetetrahydrofolate reductase polymorphisms, and breast cancer risk in women from the Malmo Diet and Cancer cohort Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology 2009 18 4 1101 10 10.1158/1055-9965.EPI-08-0401 16. Justenhoven C Hamann U Pierl CB Rabstein S Pesch B Harth V One-carbon metabolism and breast cancer risk: no association of MTHFR, MTR, and TYMS polymorphisms in the GENICA study from Germany Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology 2005 14 12 3015 8 10.1158/1055-9965.EPI-05-0592 17. Li K Li W Dong X Association of 677 C>T (rs1801133) and 1298 A>C (rs1801131) polymorphisms in the MTHFR gene and breast cancer susceptibility: a meta-analysis based on 57 individual studies PLoS One 2014 9 6 e71290 10.1371/journal.pone.0071290 24945727 18. Zhang XF Liu T Li Y Li S Association between MTHFR 677C/T and 1298A/C gene polymorphisms and breast cancer risk Genetics and Molecular Research: GMR 2015 14 4 16425 30 10.4238/2015.December.9.12 26662439 19. Lu Q Jiang K Li Q Ji YJ Chen WL Xue XH Polymorphisms in the MTHFR gene are associated with breast cancer risk and prognosis in a Chinese population Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine 2015 36 5 3757 62 10.1007/s13277-014-3016-4 25566964 20. Awwad N Yousef AM Abuhaliema A Abdalla I Yousef M Relationship between genetic polymorphisms in MTHFR (C677T, A1298C and their Haplotypes) and the incidence of breast cancer among Jordanian females—case-control study Asian Pacific Journal of Cancer Prevention: APJCP 2015 16 12 5007 11 10.7314/APJCP.2015.16.12.5007 26163632 21. Little J Bradley L Bray MS Clyne M Dorman J Ellsworth DL Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations Am J Epidemiol 2002 156 4 300 10 10.1093/oxfordjournals.aje.a000179 12181099 22. DerSimonian R Laird N Meta-analysis in clinical trials Control Clin Trials 1986 7 3 177 88 10.1016/0197-2456(86)90046-2 3802833 23. Higgins JP Thompson SG Quantifying heterogeneity in a meta-analysis Stat Med 2002 21 11 1539 58 10.1002/sim.1186 12111919 24. Egger M Davey Smith G Schneider M Minder C Bias in meta-analysis detected by a simple, graphical test BMJ 1997 315 7109 629 34 10.1136/bmj.315.7109.629 9310563 25. Akram M Malik FA Kayani MA Mutational analysis of the MTHFR gene in breast cancer patients of Pakistani population Asian Pacific Journal of Cancer Prevention: APJCP 2012 13 4 1599 603 10.7314/APJCP.2012.13.4.1599 22799374 26. Carvalho Barbosa Rde C Menezes DC Freire TF Sales DC Alencar VH Rabenhorst SH Associations of polymorphisms of folate cycle enzymes and risk of breast cancer in a Brazilian population are age dependent Mol Biol Rep 2012 39 4 4899 907 10.1007/s11033-011-1285-1 22134752 27. Chen J Gammon MD Chan W Palomeque C Wetmur JG Kabat GC One-carbon metabolism, MTHFR polymorphisms, and risk of breast cancer Cancer Res 2005 65 4 1606 14 10.1158/0008-5472.CAN-04-2630 15735051 28. Cheng CW Yu JC Huang CS Shieh JC Fu YP Wang HW Polymorphism of cytosolic serine hydroxymethyltransferase, estrogen and breast cancer risk among Chinese women in Taiwan Breast Cancer Res Treat 2008 111 1 145 55 10.1007/s10549-007-9754-x 17896178 29. Chou YC Wu MH Yu JC Lee MS Yang T Shih HL Genetic polymorphisms of the methylenetetrahydrofolate reductase gene, plasma folate levels and breast cancer susceptibility: a case-control study in Taiwan Carcinogenesis 2006 27 11 2295 300 10.1093/carcin/bgl108 16777985 30. Forsti A Angelini S Festa F Sanyal S Zhang Z Grzybowska E Single nucleotide polymorphisms in breast cancer Oncol Rep 2004 11 4 917 22 15010895 31. Gao CM Tang JH Cao HX Ding JH Wu JZ Wang J MTHFR polymorphisms, dietary folate intake and breast cancer risk in Chinese women J Hum Genet 2009 54 7 414 8 10.1038/jhg.2009.57 19557016 32. He JM Pu YD Wu YJ Qin R Zhang QJ Sun YS Association between dietary intake of folate and MTHFR and MTR genotype with risk of breast cancer Genetics and Molecular Research: GMR 2014 13 4 8925 31 10.4238/2014.October.31.7 25366783 33. Hosseini M Houshmand M Ebrahimi A MTHFR polymorphisms and breast cancer risk Archives of Medical Science: AMS 2011 7 1 134 7 10.5114/aoms.2011.20618 22291746 34. Inoue M Robien K Wang R Van Den Berg DJ Koh WP Yu MC Green tea intake, MTHFR/TYMS genotype and breast cancer risk: the Singapore Chinese Health Study Carcinogenesis 2008 29 10 1967 72 10.1093/carcin/bgn177 18669903 35. Kakkoura MG Demetriou CA Loizidou MA Loucaides G Neophytou I Marcou Y Single-nucleotide polymorphisms in one-carbon metabolism genes, Mediterranean diet and breast cancer risk: a case-control study in the Greek-Cypriot female population Genes & Nutrition 2015 10 2 453 10.1007/s12263-015-0453-7 25604861 36. Kotsopoulos J Zhang WW Zhang S McCready D Trudeau M Zhang P Polymorphisms in folate metabolizing enzymes and transport proteins and the risk of breast cancer Breast Cancer Res Treat 2008 112 3 585 93 10.1007/s10549-008-9895-6 18204969 37. Lajin B Alhaj Sakur A Ghabreau L Alachkar A Association of polymorphisms in one-carbon metabolizing genes with breast cancer risk in Syrian women Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine 2012 33 4 1133 9 10.1007/s13277-012-0354-y 22373582 38. Le Marchand L Haiman CA Wilkens LR Kolonel LN Henderson BE MTHFR polymorphisms, diet, HRT, and breast cancer risk: the multiethnic cohort study Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2004 13 12 2071 7 39. Lissowska J Gaudet MM Brinton LA Chanock SJ Peplonska B Welch R Genetic polymorphisms in the one-carbon metabolism pathway and breast cancer risk: a population-based case-control study and meta-analyses International Journal of Cancer Journal International du cancer 2007 120 12 2696 703 10.1002/ijc.22604 17311260 40. Liu Y Zhou LS Xu XM Deng LQ Xiao QK Association of dietary intake of folate, vitamin B6 and B12 and MTHFR genotype with breast cancer risk Asian Pacific Journal of Cancer Prevention: APJCP 2013 14 9 5189 92 10.7314/APJCP.2013.14.9.5189 24175799 41. Lopez-Cortes A Echeverria C Ona-Cisneros F Sanchez ME Herrera C Cabrera-Andrade A Breast cancer risk associated with gene expression and genotype polymorphisms of the folate-metabolizing MTHFR gene: a case-control study in a high altitude Ecuadorian mestizo population Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine 2015 36 8 6451 61 10.1007/s13277-015-3335-0 25801246 42. Ma E Iwasaki M Junko I Hamada GS Nishimoto IN Carvalho SM Dietary intake of folate, vitamin B6, and vitamin B12, genetic polymorphism of related enzymes, and risk of breast cancer: a case-control study in Brazilian women BMC Cancer 2009 9 122 10.1186/1471-2407-9-122 19389261 43. Mir MM Dar JA Dar NA Dar MS Salam I Lone MM Combined impact of polymorphism of folate metabolism genes; glutamate carboxypeptidase, methylene tetrahydrofolate reductase and methionine synthase reductase on breast cancer susceptibility in Kashmiri women International Journal of Health Sciences 2008 2 1 3 14 21475466 44. Ozen F Erdis E Sik E Silan F Uludag A Ozdemir O Germ-line MTHFR C677T, FV H1299R and PAI-1 5G/4G variations in breast carcinoma Asian Pacific Journal of Cancer Prevention: APJCP 2013 14 5 2903 8 10.7314/APJCP.2013.14.5.2903 23803051 45. Papandreou CN Doxani C Zdoukopoulos N Vlachostergios PJ Hatzidaki E Bakalos G Evidence of association between methylenetetrahydrofolate reductase gene and susceptibility to breast cancer: a candidate-gene association study in a South-eastern European population DNA Cell Biol 2012 31 2 193 8 10.1089/dna.2011.1292 21875371 46. Platek ME Shields PG Marian C McCann SE Bonner MR Nie J Alcohol consumption and genetic variation in methylenetetrahydrofolate reductase and 5-methyltetrahydrofolate-homocysteine methyltransferase in relation to breast cancer risk Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2009 18 9 2453 9 10.1158/1055-9965.EPI-09-0159 47. Qi J Miao XP Tan W Yu CY Liang G Lu WF Association between genetic polymorphisms in methylenetetrahydrofolate reductase and risk of breast cancer Zhonghua zhong liu za zhi (Chinese Journal of Oncology) 2004 26 5 287 9 15312365 48. Sangrajrang S Sato Y Sakamoto H Ohnami S Khuhaprema T Yoshida T Genetic polymorphisms in folate and alcohol metabolism and breast cancer risk: a case-control study in Thai women Breast Cancer Res Treat 2010 123 3 885 93 10.1007/s10549-010-0804-4 20180013 49. Sharp L Little J Schofield AC Pavlidou E Cotton SC Miedzybrodzka Z Folate and breast cancer: the role of polymorphisms in methylenetetrahydrofolate reductase (MTHFR) Cancer Lett 2002 181 1 65 71 10.1016/S0304-3835(02)00030-7 12430180 50. Shrubsole MJ Gao YT Cai Q Shu XO Dai Q Hebert JR MTHFR polymorphisms, dietary folate intake, and breast cancer risk: results from the Shanghai breast cancer study Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2004 13 2 190 6 10.1158/1055-9965.EPI-03-0273 51. Vainer AS Boiarskikh UA Voronina EN Selezneva IA Sinkina TV Lazarev AF Polymorphic variants of folate metabolizing genes (C677T and A1298C MTHFR, C1420T SHMT1 and G1958A MTHFD) are not associated with the risk of breast cancer in West Siberian Region of Russia Mol Biol 2010 44 5 816 23 52. Weiwei Z Liping C Dequan L Association between dietary intake of folate, vitamin B6, B12 & MTHFR, MTR genotype and breast cancer risk Pakistan Journal of Medical Sciences 2014 30 1 106 10 24639841 53. Wu XY Ni J Xu WJ Zhou T Wang X Interactions between MTHFR C677T-A1298C variants and folic acid deficiency affect breast cancer risk in a Chinese population Asian Pacific Journal of Cancer Prevention: APJCP 2012 13 5 2199 206 10.7314/APJCP.2012.13.5.2199 22901194 54. Xu X Gammon MD Zhang H Wetmur JG Rao M Teitelbaum SL Polymorphisms of one-carbon-metabolizing genes and risk of breast cancer in a population-based study Carcinogenesis 2007 28 7 1504 9 10.1093/carcin/bgm061 17372271 55. Ziva Cerne J Stegel V Gersak K Novakovic S Lack of association between methylenetetrahydrofolate reductase genetic polymorphisms and postmenopausal breast cancer risk Mol Med Rep 2011 4 1 175 9 21461582 56. Guo LN Methylenetetrahydrofolate reductase C677T polymorphism and cervical cancer risk: a meta-analysis Asian Pacific Journal of Cancer Prevention: APJCP 2012 13 5 2193 7 10.7314/APJCP.2012.13.5.2193 22901193 57. Larsson SC Giovannucci E Wolk A Folate intake, MTHFR polymorphisms, and risk of esophageal, gastric, and pancreatic cancer: a meta-analysis Gastroenterology 2006 131 4 1271 83 10.1053/j.gastro.2006.08.010 17030196 58. Hou XH Huang YM Mi YY Methylenetetrahydrofolate reductase gene C677T polymorphism and lung cancer: an updated meta-analysis Asian Pacific Journal of Cancer Prevention: APJCP 2012 13 5 2025 9 10.7314/APJCP.2012.13.5.2025 22901166 59. Ma C Liu Y Zhang W Liu P The association between MTHFR C677T polymorphism and ovarian cancer risk: a meta-analysis of 18,628 individuals Mol Biol Rep 2013 40 3 2061 8 10.1007/s11033-012-1970-8 23329275 60. Qi X Sun X Xu J Wang Z Zhang J Peng Z Associations between methylenetetrahydrofolate reductase polymorphisms and hepatocellular carcinoma risk in Chinese population Tumour Biology: The Journal of the International Society for Oncodevelopmental Biology and Medicine 2014 35 3 1757 62 10.1007/s13277-013-1529-x 24385382 61. Gao Y Huang YB Liu XO Chen C Dai HJ Song FJ Tea consumption, alcohol drinking and physical activity associations with breast cancer risk among Chinese females: a systematic review and meta-analysis Asian Pacific Journal of Cancer Prevention: APJCP 2013 14 12 7543 50 10.7314/APJCP.2013.14.12.7543 24460331 62. Lajous M Lazcano-Ponce E Hernandez-Avila M Willett W Romieu I Folate, vitamin B(6), and vitamin B(12) intake and the risk of breast cancer among Mexican women Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2006 15 3 443 8 10.1158/1055-9965.EPI-05-0532 63. Pizot C Boniol M Mullie P Koechlin A Boniol M Boyle P Physical activity, hormone replacement therapy and breast cancer risk: a meta-analysis of prospective studies Eur J Cancer 2016 52 138 54 10.1016/j.ejca.2015.10.063 26687833 64. Zhu H Lei X Feng J Wang Y Oral contraceptive use and risk of breast cancer: a meta-analysis of prospective cohort studies The European Journal of Contraception & Reproductive Health Care: The Official Journal of the European Society of Contraception 2012 17 6 402 14 10.3109/13625187.2012.715357 65. Sillero-Arenas M Delgado-Rodriguez M Rodigues-Canteras R Bueno-Cavanillas A Galvez-Vargas R Menopausal hormone replacement therapy and breast cancer: a meta-analysis Obstet Gynecol 1992 79 2 286 94 1530988 66. Langsenlehner T Renner W Yazdani-Biuki B Langsenlehner U Methylenetetrahydrofolate reductase (MTHFR) and breast cancer risk: a nested-case-control study and a pooled meta-analysis Breast Cancer Res Treat 2008 107 3 459 60 10.1007/s10549-007-9564-1 17453338 67. Macis D Maisonneuve P Johansson H Bonanni B Botteri E Iodice S Methylenetetrahydrofolate reductase (MTHFR) and breast cancer risk: a nested-case-control study and a pooled meta-analysis Breast Cancer Res Treat 2007 106 2 263 71 10.1007/s10549-006-9491-6 17260091 68. Yu L Chen J Association of MTHFR Ala222Val (rs1801133) polymorphism and breast cancer susceptibility: an update meta-analysis based on 51 research studies Diagn Pathol 2012 7 171 10.1186/1746-1596-7-171 23217001 69. Zhang J Qiu LX Wang ZH Wu XH Liu XJ Wang BY MTHFR C677T polymorphism associated with breast cancer susceptibility: a meta-analysis involving 15,260 cases and 20,411 controls Breast Cancer Res Treat 2010 123 2 549 55 10.1007/s10549-010-0783-5 20143151 70. Rai V Methylenetetrahydrofolate reductase A1298C polymorphism and breast cancer risk: a meta-analysis of 33 studies Annals of Medical and Health Sciences Research 2014 4 6 841 51 10.4103/2141-9248.144873 25506474

PMC005xxxxxx/PMC5002181.txt

==== Front Biomed Eng OnlineBiomed Eng OnlineBioMedical Engineering OnLine1475-925XBioMed Central London 2675861511510.1186/s12938-015-0115-4ResearchInfluence of the indirect restoration design on the fracture resistance: a finite element study Mei May Lei mei1123@hku.hk 1Chen Ya Ming yaming_chen@njmu.edu.cn 2Li Hao li.hao@hfut.edu.cn 3Chu Chun Hung +852 2859 0287chchu@hku.hk 11 Faculty of Dentistry, The University of Hong Kong, 34 Hospital Road, Hong Kong SAR, China 2 Institute of Stomatology, Nanjing Medical University, Nanjing, China 3 School of Civil and Hydraulic Engineering, Hefei University of Technology, Hefei, China 8 1 2016 8 1 2016 2016 15 1 320 7 2015 11 12 2015 © Mei et al. 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.Objectives To establish a three-dimensional (3D) finite element (FE) model of a maxillary first premolar and to evaluate the stress generated on the tooth (dentine) and on the indirect composite resin restorations by occlusal forces. Methods An embedded intact maxillary first premolar tooth was sliced serially and scanned digitally parallel to the occlusal surface. The 64 images were assembled in a 3D FE mesh and exported to generate a 3D solid tooth model. Simulated, adhesively cemented indirect mesial-occlusal-distal (MOD) inlays of 2 mm (I1), 3 mm (I2) and 4 mm (I3) in width, and MOD onlays with occlusal cusp coverage of 2 mm (O1) and 3 mm (O2) in depth were created. The peak von Mises stress values in the five tooth models resulting from static vertical and oblique occlusal forces (300 N) were evaluated using Patran FE software. Results The peak stress values generated by vertical occlusal force generated in dentine of I1, I2, I3, O1 and O2 restoration were 67, 32, 29, 38 and 27 MPa, respectively, and those generated by oblique occlusal force were 52, 114, 168, 54 and 55 MPa, respectively. The peak von Mises stress values in I1, I2, I3, O1 and O2 restoration subjected to oblique occlusal loading were 79, 120, 1740, 1400 and 1170 MPa, respectively. Conclusion A 3D FE model of a maxillary first premolar was established. Simulated cemented composite resin onlay markedly reduces occlusal stress in the underlying dentine of large MOD preparation. Oblique occlusal force imparts substantially higher stress to large composite resin inlay than to the adjacent dentine. issue-copyright-statement© The Author(s) 2016 ==== Body Background An increased interest in aesthetic dentistry and concerns about mercury in amalgam alloys have led to an increased demand from patients for metal-free restorations and a relatively simple way to restore their teeth [1]. Though the use of posterior direct composite resins is still the common practice, such restorations may lead to problems such as fracture and microleakage caused by polymerization shrinkage [2]. Those properties of direct resin may lead to postoperative sensitivity, marginal staining and secondary caries [3, 4]. Following secondary curing, indirect composite resin inlays and onlays generally show improvements in their degree of polymer conversion, mechanical properties, wear resistance and marginal microleakage [5, 6]. Correct tooth morphology also may be achieved with less difficulty for large indirect restorations. However, the use of indirect inlays to replace failed mesial-occlusal-distal (MOD) direct placement restorations in premolar teeth is questionable on biomechanical grounds [6, 7]. Concerns arise from the fact that removing and replacing such defective intracoronal restorations can lead to a significant extension of the original preparations, with an increased size, isthmus width and cusp height [8, 9]. Occlusal forces on an inlay produce stresses along the sides and base of the restoration, which may fracture the tooth [10]. Covering the occlusal surface of the intracoronal MOD preparation with metal to form an onlay is found to greatly reduce the potentially damaging effects of the occlusal forces [11]. Tooth fractures occur less readily in intact posterior teeth than following cavity preparation [12, 13]. The width and depth of premolar preparations are important factors affecting tooth strength [14, 15]. The type of cavity preparation and materials also affect the extent of tooth fracture [16, 17]. Early finite element (FE) studies found that increased cavity dimensions were associated with increased stresses in the remaining tooth structure [18–20]. More recent FE studies have evaluated several parameters of occlusal forces, MOD cavity preparation geometry and dental materials used to restore maxillary premolar teeth [20, 21]. Teeth restored with ceramic materials have showed a higher fracture resistance than those restored with composite resin in several FE studies [22, 23]. On the other hand, composite resin restorations like inlay and onlay also have their own fields of application. They are less costly, time-saving, and easy to remove when subsequent restorative treatment is necessary. Since composite resin has a lower fracture resistance than ceramics [22, 23], the optimization of cavity dimension becomes critical in making a successful resin restoration. Therefore, the question of how to control the dimension of the resin restoration was raised in the current study. Few FE studies have investigated the effects of different dimensions of resin inlay and onlay on risk of tooth fracture under different loading forces. The objective of this study was to establish a 3D FE model of a human maxillary first premolar tooth and to use this model to evaluate the effects of (a) MOD inlay restorations of different occlusal isthmus widths and (b) MOD onlay restorations of different occlusal cusp coverage depths on the stresses generated within the tooth model and restorations from occlusal forces. The null hypothesis proposed is that, when subjected to either a vertical or oblique standardized static occlusal force, there are no marked differences present in the resulting peak von Mises stress values for different sizes of indirectly fabricated MOD composite resin inlays and onlays restoring a maxillary premolar tooth model. Methods An intact human maxillary first premolar tooth extracted for orthodontic reasons was selected as the generic model. The restored tooth was at risk for cuspal fracture in clinical practice [24]. The crown dimensions of the selected tooth were similar to average values [25]. The tooth was embedded in clear acrylic resin (ProBase®Cold, Ivoclar vivadent, PR Liechtenstein), then sliced serially at 0.3 mm increments along its long axis from cusp tip to root apex using a thin diamond saw (Isomet®Low Speed Saw, Buehler Ltd, IL, USA) with copious water cooling. Each of the 64 sections obtained was scanned from the occlusal side using a digital flatbed scanner (S2W4300/3300u, BenQ Asia Pacific Corp., Taipei, Taiwan), and the images were stored as BMP files with a resolution of 1200 dots per inch (dpi). The images were assembled in a 3D wire-frame mesh structure by means of a computer software program (Photoshop 7.0, Adobe Systems Inc., San Jose, CA, USA). The 3D curves were exported using proprietary software (Rhino™ 3D, McNeel North America, Seattle, WA, USA) to generate a maxillary first premolar solid tooth model (enamel, dentine, pulp chamber and periodontal ligament) by fitting the horizontal and vertical profiles. The solid model of the maxillary premolar was subsequently exported as XT files (Patran™ 2004, MSC Software Corp, Santa Ana, CA, USA) to establish the 3D FE model. Following proportional adjustments, a virtual reality image with a criss-crossing mesh was established. The 3D solid model was used to create simulated MOD inlay preparations of three different occlusal widths: I1 (2.0 mm), I2 (3.0 mm), I3 (4.0 mm). The occlusal depth of each cavity preparation was 2.5 mm, and the gingival margins of the mesial and distal proximal boxes were placed 1.0 mm above the enamel-cemental junction. The cavity floors were flat, and the cavity walls had a 5° divergence angle. No cavosurface bevels were present (Fig. 1). The 3D solid model was also used to create simulated cusp coverage MOD onlay preparations of two different occlusal cusp coverage depths: O1 (2.0 mm) and O2 (3.0 mm) (Fig. 1). The original dimensions of each cavity preparation before cusp reduction were as for I3. An 80-µm-thick resin-based adhesive cement layer was established between the cavity walls and simulated indirectly fabricated proximal (Class II) composite resin restorations. A 130-µm-thick periodontal ligament was also established [26]. The solid models were imported into Patran 2004, redefined and meshed using a 10-noded tetrahedral element, resulting in 55,032 nodes and 37,010 elements. Sufficiently fine meshing was taken to ensure the mesh convergence of the results. Different material properties were assigned to the elements (Table 1) according to their volume definitions. All materials were assumed to be homogeneous and isotropic. Five different models were generated. To study the effects of different cavity designs, only one material of the restoration was defined to suppress interference from other factors. The boundary conditions were specified to be consistent with physiological conditions. They were set as fixed, vertical and horizontal displacements along root surfaces of the model that were restricted to simulate support from the alveolar bone [21] (Fig. 2).Fig. 1 3D models of the inlay and onlay restoration designs Table 1 Properties attributed to the tooth structure and restorative material Material Young’s modulus (GPa) Poisson’s ratio Enamel1 48 0.30 Dentine1 18 0.31 Periodontium1 0.07 0.40 Composite resin2,a 14 0.30 Resin cement2,b 6 0.30 1Ausiello et al. [37]; Couegnat et al. [21] 2Bisco, Inc., Schaumberg, IL, USA (a Tescera, b C&B Cement) Fig. 2 Illustration of load and boundary conditions: a vertical loading; b oblique loading Two types of static loading were used: (a) vertical loading on the occlusal surface perpendicular to the loading axis (Fig. 2a, b) oblique loading resolved into two oblique components acting perpendicularly to the inner cusp inclines, which lay at approximately 45° to the vertical loading axis (Fig. 2b). The force applied for both types of loading was 300 N [27]. The von Mises stress was used as the study variable, based on the distortion energy theory in engineering, using the formula σv=σ1-σ22+σ2-σ32+σ3-σ122 [28]. The equivalent stress σv is referred to as Von Mises stress. The variables σ1, σ2 and σ3 are referred to as the three principal stresses. Whatever the state of stress, when the ratio of the deformation reaches a critical value related to the property of the material, then the material begins to yield. The peak von Mises stress was evaluated under two loading conditions (oblique and vertical) for each of five restorations (3 inlays and 2 onlays). Subsequently, the peak von Mises stress in term of oblique loading condition was analysed on dentine and restorations, respectively. In addition, the von Mises stress patterns under two loading conditions (vertical and oblique) were also assessed. Results Figure 3 shows the peak von Misses stress value of dentine for the five models subjected to vertical and oblique occlusal loading. I1 has a slightly lower stress value (52 MPa) under oblique loading than under vertical loading (67 MPa). I2 has an obviously higher stress value under oblique loading (114 MPa) than under vertical loading (32 MPa). I3 has a similar pattern to I2. Similar patterns were also observed in both O1 and O2; namely, stress value was slightly higher under oblique loading than under vertical loading. In general, Peak von Mises stress values in the dentine increased for the oblique force but decreased for the vertical force, with increasing inlay cavity width in particular and with increasing onlay cusp coverage depth. Except for the 1.0-mm-wide I1, all of the other models demonstrated higher stress values resulting from oblique loading than from vertical loading. When subjected to an oblique force, peak von Mises stress values in the inlay restorations increased markedly for the 4.0-mm-wide I3 (168 MPa), being 3.25 times that for I1 (52 MPa) and 1.47 times that for I2 (114 MPa). Peak stress values also were high in the two onlay restorations, though they were lower for O2, which had an increased cusp coverage depth of 3.0 mm.Fig. 3 Peak von Mises stress values in the dentine for the five models subjected to vertical and oblique occlusal loadings of 300 N For the five models, when subjected to an oblique force, the peak von Mises stress values of the restoration and dentine are shown in Fig. 4. I1 and I2 show much less stress, not only on the restoration but also on the dentine, whereas the very wide I3 and the two onlays demonstrated much higher stress values in the restorations than in the dentine.Fig. 4 Peak von Mises stress values in the dentine and restorations for the five models subjected to oblique occlusal loadings of 300 N The distribution of the von Mises stress patterns in the five models for the vertical and oblique forces (300 N) are illustrated in Figs. 5 and 6, respectively. The stress distributions of the inlay groups are similar, even with the vertical occlusal loadings and high stress located on both buccal and lingual cusps. In the onlay groups, however, the stress on the cusps has been remarkably reduced. This may reduce the risk of tooth fracture. Moreover, the enamel-dentine junction and bottom of the cavity are also the area with high stress under vertical force. For an oblique force, the highest stresses in the cuspal dentine were associated with the widest inlay preparation.Fig. 5 Von Mises stress patterns for the five models subjected to vertical occlusal loadings of 300 N; arrows indicate the locations of peak von Mises Fig. 6 Von Mises stress patterns for the five models subjected to oblique occlusal loadings of 300 N; arrows indicate the locations of peak von Mises Discussion Indirect restorations are those that fit within the anatomic contours of the clinical crown of a tooth. They can be classified as inlays or onlays, depending on whether they have an occlusal veneer or not [29]. The biomechanical study of a tooth with an indirect restoration presents difficulties in many respects, such as the complexity of the tooth’s structure, restorative materials, boundary condition, cavity shape, dimensions and state of stress [20]. This study investigated the stress distribution of indirect restorations of different designs by means of an FE method. The reality was only reproduced by a mathematical approach using typical material properties, loading and boundary conditions. To keep the simplification effect as low as possible, calculations were not limited to the tooth and the restorations but extended to other structures such as the periodontal ligament and the alveolar bone. Though computerized tomography techniques are available for anatomical and structural data collection [30, 31], obtaining digital images of the premolar tooth in the present study was far simpler. Conventional computer-aided-design (CAD) software is able to represent the curvature of the real object [18, 32]. One of the concerns of the study is that the fracture is due to a static loading, while fatigue phenomena should be considered. In the current study, the conditions and testing of the stress distribution of the materials and tooth have been simplified to instant force. Therefore, the results cannot be extrapolated to the in vivo situation, and caution should be exercised in their interpretation. Based on the results of this study, the null hypothesis was not accepted. In the present study, as the size of the cavity and the inlay increased, the stress caused by oblique occlusal loading on the dentine increased more dramatically than that caused by vertical occlusal loading. This indicated that oblique occlusal forces in particular imparted high stresses to adhesively cemented large MOD composite resin inlays and onlays and, to a much lesser extent, to the functional remaining cusp structures present with the large inlay restoration (~160 MPa). Moreover, when a tooth with an onlay was subjected to oblique forces, the stresses could be effectively transferred to the composite resin onlay restorations rather than to the underlying tooth structure. Under oblique occlusal loading, the peak von Mises stress values in the dentine increased with the width of the inlay cavity width from I1 to I3. When the inlay was replaced by an onlay with the same cavity width, the peak von Mises stress on dentine decreased. Moreover, in large inlays and onlays under oblique occlusal force, the stress in restorations was substantially higher than the stress in dentine. This suggested that the onlay restorations offer protection to the underlying tooth structure. The Young’s modulus of the composite resin used is 14 GPa, which is similar to that of dentine (18 GPa) and less than that of porcelain (70 GPa) or Ni–Cr alloy (204 GPa) [33]. It was also shown that a less rigid restoration could relax the applied stress by means of greater elastic deformation [34]. The less rigid composite resin shows that a greater elastic deformation would result in a lower deformation of the cusps [34]. Therefore, the risk of tooth fracture associated with large cavity preparations could be reduced by the overlaying of the vulnerable cusps with proper restorative materials [35]. Different load conditions result in different concentrations of stress on a tooth. Oblique occlusal force induces a high stress concentration on the cusps and increases the risk of tooth fracture (Fig. 6). In inlay groups, the stress in dentine under oblique force was higher than that under vertical force, and the difference increased with the width of the inlay. The tooth cusps become narrow when the width of the inlay is large [22, 36]. The stress on dentine reduced remarkably when the cavity was restored by O1, which had the same width as I3 (Fig. 3). The high stress concentration on the cusps when they were restored by inlays was also relieved in onlays. Therefore, this study confirmed that preparing the cusps and making an onlay restoration for a large cavity could reduce the high stress generated by occlusal loading. In addition, the difference of stress values and distributions was not significant in the two onlay groups, therefore O1, which had a small depth, was preferred because it was more conservative than O2, which had a large depth. Conclusions In this study, a 3D FE model of a maxillary first premolar tooth was established for evaluation of the stresses generated within the tooth and the indirect restoration from occlusal forces. Oblique occlusal force imparted substantially higher stress to large composite resin inlay than to the adjacent dentine. Onlay preparation reduced occlusal stress and the risk of cusp fracture associated with large cavity preparation. Therefore, an onlay with cusp coverage should be used for restoration when the width of the cavity reaches 4 mm. Abbreviations MODmesial occlusal distal FEfinite element CADcomputer-aided design Authors’ contributions MLM: principal investigator, designed the study, collected data, performed statistical analyses and wrote the manuscript. YMC: designed the study. HL: consulted on mechanical engineering. CHC: supervised the work, evaluated the results and revised the manuscript for publication. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Fuks AB The use of amalgam in pediatric dentistry Pediatr Dent 2002 24 448 455 12412959 2. Giachetti L Scaminaci Russo D Bambi C Grandini R A review of polymerization shrinkage stress: current techniques for posterior direct resin restorations J Contemp Dent Pract 2006 7 79 88 16957794 3. Ferracane JL Resin composite–state of the art Dent Mater 2011 27 29 38 10.1016/j.dental.2010.10.020 21093034 4. Manhart J Kunzelmann KH Chen HY Mechanical properties and wear behavior of light-cured packable composite resins Dent Mater 2000 16 33 40 10.1016/S0109-5641(99)00082-2 11203521 5. Chang HS Kim JW Early hardness and shear bond strength of dual-cure resin cement light cured through resin overlays with different dentin-layer thicknesses Oper Dent 2014 39 398 406 10.2341/13-122-L 24191869 6. Fron Chabouis H Smail Faugeron V Attal JP Clinical efficacy of composite versus ceramic inlays and onlays: a systematic review Dent Mater 2013 29 1209 1218 10.1016/j.dental.2013.09.009 24119328 7. ADA Council on Scientific Affairs Direct and indirect restorative materials J Am Dent Assoc 2003 134 463 472 10.14219/jada.archive.2003.0196 12733780 8. Hernandes NM Catelan A Soares GP Influence of flowable composite and restorative technique on microleakage of class II restorations J Investig Clin Dent 2014 5 283 288 10.1111/jicd.12058 9. Lai GY Zhu LK Li MY An in vitro study on the secondary caries-prevention properties of three restorative materials J Prosthet Dent 2013 110 363 368 10.1016/j.prosdent.2013.06.011 23998624 10. Keown AJ Lee JJ Bush MB Fracture behavior of human molars J Mater Sci Mater Med 2012 23 2847 2856 10.1007/s10856-012-4756-6 22956116 11. van Dijken JW Direct resin composite inlays/onlays: an 11 year follow-up J Dent 2000 28 299 306 10.1016/S0300-5712(00)00010-5 10785294 12. Sun YS Chen YM Smales RJ Fracture resistance and microtensile bond strength of maxillary premolars restored with two resin composite inlay systems Am J Dent 2008 21 97 100 18578176 13. Chu CH Mei ML Cheung C Nalliah RP Restoring proximal caries lesions conservatively with tunnel restorations Clin Cosmet Investig Dent 2013 5 43 50 10.2147/CCIDE.S48567 14. Holberg C Winterhalder P Wichelhaus A Fracture risk of lithium-disilicate ceramic inlays: a finite element analysis Dent Mater 2013 29 1244 1250 10.1016/j.dental.2013.09.012 24119917 15. Xie KX Wang XY Gao XJ Yuan CY Li JX Chu CH Fracture resistance of root filled premolar teeth restored with direct composite resin with or without cusp coverage Int Endod J 2012 45 524 529 10.1111/j.1365-2591.2011.02005.x 22242600 16. Saridag S Sevimay M Pekkan G Fracture resistance of teeth restored with all-ceramic inlays and onlays: an in vitro study Oper Dent 2013 38 626 634 10.2341/12-211-L 23391033 17. Suyama Y Luhrs AK De Munck J Potential smear layer interference with bonding of self-etching adhesives to dentin J Adhes Dent 2013 15 317 324 23593634 18. Ausiello P Apicella A Davidson CL Effect of adhesive layer properties on stress distribution in composite restorations–a 3D finite element analysis Dent Mater 2002 18 295 303 10.1016/S0109-5641(01)00042-2 11992906 19. Ausiello P Rengo S Davidson CL Stress distributions in adhesively cemented ceramic and resin-composite Class II inlay restorations: a 3D-FEA study Dent Mater 2004 20 862 872 10.1016/j.dental.2004.05.001 15451242 20. Lin CL Chang CH Ko CC Multifactorial analysis of an MOD restored human premolar using auto-mesh finite element approach J Oral Rehabil 2001 28 576 585 10.1046/j.1365-2842.2001.00721.x 11422685 21. Couegnat G Fok SL Cooper JE Structural optimization of dental restorations using the principle of adaptive growth Dent Mater 2006 22 3 12 10.1016/j.dental.2005.04.003 16061281 22. Costa A Xavier T Noritomi P Saavedra G Borges A The Influence of elastic modulus of inlay materials on stress distribution and fracture of premolars Oper Dent 2014 39 E160 E170 10.2341/13-092-L 24967990 23. Dejak B Mlotkowski A Three-dimensional finite element analysis of strength and adhesion of composite resin versus ceramic inlays in molars J Prosthet Dent 2008 99 131 140 10.1016/S0022-3913(08)60029-3 18262014 24. Fennis WM Kuijs RH Kreulen CM A survey of cusp fractures in a population of general dental practices Int J Prosthodont 2002 15 559 563 12475162 25. Nelson SJ Wheeler’s dental anatomy 2010 Physiology and Occlusion Saunders 26. Schroeder H The periodontium 1986 Berlin Springer-Verlag 185 27. Hagberg C Electromyography and bite force studies of muscular function and dysfunction in masticatory muscles Swed Dent J Suppl 1986 37 1 64 3465056 28. Berger T Rate-distortion theory Encycl Telecommun 2003 29. Shillingburg HT Hobo S Whitsett LD Jacobi R Brackett SE Fundamentals of fixed prosthodontics 1997 Hanover Park Quintessence Publishing, Quintessence Publishing Co 30. Al-Omiri MK Rayyan MR Abu-Hammad O Stress analysis of endodontically treated teeth restored with post-retained crowns: a finite element analysis study J Am Dent Assoc 2011 142 289 300 10.14219/jada.archive.2011.0168 21357863 31. Castano MC Zapata U Pedroza A Creation of a three-dimensional model of the mandible and the TMJ in vivo by means of the finite element method Int J Comput Dent 2002 5 87 99 12680039 32. Shimizu Y Usui K Araki K Study of finite element modeling from CT images Dent Mater J 2005 24 447 455 10.4012/dmj.24.447 16279738 33. Hatamleh MM Rodrigues FP Silikas N 3D-FE analysis of soft liner-acrylic interfaces under shear loading Dent Mater 2011 27 445 454 10.1016/j.dental.2011.01.003 21354607 34. Abuelenain DA Ajaj RJ El-Bab EI Comparison of stresses generated within the supporting structures of mandibular second molars restored with different crown materials: 3-D Finite Element Analysis (FEA) J Prosthodont 2015 24 484 493 10.1111/jopr.12240 25521774 35. Ausiello P Apicellab A Davidsonc CL 3D-finite element analyses of cusp movements in a human upper premolar, restored with adhesive resin-based composites J Biomech 2001 34 1269 1277 10.1016/S0021-9290(01)00098-7 11522306 36. Fennis WM Kuijs RH Kreulen CM Fatigue resistance of teeth restored with cuspal-coverage composite restorations Int J Prosthodont 2004 17 313 317 15237878 37. Ausiello P Franciosa P Martorelli M Numerical fatigue 3D-FE modeling of indirect composite-restored posterior teeth Dent Mater 2011 27 423 430 10.1016/j.dental.2010.12.001 21227484

PMC005xxxxxx/PMC5002182.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 66210.1186/s12974-016-0662-zResearchDysregulation of the (immuno)proteasome pathway in malformations of cortical development van Scheppingen J. j.g.scheppingen@amc.uva.nl 1Broekaart D. W. M. d.w.broekaart@amc.uva.nl 1Scholl T. theresa.scholl@meduniwien.ac.at 2Zuidberg M. R. J. mark.zuidberg@student.uva.nl 1Anink J. J. j.j.anink@amc.uva.nl 1Spliet W. G. w.g.m.spliet@umcutrecht.nl 5van Rijen P. C. p.v.rijen@umcutrecht.nl 6Czech T. thomas.czech@meduniwien.ac.at 3Hainfellner J. A. johannes.hainfellner@meduniwien.ac.at 4Feucht M. martha.feucht@meduniwien.ac.at 2Mühlebner A. a.muehlebnerfahrngruber@amc.nl 1van Vliet E. A. E.A.vanVliet@uva.nl 1Aronica E. 31-20-5662943e.aronica@amc.uva.nl 1781 Academic Medical Center, Department of (Neuro)Pathology, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands 2 Department of Pediatrics, Medical University Vienna, Vienna, Austria 3 Department of Neurosurgery, Medical University Vienna, Vienna, Austria 4 Department of Pathology, Medical University Vienna, Vienna, Austria 5 Department of Pathology, Rudolf Magnus Institute for Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands 6 Department of Neurosurgery, Rudolf Magnus Institute for Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands 7 Swammerdam Institute for Life Sciences, Center for Neuroscience, University of Amsterdam, Amsterdam, The Netherlands 8 Stichting Epilepsie Instellingen Nederland (SEIN), ᅟ, The Netherlands 26 8 2016 26 8 2016 2016 13 1 2025 4 2016 18 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 The proteasome is a multisubunit enzyme complex involved in protein degradation, which is essential for many cellular processes. During inflammation, the constitutive subunits are replaced by their inducible counterparts, resulting in the formation of the immunoproteasome. Methods We investigated the expression pattern of constitutive (β1, β5) and immunoproteasome (β1i, β5i) subunits using immunohistochemistry in malformations of cortical development (MCD; focal cortical dysplasia (FCD) IIa and b, cortical tubers from patients with tuberous sclerosis complex (TSC), and mild MCD (mMCD)). Glial cells in culture were used to elucidate the mechanisms regulating immunoproteasome subunit expression. Results Increased expression was observed in both FCD II and TSC; β1, β1i, β5, and β5i were detected (within cytosol and nucleus) in dysmorphic neurons, balloon/giant cells, and reactive astrocytes. Glial and neuronal nuclear expression positively correlated with seizure frequency. Positive correlation was also observed between the glial expression of constitutive and immunoproteasome subunits and IL-1β. Accordingly, the proteasome subunit expression was modulated by IL-1β in human astrocytes in vitro. Expression of both constitutive and immunoproteasome subunits in FCD II-derived astroglial cultures was negatively regulated by treatment with the immunomodulatory drug rapamycin (inhibitor of the mammalian target of rapamycin (mTOR) pathway, which is activated in both TSC and FCD II). Conclusions These observations support the dysregulation of the proteasome system in both FCD and TSC and provide new insights on the mechanism of regulation the (immuno)proteasome in astrocytes and the molecular links between inflammation, mTOR activation, and epilepsy. Electronic supplementary material The online version of this article (doi:10.1186/s12974-016-0662-z) contains supplementary material, which is available to authorized users. Keywords ImmunoproteasomeInflammationImmunohistochemistryTuberous sclerosis complexFocal cortical dysplasiaEpilepsyAstrocytesInflammationissue-copyright-statement© The Author(s) 2016 ==== Body Background The proteasome is an evolutionarily conserved multicatalytic proteinase complex representing a major protein degradation system, present in the nucleus and cytoplasm of eukaryotic cells, that regulates diverse biological processes essential for cell survival [1–4]. The proteolytic complex of the proteasome is represented by a catalytic 20S core particle, a barrel-shaped complex consisting of four heptameric rings, composed of non-identical α or β subunits [2, 5, 6]. The catalytic activity is restricted to three of the beta subunits β1, β2, and β5, which have specific cleavage preferences, and are constitutively expressed in most tissues. Under certain conditions associated with the release of immune-modulatory cytokines (i.e., interferon-γ, IFN-γ), a specialized type of proteasome called the immunoproteasome can be generated by the incorporation of the inducible subunits, β1i (PSMB9; LMP2, low molecular weight protein 2), β2i (PSMB10; LMP10, MECL-1, multicatalytic endopeptidase complex-like 1), and β5i (PSMB8; LMP7, low molecular weight protein 7) [7, 8]. Increasing evidence supports a key role of the immunoproteasome in the regulation of immune cell function, including both the adaptive and the innate immune response [9–11]. A deregulation of the immunoproteasome system, with induction of β1i and β5i subunits in the neurons and/or glial cells, has been reported in neurodegenerative diseases ([12–14] for review, see [15, 16]). In particular, recent studies have pointed to the role of immunoproteasome in glial cells, suggesting a novel interaction between immunoproteasome and glia-mediated inflammatory response, resulting in a pro-inflammatory environment [14, 16]. Interestingly, induction of β1i and β5i subunits has been also observed in specimens of patients with pharmaco-resistent mesial temporal lobe epilepsy (MTLE; [17]). Moreover, recent experimental data support a role for the β5i subunit in modulating seizure generation in epileptic tissue, and interestingly, this subunit was not upregulated in rats exposed to pilocarpine but not developing SE and spontaneous seizures [18]. In the present study, we investigated the expression and cellular distribution of both constitutive (β1, β5) and immunoproteasome (β1i, β5i) subunits using immunohistochemistry in a large cohort of patients with malformations of cortical development (MCD; including focal cortical dysplasia (FCD) type II and tuberous sclerosis complex (TSC) cortical tubers), evaluating a possible relationship between changes in the expression of these subunits and the clinical course of epilepsy. To provide better insights into the mechanisms underlying the astroglial regulation of immunoproteasome subunits, we studied their expression in response to IL-1β stimulation in both human fetal astrocytes and FCD-derived cells. Since both FCD II and TSC are associated with constitutive activation of the mammalian target of rapamycin (mTOR) pathway [19, 20], we further evaluated the effect of rapamycin (inhibitor of the mTOR pathway) in FCD II-derived cell cultures. Methods Subjects The cases included in this study were obtained from the archives of the Departments of Neuropathology of the Academic Medical Center (AMC, University of Amsterdam, The Netherlands), the University Medical Center Utrecht (UMCU, The Netherlands), and the Medical University Vienna (MUV, Austria). A total of 23 brain tissue specimens, removed from patients undergoing surgery for intractable epilepsy, were examined. The tissue was obtained and used in accordance with the Declaration of Helsinki and the AMC Research Code provided by the Medical Ethics Committee and approved by the committee of the UMCU Biobank. This study was also approved by the Ethical Committee of the Medical University of Vienna. All cases were reviewed independently by two neuropathologists, and the diagnosis of FCD was confirmed according to the international consensus classification system recently proposed for grading FCD [21]. All patients with cortical tubers fulfilled the diagnostic criteria for TSC [22]. None of the FCD patients fulfilled the diagnostic criteria for TSC. Table 1 summarizes the clinical findings of patients with MCD and epilepsy (6 mild MCD (mMCD), 5 FCD IIa, 6 FCD IIb, 6 TSC tubers: 4 TSC2/2 TSC1; pre-operative seizure frequency/month, mean ± SEM: mMCD 19.8 ± 6.7; FCD II 149 ± 68.7; TSC 114.8 ± 24.2); seizure frequencies were recorded (video-electroencephalographic monitoring) at the time of the preoperative evaluation. One tuber specimen was obtained postmortem (age 32 years; male; TSC2). Hippocampal specimens from patients with Alzheimer’s disease (AD; n = 4; 3 females and 1 male; Braak stages V and VI, age 81.7 ± 2.8) were also examined as positive controls. In addition, normal-appearing control cortex and white matter were obtained at autopsy from six young adult control patients (Table 1), without history of seizures or other neurological diseases. All autopsies were performed within 24 h after death.Table 1 Summary of clinical findings of epilepsy patients and controls Pathology type Number of cases Gender (M/F) Mean age (years/range) Localization Mean duration of epilepsy (years/range) mMCD 6 5/1 21.5 (19–27) 3 fr/3 t 17.3 (11–32) FCD IIa 5 3/2 34.2 (18–45) 4 fr/1 t 22.4 (14–26) FCD IIb 6 2/4 33 (21–45) 4 fr/2 t 24 (15–40) Cortical tubers (TSC) 6 3/3 7.1 (3–16) 4 fr/2 t 5.8 (0.8–13) Controls/autopsy 14 8/6 27.0 (2–48) 6 fr/8 t – FCD focal cortical dysplasia, TSC tuberous sclerosis complex, mMCD mild malformations of cortical development, M male, F female, fr frontal, t temporal Tissue preparation and immunohistochemistry Brain tissue from control and MCD patients was fixed in 10 % buffered formalin and embedded in paraffin. Paraffin-embedded tissue was sectioned at 5 μm, mounted on pre-coated glass slides (Star Frost, Waldemar Knittel GmbH, Braunschweig, Germany), and used for histology and immunohistochemistry. One representative paraffin block per case was sectioned, stained, and assessed. Sections were processed for hematoxylin eosin stainings, as well as for immunohistochemical stainings for a number of neuronal and glial markers and antibodies against the constitutive (β1, β5) and immunoproteasome (β1i, β5i) subunits (Table 2). These antibodies have been extensively tested on human liver and brain tissues [23, 24], including surgical brain specimens from patients with mesial temporal lobe epilepsy revealing bands at the expected molecular weight ([17]; Additional file 1: Figure S1). To detect differences in labeling related to technical variables such as tissue fixation, we also tested the antibodies in specimens of selected regions (temporal cortex/hippocampus) collected at autopsy and immediately fixed in formalin for 24 h (same fixation time used for the surgical specimens); no differences in the immunoreactivity pattern were observed.Table 2 Immunohistochemistry: primary antibodies Antigen Primary antibody Source Dilution Glial fibrillary acidic protein (GFAP) Rabbit polyclonal DAKO, Glostrup, Denmark 1:4000 Neuronal nuclear protein (NeuN) Mouse clone MAB377 Chemicon, Temecula, CA, USA 1:2000 Phospho-S6 ribosomal protein (pS6) Ser235/236; rabbit polyclonal Cell Signaling Technology, Beverly, MA, USA 1:50 Interleukin 1β Goat polyclonal Santa Cruz Bio., Delaware CA, USA 1:70 MHC class I (HLA A, B, and C; HLA-I) Mouse clone HC-10 a 1:200 MHC class II (HLA-DP, DQ, DR; HLA-II) Mouse clone CR3/43 DAKO, Glostrup, Denmark 1:400 Proteasome β1 Mouse monoclonal IgG1 Enzo Life Sciences/Biomol, Farmingdale, NY, USA 1:200 Proteasome β5 Rabbit polyclonal Enzo Life Sciences/Biomol 1:500 Proteasome β1i Mouse monoclonal IgG1 Enzo Life Sciences/Biomol 1:200 Proteasome β5i Mouse monoclonal IgG1 Enzo Life Sciences/Biomol 1:200 MHC major histocompatibility complex aGift from Prof. J. Neefjes, Netherlands Cancer Institute, The Netherlands Single-label immunohistochemistry was performed as previously described [25]. Sections were deparaffinated in xylene, rinsed in ethanol (100, 95, and 70 %) and incubated for 20 min in 0.3 % hydrogen peroxide diluted in methanol. Antigen retrieval was performed using a pressure cooker in 0.1 M citrate buffer pH 6.0 at 120 °C for 10 min. Slides were washed with phosphate-buffered saline (PBS; 0.1 M, pH 7.4) and incubated overnight with the primary antibody in PBS at 4 °C. After washing in PBS, sections were stained with a polymer-based peroxidase immunohistochemistry detection kit (PowerVision Peroxidase System, ImmunoVision, Brisbane, CA, USA). The 3,3′-diaminobenzidine tetrahydrochloride was used as chromogen. Sections were dehydrated in alcohol and xylene and coverslipped. Double-labeling of β1, β1i, β5, or β5i with NeuN (neuronal nuclear protein (NeuN; mouse clone MAB377; Chemicon, Temecula, CA, USA; 1:2000), GFAP (polyclonal rabbit, DAKO, Glostrup, Denmark; 1:4000, or monoclonal mouse, Sigma-Aldrich, St. Louis, MO, USA; 1:4000), HLA-I (mouse clone HC-10, 1:200), or HLA-II (mouse anti-human leukocyte antigen (HLA)-DP, DQ, DR, mouse clone CR3/43; DAKO; 1:400) was performed as previously described [26]). Sections were incubated with BrightVision poly-alkaline phosphatase (AP)-anti-rabbit or anti-mouse (Immunologic, Duiven, The Netherlands) for 30 min at room temperature and washed with PBS. AP activity was visualized with the AP substrate kit III Vector Blue (SK-5300, Vector Laboratories Inc., CA, USA). To remove the first primary antibody, sections were incubated at 121 °C in citrate buffer (10 mM NaCi, pH 6.0) for 10 min. Incubation with the second primary antibody was performed overnight at 4 °C. Sections with primary antibody other than rabbit were incubated with post-antibody blocking from the BrightVision+ system (containing rabbit-α-mouse IgG; Immunologic, Duiven, The Netherlands). AP activity was visualized with the alkaline phosphatase substrate kit I Vector Red (SK-5100; Vector Laboratories Inc., CA, USA). Sections incubated without the primary antibody, with preimmune sera, or with the antibody preincubated with the antigenic peptide (for the polyclonal β5) were essentially blank. Evaluation of histology and immunohistochemistry All labeled tissue sections were evaluated by two independent observers for the presence or absence of various histopathological parameters and specific immunoreactivity (IR) for the different markers used for the diagnosis of mMCD, FCD subtypes, and TSC tubers. We also semi-quantitatively evaluated the IR (nucleus and cytoplasm in glial and neuronal cells) of β1, β1i, β5, and β5i. The intensity of the staining was evaluated using a scale of 0–3 (0: no; 1: weak; 2: moderate; 3: strong staining). All areas of the lesion were examined, and the score represents the predominant cell staining intensity found in each case. The frequency of β1, β1i, β5, or β5i positive cells ((1) rare; (2) sparse; (3) high) was also evaluated to give information about the relative number of positive cells within the lesion. We also evaluated intensity and frequency of pS6 and IL-1β staining. As described in previous studies [25, 27], the product of the intensity and frequency scores was taken to give the overall score (total score; immunoreactivity score (IRS), Table 3). Quantification of signal intensity using ImageJ software was performed for β1i and β5i subunits (Additional file 2: Figure S2).Table 3 Immunoreactivity of β1, β1i, β5, and β5i proteasome subunits in the cortex (Dysmorphic) neurons Glia Balloon/giant cells Cytoplasm Nucleus Cytoplasm Nucleus Cytoplasm Nucleus β1 Control 2 (2–3) 0 0 (0–1) 0 – – mMCD 2▪°~ 1 (1–2)*°~ 0▪°~ 0°~ – – FCDIIa 5 (4–9)* 1 (1–2)*° 2.5 (2–4)* 2.5 (2–4)*° – – FCDIIb 7.5 (3–9)* 6 (3–6)* 4 (3–6)* 5 (4–6)* 3.5 (3–6) 6 (4–9) TSC 6.5 (4–9)* 5 (4–9)* 7.5 (4–9)* 7.5 (4–9)* 6 (3–9) 9 (6–9) β1i Control 0 (0–1) 0 (0–1) 1 (0–1) 0 (0–2) – – mMCD 0▪°~ 0▪°~ 0 (0–1)▪°~ 0▪°~ – – FCDIIa 3 (2–4)* 4 (3–6)*~ 3.5 (3–4)* 3 (3–6)*~ – – FCDIIb 4 (2–4)* 6 (3–9)* 6 (3–9)* 5 (3–6)* 4 (3–4) 7.5 (3–9) TSC 3.5 (2–6)* 6 (4–9)* 5 (3–9)* 5 (2–9)* 4 (3–6) 7.5 (4–9) β5 Control 1 (0–1) 1 (0–1) 0 0 (0–1) – – mMCD 0°~ 6*▪ 0°~ 0▪°~ – – FCDIIa 3.5 (2–4)*~ 7.5 (6–9)* 1.5 (0–2)*~ 2.5 (1–4)*~ – – FCDIIb 6 (4–6)*° 9 (6–9)* 3.5 (3–4)* 6 (4–9)*° 4 (3–6) 9 (6–9) TSC 3.5 (3–4)* 7.5 (6–9)* 2.5 (2–4)* 2.5 (2–4)* 6 (4–9) 9 (6–9) β5i Control 0 0 0 (0–1) 1 (1–2) – – mMCD 0° 0▪~ 0 (0–1)°~ 0°~ – – FCDIIa 0° 1.5 (1–4)* 0 (0–2)° 0 (0–2)~ – – FCDIIb 0 (0–1)° 4 (3–6)° 2.5 (1–3)° 3 (1–6) 0 (0–1)° 6 (4–6)° TSC 6 (4–9)* 0.5 (0–1) 6 (4–9)* 1.5 (1–2)* 9 (6–9) 0.5 (0–1) Immunoreactivity score (IRS) is given as median (minimum-maximum). IRS is defined as intensity score multiplied by frequency score (see “Methods” section). Kruskall-Wallis test followed by Mann-Whitney U test *Different compared to controls; °different compared to TSC; ▪different compared to FCDIIa; ~different compared to FCDIIb, p < 0.05 Cell cultures Primary fetal astrocyte-enriched cell cultures were obtained from human fetal brain tissue (14–19 weeks of gestation) obtained from the HIS-Mouse (human immune system mouse) facility of the AMC, Amsterdam. All materials have been collected from donors from whom a written informed consent for the use of the material for research purposes had been obtained by the Bloemenhove Clinic (Heemstede, The Netherlands); these informed consents are kept together with the medical record of the donor by the clinic. The tissue was obtained in accordance with the Declaration of Helsinki and the AMC Research Code provided by the Medical Ethics Committee of the AMC. Cell isolation was performed as described elsewhere [28–30]. Briefly, after the removal of the blood vessels, the tissue was mechanically minced into smaller fragments and enzymatically digested by incubating at 37 °C for 30 min with 2.5 % trypsin (Sigma-Aldrich; St. Louis, MO, USA). The tissue was washed with incubation medium containing Dulbecco’s modified Eagle’s medium (DMEM)/HAM F10 (1:1) medium (Gibco, Life Technologies, Grand Island, New York, USA), supplemented with 50 units/ml penicillin, 50 μg/ml streptomycin, and 10 % fetal calf serum (FCS; Gibco, Life Technologies, Grand Island, New York, USA) and triturated by passing through a 70 μm mesh filter. Cell suspension was incubated at 37 °C, 5 % CO2 for 48 h to let glial cells adhere to the culture flask before it was washed with PBS to remove excess of myelin and cell debris. Cultures were subsequently refreshed twice a week. Cultures reached confluence after 2–3 weeks. Primary FCD astrocyte cultures were derived from a surgical human brain specimen obtained from a patient with FCD type IIA (age at surgery, 16 years; female; location, frontal; seizure frequency, thrice per week; duration of epilepsy, 11 years) undergoing epilepsy surgery at the Department of Pediatrics/Neurosurgery of the Medical University Vienna (Vienna, Austria). FCD astrocyte cultures were established in the same manner as described above for fetal cultures. Secondary astrocyte cultures for experimental manipulation were established by trypsinizing confluent cultures and sub-plating onto poly-l-lysine (PLL; 15 μg/ml, Sigma-Aldrich)-precoated 12- and 24-well plates (Costar, Cambridge, MA, USA; 5 × 104 cells/well in a 12-well plate for RNA isolation and PCR; 2.5 × 104 cells/well for immunocytochemistry). In the present study, astrocytes were used for analyses at passages 2–4. Cell cultures were stimulated with human recombinant (r)IL-1 β (PeproTech, Rocky Hill, NJ, USA; 10 ng/ml) or in some experiments with lipopolysaccharide (LPS; 100 ng/ml; Sigma-Aldrich, St. Louis, USA) for 24 h. Treatment of FCD-derived astrocytes with rapamycin (100 nM) was started 24 h before and continued during IL-1β stimulation. Cells were harvested 24 h after stimulation. Viability of human cell cultures was not influenced by the performed treatments (Additional file 3: Figure S3). For immunofluorescent staining of cell cultures, sections were incubated with the primary antibodies for β1, β1i, β5, or β5i for 1 h at RT, followed by 2 h of incubation at RT with Alexa Fluor® 568-conjugated anti-rabbit or Alexa Fluor® 488-conjugated anti-mouse IgG (1:200, Molecular Probes, The Netherlands) together with Alexa Fluor® 488 or 594 Phalloidin (1:200, Molecular Probes, Plaats, The Netherlands) for counterstaining actin filaments. Sections were mounted using VECTASHIELD with DAPI (Vector Laboratories Inc., Burlingame, CA, USA). Fluorescent microscopy was performed using a Leica Confocal Microscope TSC SP-8X (Leica, Son, the Netherlands) at ×40 magnification (bidirectional X, speed 600 Hz, pinhole 1.00 AU). RNA isolation and real-time quantitative PCR analysis For RNA isolation, cell culture material was homogenized in Qiazol Lysis Reagent (Qiagen Benelux, Venlo, The Netherlands). Total RNA was isolated using the miRNeasy Mini kit (Qiagen Benelux, Venlo, The Netherlands) according to the manufacturer’s instructions. The concentration and purity of RNA were determined at 260/280 nm using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA). To evaluate β1, β1i, β5 or β5i, and IFNγ mRNA expression, 200 ng of cell-culture-derived total RNA was reverse-transcribed into cDNA using oligo dT primers. PCRs were run on a Roche LightCycler 480 thermocycler (Roche Applied Science, Basel, Switzerland) using the following primers: β1 (forward: accagctcggtttccaca, reverse: cccggtatcggtaacacatc); β5 (forward: gagtctcagtgatggtctgagc, reverse: actccatggcggaacttg); β1i (forward: accaaccggggacttacc, reverse: tcaaacactcggttcaccac); β5i (forward: ccctacccacccctgttt; reverse: cacccagggactggaaga); and IFN-γ (forward: gcaagatcccatgggttgtgt; reverse: ctggctcagattgcaggcata). Quantification of data was performed using the computer program LinRegPCR in which linear regression on the log (fluorescence) per cycle number data is applied to determine the amplification efficiency per sample [31, 32]. The starting concentration of each specific product was divided by the geometric mean of the starting concentration of the reference genes (EF1α and C1orf43), and this ratio was compared between groups. Statistical analysis Statistical analyses were performed with GraphPad Prism software (Graphpad Software Inc., La Jolla, CA, USA). To assess differences in immunoreactivity score between multiple groups, non-parametric Kruskal-Wallis followed with Mann-Whitney U test was used. Correlations were assessed using Spearman’s (rho) rank correlation test. For cell culture data, Mann-Whitney U test was used to asses differences between different conditions. P<0.05 was assumed to indicate a significant difference. Results Case material and histological features The clinical features of the cases included in this study are summarized in Table 1. All operated patients had a history of chronic pharmacoresistant epilepsy. In this study, we included patients with mild degree of cortical dysplasia (mMCDs; [33]). Age at surgery, seizure duration, and seizure frequency were not statistically different between patients with FCD II and mMCD in this cohort, as well as between the FCD IIa and FCD IIb cases included in our cohort. Accordingly to the international consensus classification system of FCD [21], FCD II represents isolated focal lesions with architectural and dysmorphic abnormalities (FCD IIa with dysmorphic neurons only; FCD IIb with dysmorphic neurons and balloon cells; Figs. 1, 2, 3, and 4c–e). TSC patients were younger compared to mMCD and FCD patients. All six TSC tubers displayed similar histopathological features, including loss of lamination, astrogliosis, dysmorphic neurons, and giant cells with pale eosinophilic cytoplasm ([34]; Figs. 1, 2, 3, and 4f, g).Fig. 1 β1 proteasome subunit immunoreactivity in control, focal cortical dysplasia (FCD) type IIa, FCD type IIb, and tuberous sclerosis complex (TSC). Panels a, b (control) show weak immunoreactivity (IR) in both the cortex (a insert: neuron with weak expression of β1 subunit) and white matter (b not-detectable glial expression). Panel c (FCD IIa) shows positive dysmorphic neurons (arrows; insert: high magnification) and glial cells (arrow heads). Panels d–e (FCD IIb) show several β1 positive cells within the cortex (d) and white matter (e), including the dysmorphic neurons (arrows in d), glial cells (arrow heads and insert in d) and balloon cells (arrows in e; prominent nuclear expression; insert: high magnification). Panels f–g (TSC tuber): β1 subunit expression is observed within the tuber in dysmorphic neurons (arrows in f; insert a: co-localization with the neuronal marker NeuN; insert b, co-localization with GFAP), glial cells (arrow heads in f) and in giant cells (asterisk (g); insert: co-localization with GFAP). Additional examples of the different cell types at higher magnification in separate specimens. The inserts within panels show imagines of the different cells types at higher magnification in separate specimens. Scale bar in G: A–C, E–F: 80 μm; D–G: 40 μm Fig. 2 β1i proteasome subunit immunoreactivity in control, focal cortical dysplasia (FCD) type IIa, FCD type IIb, and tuberous sclerosis complex (TSC). a, b Control cortex (a) and with matter (b) with weak β1i expression (insert in a: negative neuron, high magnification). Panel c (FCD IIa) shows strong expression within the dysplastic region with positive dysmorphic neurons (arrows; inserts: a nuclear expression; b cytoplasmic expression) and glial cells (insert in c). Panels d–e (FCD IIb) show several β1i-positive cells within the cortex (d) and white matter (e), including dysmorphic neurons (arrow in d), glial cells (arrow heads d and e), and balloon cells (arrows in e; nuclear and cytoplasmic expression; insert: co-localization with the pS6). Panels f, g (TSC tuber): β1i subunit expression is observed within the tuber in dysmorphic neurons (f arrows, nuclear expression; arrow head, cytoplasmic expression; insert a in f: co-localization with the neuronal marker NeuN; insert b in f: co-localization with the pS6; insert c: co-localization with HLA-II), glial cells (insert b in g), and in giant cells (arrows in g; insert b: co-localization with GFAP; insert c: co-localization with HLA-I). The inserts within the panels show images of the different cell types at higher magnification in separate specimens. Scale bar in g: a–c: 80 μm; d–g: 40 μm Fig. 3 β5 proteasome subunit immunoreactivity in control, focal cortical dysplasia (FCD) type IIa, FCD type IIb, and tuberous sclerosis complex (TSC). Panels a and b: (control) show the weak immunoreactivity in both the cortex (a insert: neuron) and white matter (b not detectable glial expression). Panel c (FCD IIa) shows positive dysmorphic neurons (arrows; insert: high magnification, nuclear expression). Panels d and e (FCD IIb) show several β5 positive cells within the cortex (d) and white matter (e), including dysmorphic neurons (arrows in d and insert, nuclear and cytoplasmic expression), glial cells (arrow heads and insert in d), and balloon cells (asterisk in e; insert: high magnification, with prominent nuclear expression). Panels f and g (TSC-tuber): β5 subunit expression is observed within the tuber in dysmorphic neurons (arrows in f; insert b in panel g: co-localization with the neuronal marker NeuN; insert c in panel g, co-localization with GFAP) and in giant cells (arrow in g and insert a; insert b: co-localization with NeuN; insert c: expression in glial cells, co-localization with GFAP). The inserts within panels show images of the different cell types at higher magnification in separate specimens. Scale bar in g: a–c, g: 80 μm; d–f: 40 μm Fig. 4 β5i proteasome subunit immunoreactivity in control, focal cortical dysplasia (FCD) type IIa, FCD type IIb, and tuberous sclerosis complex (TSC). Panels a and b: control cortex (a) and with white matter (b) with weak β5i expression. Panel c (FCD IIa) shows expression within the dysplastic region with positive dysmorphic neurons (arrows and insert, nuclear expression) and glial cells (arrow heads). Panels d and e (FCD IIb) show several β5i-positive cells within the cortex (d) and white matter (e), including dysmorphic neurons (arrow in d), glial cells (arrow heads d and e; insert in d), and balloon cells (arrows in e; insert: co-localization with pS6). Panels f, and g (TSC-tuber): strong β5i subunit expression is observed within the tuber in dysmorphic neurons (f arrows and inserts a and b; b co-localization with the neuronal marker NeuN; c co-localization with the pS6; d co-localization with HLA-II), glial cells (arrows in g; insert in g, co-localization with GFAP), and in giant cells (arrows in h; insert: co-localization with HLA-I). The inserts within the panels show images of the different cell types at higher magnification in separate specimens. Scale bar in h: a–c 80 μm; d–h 40 μm Proteasome subunit expression in FCD and cortical tubers Expression of β1, β1i, β5, and β5i was observed in FCD, cortical tubers, and mMCD specimens (Figs. 1, 2, 3, and 4; Additional file 2: Figure S2 and Additional file 4: Figure S4). We observed differences in the expression level as well as in the cell-specific and subcellular distribution of the different subunits (Table 3). Constitutive proteasome catalytic subunit β1 and β5 Moderate expression of β1 and β5 subunits was observed in human control cortical specimens (Figs. 1a, b and 3a, b; Table 3). Nuclear neuronal expression was detected for β5 in MCD specimens (Table 3; Additional file 4: Figure S4E); whereas only cytoplasm expression was detected in specimens from patients with Alzheimer’s disease for both subunits (Additional file 4: Figure S4B, F). Increased expression of both constitutive subunits was observed in FCD and TSC specimens (Figs. 1c–g and 3c–g; Table 2). In the large majority of FCD and TSC cases, β1 IR was detected in the cytoplasm and nucleus of neuronal and glial cells (Fig. 1c–g; Table 3). β1 was also detected in the balloon (FCD IIb; Fig. 1e) and giant cells (TSC; Fig. 1g). FCD and TSC specimens displayed also strong β5 IR with prominent nuclear expression in both neuronal and glial cells, as well as in the balloon (FCD IIb) and giant cells (TSC; Fig. 3c–g; Table 3). A similar pattern was detected in the postmortem TSC case; double-labeling experiments confirmed the co-localization with astroglial and neuronal markers within the dysplastic area for both subunits in FCD and TSC specimens (Figs. 1f–g and 3g). Immunoproteasome subunits β1i and β5i In the large majority of control (Figs. 2a, b and 4a, b) and mMCD (Additional file 4: Figure S4C, G) specimens, the immunoproteasome subunits β1i and β5i were under the detection levels in both neuronal and glial cells (Table 3). β1i and β5i were consistently high in FCD and TSC specimens (Table 3; Additional file 2: Figure S2) with strong cytoplasmic and nuclear IR in neuronal and glial cells, in both surgical postmortem TSC specimens (whereas only cytoplasmic expression was detected in glial cells in specimens from patients with Alzheimer’s disease for both β1i and β5i subunits; Additional file 4: Figure S4D, H). A similar pattern with strong expression in FCD and TSC specimens was observed using in situ hybridization (Additional file 5: Figure S5). Double-labeling experiments confirmed the co-localization with astroglial and neuronal markers, as well as with major histocompatibility complex (MHC) class I (HLA-I; in few balloon/giant cells and in dysmorphic neurons) within the dysplastic area for both subunits in FCD and TSC specimens (Figs. 2 and 4f, g). In regions with prominent activation of microglia, IR for both β1i and β5i was also observed in cells of the microglia/macrophage lineage (HLA-II; Figs. 2 and 4). The balloon (FCD IIb; Fig. 2e) and giant cells (TSC; Fig. 4g) displayed β1i and β5i IR as well (Table 2). Co-localization was observed for both β1i and β5i with pS6 (Figs. 2 and 4). β1i and β5i expression in neurons was positively associated with pS6 expression within our MCD cohort (β1i cytoplasm, r = 0.5905, p = 0.030; β1i nucleus, r = 0.6244, p = 0.0014; β5i cytoplasm, r = 0.4510, p = 0.0065). A positive correlation was detected between β1i and β5i expression in neurons and glial cells and IL-1β IRS within the dysplastic region (neuronal β1i cytoplasm, r = 0.4287, p = 0.0413; neuronal β1i nucleus, r = 0.5090, p = 0.0131; glia β1i cytoplasm, r = 0.5298, p = 0.0093; glia β1i nucleus, r = 0.6091, p = 0.0003; neuronal β5i cytoplasm, r = 0.7322, p = <0.001; glia β5i cytoplasm, r = 0.7005, p = <0.001; glia β5i nucleus, r = 0.4210, p = 0.0455). Immunoproteasome subunit expression and clinical features We found no statistically significant association between the IRS of β1, β1i, β5 or β5i, and clinical features, such gender, age at surgery, location of the lesion, or duration of epilepsy. However, a positive correlation was observed between nuclear glial and neuronal subunit expression and the pre-operative seizure frequency (β1 and β1i neuron τ = 0.639 and τ = 0.633, p < 0.001; β1 and β1i glia τ = 0.479, p = 0.005 and τ = 0.65, p < 0.001; β1 and β1i neuron τ = 0.550, p = 0.004 and τ = 0.417, p = 0.016; β1 and β1i glia τ = 0.570, p = 0.001; τ = 0.586, p = 0.001). Regulation of immunoproteasome subunit expression in human glial cells in culture Since IL-1β is known to be strongly upregulated in FCD and TSC human brain specimens [27, 34, 35] and to play a key pathogenic role in human epilepsy (for review, see [20, 36]; we also investigated whether this inflammatory cytokine could play a role in the regulation of the expression and cellular localization of immunoproteasome subunits. qPCR analysis of astrocyte-enriched human fetal cell cultures demonstrated that exposure to IL-1β did not modify the expression of the constitutive subunits (Fig. 5a, d) but did consistently increase the expression of both immunoproteasome subunits β1i and β5i (Fig. 5b, e), increasing the β1/ β1i and β5/β5i ratios (Fig. 5c, f). Treatment with LPS, also a potent inducer of the immune response, gave comparable results (Fig. 5). Exposure to IL-1β and LPS did not significantly affect the expression of IFNγ in these cultures. Immunohistochemistry showed a translocation of the β1i and β5i subunits, shifting from cytoplasmic to perinuclear-nuclear expression following IL-1β treatment (Fig. 6d, h).Fig. 5 Effects of IL-1β and LPS stimulation on immunoproteasome subunit expression in cell culture. Quantitative real-time PCR of proteasome expression in human fetal astrocytes after the 24 h exposure to IL-1β (10 ng/ml) or LPS (100 ng/ml). a, b, d, e Stimulation with IL-1β or LPS increased expression of the β1i (b) and the β5i (e) subunits compared to control, but not of the constitutive β1 (a) and β5 (d) subunits. c, f Stimulation with either IL-1β or LPS increased the β1i/β1 (c) and the β5i/β5 (f) ratios. Data are expressed relative to the levels observed in untreated cells and are mean ± SEM (n = 5). **p < 0.01, ***p < 0.001 compared to control, Mann-Whitney U test Fig. 6 Effects of IL-1β stimulation on proteasome subunit expression in astrocytes in cell culture. Expression of β1 (a and b; green), β1i (c and d; green), β5 (e and f; red), and β5i (g and h; green) in unstimulated human fetal astrocytes (left panels) and in astrocytes after exposure to IL-1β (24 h; 10 ng/ml, right panels); increased expression of all subunits was observed. A translocation of particularly the β1i and β5i subunits, shifting from cytoplasmic to perinuclear-nuclear expression following IL-1β treatment was observed. Cells were counterstained with phalloidin (actin filaments; red in a–d and g–h, green in e and f) and diamidino-2-phenylindole, DAPI (nuclei; blue). Scale bar in a: 15 μm Effects of rapamycin on proteasome subunit expression in FCD II-derived astrocytes Since both FCD II and TSC are associated with constitutive activation of the mTOR pathway [19, 20], we investigated whether the canonical and allosteric mTOR kinase rapamycin modulates the immunoproteasome in cell cultures derived from FCD II specimens. Western blot analysis confirmed that 100 nM rapamycin reduced the phosphorylation of S6 (pS6) in human astrocytes, an indicator of mTOR activation (Additional file 3: Figure S3). Pretreatment with 100 nM rapamycin reduced the mRNA expression of β1 and β1i subunits under both unstimulated and stimulated conditions (Fig. 7a, b). β5 and β5i expression was reduced by rapamycin in the presence of IL-1β (Fig. 7c, d). Immunocytochemical analysis of these FCD cells confirmed the negative modulation of proteasome subunits by rapamycin (Fig. 8).Fig. 7 Effects of rapamycin on proteasome subunit expression in astrocytes derived from FCD type II. Quantitative real-time PCR of proteasome expression in human FCD cells after 48 h treatment with 100 nM rapamycin, under basal and stimulated (IL-1β 10 ng/ml) conditions. a, b Treatment with rapamycin decreased the expression of β1 (a) and β1i (b) subunit, both in the basal and under stimulated conditions. c, d Treatment with rapamycin decreased the expression of β5 (c) and β5i (d) subunit under stimulated, but not in basal conditions. Data are expressed relative to the levels observed in untreated cells and are mean ± SEM (n = 5). *p < 0.05, **p < 0.01 compared to control, Mann-Whitney U test Fig. 8 Effects of rapamycin treatment on proteasome subunit expression in FCD type II derived cells. Expression of β1 (a and b; green), β1i (c and d; green), β5 (e and f; red), and β5i (g and h; green) in untreated FCD cells (left panels) and in FCD cells after 48 h of treatment with 100 nM rapamycin (right panels). Expression of all subunits was decreased after treatment with rapamycin. Cells were counterstained with phalloidin (actin filaments; red in a–d and g–h, green in e and f) and diamidino-2-phenylindole, DAPI (nuclei; blue). Scale bar in a: 15 μm Discussion The present study reports in detail the expression pattern and cellular localization of the constitutive and immunoproteasome subunits in FCD II and TSC cortical tubers and mMCD. The cell-specific distribution of proteasome subunits in relation with the epileptogenicity of these developmental lesions as well as their regulation in human astrocytes is discussed in the following paragraphs. Proteasome subunits expression in malformations of cortical development: prominent expression in FCD II and TSC Our data show prominent expression of both constitutive and immunoproteasome subunits in MCD, such as FCD and TSC, associated with the mTOR pathway. In all the FCD II and TSC specimens examined, the IR for β1, β1, β5, and β5i was increased within the dysplastic regions where prominent gliosis and the presence of dysmorphic neurons and balloon or giant cells (in FCD IIb and TSC, respectively) was observed. Constitutive and particularly immunoproteasome subunits displayed increased expression compared to control but also compared to mMCD specimens from patients with chronic epilepsy. These results indicate that increased expression of proteasome subunits is not simply an effect of seizure activity; moreover, the duration of epilepsy in mMCD cases did not differ from FCDs and was even longer compared to TSC cases. However, a positive correlation was observed between nuclear glial and neuronal proteasome subunit expression and the pre-operative seizure frequency. We acknowledge limitations to the interpretation of these results; therefore, an evaluation of the real biological contribution of proteasome subunit expression to seizure generation and frequency deserves further investigation in experimental models. Several proteasome subunits show nuclear localization signaling [37], and previous studies in the human brain indicate that proteasomes are expressed in both cytoplasm and nuclei of different cell types, including glial and neuronal cells [24, 38]. Immunoproteasome expression restricted to nuclei of astrocytes has been reported in the brain after an infection with lymphocytic choriomeningitis virus, suggesting involvement of the nuclear envelope in the compartmentalization of immature proteasome precursors [39]. Whether the nuclear proteasome subunits represent (as suggested by Kremer et al. [39]) immature proteasome precursors or are proteolytically active remains still to be investigated. The nuclear proteasome subunit accumulation may reflect the induction of the proteasome system under conditions associated with cell injury and inflammation with the possibility of nucleo-cytoplasmic transfer in cells, as glial cells, undergoing cell division or during apoptosis [37]. However, the β1i subunit in the nuclear-enriched fraction has also been detected in its catalytically active form [40], and several studies indicate a possible functional role of the immunoproteasome in transcriptional regulation [41–43]. The expression pattern, either nuclear or cytoplasmic proteasome expression, can be influenced by the type and duration of fixation [37]. However, similar pattern was observed in surgical and postmortem TSC brain tissue. One of the major regulatory factors of immunoproteasome induction is inflammation [43, 44]. Several studies confirmed the occurrence of complex inflammatory changes, involving both glial and neuronal cells, and the activation of the IL-1β pathway, particularly in FCD II and TSC [20, 34, 35, 45–48]. Thus, the pro-inflammatory environment may contribute to the activation of the proteasome system, particularly to the induction and expression of the immunoproteasome subunits. Accordingly, our in vitro studies in human astrocytes and FCD cultures indicate that IL-1β treatment increases the induction of, in particular, the immunoproteasome subunits β1i and β5i, with the increase of their perinuclear-nuclear localization. This observation supports the role of astrocytes as targets of regulation of the immunoproteasome under various conditions associated with the activation of the IL-1β pathway [16] and indicates that pro-inflammatory cytokines, other than IFNγ, may regulate immunoproteasome expression. Activation of inflammatory pathways, including IL-1β, may also play a role in the regulation of immunoproteasome expression in other cell types, such as neurons. Accordingly, we found a positive correlation between the expression of immunoproteasome subunits in both glial and neuronal cells and the expression of IL-1β within the dysplastic area in FCD II and in TSC specimens. Moreover, increasing evidence supports the role of the immunoproteasome in the activation of the NF-kB pathway, modulation of pro-inflammatory cytokine production, and oxidative stress response [9, 43, 49–52]. Induction of the β5i subunit has also been shown in vivo following activation of the Toll-like receptor 4 (TLR4)-mediated NF-kB signaling pathway by LPS [53]. Thus, we may speculate about the existence of a reinforcing feedback loop between NF-kB pathway and the immunoproteasome system, which may play a crucial role in perpetuating the pro-epileptogenic inflammatory response in epilepsy. Interestingly, Mishto et al. [18] provide additional experimental evidence of the regulation of β5i subunit by TLR4 signaling in epileptogenic tissue. The immunoproteasome is known to improve MHC class I (MHC-I) antigen presentation and has been suggested to have a central function at the interface between the innate and adaptive immune system (reviewed in [11]). Interestingly, FCD II and TSC specimens are characterized by prominent activation of both innate and adaptive immune responses (for review, see [20, 36]). Moreover, recent studies provide evidence of an upregulation of MHC-I, involving also balloons/giant cells and neurons, in both FCD II and TSC specimens [54]. FCD II and TSC cases are characterized by architectural or cellular changes associated with mTOR pathway activation [20, 21]. The innate and adaptive immune responses have also been shown to be influenced by the mTOR pathway [55–57]. Moreover, the mTOR complex 1 (mTORC1) has been identified as a key regulator of autophagy [58, 59], a pathway which is defective in FCD II and TSC [60]. Increasing evidence indicates a strong relationship with tight coordination between the autophagy and the proteasome system [61]. Thus, we cannot exclude a role of mTOR in the regulation of the proteasome system, including immunoproteasome subunit expression. Accordingly, we observed a positive correlation between immunoproteasome subunit expression in neurons and pS6 expression, indicating the activation of the mTOR signal transduction pathway. The relationship between mTOR and proteasome system is also supported by the in vitro experiments showing that inhibition of the mTOR pathway by the potent allosteric mTORC1 inhibitor rapamycin was able to reduce the level of expression of inducible proteasome subunits in FCD-derived cells. This is in agreement with a recent study showing reduction of the immunoproteasome by rapamycin in H9c2 cells as well as in mouse heart in vivo [62]. Evaluation of the possible effect of rapamycin on the expression of the brain immunoproteasome in vivo deserves further studies and is presently under investigation [63]. Immunoproteasome inhibition as therapeutic strategy? An example of the possible use of inhibition of the immunoproteasome as therapeutic strategy in epilepsy is represented by the study of Mishto and colleagues [18] in which specific inhibition of the β5i subunit by ONX-0914 [64] resulted in prevention, or significant delay, of 4-aminopyridine-induced seizure-like events in acute rat hippocampal/entorhinal cortex slices, particularly in slices of epileptic rats. Clinically approved proteasome inhibitors targeting the catalytic activity of both the constitutive proteasome and the immunoproteasome have been already used in hematological malignancies [65–67]. New-generation small molecules specifically targeting the immunoproteasome are under clinical development and have been already evaluated in a large variety of animal models of autoimmune diseases and proposed as novel therapeutic approaches for patient with multiple sclerosis, as well as in neurodegenerative diseases (for reviews, see [16, 68, 69]). However, recently alternative functions for the immunoproteasome have also been considered, suggesting that the induction of the immunoproteasome may also play a role in neuronal protection and repair after injury, contributing to the preservation of cell viability upon cytokine-induced oxidative stress [49, 70, 71], which is known to be increased within the TSC tubers [72]. In particular, evidence has been provided that the immunoproteasome plays a role in the clearance of damaged proteins accumulating upon inflammation or oxidative stress (for review, see [49]), which are also detected in TSC and FCD [73]. Accordingly, the formation of aggresome-like-induced structures and increased sensitivity to apoptosis has been reported in immunoproteasome deficiency in cells and in a murine inflammation model [49, 71]. Additional studies support alternative physiological function of the immunoproteasome subunits, including also cell proliferation, cell signaling, and synaptic remodeling (for review, see [49, 74, 75]). Thus, an effective therapeutic intervention based on the immunoproteasome has to take into consideration the preservation of the potential beneficial functions of its activation, particularly during brain development. Conclusions One important question is whether the activation of the immunoproteasome system in the brain tissue may per se be responsible for an increased susceptibility to seizure activity observed in FCD and TSC. As discussed above, experimental studies in the hippocampal/entorhinal cortex slices suggest that the pharmacological inhibition of the β5i subunit may modulate seizure activity. Whether these findings can be translated to other experimental models, including models of FCD and TSC, deserves further investigation. To conclude, our observations support the occurrence of a prominent deregulation of the proteasome system in MCD. In particular, the induction of immunoproteasome subunits in both glial and neuronal cells appears to be a feature of FCD II and TSC and may represent an important accompanying feature of the immune response in these developmental lesions. Therefore, understanding the role of the immunoproteasome in epilepsy-associated pathologies may have great importance in view of the development of new therapeutic strategies. Additional files Additional file 1: Figure S1. Representative immunoblot analysis of total homogenates from (n = 3) surgical hippocampal specimens; β-subunits (β1, ~25 kDa; β1i, ~22 kDa; β5, ~25 kDa; β5i, ~25 kDa; β-actin ~42 kDa). (JPG 134 kb) Additional file 2: Figure S2. β1i and β5i intensity signal in control, mMCD, FCDII, and TSC. FCD: focal cortical dysplasia; TSC: tuberous sclerosis complex; mMCD: mild malformations of cortical development. (PPTX 61 kb) Additional file 3: Figure S3. Effect of the different treatments on fetal astrocyte cell cultures. A: scatterplots of eFluor viability dye staining as analyzed by flow cytometry after different treatments. B: Quantification of viable cells based on eFluor viability staining. Neither treatment with IL-1β nor rapamycin negatively influenced viability of cell cultures. C: Western blot analysis showed effective reduction of phosphorylated S6 after 24 h of 100 nM rapamycin treatment. FSC: forward scatter. (JPG 1420 kb) Additional file 4: Figure S4. Proteasome subunit immunoreactivity (β1, β1i β5, and β5i) in mild MCD (mMCD) and in Alzheimer’ s disease (Alz). Panels A, C, E, and G: mMCD. A: low β1 expression (insert: high magnification of a neuron, with weak nuclear expression). C: low β1i expression (insert: high magnification of a neuron). E: nuclear expression of β5 (arrows; neuron in insert). G: low β5i expression (neuron in insert). Panels B, D, F, and H (Alz; hippocampus). B: β1 expression in neurons (CA1; arrows, cytoplasmic expression) and around amyloid plaques (arrow-heads); D: β1i expression in glial cells (arrows, cytoplasmic expression). F: low β5 expression in neuronal cells (arrows). H: β5i expression in glial cells (arrows, cytoplasmic expression). Scale bar in B: A, C, F, and G: 80 μm; B, D, and H: 40 μm. (JPG 2799 kb) Additional file 5: Figure S5. In situ hybridization of β1i and β5i, proteasome subunit immunoreactivity in control, focal cortical dysplasia (FCD) type IIb, and tuberous sclerosis complex (TSC). Panels A–D: control cortex (A–C) and with matter (B–D); β1i (A–B) and β5i (C–D). Panels E–F (FCD IIb) and panels C–G (TSC) shows strong expression within the dysplastic region with several positive dysmorphic neurons (arrows and inserts in E (a) and F), giant cells (inserts in G and H (a)), and glial cells (inserts (b) in E and H). Scale bar in H: A–H: 80 μm. (JPG 3206 kb) Additional file 6: Supplementary methods: flow cytometric analysis, in situ hybridization, Western blot analysis and image quantification. (DOC 46 kb) Abbreviations FCDFocal cortical dysplasia IRSImmunoreactivity score MCDMalformations of cortical development MTLEMesial temporal lobe epilepsy PBSPhosphate-buffered saline TSCTuberous sclerosis complex Acknowledgements We acknowledge the Bloemenhove Clinic (Heemstede, The Netherlands). We are grateful to A.M. Iyer and K.W. Geijtenbeek for technical assistance. Funding This work was supported by the European Union’s Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 602391 (project acronym EPISTOP; JvS, MF, TS, EA) and grant agreement no. 602102 (project acronym EPITARGET; EvV, EA); the Austrian Science Fund (FWF): project no. J3499 (AM), and the TSC Research Award 2015 from the German Tuberous Sclerosis Foundation (TS). Availability of data and materials All data generated or analyzed during this study are included in this published article (and in Additional file 1, 2, 3, 4, 5 and 6). Authors’ contributions JS, DB, MZ, and JJA carried out the immunohistochemical analyses. JS performed the qPCR analysis. Western blot was performed by JS and DB and in situ hybridization by JJA. Analysis of the data was performed by JS, DB, and MZ. EV and AM contributed to the data analysis and interpretation. Cell culture experiments were performed by JS and TS. WS, PR, TC, JAA, MF, TS, AM, and EA helped with the selection and collection of the brain tissues. AE conceived the study and participated in its design and coordination together with EV and AM. JS, DB, AM, and EV helped EA in drafting and preparing the manuscript. All authors read, revised, 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 tissue was obtained and used in accordance with the Declaration of Helsinki and the AMC Research Code and approved by the committee of the UMCU Biobank. This study was also approved by the Ethical Committee of the Medical University of Vienna. All materials have been collected from donors from whom a written informed consent for the use of the material for research purposes had been obtained by the Bloemenhove Clinic (Heemstede, The Netherlands); these informed consents are kept together with the medical record of the donor by the clinic. ==== Refs References 1. Djaballah H Rowe AJ Harding SE Rivett AJ The multicatalytic proteinase complex (proteasome): structure and conformational changes associated with changes in proteolytic activity Biochem J 1993 292 Pt 3 857 862 10.1042/bj2920857 8318014 2. Coux O Tanaka K Goldberg AL Structure and functions of the 20S and 26S proteasomes Annu Rev Biochem 1996 65 801 847 10.1146/annurev.bi.65.070196.004101 8811196 3. Orlowski M Wilk S Catalytic activities of the 20 S proteasome, a multicatalytic proteinase complex Arch Biochem Biophys 2000 383 1 16 10.1006/abbi.2000.2036 11097171 4. Jung T Grune T The proteasome and the degradation of oxidized proteins: part I—structure of proteasomes Redox Biol 2013 1 178 182 10.1016/j.redox.2013.01.004 24024151 5. Groll M Ditzel L Lowe J Stock D Bochtler M Bartunik HD Huber R Structure of 20S proteasome from yeast at 2.4 A resolution Nature 1997 386 463 471 10.1038/386463a0 9087403 6. Tai HC, Besche H, Goldberg AL, Schuman EM: Characterization of the brain 26S proteasome and its interacting proteins. Front Mol Neurosci 2010, 3 7. Aki M Shimbara N Takashina M Akiyama K Kagawa S Tamura T Tanahashi N Yoshimura T Tanaka K Ichihara A Interferon-gamma induces different subunit organizations and functional diversity of proteasomes J Biochem 1994 115 257 269 8206875 8. Tanaka K Kasahara M The MHC class I ligand-generating system: roles of immunoproteasomes and the interferon-gamma-inducible proteasome activator PA28 Immunol Rev 1998 163 161 176 10.1111/j.1600-065X.1998.tb01195.x 9700509 9. Groettrup M Kirk CJ Basler M Proteasomes in immune cells: more than peptide producers? Nat Rev Immunol 2010 10 73 78 10.1038/nri2687 20010787 10. Hensley SE Zanker D Dolan BP David A Hickman HD Embry AC Skon CN Grebe KM Griffin TA Chen W Unexpected role for the immunoproteasome subunit LMP2 in antiviral humoral and innate immune responses J Immunol 2010 184 4115 4122 10.4049/jimmunol.0903003 20228196 11. Kruger E Kloetzel PM Immunoproteasomes at the interface of innate and adaptive immune responses: two faces of one enzyme Curr Opin Immunol 2012 24 77 83 10.1016/j.coi.2012.01.005 22296715 12. Diaz-Hernandez M Hernandez F Martin-Aparicio E Gomez-Ramos P Moran MA Castano JG Ferrer I Avila J Lucas JJ Neuronal induction of the immunoproteasome in Huntington’s disease J Neurosci 2003 23 11653 11661 14684867 13. Puttaparthi K Elliott JL Non-neuronal induction of immunoproteasome subunits in an ALS model: possible mediation by cytokines Exp Neurol 2005 196 441 451 10.1016/j.expneurol.2005.08.027 16242125 14. Orre M Kamphuis W Dooves S Kooijman L Chan ET Kirk CJ Dimayuga Smith V Koot S Mamber C Jansen AH Reactive glia show increased immunoproteasome activity in Alzheimer’s disease Brain 2013 136 1415 1431 10.1093/brain/awt083 23604491 15. Bendotti C Marino M Cheroni C Fontana E Crippa V Poletti A De Biasi S Dysfunction of constitutive and inducible ubiquitin-proteasome system in amyotrophic lateral sclerosis: implication for protein aggregation and immune response Prog Neurobiol 2012 97 101 126 10.1016/j.pneurobio.2011.10.001 22033150 16. Jansen AH Reits EA Hol EM The ubiquitin proteasome system in glia and its role in neurodegenerative diseases Front Mol Neurosci 2014 7 73 10.3389/fnmol.2014.00073 25152710 17. Mishto M Ligorio C Bellavista E Martucci M Santoro A Giulioni M Marucci G Franceschi C Immunoproteasome expression is induced in mesial temporal lobe epilepsy Biochem Biophys Res Commun 2011 408 65 70 10.1016/j.bbrc.2011.03.117 21458417 18. Mishto M Raza ML de Biase D Ravizza T Vasuri F Martucci M Keller C Bellavista E Buchholz TJ Kloetzel PM The immunoproteasome beta5i subunit is a key contributor to ictogenesis in a rat model of chronic epilepsy Brain Behav Immun 2015 49 188 196 10.1016/j.bbi.2015.05.007 26044087 19. Wong M Mammalian target of rapamycin (mTOR) activation in focal cortical dysplasia and related focal cortical malformations Exp Neurol 2013 244 22 26 10.1016/j.expneurol.2011.10.002 22015915 20. Aronica E Crino PB Epilepsy related to developmental tumors and malformations of cortical development Neurotherapeutics 2014 11 251 68 10.1007/s13311-013-0251-0 24481729 21. Blumcke I Spreafico R An international consensus classification for focal cortical dysplasias Lancet Neurol 2011 10 26 27 10.1016/S1474-4422(10)70225-8 21094091 22. Samueli S Abraham K Dressler A Groeppel G Jonak C Muehlebner A Prayer D Reitner A Feucht M Padiatrisches TSCZW Tuberous sclerosis complex: new criteria for diagnostic work-up and management Wien Klin Wochenschr 2015 127 619 630 10.1007/s00508-015-0758-y 25860851 23. Vasuri F Capizzi E Bellavista E Mishto M Santoro A Fiorentino M Capri M Cescon M Grazi GL Grigioni WF Studies on immunoproteasome in human liver. Part I: absence in fetuses, presence in normal subjects, and increased levels in chronic active hepatitis and cirrhosis Biochem Biophys Res Commun 2010 397 301 306 10.1016/j.bbrc.2010.05.104 20510670 24. Mishto M Bellavista E Santoro A Stolzing A Ligorio C Nacmias B Spazzafumo L Chiappelli M Licastro F Sorbi S Immunoproteasome and LMP2 polymorphism in aged and Alzheimer’s disease brains Neurobiol Aging 2006 27 54 66 10.1016/j.neurobiolaging.2004.12.004 16298241 25. Prabowo AS van Scheppingen J Iyer AM Anink JJ Spliet WG van Rijen PC Schouten-van Meeteren AY Aronica E Differential expression and clinical significance of three inflammation-related microRNAs in gangliogliomas J Neuroinflammation 2015 12 97 10.1186/s12974-015-0315-7 25986346 26. Iyer A van Scheppingen J Anink J Milenkovic I Kovacs GG Aronica E Developmental patterns of DR6 in normal human hippocampus and in Down syndrome J Neurodev Disord 2013 5 10 10.1186/1866-1955-5-10 23618225 27. Ravizza T Boer K Redeker S Spliet WG van Rijen PC Troost D Vezzani A Aronica E The IL-1beta system in epilepsy-associated malformations of cortical development Neurobiol Dis 2006 24 128 143 10.1016/j.nbd.2006.06.003 16860990 28. Aronica E Gorter JA Rozemuller AJ Yankaya B Troost D Interleukin-1 beta down-regulates the expression of metabotropic glutamate receptor 5 in cultured human astrocytes J Neuroimmunol 2005 160 188 194 10.1016/j.jneuroim.2004.11.015 15710472 29. Iyer A Zurolo E Prabowo A Fluiter K Spliet WG van Rijen PC Gorter JA Aronica E MicroRNA-146a: a key regulator of astrocyte-mediated inflammatory response PLoS One 2012 7 e44789 10.1371/journal.pone.0044789 23028621 30. Zurolo E de Groot M Iyer A Anink J van Vliet EA Heimans JJ Reijneveld JC Gorter JA Aronica E Regulation of Kir4.1 expression in astrocytes and astrocytic tumors: a role for interleukin-1 beta J Neuroinflammation. 2012 9 280 10.1186/1742-2094-9-280 23270518 31. Ramakers C Ruijter JM Deprez RH Moorman AF Assumption-free analysis of quantitative real-time polymerase chain reaction (PCR) data Neurosci Lett 2003 339 62 66 10.1016/S0304-3940(02)01423-4 12618301 32. Ruijter JM Ramakers C Hoogaars WM Karlen Y Bakker O van den Hoff MJ Moorman AF Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR data Nucleic Acids Res 2009 37 e45 10.1093/nar/gkp045 19237396 33. Palmini A Najm I Avanzini G Babb T Guerrini R Foldvary-Schaefer N Jackson G Luders HO Prayson R Spreafico R Vinters HV Terminology and classification of the cortical dysplasias Neurology 2004 62 S2 8 10.1212/01.WNL.0000114507.30388.7E 15037671 34. Boer K Jansen F Nellist M Redeker S van den Ouweland AM Spliet WG van Nieuwenhuizen O Troost D Crino PB Aronica E Inflammatory processes in cortical tubers and subependymal giant cell tumors of tuberous sclerosis complex Epilepsy Res 2008 78 7 21 10.1016/j.eplepsyres.2007.10.002 18023148 35. Iyer A Zurolo E Spliet WG van Rijen PC Baayen JC Gorter JA Aronica E Evaluation of the innate and adaptive immunity in type I and type II focal cortical dysplasias Epilepsia 2010 51 1763 1773 10.1111/j.1528-1167.2010.02547.x 20345941 36. Vezzani A Aronica E Mazarati A Pittman QJ Epilepsy and brain inflammation Exp Neurol 2013 244 11 21 10.1016/j.expneurol.2011.09.033 21985866 37. Wojcik C DeMartino GN Intracellular localization of proteasomes Int J Biochem Cell Biol 2003 35 579 589 10.1016/S1357-2725(02)00380-1 12672451 38. Ding Q Keller JN Proteasomes and proteasome inhibition in the central nervous system Free Radic Biol Med 2001 31 574 584 10.1016/S0891-5849(01)00635-9 11522442 39. Kremer M Henn A Kolb C Basler M Moebius J Guillaume B Leist M Van den Eynde BJ Groettrup M Reduced immunoproteasome formation and accumulation of immunoproteasomal precursors in the brains of lymphocytic choriomeningitis virus-infected mice J Immunol 2010 185 5549 5560 10.4049/jimmunol.1001517 20881186 40. Wehenkel M Ban JO Ho YK Carmony KC Hong JT Kim KB A selective inhibitor of the immunoproteasome subunit LMP2 induces apoptosis in PC-3 cells and suppresses tumour growth in nude mice Br J Cancer 2012 107 53 62 10.1038/bjc.2012.243 22677907 41. Hegde AN Upadhya SC Proteasome and transcription: a destroyer goes into construction Bioessays 2006 28 235 239 10.1002/bies.20379 16479576 42. Kodadek T No splicing, no dicing: non-proteolytic roles of the ubiquitin-proteasome system in transcription J Biol Chem 2010 285 2221 2226 10.1074/jbc.R109.077883 19955182 43. McCarthy MK Weinberg JB The immunoproteasome and viral infection: a complex regulator of inflammation Front Microbiol 2015 6 21 10.3389/fmicb.2015.00021 25688236 44. van Deventer S Neefjes J The immunoproteasome cleans up after inflammation Cell 2010 142 517 518 10.1016/j.cell.2010.08.002 20723753 45. Butler T, Li Y, Tsui W, Friedman D, Maoz A, Wang X, Harvey P, Tanzi E, Morim S, Kang Y, Mosconi L, Talos D, Kuzniecky R, Vallhabjosula S, Thesen T, Glodzik L, Ichise M, Silbersweig D, Stern E, de Leon MJ, French J. Transient and chronic seizure-induced inflammation in human focal epilepsy. Epilepsia. 2016. doi:10.1111/epi.13457. 46. Mühlebner A, van Scheppingen J, Hulshof HM, Scholl T, Iyer AM, Anink JJ, van den Ouweland AM, Nellist MD, Jansen FE, Spliet WG, Krsek P, Benova B, Zamecnik J, Crino PB, Prayer D, Czech T, Wöhrer A, Rahimi J, Höftberger R, Hainfellner JA, Feucht M, Aronica E. Novel histopathological patterns in cortical tubers of epilepsy surgery patients with tuberous sclerosis complex. PLoS One. 2016;11:e0157396. 47. Boer K Crino PB Gorter JA Nellist M Jansen FE Spliet WG van Rijen PC Wittink FR Breit TM Troost D Gene expression analysis of tuberous sclerosis complex cortical tubers reveals increased expression of adhesion and inflammatory factors Brain Pathol 2010 20 704 719 10.1111/j.1750-3639.2009.00341.x 19912235 48. Prabowo AS Anink JJ Lammens M Nellist M van den Ouweland AM Adle-Biassette H Sarnat HB Flores-Sarnat L Crino PB Aronica E Fetal brain lesions in tuberous sclerosis complex: TORC1 activation and inflammation Brain Pathol 2013 23 45 59 10.1111/j.1750-3639.2012.00616.x 22805177 49. Ebstein F Kloetzel PM Kruger E Seifert U Emerging roles of immunoproteasomes beyond MHC class I antigen processing Cell Mol Life Sci 2012 69 2543 2558 10.1007/s00018-012-0938-0 22382925 50. Basler M Kirk CJ Groettrup M The immunoproteasome in antigen processing and other immunological functions Curr Opin Immunol 2013 25 74 80 10.1016/j.coi.2012.11.004 23219269 51. Warnatsch A Bergann T Kruger E Oxidation matters: the ubiquitin proteasome system connects innate immune mechanisms with MHC class I antigen presentation Mol Immunol 2013 55 106 109 10.1016/j.molimm.2012.10.007 23140834 52. Maldonado M Kapphahn RJ Terluk MR Heuss ND Yuan C Gregerson DS Ferrington DA Immunoproteasome deficiency modifies the alternative pathway of NFkappaB signaling PLoS One 2013 8 e56187 10.1371/journal.pone.0056187 23457524 53. Pintado C Gavilan MP Gavilan E Garcia-Cuervo L Gutierrez A Vitorica J Castano A Rios RM Ruano D Lipopolysaccharide-induced neuroinflammation leads to the accumulation of ubiquitinated proteins and increases susceptibility to neurodegeneration induced by proteasome inhibition in rat hippocampus J Neuroinflammation 2012 9 87 10.1186/1742-2094-9-87 22559833 54. Prabowo AS Iyer AM Anink JJ Spliet WG van Rijen PC Aronica E Differential expression of major histocompatibility complex class I in developmental glioneuronal lesions J Neuroinflammation 2013 10 12 10.1186/1742-2094-10-12 23347564 55. Lim HK Choi YA Park W Lee T Ryu SH Kim SY Kim JR Kim JH Baek SH Phosphatidic acid regulates systemic inflammatory responses by modulating the Akt-mammalian target of rapamycin-p70 S6 kinase 1 pathway J Biol Chem 2003 278 45117 45127 10.1074/jbc.M303789200 12960176 56. Schmitz F Heit A Dreher S Eisenacher K Mages J Haas T Krug A Janssen KP Kirschning CJ Wagner H Mammalian target of rapamycin (mTOR) orchestrates the defense program of innate immune cells Eur J Immunol 2008 38 2981 2992 10.1002/eji.200838761 18924132 57. Weichhart T Saemann MD The multiple facets of mTOR in immunity Trends Immunol 2009 30 218 226 10.1016/j.it.2009.02.002 19362054 58. Bejarano E Rodriguez-Navarro JA Autophagy and amino acid metabolism in the brain: implications for epilepsy Amino Acids 2015 47 2113 2126 10.1007/s00726-014-1822-z 25145921 59. Bockaert J Marin P mTOR in brain physiology and pathologies Physiol Rev 2015 95 1157 1187 10.1152/physrev.00038.2014 26269525 60. Yasin SA Ali AM Tata M Picker SR Anderson GW Latimer-Bowman E Nicholson SL Harkness W Cross JH Paine SM Jacques TS mTOR-dependent abnormalities in autophagy characterize human malformations of cortical development: evidence from focal cortical dysplasia and tuberous sclerosis Acta Neuropathol 2013 126 207 218 10.1007/s00401-013-1135-4 23728790 61. Lilienbaum A Relationship between the proteasomal system and autophagy Int J Biochem Mol Biol 2013 4 1 26 23638318 62. Zhang HM Fu J Hamilton R Diaz V Zhang Y The mammalian target of rapamycin modulates the immunoproteasome system in the heart J Mol Cell Cardiol 2015 86 158 167 10.1016/j.yjmcc.2015.07.027 26239133 63. van Vliet EA Forte G Holtman L den Burger JC Sinjewel A de Vries HE Aronica E Gorter JA Inhibition of mammalian target of rapamycin reduces epileptogenesis and blood–brain barrier leakage but not microglia activation Epilepsia 2012 53 1254 1263 10.1111/j.1528-1167.2012.03513.x 22612226 64. Muchamuel T Basler M Aujay MA Suzuki E Kalim KW Lauer C Sylvain C Ring ER Shields J Jiang J A selective inhibitor of the immunoproteasome subunit LMP7 blocks cytokine production and attenuates progression of experimental arthritis Nat Med 2009 15 781 787 10.1038/nm.1978 19525961 65. Dou QP Zonder JA Overview of proteasome inhibitor-based anti-cancer therapies: perspective on bortezomib and second generation proteasome inhibitors versus future generation inhibitors of ubiquitin-proteasome system Curr Cancer Drug Targets 2014 14 517 536 10.2174/1568009614666140804154511 25092212 66. Miller Z Ao L Kim KB Lee W Inhibitors of the immunoproteasome: current status and future directions Curr Pharm Des 2013 19 4140 4151 10.2174/1381612811319220018 23181576 67. Miller Z Lee W Kim KB The immunoproteasome as a therapeutic target for hematological malignancies Curr Cancer Drug Targets 2014 14 537 548 10.2174/1568009614666140723113139 25059201 68. Basler M Mundt S Bitzer A Schmidt C Groettrup M The immunoproteasome: a novel drug target for autoimmune diseases Clin Exp Rheumatol 2015 33 S74 79 26458097 69. Bellavista E Santoro A Galimberti D Comi C Luciani F Mishto M Current understanding on the role of standard and immunoproteasomes in inflammatory/immunological pathways of multiple sclerosis Autoimmune Dis 2014 2014 739705 24523959 70. Ferrington DA Hussong SA Roehrich H Kapphahn RJ Kavanaugh SM Heuss ND Gregerson DS Immunoproteasome responds to injury in the retina and brain J Neurochem 2008 106 158 169 10.1111/j.1471-4159.2008.05345.x 18346202 71. Seifert U Bialy LP Ebstein F Bech-Otschir D Voigt A Schroter F Prozorovski T Lange N Steffen J Rieger M Immunoproteasomes preserve protein homeostasis upon interferon-induced oxidative stress Cell 2010 142 613 624 10.1016/j.cell.2010.07.036 20723761 72. Malik AR Liszewska E Skalecka A Urbanska M Iyer AM Swiech LJ Perycz M Parobczak K Pietruszka P Zarebska MM Tuberous sclerosis complex neuropathology requires glutamate-cysteine ligase Acta Neuropathol Commun 2015 3 48 10.1186/s40478-015-0225-z 26220190 73. Iyer A, Prabowo A, Anink J, Spliet WG, van Rijen PC, Aronica E: Cell injury and premature neurodegeneration in focal malformations of cortical development. Brain pathology 2013 74. Ding M Shen K The role of the ubiquitin proteasome system in synapse remodeling and neurodegenerative diseases Bioessays 2008 30 1075 1083 10.1002/bies.20843 18937340 75. Lazarevic V Pothula S Andres-Alonso M Fejtova A Molecular mechanisms driving homeostatic plasticity of neurotransmitter release Front Cell Neurosci 2013 7 244 10.3389/fncel.2013.00244 24348337

PMC005xxxxxx/PMC5002183.txt

==== Front Zoological LettZoological LettZoological Letters2056-306XBioMed Central London 5610.1186/s40851-016-0056-1Research ArticleSex chromosome evolution in snakes inferred from divergence patterns of two gametologous genes and chromosome distribution of sex chromosome-linked repetitive sequences Matsubara Kazumi +81 45 788 7641mbara@affrc.go.jp 125Nishida Chizuko nishida@cris.hokudai.ac.jp 3Matsuda Yoichi yoimatsu@nuagr1.agr.nagoya-u.ac.jp 24Kumazawa Yoshinori kuma@nsc.nagoya-cu.ac.jp 11 Department of Information and Basic Science and Research Center for Biological Diversity, Graduate School of Natural Sciences, Nagoya City University, 1 Yamanohata, Mizuho-cho, Mizuho-ku, Nagoya, Aichi 467-8501 Japan 2 Laboratory of Animal Genetics, Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8601 Japan 3 Department of Biological Science, Faculty of Science, Hokkaido University, North 10 West 8, Kita-ku, Sapporo, Hokkaido 060-0810 Japan 4 Avian Bioscience Research Center, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8601 Japan 5 Current affiliation: Research Center for Bioinformatics and Biosciences, National Research Institute of Fisheries Science, Japan Fisheries Research and Education Agency, Yokohama, Kanagawa 236-8648 Japan 26 8 2016 26 8 2016 2016 2 1 199 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 The discovery of differentially organized sex chromosome systems suggests that heteromorphic sex chromosomes evolved from a pair of homologous chromosomes. Whereas karyotypes are highly conserved in alethinophidian snakes, the degeneration status of the W chromosomes varies among species. The Z and W chromosomes are morphologically homomorphic in henophidian species, whereas in snakes belonging to caenophidian families the W chromosomes are highly degenerated. Snakes therefore are excellent animal models in which to study sex chromosome evolution. Herein, we investigated the differentiation processes for snake sex chromosomes using both coding and repetitive sequences. We analyzed phylogenetic relationships of CTNNB1 and WAC genes, localized to the centromeric and telomeric regions, respectively, of the long arms on snake sex chromosomes, and chromosome distribution of sex chromosome-linked repetitive sequences in several henophidian and caenophidian species. Results Partial or full-length coding sequences of CTNNB1 and WAC were identified for Z homologs of henophidian species from Tropidophiidae, Boidae, Cylindrophiidae, Xenopeltidae, and Pythonidae, and for Z and W homologs of caenophidian species from Acrochordidae, Viperidae, Elapidae, and Colubridae. Female-specific sequences for the two genes were not found in the henophidian (boid and pythonid) species examined. Phylogenetic trees constructed using each gene showed that the Z and W homologs of the caenophidian species cluster separately. The repetitive sequence isolated from the W chromosome heterochromatin of the colubrid Elaphe quadrivirgata and a microsatellite motif (AGAT)8 were strongly hybridized with W chromosomes of the viperid and colubrid species examined. Conclusion Our phylogenetic analyses suggest that the cessation of recombination between the Z and W homologs of CTNNB1 and WAC predated the diversification of the caenophidian families. As the repetitive sequences on the W chromosomes were shared among viperid and colubrid species, heterochromatinization of the proto-W chromosome appears to have occurred before the splitting of these two groups. These results collectively suggest that differentiation of the proto-Z and proto-W chromosomes extended to wide regions on the sex chromosomes in the common ancestor of caenophidian families during a relatively short period. Electronic supplementary material The online version of this article (doi:10.1186/s40851-016-0056-1) contains supplementary material, which is available to authorized users. Keywords SnakeZ chromosomeW chromosomePhylogenyEvolutionGametologRepetitive sequencesHeterochromatinGrants-in-Aid for Young Scientists (B)19770001Matsubara Kazumi Grants-in-Aid for JSPS Fellows09J05132Matsubara Kazumi issue-copyright-statement© The Author(s) 2016 ==== Body Background The discovery of differentially organized sex chromosome systems suggests that heteromorphic sex chromosomes evolved from a pair of homologous chromosomes [1, 2]. The first step is thought to have been the acquisition of a novel sex-determining gene on one member of an autosomal pair, followed by accumulation of alleles conferring an advantage to that sex [3, 4]. Meiotic recombination between the proto-sex chromosomes could have been suppressed around the heterologous region to preserve the linkage of these sexually antagonistic genes. Such suppression may have been accelerated by structural changes in chromosomes (e.g., inversion). Suppression of recombination between the sex chromosomes then favored the accumulation of repetitive DNA sequences on the non-recombining regions, increasing the extent of differentiation between sex chromosomes [5]. Extant snake species belonging to Serpentes are grouped into two infraorders, Scolecophidia and Alethinophidia. Blind snakes and thread snakes belong to the former and all other snakes belong to the latter (Fig. 1). Alethinophidian species are divided into two superfamilies, Henophidia and Caenophidia (Fig. 1). Snake karyotypes are highly conserved, and most alethinophidian species have a common karyotype whose diploid number is 36, consisting of eight pairs of macrochromosomes and ten pairs of microchromosomes [6–8]. The sex determination system is also conserved in snakes. Nearly all alethinophidian species have ZZ/ZW-type sex chromosomes, in which males have a homomorphic ZZ sex chromosome and females have a heteromorphic ZW. The Z chromosomes are the fourth or fifth largest metacentric chromosomes for most species. In contrast to the highly conserved Z chromosomes, the degeneration status of W chromosomes varies among species [2, 6, 9, 10]. The Z and W chromosomes are homomorphic in the boids and pythonids. Conversely, W chromosomes are highly degenerated and heterochromatic in the colubrids, elapids, and viperids. This characteristic makes snakes good model species for the study of sex chromosome evolution.Fig. 1 Phylogenetic relationships between snake families. Phylogeny, divergence time and classification are based on Vidal et al. [63], Pyron et al. [49], and Uetz and Hošek [65] We previously constructed, using fluorescence in situ hybridization (FISH), a cytogenetic map of the Japanese four-striped rat snake (Elaphe quadrivirgata) with more than 180 cDNA clones and found that three genes, CTNNB1, RAB5A, and WAC, were commonly mapped to Z and W chromosomes of this species [11–13]. We also compared structures of sex chromosomes, by C-banding and comparative mapping of Z-linked genes, among three snakes, E. quadrivirgata (Colubridae), Protobothrops flavoviridis (formerly Trimeresurus flavoviridis, Viperidae), and Python bivittatus (Pythonidae). The results revealed that W chromosomes of E. quadrivirgata and P. flavoviridis were highly degenerated, not only in morphology, but also in DNA sequences [12]. Gametologous genes are homologs located on opposite sex chromosomes, which arose through the lack of recombination and subsequent differentiation of sex chromosomes [14]. Although Y chromosomes of eutherian mammals and W chromosomes of neognathous birds are highly degenerated and extensively heterochromatized, the human Y chromosome still contains more than 27 homologs of X-linked single copy genes and pseudogenes [15] and the chicken (Gallus gallus) has the Z and W forms of six gametologous genes [14, 16, 17]. In the process of sex chromosome differentiation, suppression of meiotic recombination between entire or partial regions of opposite sex chromosomes facilitates sequence divergence between gametologs. Thus, the evolutionary process of sex chromosome differentiation can be examined by molecular phylogenetic analyses of the gametologs [17–23]. Ellegren and coworkers estimated the date and process of sex chromosome differentiation in birds by comparing gametologs between and within species [17, 19, 23]. Similar to in birds, the evolutionary process of sex chromosome differentiation has also been identified through comparative analysis of gametologous genes in mammals [18, 21], dioecious plants of genus Silene [20, 22], and papaya [24]. Recently, massive genome sequencing and transcriptome analyses identified putative gametologous genes in two snakes, Thamnophis elegans (Colubridae) and Sistrurus miliarius (Viperidae) [25]. The comparative analysis of gametologous genes revealed completely differentiated sex chromosomes in the two species, which suggests that suppression of recombination between the Z and W homologs began before the divergence of the two lineages. Y chromosomes of eutherian mammals and W chromosomes of neognathous birds are highly heterochromatic and rich in repetitive sequences. Accumulation of repetitive sequences, such as retrotransposons, microsatellite repeats, and ribosomal DNAs, on sex chromosomes has been reported in many species of animals and plants (e.g., [26–29]). The accumulation of repetitive sequences thus is probably associated with heterochromatinization of sex-specific chromosomes (Y and W chromosomes). Accumulation of Bkm repeats, which contain two microsatellite motifs, (GATA)n and (GACA)n, was identified on W chromosomes of several colubrid and elapid snakes [30–33], suggesting that these repeat sequences were amplified on the W chromosomes in the common ancestor of Colubridae and Elapidae. We identified amplification of two repetitive sequence families, EQU-BamHI-4 and EQU-BglI-15, on sex chromosomes of E. quadrivirgata ([12], Matsubara K. unpublished data). Whereas the EQU-BamHI-4 was amplified in the telomeric regions of both the Z and W chromosomes, the EQU-BglI-15 was intensively amplified on the W chromosome. In the present study, we sequenced Z and W homologs of CTNNB1 and WAC genes located on centromeric and telomeric regions, respectively, on Z chromosomes [12] from 16 species representing 10 snake families. We also searched for snake homologs of the genes in international nucleotide sequence databases. We constructed molecular phylogenetic trees using these genes to infer the differentiation process for the Z and W chromosomes of snakes based on the divergence patterns of the two genes. We also conducted FISH mapping of (AGAT)8 microsatellite motifs and two repetitive sequence families, EQU-BamHI-4 and EQU-BglI-15, for chromosomes of Colubridae, Viperidae, Boidae, and Pythonidae. Finally, we delineated the evolutionary process of sex chromosomes in snakes. Methods Animals Table 1 lists the snake species used for this study. One female E. quadrivirgata collected in Mie, Japan, was used for chromosome preparation. We also obtained one male, one female, and eggs from a population bred at the Japan Snake Institute. They were sacrificed to collect tissues for DNA and RNA extraction. All the other species, except for Typhlops sp., I. braminus, C. ruffus, A. arafurae, A. granulatus, and L. semicarinatum, were bred at the Japan Snake Institute, Japan. I. braminus, a pair of P. flavoviridis, and L. semicarinatum were captured at Takarajima, Amami-Oshima, and Okinawajima, Ryukyu Islands, Japan, respectively. DNA samples of Typhlops sp., C. ruffus, A. arafurae, and A. granulatus were obtained from collections of our laboratory.Table 1 Snake samples used for this study Infraoder Superfamily Family Species Abbrev. 2n a No. of used animals Scolecophidia Typhlopidae Typhlops sp. TYP 30 (M: 16, m: 14)b 1 unknow sex Indotyphlops braminus IBR 42 (M: 21, m: 21)c 1 female Alethinophidia Henophidia Tropidophiidae Tropidophis haetianus haetianus THA un 1 male Boidae Boa constrictor BCO 36 (M: 16, m: 20) 1 male, 1 female Cylindrophiidae Cylindrophis ruffus CRU un 1 unknow sex Xenopeltidae Xenopeltis unicolor XUN 36 (M: 16, m: 20)d 1 male Pythonidae Python bivittatus PBI 36 (M: 16, m: 20) 1 male, 1 female Python molurus PMO 36 (M: 16, m: 20) 1 male, 1 female Caenophidia Acrochordidae Acrochordus arafurae AAR 36e 1 male, 1 female Acrochordus granulatus AGR 36 (M: 16, m: 20)f 1 male Viperidae Protobothrops flavoviridis PFL 36 (M: 16, m: 20) 2 males, 2 females Gloydius blomhoffii GBL 36 (M: 16, m: 20) 1 male, 1 female Bitis arietans arietans BAR 36 (M: 16, m: 20) 1 male, 1 female Naja kaouthia NKA 38 (M: 16, m: 22)g 1 male, 1 female Elapidae Elaphe quadrivirgata EQU 36 (M: 16, m: 20) 1 male, 2 females, embryos Colubridae Lycodon semicarinatum LSE 34 (M: 16, m: 18)h 1 male, 1 female Rhabdophis tigrinus tigrinus RTI 40 (M: 16, m: 24) 1 male, 1 female aThe numbers of macrochromosomes (M) and microchromosomes (m) are shown in parentheses. un, the karyotypes have not been identified yet. bThe karyotype was identified in our lab [Matsubara et al., unpublished data] cThe karyotypic information is derived from Ota et al. [66] dThe karyotypic information is derived from Singh et al. [30], and Cole and Dowling [67] eThe karyotypic information is derived from CHROMOREP [68] fThe karyotypic information is derived from Sharma and Nakhasi [52, 53] gThe karyotypic information is derived from Singh [8] and Ray-Chaudhuri et al. [69] hThe karyotypic information is derived from Toriba [70] Sequencing of Z and W homologs of CTNNB1 and WAC genes Genomic DNA was extracted from blood or liver tissue by the phenol-chloroform method described by Sambrook et al. [34] or with a DNeasy kit (QIAGEN, Venlo, Netherlands), and used for templates in PCR. We determined, by primer walking, full-length nucleotide sequences of two E. quadrivirgata expressed sequence tag (EST) clones, Eq_aB_009012_N17 (BW999995) and Eq_aB_026_N02 (AU312355), previously identified as homologs of CTNNB1 and WAC genes, respectively [11, 12]. We located positions of the intron/exon boundaries on the sequences of E. quadrivirgata CTNNB1 and WAC homologs in comparison with chicken, green anole (Anolis carolinensis), and human homologs, and designed primer pairs to amplify partial exons and flanking introns (see Additional file 1 for primer sequences and Additional file 2a for their locations). PCR was conducted with a SpeedStar HS DNA polymerase (Takara, Kusatsu, Japan) under the following conditions: an initial denaturation at 94 °C for 5 min, followed by 35 cycles of 94 °C for 30 s, 50–65 °C for 30 s, 72 °C for 35 s, and 72 °C for 5 min for a final extension. Annealing temperature was changed depending on primers and target species. The PCR products were electrophoresed on 1–3 % agarose gels, and bands were isolated using a QIAquick Gel Extraction Kit (QIAGEN). Extracted DNA was directly sequenced or subcloned using the pGEM-T Easy Vector System (Promega, Madison, WI, USA). For direct sequencing, the 20–40 ng DNA fragments were labeled with a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems - Thermo Fisher Scientific, Waltham, MA, USA) using primer sets for each gene fragment based on the manufacturer’s protocol (Applied Biosystems). Both strands of the labeled products were sequenced using an ABI PRISM3700 DNA Analyzer (Applied Biosystems). The cloned DNA fragments were sequenced with T7 and Sp6 primers. All species, except for B. arietans, were used for sequencing and phylogenetic analyses of the CTNNB1 genes, whereas Typhlops sp., I. braminus, T. haetianus, B. constrictor, C. ruffus, X. unicolor, P. bivittatus, A. arafurae, A. granulatus, P. flavoviridis, N. kaouthia, and E. quadrivirgata were used for phylogenetic analyses of WAC genes. Rapid amplification of cDNA ends for E. quadrivirgata CTNNB1 and WAC homologs Total RNA was extracted from E. quadrivirgata fetal gonads using an RNeasy Kit (Qiagen). For rapid amplification of cDNA ends (RACE), cDNA was synthesized with a SMARTer® PCR cDNA Synthesis Kit (Clontech, Mountain View, CA, USA) according to the manufacturer’s protocol with the following modification. We used our own primer, 5′-GGC CAC GCG TCG ACT AGT AC(T)30 VN-3′, instead of the manufacturer’s primer for synthesis of the first strand of cDNA. Gene-specific primers were designed based on partial sequences obtained in the previous section (Additional file 1). Comparison of sequences of CTNNB1 and WAC genes among tetrapods The open reading frames (ORFs) in E. quadrivirgata Z and W homologs of CTNNB1 and WAC were predicted for the full-length cDNA sequences based on sequence similarities with homologs from humans, chickens, and green anole lizards. Putative full-length coding nucleotide and amino acid sequences of the CTNNB1 Z and W, and WAC Z and W homologs were aligned with the homologs from other tetrapods using Clustal Omega [35] at the European Bioinformatics Institute (EMBL-EBI) website. The nucleotide and amino acid sequences of the two genes from the following species were used for the comparison; CTNNB1 (XM_003223954) and WAC (XM_008112381) from A. carolinensis (Iguanidae, Squamata), CTNNB1 (KF803272) and WAC (XM_015421414) from Gekko japonicus (Gekkonidae, Squamata), CTNNB1 (NM_205081) and WAC (XM_015282076) from G. gallus (Phasianidae, Aves), CTNNB1 (XM_009687805) and WAC (XM_009679627) from Struthio camelus australis (Struthionidae, Aves), CTNNB1 (XM_006258718) and WAC (XM_014609006) from Alligator mississippiensis (Alligatoridae, Crocodilia), CTNNB1 (XM_005278593) and WAC (XM_005290938) from Chrysemys picta bellii (Emydidae, Testudines), CTNNB1 (NM_001286932) and WAC (XM_006124777) from Pelodiscus sinensis (Trionychidae, Testudines), CTNNB1 (NM_001904) and WAC (NM_016628) from Homo sapiens (Hominidae, Primates, Mammalia), and CTNNB1 (NM_001016958) and WAC (XM_012964589) from Xenopus (Silurana) tropicalis (Pipidae, Amphibia). Identification of snake CTNNB1 and WAC gene sequences in databases To obtain long coding sequences of the two genes from several snake species, we searched databases for sequences that exhibited high similarities with the E. quadrivirgata homologs. BLASTN searches were conducted on the National Center for Biotechnology Information (NCBI) website against whole genome shotgun sequences of female P. bivittatus (Pythonidae, BioProject no. PRJNA61243), male Ophiophagus hannah (Elapidae, PRJNA201683), female Crotalus mitchellii pyrrhus (Viperidae, PRJNA255393), and female Vipera berus berus (Viperidae, PRJNA170536) using the full-length cDNA sequences of E. quadrivirgata CTNNB1 and WAC homologs as queries. The contig sequences that exhibited high similarities with the E. quadrivirgata cDNA sequences, which consisted of exons, introns, and flanking regions, were selected for each species (Additional file 3). The boundaries between exons and introns within each contig were manually identified using dot-plot matrices between the cDNA sequences and the contig sequences. Next, the exon sequences were combined and the full-length or near full-length coding sequence was determined for homologs of the CTNNB1 and WAC genes in the four species based on the ORF information from other tetrapods. Available transcriptomic reads were obtained from the NCBI database for the following samples: female Boa constrictor blood (Sequence Read Archive (SRA) No. SRR941236), male Sistrurus miliarius liver (SRR941232), Xenopeltis unicolor liver (SRR629647), and male Echis coloratus brain (SRR1328164) (Viperidae). The reads were trimmed based on quality using the DynamicTrim command (h = 30), and those shorter than 20 bp were removed using the LengthSort command in SolexaQA [36]. The screened reads were assembled using Trinity [37]. Transcripts of CTNNB1 and WAC genes from each species were identified by the BLASTN search using the full-length cDNA sequences of E. quadrivirgata homologs as queries with BlastStation (TM Software, Arcadia, CA, USA). Full-length or near full-length coding sequences for the four snake species were determined based on the sequence similarities to homologs of E. quadrivirgata and other tetrapods. Multiple contigs were identified in search of B. constrictor, X. unicolor, and S. miliarius homologs of the CTNNB1 gene, and for B. constrictor homologs of the WAC gene. The variation in contigs was probably caused by the presence of transcript variants and precursor mRNA in the tissues. In these cases, transcripts that showed the highest similarity to the homologs of E. quadrivirgata and other tetrapods were selected. Two contigs of S. miliarius showed high similarities to the E. quadrivirgata CTNNB1 gene: one was homologous to two-thirds of the coding region and the other was homologous to the remaining one-third. The two contigs shared a 21-bp overlapping sequence at their ends, and thus, they were assembled and the full-length coding sequence was identified as the S. miliarius homolog. The sequences of two Thamnophis sirtalis homologs of CTNNB1 (XM_014069347, XM_014063622) and WAC (XM_014065195) were obtained from the International Nucleotide Sequence Database. Phylogenetic analysis of CTNNB1 and WAC genes Sequence alignment was performed with ClustalW [38] implemented in MEGA ver.6 [39], visually checked, and corrected. Neighbor-joining (NJ) and maximum-likelihood (ML) trees were constructed using PAUP ver.4.0a147 [40] and GARLI 2.0 [41], respectively. The most appropriate models and parameters for construction of phylogenetic trees (Additional file 4) were defined for each alignment based on the Bayesian information criterion (BIC) using the jModelTest [42, 43]. The robustness of trees was assessed by bootstrap resampling with 1000 random replications. We constructed two kinds of molecular phylogenetic trees for the two genes. One tree was constructed with a long alignment that contained full-length coding sequences of the E. quadrivirgata Z and W homologs, coding sequences of homologs for other snakes, and non-snake tetrapods identified from genomic databases and transcriptomic data. The other tree was constructed with a short alignment that covered only the amplified and sequenced region of the genes from various snake families and non-snake squamates. The alignments were constructed for only exon sequences because reliable alignments were not obtained with sequences of introns and untranslated regions (UTR). The alignment lengths were 2370, 588, 1974, and 524 sites for the CTNNB1 gene in the long alignment, the CTNNB1 gene in the short alignment, the WAC gene in the long alignment, and the WAC gene in the short alignment, respectively. The CTNNB1 sequences from A. carolinensis, Leiolepis reevesii rubritaeniata (AB490379, Agamidae, Squamata), G. japonicus, C. p. bellii, P. sinensis, A. mississippiensis, S. c. australis, G. gallus, H. sapiens, and X. tropicalis, and the WAC sequences from A. carolinensis, L. r. rubritaeniata (AB490381), C. p. bellii, P. sinensis, A. mississippiensis, Alligator sinensis (XM_014520047, Alligatoridae, Crocodilia), S. c. australis, G. gallus, H. sapiens, and X. tropicalis were used for construction of phylogenetic trees. The sequences of Pogona vitticeps (Agamidae, Squamata) homologs of CTNNB1 and WAC genes were obtained from the annotated genome (Pogona pvi1.1) through a genome browser available at https://genomics.canberra.edu.au/gbrowse/gbrowse/pogona_pvi1.1/ [44] and included in the phylogenetic analyses. Chromosome preparation and FISH Chromosome preparation and FISH were performed as described in our previous studies [11, 12, 45–47]. Chromosome slides were made from blood lymphocytes and/or fibroblast cells taken from heart tissues of B. constrictor, P. bivittatus, P. flavoviridis, B. arietans, G. blomhoffii, E. quadrivirgata, and R. tigrinus. The DNA clones of the two sex chromosome-specific repetitive elements obtained from E. quadrivirgata, EQU-BamHI-4 and EQU-BglI-15 ([12], Matsubara K. unpublished data), were labeled using a nick translation kit (Roche Diagnostics, Basel, Switzerland) with biotin-16-dUTP (Roche Diagnostics). Hybridization was conducted at 37 °C for one day. The slides were reacted with FITC-avidin (Roche Diagnostics), and then stained with propidium iodide (PI). The fluorescein-labeled oligonucleotide of (AGAT)8 was purchased from Rikaken (Nagoya, Japan) and used for FISH with the protocol described in our previous studies [46, 47]. Results Sequencing of Z and W homologs of CTNNB1 and WAC genes PCR using the three primer sets for CTNNB1 genes (Eq-CTNNB1-11-F × 13-R, Eq-CTNNB1-int12-F × 14-R, and Eq-CTNNB1-14-F × 15-R) gave rise to a band common to both males and females and a female-specific band in E. quadrivirgata (e.g., Additional file 2b). DNA fragments purified from these bands were sequenced to confirm they were parts of the CTNNB1 gene. Thus, the DNA sequences from the common bands and the female-specific bands were identified as CTNNB1 Z and W homologs, respectively. Similarly, PCR using the two primer sets for WAC (Eq-WAC-6-F × 7-R and Eq-WAC-8-F × 9-R) produced a band common to both males and females, and a female-specific band in E. quadrivirgata (data not shown). The DNA fragments from all these bands were sequenced to confirm they were parts of the WAC gene. Thus, the sequences from the common bands and the female-specific bands were identified as WAC Z and W homologs, respectively. Full-length cDNA of E. quadrivirgata CTNNB1 Z and W homologs and WAC Z and W homologs were obtained by the RACE method with specific primers for the Z and W homologs of each gene (Additional file 1), and the sequences were registered with the International Nucleotide Sequence Database Collaboration (INSDC) under the accession numbers shown in Additional file 5. PCR for the other snake species revealed that all five primer sets described above produced a band common to both males and females and a female-specific band for the acrochordid, viperid, elapid, and colubrid species examined (e.g., Additional file 2c). Sequencing the DNA fragments from these bands also demonstrated they represented Z and W homologs for these species. In contrast, all the five primer sets gave rise to a single band common to males and females for P. bivittatus and B. constrictor (e.g., Additional file 2c). The DNA fragments purified from these single bands were cloned and at least four clones were sequenced to confirm they were identical among clones and between males and females. Specifically, sex-specific sequences were not found for the two genes in P. bivittatus and B. constrictor. Only one individual of Typhlops sp., I. braminus, T. haetianus, C. ruffus, X. unicolor, and A. granulatus were used for sequencing the partial CTNNB1 and WAC gene sequences (Table 1). All five primer sets produced a single band for Typhlops sp., T. haetianus, C. ruffus, X. unicolor, and A. granulatus. In I. braminus, although the three CTNNB1 primer sets and the Eq-WAC-6-F × 7-R primer set produced single bands, the remaining primer set did not provide amplified bands (data not shown). The nucleotide sequences of amplified products were confirmed as partial sequences of the CTNNB1 and WAC genes in all six species. Other primers (snake-WAC-7-F, snake-WAC-8-R, and snake-WAC-W-8-R) were designed using available sequence data to amplify partial sequences from exon 7 to exon 8 (Additional file 1). With these new primers, partial sequences from exon 7 to exon 8 were determined in all species, except for the P. flavoviridis Z homolog, T. haetianus, and C. ruffus. All partial sequences of the two genes obtained in this study have been registered with the INSDC; accession numbers are shown in Additional file 5. Comparison of CTNNB1 and WAC sequences among tetrapods Amino acid sequences of the CTNNB1 genes were highly conserved among the homologs of tetrapod species (Fig. 2a, Additional files 6 and 7). The putative amino acid sequence of the E. quadrivirgata CTNNB1 Z homolog showed more than 99 % similarities to the homologs of the other amniotes and 97.7 % similarity to the X. tropicalis homolog. An insertion (24 bp, 8 aa) was identified at the 1684th site from a start codon in the E. quadrivirgata CTNNB1 W homolog (Fig. 2a and Additional file 7), and the W homolog exhibited approximately 97 % similarity to the homologs of the other amniotes and 95.3 % similarity to the X. tropicalis homolog (Additional file 6).Fig. 2 Comparison of partial nucleotide and amino acid sequences of CTNNB1 and WAC genes. Nucleotide and amino acid sequences are aligned between the homologs of CTNNB1 (a) and WAC (b) genes in five tetrapod species: E. quadrivirgata, A. carolinensis, G. gallus, H. sapiens and X. tropicalis. Numbers on the alignments indicate nucleotide positions from the translation initiation sites. Arrowheads in b indicate two predicted translational initiation sites In contrast to the CTNNB1 genes, amino acid sequences of WAC genes were relatively divergent among the homologs of tetrapod species compared. The amino acid sequence of the E. quadrivirgata Z homolog exhibited 92.7 % similarity to the A. carolinensis homolog, approximately 91 % similarity to the homologs of the chicken and the painted turtle, 88.7 % similarity to the human homolog, and 82.6 % similarity to the X. tropicalis homolog (Additional files 6 and 8). Two ORFs, which corresponded to two chicken transcript variants (variant X1: XM_015282076.1, variant X4: XM_015282080), two human transcript variants (variant X1: NM_016628, variant X2: NM_100264), and two X. tropicalis transcript variants (variant X1: XM_012964589, variant X3: XM_012964591), were identified in the putative amino acid sequence of the E. quadrivirgata WAC Z homolog (Fig. 2b and Additional file 8). The W homolog showed lower similarities in the amino acid sequence with the homologs of the other tetrapod species (Additional file 6). Furthermore, the W homolog did not retain a longer ORF because of a 37-bp deletion at the 42nd site from a start codon, which would cause a frameshift (Fig. 2b and Additional file 8). Although a shorter ORF starting from the second putative start codon was retained in the W homolog, a few additional deletions specific to the W homolog were also identified in the ORF sequence (Additional file 8). Molecular phylogeny of CTNNB1 and WAC genes Phylogenetic trees were constructed for each of the CTNNB1 and WAC genes (Figs. 3 and 4 for ML trees and Additional files 9 and 10 for NJ trees). Phylogenetic relationships among amniote species reconstructed using the CTNNB1 genes from the long alignment (Fig. 3a and Additional file 9a) were in good agreement with other molecular phylogenetic studies [48]. The human homolog was positioned as a sister group to reptiles, and reptiles were divided into two primary clades that corresponded to Archosauromorpha (Testudines, Crocodilia and Aves) and Squamata. Snake species are traditionally divided into three primary groups, Scolecophidia, Henophidia, and Caenophidia (Fig. 1). Although recent molecular studies [49–51] have suggested non-monophyly of the Scolecophidia and Henophidia, they established a clade comprising four henophidian families (Cylindrophiidae, Boidae, Xenopeltidae, and Pythonidae). However, the clustering of henophidian homologs was not conspicuous in the CTNNB1 tree from the long alignment (Fig. 3a and Additional file 9a). The homolog of B. constrictor diverged first from those of the other henophidian and caenophidian species, and thus, the phylogenetic relationships of three henophidian homologs did not completely match the common cladogram shown in Fig. 1. In a caenophidian clade, E. quadrivirgata Z homologs clustered with homologs of other caenophidian species. Whereas one T. sirtalis homolog (XM_014069347) was included in this cluster, the other T. sirtalis homolog (XM_014063622) formed a clade with the E. quadrivirgata W homolog (Fig. 3a and Additional file 9a), suggesting that the latter two sequences represented W homologs in Colubridae.Fig. 3 Molecular phylogenetic trees of CTNNB1 genes. Maximum-likelihood trees of CTNNB1 genes were constructed with the long alignment for 20 tetrapod species (a) and the short alignment for 26 squamate species (b). Bootstrap values (>50 %) are shown on each node. Classification is shown on the right side of species. Blue and pink bars in b show clades of Z and W homologs of caenophidian species, respectively Fig. 4 Molecular phylogenetic trees of WAC genes. Maximum-likelihood trees of WAC genes were constructed with the long alignment for 21 tetrapod species (a) and the short alignment for 21 squamate species (b). Bootstrap values (> 50 %) are shown on each node. Classification is shown on the right side of species. Blue and pink bars in b show clades of Z and W homologs of caenophidian species, respectively In the CTNNB1 trees constructed from the short alignment, the second homologs (i.e., W homologs) of eight caenophidian species from Colubridae, Viperidae, Elapidae, and Acrochordidae clearly formed a monophyletic group with 68 and 55 % bootstrap support (Fig. 3b and Additional file 9b). In addition, the first homologs (i.e., Z homologs) of species from three caenophidian families (Colubridae, Viperidae, and Elapidae) also comprised a monophyletic group with strong bootstrap values (100 and 99 %). However, as in the long-alignment trees, homologs of henophidian species from Pythonidae, Xenopeltidae, Cylindrophiidae, and Boidae were not monophyletic. It should be noted that Z homologs of acrochordid species clustered with those of other caenophidian species in the NJ tree, whereas this was not the case in the ML tree, indicating that the phylogenetic position of the acrochordid Z homologs was not resolved well with our dataset. The WAC trees from the long alignment mostly reconstructed the common cladogram of the amniotes [48] and the henophidian snakes (Fig. 1), except for the position of a human homolog, which exhibited a sister-group relationship to Archosauromorpha in ML tree (Fig. 4a and Additional file 10a). The branching patterns of caenophidian homologs were similar to those in the CTNNB1 trees from the long alignment. The E. quadrivirgata W homolog diverged from the caenophidian Z clade after splitting into the caenophidian and henophidian homologs (Fig. 4a). A T. sirtalis homolog (XM_014065195) clustered with the E. quadrivirgata Z homolog in the ML tree with 83 % bootstrap support (Fig. 4a), whereas it clustered with the E. quadrivirgata W homolog in the NJ tree, albeit with a lower (<50 %) bootstrap support (Additional file 10a). In the WAC trees from the short alignment (Fig. 4b and Additional file 10b), homologs of henophidian species from Pythonidae, Xenopeltidae, Cylindrophiidae, and Boidae also did not form a monophyletic clade. Within a caenophidian clade, the Z and W homologs of colubrid, elapid, and viperid species formed mutually monophyletic clades in both ML and NJ trees. The bootstrap values were 84 and 70 % for the Z homolog clades in ML and NJ trees, respectively, and 94 and 83 % for the W homolog clades in ML and NJ trees, respectively. The Z and W homologs of acrochordid species formed a clade separated from the Z and W homologs of the other caenophidian species in the ML tree, although bootstrap support was weak (<50 %) (Fig. 4b). In contrast, in the NJ tree, W homologs of all caenophidian species formed a clade in which the A. arafurae W homolog was a sister group to the other species with 72 % bootstrap support and the Z homologs of acrochordid species did not have a sister-group relationship with the Z homolog clade of the other caenophidian species (Additional file 10b). The phylogenetic affiliation of acrochordid Z and W homologs was not therefore conclusive in our trees. Comparative FISH mapping of sex chromosome-specific repetitive DNAs One of two E. quadrivirgata repetitive sequences, EQU-BamHI-4 repeat [12], was localized to the distal regions of the short arms of the Z and W chromosomes in all species examined: B. constrictor (Fig. 5a), R. tigrinus (Fig. 5b), and B. arietans (Fig. 5c), as well as in E. quadrivirgata, P. bivittatus, and P. flavoviridis (formerly Trimeresurus flavoviridis) [15]. The fluorescent signals were fainter in a henophidian species, B. constrictor (Boidae) (Fig. 5a) than in the colubrid R. tigrinus and the viperid B. arietans (Fig. 5b and c). These variations of signal intensities for this repeat were also observed in a previous FISH experiment [12] in which the repeat showed intense and faint signals in P. flavoviridis (Viperidae) and P. bivittatus (Pythonidae), respectively, and an intermediate intensity was observed in E. quadrivirgata (Colubridae).Fig. 5 FISH of three repetitive sequences in snakes. FITC-labeled E. quadrivirgata BamHI-4 repeat was hybridized to PI-stained metaphase spreads of B. constrictor (a), R. tigrinus (b), and B. arietans (c). E. quadrivirgata BglI-15 repeat was hybridized to metaphase spreads of R. tigrinus (d), P. flavoviridis (e), and B. arietans (f). The (AGAT)8 microsatellite motif was hybridized to metaphase spreads of E. quadrivirgata (g), R. tigrinus (h), and P. flavoviridis (i). Arrowheads indicate hybridization signals on sex chromosomes Another repetitive sequence, EQU-BglI-15 repeat, was originally identified on the long arm of the W chromosome, the centromeric regions of the Z chromosome, and on three autosome pairs (4th, 6th, and 7th) in E. quadrivirgata (Additional file 11a). In R. tigrinus, this repetitive sequence provided intense signals on the Z chromosome, the paracentric regions of one pair of small macrochromosomes (5th) and the centromeric regions of the smallest pair of macrochromosomes (7th), but not on the W chromosome (Fig. 5d). The repetitive sequence showed intense fluorescent signals on the long arms of the W chromosomes in P. flavoviridis (Fig. 5e), B. arietans (Fig. 5f), and G. blomhoffii (Additional file 11b), as in E. quadrivirgata. A microsatellite motif, (AGAT)8, showed intense hybridization signals on the long arm of W chromosomes in E. quadrivirgata (Fig. 5g) and R. tigrinus (Fig. 5h), and the telomeric regions of the long arm of W chromosomes in P. flavoviridis (Fig. 5i) and G. blomhoffii (Additional file 11c). In contrast, the EQU-BglI-15 repeat and the microsatellite motif showed no site-specific signals in B. constrictor or P. bivittatus (data not shown). Discussion Identification of two gametologous genes in snakes Except for E. quadrivirgata and P. flavoviridis, we used only one male and one female, or single individuals of each species, for PCR amplification of partial sequences of CTNNB1 and WAC genes. However, female-specific duplicate amplicons of the two genes were clearly identified in all caenophidian species, including all females of E. quadrivirgata and P. flavoviridis. Moreover, only single amplicons were recovered from males of all caenophidian species. This indicates that the female-specific amplicons are linked to the caenophidian W chromosomes. Thus, the Z and W homologs of the two genes are differentiated in caenophidian species, including acrochordid species whose Z and W chromosomes have not yet been identified morphologically [52, 53]. The presence of W homologs of the two genes was not identified by FISH in P. flavoviridis in our previous study [12]. This might have occurred because of the difficulty of gene mapping in heterochromatic regions. In two henophidian species, B. constrictor and P. bivittatus, the female-specific amplicons were not found for the CTNNB1 and WAC genes. Three explanations are possible for this result. The first is that the Z and W homologs are quite similar or not differentiated (i.e., homologous) from each other, an explanation that we favor. The second is that the primers used for PCR did not match the nucleotide sequences of the priming sites on the W homologs of the two genes. The third is that the homologs are already lost from the W chromosomes in henophidian species. However, the third explanation is unlikely because the presence of the W homologs was evidenced by genomic sequencing approaches in B. constrictor [28] and by cytogenetic analysis in P. bivittatus [15]. Because we used only one individual of the other three henophidian species (T. haetianus, C. ruffus, and X. unicolor), our results do not conclusively reveal whether the sequences for the two genes are identical in the Z and W homologs. Further research is needed for both cytogenetic and genomic characterization of sex chromosomes in henophidian species. Evolution of two gametologous genes in snakes The CTNNB1 protein is necessary for the adhesive function of cadherins, and has a role in mediating the canonical Wnt signaling pathway and regulating gene transcription [54]. Thus, this gene is ubiquitously expressed in tetrapod species, such as humans, mice, chickens, and X. tropicalis (NCBI UniGene). In the context of the sex determination pathway, ectopic stabilization of this gene causes phenotypic sex reversal from male to female in laboratory mice [55]. Although the exact function of the WAC protein is unknown, it contains a WW domain, a protein module found in a wide range of signaling proteins [56]. The WAC gene is also expressed in many tissues in tetrapod species (NCBI UniGene). Thus, the two genes probably have important functions common among tetrapod species. The snake Z homologs of the two genes, in particular the CTNNB1 gene, highly conserved amino acid sequences, suggesting that they have evolved under a strong purifying selection and retained the common function. Because the E. quadrivirgata W homologs of the two genes have retained complete ORFs, there is no evidence for their existence as pseudogenes. However, the E. quadrivirgata W homologs exhibited fewer amino acid sequence similarities with the homologs of the other tetrapod species. Specific indels were also identified in the amino acid sequences of the W homologs. Thus, the E. quadrivirgata W homologs of CTNNB1 and WAC genes may have acquired somewhat diverged functions. The phylogenetic trees of the two gametologs from the long alignments showed that the divergence between Z and W homologs of E. quadrivirgata occurred soon after the caenophidian homologs diverged from the henophidian ones (Figs. 3a and 4a; Additional files 9a and 10a). The phylogenetic trees of the two genes from the short alignment, which included more taxa, basically fit this interpretation. The clade of Z homologs of viperid, colubrid, and elapid species and that of W homologs of these species split soon after the divergence from the henophidian homologs (Figs. 3b and 4b; Additional files 9b and 10b). These suggest that the differentiation of Z and W homologs of the two genes began in an early caenophidian lineage after the divergence from henophidians. However, the phylogenetic placement of acrochordid Z and W homologs remains uncertain. Some trees clustered the acrochordid Z homolog with henophidian homologs (Fig. 3b and Additional file 10b) and others clustered it with the acrochordid W homolog (Fig. 4b), although an NJ tree (Additional file 9b) pointed to the affinity of the acrochordid Z and W homologs to Z and W homologs of other caenophidians, respectively. However, all these variations lack strong bootstrap supports and are thus unreliable. Taken together, our molecular phylogeny consistently supports the notion that differentiation between Z and W chromosomes began in an ancestral caenophidian lineage, although further data are required to resolve the phylogenetic placement of the acrochordid homologs. Heterochromatinization of W chromosomes in caenophidian snakes The nucleotide sequence and chromosomal location of the EQU-BamHI-4 repetitive sequence are highly conserved in both henophidian and caenophidian species regardless of the degeneration status of their W chromosomes (Fig. 4 in [12], Fig. 5a–c). Therefore, this repetitive sequence was probably accumulated in the telomeric regions of short arms of sex chromosomes in the common ancestor of Henophidia and Caenophidia (Fig. 6).Fig. 6 Evolution of snake sex chromosomes. The timing of evolutionary events on snake sex chromosomes inferred by this study is shown on the cladogram [49, 63]. Horizontal lines between Z and W chromosomes stand for the presence of recombination between the homologs on the chromosomes. Chromosome region with dark gray color stand for amplification of EQU-BglI-15 and (AGAT)n repeats on the W chromosomes in caenophidian species. Note that morphologies of Z and W chromosomes and locations of the EQU-BamHI-4 repeat, CTNNB1 and WAC genes in acrochordid species are not yet identified and that chromosomal locations of the two genes are also not yet identified in viperid species Extensive amplification of the EQU-BglI-15 repetitive sequence was identified on the long arm of W chromosomes in the examined colubrid and viperid species, although an exception was found in R. tigrinus (Fig. 5d–f, Additional file 11a, b). Amplification of the (AGAT)8 microsatellite repeat motif was also identified on the long arm of W chromosomes in all the examined colubrid and viperid species (Fig. 5g–i and Additional file 11c). Although the chromosomal distribution of the two repeats in acrochordid and elapid species was not examined in this study, amplification of the (AGAT)n motif was identified in several colubrid and elapid species [30–33, 57]. The amplification of repetitive sequences has frequently been associated with heterochromatinization of Y and W chromosomes in numerous animal and plant species (e.g., [28, 29, 57–59]). As nearly all colubrid, elapid, and viperid species have heterochromatic W chromosomes [12, 30, 57, 60, 61], our results suggest that heterochromatinization with accumulation of the EQU-BglI-15 and (AGAT)n sequences occurred on the W chromosomes in the common ancestor of Colubridae, Elapidae, and Viperidae (Fig. 6). It was recently reported that the dragonsnake (Xenodermus javanicus) from Xenodermatidae, another basally diverged lineage in the Caenophidia (Fig. 1), has morphologically differentiated Z and W chromosomes, and the amplification of (GATA)n was identified on the highly heterochromatic W chromosome [62]. Thus, heterochromatinization with accumulation of the (AGAT)n sequence probably occurred before the divergence of Xenodermatidae from other caenophidian families. The EQU-BglI-15 sequence does not contain the (AGAT)n repeat in its sequence. Because these two repeats showed differential distributions on the W chromosomes in the four caenophidian species (Fig. 5d–i and Additional file 11), the two repeats could have spread on the caenophidian W chromosomes independently. Differentiation process of sex chromosomes in snakes A schematic model for the differentiation process of sex chromosomes in snakes is shown in Fig. 6. At first, the EQU-BamHI-4 repetitive sequence was accumulated in the terminal regions of the short arms of Z and W chromosomes in the common ancestor of henophidians and caenophidians. The sex chromosomes of henophidian species have retained this little differentiated status to date. The size and morphology of Z chromosomes and gene orders of Z-linked genes are similar between the henophidian P. bivittatus and two caenophidian species, E. quadrivirgata and P. flavoviridis [12], suggesting that snake Z chromosomes have been conserved without large chromosomal rearrangements. In contrast, the W chromosomes of the two caenophidian species have reached a highly differentiated status. Although it is not yet clear whether the Z and W chromosomes are morphologically differentiated in acrochordid species [52, 53], our molecular phylogenetic data suggest that the cessation of recombination between the Z and W homologs of two genes (CTNNB1 and WAC) that map on the centromeric and telomeric regions, respectively, of the long arms, began in the early stage of caenophidian divergence (Fig. 6). According to the predicted divergence times between snake families [63], differentiation of the proto-Z and proto-W chromosomes may have initiated and expanded to wide regions on the sex chromosomes in the common ancestor of caenophidian families during a relatively short period, 80–100 MYA (Fig. 1). A recent study using quantitative PCR (qPCR) on six putative sex chromosome-linked genes in 37 snake species reached a similar conclusion [64]. The qPCR analyses showed that female-to-male relative gene doses of the six genes were approximately 0.5 in all examined caenophidian species, including A. granulatus and X. javanicus, suggesting that the emergence of differentiated sex chromosomes preceded the diversification of caenophidian snakes [64]. The short arm of W chromosomes is extensively degenerated and, to the best of our knowledge, no gametologous gene has been cytogenetically identified there in caenophidian species. It is thus interesting to hypothesize that the degeneration and differentiation processes started earlier on the short arm than on the long arm, although more data is needed to verify this hypothesis. Conclusions We studied the differentiation process of snake sex chromosomes using both coding sequences and repetitive sequences. We analyzed the molecular phylogeny of two gametologous genes, CTNNB1 and WAC, and chromosomal distributions of sex chromosome-linked repetitive sequences in several snake species. Our results suggest that the differentiation between the proto-Z and proto-W chromosomes and heterochromatinization of the proto-W chromosome began in ancestral caenophidian lineages after their divergence from henophidians. However, these results were obtained using only a handful of genes and repetitive elements, and many details of the differentiation process (e.g., where on the W chromosomes the differentiation process initiated, and in which direction it proceeded) are still in question. In this regard, the present study provides solid progress in methodologies and phylogenetic information for further investigation of gametologs in snakes. In the future, genomic and cytogenetic approaches will be accelerated and provide critical information to elucidate the molecular mechanisms of sex chromosome evolution in vertebrates, including snakes. Additional files Additional file 1: Primers used for this study. This table shows information of primers used for PCR amplification of partial sequences of CTNNB1 and WAC genes, and for RACE to determine full-length cDNA sequences of the two genes in E. quadrivirgata. (XLSX 32 kb) Additional file 2: Diagrams for partial gene structure, primer positions and gel electrophoresis of PCR products. This file provides information about primer positions on CTNNB1 and WAC genes, and images of gel electrophoreses for PCR amplicons of partial sequences of CTNNB1 gene in snakes. (PDF 401 kb) Additional file 3: Accession numbers of snake contigs that contain genomic sequences of CTNNB1 and WAC genes. This table shows accession numbers of contigs containing genomic sequences of CTNNB1 and WAC genes of four snake species, P. bivittatus, O. hannah, C. mitchellii and V. berus. (XLSX 33 kb) Additional file 4: Models and parameters used for estimation of molecular phylogenetic trees. This table shows models and parameters used for estimation of molecular phylogenetic trees. (XLSX 44 kb) Additional file 5: Accession numbers of CTNNB1 and WAC gene sequences for snake species determined in this study. This table shows accession numbers of partial or full-length CTNNB1 and WAC gene sequences for snake species determined in this study. (XLSX 35 kb) Additional file 6: Pairwise sequence similarities of CTNNB1 and WAC genes among nine tetrapods. This table shows pairwise sequence similarities of CTNNB1 and WAC genes among nine tetrapods. (XLSX 39 kb) Additional file 7: Alignment of full-length amino acid sequences of CTNNB1 among tetrapod species. This figure shows alignment of full-length amino acid sequences of CTNNB1 among 11 tetrapod species. (PDF 306 kb) Additional file 8: Alignment of full-length amino acid sequences of WAC among tetrapod species. This figure shows alignment of full-length amino acid sequences of WAC among 11 tetrapod species. (PDF 302 kb) Additional file 9: Molecular phylogenic trees of CTNNB1 gene. This figure shows neighbor-joining trees of CTNNB1 gene with the long alignment for 20 tetrapod species and the short alignment for 26 squamate species. (PDF 289 kb) Additional file 10: Molecular phylogenic trees of WAC gene. This figure shows neighbor-joining trees of WAC gene with the long alignment for 21 tetrapod species and the short alignment for 21 squamate species. (PDF 282 kb) Additional file 11: FISH of EQU-BglI-15 repetitive sequence and (AGAT)8 microsatellite motif in E. quadrivirgata and G. blomhoffii. This figure shows FISH images of EQU-BglI-15 repetitive sequence and (AGAT)8 microsatellite motif in E. quadrivirgata and G. blomhoffii. (PDF 814 kb) Abbreviations CTNNB1Catenin (cadherin-associated protein), beta 1, 88 kDa dUTPDeoxyuridine triphosphate EBIThe European bioinformatics institute EMBLEuropean molecular biology laboratory ESTExpressed sequence tag FISHFluorescence in situ hybridization INSDCInternational nucleotide sequence database collaboration NCBINational center for biotechnology information ORFOpen reading frame PCRPolymerase chain reaction PIPropidium iodide RACERapid amplification of cDNA ends UTRUntranslated region WACWW domain containing adaptor with coiled-coil We are deeply grateful to the late Dr. Michihisa Toriba and his colleagues at Japan Snake Institute for providing most snake samples. We also thank Dr. Hidetoshi Ota (Institute of Natural and Environmental Sciences, University of Hyogo), Dr. Seiki Katsuren (Okinawa Prefectural Institute of Health and Environment), and Dr. Shosaku Hattori (Institute of Medical Science, University of Tokyo) for providing samples of I. braminus, L. semicarinatum, and P. flavoviridis, respectively. We also thank Ms. Risako Seki and Dr. Yoshinobu Uno for their help in lab work. Funding This study was supported by Grants-in-Aid for Young Scientists (B) (No. 19770001) and JSPS Fellows (No. 09 J05132) from the Ministry of Education, Culture, Sports, Science and Technology, Japan. KM was supported by JSPS Research Fellowships for Young Scientists (No. 215132). Availability of data The snake CTNNB1 and WAC sequences obtained in this study are available in the INSDC under the accession numbers LC140988 - LC140992, LC140994, LC140996 - LC141003, LC141005 - LC141007, LC141010 - LC141014, LC141017, LC141018, LC141020, LC141023 - LC141025, LC141028, LC141031, LC141033, LC141037, LC141040, LC141043, LC141046, LC141050, LC141053, LC141057, LC141060, LC141064, LC141067, LC141070 and LC141074. Authors’ contributions The authors have made the following declarations about their contributions. KM conceived the research. KM, YM and YK designed the experiments. KM and CN performed the experiments. KM and YK analyzed the data. All authors contributed to materials/analysis tools. KM wrote the first draft and all coauthors contributed in revising it. 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 Animal care and animal experiments were done in accordance with the guidelines of the Animal Experiment Committee of Nagoya City University with permission (No. H21N-02). ==== Refs References 1. Muller HJ A gene for the fourth chromosome of Drosophila J Exp Zool 1914 17 325 336 10.1002/jez.1400170303 2. Ohno S Sex chromosomes and sex-linked genes 1967 Berlin Springer 3. Rice WR The accumulation of sexually antagonistic genes as a selective agent promoting the evolution of reduced recombination between primitive sex chromosomes Evolution 1987 41 911 914 10.2307/2408899 4. Charlesworth B The evolution of sex chromosomes Science 1991 251 1030 1033 10.1126/science.1998119 1998119 5. Charlesworth D Charlesworth B Marais G Steps in the evolution of heteromorphic sex chromosomes Heredity 2005 95 118 128 10.1038/sj.hdy.6800697 15931241 6. Beçak W Beçak ML Nazareth HRS Ohno S Close karyological kinship between the reptilian suborder Serpentes and the class Aves Chromosoma 1964 15 606 617 10.1007/BF00319994 14333153 7. Beçak W Beçak ML Cytotaxonomy and chromosomal evolution in Serpentes Cytogenetics 1969 8 247 262 10.1159/000130037 5381335 8. Singh L Evolution of karyotypes in snakes Chromosoma 1972 38 185 236 10.1007/BF00326193 5066447 9. Beçak W Beçak ML Nazareth HRS Karyotypic studies of two species of South American snakes (Boa constrictor amarali and Bothrops jararaca ) Cytogenetics 1962 1 305 313 10.1159/000129740 13970029 10. Mengden GA Linear differentiation of the C-band pattern of the W chromosome in snakes and birds Chromosoma 1981 83 275 287 10.1007/BF00286795 7273948 11. Matsuda Y Nishida-Umehara C Tarui H Kuroiwa A Yamada K Isobe T Highly conserved linkage homology between birds and turtles: Bird and turtle chromosomes are precise counterparts of each other Chromosome Res 2005 13 601 615 10.1007/s10577-005-0986-5 16170625 12. Matsubara K Tarui H Toriba M Yamada K Nishida-Umehara C Agata K Evidence for different origin of sex chromosomes in snakes, birds, and mammals and step-wise differentiation of snake sex chromosomes Proc Natl Acad Sci U S A 2006 103 18190 18195 10.1073/pnas.0605274103 17110446 13. Matsubara K Kuraku S Tarui H Nishimura O Nishida C Agata K Intra-genomic GC heterogeneity in sauropsids: evolutionary insights from cDNA mapping and GC3 profiling in snake BMC Genomics 2012 13 604 10.1186/1471-2164-13-604 23140509 14. García-Moreno J Mindell DP Rooting a phylogeny with homologous genes on opposite sex chromosomes (gametologs): a case study using avian CHD Mol Biol Evol 2000 17 1826 1832 10.1093/oxfordjournals.molbev.a026283 11110898 15. Skaletsky H Kuroda-Kawaguchi T Minx PJ Cordum HS Hillier L Brown LG The male-specific region of the human Y chromosome is a mosaic of discrete sequence classes Nature 2003 423 825 837 10.1038/nature01722 12815422 16. de Kloet RS de Kloet SR Evolution of the spindlin gene in birds: independent cessation of the recombination of sex chromosomes at the spindlin locus in neognathous birds and tinamous, a palaeognathous avian family Genetica 2003 119 333 342 10.1023/B:GENE.0000003842.72339.df 14686612 17. Handley L-JL Ceplitis H Ellegren H Evolutionary strata on the chicken Z chromosome: implications for sex chromosome evolution Genetics 2004 167 367 376 10.1534/genetics.167.1.367 15166161 18. Lahn BT Page DC Four evolutionary strata on the human X chromosome Science 1999 284 964 967 10.1126/science.286.5441.964 10542153 19. Ellegren H Carmichae A Multiple and independent cessation of recombination between avian sex chromosomes Genetics 2001 158 325 331 11333240 20. Nicolas M Marais G Hykelova V Janousek B Laporte V Vyskot B A gradual process of recombination restriction in the evolutionary history of the sex chromosomes in dioecious plants PLoS Biol 2004 3 e4 10.1371/journal.pbio.0030004 15630476 21. Sandstedt SA Tucker PK Evolutionary strata on the mouse X chromosome correspond to strata on the Human X chromosome Genome Res 2004 14 267 272 10.1101/gr.1796204 14762062 22. Bergero R Forrest A Kamau E Charlesworth D Evolutionary strata on the X Chromosomes of the dioecious plant Silene latifolia : evidence from new sex-linked genes Genetics 2007 175 1945 1954 10.1534/genetics.106.070110 17287532 23. Nam K Ellegren H The chicken (Gallus gallus ) Z chromosome contains at least three nonlinear evolutionary strata Genetics 2008 180 1131 1136 10.1534/genetics.108.090324 18791248 24. Wang J Na JK Yu Q Gschwend AR Han J Zeng F Sequencing papaya X and Yh chromosomes reveals molecular basis of incipient sex chromosome evolution Proc Natl Acad Sci U S A 2012 109 13710 13715 10.1073/pnas.1207833109 22869747 25. Vicoso B Emerson JJ Zektser Y Mahajan S Bachtrog D Comparative sex chromosome genomics in snakes: differentiation, evolutionary strata, and lack of global dosage compensation PLoS Biol 2013 11 e1001643 10.1371/journal.pbio.1001643 24015111 26. Nanda I Schartl M Feichtinger W Epplen JT Schmid M Early stages of sex chromosome differentiation in fish as analysed by simple repetitive DNA sequences Chromosoma 1992 101 301 310 10.1007/BF00346009 1576882 27. Kraemer C Schmidt ER The sex determining region of Chironomus thummi is associated with highly repetitive DNA and transposable elements Chromosoma 1993 102 553 562 10.1007/BF00368348 8243167 28. Hobza R Lengerova M Svoboda J Kubekova H Kejnovsky E Vyskot B An accumulation of tandem DNA repeats on the Y chromosome in Silene latifolia during early stages of sex chromosome evolution Chromosoma 2006 115 376 382 10.1007/s00412-006-0065-5 16612641 29. Kawai A Nishida-Umehara C Ishijima J Tsuda Y Ota H Matsuda Y Different origins of bird and reptile sex chromosomes inferred from comparative mapping of chicken Z-linked genes Cytogenet Genome Res 2007 117 92 102 10.1159/000103169 17675849 30. Singh L Purdom IF Jones KW Satellite DNA and evolution of sex chromosomes Chromosoma 1976 59 43 62 10.1007/BF00327708 1001165 31. Singh L Purdom IF Jones KW Sex chromosome associated satellite DNA: evolution and conservation Chromosoma 1980 79 137 157 10.1007/BF01175181 7398495 32. Epplen JT McCarrey JR Sutou S Ohno S Base sequence of a cloned snake W-chromosome DNA fragment and identification of a male-specific putative mRNA in the mouse Proc Natl Acad Sci U S A 1982 79 3798 3802 10.1073/pnas.79.12.3798 6954524 33. O’Meally D Patel HR Stiglec R Sarre SD Georges A Graves JAM Non-homologous sex chromosomes of birds and snakes share repetitive sequences Chromosome Res 2010 18 787 800 10.1007/s10577-010-9152-9 20734128 34. Sambrook J Fritsch EF Maniatis T Molecular cloning: a laboratory manual 1989 2 New York Cold Spring Harbor Laboratory Press 35. Sievers F Wilm A Dineen D Gibson TJ Karplus K Li W Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega Mol Syst Biol 2011 7 539 10.1038/msb.2011.75 21988835 36. Cox MP Peterson DA Biggs PJ SolexaQA: At-a-glance quality assessment of Illumina second-generation sequencing data BMC Bioinformatics 2010 11 485 10.1186/1471-2105-11-485 20875133 37. Grabherr MG Haas BJ Yassour M Levin JZ Thompson DA Amit I Full-length transcriptome assembly from RNA-seq data without a reference genome Nat Biotechnol 2011 29 644 652 10.1038/nbt.1883 21572440 38. Thompson JD Higgins DG Gibson TJ CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice Nucleic Acids Res 1994 22 4673 4680 10.1093/nar/22.22.4673 7984417 39. Tamura K Stecher G Peterson D Filipski A Kumar S MEGA6: Molecular Evolutionary Genetics Analysis version 6.0 Mol Biol Evol 2013 30 2725 2729 10.1093/molbev/mst197 24132122 40. Swofford DL PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4 2002 Sunderland, Massachusetts Sinauer Associates 41. Bazinet AL Zwickl DJ Cummings MP A gateway for phylogenetic analysis powered by grid computing featuring GARLI 2.0 Syst Biol 2014 63 812 818 10.1093/sysbio/syu031 24789072 42. Guindon S Gascuel O A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood Syst Biol 2003 52 696 704 10.1080/10635150390235520 14530136 43. Darriba D Taboada GL Doallo R Posada D jModelTest 2: more models, new heuristics and parallel computing Nat Methods 2012 9 772 10.1038/nmeth.2109 22847109 44. Georges A Li Q Lian J O'Meally D Deakin J Wang Z High-coverage sequencing and annotated assembly of the genome of the Australian dragon lizard Pogona vitticeps Giga Sci 2015 4 45 10.1186/s13742-015-0085-2 45. Matsuda Y Chapman VM Application of fluorescence in situ hybridization in genome analysis of the mouse Electrophoresis 1995 16 261 272 10.1002/elps.1150160142 7774567 46. Matsubara K Knopp T Sarre SD Georges A Ezaz T Karyotypic analysis and FISH mapping of microsatellite motifs reveal highly differentiated XX/XY sex chromosomes in the pink-tailed worm-lizard (Aprasia parapulchella , Pygopodidae, Squamata) Mol Cytogenet 2013 6 60 10.1186/1755-8166-6-60 24344753 47. Matsubara K Sarre SD Georges A Matsuda Y Graves JAM Ezaz T Highly differentiated ZW sex microchromosomes in the Australian Varanus species evolved through rapid amplification of repetitive sequences PLoS One 2014 9 e95226 10.1371/journal.pone.0095226 24743344 48. Hedges SB Kumar S The Timetree of Life 2009 New York Oxford University Press 49. Pyron RA Burbrink FT Wiens JJ A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes BMC Evol Biol 2013 13 93 10.1186/1471-2148-13-93 23627680 50. Zheng Y Wiens JJ Combining phylogenomic and supermatrix approaches, and a time-calibrated phylogeny for squamate reptiles (lizards and snakes) based on 52 genes and 4162 species Mol Phylogenet Evol 2016 94 537 547 10.1016/j.ympev.2015.10.009 26475614 51. Streicher JW Wiens JJ Phylogenomic analyses reveal novel relationships among snake families Mol Phylogenet Evol 2016 100 160 169 10.1016/j.ympev.2016.04.015 27083862 52. Sharma GP Nakhasi U Chromosomal polymorphism in three species of Indian snakes Cytobios 1979 24 167 179 527373 53. Sharma GP Nakhasi U Karyological studies on six species of Indian snakes (Colubridae: Reptilia) Cytobios 1980 27 177 192 7428439 54. Huber AH Nelson WJ Weis WI Three-dimensional structure of the armadillo repeat region of ß-catenin Cell 1997 90 871 882 10.1016/S0092-8674(00)80352-9 9298899 55. Maatouk DM DiNapoli L Alvers A Parker KL Taketo MM Capel B Stabilization of ß-catenin in XY gonads causes male-to-female sex-reversal Hum Mol Genet 2008 17 2949 2955 10.1093/hmg/ddn193 18617533 56. Xu GM Arnaout MA WAC, a novel WW domain-containing adapter with a coiled-coil region, is colocalized with splicing factor SC35 Genomics 2002 79 87 94 10.1006/geno.2001.6684 11827461 57. Matsubara K O'Meally D Azad B Georges A Sarre SD Graves JAM Amplification of microsatellite repeat motifs is associated with the evolutionary differentiation and heterochromatinization of sex chromosomes in Sauropsida Chromosoma 2016 125 111 123 10.1007/s00412-015-0531-z 26194100 58. Cioffi MB Camacho JPM Bertollo LAC Repetitive DNAs and differentiation of sex chromosomes in neotropical fishes Cytogenet Genome Res 2011 132 188 194 10.1159/000321571 21042005 59. Ezaz T Azad B O’Meally D Young MJ Matsubara K Edwards MJ Sequence and gene content of a large fragment of a lizard sex chromosome and evaluation of candidate sex differentiating gene R-spondin 1 BMC Genomics 2013 14 899 10.1186/1471-2164-14-899 24344927 60. Toriba M Geographic variation of W-chromosome in Rhabdophis tigrinus (Boie) Snake 1987 19 1 4 61. Matsubara K Uno Y Srikulnath K Seki R Nishida C Matsuda Y Molecular cloning and characterization of satellite DNA sequences from constitutive heterochromatin of the habu snake (Protobothrops flavoviridis , Viperidae) and the Burmese python (Python bivittatus , Pythonidae) Chromosoma 2015 124 529 539 10.1007/s00412-015-0529-6 26205503 62. Rovatsos M Pokorná MJ Kratochvíl L Differentiation of sex chromosomes and karyotype characterisation in the dragonsnake Xenodermus javanicus (Squamata: Xenodermatidae) Cytogenet Genome Res 2015 147 48 54 10.1159/000441646 26575989 63. Vidal N Rage J-C Couloux A Hedges SB Hedges SB Kumar S Snakes (Serpentes) The timetree of life 2009 New York Oxford University Press 390 397 64. Rovatsos M Vukić J Lymberakis P Kratochvíl L Evolutionary stability of sex chromosomes in snakes Proc R Soc B 2015 282 20151992 10.1098/rspb.2015.1992 26702042 65. Uetz P Hošek J The Reptile Database 2016 66. Ota H Hikida T Matsui M Mori A Wynn AH Morphological variation, karyotype and reproduction of the parthenogenetic blind snake, Ramphotyphlops braminus, from the insular region of East Asia and Saipan Amphibia-Reptilia 1991 12 181 193 10.1163/156853891X00158 67. Cole CJ Dowling HG Chromosomes of the sunbeam snake, Xenopeltis unicolor Reinwardt (Reptilia: Xenopeltidae) Herpetological Review 1970 2 35 36 68. CHROMOREP. A reptiles chromosomes database. http://chromorep.univpm.it/. Accessed 28 June 2016. 69. Ray-Chaudhuri SP Singh L Sharma T Evolution of sex-chromosomes and formation of W-chromatin in snakes Chromosoma 1971 33 239 251 10.1007/BF00284942 4326042 70. Toriba M Olmo E Karyotypes of some species of Japanese colubrid snakes Cytogenetics of amphibians and reptiles 1990 Basel Birkhäuser Verlag 255 259

PMC005xxxxxx/PMC5002184.txt

==== Front BMC MicrobiolBMC MicrobiolBMC Microbiology1471-2180BioMed Central London 81510.1186/s12866-016-0815-3Methodology ArticleA method for assessing efficiency of bacterial cell disruption and DNA release de Bruin Olle M. olle.debruin@dnagenotek.com Birnboim H. C. chaim.birnboim@dnagenotek.com DNA Genotek, Ottawa, ON Canada 26 8 2016 26 8 2016 2016 16 1 19725 3 2016 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 DNA-based testing is becoming the preferred method both for identifying microorganisms and for characterizing microbial communities. However, no single DNA extraction method exists that is suitable for all types of microorganisms because bacteria are variable in their susceptibility to lysis by available extraction procedures. To develop and test new DNA extraction procedures, it would be helpful to determine their efficiencies. While the amount of extracted DNA can readily be measured by different methods, calculation of true efficiency requires knowledge of the initial amount of DNA in the starting bacterial sample, which cannot be done with precision by any existing method. In the process of developing a new extraction procedure, we developed a method that can be used to determine the total amount of both DNA and RNA in bacteria. The amount of DNA can be calculated from the amount of purines released after mild acid and alkali treatment. The amount of RNA in the same extract can also be calculated from the amount of ribonucleoside monophosphates. The released purines and ribonucleoside monophosphates can be quantified by absorbance using HPLC, with reference to appropriate standards. Results The acid/HPLC method was used to measure the efficiency of commonly used bead-beating and chemical protocols for releasing DNA from a particularly hardy organism, Mycobacterium smegmatis as well as several other species (Bacillus subtilis vegetative cells and spores; Francisella philomiragia; Pseudomonas aeruginosa; Moraxella catarrhalis; Bacillus thuringiensis; Staphylococcus aureus). Surprisingly large differences in efficiency between methods were found. Conclusions The acid/HPLC method is a new tool to determine DNA extraction efficiencies and should aid in the development of improved protocols for releasing DNA from a broad range of microorganisms. Electronic supplementary material The online version of this article (doi:10.1186/s12866-016-0815-3) contains supplementary material, which is available to authorized users. http://dx.doi.org/10.13039/501100004489Mitacsissue-copyright-statement© The Author(s) 2016 ==== Body Background Nucleic acid testing (NAT) is an important alternative to traditional culture methods for identifying microorganisms of clinical relevance. Culture methods take days to weeks to provide clinically useful information whereas nucleic acid testing can provide information, such as strain and drug resistance, in few hours [1]. Sensitivity of NAT can potentially match culture methods if DNA of sufficient quantity (and purity) can be recovered from the biospecimen. The quantity of DNA may be limiting in cases where the number of microorganisms in the original biospecimen is very low or where the volume of the specimen is small. An additional factor that affects the quantity of DNA available for analysis is the efficiency of DNA release by the extraction procedure from bacteria or other microorganisms. For example, it is especially difficult to release DNA from bacteria such as Mycobacterium spp. To extract DNA from such microorganisms, a combination of physical (heat), mechanical (sonication, bead-beating) and/or chemical (pH, detergents) methods have been used [2–8]. However, studies of this type have all lacked a precise way to determine the efficiency of DNA release. High efficiency of DNA release may be expected to be particularly important in microbiome research, since profiling would benefit from obtaining DNA from all microorganisms equally in a biospecimen. Without equal efficiency of extraction, microbes that are hard to lyse may be underrepresented [2, 9]. High efficiency of DNA extraction is also desirable for NAT-based diagnosis of infectious diseases caused by difficult-to-lyse bacteria such as Mycobacterium spp., and for identification of bio-threat agents, including bacterial spores [10, 11]. To calculate the efficiency of extraction, the amount of DNA recovered and the total amount of DNA in the original sample must both be known. There is currently no generally accepted method for accurately quantifying the amount of DNA in a suspension or pellet of intact bacteria. Indirect methods for estimating cellular DNA content are based on microscopic cell counts or colony counts. However, these require accurate knowledge of the amount of DNA per cell, which is difficult to assess because the amount of DNA varies with the number of genome copies, which in turn is a function of growth rate. Counting of cells is non-trivial since many bacteria have a tendency to form clumps or chains. For these reasons, it has been very difficult to compare the efficiency of DNA released by different published extraction procedures. In this report, we describe an acid/HPLC method that uses simple chemical principles to estimate the amount of DNA and RNA in a suspension of microorganisms (see diagram below) and compare it to existing methods. Our method is based upon selective acid-catalysed depurination of DNA [12, 13]. The quantity of DNA can be calculated from the quantity of purines released, provided depurination of DNA is complete and depurination of RNA is very low. Estimation of RNA in the same sample is also possible. Treatment of RNA under mild alkaline conditions causes its degradation to ribonucleoside monophosphates, with no degradation of DNA. Purines released from DNA and ribonucleoside monophosphates from RNA can be quantified by absorbance using HPLC, using appropriate reference standards. We present evidence that the acid/HPLC technique can provide an accurate measure of the total amount of DNA in a bacterial sample, allowing a comparison of DNA release efficiency by any of the DNA extraction protocols in current use. We report surprisingly large differences between different cell disruption techniques. Although our method was developed primarily to quantify DNA in bacteria, the technique should be also applicable to other types of biological samples. Methods Reagents Materials were purchased from the indicated suppliers as follows: Canine DNA (EMD Millipore, Billerica, MA, USA); adenosine 2’(3’)-monophosphate mixed isomers, adenosine 3’-monophosphate, lysozyme, pluronic F-68, reduced Triton X-100, poly(A) and yeast RNA (ribonucleic acid type VI) (Sigma); ADA buffer, guanine and adenine (Alfa Aesar); guanidine hydrochloride (Amresco, Solon, OH, USA). Yeast RNA was treated with deoxyribonuclease to remove traces of DNA. 1.0 N HCl and 1.0 N NaOH were from Ricca Chemical. CDTA (cyclohexanediamine tetraacetate) was from GFS Chemicals. Bacteria and growth conditions Mycobacterium smegmatis (Trevisan) Lehmann and Neumann (ATCC 700044), Bacillus subtilis, Bacillus thuringiensis (ATCC 10792), Pseudomonas aeruginosa (ATCC 10145) and Staphylococcus aureus were grown on tryptic soy agar (TSA) plates at 37 °C. Liquid cultures of bacteria were grown in tryptic soy broth (TSB) at 37 °C on a shaking incubator. Moraxella catarrhalis (ATCC 25238) was grown on brain heart infusion agar and broth at 37 °C. Francisella philomiragia (ATCC 25017) was grown on TSA plates supplemented with 0.1 % cysteine and in TSB supplemented with 0.1 % cysteine at 37 °C. Spore preparation B. subtilis spores were grown in 1/10th strength Columbia broth (Difco, Sparks, MD, USA), supplemented with 100 μM manganese sulfate, for 3 days at 37 °C, as described by others [14]. Spores were harvested, washed and incubated with lysozyme at 37 °C for 1 h to remove residual vegetative cells. Weakened cells were lysed in 0.1 % sodium dodecyl sulfate (SDS). Following another wash step, DNA attached to spore cell walls was removed by treatment with DNase (10 μg/ml in 4 mM MgCl2, 1 mM CaCl2, 10 mM Tris–HCl, pH 7.5). DNase was inactivated at 75 °C for 10 min, and purity of the spore preparation was verified by microscopic examination following Schaeffer-Fulton staining [15]. Standard Acid/Alkali-treatment of bacterial cells and extracts Bacteria from liquid or agarose plate cultures were washed by centrifugation in cold phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 10 mM sodium phosphate, 1.8 mM potassium phosphate, pH 7.4). Washed pellets were well-suspended in 320 μl of water; 80 μl of 1.0 N HCl was then added and the suspension mixed. Samples were incubated for 60 min at 60 °C in a water bath, vortexing at 0, 30 and 60 min. To each sample, 133.3 μl of 1.0 N NaOH was added (80 μl to neutralize the HCl and 53.3 μl to bring the final concentration to 0.1 N). Samples were heated at 100 °C for 10 min to hydrolyze RNA to ribonucleoside 2’- and 3’-monophosphates. Samples were centrifuged at 20,817 g for 5 min to remove insoluble material. A portion (400 μl) of each supernatant was removed. Samples were brought to neutral pH by addition of 40 μl of 1.0 N HCl and 160 μl of 0.4 M ADA buffer, pH 6.6, following which they were loaded onto a HPLC column. HPLC Samples were analyzed using a Perkin Elmer HPLC system comprising a series 200 UV/VIS detector, a series 200 pump and a series 225 autosampler. The column was a reverse phase Gemini-NX-C18 with 3 μm particles (Phenomenex, Inc.). The isocratic solvent contained 2.0 % (w/v) methanol in 30 mM ammonium acetate, 1 mM CDTA, 10 mM NaH2PO4, pH 6.3. Pump speed was 0.5 ml/min, and detector wavelength was 260 nm. Up to 40 μl of sample could be injected, depending upon the expected concentration of the analyte of interest. Area under the curve (AUC) was calculated using TotalChrom Navigator version 6.3.2 software (Perkin Elmer). AUC was converted to amount of analyte from a standard curve of the analyte of interest. Preparation of adenine HPLC standards A solution of adenine used as a standard for HPLC was prepared as follows. Approximately 2.5 mg of adenine (Alfa Aesar) was mixed with 100 mL of 10 mM HCl. The mixture was heated at 50 °C for 60 min, then filtered through a 0.22 μM membrane. The concentration of adenine was calculated using 13,200 as the molar extinction coefficient at 262.5 nm under acidic conditions [16]. Serial dilutions in 10 mM HCl were prepared as required. Calculation of DNA and extraction efficiency from the amount of adenine or adenine + guanine In this report, we focus on pure cultures of microorganisms whose DNA base composition is known. The quantity of adenine released from the cells by treatment with acid and alkali, as described above, can be used to calculate the quantity of DNA present in cells. However, if the base composition is not known or if a mixed population of cells of different base compositions is analyzed, the quantity of both adenine and guanine must be determined to calculate the amount of DNA. A spreadsheet for converting nanomoles or nanograms of adenine or adenine + guanine to nanograms of DNA is provided as Additional file 1. Under our defined conditions (see Methods), guanine peaks appear at the 6 min mark of HPLC profiles of acid/alkali treated samples (see Fig. 3), and the amount of guanine can be determined using guanine standards of known concentration. Once the amount of DNA in samples prior to and after treatment with a lysis method has been calculated (either using adenine or adenine + guanine), the extraction efficiency of a lysis method then equals the amount of DNA released from the mixed population by the lysis method divided by the total amount of DNA in the mixed population. Bead-beating Two ml of a stationary phase culture of bacteria, approximately 2 × 109 cells/ml, was harvested by centrifugation, washed twice and resuspended in 1.5 ml of cold PBS. The suspension was distributed equally into three screw-cap polypropylene 1.5 ml vials (Simport, Beloeil, Quebec, Canada) containing approximately 100 μl of 100 μm glass beads (Polyscience, Inc., Warrington, PA, USA). Bead-beating was performed in a Mini-Beadbeater-16 (Biospec) in two cycles of 1 min at 3,450 oscillations/min, with a 1 min period of cooling on ice between cycles. Each sample was transferred to a fresh 1.5 ml microcentrifuge tube and centrifuged at 20,817 g for 5 min to remove unbroken cells and debris. 320 μl of the supernatant was removed and treated with standard acid/alkali, as described above. Enumeration of bacteria by plate counting Bacillus subtilis were grown in tryptic soy broth to mid-logarithmic phase and harvested by centrifugation. Cells were washed twice in cold tris-buffered saline and divided equally into six tubes. Three aliquots were serially diluted into TSB and plated on TSA. After incubation for 18 h at 37 °C, colony-forming units (CFUs) were enumerated. Bacterial pellets in the remaining three tubes were treated with standard acid/alkali as described above. Plate counting to enumerate M. smegmatis followed a similar procedure. CFU counts were converted to DNA content for comparison to acid/HPLC results based upon reported genome sequences (http://www.ncbi.nlm.nih.gov/genome/). For B. subtilis, its genome size is 4.2 Mb, which corresponds to 4.9 μg/109 cells; for M. smegmatis, its genome size is 7.1 Mb, which correspond to 8.2 μg/109 cells. Lysis of vegetative Bacillus subtilis cells with lysozyme and SDS B. subtilis was grown to exponential phase in LB broth and harvested by centrifugation. Cells were washed and resuspended in ice-cold TE. The suspension was divided equally into tubes and again centrifuged. Two of the pellets were suspended in 950 μl TE containing lysozyme (1 mg/ml), followed by incubation for 45 min at 37 °C. SDS (final concentration of 0.5 %) was added to ensure complete lysis of the cells. KCl (0.1 N) was added; after incubation on ice for 5 min, the precipitate was removed by centrifugation at 20,817 g for 10 min. An aliquot of the supernatant was transferred to a fresh tube and subjected to standard acid/alkali treatment. The remaining two pellets were suspended directly in 500 μl water and subjected to standard acid/alkali treatment. DNA extraction from M. smegmatis M. smegmatis were scraped from four TSA plates, washed twice in ice-cold water and suspended in 50 mM tris pH 8, 1 mM CDTA. For bead-beating in the presence of detergent, SDS or pluronic F-68 was added to a final concentration of 1.0 %. Bead-beating was performed as described above. Boiling was performed in a water bath for 10, 20 or 30 min. Freeze-boil was performed by placing tubes containing bacterial suspensions at −20 °C for 10 min, followed by heating in a boiling water bath for 10 min. Following each treatment, all samples were cleared by centrifugation for 5 min at 20,817 g, and then subjected to standard acid/alkali treatment. To test the effectiveness of guanidine hydrochloride as a DNA extraction reagent, a previously described procedure was followed [17]. M. smegmatis, scraped from TSA plates, was washed three times with cold PBS and distributed equally into 12 tubes. Pellets were suspended in triplicate in 200 μl of either (i) 8 M guanidine hydrochloride, 2 % reduced triton X-100, 80 mM Tris–HCl, pH 8.0, 40 mM CDTA, (ii) 2 % reduced triton X-100, 80 mM Tris–HCl, pH 8.0, 40 mM CDTA or (iii) water, then heated at 100 °C for 10 min. After heating, samples were centrifuged at 20,817 g for 5 min to remove unbroken cells and debris. Following centrifugation, the supernatant was removed and subjected to standard acid/alkali treatment as described above. Standard acid/alkali treatment of the untreated cell pellets was also carried out. The remaining three pellets were subjected directly to standard acid/alkali treatment. Results Selective release of adenine from DNA as compared to RNA DNA is degraded on exposure to relatively mild acidic conditions in two separate steps. First, purines are readily released from deoxyribose while pyrimidine-deoxyribose linkages are highly resistant. Subsequently, the resultant apurinic acid is cleaved by β-elimination to yield a series of pyrimidine isostichs [12, 13, 18]. Alkali treatment of intact, double-stranded DNA causes denaturation without degradation, but generates ribonucleoside monophosphates from RNA. Purines (adenine and guanine) and adenine ribonucleoside monophosphates (2’-AMP and 3’-AMP) can be separated and quantified by isocratic HPLC (Fig. 1a).Fig. 1 (a). Separation of purines and ribonucleoside monophosphates by reverse phase HPLC. Curve (a), adenine; curve (b), guanine; curve (c), mixture of adenine, guanine, adenosine 2’-phosphate and adenosine 3’-phosphate; curve (d), mixture of adenosine 2’-phosphate and adenosine 3’-phosphate. (b). Release of adenine from pure DNA by acid hydrolysis. The DNA was incubated in 0.15 N or 0.20 N HCl at 60 °C for 20, 40 or 60 min, then the amount of released adenine was quantified by HPLC. The quantity of DNA (calculated from the amount of adenine) was compared to the amount specified by the manufacturer (Novagen). For the 0.15 N treatment (solid circle), the amounts were 34.3, 86.7 and 94.5 %, respectively. For the 0.20 N treatment (solid square), the amounts were 99.3, 103.3 and 103.3 %, respectively. Error bars represent range of duplicate samples; where not shown, the values are within the symbol. (c). Selective release of adenine from DNA compared to RNA. DNA (curve a), RNA (curve b) and poly(rA) (curve c), were subjected to standard acid/alkali treatment, then analysed by HPLC. The expected peaks of guanine and adenine from DNA, adenosine 3’-phosphate and adenosine 2’-phosphate from poly(rA) and adenosine 3’-phosphate, adenosine 2’-phosphate and guanosine 2’-phosphate from RNA can be seen. In addition, tiny peaks of adenine can be seen in the poly(rA) and RNA samples, representing 0.25 and 1.63 % of the total adenosine phosphate from poly(A) and yeast RNA, respectively. Peaks in this and subsequent figures are identified by the following numbers: 1, adenosine 5’-phosphate; 2, guanine; 3, guanosine phosphate; 4, adenosine 3’-phosphate; 5, adenine; 6, adenosine 2’-phosphate. The guanosine phosphate isomer (Peak 3) is likely to be the 2’-phosphate, based on the order of elution of the adenosine monophosphates in our isocratic HPLC system and of the guanosine and adenosine phosphates on an ion-exchange system [22]. Absorbance at 260 nm in milliabsorbance units; adenine (AUC), Area Under the Curve for adenine in milliabsorbance units To determine conditions of acid treatment that produce complete depurination of DNA without depurination of RNA, we subjected DNA to treatment with either 0.15 N or 0.20 N HCl at 60 °C for periods up to 60 min. Release of adenine was quantified by HPLC. The data presented in Fig. 1b shows that near-maximum release of adenine occurred at 60 min in 0.15 N HCl and that maximum release of adenine occurred at 40 min and 60 min in 0.20 N HCl. We chose 0.2 N HCl for 60 min at 60 °C as standard acid treatment for all subsequent experiments. We next determined the effect of the same acid conditions on the release of purines from RNA. Yeast RNA and poly(rA) were treated with standard acid and alkali conditions and analyzed by HPLC (Fig. 1c). A very small peak of free adenine can be detected, which represents 0.25 % of the total adenine nucleotide from poly(rA) and 1.63 % from yeast RNA. These experiments demonstrate that, under the conditions described, release of purines from DNA is virtually complete while <2 % of purines are released from RNA. This difference is the basis for our acid/HPLC method to quantify DNA. Knowing the quantity of adenine (or adenine + guanine) allows calculation of the amount of DNA in a sample (see Methods and Additional file 1. To quantify RNA as well as DNA, we introduced an alkali treatment step after the acid treatment step to complete the partial hydrolysis of RNA (e.g., oligoribonucleotides, cyclic 2′,3’-nucleoside monophosphates) to ribonucleoside 2’-monophosphates and ribonucleoside 3’-monophosphates. The standard acid/HPLC protocol, as described in the Methods, includes both an acid step and an alkali step. Comparison of acid/HPLC with other methods used to quantifying the amount of DNA in a sample of bacteria As indicated above, there is no standard method for estimating the total amount of DNA in intact bacteria. We therefore compared our acid/HPLC method with three other methods that might be used for estimating the total amount of DNA in a sample of intact bacteria. The first two methods involve disrupting cells with either lysozyme (an enzyme suitable for gram-positive cells) or bead-beating (mechanical disruption by vigorous shaking with glass beads). Assuming complete disruption of the cells, released DNA can then be measured and the total amount of DNA estimated. For example, lysozyme treatment followed by SDS is a very effective disruption method of vegetative B. subtilis cells that should lead to release of all, or nearly all, DNA in the bacteria. A third way of estimating the amount of DNA in a sample of bacteria is enumeration of cell number using viable colony counts. The amount of DNA present in bacterial suspensions of B. subtilis and Mycobacterium smegmatis as determined directly by the acid/HPLC method is shown in Fig. 2. These values were compared to the amount of DNA released from disrupted cells. Disruption by lysozyme treatment was used for vegetative B. subtilis cells and bead-beating was used for M. smegmatis and B. subtilis (vegetative cells and spores). After centrifugation to remove debris and any unbroken cells, the amount of DNA in the supernatant of the disrupted cells was measured using acid/HPLC. In Fig. 2, ‘HCl’ represents the amount of DNA in ‘intact’ bacterial cells, measured directly using acid/HPLC. The results show that, in all cases except spores, the amount of DNA detected in intact cells by acid/HPLC was (within experimental error) the same as the amount of DNA released from cells first completely disrupted by lysozyme or bead-beating. For spores, acid/HPLC detected about seven times more DNA than detected in the cell supernatant after bead-beating (Fig. 2). Bacterial spores are known to represent a particularly difficult challenge for release of DNA [19, 20]. These results support the notion that the direct acid/HPLC method is detecting essentially all DNA in these microorganisms.Fig. 2 Comparison of the amount of DNA in intact cells, as detected by acid/HPLC, with the amount of DNA released after different disruption methods. Disruption methods used were lysozyme (Lys) for B.vegetative cells and bead-beating (BB) for B. subtilis spores and M. smegmatis. In each experiment, an aliquot of a cell suspension(6x108 B. subtilis vegetative or 8x108 B. subtilis spores, or 2x109 cells of M. smegmatis) was treated either by (i) the standard acid/alkali method (HCl) or (ii) disrupted as indicated. After disruption, lysates were centrifuged to remove debris, then treated by the standard acid/alkali method to release adenine from the released DNA. The amount of adenine was determined by HPLC. The mean and range of duplicate samples are shown. The right-hand axis refers to the spores samples. Adenine (AUC), Area Under the Curve for adenine in milliabsorbance units Plate counting was the third method tested as a comparison with acid/HPLC. To convert colony counts (CFUs) to total DNA requires knowledge of the amount of DNA per cell, as calculated from the genome sequence of the target organism. For B. subtilis, its genome size is 4.2 Mb, which corresponds to 4.9 μg/109 cells; for M. smegmatis, its genome size is 7.1 Mb, which correspond to 8.2 μg/109 cells (http://www.ncbi.nlm.nih.gov/genome/). Table 1 shows DNA content estimated from colony counts compared to DNA measured by acid/HPLC. In both cases, the estimated amount by CFU is less than the measured amount (81 % for B. subtilis and 38 % for M. smegmatis). A spread sheet for converting nanomoles or nanograms of adenine to nanograms of DNA is provided as Additional file 1 (see also Methods section).Table 1 Comparison of CFU and acid/HPLC to quantify the amount of DNA in bacteria Microorganism DNA content (μg) based on CFU DNA content (μg) based on acid/HPLC B. subtilis 1.32 ± 0.034 (n = 3)a 1.62 ± 0.029 (n = 3) M. smegmatis 6.10 ± 0.906 (n = 5)b 15.91 ± 0.338 (n = 5) aAssumes that Bacillus subtilis contains 4.9 μg DNA/109 cells bAssumes that Mycobacterium smegmatis contains 8.2 μg DNA/109 cells Examples of acid/alkali-treated extracts of B. subtilis, E. coli, M. smegmatis and baker’s yeast Figure 3 shows HPLC profiles obtained after acid/alkali treatment of pure cultures of microorganisms. Qualitative information about the nucleic acids in each organism can be obtained from inspection of the profiles. For example, the ratio of peak 5 (adenine) and peak 6 (ribonucleoside 2’-monophosphate) reflects the approximate relative amount of DNA and RNA, respectively, in the four different microbial species. Of the organisms tested, M. smegmatis had the lowest RNA/DNA ratio, E. coli had a slightly higher ratio and B. subtilis and S. cerevisiae had the highest ratios. The GC content of the different microorganisms can also be estimated from the relative guanine (peak 2) to adenine (peak 5) ratio. Reported GC contents for these organisms are 43.5 % for B. subtilis, 38.2 % for S. cerevisiae, 50.8 % for E. coli and 66.5 % for M. smegmatis (www.ncbi.nlm.nih.gov/genome/). Although a precise GC content was not determined in this experiment, the high gua/ade ratio of M. smegmatis can be seen in panel D.Fig. 3 Examples of HPLC profiles of acid/alkali extracts of four different types of microorganisms. Washed cell pellets, prepared as described in the Methods, were subjected to the standard acid/HPLC protocol. a S. cerevisiae (baker’s yeast); b E. coli (gram-negative bacterium); c B. subtilis (gram-positive bacterium); d M. smegmatis, (bacterium with a waxy cell wall). Numbers near the peaks refer to the identity of known components, listed in the legend to Fig. 1 Differences in the amount of DNA released from Mycobacteria smegmatis by different extraction methods Having demonstrated that the acid/HPLC method is suitable for quantifying DNA in intact bacteria, we used it to assess the efficiency of DNA release by a variety of extraction methods. An overview of the approach is shown in Fig. 4. For this experiment, we chose M. smegmatis as a target because it is a generally accepted model of a difficult-to-lyse bacterium. Bead-beating is a mechanical disruption step included in many DNA extraction protocols from Mycobacterium spp. As shown in Fig. 5a, bead-beating released about 90 % of the DNA. However, when bead-beating was preceded by heating at 100 °C for 10 min, only about 30 % of the DNA was released. The order of treatment mattered; when bead-beating was followed by heating, there was no decrease in the amount of DNA released (Fig. 5c). There was a similar diminution in the amount of DNA released when beading-beating was carried out in the presence of SDS (Fig. 5a) and triton X (Fig. 5c), but no decrease in the presence of another detergent, pluronic F-68 (Fig. 5c). Heating alone at 100 °C for 30 min or 1–2 freeze-boil cycles released 7–8 % of the total DNA from M. smegmatis (Fig. 5a).Fig. 4 Schema showing use of the acid/HPLC method to determine DNA extraction efficiency by different methods. Equal aliquots of a bacterial suspension are centrifuged and the washed pellets are treated with standard acid/alkali or subjected to an extraction method (e.g., method #1. method #2, etc.). The extract is then treated by acid/alkali. The amount of adenine in the extract is compared to the amount of adenine from the pellet of intact cells. The efficiency of DNA extraction is given by the ratio of adenine (in extract) divided by adenine (from intact cells) Fig. 5 Comparison of the efficiency of DNA release from M. smegmatis cells by procedures in current use. A and B are results from two separate experiments. Acid in A and B indicates standard treatment with acid/alkali. a DNA release from 6×109 M. smegmatis by bead-beating and boiling. > initial treatment was followed by a second treatment as indicated after the symbol. BB, bead-beating; SDS + BB, bead-beating of cells suspended in 1.0 % SDS; boil, suspension heated at 100 °C for the indicated time; Freeze-boil, cell suspension frozen at −20 °C, then heated at 100 °C for 10 min; ×2, cycle repeated once. b DNA release from 9×109 M. smegmatis by guanidine hydrochloride. Cell suspension heated at 100 °C for 10 min in: GuHCl + rTX100, 8 M guanidine hydrochloride, 2 % reduced triton X-100, 80 mM Tris–HCl, 40 mM CDTA, pH 8.0; rTX100, 2 % reduced triton X-100, 80 mM Tris–HCl, 40 mM CDTA, pH 8.0; Water, water. c DNA released from M. smegmatis by bead-beating in combination with boiling or detergents. BB, bead-beating; boil, suspension heated at 100 °C for 10 min; rTX100, 2 % reduced triton X-100; 1 % F68, pluronic F-68. See Methods for additional details. Error bars represent the mean ± range of duplicate samples. Both ‘rT100’ and ‘boil > BB’ are significantly different from ‘BB’ (P < 0.05, unpaired t-test) In the experiment shown in Fig. 5b, heating a suspension of M. smegmatis in water at 100 °C for 10 min released 4.0 % release of total DNA. Guanidine hydrochloride, a chaotropic agent commonly used for DNA extractions, is often used in combination with triton X-100 in Mycobacterium extraction protocols. Including these reagents improved extraction to 7.0 %. Surprisingly, the contribution of guanidine hydrochloride to lysis efficiency was negligible, since heating with triton X-100 alone released 6.6 % of the total DNA. Overall, our results demonstrate that commonly used DNA extraction methods can vary greatly in their ability to release DNA from a difficult-to-lyse bacterium. The acid/HPLC method is applicable to a broad-range of bacteria To demonstrate that the acid/HPLC method is applicable to a variety of bacterial species, we also treated whole cells of Francisella philomiragia, Bacillus thuringiensis, Moraxella catarrhalis, Pseudomonas aeruginosa, Staphylococcus aureus, and M. smegmatis with acid and alkali, as described above. The amount of adenine released from these samples was compared to the amount of DNA in extracts produced by bead-beating. Based upon results shown in Fig. 2, bead-beating is somewhat less than 100 % effective in disrupting cells, so extracts of bead-beaten bacteria were expected to contain less adenine than detected by acid/alkali treatment of whole bacteria. The DNA extraction efficiency of bead-beating for five different microorganisms ranged from 70 to 80 % (Table 2). These results are similar to those reported using other methods [7], supporting our assertion that the acid/HPLC method can be applied to a variety of microbial species.Table 2 Comparison of adenine in bead-beaten cell extracts with adenine in intact cellsa Organism Adenine in extracts of bead-beated cellsb (a) Adenine in intact cellsb (b) Extraction efficiency (a/b) Mycobacterium smegmatis 299.0 +/− 22.4 394.0 +/− 5.4 75.9 % Staphylococcus aureus 736.1 +/− 25.2 922.9 +/− 18.4 79.7 % Francisella philomiragia 496.6+/− 10.2 597.1 +/− 20.0 83.1 % Bacillus thuringiensis 179.6 +/− 12.4 258.6 +/− 1.0 69.5 % Moraxella catarrhalis 299.0 +/− 18.0 394 +/− 5.4 75.9 % Pseudomonas aeruginosa 501.0 +/− 20.0 624.9 +/− 4.9 80.2 % aAdenine released from DNA after acid/alkali treatment of extracts or intact cells (from approximately 109 cells), as described in Fig. 4 bArea Under Curve of adenine peak (see Methods). Mean values +/− range of two separate samples are shown Discussion We describe a method based on simple chemical principles for determining the total amount of DNA (and RNA) in an initial microbial sample. Complete release of adenine from DNA with little release from RNA is key to the acid/HPLC method. Release of guanine is expected to be similar, but this has not been proven. Once released from intracellular DNA by mild acid treatment, adenine must then be released into the extracellular milieu to be detectable. As shown in Fig. 2 and Table 1, the amount of adenine detected following acid/alkali treatment of intact cells and of disrupted cells is essentially identical, providing evidence that the acid/alkali treatment is sufficient to render cells porous to adenine. The acid/HPLC method permits detection of both intracellular DNA as well as released DNA, avoiding errors associated with use of two different DNA quantification methods (see Fig. 4). Although acid/HPLC is relatively insensitive as a detection method compared to qPCR or fluorescent dyes, it can detect single- and double-stranded DNA equally well. Finally, it also allows the quantification of intracellular RNA in the same HPLC run. In order to calculate the efficiency of any extraction procedure, knowing the initial amount of DNA (prior to extraction) is required (see Fig. 4). Currently, there exists no direct method for determining extraction efficiency, although indirect methods have been used. Colony counting is one such indirect method, but it lacks precision because the amount of DNA per genome, the number of genomes per cell and the number of cells per CFU are difficult to ascertain. An alternative approach is to determine the initial amount of DNA by exposing cells to a treatment believed to completely lyse all cells, and then quantifying the liberated DNA. For example, enzymatic lysis of cells with lysozyme is very effective but is limited to lysozyme-sensitive, gram-positive bacteria. Although bead-beating is more universally applicable, lysis efficiency can be highly variable, dependent upon equipment, buffers and cell type. For spores and other difficult-to-disrupt cells, such as Mycobacterium spp., complete cell disruption is difficult to achieve using bead-beating [7]. Because the acid/HPLC method does not suffer from any of the above limitations, we believe that it has wide applicability. The sensitivity of the method using HPLC with UV detection is 68 nanograms of adenine, or approximately 107 cells. Other methods of detection, such as fluorescence and mass spectrometry, may decrease the number of cells required for analysis. We provide several examples of the broad utility of the acid/HPLC method. We demonstrate that B. subtilis spores are relatively resistant to disruption by bead-beating, presumably due to their strong cell wall, small size and quasi-spherical shape. Calculating the amount of DNA in bacteria by colony counting significantly underestimates the amount of DNA in B. subtilis and M. smegmatis. The observed differences may be due to propensity of bacteria, particularly waxy microbes such as M. smegmatis, to form clumps. It should be noted, however, that colony counting measures only viable cells while the acid/HPLC method measures the DNA in both viable and nonviable cells. We demonstrate that minor differences in the bead-beating protocol for M. smegmatis, can produce large differences in the extraction efficiency (Fig. 5). Under optimal conditions, bead-beating was indeed an effective way to release DNA from these cells. However, when preceded by heat treatment (100 °C for 10–20 min), the efficiency of DNA release was markedly reduced. Reversing the order of treatments (bead-beating, then heating) did not reduce the effectiveness of bead-beating. This may be due to the formation of a flocculent suspension on heating. The presence of SDS, an anionic detergent commonly used in extraction protocols [21], produced a similar reduction in extraction efficiency. The likely cause is foaming, as a non-foaming detergent, pluronic F-68, did not affect extraction efficiency. These observations are of interest because bead-beating is often used in combination with heat treatment or SDS. The inefficiency of other methods for releasing DNA, such as heating in water at 100 °C, freeze-thawing or treatment with guanidine hydrochloride, was also demonstrated. Conclusions Maximizing DNA extraction efficiency is a highly desirable goal for microbiome studies and for analyzing clinical samples. For microbiome studies, high extraction efficiency may increase the detection of difficult-to-lyse bacteria and the profile of extracted DNA may more closely resemble the actual profile of different species in the original sample. For clinical samples, high efficiency of extraction is expected to increase the likelihood of a correct diagnosis by recovering microbial DNA from hard-to-lyse organisms present in low numbers. It is our expectation that the acid/HPLC method will assist in the development of new, more efficient DNA extraction methods for use in these applications. Additional file Additional file 1: Spreadsheet for converting nanomoles or nanograms of adenine to nanograms of DNA. (XLSX 17 kb) Acknowledgements We thank Jessica Gage-LaVecchia (DNA Genotek) for technical assistance and Richard Kibbee (Carleton University) for assistance with growing M. smegmatis. Funding The Mitacs-Accelerate Program provided post-doctoral support funds to O.M.B. Authors’ contributions OMB planned and executed the experiments presented. HCB provided expertise concerning depurination of DNA, degradation of RNA and contributed establishing the HPLC analysis. Both authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Niemz A Ferguson TM Boyle DS Point-of-care nucleic acid testing for infectious diseases Trends Biotechnol 2011 29 240 50 10.1016/j.tibtech.2011.01.007 21377748 2. Yuan S Cohen DB Ravel J Abdo Z Forney LJ Evaluation of methods for the extraction and purification of DNA from the human microbiome PLoS One 2012 7 e33865 10.1371/journal.pone.0033865 22457796 3. Zhang ZQ Ishaque M Evaluation of methods for isolation of DNA from slowly and rapidly growing mycobacteria Int J Lepr Other Mycobact Dis 1997 65 469 76 9465157 4. Kaser M, Ruf MT, Hauser J, Pluschke G. Optimized DNA preparation from mycobacteria. Cold Spring Harb Protoc. 2010;2010:db. doi: 10.1101/pdb.prot5408 5. Aldous WK Pounder JI Cloud JL Woods GL Comparison of six methods of extracting Mycobacterium tuberculosis DNA from processed sputum for testing by quantitative real-time PCR J Clin Microbiol 2005 43 2471 3 10.1128/JCM.43.5.2471-2473.2005 15872286 6. Vandeventer PE Weigel KM Salazar J Erwin B Irvine B Doebler R Mechanical disruption of lysis-resistant bacterial cells by use of a miniature, low-power, disposable device J Clin Microbiol 2011 49 2533 9 10.1128/JCM.02171-10 21543569 7. Olson ND Morrow JB DNA extract characterization process for microbial detection methods development and validation BMC Res Notes 2012 5 668 10.1186/1756-0500-5-668 23206278 8. Amaro A Duarte E Amado A Ferronha H Botelho A Comparison of three DNA extraction methods for Mycobacterium bovis, Mycobacterium tuberculosis and Mycobacterium avium subsp. avium Lett Appl Microbiol 2008 47 8 11 10.1111/j.1472-765X.2008.02372.x 18498320 9. Carrigg C Rice O Kavanagh S Collins G O’Flaherty V DNA extraction method affects microbial community profiles from soils and sediment Appl Microbiol Biotechnol 2007 77 955 64 10.1007/s00253-007-1219-y 17960375 10. Tenover FC Developing molecular amplification methods for rapid diagnosis of respiratory tract infections caused by bacterial pathogens Clin Infect Dis 2011 52 Suppl 4 S338 45 10.1093/cid/cir049 21460293 11. Firmani MA Broussard LA Molecular diagnostic techniques for use in response to bioterrorism Expert Rev Mol Diagn 2003 3 605 16 10.1586/14737159.3.5.605 14510181 12. Burton K, Lunt MR, Peterson GB, Siebke JC. Studies of nucleotide sequences in deoxyribonucleic acid. Cold Spring Harb Symp Quant Biology. 1963;28:27–34. 13. Suzuki T Ohsumi S Makino K Mechanistic studies on depurination and apurinic site chain breakage in oligodeoxyribonucleotides Nucleic Acids Res 1994 22 4997 5003 10.1093/nar/22.23.4997 7800492 14. Majcher MR Bernard KA Sattar SA Identification by quantitative carrier test of surrogate spore-forming bacteria to assess sporicidal chemicals for use against Bacillus anthracis Appl Environ Microbiol 2008 74 676 81 10.1128/AEM.01715-07 18083869 15. Doetsch RN Gerhardt P Murray RGE Costilow RN Nester EW Wood WA Krieg NR Determinative methods of light microscopy Manuals of Methods for General Bacteriology 1981 Washington, D.C American Society for Microbiology 21 3 16. Voet D Gratzer WB Cox RA Doty P Absorption spectra of nucleotides, polynucleotides, and nucleic acids in the far ultraviolet Biopolymers 1963 1 193 208 10.1002/bip.360010302 17. Michelon CT Rosso F Schmid KB Sperhacke RD Oliveira MM Kritski AL Colorimetric microwell plate reverse-hybridization assay for Mycobacterium tuberculosis detection Mem Inst Oswaldo Cruz 2011 106 194 9 10.1590/S0074-02762011000200013 21537680 18. Tamm C Hodes ME Chargaff E The formation of apurinic acid from the desoxyribonucleic acid of calf thymus J Biol Chem 1952 195 49 63 14938354 19. Pribylka A Almeida AV Altmeyer MO Petr J Sevcik J Manz A Fast spore breaking by superheating Lab Chip 2013 13 1695 8 10.1039/c3lc41305e 23474861 20. Belgrader P Hansford D Kovacs GT Venkateswaran K Mariella R Jr Milanovich F A minisonicator to rapidly disrupt bacterial spores for DNA analysis Anal Chem 1999 71 4232 6 10.1021/ac990347o 10517145 21. Kaser M Ruf MT Hauser J Marsollier L Pluschke G Optimized method for preparation of DNA from pathogenic and environmental mycobacteria Appl Environ Microbiol 2009 75 414 8 10.1128/AEM.01358-08 19047396 22. Cohn WE Methods of isolation and characterization of mono- and polynucleotides by ion exchange chromatography Meth Enzymology 1957 3 724 43 10.1016/S0076-6879(57)03450-3

PMC005xxxxxx/PMC5002185.txt

==== Front J Diabetes Metab DisordJ Diabetes Metab DisordJournal of Diabetes and Metabolic Disorders2251-6581BioMed Central London 2757074825710.1186/s40200-016-0257-3Research ArticleThe effect of vitamin D on COPD exacerbation: a double blind randomized placebo-controlled parallel clinical trial Sanjari Mojgan mjnsanjari@yahoo.com 1Soltani Akbar SoltaniA@sina.tums.ac.ir 2Habibi Khorasani Abdolrahim habibikhorasani@kmu.ac.ir 3Zareinejad Maryam +989131404529kerman.physiology@gmail.com 41 Endocrinology and Metabolism Research Center, Basic and Clinical Physiology Institute, Kerman University of Medical Sciences, Kerman, Iran 2 Endocrinology and Metabolism Research Center, Endocrinology and Metabolism Clinical Sciences Institute, EBM Group, Tehran University of Medical Sciences, Tehran, Iran 3 Department of Pulmonary Medicine, Kerman University of medical sciences, Kerman, Iran 4 Physiology Research Center, Institute of Neuropharmacology, Kerman University of Medical Sciences, Kerman, Iran 26 8 2016 26 8 2016 2016 15 3316 2 2016 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 To investigate the effect of supplementation of standard treatment (inhaled long-acting β2 agonists, anticholinergics and corticosteroids) with vitamin D on C reactive protein and pulmonary function tests in patients with COPD exacerbation. Methods Design: Randomized, single-center, double-blind, placebo-controlled parallel trial. One teaching hospital Participants: 135 patients in pulmonary ward with moderate to severe COPD and exacerbations.120 patients fulfilled the study protocol. Interventions: Patients were randomly divided into three groups receiving 7 day treatment with 0.25 μg calcitriol daily (n = 45), 50000 IU daily of vitamin D (n = 45) or placebo (n = 45). An independent nurse was responsible for allocation, preparation, and accounting of trial medications. Main Outcome measures: Maximal expiratory flow volume (FEV1) and forced volume capacity curves (FVC) and Modified Medical Research Council (MMRC) scale. Results Out of 135 patients who were recruited consecutively, 45 patients randomly were randomly assigned in three groups (balance blocked randomization.15 patients were dropped out due to non-compliance for second PFT. Intention to treat analysis was carried out for 120 participants. The difference between before and after treatment FEV1 and FEV1/FVC ratio had no significant difference between treatment groups and placebo. (P = 0.43, P = 0.51, respectively)but clinical improvement was significant in patients who received calcitriol. No side effects were reported. Conclusions Short term treatment with either calcitriol or 25(OH) 2Vit D didn’t changed FEV1 or FVC in vitamin D sufficient patients with COPD exacerbation; nevertheless it can provide clinical benefit. Trial registration Trial registration: Iranian Registry of Clinical Trials no. IRCT138712271774N1. Registered 10 April 2011. Keywords CalcitriolCOPDForced expiratory volumeForced Vital CapacityPulmonary function testsVitamin DThis study was a part of a joint project supported by the vise chancellor and physiology research center Kerman University of medical sciences and Endocrine and Metabolism Research Center, Tehran University of Medical Sciencesk/88/85Sanjari Mojgan issue-copyright-statement© The Author(s) 2016 ==== Body Background Chronic obstructive pulmonary disease (COPD) is the fourth main cause of death in the United States [1]. Many studies have shown 25-OHD serum levels in patients suffering from COPD. Vitamin D deficiency in African-Americans is closely associated with higher susceptibility to COPD [2, 3]. Also, the studies suggest that there is a link between genetic variants in the vitamin D pathway and COPD. For example, studies suggest that a single nucleotide polymorphism (SNP) of the vitamin D binding protein protects COPD through a mechanism that is still unknown to researches [4]. The VDR levels is significantly low in lung tissues of COPD patients [5]. Active metabolite of vitamin D (1, 25 dihydroxy vitamin D3) has a vital role in cellular metabolism and differentiation by means of its nuclear receptor (VDR) that is associated with a number of other chromatin modification enzymes (histone acetyltrans-ferases and histone deacetylases), and therefore involved in the complex epigenetic events in vitamin D signaling and metabolism [6]. Recently, studies have found a linear correlation between higher risks of vitamin D deficiency in patients with COPD and the function of lung and systemic levels of vitamin [7, 8]. Nevertheless in a cohort study carried out on patients with severe COPD, baseline 25 (OH) D levels could not predict the ensuing acute exacerbation of COPD [8]. Patients with COPD are subjected to exacerbation. The symptoms of exacerbations are normally severe deterioration of symptoms, intense airway inflammation and physiological worsening [9]. They are assumed to act as a reference in the therapeutic study of COPD. Recently, some studies have found the effect of exacerbations on disease progression, which can be manifested by accelerated forced expiratory volume in one second (FEV1), and reduced characteristic of COPD [10, 11]. Brehm showed that vitamin D deficiency is related to severe of allergy and asthma attacks. There was a significant and inverse relationship between the level of Serum vitamin D and increased airway responsiveness [12]. Exacerbations is estimated to be the cause of nearly 25 % of the FEV1 decline in COPD [10]. With the exception of smoking quitting, no other proved treatment has been found for hindering COPD disease progression; however, two recent large randomized controlled experiments have shown the positive impacts of long-acting bronchodilators on the rate of exacerbation, life quality, and survival [11, 13]. It is while the inefficacy of a host of new therapeutics or recombinant antibodies have been proved. In light of the unsatisfactory results of these expensive medications, greater attention should be paid to the potential of traditional medications. The potential therapeutic effect of drugs such as statins, angiotensin-converting enzyme (ACE)–inhibitors, and heparin have not been recognized and they may attract greater attention in future therapy of COPD [14, 15]. This study explores the potentials of another, even older molecule, vitamin D. A cross-sectional, retrospective study of children there is a relationship between in Costa Rica show that vitamin D deficiency and higher symptoms of asthma severity [12]. Another study suggested the higher risk of severe asthma exacerbation in patients suffering from vitamin D insufficiency [16]. The aim of this study is to evaluate the role of exogenous addition of vitamin D and 1, 25 dihydroxy vitamin D on the function of lung in the patients suffering from COPD exacerbation. Methods Study Design and Participants: The current study was a single-center, double-blind research, which used parallel group and randomized placebo-controlled trial. Two hundred and five patient were assessed for eligibility (Not meeting inclusion criteria, n = 38 & Declined to participate, n = 32). A total of 135 patients aged between 40 and 60 years were recruited consecutively diagnosed as having COPD from the emergency department of Kerman Afzalipour university Hospital in Kerman city (the largest province in south east of Iran) during winter of 2011. All subjects were ex-smokers with a history of chronic cough and expectoration who also had exertional dyspnea and were admitted with the diagnosis of COPD exacerbation. Based on the ERS criteria, COPD defined when FEV1 % prediction is <88 % for men, or <89 % for women [17, 18]. An acute exacerbation of COPD can be defined as a dramatic degradation of COPD symptoms (for example, the quantity and the color of phlegm or shortness of breath) that last for a couple of days [19]. Patients were excluded if the following were presented: unwillingness to participate in the study, unable to perform spirometry, A history of asthma symptoms, existence of other respiratory disorders including bronchial carcinoma, a history of hospitalization (within 4 weeks) for COPD, any medical condition which needed more invasive respiratory support, symptoms of lower respiratory tract infection or other kinds of simultaneous systemic disease such as hypercalcemia, renal failure, hyperparathyroidism, malignancy, history of renal stone, cardiac arrhythmia, or patients who were using lithium. All subjects signed an informed written consent to participate in the study. The protocols of the study were confirmed by the ethics committee of KUMS [Kerman university of medical sciences (Code 88/17)], and this trial was registered on the IRCT [Iranian Registry of Clinical Trials] website (the registration number of the trial is IRCT138712271774N1). We tried this randomized controlled trials conform to the CONSORT statement and flow diagram (Fig. 1) [19].Fig. 1 Study flow diagram Full trial protocol can be accessed in Iranian Registry of Clinical Trials website. Randomization and Masking: In this study, we used a double-blind randomized, parallel group, placebo-controlled design. After the preliminary evaluation, the qualified patients were randomly assigned to one of the three treatment groups and followed up for one week. Participants were randomized by a computer-generated list to 3 equal groups. Patients were divided into 3 groups according to the balance blocked randomization. Out of 135 patients, 45 patients randomly were selected in three groups of A, B and C. Patients in each group received same envelopes containing different drugs. Both patients and the physicians and epidemiologist were blinded to the treatment received. A nurse who was not involved in care of the trial patients and independent of the site investigator was responsible for allocation, preparation, and accounting of trial medications. The trial medications and placebo was prepared at a separate site, and then taken to the clinic every week. The randomization schedule was thus concealed from all care providers, ward physicians, and other research personnel. The patients who left the study or for whom the experiment could not be completed were excluded, and finally3, 6 and 6 patients were eliminated in groups A, B and C respectively. No side effect or harm was reported. In phone recall all of eliminated patients were alive and improved clinically after one week but they didn’t participate to the last part of the study (Second PFT and Lab Tests) due to personal problems. Procedures After randomizing the patients, there were follow-up visits at the beginning and the end of study (day 0 and 8). Patient received the following treatment. Group A: Placebo (n = 42) - The patients in this group received standard treatment for COPD according to the guidelines of GOLD [20]. Also, these patients received placebo capsules, which were similar to active drug in group B and C, daily for a week. Group B: Treatment with vitamin D supplementation (N = 39. Similar to in group A, these patients received standard treatment for COPD in addition to25 (OH)-D3 (Pourateb, Iran) 50000 IU capsules Daily for seven days. Group C: Treatment supplemented with calcitriol (n = 39). Standard COPD treatment was given to these patients according to the guidelines of GOLD [20, 21] plus calcitriol (Pourateb, Iran) 0.25 μg capsules daily, for seven days. If necessary, inhaled salbutamol was permitted. Also, to control exacerbations, additional bronchodilators, oral steroids and antibiotics were used. Corticosteroid therapy (oral or inhaled), if present, could continue without any change. Spirometry: Spirometry was carried out at the beginning and the end of study (day 0 and 8). The experiment was carried out on a dry, rolling-seal spirometer Benchmark model (Spirolab, UK). According to the guidelines of American Thoracic Society/European Respiratory Society Maximal expiratory flow volume and forced volume capacity curves were attained. For each subject, three acceptable and a minimum of two reproducible curves were obtained. The greatest values of FEV1 and FVC were chosen for analysis. we used the Modified Medical Research Council (MMRC) scale for the evaluation of dyspnea in everyday activities at day 0 and 8 in each group [22] [Table 1].Table 1 The modified Medical Research Council dyspnea scale (MMRC) Grade of dyspnea Symptom Grade 0 Not troubled by breathlessness except on strenuous exercise Grade 1 Short of breath when hurrying or walking up a slight hill Grade 2 Walks slower than contemporaries on the level because of breathlessness or has to stop for breath when walking at own pace Grade 3 Stops for breath after walking 100 m or after a few minutes on the level Grade 4 Too breathless to leave the house or breathless when dressing or undressing Biochemical estimations: All serum samples were stored frozen at −80 C. Serum calcium was measured for excluding hypercalcemic patients and albumin was checked for interpreting the calcium results. C-reactive protein (CRP) was measured when the follow-up period was terminated. High sensitivity CRP concentrations were measured on a Hitachi 911 analyzer by a latex-enhanced immuneturbidimetric assay from Roche. Circulating25 (OH) D3, and1, 25(OH) 2VITD was measured by Stat Fax Instrument and Immunodiagnostic System Kit using an Enzyme immunoassay method. In each set of samples, we measured two internal control specimens that the reagent manufacturer had provided. The total variation coefficient of for measured internal controls was 9.3. In the final stage of treatment, responses were assessed by a physician, who asked patients about the improvement, stability or deterioration of their condition. The primary end points were FEV1 and FVC and clinical response. Secondary end points were serum level of C reactive protein, totalanti-oxidant, 25 and 1&25OH- vitamin D. Statistical Analysis: For convenience, STATA statistical software was used for determining the sample size. SPSS 16 (Chicago, IL) was used for statistical analysis. Paired t test and Mann Whitney test were respectively used for data with and without normal distribution. To compare categorical variables within three groups, Chi square test (and Fisher exact test if required) was used. Similarly the changes in the outcome parameters were compared by univariate analysis of variance (ANOVA) and followed by Bonferroni correction. To control the effects of first value FEV1 and FEV1/FVC, two-way ANOVA (ANCOVA) was used. All collected data were included in the intention to treat analysis. The statistical significance level was considered less than 0.05 (P value < 0.05). Results One hundred and twenty patients who had COPD exacerbation met the eligibility criteria and completed the trial. The details of the population are summarized in [Table 2]. Statistically, treatment arms were not significantly different with respect to age and gender and severity of disease which is shown pre-treatment FEV1 and FEV1/FVC ratio. There was not a significant statistical difference between Calcium and Albumin and C reactive protein (P value > 0.05). At the end of study, none of the patient developed hypercalcemia [Table 3].Table 2 Baseline characteristics of the study population Variables Placebo (n = 42) Vitamin D (n = 39) Calcitriol (n = 39) P value Age, 58.4 ± 9.5 55.8 ± 9.5 55.6 ± 10.4 0.382 Sex, %  Male 69.2 71.4 87.2 0.128  Female 30.8 28.6 12.8 FEV1  Baseline 46.7 ± 20.4 45.3 ± 18.9 46.3 ± 17.5 0.439  End 62.3 ± 22.9 64.5 ± 22.8 64.1 ± 21.3 FEV1/FVC  Baseline 70.2 ± 16.9 68.2 ± 16.9 69.1 ± 17.2 0.513  End 72.9 ± 14.1 73.1 ± 14.2 74.4 ± 15.4 MMRC 3.62 ± 0.49 3.69 ± 0.46 3.6 ± 0.49* 0.001 2.66 ± 0.74 2.65 ± 0.67 1.98 ± 0.65 Data are expressed as mean ± SD *different from placebo p ≤ 0.05 Table 3 Effect of treatments on Calcium, Albumin,CRP and different forms of vitamin D Variables Placebo (n = 42) Vitamin D (n = 39) Calcitriol (n = 39) P. Value Albumin (g/l)  Baseline 39.05 ± 4.37 38.71 ± 5.7 39.7 ± 4.9 0.66  Post-Baseline(End) 43.2 ± 4.1 42.1 ± 4.2 43.4 ± 5.1 0.33  Difference (change) 4.2 ± 3.9 3.4 ± 3.7 3.76 ± 4.8 0.67 Calcium (mmol/l)  Baseline 2.4 ± 0.19 2.46 ± 0.13 2.43 ± 0.18 0.43  Post-Baseline(End) 2.6 ± 0.15 2.6 ± 0.12 2.6 ± 0.12 0.076  Difference (change) 0.19 ± 0.18 0.20 ± 0.16 0.17 ± 0.14 0.78 CRPa (Mg/dL)  Baseline 70.1 ± 61.5 70.9 ± 52.1 67.3 ± 57.1 0.95 25OH-D(nmol/l)  Baseline 59.9 ± 26 58.9 ± 27 54.9 ± 34.9 <0.001  End 65.2 ± 39.2 97.7 ± 52.2* 57.1 ± 44.4 1,25OH-D(pmol/l)  Baseline 91. ±13.5 110.2 ± 71.2 103.5 ± 83.7 0.563  End 108.7 ± 59.5 128.2 ± 121.9 101.9 ± 81.1 aCRP:Creactive protein *P < 0.001 The baseline-anticipatedFEV1 percent was the same in the three groups (Table 2). The FEV1 percent predicted in Calcitroil group showed an increase from 46.3 ± 17.5 at baseline to 64.1 ± 21.3 after treatment (P < 0.01). In Vitamin D group, it rose from 45. 3 ± 18.9 to 64.5 ± 22.8 after treatment (P < 0.01), and in Placebo group, 46.7 ± 20.4 reached to 62.3 ± 22.9 (P < 0.01). The difference between before and after treatment FEV1 and FEV1/FVC ratio was calculated and compared between groups by ANOVA and it no significant difference was observed between treatment groups and placebo.(P = 0.43, P = 0.51, respectively) After adjusting the effect of baseline FEV1 and FEV1/FVC ratio by ANCOVA the result didn’t changed too (P = 0.44, P = 0.54, respectively). The magnitude of increase in FEV1 and FEV1/FVC ratio was corresponding in the three groups. However, clinically improved lung function(Reduction in MMRC) was observed in all cases and was significant in the majority of patients. The MMRC in Calcitroil group showed a significant decrease from 3.6 ± 0.49 at baseline to 1.98 ± 0.65after treatment (P < 0.0001). In Vitamin D group, it decreased from 3.69 ± 0,46 to 2.65 ± 0.67 after treatment (P < 0.001), and in Placebo group, 3.62 ± 0.49reached to 2.62 ± 0.74 (P < 0.01). The difference between before and after treatment MMRC was calculated and compared between groups by ANOVA and Tukey test and significant difference was observed between calcitriol group vs. placebo (1.46 vs. 0.71 P < 0.001). Maximum reduction in MMRC was seen in calcitriol group and then vitamin D and placebo group respectively (P < 0.001). Statistically speaking, there was no relationship between FEV1 and changed FEV1with serum level of 25 OH-vitamin D and 1,25 (OH)2 vitamin D before and after treatment (P > 0.1) there was a significant rise in the levels of plasma vitamin 25OH-D in Vitamin D group from58.9 ± 27 nmol/l at baseline to 97.7 ± 52.2 nmol/l after treatment (P < 0.001). In Calcitriol group, however, the variation from 54.9 ± 34.9 to 57.1 ± 44.4 nmol/l was not significant. The same result was seen in placebo group. The plasma level of 1,25OH-D was not affected significantly by any of treatments [Table 3]. Discussion This study revealed that short term treatment with neither calcitriol nor vitamin D has any effect on pulmonary function of COPD patients during exacerbation, nevertheless the MMRC in Calcitroil group showed a significant decrease. The level of calcitriol did not change during this study, but the increase in 25(OH) vitamin D in group B indicated the compliance of the patient in ingesting the drug. A number of studies have stated 25-OHD serum levels in COPD patients. One recently published study showed that aspiratory muscle strength and maximal oxygen uptake in patients receiving vitamin D improved significantly compared with placebo group, but two groups were not statistically different in terms of improved quadriceps strength or six minutes walking distance [23]. Also, high doses of vitamin D(100 000 IU every 4 weeks for 1 year) in vitamin D sufficient patients did not have a significant effect on mean time before the appearance of the first exacerbation, exacerbation rates, FEV1, hospitalization, life quality, and death. In participants who had acute vitamin D deficiency at baseline, the use of supplementation could reduce exacerbations [21]. The studies of Forli and colleagues showed that more than 50 % of patients enlisted for lung transplantation had vitamin D deficiency (<20 ng/ml) [24]. Likewise, a study on community-dwelling patients in Denmark who suffered from COPD revealed 68 % of the participants had untreated osteoporosis or osteopenia [25]. It is worth noting that insufficient or low levels of vitamin D are not particular to COPD. Patients with COPD are more likely to have vitamin D deficiency for several reasons [3]. Vitamin D deficiency is a common problem in most population yet [26, 27]. The current supplementation with a daily dose of 800 to 1,000 IU of vitamin D makes up for low levels of serum 25-OHD in the majority of adults in concentrations greater than 20 ng/ml, but higher doses may be needed to increase 25-OHD levels proportional to the needs of high-risk patients who don’t have hypocalcaemia [28]. One recent study suggest that “the current vitamin D recommended dietary allowance (RDA) (600 IU/day) is insufficient to cover the skeletal health needs of at least 50 % of black and white children” [29]. In this study most patients were vitamin D sufficient maybe due to extra use of vitamin D supplementation in old population in our country [30]. It was shown that “ Vitamin D intake decreased COPD exacerbation and improved FEV1 in the patients with severe and very severe COPD” [31]. Unfortunately in this study which was carried out in our country the baseline level of vitamin D was not determined. It was shown that “vitamin D3 was protective against moderate or severe exacerbation in participants with baseline serum 25-hydroxyvitamin D concentrations of less than 50 nmol/L” too [32]. Our study showed no significant relationship between vitamin D level and FEV1 and or FEV1/FVC, it may be due to sufficient leveled vitamin D in the population of our study. Unfortunately there were not enough vitamin D deficient patients to do any statistical analysis in this subgroup. The present study showed that any relationship between vitamin D level and FEV1 or airway responsiveness may be due to high levels of vitamin D in the patients. The more important cause can be explained by the difference between the pathophysiology of asthma and chronic obstructive lung diseases a non-reversible airway disease. The level of CRP didn`t significant difference between three groups it may emphasis that this dose and duration of vitamin D and calcitriol can`t be effective on reducing the inflammation. Due to the great possibility of genetic involvement in Vitamin D's effect on respiratory disease and the high prevalence of Vitamin D deficiency in Iran, it is suggested to conduct clinical trials among the Iranian people and evaluate the efficacy of Vitamin D [33]. Our study had some limitations. First most of our patients were Vitamin D sufficient so we couldn’t determine the effect of short term treatment in patient Vitamin D deficiency (which may be more effective). Longer duration with larger sample size study in vitamin D deficient patients with COPD suggests. Second our study was too short to follow the patient for readmission so we recommend study the effect of calcitriol on the frequency of hospitalization. Third as this study were done in normocalcemic patient we had chosen the lowest dosage of calcitriol, higher doses of calcitriol can be used in future studies. Given the inefficacy of current treatments in COPD, there is still a need for further randomized control trials with vitamin D supplementation. In such randomized controlled trials in which vitamin D treatment is only based on bone maintenance, defining the optimal serum levels for 25-OHD in this disease as reference for other calcemic endpoints is a challenging task. Future clinical trials are suggested to evaluate the impact of higher levels of vitamin D and especially calcitriol on lung function during exacerbation. We suggest investigating this correlation in vitamin D deficient patients. Conclusion Short term treatment with either calcitriol or 25(OH) 2Vit D didn`t changed FEV1 or FVC in vitamin D sufficient patients with COPD exacerbation; nevertheless it can provide clinical benefit. Abbreviations AlbAlbumin CaCalcium 25OH-DVitamin D 1,25OH-D1,25Vitamin D CRPC reactive protein CalcitriolCalcitriol Acknowledgment This study was a part of a joint project supported by the vise chancellor and physiology research center Kerman University of medical sciences and Endocrine and Metabolism Research Center, Tehran University of Medical Sciences. This study approved as an Internal medicine resident thesis (proposal no.88/85). The authors would like to thank all participant of this study. The authors would like to thank the director and laboratory personnel of Mehrgan lab (especially Mrs. Zohreh Safi). Funding This study was a part of a joint project supported by the vise chancellor and physiology research center Kerman University of medical sciences and Endocrine and Metabolism Research Center, Tehran University of Medical Sciences. Availability of data and materials Copies informed consent form, lab data and SPSS files are available in Endocrinology and metabolism Research Center, institute of Basic and clinical physiology Sciences, Kerman University of Medical Sciences. Authors’ contribution MS designed the study, worked on design and data collection. AHK and AS advised scientific guidance. MS and MZ did the analysis with input from all other authors. The initial version of this article was written by MS, MZ and AS. All other authors contributed to the writing of the paper. All authors read and approved the final manuscript. Authors' information 1Associate Professor Endocrinology and Metabolism Research Center, Basic and Clinical Physiology Institute, Kerman University of Medical Sciences, Kerman, Iran Email: mjnsanjari @yahoo.com. 2Associate Professor, Endocrinology and Metabolism Research Center, Endocrinology and Metabolism Clinical Sciences Institute, EBM Group, Tehran University of Medical Sciences, Tehran, Iran Email: SoltaniA @sina.tums.ac.ir, 3 Assistant Professor Department of Pulmonary Medicine, Kerman University of medical sciences, Kerman, Iran. Email: habibikhorasani@kmu.ac.ir, 4 Physiology Research Center, Institute of Neuropharmacology, Kerman University of Medical Sciences, Kerman, Iran.. Email: kerman.physiology@gmail.com Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable (Design is original article not case report or series. Consent for publication had been a part for consent for participates). Ethics approval and consent to participate The study was approved by Ethics Committee of Kerman University of Medical Sciences; license number k/88/85. Written informed consent was obtained from all cases and controls. ==== Refs References 1. Gan WQ Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-analysis Thorax 2004 59 7 574 580 10.1136/thx.2003.019588 15223864 2. Bouillon R Bischoff‐Ferrari H Willett W Vitamin D and health: perspectives from mice and man J Bone Miner Res 2008 23 7 974 979 10.1359/jbmr.080420 18442312 3. Holick MF Vitamin D deficiency N. Engl. J. Med. 2007 357 3 266 281 10.1056/NEJMra070553 17634462 4. Lips P Vitamin D physiology Prog Biophys Mol Biol 2006 92 1 4 8 10.1016/j.pbiomolbio.2006.02.016 16563471 5. Sundar IK Deletion of vitamin D receptor leads to premature emphysema/COPD by increased matrix metalloproteinases and lymphoid aggregates formation Biochem Biophys Res Commun 2011 406 1 127 133 10.1016/j.bbrc.2011.02.011 21300024 6. Sundar IK Rahman I Vitamin D and susceptibility of chronic lung diseases: role of epigenetics Front Pharmacol 2011 30 2 50 7. Persson LJP Chronic obstructive pulmonary disease is associated with low levels of vitamin D PLoS One 2012 7 6 e38934 10.1371/journal.pone.0038934 22737223 8. Kunisaki KM Niewoehner DE Connett JE Vitamin D levels and risk of acute exacerbations of chronic obstructive pulmonary disease: a prospective cohort study Am J Respir Crit Care Med 2012 185 3 286 290 10.1164/rccm.201109-1644OC 22077070 9. Bhowmik A Relation of sputum inflammatory markers to symptoms and lung function changes in COPD exacerbations Thorax 2000 55 2 114 120 10.1136/thorax.55.2.114 10639527 10. Jones P Agusti A Outcomes and markers in the assessment of chronic obstructive pulmonary disease Eur Respir J 2006 27 4 822 832 10.1183/09031936.06.00145104 16585091 11. Rabe KF Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary Am J Respir Crit Care Med 2007 176 6 532 555 10.1164/rccm.200703-456SO 17507545 12. Brehm JM Serum vitamin D levels and markers of severity of childhood asthma in Costa Rica Am J Respir Crit Care Med 2009 179 9 765 771 10.1164/rccm.200808-1361OC 19179486 13. Seemungal TA Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease Am J Respir Crit Care Med 1998 157 5 1418 1422 10.1164/ajrccm.157.5.9709032 9603117 14. Cazzola M Targeting systemic inflammation: novel therapies for the treatment of chronic obstructive pulmonary disease 2007 15. Mancini GJ Reduction of morbidity and mortality by statins, angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers in patients with chronic obstructive pulmonary disease J Am Coll Cardiol 2006 47 12 2554 2560 10.1016/j.jacc.2006.04.039 16781387 16. Ciaccio CE Portnoy JM Vitamin D insufficiency and risk of severe asthma exacerbations Curr Allergy Asthma Rep 2010 10 6 387 388 10.1007/s11882-010-0144-9 20725812 17. Nathell L COPD diagnosis related to different guidelines and spirometry techniques Respir Res 2007 8 1 89 10.1186/1465-9921-8-89 18053200 18. Quajer P Lung volumes and forced ventilatory flows. Report working party. Standardization of lung function test European Community for steel and oral official statement of the European Respiratory Society Eur Respir J 1993 6 5 40 10.1183/09041950.005s1693 19. Schulz KF Altman DG Moher D CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials Ann Intern Med 2010 152 11 726 732 10.7326/0003-4819-152-11-201006010-00232 20335313 20. Rahman I MacNee W Reactive oxygen species 2002 London Academic 21. Lehouck A High doses of vitamin D to reduce exacerbations in chronic obstructive pulmonary disease: a randomized trial Ann Intern Med 2012 156 2 105 114 10.7326/0003-4819-156-2-201201170-00004 22250141 22. Mahler D Wells CK Evaluation of clinical methods for rating dyspnea CHEST J 1988 93 3 580 586 10.1378/chest.93.3.580 23. Hornikx M Vitamin D supplementation during rehabilitation in COPD: a secondary analysis of a randomized trial Respir Res 2012 13 1 84 10.1186/1465-9921-13-84 23006613 24. Førli L Vitamin D deficiency, bone mineral density and weight in patients with advanced pulmonary disease J Intern Med 2004 256 1 56 62 10.1111/j.1365-2796.2004.01337.x 15189366 25. Jørgensen NR The prevalence of osteoporosis in patients with chronic obstructive pulmonary disease—a cross sectional study Respir Med 2007 101 1 177 185 10.1016/j.rmed.2006.03.029 16677808 26. Cashman KD, Dowling KG, Škrabáková Z, Gonzalez-Gross M, Valtueña J, De Henauw S, Moreno L, Damsgaard CT, Michaelsen KF, Mølgaard C, Jorde R. Vitamin D deficiency in Europe: pandemic? Am J Clin Nutr. 2016;103(4):1033-44. 27. Souberbielle JC, Massart C, Brailly-Tabard S, Cavalier E, Chanson P. Prevalence and determinants of vitamin D deficiency in healthy French adults: the VARIETE study. Endocrine. 2016;22:1–8. 28. Vieth R Vitamin D toxicity, policy, and science J Bone Miner Res 2007 22 S2 V64 V68 10.1359/jbmr.07s221 18290725 29. Rajakumar K, Moore CG, Yabes J, Olabopo F, Haralam MA, Comer D, Holick MF, Greenspan SL. Estimations of dietary vitamin D requirements in black and white children. Pediatr Res. 2016;80(1):14-20. doi:10.1038/pr.2016.46. Epub 2016. 30. Hashemipour S Vitamin D deficiency and causative factors in the population of Tehran BMC Public Health 2004 4 1 1 10.1186/1471-2458-4-38 14706119 31. Zendedel A Effects of Vitamin D Intake on FEV1 and COPD Exacerbation: A Randomized Clinical Trial Study Global J. Health Sci. 2015 7 4 243 10.5539/gjhs.v7n4p243 32. Martineau AR Vitamin D 3 supplementation in patients with chronic obstructive pulmonary disease (ViDiCO): a multicentre, double-blind, randomised controlled trial Lancet Respir Med 2015 3 2 120 130 10.1016/S2213-2600(14)70255-3 25476069 33. Hejazi ME Modarresi-Ghazani F Entezari-Maleki T A review of Vitamin D effects on common respiratory diseases: Asthma, chronic obstructive pulmonary disease, and tuberculosis J Res Pharm pract 2016 5 1 7 10.4103/2279-042X.176542 26985430

PMC005xxxxxx/PMC5002186.txt

==== Front Cardiovasc DiabetolCardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 43610.1186/s12933-016-0436-zOriginal InvestigationAssociation between dietary acid load and the risk of cardiovascular disease: nationwide surveys (KNHANES 2008–2011) Han Eugene eghan@yuhs.ac 12Kim Gyuri gyuri.kim.md@yuhs.ac 12Hong Namki nkhong84@yuhs.ac 12http://orcid.org/0000-0002-6219-4942Lee Yong-ho yholee@yuhs.ac 13Kim Dong Woo kimdow@knou.ac.kr 4Shin Hyun Joon hys219@mail.harvard.edu 56Lee Byung-Wan bwanlee@yuhs.ac 123Kang Eun Seok edgo@yuhs.ac 123Lee In-Kyu leei@knu.ac.kr 7Cha Bong-Soo bscha@yuhs.ac 1231 Division of Endocrinology, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Korea 2 Graduate School, Yonsei University College of Medicine, Seoul, Korea 3 Institute of Endocrine Research, Yonsei University College of Medicine, Seoul, Korea 4 Department of Home Economics, Food and Nutrition, Korea National Open University, Seoul, Korea 5 Department of Nutrition, Harvard School of Public Health, Boston, MA USA 6 Department of Medicine, Baylor University Medical Center and Baylor Jack and Jane Hamilton Heart and Vascular Hospital, Dallas, TX USA 7 Division of Endocrinology, Department of Internal Medicine, Kyungpook National University School of Medicine, Daegu, Korea 26 8 2016 26 8 2016 2016 15 1 12221 6 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 Acid–base imbalance has been reported to increase incidence of hypertension and diabetes. However, the association between diet-induced acid load and cardiovascular disease (CVD) risk in the general population has not been fully investigated. Methods This was a population-based, retrospectively registered cross-sectional study using nationally representative samples of 11,601 subjects from the Korea National Health and Nutrition Examination Survey 2008–2011. Individual CVD risk was evaluated using atherosclerotic cardiovascular disease (ASCVD) risk equations according to 2013 ACC/AHA guideline assessment in subjects aged 40–79 without prior CVD. Acid–base status was assessed with both the potential renal acid load (PRAL) and the dietary acid load (DAL) scores derived from nutrient intake. Results Individuals in the highest PRAL tertile had a significant increase in 10 year ASCVD risks (9.6 vs. 8.5 %, P < 0.01) and tended to belong to the high-risk (10 year risk >10 %) group compared to those in the lowest PRAL tertile (odds ratio [OR] 1.23, 95 % confidence interval [CI] 1.22–1.35). The association between higher PRAL score and high CVD risk was stronger in the middle-aged group. Furthermore, a multiple logistic regression analysis also demonstrated this association (OR 1.20 95 % CI 1.01–1.43). Subgroup analysis stratified obesity or exercise status; individuals in unhealthy condition with lower PRAL scores had comparable ASCVD risk to people in the higher PRAL group that were in favorable physical condition. In addition, elevated PRAL scores were associated with high ASCVD risk independent of obesity, exercise, and insulin resistance, but not sarcopenia. Similar trends were observed with DAL scores. Conclusion Diet-induced acid load was associated with increased risk of CVD, independent of obesity and insulin resistance. Keywords DietAtherosclerosisRisk factorsEpidemiologyKorea Ministry of Health and WelfareHI14C2476HI16C1501Lee Yong-ho issue-copyright-statement© The Author(s) 2016 ==== Body Background Cardiovascular disease (CVD) has been a major cause of death around the world [1]. Although CVD mortality has decreased in developed countries in recent decades [1], CVD accounts for 46 % of total deaths in Europe [2] and creates high socioeconomic burdens, costing up to $320.1 billion annually in the US [3]. There is no exception for Asian countries: CVD ranks the second for mortality cause in South Korea [4], and explains 25 % for all-cause death in Japan [5]. As nutritional imbalance is a risk factor for CVD, the American Heart Association (AHA) encourages adequate intake of fruits and vegetables [3]. However, a Western diet induces an acid load that overwhelms the base production from vegetables, leading to chronic metabolic acidosis [6]. Recently, acid–base imbalance was suggested to be a risk factor for metabolic disorders [7–9]; higher dietary acid load increased the incidence of type 2 diabetes and hypertension in prospective studies [7, 8] and was associated with insulin resistance [9]. The diet-induced acid load is estimated using a formula that accounts for organic compounds, including the potential renal acid load (PRAL), dietary acid load (DAL), and net endogenous acid production (NEAP). In comparison with NEAP, which is calculated from the ratio of ingested protein and potassium, PRAL and DAL calculations include other materials (calcium, phosphorus, and magnesium) along with protein and potassium. This discrepancy accounts for the bioavailability of nutrients, enabling PRAL and DAL scores to give more accurate predictions of dietary effects on body acidity [10, 11]. Negative values of PRAL and lower DAL values indicate base-forming potential, while positive PRAL scores and higher DAL scores reflect acid-forming potential [12]. To date, studies on acid–base homeostasis have focused on the relationship between dietary patterns and the risk of hypertension or metabolic diseases, but not CVDs. We hypothesized that diet-induced acid load increases the CVD risks regardless of other metabolic conditions. The aim of this study was to investigate the association between diet-induced acid load, using both PRAL and DAL scores, with CVD risk in the general population. Methods Study population This cross-sectional study extracted participant results from the Korea National Health and Nutrition Examination Surveys (KNHANES) 2008–2011. As previously described in detail, the KNHANES is a nationwide, population-based, cross-sectional health examination and annual survey. The survey is conducted by the Division of Chronic Disease Surveillance of the Korea Centers for Disease Control and Prevention in the Ministry of Health and Welfare to monitor the public health and nutrition in South Korea [13]. Each KNHANES is composed of independent data sets from the general population of Korea. As described in Fig. 1, of 37,753 participants from KNHANES 2008–2011, we initially selected those aged 40–79 without prior history of CVD. Subjects excluded were those with missing data for dietary intake and CVD risk assessment components. Ultimately, 11,601 subjects (4813 men and 6788 women) were included in the analysis. All participants provided written informed consent. The survey protocol was approved by the institutional review board of the Korean Centers for Disease Control and Prevention (2008-04EXP-01-C, 2009-07CON-03-2C, 2010-02CON-21-C, and 2011-02CON-06C).Fig. 1 The flow diagram of subject inclusion and exclusion in the Korean National Health and Nutrition Examination Surveys (KNHANES IV and V) Measurement of clinical and laboratory parameters KNHANES data includes a 3-part medical history, nutritional status, and laboratory tests. Subjects’ medical history was evaluated, including smoking, alcohol consumption, exercise, and disease diagnosis or treatment, based on direct interviews and self-reporting. Regular exercise was defined as more than 20 min per session and at least 3 times per week, and heavy drinkers were defined as those whose alcohol consumption was >140 g/week for men and >70 g/week for women. Blood pressure was manually measured using mercury sphygmomanometers (Baumanometer; W.A. Baum, Copiague, NY) three times on the right arms of people in resting, seated positions, and final blood pressure values were assessed by averaging the second and third blood pressure readings. Appendicular skeletal muscle (ASM) was measured using dual-energy X-ray absorptiometry (DXA, QDR 4800A; Hologic Inc., Bedford, MA, USA) and was limited in 8690 subjects. Overnight (8 h) fasting blood and spot urine samples were collected, refrigerated, and transported to the central laboratory institute (NeoDin Medical Institute, Seoul, South Korea) within 24 h. All biochemical samples were measured as previously described [14]. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology collaboration (CKD-EPI) equation [15]. The homeostasis model assessment of insulin resistance (HOMA-IR) was subtracted from the formula as previously described [16]. Cardiovascular disease definition and assessment CVD risk was evaluated using various approaches, including the 10 year atherosclerotic vascular disease (ASCVD) risk score from the 2013 American College of Cardiology (ACC)/AHA guidelines, as well as the Framingham 10 year CVD risk score [17, 18]. In addition, individuals exhibiting >10 % of ACC/AHA ASCVD 10 year risk and >20 % of Framingham CVD risk were classified into the high-risk group [17, 18]. Impaired fasting glucose was defined as fasting plasma glucose ranging from 100 to 125 mg/dL without prior diabetes diagnosis. Diabetes was defined in subjects who were using oral anti-diabetic agents, or exhibited fasting plasma glucose ≥126 mg/dL or glycated hemoglobin (HbA1c) ≥6.5 %. Subjects were considered to have hypertension if their systolic blood pressure was ≥140 mmHg, or their diastolic blood pressure was ≥90 mmHg, or if they were taking anti-hypertensive medications. Body mass index (BMI) was evaluated using the ratio of weight to height (kg/m2), and overweight was defined based on the criteria in the Asians (BMI ≥23 kg/m2) [19]. Chronic kidney disease (CKD) was characterized in subject with less than 60 mL/min/1.73 m2, accordance with CKD guideline [20]. Sarcopenia was defined as an ASM divided by BMI (ASM/BMI) <0.789 for men and <0.512 for women, as recommended by the international consensus meeting from the Foundation for the National Institutes of Health (FNIH) [21]. Because a standard definition of sarcopenia has not yet been established, ASM/height2 and ASM/weight estimations were applied [22, 23]. Sarcopenia was defined as <2 standard deviations (SD) below the sex-specific average for a young (age 18–39), healthy reference population from the datasets (2364 men and 3093 women) [22–24]. The ASM/height2 cutoff was 6.44 kg/m2 for men and 4.48 kg/m2 for women, whereas the ASM/weight cutoff was 28.80 and 22.72 % for men and women, respectively [25]. An age cutoff point to define younger or older adults was 55, the median value. Dietary acid load score assessment Dietary intake information was collected by 1 day face-to-face 24 h recall method by well-trained dietary interviewers aided by various measuring devices (a common size container, standard measuring cups and spoons, a three-dimensional food model, or two-dimensional aids). Moreover, the quality control on interview at filed was conducted throughout the survey by the Center for Nutrition Policy and Promotion at the Korea Health Industry Development Institute with the checklist of the control manual. The 24 h recall method provides high-quality nutrient data and has demonstrated its accuracy in other studies [26, 27]. Detailed information and data resource profile of the survey has been provided elsewhere [28]. To calculate PRAL, intake data were needed for protein, phosphorus, potassium, calcium, and magnesium. Because magnesium intake data were not available in KNHANES, daily magnesium intake values were estimated using an external magnesium content database developed by another research group [29]. We matched the 5105 food items contained in the KNHANES 2008–2011 to the magnesium content database of 366 foods commonly consumed by Koreans. Foods not included in the database were indirectly matched with similar foods in the database. Consumption frequencies of foods by this matching process contributed 65.4 % of the total consumption frequency in KNHANES. PRAL and DAL scores were derived from the equations of nutrient intakes and tertiles of the scores were used for statistical analysis: PRAL (mEq/day) = (0.49 × protein [g/day]) + (0.037 × phosphorus [mg/day]) − (0.021 × potassium [mg/day]) − (0.026 × magnesium [mg/day]) − (0.013 × calcium [mg/day]) [30], and DAL (mEq/day) = PRAL + (body surface area [m2] × 41 [mEq/day]/1.73 m2) [10]. Both PRAL and DAL score was categorized into sex-specific tertiles. Body surface area was calculated by the Du Bois formula: 0.007184 × height0.725 × weight0.425 [31, 32]. Because data on body surface area were missing for 36 subjects, 11,565 subjects with DAL scores were newly categorized according to DAL tertiles. Statistical analysis Data are presented as the mean ± SD for continuous variables, and presented as number (N), or percent (%) for categorical variables. We analyzed the study participants’ characteristics according to PRAL tertiles, using one-way analysis of variance (ANOVA) to compare continuous variables, χ2 tests for categorical variables followed by post hoc analyses with the Bonferroni method. To evaluate the association between acid load and CVD risk, the effects of comorbidities should be minimized. Subjects were divided according to the presence of obesity, sarcopenia, and regular exercise, and then χ2 tests were applied for each group. The higher PRAL group included both the second and highest PRAL tertiles, and this group was compared against the lowest PRAL tertile. Multiple logistic regression analysis was used to assess the independent association between acid load and high CVD risk (>10 % of ACC/AHA ASCVD 10 year risk score or >20 % of Framingham 10 year risk score) including other covariates. Fasting insulin, HOMA-IR, total cholesterol, triglyceride, HDL cholesterol, and LDL cholesterol values were not normally distributed; analyses were performed using log- and back-transformed data. Statistical analyses were performed using IBM SPSS version 20.0 for Windows (IBM Corp., Armonk, NY, USA); P < 0.05 was considered statistically significant. Results Diet-induced acid load and baseline characteristics In total, 11,601 individuals free from CVD were analyzed in the present study (Fig. 1). Individuals with a higher diet-induced acid load (higher PRAL and DAL scores) tended to have increased blood pressure, triglyceride levels, and metabolic syndrome (Tables 1, 2). They also reported higher incidences of smoking cigarettes, drinking alcohol, and exercising less. Estimated dietary acid–base load also corresponded with urine acidity, reflecting the body’s acid–base homeostasis. In contrast, mean BMI, LDL, and HDL cholesterol and the proportions of CKD were similar across the PRAL tertiles. Regarding DAL scores, more discrepancies among the tertiles were observed; higher BMI, lower HDL and increased insulin resistance were found in the highest DAL tertile.Table 1 Baseline characteristics of study population by categories of PRAL Tertiles of dietary PRAL (mEq/day) P value* T1 (n = 3859) T2 (n = 3540) T3 (n = 4202) PRAL (mEq/d)a −21.8 (−199.8 ~ −7.5) −2.9‡ (−11.3 ~ 8.2) 11.2‡ (−10.4 ~ 181.4) <0.001 DAL (mEq/d) 11.1 ± 20.8 35.9 ± 6.5‡ 53.4 ± 14.6‡ <0.001 Magnesium (mg/d) 388.0 ± 186.4 278.1 ± 139.4‡ 307.8 ± 177.2‡ <0.001 Age (year) 56.2 ± 10.3 57.1 ± 11.0‡ 56.8 ± 11.2‡ 0.001 Male (%) 41.4 41.3 41.8 0.881 Waist circumference (cm) 82.7 ± 9.0 82.5 ± 9.2 82.9 ± 9.3 0.255 BMI (kg/m2) 24.1 ± 3.0 23.9 ± 3.1‡ 24.0 ± 3.1 0.041 Systolic blood pressure (mmHg) 120.4 ± 17.2 122.1 ± 18.0‡ 122.2 ± 17.6‡ <0.001 Diastolic blood pressure (mmHg) 76.3 ± 10.1 76.8 ± 10.6 77.2 ± 10.5‡ 0.001 Fasting plasma glucose (mg/dL) 101.4 ± 27.5 100.5 ± 23.3 100.6 ± 24.9 0.213 Total cholesterol (mg/dL)b 193.2 ± 36.1 194.1 ± 36.7 194.3 ± 36.5 0.436 HDL cholesterol (mg/dL)b 51.1 ± 12.3 51.1 ± 12.6 51.2 ± 12.5 0.842 Triglycerides (mg/dL)b 138.8 ± 102.7 145.5 ± 119.8‡ 144.7 ± 113.5‡ 0.004 LDL cholesterol (mg/dL)b 119.0 ± 32.4 119.2 ± 32.6 119.0 ± 32.3 0.043 Creatinine (mg/dL) 0.8 ± 0.2 0.8 ± 0.2 0.8 ± 0.3 0.903 eGFR (mL/min/1.73 m2) 88.3 ± 14.9 87.9 ± 15.0 88.5 ± 15.5 0.193 Insulin (ųU/mL)b 9.8 ± 5.2 9.8 ± 5.9 10.0 ± 6.5 0.807 HOMA-IRb 2.5 ± 1.9 2.5 ± 1.8 2.5 ± 2.5 0.962 Urine pH 5.8 ± 0.9 5.7 ± 0.9 5.7 ± 0.8‡ 0.013 Heavy drink (%) 12.3 14.1 17.0‡ <0.001 Current smoking (%) 15.4 17.8‡ 20.0‡ <0.001 Exercise (%) 28.3 24.6‡ 24.6‡ <0.001 Hypertension (%) 34.4 37.9‡ 38.5‡ <0.001 IFG/Diabetes (%) 23.3/13.4 24.2/13.4 25.0/12.7 0.427 Metabolic syndrome (%) 37.1 38.8 40.2‡ 0.014 Chronic kidney disease (%) 3.8 4.3 4.5 0.263 Region (Metro/City/Rural,  %) 43.3/30.2/26.5 42.3/31.5/26.3 41.7/30.2/28.1 0.237 Family history of cardio- or cerebro-vascular disease (%) 6.6 6.3 6.7 0.811 PRAL potential renal acid load; DAL dietary acid load; BMI body mass index; HDL cholesterol, high density lipoprotein cholesterol; LDL cholesterol, low density lipoprotein cholesterol; eGFR estimated glomerular filtration rate; HOMA-IR homeostasis model assessment of insulin resistance; IFG impaired fasting glucose ‡P < 0.05 by post hoc analyses when compared with lowest tertiles *Chi square tests for qualitative variables and ANOVA tests for quantitative variables aValues are medians per tertile bLog-transformed Table 2 Baseline characteristics of study population by categories of DAL Tertiles of dietary DAL (mEq/day) P value* T1 (n = 3900) T2 (n = 3839) T3 (n = 3826) DAL (mEq/d)a 16.9 (−165.5 ~ 33.9) 36.1‡ (25.9 ~ 48.5) 51.9‡ (39.3 ~ 221.1) <0.001 PRAL (mEq/d) −27.5 ± 20.5 −2.2 ± 5.3‡ 15.4 ± 13.4‡ <0.001 Magnesium (mg/d) 381.4 ± 187.5 284.0 ± 146.0‡ 310.4 ± 177.0‡ <0.001 Age (year) 56.8 ± 10.5 57.4 ± 11.1‡ 55.9 ± 10.9‡ <0.001 Male (%) 40.8 40.5 43.0 0.062 Waist circumference (cm) 81.7 ± 8.8 82.1 ± 9.1 84.4 ± 9.4‡ <0.001 BMI (kg/m2) 23.7 ± 2.9 23.7 ± 3.1 24.5 ± 3.2‡ <0.001 Systolic blood pressure (mmHg) 120.5 ± 17.3 122.2 ± 18.3‡ 121.9 ± 17.2‡ <0.001 Diastolic blood pressure (mmHg) 76.1 ± 10.1 76.6 ± 10.5 77.6 ± 10.5‡ <0.001 Fasting plasma glucose (mg/dL) 101.3 ± 27.7 100.4 ± 24.5 100.9 ± 24.5 0.268 Total cholesterol (mg/dL)b 193.0 ± 36.2 194.2 ± 36.8 194.5 ± 36.2 0.146 HDL cholesterol (mg/dL)b 51.5 ± 12.4 51.3 ± 12.6 50.7 ± 12.3‡ 0.031 Triglycerides (mg/dL)b 137.4 ± 110.9 142.8 ± 106.1‡ 148.7 ± 118.9‡ <0.001 LDL cholesterol (mg/dL)b 118.7 ± 32.5 119.4 ± 32.8 119.0 ± 32.0 0.613 Creatinine (mg/dL) 0.8 ± 0.2 0.8 ± 0.2 0.9 ± 0.3 0.085 GFR, EPI (mL/min/1.73 m2) 88.3 ± 14.9 87.7 ± 15.1 88.8 ± 15.5 0.011 Insulin (μU/mL)b 9.5 ± 4.8 9.8 ± 5.8 10.3 ± 6.7‡ <0.001 HOMA-IRb 2.4 ± 1.8 2.5 ± 1.8 2.6 ± 2.6‡ <0.001 Urine pH 5.8 ± 0.9 5.7 ± 0.9 5.7 ± 0.8‡ 0.001 Heavy drink (%) 12.1 13.9 17.8‡ <0.001 Current smoking (%) 15.7 17.4 20.3‡ <0.001 Exercise (%) 27.9 24.4‡ 25.3 <0.001 Hypertension (%) 34.4 37.6‡ 38.7‡ <0.001 IFG/diabetes (%) 22.8/13.3 24.0/13.2 25.9/13.0 0.918 Metabolic syndrome (%) 35.1 38.1‡ 43.1‡ <0.001 Chronic kidney disease (%) 3.8 4.5 4.2 0.376 Region (Metro/City/rural,  %) 42.5/30.2/27.3 43.2/30.6/26.3 41.6/30.9/27.5 0.613 Family history of cardio- or cerebro-vascular disease (%) 6.5 6.6 6.6 0.984 ‡P < 0.05 by post hoc analyses when compared with lowest tertiles *Chi square tests for qualitative variables and ANOVA tests for quantitative variables PRAL potential renal acid load; DAL dietary acid load; BMI body mass index; HDL cholesterol, high density lipoprotein cholesterol; LDL cholesterol, low density lipoprotein cholesterol; eGFR estimated glomerular filtration rate; HOMA-IR homeostasis model assessment of insulin resistance; IFG impaired fasting glucose aValues are medians per tertile bLog-transformed Diet-induced acid load and cardiovascular disease risks The average 10 year ACC/AHA ASCVD risk score increased as PRAL increased (from 8.5 to 9.6 %, P < 0.001, Fig. 2a), and the discrepancy among the tertiles was greater in the older age group (≥55 years) (Fig. 2b). Overall, the proportion of patients in the high-risk group (predicted 10 year risk of hard ACC/AHA ASCVD event >10 %) in our population was 32.6 %, and the distribution of high-risk group members and PRAL tertiles showed the same trend as ACC/AHA ASCVD scores (29.5 and 34.0 % for lowest and highest tertiles, respectively, P < 0.001, Fig. 2c). The link between higher PRAL score and high-risk group classification was observed in older age group (Fig. 2d). Similarly, higher PRAL scores were associated with CVD risk that was estimated by a different method, the Framingham 10 year risk score (Fig. 3). Regarding DAL score, the correlation between higher DAL score and increased ACC/AHA ASCVD score was only observed in young age group (Fig. 4). As CVD risk gradually increases according to age, we stratified subjects using 10 year units, and evaluated PRAL scores and ASCVD risk scores (Fig. 5). The proportion of individuals with higher ASCVD risk was increased in the higher PRAL population, especially among middle-aged individuals (50–59 years old). A similar trend was also observed using DAL scores.Fig. 2 Differences in CVD risk according to PRAL tertiles. a Average ACC/AHA ASCVD 10 year risk scores. b Average ACC/AHA ASCVD 10 year risk scores in age groups. c Proportion of individuals with high ACC/AHA ASCVD 10 year risk (>10 %). d Proportion of individuals with high ACC/AHA ASCVD 10 year risk (>10 %) in age groups. Mean ± standard errors, *P < 0.05, **P < 0.001 Fig. 3 Differences in the CVD risk according to PRAL tertiles. a Average Framingham 10 year risk scores. b Average Framingham 10 year risk scores in age groups. c Proportion of individuals with high Framingham 10 year risk (>20 %). d Proportion of individuals with high Framingham 10 year risk (>20 %) in age groups. Mean ± standard errors, *P < 0.05, **P < 0.001 Fig. 4 Differences in the CVD risk according to DAL tertiles. a Average ACC/AHA ASCVD 10 year risk scores. b Average ACC/AHA ASCVD 10 year risk scores in age groups. c Proportion of individuals with high ACC/AHA ASCVD 10 year risk (>10 %). d Proportion of individuals with high ACC/AHA ASCVD 10 year risk (>10 %) in age groups. Mean ± standard errors, *P < 0.05, **P < 0.001 Fig. 5 The proportion of individuals with high ACC/AHA ASCVD 10 year risk (>10 %) stratified by age groups. a Diet-induced acid load defined by PRAL score. b Diet-induced acid load defined by DAL score. Dark and light boxes indicate the high-score group and the lowest score group, respectively. *P < 0.05, **P < 0.001 Diet-induced acid load is linked with CVD risk independent of obesity, exercise, and insulin resistance, but not sarcopenia As overweight and low exercise are considered important CVD risk factors, we stratified the study population according to these variables. Higher PRAL scores raised the proportion of high ACC/AHA ASCVD risk regardless of BMI (Odds Ratio [OR] ranged from 1.19 to 1.37, P < 0.05 for all groups), and this association was stronger in the lean group (Fig. 6a). Interestingly, lean subjects with higher PRAL scores exhibited a similar proportion of higher CVD risk compared to those in the obese group. Although it was not statistically significant, more lean subjects with higher PRAL scores were in the ACC/AHA ASCVD high-risk group than obese subjects with lower PRAL scores. A high-PRAL diet increased ACC/AHA ASCVD risk group distribution independent of exercise status (OR ranged from 1.23 to 1.27, P < 0.05 for all groups, Fig. 6b). Moreover, the lowest PRAL individuals who did not exercise showed comparable ACC/AHA ASCVD risk with those in the higher PRAL group who regularly exercised. Increased ACC/AHA ASCVD risks in the higher PRAL group were also observed regardless of insulin resistance status assessed by HOMA-IR (Fig. 6c). Regarding skeletal muscle mass, the association between PRAL and high-risk ACC/AHA ASCVD was stronger among subjects with preserved skeletal muscle mass (OR 1.17, 95 % CI 1.06–1.30, Fig. 6d); this association was not observed among sarcopenic subjects (OR 1.17, 95 % CI 0.81–1.69). This trend was consistently observed when other sarcopenia definitions were applied (Fig. 7).Fig. 6 Difference in CVD risk according to PRAL scores, stratified by metabolic status and physical activity. a Proportion of individuals with high ACC/AHA ASCVD 10 year risk stratified by overweight defined as BMI ≥23 kg/m2, b regular exercise, c HOMA-IR with a cutoff point of 2.5, and d sarcopenia defined as ASM/BMI definition. The data are presented as OR with 95 % CI, NS non-significance; *P < 0.05, **P < 0.001 Fig. 7 Proportion of individuals with high ACC/AHA ASCVD 10 year risk according to sarcopenic status. a ASM/height2 definition. b ASM/weight definition. The data are presented as OR with 95 % CI, NS non-significance; *P < 0.05, **P < 0.001 Higher PRAL and DAL scores are associated with elevated predicted risk of CVD regardless of hypertension or diabetes We assessed high predicted ASCVD risk and diet-induced acid load after adjusting for confounding factors, including age, sex, and other clinic-laboratory parameters. Multiple logistic regression analysis showed that higher PRAL scores was independently associated with high risk of CVD evaluated from both ACC/AHA assessment and Framingham risk score (Tables 3, 4). The effects of PRAL on ACC/AHA ASCVD risk showed no significant difference in patients categorized by hypertension, by age or by overweight (hypertension; P = 0.345, age; P = 0.112, overweight; P = 0.150, respectively, for interaction). Although this statistical significance was weakened in DAL score, the similar trend was observed in DAL tertiles (Table 5).Table 3 Odds ratio and 95 % confidential interval of high ASCVD risk (>10 % 2013 ACC/AHA score) according to PRAL tertiles in adults Lowest tertile Second tertile Highest tertile Crude 1 (referent) 1.27 (1.15–1.40) 1.23 (1.22–1.35) Model 1 1 (referent) 1.29 (1.11–1.51) 1.18 (1.02–1.36) Model 2 1 (referent) 1.28 (1.10–1.50) 1.17 (1.01–1.35) Model 3 1 (referent) 1.31 (1.09–1.58) 1.20 (1.01–1.43) Model 1: adjusted for age (per 5 years), and sex Model 2: adjusted for age (per 5 years), sex, exercise, and family history of cardio- and cerebro-vascular disease Model 3: adjusted for age (per 5 years), sex, exercise, family history of cardio- and cerebro-vascular disease, diabetes, hypertension, LDL cholesterol, eGFR, and urine pH ASCVD 10 year atherosclerotic vascular disease; PRAL potential renal acid load; eGFR estimated glomerular filtration rate; LDL cholesterol; low density lipoprotein cholesterol Table 4 Odds ratio and 95 % confidential interval of high ASCVD risk (>10 % 2013 ACC/AHA score) according to DAL tertiles in adults Lowest tertile Second tertile Highest tertile Crude 1 (referent) 1.21 (1.10–1.33) 1.00 (0.90–1.09) Model 1 1 (referent) 1.24 (1.07–1.44) 1.16 (1.00–1.35) Model 2 1 (referent) 1.23 (1.06–1.43) 1.15 (0.99–1.34) Model 3 1 (referent) 1.25 (1.04–1.46) 1.07 (0.89–1.28) Model 1: adjusted for age (per 5 years), and sex Model 2: adjusted for age (per 5 years), sex, exercise, and family history of cardio- and cerebro-vascular disease Model 3: adjusted for age (per 5 years), sex, exercise, family history of cardio- and cerebro-vascular disease, diabetes, hypertension, LDL cholesterol, eGFR, and urine pH ASCVD 10 year atherosclerotic vascular disease; DAL dietary acid load; eGFR estimated glomerular filtration rate; LDL cholesterol; low density lipoprotein cholesterol Table 5 Odds ratio and 95 % confidential interval of high Framingham 10 year risk (>20 %) according to PRAL tertiles in adults Lowest tertile Second tertile Highest tertile Crude 1 (referent) 1.22 (1.10–1.36) 1.18 (1.06–1.31) Model 1 1 (referent) 1.19 (1.03–1.36) 1.16 (1.01–1.32) Model 2 1 (referent) 1.18 (1.03–1.36) 1.15 (1.01–1.31) Model 3 1 (referent) 1.25 (1.05–1.49) 1.19 (1.01–1.41) Model 1: adjusted for age (per 5 years), and sex Model 2: adjusted for age (per 5 years), sex, exercise, and family history of cardio- and cerebro-vascular disease Model 3: adjusted for age (per 5 years), sex, exercise, family history of cardio- and cerebro-vascular disease, diabetes, hypertension, LDL cholesterol, eGFR, and urine pH ASCVD 10 year atherosclerotic vascular disease; PRAL potential renal acid load; eGFR estimated glomerular filtration rate; LDL cholesterol; low density lipoprotein cholesterol Discussion In this current large, nationally representative, population-based study, we demonstrated that people with higher diet-induced acid load had higher CVD risks in the general population. Diet-induced acid load was closely linked with CVD risk especially among middle-aged individuals. In addition, higher PRAL scores were associated with CVD risk independent of obesity, exercise, and insulin resistance, but not sarcopenia. This association remained significant after adjusting for other essential confounding factors. To date, the health effects of acid–base imbalance were mainly investigated in bone mass [33], and kidney stones [34]. Recently, studies on systemic metabolism were conducted that extended the effects of diet-induced acid load on one’s body. For example, a prospective study showed that higher PRAL scores correlated with the incidence of type 2 diabetes (hazard ratio = 1.56, 95 % CI 1.29‒1.90) [7]. Similarly, a Japanese study reported men with highest PRAL quartile had 61 % increased type 2 diabetes prevalence in over 5 years follow-up period [35]. In addition, individuals with higher diet-induced acid load were reported to exhibit 27 % increases on hypertension development regardless of age [8], and this positively correlated with the prevalence of nonalcoholic fatty liver disease [36]. In this study, we also demonstrated higher incidences of increased blood pressure, prevalence of hypertension and metabolic syndrome in conjunction with higher PRAL and DAL scores. In addition, individuals with higher diet-induced acid load tended to have unhealthy lifestyle patterns during relatively younger ages, including less physical activity, stricter adherence to Western diet patterns, and higher BMI, in accordance with previous studies [7, 8, 36]. This was consistent with our findings that in the highest PRAL score group, individuals were more likely to exhibit sedentary patterns, increased waist circumferences, and high prevalence of smoking and drinking alcohol. Diet-induced acid load, cardiovascular diseases and possible mechanisms The mechanism linking diet-induced acid load and metabolic disease incidence is mainly reported as insulin resistance [7, 9]. Toward that end, insulin binding affinity to its receptor was markedly decreased in individuals with metabolic acidosis [37]. Thus, even in healthy individuals, a very low degree of metabolic acidosis could lead to insulin resistance, resulting hyperglycemia [38]. Moreover, a previous study reported that attenuating metabolic acidosis could increase insulin sensitivity [39]. In accordance with this evidence, our study results demonstrated that the highest DAL tertile was associated with increased values of HOMA-IR. Regarding the effect of diet-induced acid load on blood pressure, several mechanisms have been suggested. A diet depleted of potassium could affect vasodilatation and be toxic to the blood vessels [40]. Potassium restriction results in intracellular potassium deficits and causes compensatory sodium gains in cells to maintain tonicity and volume [40]. In a human study, even a 10 day period of low potassium intake increased systolic blood pressure by 5 mmHg (P < 0.02) and modified salt sensitivity, resulting in exacerbation of hypertension [41]. Conversely, hypertensive rats fed a potassium-rich diet showed decreases in blood pressure and stroke development [42, 43]. Likewise, our study demonstrated that PRAL and DAL scores were positively associated with both systolic and diastolic blood pressure values, as well as hypertension prevalence (OR 1.19, 95 % CI 1.09‒1.31). In addition, systemic metabolic acidosis, caused by the Western diet, is associated with excessive cortisol levels and leads to ammoniagenesis, which may lead to loss of renal function [12, 44]. Diet-induced insulin resistance could autonomously promote cardiovascular disease in various pathways; impair coronary microcirculatory function [45], stimulate conduction dysfunction and increase arrhythmogenesis [46]. Diet-induced acid load in unhealthy states Previously, the correlation between diet-acid load and metabolic disease was stronger in non-obese individuals, suggesting that it is independent of adiposity [7, 9]. We evaluated the PRAL score and ACC/AHA ASCVD risk according to BMI as well as insulin resistance. Although this association remained significant in both lean/overweight and insulin sensitive/resistant groups, more substantial correlations were observed in lean individuals (OR 1.37 vs 1.19) and in the insulin sensitive group (OR 1.34 vs 1.21). Moreover, there was no significant difference in risk of ACC/AHA ASCVD between lean subjects with higher PRAL scores and those in the overweight group in the lowest PRAL tertile, suggesting that CVD risk of lean individuals would increase if diet-induced acid loads were elevated. An epidemiologic study that reported that dietary patterns could increase the risk of obesity, even in lean individuals, suggests that diet affects metabolism in non-obese populations [47]. In addition, if sedentary individuals lowered their intake of dietary acid, they might exhibit comparable CVD risks with people who regularly exercised. While obesity and exercise are major contributing factors to CVD risk, diet-induced acid load might modify these effects. Regardless, the association between diet-induced acid load and CVD risk with respect to muscle mass seems compelling. There has been some evidence that sarcopenia could influence CVD, such as increased arterial stiffness and inflammatory markers in people with sarcopenia [48, 49]. Our results also demonstrated increased CVD risk in the sarcopenic group compared to the non-sarcopenic group; however, the association between PRAL score and ACC/AHA ASCVD risk did not remain significant among sarcopenic individuals. Furthermore, subjects with higher PRAL scores but preserved muscle mass had lower CVD risk than those in the sarcopenic group with lower PRAL scores. This might suggest that skeletal muscle mass is more closely associated with lower CVD risks compared to healthy diet patterns. This study has some limitations. First, our cross-sectional study design could not conclude causality between diet-induced acid load and CVD risk. Second, PRAL and DAL scores were derived from self-reports using a 24 h recall method, which only confirmed the short term dietary intakes of the study participants; the daily variation in food consumption could not be considered. Third, we could not assess the quality of protein that was used to calculate PRAL and DAL scores. In addition, KHNASE did not collect each individual’s medication information, and not evaluate physical performance status, and muscle strength. Sarcopenic status was classified by muscle mass alone, not considering muscle strength or physical performance. The potential medication effect on CVD risk was not considered in our results. Despite these limitations, the current study had several strengths. First, it was a large population study based on national data, guaranteeing the statistical reliability of results. As KNHANES represents the non-institutionalized general population, selection bias was minimized. Second, this investigation provided strong evidence of a close relationship between diet-induced acid load and CVD risk by adjusting for other confounding factors and conducting stratification analyses. To the best of our knowledge, this is the first large-scale study to estimate the CVD risk associated with diet-induced acid load. We assessed diet-induced acid load using variable methods: PRAL and DAL scores. The correlation between PRAL and DAL scores was high (Pearson correlation coefficient = 0.99, P < 0.001) in our study, which was consistent with previous studies [10]. Third, individual CVD risks were estimated using various equations, and comparable results were derived regardless of these different models. Fourth, although the current study was cross-sectional, we limited the study population by excluding those with prior CVD history to prevent reverse causality. In the light of dietary pattern and CVD risk, our study results demonstrate that individuals with higher diet-induced acid load are more vulnerable to CVD risks. Despite advances in CVD diagnostic and treatment modalities, prevention is the most high-leverage action, and therefore, modifiable risk factors are important. Prospective, well-designed, longitudinal studies with sufficient laboratory and cardiovascular-imaging resources are warranted to elucidate the complex relationship between diet-induced acid load and CVD risk. If the causality between diet-induced acid load and CVD risk, as well as the cutoff value of diet-induced acid load are clearly investigated, then meaningful, comprehensive, and practical dietary recommendations would be provided to the public. Conclusions This nationwide survey of a representative sample of the Korean population demonstrated that diet-induced acid load was associated with increased risks of CVD, independent of other cardiovascular factors. Abbreviations ACC/AHAAmerican College of Cardiology/American Heart Association ASCVDatherosclerotic vascular disease BMIbody mass index CIconfidential interval CVDcardiovascular disease DALdietary acid load HDLhigh density lipoprotein HOMA-IRhomeostasis model assessment of insulin resistance KNHANESThe Korea National Health and Nutrition Examination Surveys LDLlow density lipoprotein PRALpotential renal acid load Authors’ contributions Study concept and design: EH, YHL and DWK. Analysis and interpretation of data: EH, GK, NH, DWK and YHL. Drafting of the manuscript: EH and YHL. Critical revision of the manuscript for important intellectual content: HJS, BWL, ESK, IKL and BSC. Statistical analysis: EH and YHL. Administrative, technical, or material support: BWL, ESK, IKL and BSC. All authors read and approved the final manuscript. Acknowledgements The authors are very grateful to officers who conducted KNHANES. Competing interests The authors declare that they have no competing interests. Availability of data and material All data files are derived from the Division of Chronic Disease Surveillance of the Korea Centers for Disease Control and Prevention in the Ministry of Health and Welfare database, http://knhanes.cdc.go.kr/knhanes/index/do, which is publically available to researchers who agree to conform to ethical research principles. Consent for publication Not applicable. Ethics approval and consent to participate All participants provided written informed consent. The survey protocol was approved by the institutional review board of the Korean Centers for Disease Control and Prevention (2008-04EXP-01-C, 2009-07CON-03-2C, 2010-02CON-21-C, and 2011-02CON-06C). Funding This research was supported by the grant from the Korea Healthcare Technology Research & Development (R&D) Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (HI14C2476 and HI16C1501) and Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, Information and Communication Technology (ICT) & Future Planning (2015R1C1A1A01052558). ==== Refs References 1. Mortality GBD Causes of Death C Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013 Lancet 2015 385 9963 117 171 10.1016/S0140-6736(14)61682-2 25530442 2. Nichols M Townsend N Scarborough P Rayner M Cardiovascular disease in Europe 2014: epidemiological update Eur Heart J 2014 35 42 2950 2959 10.1093/eurheartj/ehu299 25139896 3. Mozaffarian D Benjamin EJ Go AS Arnett DK Blaha MJ Cushman M de Ferranti S Despres JP Fullerton HJ Howard VJ Heart disease and stroke statistics–2015 update: a report from the American Heart Association Circulation 2015 131 4 e29 e322 10.1161/CIR.0000000000000152 25520374 4. Lee KS Park JH Burden of disease in Korea during 2000–10 J Public Health 2014 36 2 225 234 10.1093/pubmed/fdt056 5. Ohira T Iso H Cardiovascular disease epidemiology in Asia: an overview Circ J 2013 77 7 1646 1652 10.1253/circj.CJ-13-0702 23803294 6. Frassetto LA Todd KM Morris RC Jr Sebastian A Estimation of net endogenous noncarbonic acid production in humans from diet potassium and protein contents Am J Clin Nutr 1998 68 3 576 583 9734733 7. Fagherazzi G Vilier A Bonnet F Lajous M Balkau B Boutron-Rualt MC Clavel-Chapelon F Dietary acid load and risk of type 2 diabetes: the E3N-EPIC cohort study Diabetologia 2014 57 2 313 320 10.1007/s00125-013-3100-0 24232975 8. Zhang L Curhan GC Forman JP Diet-dependent net acid load and risk of incident hypertension in United States women Hypertension 2009 54 4 751 755 10.1161/HYPERTENSIONAHA.109.135582 19667248 9. Akter S Eguchi M Kuwahara K Kochi T Ito R Kurotani K Tsuruoka H Nanri A Kabe I Mizoue T High dietary acid load is associated with insulin resistance: The Furukawa Nutrition and Health Study Clin Nutr 2016 35 453 459 10.1016/j.clnu.2015.03.008 25863769 10. Remer T Manz F Potential renal acid load of foods and its influence on urine pH J Am Diet Assoc 1995 95 7 791 797 10.1016/S0002-8223(95)00219-7 7797810 11. Michaud DS Troiano RP Subar AF Runswick S Bingham S Kipnis V Schatzkin A Comparison of estimated renal net acid excretion from dietary intake and body size with urine pH J Am Diet Assoc 2003 103 8 1001 1007 10.1016/S0002-8223(03)00469-3 12891148 12. Engberink MF Bakker SJ Brink EJ van Baak MA van Rooij FJ Hofman A Witteman JC Geleijnse JM Dietary acid load and risk of hypertension: the Rotterdam Study Am J Clin Nutr 2012 95 6 1438 1444 10.3945/ajcn.111.022343 22552032 13. Lee YH Kim SU Song K Park JY Kim do Y Ahn SH Lee BW Kang ES Cha BS Han KH Sarcopenia is associated with significant liver fibrosis independently of obesity and insulin resistance in nonalcoholic fatty liver disease: Nationwide surveys (KNHANES 2008–2011) Hepatology 2016 63 3 776 786 10.1002/hep.28376 26638128 14. Lee YH Armstrong EJ Kim G Oh J Kang SM Lee BW Ahn CW Cha BS Lee HC Mantzoros CS Undiagnosed diabetes is prevalent in younger adults and associated with a higher risk cardiometabolic profile compared to diagnosed diabetes Am Heart J 2015 170 4 760 769 10.1016/j.ahj.2015.07.024 26386800 15. Levey AS Stevens LA Schmid CH Zhang YL Castro AF 3rd Feldman HI Kusek JW Eggers P Van Lente F Greene T A new equation to estimate glomerular filtration rate Ann Intern Med 2009 150 9 604 612 10.7326/0003-4819-150-9-200905050-00006 19414839 16. Yokoyama H Emoto M Fujiwara S Motoyama K Morioka T Komatsu M Tahara H Shoji T Okuno Y Nishizawa Y Quantitative insulin sensitivity check index and the reciprocal index of homeostasis model assessment in normal range weight and moderately obese type 2 diabetic patients Diabetes Care 2003 26 8 2426 2432 10.2337/diacare.26.8.2426 12882874 17. Goff DC Jr Lloyd-Jones DM Bennett G Coady S D’Agostino RB Gibbons R Greenland P Lackland DT Levy D O’Donnell CJ 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines Circulation 2014 129 25 Suppl 2 S49 S73 10.1161/01.cir.0000437741.48606.98 24222018 18. D’Agostino RB Sr Vasan RS Pencina MJ Wolf PA Cobain M Massaro JM Kannel WB General cardiovascular risk profile for use in primary care: the Framingham Heart Study Circulation 2008 117 6 743 753 10.1161/CIRCULATIONAHA.107.699579 18212285 19. Consultation WHOE Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies Lancet 2004 363 9403 157 163 10.1016/S0140-6736(03)15268-3 14726171 20. National Kidney F K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification Am J Kidney 2002 39 2 Suppl 1 S1 266 21. Studenski SA Peters KW Alley DE Cawthon PM McLean RR Harris TB Ferrucci L Guralnik JM Fragala MS Kenny AM The FNIH sarcopenia project: rationale, study description, conference recommendations, and final estimates J Gerontol A Biol Sci Med Sci. 2014 69 5 547 558 10.1093/gerona/glu010 24737557 22. Baumgartner RN Koehler KM Gallagher D Romero L Heymsfield SB Ross RR Garry PJ Lindeman RD Epidemiology of sarcopenia among the elderly in New Mexico Am J Epidemiol 1998 147 8 755 763 10.1093/oxfordjournals.aje.a009520 9554417 23. Janssen I Heymsfield SB Ross R Low relative skeletal muscle mass (sarcopenia) in older persons is associated with functional impairment and physical disability J Am Geriatr Soc 2002 50 5 889 896 10.1046/j.1532-5415.2002.50216.x 12028177 24. Cruz-Jentoft AJ Baeyens JP Bauer JM Boirie Y Cederholm T Landi F Martin FC Michel JP Rolland Y Schneider SM Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People Age Ageing 2010 39 4 412 423 10.1093/ageing/afq034 20392703 25. Lee YH Kim JE Roh YH Choi HR Rhee Y Kang DR Lim SK The combination of vitamin D deficiency and mild to moderate chronic kidney disease is associated with low bone mineral density and deteriorated femoral microarchitecture: results from the KNHANES 2008-2011 J Clin Endocrinol Metab 2014 99 10 3879 3888 10.1210/jc.2013-3764 24878040 26. Schatzkin A Kipnis V Carroll RJ Midthune D Subar AF Bingham S Schoeller DA Troiano RP Freedman LS A comparison of a food frequency questionnaire with a 24-hour recall for use in an epidemiological cohort study: results from the biomarker-based Observing Protein and Energy Nutrition (OPEN) study Int J Epidemiol 2003 32 6 1054 1062 10.1093/ije/dyg264 14681273 27. Thompson FE Dixit-Joshi S Potischman N Dodd KW Kirkpatrick SI Kushi LH Alexander GL Coleman LA Zimmerman TP Sundaram ME Comparison of interviewer-administered and automated self-administered 24-hour dietary recalls in 3 diverse integrated health systems Am J Epidemiol 2015 181 12 970 978 10.1093/aje/kwu467 25964261 28. Kweon S Kim Y Jang MJ Kim Y Kim K Choi S Chun C Khang YH Oh K Data resource profile: the Korea National Health and Nutrition Examination Survey (KNHANES) Int J Epidemiol 2014 43 1 69 77 10.1093/ije/dyt228 24585853 29. Bae YJ Kim MH Choi MK Analysis of magnesium contents in commonly consumed foods and evaluation of its daily intake in Korean independent-living subjects Biol Trace Elem Res 2010 135 1–3 182 199 10.1007/s12011-009-8511-x 19756403 30. Remer T Dimitriou T Manz F Dietary potential renal acid load and renal net acid excretion in healthy, free-living children and adolescents Am J Clin Nutr 2003 77 5 1255 1260 12716680 31. DuBois D DuBois EF A formula to estimate the approximate surface area if height and weight be known Arch Intern Med 1916 17 6 863 871 10.1001/archinte.1916.00080130010002 32. Verbraecken J Van de Heyning P De Backer W Van Gaal L Body surface area in normal-weight, overweight, and obese adults. A comparison study Metab Clin Exp 2006 55 4 515 524 10.1016/j.metabol.2005.11.004 16546483 33. Alexy U Remer T Manz F Neu CM Schoenau E Long-term protein intake and dietary potential renal acid load are associated with bone modeling and remodeling at the proximal radius in healthy children Am J Clin Nutr. 2005 82 5 1107 1114 16280446 34. Reddy ST Wang CY Sakhaee K Brinkley L Pak CY Effect of low-carbohydrate high-protein diets on acid-base balance, stone-forming propensity, and calcium metabolism Am J Kidney Dis 2002 40 2 265 274 10.1053/ajkd.2002.34504 12148098 35. Akter S Kurotani K Kashino I Goto A Mizoue T Noda M Sawada N Tsugane S Japan Public Health Center-based Prospective Study G High dietary acid load score is associated with increased risk of type 2 diabetes in Japanese men: the Japan Public Health Center-based Prospective Study J Nutr 2016 146 5 1076 1083 10.3945/jn.115.225177 27052540 36. Chan R Wong VW Chu WC Wong GL Li LS Leung J Chim AM Yeung DK Sea MM Woo J Higher estimated net endogenous Acid production may be associated with increased prevalence of nonalcoholic Fatty liver disease in chinese adults in Hong Kong PLoS One 2015 10 4 e0122406 10.1371/journal.pone.0122406 25905490 37. Whittaker J Cuthbert C Hammond VA Alberti KG The effects of metabolic acidosis in vivo on insulin binding to isolated rat adipocytes Metab Clin Exp 1982 31 6 553 557 10.1016/0026-0495(82)90094-4 6281619 38. DeFronzo RA Beckles AD Glucose intolerance following chronic metabolic acidosis in man Am J Physiol 1979 236 4 E328 E334 434194 39. Mak RH Effect of metabolic acidosis on insulin action and secretion in uremia Kidney Int 1998 54 2 603 607 10.1046/j.1523-1755.1998.00023.x 9690228 40. Adrogue HJ Madias NE Sodium and potassium in the pathogenesis of hypertension N Engl J Med 2007 356 19 1966 1978 10.1056/NEJMra064486 17494929 41. Coruzzi P Brambilla L Brambilla V Gualerzi M Rossi M Parati G Di Rienzo M Tadonio J Novarini A Potassium depletion and salt sensitivity in essential hypertension J Clin Endocrinol Metab 2001 86 6 2857 2862 10.1210/jcem.86.6.7601 11397900 42. Tobian L Lange J Ulm K Wold L Iwai J Potassium reduces cerebral hemorrhage and death rate in hypertensive rats, even when blood pressure is not lowered Hypertension 1985 7 3 Pt 2 I110 I114 10.1161/01.HYP.7.3_Pt_2.I110 3997230 43. Liu DT Wang MX Kincaid-Smith P Whitworth JA The effects of dietary potassium on vascular and glomerular lesions in hypertensive rats Clin Exp Hypertens 1994 16 4 391 414 10.3109/10641969409067953 7920452 44. Tanner GA Renal regulation of acid-base balance: ammonia excretion Physiologist 1984 27 2 95 97 6728923 45. van Haare J Kooi ME Vink H Post MJ van Teeffelen JW Slenter J Munts C Cobelens H Strijkers GJ Koehn D Early impairment of coronary microvascular perfusion capacity in rats on a high fat diet Cardiovasc Diabetol 2015 14 150 10.1186/s12933-015-0312-2 26576929 46. Axelsen LN Calloe K Braunstein TH Riemann M Hofgaard JP Liang B Jensen CF Olsen KB Bartels ED Baandrup U Diet-induced pre-diabetes slows cardiac conductance and promotes arrhythmogenesis Cardiovasc Diabetol 2015 14 87 10.1186/s12933-015-0246-8 26169175 47. Okubo H Sasaki S Murakami K Kim MK Takahashi Y Hosoi Y Itabashi M Freshmen in Dietetic Courses Study IIg Three major dietary patterns are all independently related to the risk of obesity among 3760 Japanese women aged 18–20 years Int J Obes 2008 32 3 541 549 10.1038/sj.ijo.0803737 48. Sanada K Miyachi M Tanimoto M Yamamoto K Murakami H Okumura S Gando Y Suzuki K Tabata I Higuchi M A cross-sectional study of sarcopenia in Japanese men and women: reference values and association with cardiovascular risk factors Eur J Appl Physiol 2010 110 1 57 65 10.1007/s00421-010-1473-z 20390291 49. Cesari M Kritchevsky SB Baumgartner RN Atkinson HH Penninx BW Lenchik L Palla SL Ambrosius WT Tracy RP Pahor M Sarcopenia, obesity, and inflammation–results from the Trial of Angiotensin Converting Enzyme Inhibition and Novel Cardiovascular Risk Factors study Am J Clin Nutr 2005 82 2 428 434 16087989

PMC005xxxxxx/PMC5002187.txt

==== Front J Occup Med ToxicolJ Occup Med ToxicolJournal of Occupational Medicine and Toxicology (London, England)1745-6673BioMed Central London 13110.1186/s12995-016-0131-8ResearchIs outdoor work associated with elevated rates of cerebrovascular disease mortality? A cohort study based on iron-ore mining http://orcid.org/0000-0002-6080-5973Björ Ove +46 90 785 6878ove.bjor@vll.se 1Jonsson Håkan hakan.s.jonsson@vll.se 1Damber Lena lena.damber@onkologi.umu.se 1Burström Lage lage.burstrom@umu.se 2Nilsson Tohr tohr.nilsson@envmed.umu.se 21 Department of Radiation Sciences (Oncology), Umeå University, SE-901 85 Umeå, Sweden 2 Department of Public Health & Clinical Medicine, Occupational and Environmental Medicine, Umeå University, SE-901 85 Umeå, Sweden 26 8 2016 26 8 2016 2016 11 1 4018 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 A cohort study that examined iron ore mining found negative associations between cumulative working time employed underground and several outcomes, including mortality of cerebrovascular diseases. In this cohort study, and using the same group of miners, we examined whether work in an outdoor environment could explain elevated cerebrovascular disease rates. Methods This study was based on a Swedish iron ore mining cohort consisting of 13,000 workers. Poisson regression models were used to generate smoothed estimates of standardized mortality ratios and adjusted rate ratios, both models by cumulative exposure time in outdoor work. Results The adjusted rate ratio between employment classified as outdoor work ≥25 years and outdoor work 0–4 years was 1.62 (95 % CI 1.07–2.42). The subgroup underground work ≥15 years deviated most in occurrence of cerebrovascular disease mortality compared with the external reference population: SMR (0.70 (95 % CI 0.56–0.85)). Conclusions Employment in outdoor environments was associated with elevated rates of cerebrovascular disease mortality. In contrast, work in tempered underground employment was associated with a protecting effect. Keywords Poisson regressionOccupationStandardized mortality ratioMortalityCohort studyMiningEuropean Union (Kolarctic ENPI CBC - MineHealth)02/2011/043/KO303issue-copyright-statement© The Author(s) 2016 ==== Body Background A cohort study of Swedish iron-ore miners reported that duration of employment time among blue collar workers, especially in underground mining, was associated with lower risks of mortality and incidental cancer [1]. Compared with other workers, mortality from cerebrovascular diseases was lower among long term underground mining blue-collar workers (Rate Ratio: 0.66 (95 % ci 0.49–0.89)). Three explanations for these inverse associations were suggested: a protecting effect from working in a tempered underground environment, a protecting effect from heavy physical workloads, or, a healthy worker survivor effect (HWSE). The lower risks of cerebrovascular disease associated with duration of underground mining work could be explained by a protecting effect from the more tempered underground climate compared with the outdoor climate. Acute effects of cerebrovascular disease mortality have been explained by cold outdoor temperatures [2, 3]. Workers that remain in longer term employment underground could also be individuals that have better health because of lifestyle factors, beneficial effects from long-term employment or genetic factors. This selection phenomenon of healthier workers remaining in the workforce with a resulting inverse association between cumulative exposure and outcome is called the HWSE [4, 5]. A recent paper from the Kiruna- and Malmberget mines in Sweden showed that adjusting for the HWSE by g-estimation of accelerated failure-time modeling resulted in elevated risks of mortality from respirable dust [6]. Historically, high physical workloads have been associated with mining activities [7]. Several papers have suggested that physically active occupations and lifestyles protect from stroke [8, 9], and colon and rectal cancer [10, 11]. The aim of this study was, to examine if outdoor mining work was associated with elevated rates of cerebrovascular disease mortality. This study will also consider underground work, due to the results from the previous study on this cohort that found a negative association with cerebrovascular disease mortality [1]. Methods The cohort The Kiruna and Malmberget iron-ore mines are located in the northernmost part of Sweden. They have been in operation since the early 19th century. The mines have similar employment policies and safety routines, and use similar types of equipment. Malmberget has been an underground mine since it began operation. Kiruna was an open-pit mine until the 1950s, when the mining activities successively went underground. Employees were included in the study cohort if they had been employed for at least one year between 1923–1998 at the Kiruna mine, and between 1923–1996 at the Malmberget mine. The cohort has previously been described in detail [12, 13]. Cohort mortality was followed from 1952–2006. Data was collected from the Swedish national cause of death register. The mortality endpoint was cerebrovascular diseases (ICD10: I60–I69), subgroups’ cerebral infarction (ICD10: I63–I64), and cerebral haemorrhage (ICD10: I61). Information on cerebral haemorrhage was only available from 1965. A unique Swedish personal identification number was used to link dates of events and ICD-codes for mortality to each cohort member. Subjects that were not found in the cause of death register or the national population register (N = 584) were excluded from the cohort because they had died before 1952 or had emigrated from Sweden. Time at emigration from Sweden after 1952 was censured in the statistical analysis (N = 54). Workers for whom information on occupational class (blue-collar worker, white-collar worker, miscellaneous), or work location (inside, outdoor, mining) was not available were excluded (N = 130). Work periods were categorized as working underground, working 50 % of the time underground, or other work. We used this work period variable to represent underground work, and the classification procedure has been previously described [12, 13]. Of the remaining workers, 493 of them were employed only in white collar work, and were excluded from the study. The final cohort consisted of 13,000 male workers. Classification of work periods into outdoor work Classification of work location was based on job title, workplace, and year. This information was collected from written and computerized company records. Work location categories were working outdoors, working inside, or working in the mines. Due to the successive change from open to underground mining, all pre-1954 Kiruna mine work periods that were not registered as working inside were classified as working outdoors. This method of classifying work periods into outdoor work implied a possible overlap between the previously defined underground work and this new variable, outdoor work. From 1954 to 1958, blue-collar mining work periods were increasingly classified as working underground with the following weights: 0.1, 0.3, 0.5, 0.7, and 0.9. Classification of work location was performed by three researchers in occupational medicine that were knowledgeable about mine conditions, guided by descriptions of mine history [7]. Statistical methods Two log-linear Poisson regression models were calculated. One model was based on cohort data only. The other model also incorporated external rates from the reference population of northern Sweden (male population 1980: 460445). The model limited to comparisons within the cohort was: logEdkl=lognkl+α+xkβ+zlβ* where E(dkl): predicted number of outcomes in the cohort per nuisance variable, k, and exposure level variable, l nkl: person years from the cohort per categories, k, and l α: constant xkβ: is a vector of nuisance variables attained age, calendar year, year of first employment, also having been employed as a white collar worker, mine, and cumulative employment time underground and β represents the corresponding parameter estimates. Z1β*: is a vector representing cumulative time working outdoors and β is the corresponding parameter estimates. The main predictor (zlβ* ) estimated the ratio between categories 5–14, 15–24, and ≥25 years of cumulative employment time outdoor and the reference category 0–4 years. All variables representing cumulative employment time, calendar year, attained age, cumulative employment time underground, and white collar worker (yes, from the first day of employment as a white collar worker and forward, or, no) were analyzed as time-dependent variables. Test of trend was performed by entering a single continuous covariate in this Poisson model, with values representing the mean duration of employment for each category. To evaluate if acute effects of outdoor work could explain elevated rates (deaths occurring close to being employed), a variable coded as 1 (employed) or 0 (not employed) for each time period was included in the model to be able to adjust for possible differences due to deaths that occurred while employed. All deaths within 20 days from the last date of employment were counted as death while employed. The model that incorporated external rates was: logEdijl=logexpij+α+fzl where E(dijl): predicted number of outcomes in the cohort per year, i, age group, j and category of outdoor work expij: expected number of outcomes in the cohort per year, i, age group, j. The expected values were derived from mortality and cancer incidence risks based on the reference population of northern Sweden α: constant f(zl): a cubic spline function of cumulative time in outdoor work This model generated smoothed estimates [14] of standardized mortality ratios (SMR) by cumulative employment time categorized into outdoor work. The cubic spline function (f (zl)) was based on 3 knots: at 0 years of exposure, at the median number of years of exposure, and at 35 years of exposure. The function was linear from year 35. To complement the adjusted Poisson model described earlier, and to give priority to precision of the estimated SMR, we chose not to include nuisance parameters in this model. Because of the cohort inclusion criterion of a one year minimum employment time, first year of employment was not included in any statistical analysis as time at risk. The potential for overdispersion was accounted for during Poisson regression estimation of all confidence intervals. By calculating SMR based on cerebrovascular disease mortality as the outcome, different subgroups of the cohort were compared with the population of northern Sweden. SMR was calculated from cell-specific external rates represented by calendar year in one year classes and age in five year classes (maximum age class, ≥85 years). Because of the wide upper category of ≥85 years in the reference rates, the analyses on subgroups were also performed limited to <85 years of age. 95 % confidence intervals were calculated by assuming that the number of deaths in the cohort followed a Poisson distribution. Data were analyzed with the statistical software package R (version 3.1.1; R Development Core Team, R foundation for Statistical Computing, Vienna, Austria). Results The study cohort consisted of 13,000 workers representing 443,930 person years. The average age at death of all causes was 67.3 (standard deviation (SD), 14.3) years. The distribution for key variables based on mortality of cerebrovascular diseases as outcome is presented in Table 1.Table 1 Mean (Standard Deviation), age at death and distribution of number of deaths of cerebrovascular diseases Cerebrovascular diseases Mean (SD) age at outcome 72.0 (12.4) Total number of outcomes 403 Number of outcomes per birth year −1899 49 1900–1909 106 1910–1919 112 1920–1929 85 ≥1930 51 Number of outcomes per year of first employment ≤1929 112 1930–1939 60 1940–1949 62 1950–1959 137 ≥1960 32 Number of outcomes per cumulative employment time in years 0–4 51 5–14 75 15–29 111 ≥30 166 Number of outcomes per mine Kiruna 246 Malmberget 157 The distribution of time classified as outdoor and underground work Nineteen percent of the person years, as employed, were classified as “outdoor workers”, compared with 46 % “working inside”, 31 % “mining”, and 3 % “missing”. Working outdoors was most common before 1960 because the Kiruna mine was an open pit mine prior to 1960 (Fig. 1). Of all person time classified as outdoor, 26 % was simultaneously classified as underground work. Of all person years representing employment, 43 % was neither classified as outdoor or underground work. The classification procedures were performed by different experts, and the transition between outdoor and underground work meant that the specific location could be interpreted as outdoor and underground, depending on the criteria for location.Fig. 1 Person year distribution of outdoor and underground work. Gray area represents the number of employed workers Outdoor work and the association with cerebrovascular disease mortality The adjusted rate ratio estimates for mortality from cerebrovascular diseases increased with cumulative outdoor employment time (Fig. 2 and Table 2). An expanded analysis that included an interaction term between year of first employment and cumulative outdoor employment indicated that rates for workers employed for the first time in later years (1930–1949) more clearly was associated with cumulative outdoor employment compared with earlier years (<1930). This interaction effect between year of first employment and cumulative employment time outdoors did not significantly contribute to the model (p = 0.139, based on a likelihood ratio test for inclusion of the interaction effect). However, the rate ratio estimate for outdoor work ≥25 years compared with 0–4 years was 3.45 (95 % CI 1.55–7.65; 11 deaths: ≥25 years) for workers employed first time 1930–1949. The corresponding estimate was 1.16 (95 % CI 0.73–1.86; 37 deaths: ≥25 years) for workers that were employed first time before 1930. Because of the limited number of outcomes, it was not possible to calculate interaction term-estimates for later years of first employment. We also calculated adjusted rate ratio estimates from Poisson regression by delimiting outcomes to subgroups’ cerebral haemorrhages and cerebral infarctions. Due to the limited number of outcomes, long-term outdoor work was defined as outdoor work ≥15 years. For mortality of cerebral haemorrhages, the adjusted rate ratio for outdoor work 5–14 years (12 deaths) compared to outdoor work 0–4 years (46 deaths) was 1.21 (95 % CI 0.60–2.28) and the corresponding ratio for outdoor work ≥15 years (18 deaths) was 2.52 (95 % CI 1.23–5.04). For mortality of cerebral infarctions, the adjusted rate ratio for outdoor work 5–14 years (23 deaths) compared to outdoor work 0–4 years (71 deaths) was 1.38 (95 % CI 0.83–2.22) and the corresponding ratio for outdoor work ≥15 years (21 deaths) was 1.60 (95 % CI 0.88–2.81).Fig. 2 Smoothed estimates of standardized mortality ratios (SMRs) with 95 % confidence intervals (lines) and adjusted point estimates of rate ratios in relation to cumulative time employed outdoors The categorical estimates in the figure are located at the mean employment time (1.0, 10.7, 21.0, 33.1 years). The location of the mortality outcomes in relation to cumulative time employed outdoors is marked on the top of the figure Table 2 Adjusted rate ratio estimates derived from Poisson regression models, with 95 % confidence intervals (CI) 0–4 years 5–14 years 15–24 years ≥25 years na Rate ratio na Rate ratio 95 % CI na Rate ratio 95 % CI na Rate ratio 95 % CI Test of trend, p-value Cerebrovascular diseases 228 1 77 1.22 0.90–1.62 50 1.42 0.97–2.04 48 1.62 1.07–2.42 0.010 Cerebrovascular diseases limited to age ≥ 55 195 1 73 1.26 0.91–1.71 49 1.45 0.97–2.13 47 1.62 1.04–2.47 0.013 Cerebrovascular diseases adjusted for acute effectsb 228 1 77 1.22 0.91–1.62 50 1.42 0.98–2.04 48 1.63 1.08–2.41 0.008 aNumber of deaths. bEstimated from Poisson model, included a variable coded 1 if employed and 0 otherwise to control for death occurring while employed The results for SMR based on splines did not suggest as strong association between outdoor work and cerebrovascular disease mortality as suggested by the adjusted rate ratio estimates (Fig. 2). For short term outdoor work, the rates were lower than the corresponding expected rates derived from the reference population of northern Sweden which contributed to that the SMR, by design, was closer to unit for long term outdoor work compared to the category estimates. Figure 3 shows the adjusted rate ratio estimates for cumulative outdoor work together with the corresponding estimates for underground work. The figure clearly illustrates the diverging trends when comparing the estimates of outside and underground work.Fig. 3 Adjusted rate ratio estimate for cumulative outdoor and underground employment time based on Poisson regression. The estimates are derived from the same model based on the same category levels and the points are located at the mean employment time per category Delimiting the adjusted Poisson models to include only age ≥ 55 years did not change the interpretations of the results (Table 2). Adjusting for current employment status in the Poisson model had only small effects on the results (Table 2). SMR based on subgroups from the cohort The result for the estimated SMR of mortality from cerebrovascular disease revealed that the rates associated with underground work differed most from the expected rates (Fig. 4). Underground work ≥15 years represented 30 % lower rates: 0.70 (95 % CI 0.56–0.85). Corresponding SMR for outdoor work ≥15 years was 1.16 (95 % CI 0.94–1.40). Delimiting the calculations to include only ages < 85 years yielded similar conclusions, but with slightly lower levels of the estimates overall (data not shown).Fig. 4 Standardized mortality ratios (SMRs) for cerebrovascular disease based on northern Sweden as reference population. The numbers within the graph are the estimated SMRs Discussion The results of this study suggested that there was an association between employments represented by temperature varying outdoor work environments and elevated rates of cerebrovascular disease mortality. The rate of cerebrovascular disease mortality representing cumulative outdoor employment time 25 years or more was 62 % higher (rate ratio: 1.62 (95 % CI 1.07–2.42)) compared with outdoor employment 0 to 4 years (Fig. 2 and Table 2). Moreover, diverging trends between outdoor work and underground work became evident when both variables were included in the same adjusted Poisson model (Fig. 3). There was an indication that the elevated rates for outdoor work were most pronounced for long term outdoor workers employed first time 1930–1949, compared with corresponding workers employed first time before 1930. The adjusted categorical rate ratios for long term outdoor work were noticeably more elevated than the corresponding SMRs based on the reference population (Fig. 2). The rates for short term outdoor work were clearly lower than the reference population of northern Sweden, whereas, the lower limit of the point estimated rate ratio is by design, equal to one. For outdoor work ≥15 years, the rate of cerebral haemorrhages (i.e., the delimiting outcomes for cerebral haemorrhages and cerebral infarctions) was statistically significantly higher than outdoor work 0–4 years – 2.52 (95 % CI 1.23–5.04). The corresponding rate of cerebral infarctions was also elevated, but not statistically significant – 1.60 (95 % CI 0.88–2.81). The results of the cohort subgroup comparisons of the observed rates and expected rates of cerebrovascular disease mortality derived from the reference population showed that cumulative employment time underground most clearly deviated from the expected rates (Fig. 4). However, the causal link explaining these results could have been connected to both outdoor and underground work. Although adjustments were made for year of employment, attained age, and calendar year in the analysis, outdoor and underground work represented different time-windows of exposure (Fig. 1) as mining methods changed and safety precautions improved. However, the emphasis of the analysis was on earlier years of employment in both groups as the largest proportion of deaths as the result of cerebrovascular diseases occurred among workers who had been employed before 1960 (>90 % in both groups). The relatively smaller deviation in SMR from the reference population for outdoor workers compared with underground workers (Fig. 4) may have been diluted by the 26 % overlap with underground work in the classification procedure. The overlapping work-periods may be explained by the crude classification of outdoor mining, which was defined as all mining in Kiruna before the late 1950s. Moreover, the classification was affected by the fact that some of the mining performed in the earlier periods in Kiruna was drift mining that could be described as a sub-underground mining method. Cold winter-time temperatures and large seasonal variations in sun light are characteristic of outdoor work in the northernmost part of Sweden. Exposure to cold temperature has been shown to increase blood pressure [15], with several potential mechanisms leading to cerebrovascular disease [16]. However, cold temperatures that result in elevated mortality from cerebrovascular disease act mainly through acute or subacute effects [2]. The monthly distribution of the 304 deaths during one season showed that the three months with the highest frequency of deaths were December, January, and February (31 % of all deaths). Although, adjusting for current status (employed or not employed) did not weaken the association between outdoor work and mortality (Table 2), which implies that outdoor-work employments could have a preserving effect on mortality. We believe that temperature caused hypertension is the risk factor from work that likely best explains these results even if this explanation does require that work has a preserving effect on hypertension. Smoking increases the risk of cerebrovascular diseases and the associations are strongest in younger ages [9]. The different working conditions and regulations of outdoor and underground work could contribute to different smoking habits by making smoking more or less accessible. We did not have data to adjust for smoking in the analysis. However, a previous study on this cohort showed a SIR of 1.74 (95 % CI 1.52–1.99) for lung cancer among underground workers [1]. The corresponding SIR for outdoor workers was 1.86 (95 % CI 1.53–2.23), a comparable SIR to underground workers. Smoking is a considerably stronger risk factor for lung cancer than it is for cerebrovascular diseases, which indirectly supports that smoking is not a strong confounder in this study. Also, a previous study on workers in the Malmberget mines, which used data on smoking for 2310 workers, found that 75 % of the smokers (1972–1992) had worked underground compared to 71 % of the non-smokers [13], indicating that a possible prohibition on smoking on underground workers did not result in smoking being less common among underground workers. Other known risk factors not possible to adjust for in this study include diabetes, atrial fibrillation, dyslipidemia, asymptomatic carotid stenosis, heredity, and an unhealthy lifestyle factors such as alcohol consumption [9]; these risk factors are harder to causally link with working outdoor or underground (but could still be possible confounders). The hypothesis that excess sunlight could be a cause of mortality from cerebrovascular disease has not been not supported by the literature [17]. A weakness of using cause-specific mortality as a measure has been that the causes of death register contains misclassification errors. In a validation study based on the Swedish cause of death register for 1094 deaths registered in 1995, 0.68 % (95 % CI 0.60–0.77) of all deaths from cerebrovascular underlying causes was correctly classified [18]. The healthy worker survivor effect A healthy worker survivor effect (HWSE) [4, 5] could contribute to the reduced mortality rates associated with underground work found in the previous study of this cohort [1]. For cerebrovascular disease, we observed simultaneously elevated rates with outdoor work and reduced rates with underground work, as was found in the previous study. If the HWSE explains this inverse association with underground mining work, it would mirror a shift from underground mining to other categories of employment in the mines that is correlated with increased cerebrovascular disease mortality. The pattern of lower risks for several diseases with longer employment as a blue collar worker, especially underground, that was reported in the previous paper [1], elucidates the need to account for the HWSE when performing studies that focus on cumulative exposures in occupational cohorts. Validity of this study This study was based on iron ore miners from the Kiruna and Malmberget mines and consisted of a relatively large cohort with 443,930 person years and 13,000 workers. The cohort data are of very high quality because of the accuracy of the records of the mining company and because the exposure assessments were performed by occupational health researchers with knowledge about the mines. A study based on workers located in the far reaches of Sweden could be expected to represent a relatively homogeneous population in terms of socio-economic and lifestyle factors, potentially reducing bias from residual confounding. We only included workers who had a valid Swedish unique personal identification number. That is, only Swedish citizens are in the cohort as immigration to Sweden from outside Europe was limited before the 1980s. We did not have any information on the workers’ ethnicity, but a study by Sjölander [19] found that the risk of cardiovascular diseases among the Swedish Sami was slightly lower than the majority population, a finding explained by the active reindeer herding lifestyle. A weakness could have been misclassification of both exposures and for outcomes. As in most epidemiological studies, unmeasured confounding could have influenced the results. Information on hypertension, physical activity, and lifestyle factors other than work would have increased study validity. Conclusions The results of this cohort study revealed that rates of cerebrovascular disease mortality increased with time of employment categorized as outdoor work. This association was also in contrast with the previous observation that work performed in a tempered underground environment has a protecting effect on cerebrovascular disease mortality. However, compared with the northern Sweden reference population, rates for long term outdoor employment were only marginally elevated. Abbreviations HWSEHealthy worker survivor effect ICDInternational classification of disease SMRStandardized mortality ratio Acknowledgements Not applicable. Funding This document was been produced with financial assistance from the European Union (Kolarctic ENPI CBC Project 02/2011/043/KO303- MineHealth). The contents of this document are the sole responsibility of Umeå University and under no circumstances can be regarded as reflecting the position of the European Union. The funding body has had no role in the collection, analysis, interpretation of data or in writing the manuscript. Availability of data and materials This study was approved by the Regional Ethical Review Board of Umeå, provided there was no risk of identifying any individual in the publication. However, as our cohort data contains detailed information on each workers job-history and possible occurrence of cancer, identification of an individual is a possibility if raw data would be freely available. Raw data from this cohort is available in our repository by writing to the corresponding author, but researchers who would like to use the data need to apply for an ethical permission. Authors’ contributions All authors (OB, HJ, LD, LB and TN) have contributed according to the following criteria: Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; Drafting the work or revising it critically for important intellectual content; 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 This work was approved by the Regional Ethical Review Board of Umeå: §59/03, dnr 03–040 (2003-02-11). The approval contains a stipulation that the results can only be presented if there is no risk of identifying any individual involved in the study. All workers in the company were informed of this study by their superiors, their union representatives and newsletters, and the work environment committee for the mines. ==== Refs References 1. Bjor O Jonsson H Damber L Wahlstrom J Nilsson T Reduced mortality rates in a cohort of long-term underground iron-ore miners Am J Ind Med 2013 56 5 531 40 10.1002/ajim.22168 23450695 2. Analitis A Katsouyanni K Biggeri A Baccini M Forsberg B Bisanti L Effects of cold weather on mortality: results from 15 European cities within the PHEWE project Am J Epidemiol 2008 168 12 1397 408 10.1093/aje/kwn266 18952849 3. Rocklov J Ebi K Forsberg B Mortality related to temperature and persistent extreme temperatures: a study of cause-specific and age-stratified mortality Occup Environ Med 2011 68 7 531 6 10.1136/oem.2010.058818 20962034 4. Arrighi HM Hertz-Picciotto I The evolving concept of the healthy worker survivor effect Epidemiology 1994 5 2 189 96 10.1097/00001648-199403000-00009 8172994 5. Kolstad HA Olsen J Why do short term workers have high mortality? Am J Epidemiol 1999 149 4 347 52 10.1093/oxfordjournals.aje.a009819 10025477 6. Björ O Damber L Jonsson H Nilsson T A comparison between standard methods and structural nested modeling when bias from a healthy worker survivor effect is suspected: an iron-ore mining cohort study Occup Environ Med 2015 72 7 536 42 10.1136/oemed-2014-102251 25713154 7. Eriksson U Gruva och arbete: Kiirunavaara 1890–1990 1991 Uppala Uppsala University 8. Wendel-Vos GC Schuit AJ Feskens EJ Boshuizen HC Verschuren WM Saris WH Physical activity and stroke. A meta-analysis of observational data Int J Epidemiol 2004 33 4 787 98 10.1093/ije/dyh168 15166195 9. Goldstein LB Bushnell CD Adams RJ Appel LJ Braun LT Chaturvedi S Guidelines for the primary prevention of stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association Stroke 2011 42 2 517 84 10.1161/STR.0b013e3181fcb238 21127304 10. Friedenreich CM Neilson HK Lynch BM State of the epidemiological evidence on physical activity and cancer prevention Eur J Cancer 2010 46 14 2593 604 10.1016/j.ejca.2010.07.028 20843488 11. Moradi T Gridley G Bjork J Dosemeci M Ji BT Berkel HJ Occupational physical activity and risk for cancer of the colon and rectum in Sweden among men and women by anatomic subsite Eur J Cancer Prev 2008 17 3 201 8 10.1097/CEJ.0b013e3282b6fd78 18414190 12. Bergdahl IA Jonsson H Eriksson K Damber L Jarvholm B Lung cancer and exposure to quartz and diesel exhaust in Swedish iron ore miners with concurrent exposure to radon Occup Environ Med 2010 67 8 513 8 10.1136/oem.2009.047456 20519746 13. Jonsson H Bergdahl IA Akerblom G Eriksson K Andersson K Kagstrom L Lung cancer risk and radon exposure in a cohort of iron ore miners in Malmberget, Sweden Occup Environ Med 2010 67 8 519 25 10.1136/oem.2009.047449 20647379 14. Wood SN Generalized Additive Models. An Introduction with R 2006 1 Boca Raton Chapman & Hall/CRC 15. Halonen JI Zanobetti A Sparrow D Vokonas PS Schwartz J Relationship between outdoor temperature and blood pressure Occup Environ Med 2011 68 4 296 301 10.1136/oem.2010.056507 20864465 16. Brickman AM Reitz C Luchsinger JA Manly JJ Schupf N Muraskin J Long-term blood pressure fluctuation and cerebrovascular disease in an elderly cohort Arch Neurol 2010 67 5 564 9 10.1001/archneurol.2010.70 20457955 17. Kent ST McClure LA Judd SE Howard VJ Crosson WL Al-Hamdan MZ Short- and long-term sunlight radiation and stroke incidence Ann Neurol 2013 73 1 32 7 10.1002/ana.23737 23225379 18. Johansson LA Bjorkenstam C Westerling R Unexplained differences between hospital and mortality data indicated mistakes in death certification: an investigation of 1,094 deaths in Sweden during 1995 J Clin Epidemiol 2009 62 11 1202 9 10.1016/j.jclinepi.2009.01.010 19364635 19. Sjolander P. What is known about the health and living conditions of the indigenous people of northern Scandinavia, the Sami? Glob Health Action. 2011;4. doi:10.3402/gha.v4i0.8457.

PMC005xxxxxx/PMC5002188.txt

==== Front BMC Musculoskelet DisordBMC Musculoskelet DisordBMC Musculoskeletal Disorders1471-2474BioMed Central London 119810.1186/s12891-016-1198-1CorrespondenceThe relationship between Modic changes and intervertebral disc degeneration Määttä Juhani H. Juhani.maatta@fimnet.fi 1MacGregor Alex A.Macgregor@uea.ac.uk 2Karppinen Jaro Jaro.Karppinen@ttl.fi 134Williams Frances M. K. Frances.williams@kcl.ac.uk 21 Medical Research Center Oulu, Oulu University Hospital and University of Oulu, Oulu, Finland 2 Department of Twin Research and Genetic Epidemiology, King’s College London, St Thomas’ Hospital, London, SE1 7EH UK 3 Finnish Institute of Occupational Health, Oulu, Finland 4 Faculty of Medicine, Center for Life Course Health Research, University of Oulu, Oulu, Finland 26 8 2016 26 8 2016 2016 17 1 37122 4 2016 2 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 Recent reported results have added to the weight of evidence supporting association between disc degeneration and Modic changes. Endplate or Modic changes are also associated with increased body mass index. The most recent study from Teichtahl et al. titled ‘Modic changes in the lumbar spine and their association with body composition, fat distribution and intervertebral disc height – a 3.0 T-MRI study’ showed associations of Modic changes with quantitatively measured reduced disc height and fat mass index. However, there were some facts, which we would like to address in this Correspondence to their article. Discussion The different components of intervertebral disc degeneration such as loss of disc height and disc signal intensity have already been shown associated with endplate changes – but not disc height if it is assessed using newer more precise methods of quantitation of disc height. A possible protective effect of different adiposity distribution in the body to Modic change development would be of interest if observed in a longitudinal study in the future. Summary Modic changes have been associated with different components of intervertebral disc degeneration such as loss of disc height and disc signal intensity previously. The influence of body fat distribution on endplate changes would be interesting to study longitudinally. Keywords Body mass indexDisc degenerationModic changeissue-copyright-statement© The Author(s) 2016 ==== Body Background We were interested to read the recently published study on Modic changes by Teichtahl and Urquhart et al. [1] which used 3 T magnetic resonance imaging of the spine in a small population sample. There is a pressing need to understand better the mechanisms behind the common and costly social problem of low back pain. We wonder what the rationale for studying fat distribution was when the authors reference a paper showing no evidence of an effect of body mass index (BMI) [2]. In fact, there are several studies published already which have examined endplate changes (i.e. Modic changes). They provide evidence for increasing BMI being associated with endplate changes. Body mass index and waist circumference were both associated with Modic type 2 change among middle-aged male workers in Finland [3]; and among Spanish chronic low back pain patients any Modic change was associated with increasing BMI [4]. Finally, our recent study of a predominantly female twin sample (TwinsUK) found association with increasing BMI and endplate changes [5]. More recently, obesity was proposed to affect to Modic change development through increased spinal forces, i.e. hyperloading, but also may exert influence on adipogenesis, hematopoiesis and osteogenesis [6]. Our work using the TwinsUK registry [5], as well as that of Kerttula et al. [7], examined endplate changes and features of intervertebral disc and found an association with both disc height loss and change in disc signal intensity. While our work lacks T1-weighted images and therefore isn’t strictly as Modic described, the sample is more than ten fold larger than the reported Australian study and likely presents robust findings. Even though disc height loss has been associated with Modic changes, it is true that previous studies have evaluated disc height by semi-quantitative scale as applied by trained personnel using a reference atlas, and not quantitatively. Quantitative measurement is an advantage in the study of Teichtahl et al. [1], but it should be noted that the association of disc height loss and Modic changes have been reported previously. It would be of interest to determine whether different distributions of adiposity really have opposite effects or whether this reflects chance findings in a cross-sectional snapshot in endplate progression. In TwinsUK data we observed an association between disc height and disc bulge at baseline and an incident endplate change during follow-up of over a decade. Finally, most importantly of all, we showed both intervertebral disc degeneration and endplate changes to be independent predictors of episodes of severe and disabling low back pain. Acknowledgements Not applicable. Funding Not applicable. Availability of data and materials Not applicable. Authors’ contributions JHM, AM, JK and FMKW contributed in drafting the manuscript and revised it critically for important intellectual content and gave final approval of the version to be published. Authors’ information Not applicable. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. Response to “The relationship between Modic changes and intervertebral disc degeneration” Teichtahl Andrew J. 56Urquhart Donna M. 5Wang Yuanyuan 5Wluka Anita E. 5O’Sullivan Richard 78Jones Graeme 9Cicuttini Flavia M. flavia.cicuttini@monas.edu 55 Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Alfred Hospital, Melbourne, VIC 3004 Australia 6 Baker IDI Heart and Diabetes Institute, Commercial Road, Melbourne, VIC 3004 Australia 7 Healthcare Imaging Services, Epworth Hospital, Richmond, Melbourne, VIC 3121 Australia 8 Department of Medicine, Central Clinical School, Monash University, Melbourne, VIC 3004 Australia 9 Menzies Research Institute, Private bag 23, Hobart, TAS 7000 Australia We thank Määttä et al. for their interest in our work [1]. We agree that a strength of our study is the quantitative measure of intervertebral disc height. This complements the findings from the TwinsUK registry whereby endplate changes (i.e. Modic changes) were associated with semi-quantitative features of intervertebral disc degeneration [5]. We offer intervertebral disc height as a validated, quantitative, easily measured and sensitive alternative to qualitative or semi-quantitative measures of disc degeneration [8]. A major aim of our study was to determine whether body composition was a risk factor for Modic change, with the hypothesis being that adiposity would be a risk factor for such lesions. Previous attempts to examine the association between body mass index (BMI) and Modic change, whether demonstrating significant [3, 4] or non-significant [2] findings, cannot specifically comment on the contribution of adiposity: BMI cannot discriminate adipose or non-adipose mass, nor can it measure the distribution of fat. This is of particular interest since Modic type 2 change histologically represents fat replacement of the vertebral endplate [9]. Additionally, we agree that it would be of interest to further corroborate the differential effect of android and gynoid fat distribution. Nevertheless, our data is compelling: we have shown statistically significant associations (p = 0.01) in different biological directions for the exposure of android and gynoid fat. It would seem unlikely that these disparate associations occurred by chance alone, particularly given our modest sample size which would only serve to reduce our chance of showing significant results. We welcome further research efforts examining these associations, and agree that longitudinal studies would be particularly helpful in determining the predictive nature of not only structural abnormalities, but also symptoms. Abbreviation BMI, body mass index ==== Refs References 1. Teichtahl AJ Urquhart DM Wang Y Wluka AE O’Sullivan R Jones G Modic changes in the lumbar spine and their association with body composition, fat distribution and intervertebral disc height - a 3.0 T-MRI study BMC Musculoskelet Disord 2016 17 1 92 10.1186/s12891-016-0934-x 26891686 2. Karchevsky M Schweitzer ME Carrino JA Zoga A Montgomery D Parker L Reactive endplate marrow changes: a systematic morphologic and epidemiologic evaluation Skeletal Radiol 2005 34 3 125 9 10.1007/s00256-004-0886-3 15647940 3. Kuisma M Karppinen J Haapea M Niinimäki J Ojala R Heliövaara M Are the determinants of vertebral endplate changes and severe disc degeneration in the lumbar spine the same? A magnetic resonance imaging study in middle-aged male workers BMC Musculoskelet Disord 2008 9 51 10.1186/1471-2474-9-51 18416819 4. Arana E Kovacs FM Royuela A Estremera A Asenjo B Sarasibar H Modic changes and associated features in Southern European chronic low back pain patients Spine J 2011 11 5 402 11 10.1016/j.spinee.2011.03.019 21558034 5. Määttä JH Wadge S MacGregor A Karppinen J Williams FM ISSLS Prize Winner: Vertebral Endplate (Modic) Change is an Independent Risk Factor for Episodes of Severe and Disabling Low Back Pain Spine (Phila Pa 1976) 2015 40 15 1187 93 10.1097/BRS.0000000000000937 25893353 6. Dudli S, Fields AJ, Samartzis D, Karppinen J, Lotz JC. Pathobiology of Modic changes. Eur Spine J. 2016 [Epub ahead of print]. 7. Kerttula L Luoma K Vehmas T Grönblad M Kääpä E Modic type I change may predict rapid progressive, deforming disc degeneration: a prospective 1-year follow-up study Eur Spine J 2012 21 6 1135 42 10.1007/s00586-012-2147-9 22249308 8. Teichtahl AJ Urquhart DM Wang Y Wluka AE Heritier S Cicuttini FM A Dose–response relationship between severity of disc degeneration and intervertebral disc height in the lumbosacral spine Arthritis Res Ther 2015 17 297 10.1186/s13075-015-0820-1 26498120 9. Modic MT Steinberg PM Ross JS Masaryk TJ Carter JR Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging Radiology 1988 166 1 Pt 1 193 9 10.1148/radiology.166.1.3336678 3336678

PMC005xxxxxx/PMC5002189.txt

==== Front Lipids Health DisLipids Health DisLipids in Health and Disease1476-511XBioMed Central London 31110.1186/s12944-016-0311-7ResearchReduced circulating adiponectin levels are associated with the metabolic syndrome independently of obesity, lipid indices and serum insulin levels: a cross-sectional study Ntzouvani Agathi 1Fragopoulou Elisabeth 1Panagiotakos Demosthenes 2Pitsavos Christos 3Antonopoulou Smaragdi +30 2109549230antonop@hua.gr 11 Laboratory of Biology, Biochemistry, Physiology and Microbiology, Department of Nutrition and Dietetics, School of Health Science and Education, Harokopio University, Eleftheriou Venizelou 70, Athens, 17671 Greece 2 Department of Nutrition and Dietetics, School of Health Science and Education, Harokopio University, Athens, Greece 3 First Cardiology Clinic, School of Medicine, National and Kapodistrian University of Athens, Athens, Greece 27 8 2016 27 8 2016 2016 15 1 14014 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 Given the increasing rate of overweight and the burden of metabolic syndrome (MetS) on cardiovascular disease development, better understanding of the syndrome is of great importance. Therefore, the objectives were to examine whether interleukin-6 (IL-6) and adiponectin are associated with MetS, and whether this association is mediated by components of the MetS. Methods During 2011–2012, 284 individuals (159 men, 53 ± 9 years, 125 women 52 ± 9 years) without cardiovascular disease, type 1 diabetes mellitus, high-grade inflammatory disease, living in the greater Athens area, Greece, participated in clinical examination. Adiponectin and IL-6 were measured in fasting plasma samples. MetS was defined according to the International Diabetes Federation (IDF) and the American Heart Association/National Heart, Lung, and Blood Institute (AHA/NHLBI) criteria. Results MetS was present in 37 % (IDF) and 33 % (AHA/NHLBI) of the study population (P < 0.001). Adiponectin was inversely associated with MetS (odds ratio, 95 % confidence interval: 0.829, 0.762- 0.902 for MetS-IDF, and 0.840, 0.772- 0.914 for MetS-AHA/NHLBI). Body mass index (BMI), waist circumference, high density lipoprotein (HDL)-cholesterol, triglyceride and insulin concentration mediated the association between adiponectin and MetS-IDF (z-test, standard error, P-value: 2.898, 0.012, 0.004, for BMI; 2.732, 0.012, 0.006 for waist circumference; 2.388, 0.011, 0.017 for HDL-cholesterol; 2.163, 0.010, 0.031 for triglyceride; 2.539,  0.010,  0.011 for insulin). Similarly, BMI, waist circumference, HDL-cholesterol and insulin concentration mediated the association between adiponectin and MetS-AHA/NHLBI (z-test, standard error, P-value: 2.633, 0.011, 0.008 for BMI; 2.441, 0.011, 0.015 for waist circumference; 1.980, 0.010, 0.048 for HDL-cholesterol; 2.225, 0.009, 0.026 for insulin). However, adiponectin remained significantly associated with MetS. IL-6 was not significantly associated with MetS. Conclusion MetS components, in particular obesity and lipid indices, as well as serum insulin levels, mediate the association between adiponectin and MetS as defined by both the IDF and AHA/NHLBI criteria. Keywords Metabolic syndromeAdiponectinInterleukin-6Mediation effectissue-copyright-statement© The Author(s) 2016 ==== Body Background Metabolic syndrome (MetS) is a cluster of factors of clinical importance that increases the risk of cardiovascular disease (CVD). These factors are widely accepted indices of obesity, metabolic function and blood pressure homeostasis [1]. Cardiovascular disease (CVD) incidence has been recently associated with dyslipidemia, diabetes mellitus and hypertension in a cohort study of CVD epidemiology in Greece [2]. Several organizations formulated simple criteria for the diagnosis of MetS in the clinical practice in order to modify risk factors for CVD development through lifestyle changes. The World Health Organization, WHO, (1998) consultation group, the European Group for Study of Insulin Resistance, EGIR, (1999), and the American Association of Clinical Endocrinologists, AACE, (2003) emphasized insulin resistance as the underlying cause of MetS and required evidence of insulin resistance for diagnosis. The National Cholesterol Education Program Adult Treatment Panel III, NCEP ATP III, (2001) aimed at identifying people at higher long-term risk for atherosclerotic cardiovascular disease who deserved lifestyle intervention to reduce risk. The International Diabetes Federation, IDF, writing group (2005) considered that abdominal obesity is strongly correlated with insulin resistance, and made the presence of abdominal obesity a prerequisite for the diagnosis of MetS. The American Heart Association/National Heart, Lung, and Blood Institute, AHA/NHLBI, (2005) scientific statement revised the NCEP ATP III criteria as regards the threshold for impaired fasting glucose. Neither the NCEP ATP III nor the AHA/NHLBI criteria drew conclusions on mechanistic pathogenesis [3]. The occurrence of MetS has been characterized as a global epidemic. The two most widely used definitions of MetS are based on the NCEP ATP III (2001) and the IDF (2005) criteria [1]. The prevalence of MetS in the Greek population was evaluated in two population-based epidemiological studies, the ATTICA study [4] and the MetS Greece Study [5]. The prevalence of MetS was 23.6 % [5], according to the NCEP ATP III definition, and 48.9 % according to the IDF definition [6]. The most prevalent abnormality among subjects with the MetS was obesity, particularly abdominal obesity. Abdominal obesity is considered the predominant underlying cause of MetS and is associated with both insulin resistance and low-grade chronic inflammation [3]. Waist circumference is a widely used index of abdominal obesity, and was found to be a better predictor of MetS compared with BMI, waist-to-hip ratio and waist-to-height ratio [7]. Metabolic health has been associated with lower concentrations of pro-inflammatory cytokines (e.g. IL-6), and higher concentrations of anti-inflammatory adipokines (e.g. adiponectin) in both obese and non-obese adults [8]. Presence of MetS and its components have been associated with increased levels of IL-6 and decreased levels of adiponectin [3, 9]. A cross-sectional data analysis from the Diet and Omega-3 Intervention Trial on Atherosclerosis (DOIT) showed that serum levels of IL-6 were significantly higher in subjects with MetS compared to those without MetS, but there was no significant association between IL-6 and increasing MetS components [10]. On the contrary, the proportion of subjects with MetS, declined across sex-specific adiponectin quartiles in the context of the Carotid Ultrasound Disease Assessment Study [11]. Given the increasing rates of overweight and obesity, as well as the burden of MetS on cardiovascular disease development, better understanding of the syndrome is of great importance. Thus, the present study evaluated the prevalence of MetS in a sample of the Greek population, using two definitions which include an index of abdominal obesity and differ only by the waist circumference criteria. The hypothesis was that adiponectin and IL-6 plasma concentration is associated with MetS through its components. Therefore, the aims of the present study were to examine i) whether IL-6 or adiponectin concentration is associated with MetS, and ii) whether this association is mediated by components of the MetS. The ability of IL-6 or adiponectin concentration in identifying individuals with MetS was also evaluated. Methods Participants This was a cross-sectional study carried out in the greater area of Athens (78 % urban and 22 % rural regions) during 2011–2012. The study population consisted of individuals aged > 30 years from the general population. Participants responded to an invitation to health evaluation which was published at the participants’ workplace. Five hundred individuals participated in the initial evaluation (Fig. 1). The sampling was based on a feasibility basis, and the evaluation was performed at each participant’s workplace or home by trained personnel (cardiologists, general practitioners, dietitians and nurses).Fig. 1 Study flowchart Participants diagnosed with cardiovascular disease (i.e. myocardial infarction, angina pectoris, other identified forms of ischemia; coronary revascularization: coronary artery bypass surgery and percutaneous coronary intervention, heart failure of different types, chronic arrhythmias, or stroke) were excluded from the study. Other exclusion criteria were presence of high-grade chronic inflammatory disease (e.g. rheumatoid arthritis, inflammatory bowel disease, atopic dermatitis, and asthma), viral infections, cold or flu, acute respiratory infection, dental problems, any type of surgery the month preceding the study, and type 1 diabetes mellitus. Two hundred eighty four participants who were eligible to participate in the study and had complete lifestyle, clinical and biochemical data were included in the present study (Fig. 1); 159 participants were men (53 ± 9 years) and 125 were women (52 ± 9 years). No significant differences were observed between participants who were finally included in the study and the rest of the participants who were excluded, as regards age and sex (P > 0.30, for all). Lifestyle evaluation Dietary habits were evaluated with a validated semi-quantitative food-frequency questionnaire [12]; overall dietary habits were evaluated using the MedDietScore (range 0–55) that incorporates the inherent characteristics of the Mediterranean diet [13]. Smokers were defined as those who were smoking at least one cigarette per day during the past year or had recently stopped smoking (within the last 12 months); the rest of the participants were defined as non-smokers. Passive smokers were defined as those who were exposed to other peoples’ smoke for more than 30 min/day and more than 5 days/week. The criteria for defining passive smoking were based on literature addressing the biologic effects of second-hand smoke on the cardiovascular system [14, 15]. Physical activity level was evaluated with the International Physical Activity Questionnaire (IPAQ), modified and adapted for the Greek population. Physical activity was classified as vigorous, moderate and walking physical activity and expressed in MET-minutes per week (MET.min.wk−1). Total physical activity level and sitting hours per day were also evaluated [16]. Clinical evaluation Resting blood pressure was measured twice on the right arm with an electronic monitor device. All participants were at least 30 min at rest before measurement which was performed in sitting position. Diagnosis and current medication treatment for hypertension, hyperlipidemia and type 2 diabetes mellitus were recorded in a self-administered questionnaire. Waist circumference (in centimeters, cm) was measured in the middle between the 12th rib and the iliac crest. Hip circumference (in cm) was measured around the buttocks at the level of the maximum extension. Height was measured to the nearest 0.5 cm, without shoes, back against the wall tape, and eyes looking straight ahead. Weight was measured with a lever balance, to the nearest 100 g, without shoes, in light undergarments. BMI was calculated as weight (in kilograms, kg) divided by height (in meters squared, m2). Overweight was defined as BMI between 25 and 29.9 kg/m2, while obesity as BMI greater than 29.9 kg/m2, based on the WHO criteria. Blood collection and biochemical analyses Venous blood samples were collected between 08:00 and 10:00, after 12 h overnight fast, with the participant in sitting position. Instructions about the preparation before blood collection were given to each participant either by telephone or by e-mail. Fasting serum was obtained by collecting blood into silicone coated Vacutainer Tubes (Becton Dickinson) with clot activator. Blood was allowed to clot at room temperature (18 – 25 °C) for 60 min and immediately centrifuged for 10 min at 1,500xg before isolation of the serum fraction. Fasting plasma was obtained by collecting blood into K2-EDTA (EDTA-dipotassium salt) Vacutainer Tubes (Becton Dickinson); the final EDTA concentration in the samples was 4 mmol/L. The EDTA blood samples were centrifuged within 60 min at room temperature for 10 min at 1,500xg. Plasma and serum aliquots were stored at −80 °C until use. Serum total cholesterol, HDL-cholesterol, triglyceride, and glucose concentration were measured on a COBAS 8000/ ROCHE analyzer, based on colorimetric detection. The CHOD-PAP method was applied for total cholesterol (2.06 % intra-assay coefficient of variation-CV, 0.94 % inter-assay CV) and HDL cholesterol (1.50 % intra-CV, 0.80 % inter-CV), the GPO-PAP method for triglycerides (1.80 % intra-CV, 1.98 % inter-CV), and the GOD-PAP method for glucose (1.97 % intra-CV, 1.28 % inter-CV). All measurements were carried out at the same laboratory (BIOMED S.A., accreditation standard ELOT EN ISO 15189, Hellenic Accreditation System – E.SY.D.). None of the study participants had triglyceride values >4.5 mmol/L. LDL-cholesterol was estimated with the Friedewald equation: (total cholesterol) – (HDL cholesterol) – (triglycerides/2.2) [17]. All biochemical indices were measured in duplicate and are expressed in mmol/L. Serum insulin concentration was measured on a TOSOH AIA-600 II automated enzyme immunoassay analyzer using a two-site immune-enzymometric assay. The intra- and inter-assay CVs were <3 %. Insulin concentration is expressed in mU/L. IL-6 and adiponectin concentrations were measured in plasma EDTA samples. IL-6 was measured by a commercially available ELISA method (Quantikine HS, R&D Systems Europe Ltd., Abingdon, U.K.) with an assay range of 0.156 -10 pg/ml. Adiponectin was measured by a commercially available ELISA method (Quantikine, R&D Systems Europe Ltd., Abingdon, U.K.) with an assay range of 3.9 - 250 ng/ml. The intra- and inter-assay CVs were <10 % for IL-6 and <7 % for adiponectin. Il-6 and adiponectin concentrations are expressed in pg/ml and μg/ml, respectively. Definition of metabolic syndrome Metabolic syndrome was defined by the AHA/NHLBI and the IDF criteria [3]. According to the AHA/NHLBI definition, the diagnosis of MetS is established when 3 of 5 factors are present: abdominal obesity, elevated triglycerides, reduced HDL cholesterol, elevated blood pressure, elevated fasting glucose or type 2 diabetes mellitus. The diagnostic criteria are: waist circumference ≥102 cm for men and ≥88 cm for women, triglycerides ≥1.7 mmol/L, HDL-cholesterol <1.03 mmol/L for men and <1.3 mmol/L for women, systolic blood pressure (SBP) ≥130 mmHg or diastolic blood pressure (DBP) ≥85 mmHg, and fasting blood glucose ≥5.6 mmol/L. Participants on drug treatment for elevated triglycerides, reduced HDL-cholesterol, elevated blood pressure or hyperglycemia were considered as meeting the aforementioned criteria, respectively. Presence of MetS was also defined by the IDF criteria. According to the IDF, one must have abdominal obesity (waist circumference ≥94 cm for men, and ≥80 cm for women of Europid origin) or BMI ≥30 kg/m2 and any two of the other risk factors mentioned in the AHA/NHLBI definition. Bioethics The study adhered to the Declaration of Helsinki principles, and it was approved by the Bioethics Committee of Harokopio University, Athens. Participants were informed about the aims and the procedure of the study and they provided their written consent prior to the collection of any information. Statistical analysis The prevalence of MetS was determined as frequencies, using both definitions. Continuous variables are expressed as median and interquartile range (IQR). Categorical variables are expressed as frequencies. Normal distribution of continuous variables was tested with the Kolmogorov-Smirnov test and P-P plots. Adiponectin and IL-6 had a rightly skewed distribution and the values were log-transformed (log10). The association of continuous variables (i.e. age, HDL-cholesterol, LDL-cholesterol, triglycerides, glucose, insulin, SBP, DBP, BMI, waist circumference, IL-6 and adiponectin) with MetS was evaluated using the non-parametric Kolmogorov - Smirnov Z-test (due to the skewed distribution of the variables), whereas the Pearson Chi-Square test was used for the categorical variables (i.e. sex and medical treatment for hypertension, hypercholesterolemia, type 2 diabetes mellitus). Pearson’s unadjusted and partial (adjusted for sex and age) correlation coefficient (r) was applied to evaluate correlations between adiponectin, IL-6 and MetS components (i.e. waist circumference, BMI, HDL-cholesterol, triglycerides, glucose, insulin, SBP and DBP). Linear regression was applied, where adiponectin or IL-6 was the dependent variable and sex, age and MetS components were the independent variables; results are presented as beta coefficients and SE. The collinearity statistics (tolerance and variance inflation factor, VIF) showed that there was no problem with multicollinearity. Logistic regression was applied, where MetS was the dependent variable and adiponectin or IL-6 was the independent variable, and adjusted for sex and age, and further adjusted for BMI, waist circumference, HDL-cholesterol, triglyceride and insulin concentration in order to test for the potential mediating effect of these factors as regards the adiponectin - MetS and IL-6-MetS association. Results are expressed as odds ratio (OR) and the corresponding 95 % confidence interval (CI). The Sobel Test was used to determine whether the indirect association between the independent variable (i.e., IL-6 or adiponectin) and the dependent variable (i.e., MetS) via the mediator is significantly different from zero [18]. The test was performed online http://quantpsy.org/sobel/sobel.htm. The receiver operating characteristic (ROC) analysis was used to assess the ability of adiponectin to discriminate between participants with MetS and participants without MetS after adjustment for sex, age, and further adjustment for BMI, waist circumference and insulin concentration. Results are expressed as the area under the ROC curve (AUC) and the corresponding 95 % CI. The AUC provides a scale from 0.5 to 1.0 (i.e. 0.5 represents random chance and 1.0 indicates perfect discrimination) by which to appraise the accuracy of adiponectin. ROC analysis was performed using R (version 3.2.4 for Windows), package “pROC” [19]. SPSS version 21 (Statistical Package for Social Sciences, SPSS Inc., Chicago, IL, USA) was used for all other statistical analyses. All reported P-values are based on two-sided tests, and statistical significance was set at P < 0.05. Results Prevalence of MetS and comparison for lifestyle, clinical and biochemical factors Table 1 shows the general characteristics of the study population, and the comparison between participants with MetS and participants without MetS. The prevalence of MetS was 37 % of the total study population (32 women and 74 men) according to the IDF criteria, and 33 % (29 women and 66 men) according to the AHA/NHLBI criteria. The prevalence of MetS differed significantly between the two definitions (P < 0.001). Participants with MetS did not differ significantly as regards current smoking habits (smokers vs. non-smokers, cigarettes per day and years of smoking), passive smoking (passive smokers vs. non-passive smokers), physical activity level (MET.min.wk−1), sitting hours (h/day) or dietary habits (portions of food group consumption per wk and MedDietScore) compared with participants without MetS as defined by either the IDF or the AHA/NHLBI criteria (data not shown).Table 1 Clinical and biochemical factors in the total study sample, participants without metabolic syndrome and participants with metabolic syndrome IDF (n, %) AHA/NHLBI (n, %) All no MetS (178, 63) MetS (106, 37) no MetS (189, 67) MetS (95, 33) Age (years) 52.00 (46.00, 59.88) 51.00 (45.00, 56.00) 56.50 (49.00, 62.25)b 51.00 (45.00, 56.50) 56.00 (49.00, 62.00)b Sex (men, women), n (%) 159 (56)/ 125(44) 85 (48)/ 93 (52) 74 (70)/ 32 (30)a 93 (49)/ 96 (51) 66 (69)/ 29 (31)a HDL-cholesterol (mmol/L) 1.29 (1.08, 1.51) 1.38 (1.21, 1.58) 1.09 (0.94, 1.29)b 1.38 (1.20, 1.58) 1.08 (0.94, 1.25)b LDL-cholesterol (mmol/L) 3.43 (2.91, 4.04) 3.46 (2.97, 4.06) 3.30 (2.73, 3.92) 3.45 (2.94, 3.95) 3.38 (2.90, 4.10) Triglycerides (mmol/L) 1.15 (0.86, 1.62) 1.05 (0.76, 1.30) 1.61 (1.14, 2.18)b 1.05 (0.77, 1.30) 1.70 (1.13, 2.33)b Glucose (mmol/L) 4.99 (4.72, 5.35) 4.90 (4.69, 5.17) 5.22 (4.86, 5.71)b 4.90 (4.68, 5.17) 5.24 (4.90, 5.82)b Insulin (mU/L) 8.20 (5.80, 11.85) 6.80 (4.90, 9.20) 11.30 (8.65, 14.00)b 6.80 (5.00, 9.05) 11.80 (9.10, 14.87)b Medical treatment for:  Hypertension (no/yes), n (%) 213 (75)/ 49 (17.3) 152 (85.4)/ 14 (7.9) 61 (57.5)/ 35 (33)a 155 (82)/ 20 (10.6) 58 (61.1)/ 29 (30.5)a  Hypercholesterolemia (no/yes), n (%) 161 (56.7)/ 109 (38.4) 118 (66.3)/ 52 (29.2) 43 (40.6)/ 57 (53.8)a 122 (64.6)/ 57 (30.2) 39 (41.1)/ 52 (54.7)a  Type 2 diabetes mellitus (no/yes), n (%) 250 (88)/ 14 (4.9) 165 (92.7)/ 2 (1.1) 85 (80.2)/ 12 (11.3)a 175 (92.6)/ 2 (1.1) 75 (78.9)/ 12 (12.6)a  SBP (mmHg) 123.00 (114.00, 133.50) 118.00 (109.75, 127.00) 133.00 (124.75, 138.00)b 118.50 (110.00, 128.00) 133.00 (125.25, 138.00)b  DBP (mmHg) 79.00 (70.00, 86.00) 75.25 (69.00, 82.40) 84.00 (77.00, 90.00)b 76.00 (69.00, 83.00) 85.00 (77.00, 90.00)b  BMI (kg/m2) 27.25 (24.73, 31.00) 25.40 (23.90, 29.00) 29.80 (27.20, 32.40)b 25.70 (23.95, 28.85) 30.00 (28.00, 32.70)b  Waist circumference (cm) 96.00 (86.00, 106.00) 90.00 (81.25, 98.88) 103.00 (97.75, 111.25)b 90.00 (82.00, 99.00) 105.00 (98.00, 112.00)b  IL-6 plasma concentration (pg/mL) 1.621 (0.983, 2.696) 1.392 (0.897, 2.209) 2.152 (1.274, 3.026)b 1.394 (0.897, 2.329) 2.099 (1.391, 3.036)b  Adiponectin plasma concentration (μg/mL) 7.419 (4.763, 10.516) 8.471 (5.883, 12.944) 5.546 (3.799, 8.693)b 8.473 (5.680, 12.579) 5.476 (4.072, 8.454)b Quantitative variables are expressed as median (25th, 75th IQR) aPearson Chi-Square, P ≤ 0.001 for comparison between no MetS and MetS separately for two definitions of MetS bKolmogorov-Smirnov Z Test, P ≤ 0.001 for comparison between no MetS and MetS separately for two definitions of MetS According to the unadjusted bivariate associations, adiponectin was positively correlated with HDL-cholesterol (r = 0.385, P < 0.001) and negatively correlated with triglycerides (r = −0.362, P < 0.001), glucose (r = −0.219, P < 0.001), insulin (r = −0.312, P < 0.001), SBP (r = −0.123, P < 0.05), DBP (r = −0.128, P < 0.05), BMI (r = −0.171, P < 0.01) and waist circumference (r = −0.391, P < 0.001). On the contrary, IL-6 was negatively correlated with HDL-cholesterol (r = −0.308, P < 0.001) and positively correlated with triglycerides (r = 0.219, P < 0.001), insulin (r = 0.125, P < 0.05), SBP (r = 0.124, P < 0.05), BMI (r = 0.267, P < 0.001) and waist circumference (r = 0.298, P < 0.001). Age was significantly correlated with triglyceride, glucose and insulin concentration (r = 0.164, P < 0.01, r = 0.265, P < 0.001, and r = 0.132, P < 0.05 respectively) in the total study population. Age was also significantly correlated with SBP and waist circumference (r = 0.346, P < 0.001 and r = 0.132, P <0.05, respectively). Men had significantly higher triglyceride, glucose and insulin concentration, SBP and DBP, BMI and waist circumference than women. On the contrary, men had lower HDL-cholesterol than women (P < 0.001, for all). After adjustment for age and sex, the association between adiponectin and HDL-cholesterol (r = 0.282), triglycerides (r = −0.319), glucose (r = −0.140), insulin (r = −0.305), BMI (r = −0.161) and waist circumference (r = −0.208) remained statistically significant. The association between IL-6 and HDL-cholesterol (r = −0.335), triglycerides (r = 0.199), insulin (r = 0.130), BMI (r = 0.264) and waist circumference (r = 0.333) remained statistically significant after adjustment for sex and age. There was no significant association between adiponectin and IL-6 either in the unadjusted or the adjusted model. In the multivariate adjusted models, adiponectin was significantly associated with serum triglyceride concentration (Table 2; Model 1 and Model 2) and waist circumference (Table 2; Model 2). Insulin was an independent associate of adiponectin (Table 2; Model 3, β = −0.164, P = 0.006, and Model 4, β = −0.148, P = 0.011) after adjusting for sex, age and MetS components. Waist circumference did not remain an independent associate of adiponectin after adjustment for insulin (Table 2; Model 4). IL-6 was significantly associated with HDL-cholesterol, BMI and waist circumference in the multivariate adjusted models (Table 3). Insulin was not an independent associate of IL-6 after adjustment for sex, age and MetS components (Table 3; Model 3 and Model 4).Table 2 Multivariate adjusted associations between metabolic syndrome components and log10(adiponectin) Model 1 R 2 = 0.376 Model 3 R 2 = 0.393 B SE β t P B SE β t P Sex (women/men) −0.194 0.025 −0.405 −7.618 <0.001 −0.189 0.025 −0.395 −7.456 <0.001 Age (years) 0.005 0.001 0.210 4.061 <0.001 0.005 0.001 0.202 3.942 <0.001 HDL-cholesterol (mmol/L) 0.002 0.001 0.119 1.931 0.055 0.002 0.001 0.119 1.948 0.053 Triglycerides (mmol/L) −0.001 0.000 −0.213 −3.538 <0.001 −0.001 0.000 −0.169 −2.736 0.007 Glucose (mmol/L) −0.001 0.001 −0.046 −0.842 0.401 −0.001 0.001 −0.040 −0.741 0.459 Body mass index (kg/m2) −0.004 0.003 −0.074 −1.432 0.153 0.000 0.003 −0.004 −0.074 0.941 Insulin (mU/L) - - - - - −0.007 0.002 −0.164 −2.755 0.006 Model 2 R 2 = 0.380 Model 4 R 2 = 0.394 B SE β t P B SE β t P Sex (women/men) −0.172 0.028 −0.359 −6.206 <0.001 −0.179 0.028 −0.375 −6.410 <0.001 Age (years) 0.006 0.001 0.211 4.103 <0.001 0.005 0.001 0.205 3.999 <0.001 HDL-cholesterol (mmol/L) 0.002 0.001 0.119 1.954 0.052 0.002 0.001 0.117 1.912 0.057 Triglycerides (mmol/L) −0.001 0.000 −0.201 −3.313 0.001 −0.001 0.000 −0.165 −2.670 0.008 Glucose (mmol/L) −0.001 0.001 −0.043 −0.784 0.434 −0.001 0.001 −0.037 −0.674 0.501 Waist circumference (cm) −0.002 0.001 −0.117 −1.975 0.049 −0.001 0.001 −0.054 −0.848 0.398 Insulin (mU/L) - - - - - −0.006 0.002 −0.148 −2.558 0.011 Table 3 Multivariate adjusted associations between metabolic syndrome components and log10(IL-6) Model 1 R 2 = 0.199 Model 3 R 2 = 0.220 B SE β t P B SE β t P Sex (women/men) −0.079 0.040 −0.117 −1.951 0.052 −0.090 0.040 −0.134 −2.224 0.027 Age (years) 0.008 0.002 0.211 3.605 <0.001 0.008 0.002 0.216 3.720 <0.001 HDL-cholesterol (mmol/L) −0.008 0.002 −0.265 −3.817 <0.001 −0.008 0.002 −0.286 −4.110 <0.001 Triglycerides (mmol/L) 0.000 0.000 0.067 0.985 0.326 0.000 0.000 0.039 0.564 0.573 Glucose (mmol/L) −0.002 0.002 −0.050 −0.803 0.422 −0.002 0.002 −0.059 −0.953 0.342 Body mass index (kg/m2) 0.013 0.004 0.201 3.483 0.001 0.011 0.004 0.172 2.793 0.006 Insulin (mU/L) - - - - - 0.005 0.004 0.084 1.296 0.196 Model 2 R 2 = 0.236 Model 4 R 2 = 0.253 B SE β t P B SE β t P Sex (women/men) −0.171 0.043 −0.256 −3.991 <0.001 −0.179 0.043 −0.268 −4.123 <0.001 Age (years) 0.007 0.002 0.205 3.586 <0.001 0.008 0.002 0.214 3.740 <0.001 HDL-cholesterol (mmol/L) −0.007 0.002 −0.261 −3.884 <0.001 −0.008 0.002 −0.287 −4.242 <0.001 Triglycerides (mmol/L) 0.000 0.000 0.040 0.591 0.555 0.000 0.000 0.027 0.397 0.692 Glucose (mmol/L) −0.002 0.002 −0.067 −1.110 0.268 −0.002 0.002 −0.072 −1.184 0.238 Waist circumference (cm) 0.008 0.002 0.320 4.872 <0.001 0.008 0.002 0.312 4.396 <0.001 Insulin (mU/L) - - - - - 0.001 0.004 0.025 0.382 0.703 Assessing the potential mediation effect of insulin, BMI, waist circumference, HDL-cholesterol and triglycerides Plasma adiponectin concentration was significantly associated with MetS (OR, 95 % CI: 0.829, 0.762- 0.902 for MetS-IDF, and 0.840, 0.772- 0.914 for MetS-AHA/NHLBI) (Table 4), whereas plasma IL-6 concentration was not significantly associated with MetS (Table 5). Adiponectin remained significantly associated with MetS after controlling for insulin, BMI, waist circumference, HDL-cholesterol or triglyceride concentration. However, adiponectin was not significantly associated with MetS-AHA/NHLBI after controlling for triglyceride concentration (Table 4). Results of the Sobel Test showed that the indirect association between adiponectin and MetS-IDF was significantly different from zero after controlling for insulin (z-test = 2.539, SE = 0.010, P = 0.011), BMI (z-test = 2.898, SE = 0.012, P = 0.004), waist circumference (z-test = 2.732, SE = 0.012, P = 0.006), HDL-cholesterol (z-test = 2.388, SE = 0.011, P = 0.017) or triglyceride concentration (z-test = 2.163, SE = 0.010, P = 0.031). Similarly, the indirect association between adiponectin and MetS-AHA/NHLBI was significantly different from zero after controlling for insulin (z-test = 2.225, SE = 0.009, P = 0.026), BMI (z-test = 2.633, SE = 0.011, P = 0.008), waist circumference (z-test = 2.441, SE = 0.011, P = 0.015) or HDL-cholesterol (z-test = 1.980, SE = 0.010, P = 0.048).Table 4 Association between plasma adiponectin concentration and metabolic syndrome, and the mediation effect of MS components OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI) MetS-IDF Age (years) 1.087 (1.052- 1.123) 1.084 (1.048- 1.122) 1.088 (1.051- 1.126) 1.084 (1.045- 1.124) 1.089 (1.051- 1.128) 1.075 (1.037- 1.113) Sex (men) 0.757 (0.404- 1.418) 0.685 (0.353-1.330) 0.652 (0.330- 1.290) 1.656 (0.784- 3.497) 0.976 (0.487- 1.956) 0.743 (0.372- 1.486) Adiponectin (μg/ml) 0.829 (0.762- 0.902) 0.870 (0.797-0.949) 0.837 (0.764- 0.916) 0.846 (0.769- 0.930) 0.875 (0.798- 0.958) 0.894 (0.818- 0.977) Insulin (mU/L) - 1.131 (1.069-1.197) - - - - BMI (kg/m2) - - 1.193 (1.113- 1.278) - - - Waist circumference (cm) - - 1.098 (1.064- 1.133) - - HDL-cholesterol (mmol/L) - - - 0.904 (0.873- 0.937) - Triglycerides (mmol/L) - - - - 1.020 (1.013- 1.027) MetS-AHA/NHLBI Age (years) 1.082 (1.047- 1.117) 1.080 (1.043-1.118) 1.083 (1.046- 1.122) 1.078 (1.039- 1.117) 1.086 (1.047- 1.126) 1.070 (1.031- 1.110) Sex (men) 0.803 (0.425- 1.517) 0.720 (0.366-1.418) 0.667 (0.331- 1.343) 1.811 (0.844- 3.886) 1.103 (0.534- 2.276) 0.779 (0.377- 1.609) Adiponectin (μg/ml) 0.840 (0.772- 0.914) 0.887 (0.813-0.969) 0.850 (0.776- 0.931) 0.861 (0.783- 0.947) 0.897 (0.817- 0.985) 0.921 (0.842- 1.007) Insulin (mU/L) - 1.147 (1.083-1.215) - - - - BMI (kg/m2) - - 1.212 (1.129- 1.302) - - - Waist circumference (cm) - - 1.105 (1.070- 1.141) - - HDL-cholesterol (mmol/L) - - - 0.884 (0.850- 0.919) - Triglycerides (mmol/L) - - - - 1.023 (1.016- 1.031) Table 5 Association between plasma IL-6 concentration and metabolic syndrome, and the mediation effect of MS components OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI) MetS-IDF Age (years) 1.072 (1.040- 1.105) 1.072 (1.038-1.108) 1.076 (1.042- 1.111) 1.072 (1.037- 1.109) 1.091 (1.053- 1.129) 1.068 (1.032- 1.105) Sex (men) 0.357 (0.205- 0.624) 0.405 (0.223-0.738) 0.345 (0.190- 0.626) 0.957 (0.481- 1.907) 0.690 (0.364- 1.309) 0.493 (0.262- 0.927) IL-6 (pg/ml) 1.086 (0.975- 1.210) 1.051 (0.935-1.182) 1.054 (0.934- 1.191) 1.021 (0.894- 1.166) 0.968 (0.853- 1.099) 1.033 (0.907- 1.176) Insulin (mU/L) - 1.154 (1.090-1.221) - - - - BMI (kg/m2) - - 1.139 (1.073- 1.209) - - - Waist circumference (cm) - - - 1.097 (1.065- 1.131) - - HDL-cholesterol (mmol/L) - - - - 0.893 (0.861- 0.925) - Triglycerides (mmol/L) - - - - - 1.023 (1.015- 1.030) MetS-AHA/NHLBI Age (years) 1.066 (1.034- 1.098) 1.067 (1.033-1.103) 1.072 (1.038- 1.108) 1.068 (1.031- 1.105) 1.090 (1.051- 1.131) 1.063 (1.026- 1.102) Sex (men) 0.395 (0.224- 0.695) 0.456 (0.246-0.844) 0.372 (0.201- 0.689) 1.171 (0.573- 2.395) 0.923 (0.467- 1.824) 0.579 (0.299- 1.123) IL-6 (pg/ml) 1.064 (0.954- 1.186) 1.024 (0.907-1.156) 1.019 (0.894- 1.162) 0.983 (0.854- 1.131) 0.918 (0.805- 1.048) 0.995 (0.870- 1.139) Insulin (mU/L) - 1.168 (1.103-1.237) - - - - BMI (kg/m2) - - 1.167 (1.094- 1.244) - - - Waist circumference (cm) - - - 1.110 (1.075- 1.146) - - HDL-cholesterol (mmol/L) - - - - 0.865 (0.830- 0.902) - Triglycerides (mmol/L) - - - - - 1.026 (1.018- 1.034) Accuracy of factors associated with the identification of MetS The AUC and the 95 % CI were calculated for adiponectin after adjustment for sex, age, and further adjustment for insulin, BMI or waist circumference, in order to assess the ability of adiponectin to discriminate between participants with MetS and participants without MetS. Adiponectin performed better than chance (i.e., the AUC was significantly greater than 0.5) in classifying correctly subjects with MetS, both in the model that adjusted only for sex and age, as well as in the model that further adjusted for insulin, BMI or waist circumference (Table 6). The results were similar independently of the definition used. Since IL-6 concentration was not significantly associated with MetS, the AUC was not calculated for this biomarker.Table 6 Discriminative accuracy of adiponectin, insulin, BMI and waist circumference in the prediction of prevalent metabolic syndrome AUC 95 % CI MetS-IDF Adiponectin 0.744 0.685–0.801 Insulin 0.783 0.730–0.837 BMI 0.752 0.688–0.815 Waist circumference 0.798 0.743–0.853 Adiponectin and insulin 0.813 0.760–0.858 Adiponectin and BMI 0.793 0.734–0.850 Adiponectin and waist circumference 0.825 0.767–0.883 Adiponectin and insulin and BMI 0.829 0.780–0.874 Adiponectin and insulin and waist circumference 0.858 0.814–0.896 MetS-AHA/NHLBI Adiponectin 0.731 0.672–0.791 Insulin 0.789 0.734–0.843 BMI 0.758 0.696–0.817 Waist circumference 0.803 0.746–0.859 Adiponectin and insulin 0.818 0.763–0.862 Adiponectin and BMI 0.791 0.732–0.849 Adiponectin and waist circumference 0.822 0.767–0.879 Adiponectin and insulin and BMI 0.840 0.793–0.884 Adiponectin and insulin and waist circumference 0.860 0.813–0.899 All models were adjusted for sex and age Discussion The major findings of the present study are summarized as follows. Adiponectin was significantly associated with prevalent MetS, whereas IL-6 was not significantly associated with MetS. MetS components mediated the association between adiponectin and MetS, but adiponectin remained significantly associated with MetS. Adiponectin had a high discriminative accuracy for MetS independently of the definition used. All the results were similar regardless of the definition used. Even though abdominal obesity is a prerequisite for the IDF definition and the cut-off value of waist circumference is different between the two definitions used in the present study, the results indicated that the IDF and AHA/NHLBI definitions for evaluating presence of MetS are practically identical. However, we could not draw definite conclusions based on the results of the present study due to two main limitations, i.e. the cross-sectional design of the study and the relatively small sample size. Nevertheless, the importance of the present study lies in exploring MetS beyond its clinical and biochemical constituents. Obesity, particularly abdominal obesity, has been identified as a significant constituent of MetS. According to the IDF definition, abdominal obesity, as assessed by waist circumference, is an essential diagnostic criterion because of the strength of the evidence linking waist circumference with metabolic abnormalities [20]. Dyslipidaemia was the major metabolic abnormality and the second most frequent MetS component after abdominal obesity, found in our sample. Dyslipidaemia, which is primarily characterized by elevated plasma FFA and triglycerides, decreased levels of HDL-cholesterol, and abnormal LDL composition, is a main risk factor for CVD incidence and mortality. Normalization of fasting blood lipid levels can improve the cardiovascular risk profile of individuals. The Mediterranean dietary pattern is a lifestyle factor which consists of food groups that are already known to ameliorate dyslipidaemia and decrease the incidence of cardiovascular events, probably by means of their nutrient content, such as resveratrol and polyphenols in red wine, fish oil and proteins, polyphenols, and phytosterols in fruits and vegetables [21]. In the present study, participants with MetS did not differ significantly as regards food group consumption or their overall score of adherence to the Mediterranean dietary pattern compared with participants without MetS. However, this was a cross-sectional study, and the participants’ dietary habits were evaluated once using a semi-quantitative food-frequency questionnaire, therefore evaluation of true intake may be inaccurate due to daily and seasonal effects, as well as other personal characteristics [12]. Metabolic health has been associated with an altered secretion pattern of adipokines. Specifically, production of pro-inflammatory cytokines is enhanced, whereas production of adiponectin is inhibited in the adipose tissue in the presence of obesity [22]. Furthermore, certain adipokines, such as leptin, are mainly associated with total obesity, whereas others, such as IL-6 and adiponectin may be more closely linked with abdominal obesity [23]. It is estimated that 15–35 % of total IL-6 concentration in the circulation originates from the adipose tissue where it is produced by non-adipocytes, such as fibroplasts, endothelial cells, and monocytes [24]. Adiponectin is abundantly expressed in the white adipose tissue by mature adipocytes [25, 26], and it circulates in different oligomeric forms, i.e. low-molecular weight (LMW) trimers, medium-molecular weight (MMW) hexamers, and high-molecular weight (HMW) oligomers [27]. HMW adiponectin has been suggested as the biologically active form of adiponectin [28], and circulating HMW adiponectin, rather than total adiponectin, has been associated with insulin sensitivity [29], as well as with anteroposterior diameter of infra-renal abdominal aorta (APAO), an ultrasound early marker of atherosclerosis [30]. Nevertheless, there has been found a strong association between total and HMW adiponectin in circulation [31, 32]. In addition, both total and HMW adiponectin blood levels have been inversely associated with biomarkers of inflammation, endothelial dysfunction, and insulin resistance [33], as well as with prevalent or incident MetS [34, 35]. There has been evidence for a reciprocal association between adiponectin, pro-inflammatory cytokines (e.g. IL-6 and TNF-a), and the acute-phase reactant CRP, each regulating the expression of the others in a feedback loop [36]. IL-6 is a central mediator of the acute-phase response because it can induce the production of CRP from the liver [37] and the content of adipose tissue in IL-6 is significantly associated with the concentration of IL-6 and CRP in the circulation [38]. Adiponectin, as an anti-inflammatory adipokine, can inhibit IL-6 production [39], suppress TNF-a levels [40], and reduce IL-6-stimulated CRP secretion [41]. On the other hand, adiponectin gene expression and secretion have been found to be downregulated by IL-6, TNF-a [28], and CRP [42]. Based on the literature discussed so far, we measured total adiponectin and IL-6 as representative inflammatory indices in order to study the association with metabolic indices and the prevalence of MetS. In the present study, adiponectin was significantly and inversely associated with waist circumference, whereas IL-6 was significantly and positively associated with BMI and waist circumference in the multivariate linear regression models. These results are in accordance with previous studies which have shown that adiponectin is inversely associated with markers of adiposity such as fat mass, BMI, waist circumference and waist-to-hip ratio in multivariate linear regression models [43–46]. Central obesity has been recognized as significant predictor of high IL-6 levels in the circulation even after adjustment for markers of lipid and glucose metabolism, blood pressure, smoking status, sex and age [47]. Abdominal obesity, assessed as waist circumference, has been found to explain most of the variance in IL-6 levels after adjustment for age, sex, smoking habits, alcohol consumption, and physical activity level [48]. In the present study, adiponectin was significantly and negatively associated with serum triglycerides in the multivariate adjusted model, and this association was attenuated, but remained significant after adjustment for serum insulin concentration. IL-6 was significantly and negatively associated with serum HDL-cholesterol in the multivariate adjusted model, but this association was not affected after further adjustment for serum insulin concentration. Previous studies have shown that adipokines are associated with insulin signalling and actions, and can therefore affect lipid metabolism. Specifically, adiponectin can have beneficial effects on lipid metabolism because of its insulin-sensitizing effects, whereas IL-6 has been involved in the dysregulation of insulin signalling and has been associated with dyslipidaemia [49–52]. Since this was a cross-sectional study, measurement of blood lipid, insulin and adipokine levels are based on a single sample, and therefore we cannot draw conclusions regarding cause-effect. In the present study, adiponectin levels were inversely associated with prevalent MetS even after adjustment for MetS components using either definition. Similarly, previous studies have found that the odds ratio for either prevalent or incident MetS is lower for the top quartile of adiponectin compared with the bottom quartile even after adjustments for age, sex, MetS components and other MetS-related factors [11, 53–56]. In the present study, IL-6 levels were not associated with MetS. Previous study has found that IL-6 is associated with MetS, and that the top tertile of IL-6 is associated with increased odds ratio for MetS compared with the bottom tertile after adjustment for age and sex. Nevertheless, further adjustment for BMI and insulin has attenuated that association [57]. Similarly, IL-6 levels have not been associated with MetS after adjustment for age, sex and MetS components [58]. In the present study, adjustments for MetS components (i.e. BMI, waist circumference, HDL-cholesterol, triglyceride and insulin concentration) were made in order to test the potential mediating effect of these factors as regards the adiponectin-MetS association. Furthermore, the indirect association between adiponectin and MetS via the MetS components was determined in order to provide for an explanation regarding the association observed between adiponectin and MetS. As far as we are aware of, this is the first study which examines the mediation effect of MetS components in the association between adiponectin and MetS found in previous studies and in the present study as well. In the present study, adiponectin had a high discriminative accuracy for MetS after adjustment for sex, age, and after further adjustment for BMI, waist circumference and insulin concentration, regardless of the definition used. Previous studies have also found that adiponectin has a high discriminative accuracy for MetS. In particular, plasma adiponectin quartiles have been found to classify correctly subjects with prevalent MetS after adjustment for sex, age, BMI, and MetS-related factors [11]. Similarly, plasma adiponectin has been found to have high discriminative accuracy for prevalent MetS, and it has been observed to be similar to the discriminative accuracy of BMI in age and sex adjusted models [53]. Finally, a previous study has found that adiponectin levels have a high discriminative accuracy for incident MetS [56]. As far as we are aware of, this is the first study that used two definitions in order to diagnose presence of MetS, and that evaluated the discriminative accuracy of adiponectin in classifying correctly subjects with MetS by both definitions. Limitations of the present study have to be taken into account. Firstly, this was a cross-sectional study which does not allow for conclusions regarding causality. Secondly, this study included middle-aged white individuals, without prevalent cardiovascular disease. Caution is thus needed in the extrapolation of the findings to other populations, i.e. younger, of other ethnicity or with cardiovascular disease. Moreover, type 2 diabetes mellitus and impaired fasting glucose were less prevalent than dyslipidemia or hypertension in the study population. Therefore, we cannot exclude blood glucose as a significant mediator between adiponectin and MetS. Finally, the adipokines measured are only a small fraction of the wide array of pro- and anti-inflammatory biochemical indices that are produced by the adipose tissue. This study has also strengths. Firstly, the measurement of biochemical indices was performed by a single certified laboratory. Secondly, the prevalence of MetS was evaluated by two definitions which are closely related and differ only by the abdominal obesity criterion. This allows for safer comparison between two existing definitions of MetS, and tests for the role of abdominal obesity in MetS. Furthermore, the adipokines measured were selected because of their proved and consistent association with adiposity, metabolism and inflammation. In addition, adiponectin and IL-6 were chosen to be measured because of their opposite functions and effects in metabolism and inflammation. Finally, this is one of the few studies which provide for a plausible explanation for the association between adiponectin and MetS. Conclusions Adiponectin was identified as a significant and consistent parameter associated with MetS, whereas IL-6 was suggested to be only a biomarker of the MetS state. Moreover, MetS components which are indices of obesity and lipid metabolism (i.e. waist circumference, BMI, HDL-cholesterol and triglycerides), as well as insulin levels mediate, but do not abolish the association between adiponectin and MetS. Abbreviations MetSMetabolic syndrome CVDCardiovascular disease IL-6Interleukin 6 IDFInternational Diabetes Federation AHA/NHLBIAmerican Heart Association/National Heart, Lung, and Blood Institute BMIBody mass index HDLHigh-density lipoprotein SBPSystolic blood pressure DBPDiastolic blood pressure IQRInterquartile range OROdds ratio 95 % CI95 % confidence interval ROCReceiver operating characteristic analysis AUCArea under the ROC curve Acknowledgements The authors would like to thank the cardiologists, general practitioners, nurses and dietitians who collaborated on this project, as well as the participants for their cooperation during the study. Finally, we acknowledge Efstathia Giannopoulou who provided materials for the study. Funding The present study was supported by the Postgraduate Program of the Department of Nutrition and Dietetics, School of Health Science and Education, Harokopio University, Athens, and was partly funded by the Hellenic Atherosclerosis Society (HAS). The funders had no role in the design and analysis of the study or in the writing of this article. Availability of data and materials The datasets analysed during the current study are available from the corresponding author on reasonable request. Authors’ contributions SA, DP and CP contributed to the study design; SA, DP and AN had the concept of the paper; AN conducted research; SA, EF, DP and AN analyzed data or performed statistical analysis; SA, DP and AN interpreted the results; AN wrote the paper; SA had the primary responsibility for final content; SA, EF and DP critically revised the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Alberti KG Eckel RH Grundy SM Zimmet PZ Cleeman JI Donato KA Fruchart JC James WP Loria CM Smith SC Jr International Diabetes Federation Task Force on Epidemiology and Prevention, National Heart, Lung, and Blood Institute, American Heart Association, World Heart Federation, International Atherosclerosis Society, International Association for the Study of Obesity 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 1640 5 10.1161/CIRCULATIONAHA.109.192644 19805654 2. Nomikos T Panagiotakos D Georgousopoulou E Metaxa V Chrysohoou C Skoumas I Antonopoulou S Tousoulis D Stefanadis C Pitsavos C ATTICA Study group Hierarchical modelling of blood lipids’ profile and 10-year (2002–2012) all cause mortality and incidence of cardiovascular disease: the ATTICA study Lipids Health Dis 2015 14 108 10.1186/s12944-015-0101-7 26370413 3. Grundy SM Cleeman JI Daniels SR Donato KA Eckel RH Franklin BA Gordon DJ Krauss RM Savage PJ Smith SC Association/National Heart, Lung, and Blood Institute Scientific Statement Diagnosis and Management of the Metabolic Syndrome: An American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement Circulation 2005 112 2735 52 10.1161/CIRCULATIONAHA.105.169404 16157765 4. Pitsavos C Panagiotakos DB Chrysohoou C Stefanadis C Study protocol. Epidemiology of cardiovascular risk factors in Greece: aims, design and baseline characteristics of the ATTICA study BMC Public Health 2003 3 32 10.1186/1471-2458-3-32 14567760 5. Athyros VG Bouloukos VI Pehlivanidis AN Papageorgiou AA Dionysopoulou SG Symeonidis AN Petridis DI Kapousouzi MI Satsoglou EA Mikhailidis DP for The MetS-Greece Collaborative Group The prevalence of the metabolic syndrome in Greece: The MetS-Greece Multicentre Study Diabetes Obes Metab 2005 7 397 405 10.1111/j.1463-1326.2004.00409.x 15955126 6. Panagiotakos DB Pitsavos C Das UN Skoumas Y Stefanadis C The implications of anthropometric, inflammatory and glycaemic control indices in the epidemiology of the metabolic syndrome given by different definitions: a classification analysis Diabetes Obes Metab 2007 9 660 8 10.1111/j.1463-1326.2006.00640.x 17697058 7. Bener A Yousafzai MT Darwish S Al-Hamaq A Nasralla EA Abdul-Ghani M Obesity index that better predict metabolic syndrome: body mass index, waist circumference, waist hip ratio, or waist height ratio J Obes 2013 269038 9 8. Phillips CM Perry IJ Does inflammation determine metabolic health status in obese and nonobese adults? Clin Endocrinol Metab 2013 98 E1610 E9 10.1210/jc.2013-2038 9. Marques-Vidal P Bastardot F von Känel R Paccaud F Preisig M Waeber G Vollenweider P Association between circulating cytokine levels, diabetes and insulin resistance in a population-based sample (CoLaus study) Clin Endocrinol (Oxf) 2013 78 232 41 10.1111/j.1365-2265.2012.04384.x 22409372 10. Weiss TW Arnesen H Seljeflot I Components of the Interleukin-6 trans-signaling system are associated with the metabolic syndrome, endothelial dysfunction and arterial stiffness Metab Clin Exp 2013 62 1008 13 10.1016/j.metabol.2013.01.019 23428306 11. Hung J McQuillan BM Thompson PL Beilby JP Circulating adiponectin levels associate with inflammatory markers, insulin resistance and metabolic syndrome independent of obesity Int J Obes (Lond) 2008 32 772 9 10.1038/sj.ijo.0803793 18253163 12. Bountziouka V Bathrellou E Giotopoulou A Katsagoni C Bonou M Vallianou N Barbetseas J Avgerinos PC Panagiotakos DB Development, repeatability and validity regarding energy and macronutrient intake of a semiquantitative food frequency questionnaire: Methodological considerations Nutr Metab Cardiovasc Dis 2012 22 659 67 10.1016/j.numecd.2010.10.015 21269818 13. Panagiotakos DB Pitsavos C Stefanadis C Dietary patterns: a Mediterranean diet score and its relation to clinical and biological markers of cardiovascular disease risk Nutr Metab Cardiovasc Dis 2006 16 559 68 10.1016/j.numecd.2005.08.006 17126772 14. Panagiotakos DB Pitsavos C Chrysohoou C Skoumas J Masoura C Toutouzas P Stefanadis C Effect of exposure to secondhand smoke on markers of inflammation: the ATTICA study Am J Med 2004 116 145 50 10.1016/j.amjmed.2003.07.019 14749157 15. Barnoya J Glantz SA Cardiovascular effects of secondhand smoke. Nearly as large as smoking Circulation 2005 111 2684 98 10.1161/CIRCULATIONAHA.104.492215 15911719 16. Papathanasiou G Georgoudis G Papandreou M Spyropoulos P Georgakopoulos D Kalfakakou V Evangelou A Reliability measures of the short International Physical Activity Questionnaire (IPAQ) in Greek young adults Hellenic J Cardiol 2009 50 283 94 19622498 17. Friedewald WT Levy RI Fredrickson DS Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge Clin Chem 1972 18 499 502 4337382 18. Preacher KJ Hayes AF SPSS and SAS procedures for estimating indirect effects in simple mediation models Behav Res Methods Instrum Comput 2004 36 717 31 10.3758/BF03206553 15641418 19. Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC, Müller M. Package ‘pROC’: Display and Analyze ROC Curves (Version 1.8), 2015. Available at: https://cran.r-project.org/web/packages/pROC/pROC.pdf. Accessed 24 Mar 2016. 20. Maury E Brichard SM Adipokine dysregulation, adipose tissue inflammation and metabolic syndrome Mol Cell Endocrinol 2010 314 1 16 10.1016/j.mce.2009.07.031 19682539 21. Scicchitano P Cameli M Maiello M Modesti PA Muiesan ML Novo S Palmiero P Saba PS Pedrinelli R Ciccone MM on behalf of the “Gruppo di Studio Ipertensione, Prevenzione e Riabilitazione”, Società Italiana di Cardiologia Nutraceuticals and dyslipidaemia: Beyond the common therapeutics J Funct Foods 2014 6 11 32 10.1016/j.jff.2013.12.006 22. You T Yang R Lyles MF Gong D Nicklas BJ Abdominal adipose tissue cytokine gene expression: relationship to obesity and metabolic risk factors Am J Physiol Endocrinol Metab 2005 288 E741 E7 10.1152/ajpendo.00419.2004 15562250 23. You T Nicklas BJ Ding J Penninx BWJH Goodpaster BH Bauer DC Tylavsky FA Harris TB Kritchevsky SB for the Health, Aging and Body Composition Study The metabolic syndrome is associated with circulating adipokines in older adults across a wide range of adiposity J Gerontol A Biol Sci Med Sci 2008 63 414 9 10.1093/gerona/63.4.414 18426966 24. Mohamed-Ali V Goodrick S Rawesh A Katz DR Miles JM Yudkin JS Klein S Coppack SW Subcutaneous adipose tissue releases interleukin-6, but not tumor necrosis factor-alpha, in vivo J Clin Endocrinol Metab 1997 82 4196 200 9398739 25. Lago F Dieguez C Gómez-Reino J Gualillo O The emerging role of adipokines as mediators of inflammation and immune responses Cytokine Growth Factor Rev 2007 18 313 25 10.1016/j.cytogfr.2007.04.007 17507280 26. Halleux CM Takahashi M Delporte ML Detry R Funahashi T Matsuzawa Y Brichard SM Secretion of adiponectin and regulation of apM1 gene expression in human visceral adipose tissue Biochem Biophys Res Commun 2001 288 1102 7 10.1006/bbrc.2001.5904 11700024 27. Ruan H Dong LQ Adiponectin signaling and function in insulin target tissues J Mol Cell Biol 2016 8 101 9 10.1093/jmcb/mjw014 26993044 28. Caselli C Role of adiponectin system in insulin resistance Mol Genet Metab 2014 113 155 60 10.1016/j.ymgme.2014.09.003 25242063 29. Tsao TS Assembly of adiponectin oligomers Rev Endocr Metab Disord 2014 15 125 36 10.1007/s11154-013-9256-6 23990400 30. Ciccone MM, Faienza MF, Altomare M, Nacci C, Montagnani M, Federica V, Cortese F, Gesualdo M, Zito A, Mancarella R, Leogrande D, Viola D, Scicchitano P, Giordano P. Endothelial and Metabolic Function Interactions in Overweight/ Obese Children: The Role of High-Molecular Weight Adiponectin. J Atheroscler Thromb. 2016. [Epub ahead of print]. 31. Kizer JR Arnold AM Benkeser D Ix JH Djousse L Zieman SJ Barzilay JI Tracy RP Mantzoros CS Siscovick DS Mukamal KJ Total and high-molecular-weight adiponectin and risk of incident diabetes in older people Diabetes Care 2012 35 415 23 10.2337/dc11-1519 22148099 32. Saito I Yamagishi K Chei CL Cui R Ohira T Kitamura A Kiyama M Imano H Okada T Kato T Hitsumoto S Ishikawa Y Tanigawa T Iso H Total and high molecular weight adiponectin levels and risk of cardiovascular disease in individuals with high blood glucose levels Atherosclerosis 2013 229 222 7 10.1016/j.atherosclerosis.2013.04.014 23676254 33. Fargnoli JL Sun Q Olenczuk D Qi L Zhu Y Hu FB Mantzoros CS Resistin is associated with biomarkers of inflammation while total and HMW adiponectin are associated with biomarkers of inflammation, insulin resistance, and endothelial function Eur J Endocrinol 2010 162 281 8 10.1530/EJE-09-0555 19920090 34. Nakashima R Yamane K Kamei N Nakanishi S Kohno N Low serum levels of total and high-molecular-weight adiponectin predict the development of metabolic syndrome in Japanese-Americans J Endocrinol Invest 2011 34 615 9 21164278 35. von Frankenberg AD do Nascimento FV Gatelli LE Nedel BL Garcia SP de Oliveira C Saddi-Rosa P Reis AF Canani LH Gerchman F Major components of metabolic syndrome and adiponectin levels: a cross-sectional study Diabetol Metab Syndr 2014 6 26 10.1186/1758-5996-6-26 24568287 36. Knights AJ Funnell A Pearson R Crossley M Bell-Anderson K Adipokines and insulin action. A sensitive issue Adipocyte 2014 3 88 96 10.4161/adip.27552 24719781 37. Ridker PM Clinical application of C-reactive protein for cardiovascular disease detection and prevention Circulation 2003 107 363 9 10.1161/01.CIR.0000053730.47739.3C 12551853 38. Maachi M Piéroni L Bruckert E Jardel C Fellahi S Hainque B Capeau J Bastard JP Systemic low grade inflammation is related to both circulating and adipose tissue TNF-a, leptin and IL-6 levels in obese women Int J Obes 2004 28 993 7 10.1038/sj.ijo.0802718 39. Eder K Baffy N Falus A Fulop AK The major inflammatory mediator interleukin-6 and obesity Inflamm Res 2009 58 727 36 10.1007/s00011-009-0060-4 19543691 40. Shehzad A Iqbal W Shehzad O Lee YS Adiponectin: Regulation of its production and its role in human diseases Hormones (Athens) 2012 11 8 20 22450341 41. Sun H Zhang Y Gao P Li Q Sun Y Zhang J Xu C Adiponectin reduces C-reactive protein expression and downregulates STAT3 phosphorylation induced by IL-6 in HepG2 cells Mol Cell Biochem 2011 347 183 9 10.1007/s11010-010-0627-y 20978825 42. Yuan G Jue Jia J Di L Zhou L Dong S Ye J Wang D Yang L Wang J Li L Yang Y Mao C Chen M Effects of C-reactive protein on adipokines genes expression in 3 T3-L1 adipocytes Biochem Biophys Res Commun 2012 424 462 8 10.1016/j.bbrc.2012.06.133 22771801 43. Vilarrasa N Vendrell J Maravall J Broch M Estepa A Megia A Soler J Simón I Richart C Gómez JM Distribution and determinants of adiponectin, resistin and ghrelin in a randomly selected healthy population Clin Endocrinol (Oxf) 2005 63 329 35 10.1111/j.1365-2265.2005.02346.x 16117822 44. Gannagé-Yared MH Khalife S Semaan M Fares F Jambart S Halaby G Serum adiponectin and leptin levels in relation to the metabolic syndrome, androgenic profile and somatotropic axis in healthy non-diabetic elderly men Eur J Endocrinol 2006 155 167 76 10.1530/eje.1.02175 16793964 45. Maahs DM Ogden LG Snell-Bergeon JK Kinney GL Wadwa RP Hokanson JE Dabelea D Kretowski A Eckel RH Rewers M Determinants of serum adiponectin in persons with and without type 1 diabetes Am J Epidemiol 2007 166 731 40 10.1093/aje/kwm125 17591595 46. Salas-Salvadó J Granada M Bulló M Corominas A Casas P Foz M Plasma adiponectin distribution in a Mediterranean population and its association with cardiovascular risk factors and metabolic syndrome Metabolism 2007 56 1486 92 10.1016/j.metabol.2007.06.014 17950098 47. Thewissen MM Damoiseaux JG Duijvestijn AM van Greevenbroek MM van der Kallen CJ Feskens EJ Blaak EE Schalkwijk CG Stehouwer CD Tervaert JWC Abdominal fat mass is associated with adaptive immune activation: the CODAM study Obesity (Silver Spring) 2011 19 1690 8 10.1038/oby.2010.337 21253003 48. Thorand B Baumert J Döring A Herder C Kolb H Rathmann W Giani G Koenig W for the KORA Group Sex differences in the relation of body composition to markers of inflammation Atherosclerosis 2006 184 216 24 10.1016/j.atherosclerosis.2005.04.011 15993885 49. Jung UJ Choi MS Obesity and its metabolic complications: the role of adipokines and the relationship between obesity, inflammation, insulin resistance, dyslipidemia and nonalcoholic fatty liver disease Int J Mol Sci 2014 15 6184 223 10.3390/ijms15046184 24733068 50. Rabe K Lehrke M Parhofer KG Broedl UC Adipokines and insulin resistance Mol Med 2008 14 741 51 10.2119/2008-00058.Rabe 19009016 51. Hoene M Weigert C Appetite regulatory peptides. The role of interleukin-6 in insulin resistance, body fat distribution and energy balance Obes Rev 2008 9 20 9 17956545 52. Fisman EZ Tenenbaum A Adiponectin: a manifold therapeutic target for metabolic syndrome, diabetes, and coronary disease? Cardiovasc Diabetol 2014 13 103 10.1186/1475-2840-13-103 24957699 53. Henneman P Janssens ACJW Zillikens MC Frolich M Frants RR Oostra BA van Duijn CM van Dijk KW Menopause impacts the relation of plasma adiponectin levels with the metabolic syndrome J Intern Med 2010 267 402 9 10.1111/j.1365-2796.2009.02162.x 19912464 54. Koh SB Yoon J Kim JY Yoo BS Lee SH Park JK Choe KH Relationships between serum adiponectin with metabolic syndrome and components of metabolic syndrome in non-diabetic Koreans: ARIRANG study Yonsei Med J 2011 52 234 41 10.3349/ymj.2011.52.2.234 21319340 55. Juonala M Saarikoski LA Viikari J Oikonen M Lehtimäki T Lyytikäinen LP Huupponen R Magnussen CG Koskinen J Laitinen T A longitudinal analysis on associations of adiponectin levels with metabolic syndrome and carotid artery intima-media thickness. The cardiovascular risk in young Finns study Atherosclerosis 2011 217 234 9 10.1016/j.atherosclerosis.2011.03.016 21481397 56. Kim JY Ahn SV Yoon JH Koh SB Yoon J Yoo BS Lee SH Park JK Choe KH Guallar E Prospective study of serum adiponectin and incident metabolic syndrome: the ARIRANG study Diabetes Care 2013 36 1547 53 10.2337/dc12-0223 23275369 57. Hung J McQuillan BM Chapman C Thompson PL Beilby JP Elevated interleukin-18 levels are associated with the metabolic syndrome independent of obesity and insulin resistance Arterioscler Thromb Vasc Biol 2005 25 1268 73 10.1161/01.ATV.0000163843.70369.12 15790931 58. Choi KM Ryu OH Lee KW Kim HY Seo JA Kim SG Kim NH Choi DS Baik SH Serum adiponectin, interleukin-10 levels and inflammatory markers in the metabolic syndrome Diabetes Res Clin Pract 2007 75 235 40 10.1016/j.diabres.2006.06.019 16872712

PMC005xxxxxx/PMC5002190.txt

==== Front BMC CancerBMC CancerBMC Cancer1471-2407BioMed Central London 273210.1186/s12885-016-2732-0Research ArticleWorse outcome in breast cancer with higher tumor-infiltrating FOXP3+ Tregs : a systematic review and meta-analysis Shou Jiafeng 3100102375@zju.edu.cn 1Zhang Zhigang zzg2011@zju.edu.cn 1Lai Yucheng 704960737@qq.com 1Chen Zhigang pridechen@126.com 12Huang Jian drhuangjian@zju.edu.cn 121 Cancer Institute (Key Laboratory of Cancer Prevention and Intervention, National Ministry of Education, Provincial Key Laboratory of Molecular Biology in MedicalSciences), The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, 31009 China 2 Department of Oncology, Second Affiliated Hospital, ZhejiangUniversity School of Medicine, Hangzhou, 310009 China 26 8 2016 26 8 2016 2016 16 1 68717 6 2015 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 Forkhead box P3(FOXP3) is known as the optimum maker for regulatory T cells(Tregs), which are conventionally thought to induce immune tolerance to disturb the antitumor immunity. However, the research on the prognostic significance of tumor-infiltrating FOXP3+ Tregs in breast cancer is still limited and the results are controversial. Methods We searched for studies in PubMed, EMBASE and Web of Science prior to January 2015. The correlation between FOXP3+ tumor-infiltrating lymphocytes(TILs) and breast cancer prognosis was analyzed. The meta-analysis was performed using STATA 11.0. Pooled hazard ratios (HRs) with 95 % confidence intervals (CIs) were used to estimate the degree of the association between FOXP3+ TILs and prognosis of breast cancers, while relative ratios (RRs) were used to evaluate the relationship between FOXP3+ TILs and clinicopathological features of breast cancers. Result A total of 15 studies comprising 8666 breast cancer patients met the inclusion criteria. Our results showed that higher FOXP3+ TILs level was significantly associated with poor prognosis in terms of overall survival (OS) (pooled HR:1.60, 95 % CI:1.06–2.42; P < 0.05). We found that breast cancer with higher FOXP3+ TILs level was positively correlated with c-erbB-2 positive status (pooled RR:1.52, 95 % CI:1.32–1.75; P < 0.05), lymph node positive status(pooled RR:1.17, 95 % CI:1.04–1.32; P < 0.05) while there was a negative association with ER positive status(pooled RR:0.65, 95 % CI:0.56–0.76; P < 0.05) and PR positive status(pooled RR:0.66, 95 % CI:0.51–0.87; P < 0.05). Conclusion The present results of meta-analysis showed that higher FOXP3+ TILs level in patients with breast cancer led to poor overall survival (OS) and was significantly associated with c-erbB-2 status, lymph node status, ER status and PR status. FOXP3+ TILs level is a promising prognostic factor in breast cancer. Electronic supplementary material The online version of this article (doi:10.1186/s12885-016-2732-0) contains supplementary material, which is available to authorized users. Keywords Breast cancerFOXP3TILPrognosisissue-copyright-statement© The Author(s) 2016 ==== Body Background Breast cancer is one of the most common malignancies and expected to account for 29 % of all newly diagnosed cancers and 14 % of all cancer deaths in women worldwide. Though death rate for female breast cancer has decreased by 35 % from peak rate, it is still the leading cause of cancer death in women from aged 20 to 59 years [1]. In most cancers, including breast cancer [2–4], infiltrating of FOXP3+ regulatory TILs have been reported to be associated with worse clinical outcome. However, MJM Gooden et al. reported that FOXP3+ regulatory TILs were not linked to the overall survival in cancers [2]. FOXP3 is a forkhead box transcription factor containing a DNA-binding domain that suppresses the expression of target genes [3]. The regulatory T lymphocytes (Tregs), subpopulation of CD4(+) T lymphocytes, is an important obstacle in antitumor immunity by suppression of tumor antigen reactive T lymphocytes [4–6]. The transcription factor FOXP3, known as the most specific marker of Tregs [7–9], plays a crucial role in the development and function of Tregs,. FOXP3 is constitutively expressed in the nucleus of human Tregs [10, 11]. Several studies showed that higher FOXP3+ Tregs indicated poor prognosis [12–14] in patients with breast cancer while some revealed no direct association between them [15]. In short, studies to confirm the clinical significance of FOXP3+ TILs in breast cancer are still insufficient and the prognostic value still lacks assessment, especially in different molecule types of breast cancer. Therefore, we find it necessary to further assess the association between tumor-infiltrating FOXP3+ Tregs and prognosis of breast cancer by conducting a meta-analysis with a large sample size (N = 8666). The relationship between FOXP3+ TILs and several clinicopathological features of breast cancer was also evaluated. Methods Literature search strategy We searched PubMed, EMBASE and Web of Science for relevant studies before January 2015, using the terms: “FOXP3 lymphocytes” or “FOXP3 regulatory T cell” or “FOXP3 TIL” or “FOXP3 tumor-infiltrating lymphocytes” and “breast cancer”. We searched the references of all retrieved publications and conference proceedings again to identify additional relevant studies. Selection criteria The studies included in our meta-analysis should meet following criteria: [1] the study must be conducted on the human beings; [2] the study must assess the association between FOXP3+ TILs level and the prognosis of breast cancer; [3] The count of FOXP3+ TIL include either tumor bed or tumor peripheral lymphocytes [4] the study must contain sufficient published data to determine an estimate of hazard ratio(HR) and a 95 % confidence interval(95 % CI); [5] original research article must be published in English. Data extraction and quality assessment Data were extracted from the eligible studies by two investigators independently. Discrepancies were resolved by consensus. The following information was abstracted from all included publications: author, year of publication, country of study, tumor type, median follow-up time, cut-offs for positive expression, number of TILs-low and TILs-high patients, number of patients with different clinicopathological features, outcome of analysis (including HRs and 95 % CIs). Quality assessment was conducted for each of the included studies using the Newcastle-Ottawa quality assessment scale [16]. The score assessed eight items on methodology that were categorized into three dimensions, including selection, comparability and outcome. Interpretation of the scale is performed by awarding “stars”, for high-quality element. The studies with 6 scores or more were regarded as high-quality ones in the scale. Statistical analysis Survival outcome data were synthesized using the HR and its 95 % CI to assess the impact of higher FOXP3+ TILs level on the overall survival (OS) and relapse free survival (RFS) of patients with breast cancer. Several included studies have provided HRs and their 95 % CIs, which we extracted directly from the papers. Otherwise, we calculated the HRs and 95 % CIs from the Kaplan-Meier survival curve using the software Engauge Digitizer version 4.1 (http://digitizer.sourceforge.net/) [17, 18]. The relative ratio (RR) was used to evaluate the association between FOXP3+ TILs and clinicopathological features of breast cancer (including tumor category(T), lymph node category(N), c-erbB-2, ER and PR status). Heterogeneity across studies was evaluated using a Chi-square-based Q statistical test [19], and the I2 value was used to quantify the heterogeneity [20]. The random-effect model (the DerSimonian and Lairdmethod) was used for meta-analysis [21]. Publication bias was assessed using Begg’s test [22]. In addition, sensitivity analysis was performed to examine the stability of the pooled results. The statistical analyses were conducted using STATA11.0. All P values were two-sided, and P < 0.05 was considered to be statistically significant. Results Study selection and characteristics As shown in Fig. 1, a total of 284 articles were initially retrieved from the above databases using the search strategy described above. From candidate publications following exclusions, our search found 15 studies satisfying the inclusion criteria. The main characteristics of included studies were presented in Additional file 1: Table S1. An additional file shows this in more detail (see Additional file 1). These studies were published from 2006 to 2014 and total 8666 patients with breast cancer were enrolled. Individual sample size ranged from 72 to 3992 (mean 578). These studies were conducted in nine counties (the United Kingdom, the United States, Turkey, Korea, China, Canada, Japan, Australia and France). Overall, the included studies in meta-analysis were of high quality, six studies scored 8, six studies scored 7 and three studies scored 6.Fig. 1 Flow diagram showing the study selection procedure FOXP3+ TILs level and prognosis of breast cancers We pooled overall survival (OS) and relapse free survival (RFS) to assess the impact of FOXP3+ TILs level on the prognosis of breast cancers. 8 studies evaluated the relationship between FOXP3+ TILs and OS [12, 13, 23–28]. The pooled HR is 1.60(95 % CI:1.06–2.42; P < 0.05) (Fig. 2). It indicated that higher FOXP3+ TILs level was statistically related to a poor OS rate. 5 studies assessed the association between FOXP3+ TILs and RFS [14, 23, 28–30]. The pooled HR is 1.06(95 % CI: 0.64 ~ 1.74; P > 0.05) (Fig. 3) which indicated that FOXP3+ TILs level was not correlated with RFS of breast cancers.Fig. 2 Forest plot of the hazard ratio (HR) for the association of FOXP3+ level with overall survival (OS) in breast cancer patients Fig. 3 Forest plot of the hazard ratio (HR) for the association of FOXP3+ level with relapse free survival (RFS) in breast cancer patients Correlation of FOXP3+ TILs level with clinicopathological parameters The meta-analysis also estimated the association between FOXP3+ TILs level and certain clinicopathological parameters. As shown in Fig. 4c and e, breast cancers with higher FOXP3+ TIL level were positively correlated with c-erbB-2 status(pooled RR:1.52, 95 % CI:1.32–1.75; P < 0.05), lymph node status(pooled RR:1.17, 95 % CI:1.04–1.32; P < 0.05). Compared with ER-negative breast cancers, ER-positive breast cancers showed lower FOXP3+ TILs level (pooled RR:0.65, 95 % CI:0.56–0.76; P < 0.05) (Fig. 4a). Similar phenomenon was observed in PR-positive breast cancers(pooled RR:0.66, 95 % CI:0.51–0.87; P < 0.05) (Fig. 4b). This indicated that in c-erbB-2 positive or lymph node positive breast cancers, there was higher FOXP3+ TILs level. On the other hand, ER positive or PR positive breast cancers were accompanied with lower FOXP3+ TILs level. In addition, we found no association between higher FOXP3+ TIL level and tumor category of breast cancers (pooled RR:1.08, 95 % CI:0.98–1.19; P > 0.05) (Fig. 4d).Fig. 4 Forest plot assessing the FOXP3+ TILs and clinicopathological features such as ER status (a), PR status (b), c-erbB-2 status (c), tumor category (d), lymph node category (e) Publication bias Begg’s test indicated negligible publication bias after assessing the funnel plot for the studies included in our meta-analysis (Figs. 5 and 6).Fig. 5 Begg’s test results of the OS (a) and RFS (b) Fig. 6 Begg’s test assessing the FOXP3+ TILs and clinicopathological features such as ER status (a), PR status (b), c-erbB-2 status (c), tumor category (d), lymph node category (e) Sensitivity analysis In the sensitivity analysis, the influence of each study on the pooled HR of OS and RFS was assessed by repeating the meta-analysis while sequentially omitting each study. The results showed that the association did not change significantly after removing any study. An additional shows this in more detail (see Additional file 2). Discussion FOXP3+ Tregs, part of tumor-infiltrating lymphocytes (TILs), play a critical role in immune tolerance and deficiency of anti-tumor immunity [4–6]. Several studies revealed that FOXP3+ TILs level had a negative impact on the prognosis of breast cancer [14, 31, 32]. However, the exact prognostic significance was still unclear and the information was limited. In our meta-analysis, 15 eligible articles were included which evaluated the association between FOXP3+ TILs level and prognosis of breast cancer. Ezzeldin M et al. and X. Yu et al. have reported that TILs were associated with prognosis of breast cancer in their meta-analysis [33, 34]. However, TILs in our meta-analysis referred to FOXP3+ TILs specifically. Thus, our meta-analysis seems more applicable in clinical practice. The pooled results showed that a higher density of FOXP3+ lymphocytes in tumor tissue was a promising prognostic factor for OS of breast cancers. However, the FOXP3+ TILs level was proved not having significantly higher risk of relapse. These results are potentially important for prognostic and treatment reasons. FOXP3 is the most specific marker of CD4 + CD25+ Tregs and appears to be critical for the development and function of Tregs derived from thymus [8, 35, 36]. Tumor-infiltrating lymphocytes(TILs) in our meta-analysis refer to the lymphocytes either infiltrating in tumor bed or tumor periphery. FOXP3+ Tregs was an important group of TILs and Udaya K et al. has reported that FOXP3 + Tregs prevalence was 20.2 % in TILs of breast adenocarcinoma [37]. FOXP3+ Tregs suppressed the function of effector T cells to destroy maintenance of immune balance which resulted in the escape of tumor immunological surveillance [8, 38]. FOXP3+ Tregs secreted TGF-β, which indicated that the suppression of anti-tumor immunity of FOXP3+ Tregs may be cytokine-dependent [39]. FOXP3+ Tregs also secreted IL-10 to suppress Th1/2 cell proliferation via inhibition of IL-2 and down-regulating MHC class II on monocytes [40, 41]. In addition, FOXP3+ Tregs inhibit the proliferation and IFN-y secretion by activated CD8+ T cells and CD4+ helper T cells [37]. FOXP3+ Tregs express CD39/CD73 that generates adenosine, which can down-modulate immune function [42, 43]. These may account for the poor overall survival in breast cancers with higher FOXP3+ TILs level. Nevertheless, other immune cells might have impact on the prognosis of breast cancer. Natural killer (NK) cells contributed to a pathological complete response (pCR) in breast cancers following neoadjuvant chemotherapy [44]. Lymphocytes infiltrating human breast cancers showed lower levels of NK-cell activity [45]. Tiainen S et al. reported that higher level of infiltrated macrophage, especially M2- like, correlated with poor outcome in breast cancer [46]. Thus, a more effective model needs to be established to predict the prognosis of breast cancer. No association was observed between higher FOXP3+ TILs level and RFS of breast cancers in our meta-analysis. In addition, we investigated the association between FOXP3+ TILs and some clinicopathological parameters of breast cancer. Significant positive correlations were found between high FOXP3+ TILs with c-erbB-2 status and lymph node status while negative correlations with ER and PR status. The result implied that ER+ breast cancer was accompanied with lower FOXP3+ TILs. N R West et al. [30] demonstrated a similar phenomenon with our result. Another point worth attention was that in c-erbB-2+ breast cancer, there was a significantly higher level of FOXP3+ TILs, also reported by Sahar M. A et al. [15]. However, there are still some limitations in our meta-analysis. Firstly, the included studies were restricted within English articles and completed in 9 countries, which may result in potential publication bias while Begg’s test showed no evidence. Secondly, heterogeneity was found in the meta-analysis, which may be due to different cutoff values of FOXP3+ TILs in the included studies. Though the detecting technology of FOXP3+ TILs in included studies was all immunohistochemistry (IHC), the different antibody type and antibody concentration may contribute to the heterogeneity of included studies. Thirdly, the HRs and 95 % CIs extracted from survival curves may be less reliable than those directly obtained from the articles. Conclusion Despite the limitation listed above, our meta-analysis based on currently published articles indicted that higher FOXP3+ TILs level was inclined to poor OS while no significant association was observed with RFS. Accordingly, FOXP3+ TILs can be considered as a promising therapeutic target. Additional files Additional file 1: Table S1. Characteristics of the included studies. (DOCX 39 kb) Additional file 2: Sensitivity analysis of forest plots included in our meta-analysis. (PDF 43 kb) Abbreviations FOXP3Forkhead box P3 TregsRegulatory T cells TILsTumor-infiltrating lymphocytes HRHazard ratio CIConfidence interval RRRelative ratio OSOverall survival RFSRelapse free survival TTumor category NLymph node category NKNatural killer pCRPathological complete response IHCImmunohistochemistry The authors would like to thank the members of the research group for useful discussions. Funding This study was supported by a grant from the National Natural Science Foundation of China (No. 81301654) and the Foundation of Education Department of Zhejiang Province (No. Y201328265). Availability of data and materials The datasets supporting the conclusions of this article are included within the article. Authors’ contributions JS participated in the design of the study and performed the statistical analysis. ZZ and ZC and YL carried out the data extraction. JH conceived of the study, and participated in its design and coordination and helped to draft 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 because this manuscript does not contain any individual persons data. ==== Refs References 1. Siegel RL Miller KD Jemal A Cancer statistics, 2015 CA Cancer J Clin 2015 65 1 5 29 10.3322/caac.21254 25559415 2. Gooden MJ de Bock GH Leffers N Daemen T Nijman HW The prognostic influence of tumour-infiltrating lymphocytes in cancer: a systematic review with meta-analysis Br J Cancer 2011 105 1 93 103 10.1038/bjc.2011.189 21629244 3. Lopes JE Torgerson TR Schubert LA Anover SD Ocheltree EL Ochs HD Analysis of FOXP3 reveals multiple domains required for its function as a transcriptional repressor J Immunol 2006 177 5 3133 42 10.4049/jimmunol.177.5.3133 16920951 4. DiPaolo RJ Glass DD Bijwaard KE Shevach EM CD4 + CD25+ T cells prevent the development of organ-specific autoimmune disease by inhibiting the differentiation of autoreactive effector T cells J Immunol 2005 175 11 7135 42 10.4049/jimmunol.175.11.7135 16301616 5. Josefowicz SZ Rudensky A Control of regulatory T cell lineage commitment and maintenance Immunity 2009 30 5 616 25 10.1016/j.immuni.2009.04.009 19464984 6. Kryczek I Liu R Wang G Wu K Shu X Szeliga W FOXP3 defines regulatory T cells in human tumor and autoimmune disease Cancer Res 2009 69 9 3995 4000 10.1158/0008-5472.CAN-08-3804 19383912 7. Hori S Nomura T Sakaguchi S Control of regulatory T cell development by the transcription factor Foxp3 Science 2003 299 5609 1057 61 10.1126/science.1079490 12522256 8. Fontenot JD Gavin MA Rudensky AY Foxp3 programs the development and function of CD4 + CD25+ regulatory T cells Nat Immunol 2003 4 4 330 6 10.1038/ni904 12612578 9. Beyer M Schultze JL Regulatory T cells in cancer Blood 2006 108 3 804 11 10.1182/blood-2006-02-002774 16861339 10. Martin F Ladoire S Mignot G Apetoh L Ghiringhelli F Human FOXP3 and cancer Oncogene 2010 29 29 4121 9 10.1038/onc.2010.174 20498631 11. Curiel TJ Coukos G Zou L Alvarez X Cheng P Mottram P Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival Nat Med 2004 10 9 942 9 10.1038/nm1093 15322536 12. Takenaka M Seki N Toh U Hattori S Kawahara A Yamaguchi T FOXP3 expression in tumor cells and tumor-infiltrating lymphocytes is associated with breast cancer prognosis Mol Clin Oncol 2013 1 4 625 32 24649219 13. Sun S Fei X Mao Y Wang X Garfield DH Huang O PD-1(+) immune cell infiltration inversely correlates with survival of operable breast cancer patients Cancer Immunol Immunother 2014 63 4 395 406 10.1007/s00262-014-1519-x 24514954 14. Maeda N Yoshimura K Yamamoto S Kuramasu A Inoue M Suzuki N Expression of B7–H3, a potential factor of tumor immune evasion in combination with the number of regulatory T cells, affects against recurrence-free survival in breast cancer patients Ann Surg Oncol 2014 21 Suppl 4 546 54 10.1245/s10434-014-3564-2 15. Mahmoud SMA Paish EC Powe DG Macmillan RD Lee AHS Ellis IO An evaluation of the clinical significance of FOXP3(+) infiltrating cells in human breast cancer Breast Cancer Res Treat 2011 127 1 99 108 10.1007/s10549-010-0987-8 20556505 16. Stang A Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses Eur J Epidemiol 2010 25 9 603 5 10.1007/s10654-010-9491-z 20652370 17. Parmar MKTV Stewart L Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints Stat Med 1998 17 2815 34 10.1002/(SICI)1097-0258(19981230)17:24<2815::AID-SIM110>3.0.CO;2-8 9921604 18. Tierney JF Stewart LA Ghersi D Burdett S Sydes MR Practical methods for incorporating summary time-to-event data into meta-analysis Trials 2007 8 16 10.1186/1745-6215-8-16 17555582 19. Handoll HH Systematic reviews on rehabilitation interventions Arch Phys Med Rehabil 2006 87 6 875 10.1016/j.apmr.2006.04.006 16731227 20. Higgins JP Thompson SG Deeks JJ Altman DG Measuring inconsistency in meta-analyses BMJ 2003 327 7414 557 60 10.1136/bmj.327.7414.557 12958120 21. 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 22. Begg CB Mazumdar M Operating characteristics of a rank correlation test for publication bias Biometrics 1994 50 4 1088 101 10.2307/2533446 7786990 23. Bates GJ Fox SB Han C Leek RD Garcia JF Harris AL Quantification of regulatory T cells enables the identification of high-risk breast cancer patients and those at risk of late relapse J Clin Oncol 2006 24 34 5373 80 10.1200/JCO.2006.05.9584 17135638 24 Kim S Lee A Lim W Park S Cho MS Koo H Zonal difference and prognostic significance of foxp3 regulatory T cell infiltration in breast cancer J Breast Cancer 2014 17 1 8 17 10.4048/jbc.2014.17.1.8 24744792 25 Lee S Cho EY Park YH Ahn JS Im YH Prognostic impact of FOXP3 expression in triple-negative breast cancer Acta Oncol 2013 52 1 73 81 10.3109/0284186X.2012.731520 23075422 26 Liu F Lang R Zhao J Zhang X Pringle GA Fan Y CD8(+) cytotoxic T cell and FOXP3(+) regulatory T cell infiltration in relation to breast cancer survival and molecular subtypes Breast Cancer Res Treat 2011 130 2 645 55 10.1007/s10549-011-1647-3 21717105 27 Gobert M Treilleux I Bendriss-Vermare N Bachelot T Goddard-Leon S Arfi V Regulatory T cells recruited through CCL22/CCR4 are selectively activated in lymphoid infiltrates surrounding primary breast tumors and lead to an adverse clinical outcome Cancer Res 2009 69 5 2000 9 10.1158/0008-5472.CAN-08-2360 19244125 28 Demir L Yigit S Ellidokuz H Erten C Somali I Kucukzeybek Y Predictive and prognostic factors in locally advanced breast cancer: effect of intratumoral FOXP3+ Tregs Clin Exp Metastasis 2013 30 8 1047 62 10.1007/s10585-013-9602-9 23836289 29 Liu S Foulkes WD Leung S Gao D Lau S Kos Z Prognostic significance of FOXP3+ tumor infiltrating lymphocytes in breast cancer depends on estrogen receptor and human epidermal growth factor receptor-2 expression status and concurrent cytotoxic T-cell infiltration Breast Cancer Res 2014 16 5 432 10.1186/s13058-014-0432-8 25193543 30 West NR Kost SE Martin SD Milne K Deleeuw RJ Nelson BH Tumour-infiltrating FOXP3(+) lymphocytes are associated with cytotoxic immune responses and good clinical outcome in oestrogen receptor-negative breast cancer Br J Cancer 2013 108 1 155 62 10.1038/bjc.2012.524 23169287 31 Bohling SD Allison KH Immunosuppressive regulatory T cells are associated with aggressive breast cancer phenotypes: a potential therapeutic target Mod Pathol 2008 21 12 1527 32 10.1038/modpathol.2008.160 18820666 32 Yan M Jene N Byrne D Millar EK O’Toole SA McNeil CM Recruitment of regulatory T cells is correlated with hypoxia-induced CXCR4 expression, and is associated with poor prognosis in basal-like breast cancers Breast Cancer Res 2011 13 2 R47 10.1186/bcr2869 21521526 33 Ibrahim EM Al-Foheidi ME Al-Mansour MM Kazkaz GA The prognostic value of tumor-infiltrating lymphocytes in triple-negative breast cancer: a meta-analysis Breast Cancer Res Treat 2014 148 3 467 76 10.1007/s10549-014-3185-2 25361613 34 Yu X Zhang Z Wang Z Wu P Qiu F Huang J Prognostic and predictive value of tumor-infiltrating lymphocytes in breast cancer: a systematic review and meta-analysis Clin Transl Oncol. 2016 18 497 506 10.1007/s12094-015-1391-y 26459255 35 Romero P Dunbar PR Valmori D Pittet M Ogg GS Rimoldi D Chen DL JL Cerottini JC Cerundolo V Ex Vivo Staining of Metastatic Lymph Nodes by Class I Major Histocompatibility Complex Tetramers Reveals High Numbers of Antigen-experienced Tumor-specific Cytolytic T Lymphocytes J Exp Med 1998 188 1641 10.1084/jem.188.9.1641 9802976 36 Khattri R Cox T Yasayko SA Ramsdell F An essential role for Scurfin in CD4 + CD25+ T regulatory cells Nat Immunol 2003 4 4 337 42 10.1038/ni909 12612581 37 Liyanage UK Moore TT Joo HG Tanaka Y Herrmann V Doherty G Prevalence of Regulatory T Cells Is Increased in Peripheral Blood and Tumor Microenvironment of Patients with Pancreas or Breast Adenocarcinoma J Immunol 2002 169 5 2756 61 10.4049/jimmunol.169.5.2756 12193750 38 Zou W Regulatory T, cells, tumour immunity and immunotherapy Nat Rev Immunol 2006 6 4 295 307 10.1038/nri1806 16557261 39 Viguier M Lemaitre F Verola O Cho MS Gorochov G Dubertret L Foxp3 Expressing CD4 + CD25high Regulatory T Cells Are Overrepresented in Human Metastatic Melanoma Lymph Nodes and Inhibit the Function of Infiltrating T Cells J Immunol 2004 173 2 1444 53 10.4049/jimmunol.173.2.1444 15240741 40 Ostmann A Paust HJ Panzer U Wegscheid C Kapffer S Huber S Regulatory T cell-derived IL-10 ameliorates crescentic GN J Am Soc Nephrol 2013 24 6 930 42 10.1681/ASN.2012070684 23641052 41 Vieira PL Christensen JR Minaee S O’Neill EJ Barrat FJ Boonstra A IL-10-secreting regulatory T cells do not express Foxp3 but have comparable regulatory function to naturally occurring CD4+CD25+ regulatory T cells J Immunol 2004 172 10 5986 93 10.4049/jimmunol.172.10.5986 15128781 42 Wang YM McRae JL Robson SC Cowan PJ Zhang GY Hu M Regulatory T cells participate in CD39-mediated protection from renal injury Eur J Immunol 2012 42 9 2441 51 10.1002/eji.201242434 22684996 43 Barron L Dooms H Hoyer KK Kuswanto W Hofmann J O’Gorman WE Cutting edge: mechanisms of IL-2-dependent maintenance of functional regulatory T cells Journal of immunology (Baltimore, Md : 1950) 2010 185 11 6426 30 10.4049/jimmunol.0903940 44 Verma C Kaewkangsadan V Eremin JM Cowley GP Ilyas M El-Sheemy MA Natural killer (NK) cell profiles in blood and tumour in women with large and locally advanced breast cancer (LLABC) and their contribution to a pathological complete response (PCR) in the tumour following neoadjuvant chemotherapy (NAC): differential restoration of blood profiles by NAC and surgery J Transl Med 2015 13 180 10.1186/s12967-015-0535-8 26040463 45 Eremin O Coombs RR Ashby J Lymphocytes infiltrating human breast cancers lack K-cell activity and show low levels of NK-cell activity Br J Cancer 1981 44 2 166 76 10.1038/bjc.1981.167 6268135 46 Tiainen S Tumelius R Rilla K Hamalainen K Tammi M Tammi R High numbers of macrophages, especially M2-like (CD163-positive), correlate with hyaluronan accumulation and poor outcome in breast cancer Histopathology 2015 66 6 873 83 10.1111/his.12607 25387851

PMC005xxxxxx/PMC5002191.txt

==== Front BMC GenomicsBMC GenomicsBMC Genomics1471-2164BioMed Central London 302510.1186/s12864-016-3025-3Methodology ArticleMolecular and structural considerations of TF-DNA binding for the generation of biologically meaningful and accurate phylogenetic footprinting analysis: the LysR-type transcriptional regulator family as a study model Oliver Patricia poliverster@gmail.com 1Peralta-Gil Martín martinpergil@gmail.com 2Tabche María-Luisa tabche@ibt.unam.mx 1http://orcid.org/0000-0002-0374-0640Merino Enrique merino@ibt.unam.mx 11 Departmento de Microbiología Molecular, Instituto de Biotecnología, Universidad Nacional Autónoma de México, Cuernavaca, Morelos México 2 Escuela Superior de Apan de la Universidad Autónoma del Estado de Hidalgo, Carretera Apan-Calpulalpan, Km 8, Chimalpa Tlalayote s/n, Colonia Chimalpa, Apan, Hidalgo México 27 8 2016 27 8 2016 2016 17 1 68618 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 The goal of most programs developed to find transcription factor binding sites (TFBSs) is the identification of discrete sequence motifs that are significantly over-represented in a given set of sequences where a transcription factor (TF) is expected to bind. These programs assume that the nucleotide conservation of a specific motif is indicative of a selective pressure required for the recognition of a TF for its corresponding TFBS. Despite their extensive use, the accuracies reached with these programs remain low. In many cases, true TFBSs are excluded from the identification process, especially when they correspond to low-affinity but important binding sites of regulatory systems. Results We developed a computational protocol based on molecular and structural criteria to perform biologically meaningful and accurate phylogenetic footprinting analyses. Our protocol considers fundamental aspects of the TF-DNA binding process, such as: i) the active homodimeric conformations of TFs that impose symmetric structures on the TFBSs, ii) the cooperative binding of TFs, iii) the effects of the presence or absence of co-inducers, iv) the proximity between two TFBSs or one TFBS and a promoter that leads to very long spurious motifs, v) the presence of AT-rich sequences not recognized by the TF but that are required for DNA flexibility, and vi) the dynamic order in which the different binding events take place to determine a regulatory response (i.e., activation or repression). In our protocol, the abovementioned criteria were used to analyze a profile of consensus motifs generated from canonical Phylogenetic Footprinting Analyses using a set of analysis windows of incremental sizes. To evaluate the performance of our protocol, we analyzed six members of the LysR-type TF family in Gammaproteobacteria. Conclusions The identification of TFBSs based exclusively on the significance of the over-representation of motifs in a set of sequences might lead to inaccurate results. The consideration of different molecular and structural properties of the regulatory systems benefits the identification of TFBSs and enables the development of elaborate, biologically meaningful and precise regulatory models that offer a more integrated view of the dynamics of the regulatory process of transcription. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-3025-3) contains supplementary material, which is available to authorized users. Keywords Phylogenetic footprinting analysisMotif profileTranscription factorsBinding sitesTranscription regulationLysR-type transcription regulator familyLTTRhttp://dx.doi.org/10.13039/501100003141Consejo Nacional de Ciencia y Tecnología235817Merino Enrique Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica (PAPIIT)IN201714Merino Enrique issue-copyright-statement© The Author(s) 2016 ==== Body Background Gene regulation is a key feature of all organisms in response to intracellular and environmental changes. Bacterial gene regulation primarily occurs at the beginning of transcription by transcription factors (TFs) that recognize specific regions near promoter sequences and results in the activation or repression of the transcription of the nearby genes. The number of TFs in prokaryote genomes typically scales as the square of the total number of their genes [1]. For example, for the model organism Escherichia coli, with 4,405 genes, approximately 8 % of these genes have been estimated to code for predicted or known TFs [2], of which 35 % correspond to activators and 43 % to repressors, while 22 % have dual activities [3]. The in silico identification of transcription factor-binding sites (TFBSs) is a key issue for many molecular biology studies aimed at characterizing regulatory elements in genome sequences. These analyses have been performed by considering either different co-regulated genes in one genome [4] or a set of upstream regions of orthologous genes in closely related genomes, a procedure known as phylogenetic footprinting analysis [5–8]. In any case, it is assumed that the nucleotide conservation of a specific region in the set of sequences is indicative of a selective pressure required for the recognition of TFs for their corresponding TFBSs. Based on this principle, the goal of many programs that have been developed to find TFBSs has been the identification of discrete sequence motifs that are significantly over-represented in a given set of sequences where a TF is expected to bind. These motifs are considered to be part of the TFBSs and are commonly represented as position-specific scoring matrices (PSSMs). TFBSs and their corresponding PSSMs have been compiled in a number of different databases, such as RegulonDB [9], EcoCyc [10], RegPrecise [11], Prodoric [12] and Tractor_DB [13]. To evaluate the significance of these TFBS predictions, different approaches have been developed based on theoretical models, such as log-odds, entropy-weighted values [14] or the combination of theoretical and empirical score distributions [15]. Despite their extensive use, the accuracies reached with these programs remain low. In many cases, the true TFBSs are excluded from the identification process or are imprecisely identified, especially when they correspond to low-affinity but important binding sites of the regulatory systems. In other words, the significance of a motif given its over-representation in a set of sequences of co-regulated genes is not necessarily the best way to identify the set of TFBSs for a given regulon. Herein, we present a new computational protocol, termed Phylogenetic Profile of Consensus Motifs (PProCoM), which is based on the construction of profiles obtained from a set of consensus motifs of canonical phylogenetic footprinting techniques, using analysis windows of different incremental sizes. This profile of motifs was further examined considering the fundamental aspects of the TF-DNA binding process, such as: i) the active homodimeric conformations of some TFs impose symmetric structures on the DNA binding sites (these symmetric DNA sites can be classified as directed repeat (DR) or inverted repeat (IR) sequences); ii) the possibility of the TFs binding cooperatively in adjacent low-affinity sites (these low-affinity sites are commonly not included in the reported TFBS; see Fig. 1a and b); iii) the effects of the presence or absence of low-molecular weight effectors, known as co-inducers; iv) the incorrect assignment of very long spurious motifs due to the proximity of two or more TFBSs or the proximity of one TFBS to a promoter sequence (see Fig. 1c and d); v) the improper inclusion of AT-rich sequences in the reported motifs not recognized by the TF but that are over-represented because they provide the DNA flexibility required for TF binding (see Fig. 1a to d); and vi) the dynamic order in which the different binding events take place to determine the appropriate regulatory response (i.e., activation or repression) of the system.Fig. 1 Common types of incorrectly identified regulatory motifs in phylogenetic footprinting analyses that do not correspond to real TFBSs of the LysR-type family in Gammaproteobacteria. The TFs belonging to the LysR-type family in Gammaproteobacteria are commonly transcribed in a divergent orientation with respect to their TGs. In the intergenic region of the TF-TG, there are two to three IRs, represented by purple (IR1), green (IR2) and red (IR3) boxes. The -35 and -10 boxes of the TG are represented by cream rectangles. The question marks represent motif regions that are not commonly identified, while the exclamation marks represent DNA regions that are not part of the regulatory motif but were identified as such. Due to the molecular bases of the regulatory systems, each TFBS was recognized with different affinities by their corresponding TF. Therefore, their sequence conservation varies, wherein IR1 is the most conserved sequence and IR2 is the least conserved sequence. Additionally, the sequence conservation within an IR sequence also presents important differences. Colored spaces within the boxes represent nucleotides of the motif that are more conserved, while white spaces represent poorly conserved nucleotides. Additionally, nucleotide conservation levels of the motifs are represented with plus signs (+), with +++ (three plus signs) indicating DNA regions with the most conserved nucleotides and + (one plus sign) indicating less conserved DNA regions. In each example, the name of the TF of the regulatory system and its corresponding references are indicated. a Only IR1, the most conserved of the IRs, is identified. b Only the most conserved parts of IR1 and IR3 are identified. c A large DNA region including IR1, the most conserved part of the IR2 and IR3 are identified. Additionally, the DNA regions between IR1 and IR2 that are not recognized by the TF are also incorrectly included. d A contiguous long DNA region, including the contiguous IR1 and IR2 sequences and the sequence between them, are reported as the TF-binding sequence To evaluate the performance of our protocol, we analyzed the regulatory system of six members of the LysR-type transcriptional regulator family in Gammaproteobacteria. This family represents one of the most important families of TFs in bacteria with poorly conserved TFBSs. The members of this family have three domains: the N-terminal domain, which contains a helix-turn-helix motif for DNA binding; a central domain involved in co-inducer recognition; and a C-terminal domain required for both DNA binding and co-inducer response [16]. For most of the cases in our study, we identified TFBSs with different sequence-conservations and, thus, different affinity strengths. In our study, we found that all identified TFBSs were biologically meaningful and allowed us to propose precise dynamic regulatory models. Results To assess the performance of our protocol, we performed in silico identifications of the binding sites of TFs of six regulatory systems that are members of the LysR-type family in Gammaproteobacteria, with target genes (TGs) commonly transcribed in divergent orientations. For comparative purposes, we divided these systems into three different groups in accordance with the regulatory activity of the TF on its TG and the position of the TFBSs with respect to the promoter sequences of the regulated genes. Group one: GcvA and MetR The group one is composed of two regulatory systems with the TFs GcvA and MetR. In the intergenic sequences of these regulatory systems, our PProCoM analysis identified two TFBSs (IR1 and IR2, Figs. 2 and 3). The genes coding for the TFs and TGs are transcribed in opposite directions. The transcriptional activation of the TG occurs when one dimer of the TF binds to a TFBS located adjacent the -35 box of the TG promoter and interacts with the RNA polymerase (RNAP; IR2 of Figs. 2 and 3). Simultaneously, the self-repression of the TF occurs when it binds to a TFBS that overlaps its own promoter located on the opposite strand of the DNA (IR1 of Figs. 2 and 3).Fig. 2 PProCoM analysis of the gcvA-gcvB intergenic regions in Gammaproteobacteria. a Profile of multiple consensus sequences of increasing length positioned relative to the E. coli K12 gcvA-gcvB intergenic region. In the left column, separated by a pipe, the window width used in each MEME analysis, the E-value obtained for each motif and the number of organisms presenting the identified motif (out of 150 Gammaproteobacteria used in our analysis) are indicated. The last of these consensus sequences is indicated as dm and corresponds to the default motif without forcing the size of the analysis window (see the Methods section). The consensus motifs of the IR sequences (IR1 and IR2) are displayed at the top of the figure and are represented with inverted black arrows. b TFBSs with experimental reported evidence, with references cited on the left side of the figure. c Each one of the identified motifs was mapped into the E. coli K12 gcvA-gcvB intergenic region and was used as a reference. Black arrows indicate TSSs that had been previously identified or proposed in our study. The -35 and -10 promoter boxes are indicated with yellow boxes. TSSs and -35 and -10 promoter boxes are indicated with solid lines if these elements had been previously reported and with dashed lines if these elements were identified based on our PProCoM analysis. The center positions of the IR motifs related to the beginnings of transcription of the genes coding for the TF or TG are indicated. The nucleotides of the E. coli IR1 sequence, matching the consensus, are underlined with red lines. d A LOGO corresponding to a representative consensus was selected from the profile of a consensus of the section (marked with a red asterisk) and is shown. This LOGO includes all of the regulatory motifs of the intergenic region of study Fig. 3 PProCoM analysis of the metR-metE intergenic regions in Gammaproteobacteria. The descriptions of sections (a to d) and the symbols are the same as those of Fig. 2. d None of the motifs obtained using the different analysis window sizes include all IR sequences of the intergenic metR-metE region; therefore, the LOGOs of two different window sizes are included The GcvA regulatory system GcvA, Glycine Cleavage A, is a TF that regulates the transcription of genes involved in the serine-glycine pathway of E. coli [17, 18]. This regulator is encoded by the divergent operon gcvA-gcvB from overlapping promoters and has a common regulatory mechanism. In the presence of glycine, GcvA is negatively auto-regulated and coordinately increases the transcription of the gcvB divergent gene coding for a small RNA [19] by direct interaction with α and β RNAP [20]. Additionally, GcvA regulates the transcription of the gcvTHP operon [18]. Using DNase I footprint analysis, Wilson et al. reported that in E. coli, GcvA protects a large 48-bp sequence in the intergenic region of gcvA-gcvB and two other sequences 35 and 57 bp upstream the gcvTHP operon [18]. Alignment of these sequences revealed a conserved 5′-CTAAT-3′ motif, which was subsequently determined by site-directed mutagenesis to be important for GcvA binding and the negative regulation of the gcvTHP and gcvA transcription units [18, 21, 22]. In general, the GcvA-binding sites do not present a clear sequence conservation, except for a short 5′-CTAAT-3′ motif. Additionally, the protected regions of GcvA contain the IR sequence, 5′-ATTA-n7-TAAT-3′ [18], which is coincident with the GcvA-binding site reported in the RegPrecise database [11]. Our PProCoM analysis of the gcvB-gcvA intergenic region identified the presence of two 15-bp IR sequences (5′-ATTAG-n5-CTAAT-3′, see Fig. 2). These IR sequences include the previously mentioned motifs reported by Wilson et al., 5′-CTAAT-3′ and 5′-ATTA-n7-TAAT-3′ [18]. Considering the E. coli gcvA-gcvB intergenic region, the central positions of the predicted IR1 and IR2 motifs are located -65 and -43 bp from the gcvB transcription start site (TSS), respectively (see Fig. 2). It is important to remark that the sequences shown in this figure are not the result of a standard sequence alignment but are obtained from the relative location of conserved motifs of different sizes to the E. coli gcvB-gcvA regulatory region (see the Methods section). The MetR regulatory system MetR is a TF that regulates the expression of genes involved in methionine biosynthesis and protection against nitric oxide [23–28]. The transcriptional activity of MetR is modulated by homocysteine, the metabolic precursor of methionine. In the presence of homocysteine, MetR activates the transcription of some genes, such as metE and glyA, and represses the transcription of a few others, such as metH, metA, and hmp, along with its own transcription [23–31]. In E. coli and Salmonella typhimurium, the metE and metR genes are divergently transcribed from overlapping promoters; thus, they share a common regulatory region [23–29]. DNase I footprint and mutational analyses in S. typhimurium showed that MetR binds to two IR sequences arranged in tandem with different affinities to consensus sequence 5′-TGAAnnTnnTTCA-3′ [29]. In E. coli, two binding sites with the same characteristics have been reported in the regulatory region of the divergently transcribed hmp-glyA genes regulated by MetR, [30, 31]. The presence of homocysteine has been postulated to enhance the affinity of MetR to these contiguous DNA-binding sites to activate metE and repress metR transcription [29]. To date, no experimental evidence supports the existence of two MetR-binding sites on the E. coli metR-metE intergenic region. Nevertheless, based on our PProCoM analysis, we identified two 15-bp IR sequences with consensus sequence 5′-ATGAA-n5-TTCAT-3′, which is the reported size of the TFBSs of the LysR-type TF family [16]. Based on the E. coli reference sequence, we localized the distal and more conserved IR1 site -63 bp from the metE TSS, while the proximal and less conserved site, IR2, was located -41 bp from the TSS of metE (Fig. 3). These central locations are among the preferred positions of the transcriptional activators in E. coli [32, 33]. As shown in Fig. 3, the E. coli IR1-IR2 inter-motifs sequence is one base shorter than the IR1-IR2 inter-motifs sequence of the overall PProCoM motif alignment (see Fig. 3, sloped-dotted lines). The effect, if any, of this one missing base in the E. coli metR-metE inter-motifs space on the system regulation is not clear. Nevertheless, longer variations in the inter-motifs space, such as 6 bases (half-helix turn), have been demonstrated to have a negative effect on S. typhimurium metE transcription [34]. Additionally, point mutations in any of the two proposed TFBSs have also been reported to decrease the metE transcription, indicating that both TFBSs are required for full metE activation [29]. The 15-bp consensus sequence obtained from our PProCoM analysis was coincident with that reported for MetR in the RegPrecise database, i.e., 5′-ATGAAAATTTTTCAT-3′ [11]. Group two: OxyR, IlvY and CynR Group two is composed of three regulatory systems with the TFs OxyR, IlvY and CynR. Our PProCoM analyses of these regulatory systems identified three TFBSs in their corresponding intergenic regulatory regions (IR1, IR2 and IR3; Figs. 4, 5 and 6). The transcriptional activation of the TG occurs by the cooperative binding of two dimers that recognize IR1 and IR2 in the presence of their respective inducers. Because IR2 is located adjacent to the -35 box of the TG promoter, the TF dimer bound to this site promotes TG transcription by its interaction with the RNAP. Additionally, the self-repression of the TF simultaneously occurs because IR1 overlaps the TF promoter located on the opposite strand of the DNA (IR1 of Figs. 4, 5 and 6). The main difference with respect to our group one of regulatory systems is the presence of a third TFBS that overlaps the -35 box of the TG promoter (IR3; Figs. 4, 5 and 6). A remarkable characteristic of this group is that this third TFBS (IR3, used for TG repression) partially overlaps the second TFBS (IR2, used for TG activation) in such a way that the binding of the TF to one of these two mutually exclusive sites determines the transcriptional regulatory activity (i.e., activation or repression) on the TF over the TG.Fig. 4 PProCoM analysis of the oxyR-oxyS intergenic regions in Gammaproteobacteria. The descriptions of sections (a to d) and the symbols are the same as those of Fig. 2 Fig. 5 PProCoM analysis of the ilvY-ilvC intergenic regions in Gammaproteobacteria. The descriptions of sections (a to d) and the symbols are the same as those of Fig. 2 Fig. 6 PProCoM analysis of the cynR-cynT intergenic regions in Gammaproteobacteria. The descriptions of sections (a to d) and the symbols are the same as those of Fig. 2. d None of the motifs obtained using the different analysis window sizes includes all IR sequences of the intergenic cynR-cynT region; therefore, the LOGOs of two different window sizes are included The OxyR regulatory system OxyR is a TF that regulates the expression of genes involved in oxidative stress protection, redox balance, and manganese uptake [35–38]. The transcriptional activity of OxyR depends on its oxidized state, which determines the reversible disulfide bond formation of a pair of cysteine residues in its amino acid sequence [39]. In its oxidized state, OxyR activates the transcription of the divergent small RNA gene oxyS. Additionally, OxyR represses its own expression under oxidizing and reducing conditions [40]. Based on DNase I footprint analyses, Tartaglia et al. showed that OxyR binds to an unusually long DNA region that spans over 45 bp, with putative OxyR-binding sites with no obvious sequence similarity [41]. Using an in vitro binding assay of OxyR to random oligonucleotides and DNase I footprint analyses, Toledano et al. showed that the DNA recognition of OxyR depends on its oxidized/reduced states. In its oxidized form, OxyR recognizes a DNA region that includes four repetitions of the 5′-ATAGnt-3′ sequences located in four contiguous major grooves on one face of the DNA helix. In its reduced form, OxyR binds two repetitions of the 5′-ATAGnt-3′ sequences located at two pairs of major grooves separated by one helical turn [40]. Our PProCoM analysis of the oxyR-oxyS intergenic region identified the presence of three 15-bp IR sequences (5′-ATAG-n7-CTAT-3′). Considering the E. coli oxyR-oxyS intergenic region, the central positions of the predicted IR1, IR2 and IR3 motifs are located -66, -44 and -35 bp from the oxyS TSS, respectively (see Fig. 4). The IlvY regulatory system IlvY positively regulates the transcription of ilvC, a gene involved in isoleucine and valine biosynthesis. The transcriptional activation of ilvC by IlvY depends on the presence of an IlvY inducer, such as acetolactate or acetohydroxybutyrate. At the same time, IlvY negatively regulates its own transcription in an inducer-independent manner [42, 43]. The ilvY and ilvC genes are divergently transcribed from overlapping promoters. Using DNase I footprint analyses, Wek and Hatfield proposed that IlvY binds to two 27-bp operator sequences, named O1 and O2, in the ilvY-ilvC intergenic region [43]. These regions are arranged in tandem and possess imperfect 21-bp inverted repeat motifs: O1, 5′-ACgTTGCAAaaaTTGCAAtGT-3′ (centered at position +17 relative to the ilvY TSS), and O2, 5′-aTATatCaatttccGcaATAa-3′ (which overlaps the proposed -10 and -35 promoter boxes of ilvY and the -35 promoter box of ilvC). The consensus IlvY-binding motif common to the O1 and O2 operators is 5′-A[C/T]ATTGCAA-3′ [43]. These authors proposed that IlvY represses its own transcription in an inducer-independent manner when IlvY binds to O1 and activates transcription of ilvC when IlvY binds to the O1 and O2 operators in a cooperative dependent manner in the presence of the system inducers. In this condition, the transcriptional activation of ilvC was proposed to result from IlvY-RNAP interactions when IlvY was bound to O2 or by a change in the DNA conformation at the ilvC -35 promoter box. Following this reasoning, Rhee et al. proposed that the transcription of the divergent genes ilvY and ilvC is coupled in a DNA supercoiling-dependent manner that increases the binding of the RNAP at this promoter by nearly 100-fold [42]. Our PProCoM analysis of the ilvY-ilvC intergenic region identified the presence of three 15-bp IR sequences (5′-TTGCA-n5-TGCAA-3′; see Fig. 5). Considering the E. coli ilvY-ilvC intergenic region, the central positions of these predicted IR1, IR2 and IR3 motifs are located -65, -43 and -34 bp from the ilvC TSS, respectively (see Fig. 5). The CynR regulatory system CynR is a TF that regulates the transcription of the cynTSX operon, which is involved in cyanate detoxification. Cyanate is also used as a nitrogen source due to its hydrolysis to ammonia and bicarbonate [44]. This activation of the cynTSX operon by CynR depends on the presence of cyanate. CynR also negatively regulates its own transcription in a cyanate-independent manner [44]. As in the case of the abovementioned LysR-type regulatory systems, the gene coding for the TF (cynR) and its regulatory TGs (cynTSX) are transcribed in opposite directions, and their corresponding promoters overlap [45, 46]. Using DNase I digestion analyses, Lamblin and Fuchs showed that CynR binds to a 60-bp region in the cynR-cynTSX intergenic region and proposed that this region contains two putative binding sites with different affinities [46]. The first of these regions, R1 (5′-ATAAGTAAA-3′), was proposed to have the highest binding affinity, whereas the second region, R2 (5′-ATAAGGTAA-3′), was proposed to overlap the entire cynR promoter sequence and the -35 promoter region of the cynTSX operon [45, 46]. These authors suggested that in a first instance, a CynR dimer could bind to R1 (i.e., the most conserved region), and in a second but almost simultaneous instance, another CynR dimer could bind to R2 in a strong cooperative manner. These authors also proposed that the transcriptional activation of the cynTSX operon takes place in the presence of cyanate, which was believed to trigger a conformational change in CynR, modifying its interaction with DNA [46]. Our PProCoM analysis of the cynR-cynTSX intergenic region identified the presence of three 15-bp IR sequences (5′-ATAA-n7-TTAT-3′), including the sequences proposed by Lamblin and Fuchs (see Fig. 6). Considering the E. coli cynR-cynTSX intergenic region, the central positions of the predicted IR1, IR2 and IR3 motifs are located -66, -44 and -34 bp from the cynTSX TSS, respectively (see Fig. 6). Group three: LysR The group three is composed of one regulatory system with the TF, LysR. Our PProCoM analysis of this regulatory system identified three TFBSs in the lysR-lysA intergenic regulatory region (IR1, IR2 and IR3; Fig. 7). As observed in the groups one and two of the LysR-Type family, the transcriptional activation of the TG (lysA) occurs by the cooperative binding of two dimers, which recognize the IR1 and IR2 sites in the presence of a LysR inducer (i.e., diaminopimelic acid). The self-repression of the TF (lysR) occurs simultaneously because IR1 overlaps the TF promoter located on the opposite strand of the DNA (IR1 of Fig. 7). The main difference with respect to the regulatory system of the group two is that the second and third TFBSs do not overlap (IR2 and IR3; Fig. 7).Fig. 7 PProCoM analysis of the lysR-lysA intergenic regions in Gammaproteobacteria. The descriptions of sections (a to d) and the symbols are the same as those of Fig. 2 The LysR regulatory system LysR is a TF that regulates the transcription of lysA, which encodes an enzyme that catalyzes the final step of lysine biosynthesis. LysR negatively regulates its own transcription and positively regulates the transcription of lysA in the presence of its inducer, diaminopimelic acid [47–49]. As in the previous cases, the genes coding for the TF (lysR) and its regulatory TG (lysA) are transcribed in opposite directions. The TFBSs of LysR and their regulatory mechanism have not yet been identified. However, the LysR-binding sites have been determined to be within a 73-bp DNA fragment located 48 bp upstream of the lysA structural gene [48]. The intracellular concentration of active LysR could be limiting because its regulatory role is diminished when the abovementioned fragment is cloned on plasmids [48]. Based on experimental analyses, the lysR TSS has been predicted to be located 26 bp upstream of its structural gene [9]. However, a putative lysA promoter, with a -35 box (TTGcat) and a -10 box (TATTTT), has been predicted to be located 52 bp from the lysA coding region [50]. The corresponding TSS has been proposed to be located 3 bp downstream of the -10 box of the predicted promoter [50]. Our PProCoM analysis of the lysR-lysA intergenic region identified the presence of three 15-bp IR sequences (5′-ATATC-n5-GATAT-3′, see Fig. 7). Considering the E. coli lysR-lysA intergenic region, the central positions of the predicted IR1, IR2 and IR3 motifs are located -64, -43 and -9 bp from the lysA TSS, respectively (see Fig. 7). Based on the positions of these predicted TFBSs, we postulate that the lysA TSS is located 22 bp upstream of its structural gene. Discussion Common sequence motifs of the TFBSs of the LysR-type TF family Our systematic analyses of the six representative members of the LysR-Type family in Gammaproteobacteria using our PProCoM protocol allowed us to precisely identify their corresponding TFBSs and common characteristics, which are summarized in Fig. 8. The genes coding for the TFs and their TGs are transcribed in divergent orientations. Their intergenic regions present two or three inverted repeated motifs, i.e., IR1, IR2 and IR3. Based on this figure, the IR sequences are clearly conserved in terms of their length and common inter-motif distances and also show similarities in their molecular regulatory mechanisms. In addition to these specific sequence conservations, the correct identification of the TFBSs of these LysR-type TFs enables us to identify conserved sequence similarities represented by the consensus sequence 5′-CTATA-n9-TATAG-3′, as shown in Fig. 9. This consensus sequence can be considered an extended version of the T-n11-A “LysR motif”, originally postulated by Goethals et al. and based on the analyses of the TFBSs of NodD in Azorhizobium and other members of the LysR-type family [51]. This relevant sequence conservation of the TFBS among members of the LysR-type family can be explained when considering that new genomes frequently acquire these TFs via horizontal gene transfer. Furthermore, these TFs evolved from a common ancestor, as evidenced by the conservation of similar sequences for the binding motifs and the similar molecular mechanisms that regulate the transcriptional responses to a variety of stimuli and functions, including metabolism, quorum sensing, motility and virulence, among others (reviewed in [52]). From Fig. 9, it can also be seen that the most common interspace size between the 5′-CTATA-3′ and 5′-TATAG-3′ monomer recognition motifs is 9 nucleotides. Minor variations of this length can be found; the largest variation was observed for IlvY, with an interspace of 11 nucleotides. Our PProCoM analyses show that this inter-monomer motif sequence has a high AT content to provide the DNA the flexibility required for proper TF-DNA recognition. A few variations to this 5′-CTATA-n9-TATAG-3′ consensus exist because they are required for the specific recognition of a TF by its corresponding TFBSs (see Figs. 2, 3, 4, 5, 6 and 7). Figure 9 also includes representative examples of other members of the LysR-type with experimentally characterized TFBSs. These TFBSs are consistent with our extended LysR motif consensus. For example, in the catR-catBC intergenic region of Pseudomonas putida, the distal TFBS of catBC, also known as the repressor-binding site, has an imperfect palindromic sequence 5′-tcAgA-n9-TATgG-3′ (note the underlined g nucleotide) that resembles our extended LysR motif, 5′-CTATA-n9-TATAG-3′. Site-directed mutagenesis of G➔T in the fourth nucleotide of this motif created a sequence most similar to the consensus and resulted in increased binding of the CatR and increased the transcription level of the catBC operon. However, substitutions of the first A➔T in the same 5′-tcAgA-n9-TATgG-3′ TFBS (note the underlined A nucleotide) made this sequence less similar to the consensus and resulted in decreased binding of CatR, with concomitantly decreased transcription of the catBC operon [53]. Our second example corresponds to the OccR regulatory system in Agrobacterium tumefaciens. By cloning discrete regions of the occR-occQ intergenic region and characterizing them using DNase I footprinting and gel mobility shift assays, Wang et al. defined five binding sites of OccR and their relative affinities [54, 55]. Sites 1 and 2 formed an IR located -33 bp from the occQ TSS and corresponded to the IR3 site of the IRs used in this work. Sites 4 and 5 form another IR located -64 bp from the occQ TSS, corresponding to IR1, the site with the greatest affinity for OccR. Sites 3 and 2 form an IR that corresponds with IR2, i.e., the site with the least affinity of the three IRs of the system [54]. The replacement of IR3 with IR1, i.e., the IR with the greatest affinity, resulted in an enhanced binding of OccR and a greater transcription repression of the occQ TG [55]. OccR only binds IR2, i.e., the IR with the smaller affinity, in a cooperative dependent manner in the presence of the system inducer, octopine [55]. Nevertheless, the replacement of IR2 with IR1, i.e., the IR with the greatest affinity, resulted in a partial octopine-independent binding of OccR to this site [55]. Finally, our third example corresponds to the pcaQ-pcaMNVWX intergenic region in Sinorhizobium meliloti. Based on site-directed mutagenesis, McLean et al. proposed that the PcaQ-binding site corresponded to the sequence 5′-ATAaccgggggatTAT-3′ which central position is located 65.5 bp upstream of the structural gene (see Fig. 9). Relevant changes in the nucleotides for TF recognition in our 5′-CTATA-n9-TATAG-3′ consensus resulted in decreased transcriptional activation of the target pcaMNVWX operon in the presence of its inducer [56]. These mutations involved the A➔G changes at the underlined nucleotides of the sequence 5′-ATA-n10-TAT-3′, generating the sequences 5′-GTA-n10-TAT-3′, 5′-ATG- n10-TAT-3′ and 5′-ATA- n10-TGT-3′ [56].Fig. 8 Architecture of the TFBSs of the LysR-type TF family in Gammaproteobacteria revealed by PProCoM analysis. A common characteristic of the members of the LysR-type family in Gammaproteobacteria is that their coding genes and corresponding target genes are transcribed in divergent orientations, and their intergenic regions present two or three inverted repeated motifs, IR1, IR2 and IR3. The architecture of the intergenic regions of the six TFs analyzed in our study is summarized. Clear conservations of motif length and inter-motif distance suggest that there are similarities in their molecular regulatory mechanisms Fig. 9 Consensus sequence for the TFBSs of the LysR-type TF family. The T-n11-A motif was originally proposed by Goethals et al. [51] as the consensus sequence recognized by members of the LysR-type family. Considering the results of our PProCoM analysis of the TFBSs of six representative members of this family in Gammaproteobacteria, we defined a new and extended version of this motif: 5′-CTATA-n9-TATAG-3′. Additionally, examples of the sequence consensus of the TFBSs of other members of the LysR-type family that have been experimentally verified are also shown and include the distal TFBSs of CatR of the Gammaproteobacteria Pseudomonas putida [53], OccR of the Alphaproteobacteria Agrobacterium tumefaciens [53, 54] and PcaQ of the Alphaproteobacteria Sinorhizobium meliloti [56]. Dots within the inter-motif sequences were used to align the conserved nucleotides of the consensus sequences Dynamic models of regulation In addition to a static description of the TFBSs identified by our PProCoM analyses, dynamic models of regulation for each one of our systems can be elucidated based on the characteristics of the elements of the regulatory system, which are as follows:The intergenic sequences of the regulatory systems of group one (metR-metE and gcvA-gcvB), contained two IR motifs, whilst the regulatory systems of group two (oxyR-oxyS, ilvY-ilvC, and cynR-cynT) and group three (lysR-lysA) contained three IR motifs. In all these cases, the IR motifs show different sequence conservation, and thus, different affinity. In group one, IR1 is the most conserved, and IR2 is the least conserved motif. In groups two and three, IR1 and IR3 are the most conserved, and IR2 is the least conserved motif. All the TFs analyzed; GcvA, MetR, OxyR, IlvY, CynR and LysR, adopt two different conformations depending on the presence or absence of their corresponding inducers: glycine, homocysteine, reactive oxygen species, acetolactate, cyanate and diaminopimelic acid, respectively. Without the system inducers, the TFs bind as dimers, preferentially to IR1, in the case of group one, and to IR1 and IR3, in the case of groups two and three. In accordance with this binding, footprinting assays with LysR family members show a hypersensitive region 50 bp upstream of the TSS of IlvY [42, 43], CynR [46], OccR [54, 57]. Similar results have been observed in studies with other regulatory TFs of the LysR family such as ClcR [58], CatR [58] and PcaQ [59]. In the case of CynR, this hypersensitive region corresponds to the region where the DNA curves with the binding of CynR [46]. In the presence of the system inducers, the TFs bind DNA as dimers of dimers in a cooperative manner. Only through this cooperative binding the TFs can recognize IR2, the less conserved of the TFBSs. This kind of binding for members of the LysR TF family has been demonstrated by footprinting assays [18, 29, 34, 40, 42, 43, 54, 57] and site directed mutagenesis analysis [21, 29, 34, 40, 53–56]. As a consequence of this binding, the hypersensitive DNA regions located around -50 bp upstream the TSSs markedly decrease. In addition, it has been shown that altering the distance between IR1 and IR2 reduces the cooperative binding of the TFs [40, 54, 57]. A TFs acts as transcription repressor or activator of the TF or TG genes depending on the position of the IR to which it binds. The IR1 motifs are downstream or overlap the -10 box of the TF promoters, therefore, the auto-repression of transcription takes place when the TFs are bound to IR1 sites. The IR2 motifs overlap the TFs promoters and are also immediately downstream of the -35 box of the TGs promoters, therefore, a TF bound to IR2 represses the TF transcription and activates the TG transcription. In the case of group two, the IR3 motif overlaps the TF and TG promoters, hence, a TF bound to this site simultaneously blocks the transcription of the TF and TG genes. In the case of group three, the IR3 motif only overlaps the TG promoter, accordingly, a TF bound to this site exclusively blocks TF transcription. In addition to the above-mentioned regulatory outcomes, it is worth mentioning that in the case of group two, the IR2 and IR3 sites overlap, therefore, the binding of TFs to these sites are mutually exclusive. In the absence of the system inducers, the TFs would preferentially bind IR3 since this site has greater sequence conservation than IR2; nevertheless, in the presence of the system inducers, the TFs would bind cooperatively as a dimer of dimers to IR1 and IR2. In this case, the binding of the TFs to IR2 would have two positive effects on TG transcription; directly by its interaction with the RNA polymerase, and indirectly, by blocking the binding of the TFs to IR3, an event that otherwise would repress TF transcription. Representative regulatory models of the LysR-type TF family in Gammaproteobacteria revealed by PProCoM analyses In addition to the sequence motifs identified in the TFBSs of the members of the LysR-type TF family, our regulatory model also includes the effects of TF binding on the DNA curvature. These effects have also been reported for several TFs, such as GcvA [18], MetR [30], OxyR [40], IlvY [42], CynR [46], LysR [52], CysB [60], CatR [53], ClcR [58], and OccR [57]. In the above-mentioned regulatory systems, based on DNase footprinting analyses, it has been reported that in the absence of system inducers, TFs bind to long regions of DNA. Conversely, in the presence of inducers, the protected area of the DNA in the footprinting analyses significantly decreases. For example, in absence of the inducer, OccR protects a region of approximately 60 bp, resulting in DNA with a curvature angle of 62°, showing hypersensitive regions around the -50 position [61]. In the presence of the inducer, the angle decreases to 46°, shortening the length of protected DNA in the DNase footprinting assay to 50 bp, decreases the hypersensitive region [57, 61]. Toledano et al. proposed that this reduction in the length of protected DNA was caused by the rearrangement of the dimers of dimers of the TFs. In the absence of inducers, dimers bind to distal sites, e.g., IR1 and IR3. A single turn of the separation between the two dimers causes a bend in the DNA and, consequently, the inhibition of the transcription of divergent transcription units [40]. A similar regulatory model was proposed by Wang and Winans in the occR-occQ regulatory system [57]. The results obtained with our PProCoM protocol, summarized as our 5′-CTATA-n9-TATAG-3′ extended LysR-type TFBS motif consensus (Fig. 9) and schematized in our model (Fig. 10), are consistent with the observations on DNA curvature available in the literature.Fig. 10 Representative regulatory models of the LysR-type TF family in Gammaproteobacteria revealed by PProCoM analyses. a A typical architecture of the regulatory regions of these TFs is the presence of three IR sequences, represented by blue (IR1), green (IR2) and red (IR3) boxes. Some regulatory systems, such as those of our first analysis group, GcvA and MetR, lack the third IR element. b Because the sequence affinities of IR1 and IR3 (observed as sequence conservation of the motifs) are greater than the one for IR2, in the absence of the inducer, the TF of the system only binds to the IR1 and IR3 sites. The positions of IR1 and IR3 are critical for the transcriptional repression of divergent systems. IR1 overlaps the TF promoter, while IR3 overlaps the TF and the TG promoters. c In the presence of the system inducer, the TF dimer can bind cooperatively to a less conserved and less affine IR in the system, i.e., IR2. A remarkable characteristic of several regulatory systems in this family is that IR2 partially overlaps IR3; therefore, a first consequence of the binding of the TF to IR2 is the steric displacement of the TF that was bound to IR3, resulting in TG transcription repression. In addition to this de-repression effect, a second effect of the binding of the TF to IR2 is direct transcriptional activation of the TG due to the position of IR2 immediately upstream of the -35 promoter box of the TG, where the TF interacts with the RNAP. Figure modified from [57] Potential use of PProCoM to identify TFBSs of other regulatory systems different to those of the LysR-Family Our PProCoM protocol can be used to identify TFBSs of almost any bacterial regulatory system if the characteristics of their TFs are considered. For example, in addition to of the LysR regulatory system, we currently conduct a study to identify the binding sites of the TF members of the AraC/XylS family [62]. These TFs usually bind DNA as dimers to repeated direct asymmetric contiguous TFBSs, being the distal one the most conserved site and proximal site the less conserved. The problem in identifying TFBS of the AraC/XylS family is the low conservation and asymmetry of these proximal TFBSs. Nevertheless, we believe that PProCoM is particularly useful identifying such low conserved binding sites since its accuracy does not exclusively depend on the sequence conservation of the TFBSs, but on the molecular properties of the TFs and their interactions between themselves, with the DNA and with the DNA polymerase. Regarding the use our PProCoM protocol for identifying TFBSs in eukaryotic organisms with small intergenic regions, such as yeast, we consider the possibility of obtaining positive results as obtained so far in prokaryotic organisms. Currently we perform site directed mutagenesis and transcriptional quantification of our regulatory systems for experimental verification of our theoretical predictions. Conclusions PProCoM represents an unconventional multiple motif alignment representing a set of consensus sequences of increasing length, which are arranged according to reference nucleotide intergenic region – E. coli sequences in our examples (Figs. 2, 3, 4, 5, 6 and 7). This strategy enables the merging of the represented motifs (with significant E-values) with less conserved motifs that play important roles in dynamic transcription regulation systems. These less conserved motifs have generally not been identified or included in previous studies, even in cases with experimental analyses, such as DNase footprinting analysis. Our PProCoM analysis of six members of the LysR-type TF family have made evident the high relevance of the less conserved motifs in the intergenic regions of their regulatory sequences. This approach enables the comprehension of the homodimeric nature of these TFs and provides a more integrated and complete picture of their regulatory processes. Methods In general, our computational PProCoM protocol is an extension of phylogenetic footprinting analysis [5]. Briefly, different consensus motifs are obtained using analysis windows with different incremental sizes that are then aligned to build a profile of consensus motifs. This use of windows of different sizes allows for the identification of the most represented (i.e., most significant) sequence motifs and includes other less represented (i.e., less significant) motifs that are nevertheless of equal importance. The fundamental computational steps of our approach are illustrated in Fig. 11 and are described as follows.Fig. 11 PProCoM workflow. The PProCoM workflow includes four main steps, represented by the units (a to d). These steps are fully described in the Methods section Retrieval of orthologous non-coding regulatory sequences of non-redundant organisms To avoid bias introduced by the sequencing of preferential model organisms, non-redundant genomes were selected from the KEGG database (release 2015) based on their phylogenetic distances, which were evaluated using the PROTDIST program [63] from a multiple alignment of concatenated sequences of a set of 31 “house-keeping” proteins defined by Ciccarelli et al. [64] (see Fig. 11). The phylogenetic group considered in our study was Gammaproteobacteria. Orthologous genes were defined using “bidirectional best hits” criteria [65] in BLAST [66]. Only intergenic regions with greater length to 10 nucleotides were considered. 150 intergenic regions were considered for analysis. The list of these organisms is presented in Additional file 1: Table S1. Obtaining the profile of significantly over-represented motifs from phylogenetic footprinting analysis Over-represented motifs in each orthologous set of regulatory regions were obtained using the Multiple EM for Motif Elicitation (MEME) program [67] considering the following parameter values:Length of the analysis window. Although MEME can automatically set the size of the analysis window to define the value at which over-representation of a motif is most significant, in our PProCoM protocol, the MEME analysis was repeatedly performed using analysis windows of different sizes, from the smallest, 10 bp, to the largest, 100 bp, in increments of 2 bp per cycle, or in the case that the results of the analysis remains unchanged despite the increment of the two pair bases. The sizes of the analysis windows were defined using the –w argument of MEME. In addition, we also include the result of a MEME analysis without forcing the size of the analysis window. In Figs. 2, 3,4, 5, 6 and 7, these motifs are indicated as dm (default motif without forcing the size of the analysis window). E-value of the motifs. Unlike most computational methods that use the E-value to define a motif as significant, in our PProCoM protocol, the E-value is considered as one, among other different criteria, for the selection of significant motifs. The above consideration is because the E-value of a motif might vary depending on the affinity of the TFBSs (high or low), the size of the analysis window, the number of sequences analyzed and on the phylogenetic distances between the organisms in the study. Nevertheless, as a first filter to define a motif as significant, the E-value was set to 1e-6 using the –evt argument of MEME. Figures 2, 3, 4, 5, 6 and 7 include the E-values obtained for each of the analysis window of our six regulatory systems. In all these cases, the E-values were statistically significant (E-values < 1e-20). Number of motifs identified. To build the PProCoM profile, only the most significant motif is considered per analysis window. This was specified setting the –nmotifs argument of MEME to 1. Motif symmetry. Considering that some homodimeric TFs, as those of the LysR-family, recognize palindromic DNA sequences, the –pal argument of MEME was used to force this symmetry in the identified motifs. Distribution of motifs. To specify that the distribution of the motifs to be found by MEME in the set of regulatory sequences corresponded to zero or one per sequence, the –mod argument of MEME was set to zoops. Background Markov model. In order to avoid the bias originated by the unbalance distribution of the nucleotides (i.e. low or high %GC) in the regulatory sequences, we build a Markov model file for each one of our six regulatory systems. The names of these files were specified using the –bfile argument of MEME. Alphabet of the sequences. The –dna argument of MEME was used to specify the nature of the nucleotide sequences used in our study. Mapping the significant motifs onto a reference sequence To identify the relative positions of the different motifs identified in the previous steps of our protocol, every motif was mapped to a reference intergenic region of a model organism. In our case, we selected E. coli K12 because it is one of the best-characterized organisms among the Gammaproteobacteria. As a result of this mapping step, a PProCoM was obtained. Integration of the mapped motifs with biological knowledge of the regulatory system and construction of dynamic models of the regulatory system To properly interpret the results obtained in the previous steps represented as a PProCoM, the molecular characteristics of the TF in the study were considered. The characteristics of the TFs belonging to the LysR-type family are listed in the Background section and include the following properties: the TF-TG divergent transcriptional orientations, the tandem arrangement of TFBSs, the inverted repeat symmetry and length of the TFs, the cooperative binding of the TFs in the presence of a specific inducer, the relative degrees of sequence conservation (i.e., binding affinities) of the TFBSs and their positions with respect to their promoters, and the spaces between the TFBSs that determine their relative orientations in terms of helix-turns. Additional file Additional file 1: List of the 150 organisms whose genomic sequences were used in our study. (XLSX 69 kb) Acknowledgements We wish to thank Ricardo Ciria for computer support and Shirley Ainsworth for bibliographical assistance. Patricia Oliver is a doctoral student from Programa de Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México (UNAM), and received a CONACyT fellowship (23556), scholar No. 45230. Funding This work was supported by Consejo Nacional de Ciencia y Tecnologıa (CONACyT) (235817) and PAPIIT (IN201714) grants to EM. Availability of data and materials The datasets of intergenic sequences of the TF-TG of the set of non-redundant Gammaproteobacteria genomes that were used as input sequences in our study are available at our web page http://www.ibt.unam.mx/biocomputo/pprocom_organisms.html. Authors’ contributions PO, MP and EM co-developed the project idea, designed and performed the analyses, interpreted the biological significance of the results, and wrote the manuscript. MLT assisted in the biological interpretation of the results. EM also coordinated the study. All authors participated in discussions and read and approved the final manuscript. Competing interests The authors have declared that no competing interests exist. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. ==== Refs References 1. Molina N van Nimwegen E Scaling laws in functional genome content across prokaryotic clades and lifestyles Trends Genet. 2009 25 243 7 10.1016/j.tig.2009.04.004 19457568 2. Martínez-Antonio A Collado-Vides J Identifying global regulators in transcriptional regulatory networks in bacteria Curr Opin Microbiol. 2003 6 482 9 10.1016/j.mib.2003.09.002 14572541 3. Pérez-Rueda E Collado-Vides J The repertoire of DNA-binding transcriptional regulators in Escherichia coli K-12 Nucleic Acids Res. 2000 28 1838 47 10.1093/nar/28.8.1838 10734204 4. Huerta AM Salgado H Thieffry D Collado-Vides J RegulonDB: a database on transcriptional regulation in Escherichia coli Nucleic Acids Res. 1998 26 55 9 10.1093/nar/26.1.55 9399800 5. Tagle DA Koop BF Goodman M Slightom JL Hess DL Jones RT Embryonic epsilon and gamma globin genes of a prosimian primate (galago crassicaudatus). Nucleotide and amino acid sequences, developmental regulation and phylogenetic footprints J Mol Biol 1988 203 439 55 10.1016/0022-2836(88)90011-3 3199442 6. Tan K Moreno-Hagelsieb G Collado-Vides J Stormo GD A comparative genomics approach to prediction of new members of regulons Genome Res. 2001 11 566 84 10.1101/gr.149301 11282972 7. Tan K McCue LA Stormo GD Making connections between novel transcription factors and their DNA motifs Genome Res. 2005 15 312 20 10.1101/gr.3069205 15653829 8. Janky R van Helden J Evaluation of phylogenetic footprint discovery for predicting bacterial cis-regulatory elements and revealing their evolution BMC Bioinformatics. 2008 9 37 10.1186/1471-2105-9-37 18215291 9. Salgado H Peralta-Gil M Gama-Castro S Santos-Zavaleta A Muniz-Rascado L Garcia-Sotelo JS RegulonDB v8.0: omics data sets, evolutionary conservation, regulatory phrases, cross-validated gold standards and more Nucleic Acids Research 2013 41 D203 13 10.1093/nar/gks1201 23203884 10. Keseler IM Mackie A Peralta-Gil M Santos-Zavaleta A Gama-Castro S Bonavides-Martínez C EcoCyc: fusing model organism databases with systems biology Nucleic Acids Res. 2013 41 D605 12 10.1093/nar/gks1027 23143106 11. Novichkov PS Brettin TS Novichkova ES Dehal PS Arkin AP Dubchak I RegPrecise web services interface: programmatic access to the transcriptional regulatory interactions in bacteria reconstructed by comparative genomics Nucleic Acids Res. 2012 40 W604 8 10.1093/nar/gks562 22700702 12. Grote A Klein J Retter I Haddad I Behling S Bunk B PRODORIC (release 2009): a database and tool platform for the analysis of gene regulation in prokaryotes Nucleic Acids Res 2009 37 D61 5 10.1093/nar/gkn837 18974177 13. Pérez AG Angarica VE Vasconcelos AT Collado-Vides J Tractor_DB (version 2.0): a database of regulatory interactions in gamma-proteobacterial genomes Nucleic Acids Res 2007 35 D132 6 10.1093/nar/gkl800 17088283 14. Oberto J FITBAR: a web tool for the robust prediction of prokaryotic regulons BMC Bioinformatics. 2010 11 554 10.1186/1471-2105-11-554 21070640 15. Medina-Rivera A Abreu-Goodger C Thomas-Chollier M Salgado H Collado-Vides J Van Helden J Theoretical and empirical quality assessment of transcription factor-binding motifs Nucleic Acids Res. 2011 39 808 24 10.1093/nar/gkq710 20923783 16. Schell MA Molecular biology of the LysR family of transcriptional regulators Annu Rev Microbiol. 1993 47 597 626 10.1146/annurev.mi.47.100193.003121 8257110 17. Wilson RL Steiert PS Stauffer GV Positive regulation of the Escherichia coli glycine cleavage enzyme system J Bacteriol. 1993 175 902 4 8423160 18. Wilson RL Urbanowski ML Stauffer GV DNA binding sites of the LysR-type regulator GcvA in the gcv and gcvA control regions of Escherichia coli J Bacteriol. 1995 177 4940 6 7665470 19. Urbanowski ML Stauffer LT Stauffer GV The gcvB gene encodes a small untranslated RNA involved in expression of the dipeptide and oligopeptide transport systems in Escherichia coli Mol Microbiol. 2000 37 856 68 10.1046/j.1365-2958.2000.02051.x 10972807 20. Stauffer LT Stauffer GV GcvA interacts with both the alpha and sigma subunits of RNA polymerase to activate the Escherichia coli gcvB gene and the gcvTHP operon FEMS Microbiol Lett. 2005 242 333 8 10.1016/j.femsle.2004.11.027 15621456 21. Jourdan AD Stauffer GV Genetic analysis of the GcvA binding site in the gcvA control region Microbiol. 1999 145 2153 62 10.1099/13500872-145-8-2153 22. Wonderling LD Urbanowski ML Stauffer GV GcvA binding site 1 in the gcvTHP promoter of Escherichia coli is required for GcvA-mediated repression but not for GcvA-mediated activation Microbiology. 2000 146 2909 18 10.1099/00221287-146-11-2909 11065369 23. Cai XY Maxon ME Redfield B Glass R Brot N Weissbach H Methionine synthesis in Escherichia coli: effect of the MetR protein on metE and metH expression Proc Natl Acad Sci USA 1989 86 4407 11 10.1073/pnas.86.12.4407 2543976 24. Weissbach H Brot N Regulation of methionine synthesis in Escherichia coli Mol Microbiol. 1991 5 1593 7 10.1111/j.1365-2958.1991.tb01905.x 1943695 25. Flatley J Barrett J Pullan ST Hughes MN Green J Poole RK Transcriptional responses of Escherichia coli to S-nitrosoglutathione under defined chemostat conditions reveal major changes in methionine biosynthesis J Biol Chem. 2005 280 10065 72 10.1074/jbc.M410393200 15647275 26. Membrillo-Hernández J Coopamah MD Channa A Hughes MN Poole RK A novel mechanism for upregulation of the Escherichia coli K-12 hmp (flavohaemoglobin) gene by the “NO releaser”, S-nitrosoglutathione: nitrosation of homocysteine and modulation of MetR binding to the glyA-hmp intergenic region Mol Microbiol. 1998 29 1101 12 10.1046/j.1365-2958.1998.01000.x 9767577 27. Maxon ME Redfield B Cai XY Shoeman R Fujita K Fisher W Regulation of methionine synthesis in Escherichia coli: effect of the MetR protein on the expression of the metE and metR genes Proc Natl Acad Sci U S A. 1989 86 85 9 10.1073/pnas.86.1.85 2643109 28. Jafri S Urbanowski ML Stauffer GV A mutation in the rpoA gene encoding the alpha subunit of RNA polymerase that affects metE-metR transcription in Escherichia coli J Bacteriol. 1995 177 524 9 7836282 29. Wu WF Urbanowski ML Stauffer GV Characterization of a second MetR-binding site in the metE metR regulatory region of salmonella typhimurium J Bacteriol. 1995 177 1834 9 7896708 30. Lorenz E Stauffer GV Characterization of the MetR binding sites for the glyA gene of Escherichia coli J Bacteriol. 1995 177 4113 20 7608086 31. Lorenz E Stauffer GV Cooperative MetR binding in the Escherichia coli glyA control region FEMS Microbiol Lett. 1996 137 147 52 10.1111/j.1574-6968.1996.tb08097.x 8998977 32. Harari O del Val C Romero-Zaliz R Shin D Huang H Groisman EA Identifying promoter features of co-regulated genes with similar network motifs BMC Bioinformatics 2009 10 Suppl 4 S1 10.1186/1471-2105-10-S4-S1 19426448 33. Collado-Vides J Salgado H Morett E Gama-Castro S Jiménez-Jacinto V Martínez-Flores I Bioinformatics resources for the study of gene regulation in bacteria J Bacteriol. 2009 191 23 31 10.1128/JB.01017-08 18978060 34. Urbanowski ML Stauffer GV Genetic and biochemical analysis of the MetR activator-binding site in the metE metR control region of Salmonella typhimurium J Bacteriol 1989 171 5620 9 2676984 35. Anjem A Varghese S Imlay JA Manganese import is a key element of the OxyR response to hydrogen peroxide in Escherichia coli Mol Microbiol. 2009 72 844 58 10.1111/j.1365-2958.2009.06699.x 19400769 36. Storz G Tartaglia LA Ames BN The OxyR regulon Antonie van Leeuwenhoek 1990 58 157 61 10.1007/BF00548927 2256675 37. Zheng M Wang X Templeton LJ Smulski DR LaRossa RA Storz G DNA microarray-mediated transcriptional profiling of the Escherichia coli response to hydrogen peroxide J Bacteriol. 2001 183 4562 70 10.1128/JB.183.15.4562-4570.2001 11443091 38. Mongkolsuk S Helmann JD Regulation of inducible peroxide stress responses Mol Microbiol. 2002 45 9 15 10.1046/j.1365-2958.2002.03015.x 12100544 39. Zheng M Aslund F Storz G Activation of the OxyR transcription factor by reversible disulfide bond formation Science. 1998 279 1718 21 10.1126/science.279.5357.1718 9497290 40. Toledano MB Kullik I Trinh F Baird PT Schneider TD Storz G Redox-dependent shift of OxyR-DNA contacts along an extended DNA-binding site: A mechanism for differential promoter selection Cell. 1994 78 897 909 10.1016/S0092-8674(94)90702-1 8087856 41. Tartaglia LA Gimeno CJ Storz G Ames BN Multidegenerate DNA recognition by the OxyR transcriptional regulator J Biol Chem. 1992 267 2038 45 1730735 42. Rhee KY Senear DF Hatfield GW Activation of gene expression by a ligand-induced conformational change of a protein-DNA complex J Biol Chem. 1998 273 11257 66 10.1074/jbc.273.18.11257 9556617 43. Wek RC Hatfield GW Transcriptional activation at adjacent operators in the divergent-overlapping ilvY and ilvC promoters of Escherichia coli J Mol Biol. 1988 203 643 63 10.1016/0022-2836(88)90199-4 3062177 44. Sung YC Fuchs JA Characterization of the cyn operon in Escherichia coli K12 J Biol Chem. 1988 263 14769 75 3049588 45. Lamblin AF Fuchs JA Expression and purification of the CynR regulatory gene product: CynR is a DNA-binding protein J Bacteriol. 1993 175 7990 9 8253686 46. Lamblin AF Fuchs JA Functional analysis of the Escherichia coli K-12 cyn operon transcriptional regulation J Bacteriol. 1994 176 6613 22 7961413 47. Stragier P Richaud F Borne F Patte JC Regulation of diaminopimelate decarboxylase synthesis in Escherichia coli. I. Identification of a lysR gene encoding an activator of the lysA gene J Mol Biol 1983 168 307 20 10.1016/S0022-2836(83)80020-5 6411928 48. Stragier P Danos O Patte JC Regulation of diaminopimelate decarboxylase synthesis in Escherichia coli. II. Nucleotide sequence of the lysA gene and its regulatory region J Mol Biol 1983 168 321 31 10.1016/S0022-2836(83)80021-7 6350601 49. Stragier P Patte JC Regulation of diaminopimelate decarboxylase synthesis in Escherichia coli. III. Nucleotide sequence and regulation of the lysR gene J Mol Biol 1983 168 333 50 10.1016/S0022-2836(83)80022-9 6350602 50. Huerta AM Collado-Vides J Sigma70 promoters in Escherichia coli: specific transcription in dense regions of overlapping promoter-like signals J Mol Biol. 2003 333 261 78 10.1016/j.jmb.2003.07.017 14529615 51. Goethals K Van Montagu M Holsters M Conserved motifs in a divergent nod box of Azorhizobium caulinodans ORS571 reveal a common structure in promoters regulated by LysR-type proteins Proc Natl Acad Sci U S A. 1992 89 1646 50 10.1073/pnas.89.5.1646 1542656 52. Maddocks SE Oyston PC Structure and function of the LysR-type transcriptional regulator (LTTR) family proteins Microbiology. 2008 154 3609 23 10.1099/mic.0.2008/022772-0 19047729 53. Parsek MR Ye RW Pun P Chakrabarty AM Critical nucleotides in the interaction of a LysR-type regulator with its target promoter region. catBC promoter activation by CatR J Biol Chem 1994 269 11279 84 8157659 54. Wang L Winans SC The sixty nucleotide OccR operator contains a subsite essential and sufficient for OccR binding and a second subsite required for ligand-responsive DNA bending J Mol Biol. 1995 253 691 702 10.1006/jmbi.1995.0583 7473744 55. Akakura R Winans SC Mutations in the occQ operator that decrease OccR-induced DNA bending do not cause constitutive promoter activity J Biol Chem. 2002 277 15773 80 10.1074/jbc.M200109200 11877409 56. MacLean AM Haerty W Golding GB Finan TM The LysR-type PcaQ protein regulates expression of a protocatechuate-inducible ABC-type transport system in Sinorhizobium meliloti Microbiology. 2011 157 2522 33 10.1099/mic.0.050542-0 21700663 57. Wang L Winans SC High angle and ligand-induced low angle DNA bends incited by OccR lie in the same plane with OccR bound to the interior angle J Mol Biol. 1995 253 32 8 10.1006/jmbi.1995.0533 7473714 58. Parsek MR McFall SM Shinabarger DL Chakrabarty AM Interaction of two LysR-type regulatory proteins CatR and ClcR with heterologous promoters: functional and evolutionary implications Proc Natl Acad Sci U S A 1994 91 12393 7 10.1073/pnas.91.26.12393 7809047 59. MacLean AM Anstey MI Finan TM Binding site determinants for the LysR-type transcriptional regulator PcaQ in the legume endosymbiont Sinorhizobium meliloti J Bacteriol 2008 190 1237 46 10.1128/JB.01456-07 18055594 60. Hryniewicz MM Kredich NM Stoichiometry of binding of CysB to the cysJIH, cysK, and cysP promoter regions of salmonella typhimurium J Bacteriol. 1994 176 3673 82 8206845 61. McFall SM Klem TJ Fujita N Ishihama A Chakrabarty AM DNase I footprinting, DNA bending and in vitro transcription analyses of ClcR and CatR interactions with the clcABD promoter: evidence of a conserved transcriptional activation mechanism Mol Microbiol 1997 24 965 76 10.1046/j.1365-2958.1997.4041763.x 9220004 62. Gallegos MT Schleif R Bairoch A Hofmann K Ramos JL Arac/XylS family of transcriptional regulators Microbiol Mol Biol Rev 1997 61 393 410 9409145 63. Felsenstein J Evolutionary trees from DNA sequences: a maximum likelihood approach J Mol Evol. 1981 17 368 76 10.1007/BF01734359 7288891 64. Ciccarelli F Doerks T Von MC Toward automatic reconstruction of a highly resolved tree of life Science. 2006 311 1283 7 10.1126/science.1123061 16513982 65. Zhang M Leong HW Bidirectional best hit r-window gene clusters BMC Bioinformatics 2010 11 Suppl 1 S63 10.1186/1471-2105-11-S1-S63 20122239 66. Altschul SF Madden TL Schäffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res. 1997 25 3389 402 10.1093/nar/25.17.3389 9254694 67. Bailey TL Boden M Buske FA Frith M Grant CE Clementi L MEME suite: tools for motif discovery and searching Nucleic Acids Res. 2009 37 W202 8 10.1093/nar/gkp335 19458158

PMC005xxxxxx/PMC5002192.txt

==== Front BMC MicrobiolBMC MicrobiolBMC Microbiology1471-2180BioMed Central London 81410.1186/s12866-016-0814-4Research ArticleSynergism and the mechanism of action of the combination of α-mangostin isolated from Garcinia mangostana L. and oxacillin against an oxacillin-resistant Staphylococcus saprophyticus Phitaktim Sineewan sineewan_2010@yahoo.com 1Chomnawang Mullika dr.mullika_chomnawang@yahoo.co.th 2Sirichaiwetchakoon Kittipot kittipotsirichai@gmail.com 1Dunkhunthod Benjawan dunkhunthod_benjawan@yahoo.com 1Hobbs Glyn g.hobbs@ljmu.ac.uk 3Eumkeb Griangsak griang@sut.ac.th 11 School of Pharmacology, Institute of Science, Suranaree University of Technology, Nakhon Ratchasima, 30000 Thailand 2 Department of Microbiology, Faculty of Pharmacy, Mahidol University, Rajathevi, Bangkok Thailand 3 School of Pharmacy and Biomolecular Sciences, Liverpool John Moores University, Byrom Street, Liverpool, L3 3AF UK 26 8 2016 26 8 2016 2016 16 1 19523 2 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 Globally, staphylococci have developed resistance to many antibiotics. New approaches to chemotherapy are needed and one such approach could be to use plant derived actives with conventional antibiotics in a synergestic way. The purpose of this study was to isolate α-mangostin from the mangosteen (Garcinia mangostana L.; GML) and investigate antibacterial activity and mechanisms of action when used singly and when combined with oxacillin against oxacillin-resistant Staphylococcus saprophyticus (ORSS) strains. The isolated α-mangostin was confirmed by HPLC chromatogram and NMR spectroscopy. The minimum inhibitory concentration (MIC), checkerboard and killing curve were determined. The modes of action of these compounds were also investigated by enzyme assay, transmission electron microscopy (TEM), confocal microscopic images, and cytoplasmic membrane (CM) permeabilization studies. Results The MICs of isolated α-mangostin and oxacillin against these strains were 8 and 128 μg/ml, respectively. Checkerboard assays showed the synergistic activity of isolated α-mangostin (2 μg/ml) plus oxacillin (16 μg/ml) at a fractional inhibitory concentration index (FICI) of 0.37. The kill curve assay confirmed that the viability of oxacillin-resistant Staphylococcus saprophyticus DMST 27055 (ORSS-27055) was dramatically reduced after exposure to isolated α-mangostin (2 μg/ml) plus oxacillin (16 μg/ml). Enzyme assays demonstrated that isolated α-mangostin had an inhibitory activity against β-lactamase in a dose-dependent manner. TEM results clearly showed that these ORSS-27055 cells treated with this combination caused peptidoglycan and cytoplasmic membrane damage, irregular cell shapes and average cell areas were significantly larger than the control. Clearly, confocal microscopic images confirmed that this combination caused considerable peptidoglycan damage and DNA leakage. In addition, the CM permeability of ORSS-27055 was also increased by this combination of actives. Conclusions These findings provide evidence that isolated α-mangostin alone has not only some activity but also shows the synergistic activity with oxacillin against ORSS-27055. The chromone and isoprenyl structures could play a significant role in its action. This synergistic activity may involve three mechanisms of action. Firstly, potential effects of cytoplasmic membrane disruption and increases permeability. Secondly, inhibit β-lactamase activity. Finally, also damage to the peptidoglycan structure. We proposes the potential to develop a novel adjunct phytopharmaceutical to oxacillin for the treatment of ORSS. Future studies require clinical trials to establish if the synergy reported can be translated to animals and humans. Keywords α-mangostinGarcinia mangostanaOxacillinOxacillin-resistant S. saprophyticusSynergistic activityMechanism of actionhttp://dx.doi.org/10.13039/501100004396Thailand Research FundPHD58I0015 Code 5712035Sirichaiwetchakoon Kittipot issue-copyright-statement© The Author(s) 2016 ==== Body Background Antibiotic resistance in staphylococci has been globally documented [1]. The resistance of these strains to board spectraum β-lactam antibiotics, such as methicillin, oxacillin, and flucloxacillin, have emerged rapidly only a few years after the introduction of the first drug in this class and there has been a steady risen in the incidence of methicillin-resistant S. aureus (MRSA) in clinical isolates [1, 2]. A previous study of 12 hospitals in Virginia, USA found that overall 53 % of S. aureus isolates were resistant to oxacillin [3]. In addition, to pathogenic S. aureus, currently S. saprophyticus, a coagulase- negative staphylococcus that frequently causes community-associated uncomplicated urinary tract infection (UTI) in young and middle-aged women, has become resistant to β-lactam antibiotics, such as methicillin by acquisition of staphylococcal cassette chromosome mec (SCCmec) element [4, 5]. These problems present an urgent need to search for new antibiotics and novel approaches to treating these bacterial infections. Plant-derived antimicrobials are a potential source of novel therapeutics because plants are known to produce various antimicrobial molecules to protect themselves from plant or environmental pathogens [6]. Furthermore, drug combination strategies, in particular, phytochemical and antibiotic combination approachs have been recommended in several studies to combat multiple drug-resistant bacteria [7–9]. Mangosteen (the queen of fruit), belonging to the family Guttiferae, is a tropical evergreen tree that is widely cultivated throughout India, Myanmar, Malaysia, Philippines, Sri Lanka, and Thailand [10]. The pericarp (peel, rind, and hull) or the ripe fruit of GML has been traditionally used for the treatment of diarrhea, inflammation, abdominal pain, dysentery, wound infection, suppuration and chronic ulcers [11]. The α-mangostin, a xanthone derivative, has been found to possess several beneficial biological activities, such as a competitive antagonist of the histamine H1 receptor and weak antioxidant properties [12], antibacterial activity against Helicobacter pylori, anti-inflammatory activities, inhibition of oxidative damage by human low-density lipoproteins (LDL), antimicrobial activity against methicillin-resistant Staphylococcus aureus [13]. However, the recent studies have not reported on the antibacterial activity of α-mangostin and synergism with oxacillin against oxacillin-resistant S. saprophyticus. To this end, the present study was instigated to elucidate antibacterial and synergistic activity of α-mangostin isolated from the GML pericarp and oxacillin either alone or in combination against this strain. The antibacterial actions and cell line toxicity of these compounds were also investigated. Methods Plant materials, β-lactam antibiotics, bacterial strains, and cell line The dried fruit hulls of mangosteen were purchased locally in Nakhon-Ratchasima, Thailand. The samples were identified by Dr. Paul J. Grote, Suraneree University of Technology. The voucher specimens (SGM0804U) were deposited in the School of Pharmacology, Institute of Science, Suranaree University of Technology, Nakhon-Ratchasima, Thailand. The mature fruit was cleaned. The fruit rinds were cut into small pieces, dried in a hot oven at 50 °C for 72 h and ground into powder, passed through a sieve (20 mesh). The powdered sample was kept in an airtight container protected from light until used. All clinical isolates of S. saprophyticus were obtained from the Department of Medical Science, National Institute of Health, Ministry of Public Health, Bangkok, Thailand. The susceptible strain S. aureus ATCC 29213, a reference strain, was obtained from the American Type Culture Collection (ATCC). Oxacillin, Nisin, o-nitrophenol-β-D-galactoside (ONPG), α-mangostin standard, and β-lactamase type IV isolated from E. cloacae were obtained from Sigma-Aldrich, UK. Meuller-Hinton broth (MHB) and Mueller-Hinton agar (MHA) were purchased from Oxoid (Basingstoke, UK). The 3T3-L1 mouse embryonic fibroblasts and bovine calf serum (CBS) were purchased from the American Type Culture Collection (ATCC, USA). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), Dulbecco’s Modified Eagle’s medium (DMEM), Fetal bovine serum (FBS), Penicillin-Streptomycin, N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid (HEPES) were obtained from Gibco Invitrogen (Grand Island, NY, USA). Isolation and purification of α-mangostin α-mangostin from the pericarp of GML was isolated and purified according to previous methods with some modifications [12, 14]. Briefly, 1 kg of dried pericarp powder of GML was extracted successively with n-hexane, dichloromethane (CH2Cl2), ethanol, and acetone using a Soxhlet extractor. The filtered extracts were then concentrated using a rotatory evaporator to provide a yellowish power for n-hexane (84 g) and dichloromethane (106 g), brown sticky semi-solid for ethanol (262 g) and acetone (130 g) respectively. The CH2Cl2 crude extract was further separated by silica gel column chromatography to yield 11 fractions. These fractions were subjected to HPLC (C18 column, a mobile phase of methanol-water (85:15) and a flow rate of 1.0 ml/min, a photodiode array detector) and purified using preparative thin layer chromatography to obtain isolated α-mangostin. The α-mangostin was analyzed by HPLC and its chemical structure was elucidated by 1H NMR and 13C NMR. The spectrum structure data of this compound was compared with those previously reported [15]. Standardised bacterial suspensions To select bacterial suspensions with a known viable count, the method of Liu et al. [16] was followed with little modifications. MHA and Cation-adjusted Mueller-Hinton broth (CAMHB) were used as a medium. Minimum inhibitory concentration (MIC) and checkerboard determinations The antibacterial activity and drug interaction of isolated α-mangostin from the pericarp of GML with oxacillin were performed by MIC and checkerboard assays, respectively using broth macrodilution procedure. These assays were conducted following the methods of Clinical and Laboratory Standard Institute guidelines [16, 17]. In summary, 0.25 ml of 5 × 106 cfu/ml bacterial suspensions was added to a series of 2.25 ml CAMHB plus 1 in 10 serial dilutions of the α-mangostin plus oxacillin combinations to give 5x105 cfu/ml. Tubes of the broth without antimicrobialsl were used as the control. The cultures were incubated for 24 h at 37 °C. The tests were carried out in triplicate. MICs were determined for each antibacterial combination and the isobolograms were plotted. The interaction between the two agents was calculated by the fractional inhibitory concentration (FIC) index of the combination. The FIC of each agent was calculated by the complete growth inhibition of microorganism in the combination tube. The following formula was used for FIC index calculation: FIC of α-mangostin = MIC α-mangostin in the combination/MIC of α-mangostin alone; FIC of oxacillin = MIC of oxacillin in the combination/MIC of oxacillin alone; therefore, FIC index = FIC of α-mangostin + FIC of oxacillin. When the FIC index of the combination is equal to or less than 0.5, the combination is defined as synergistic; when the FIC index falls between 0.5 and 4.0, it indicates ‘no interaction’ between the agents and a value above four is considered to show antagonism between the two compounds [18]. S. aureus ATCC 29213 was used as positive control. The MICs and FIC index is presented as the median values obtained in duplicates from three independent experiments. Kill curve determinations (Viable counts) The experiment was carried out to confirm antibacterial and synergistic activities of isolated α-mangostin from the pericarp of GML when used singly and in combination with oxacillin as previously described by Mun et al. and Richards et al. [19, 20]. Compounds were used at the half minimal inhibitory concentration (1/2-MICs) when each compound was assessed alone. However, to study the effect of the compounds in combination, each compound was used at the MIC that yielded synergism. Enzyme assays The ability of isolated α-mangostin from the pericarp of GML to inhibit the activity of β-lactamase type IV isolated from E. cloacae was determined in accordance with the methods of Eumkeb et al. and Richards et al. [21, 22]. Concisely, benzylpenicillin, a substrate for β-lactamase type IV, was adjusted to concentrations sufficient to hydrolyze 50-60 % substrate within 5 min, β-lactamase at 100 μg/ml was used. The α-mangostin at 1, 2, 4 and 8 μg/ml were preincubated with the enzyme in 50 mM sodium phosphate buffer (pH 7.0) at 37 °C for 5 min before adding a substrate. A time - course assay was performed at 0, 5, 10, 15 and 20 min using methanol/acetic acid (100:1) as a stopping agent. Aliquots (10 μl) of each sample were injected onto a reverse-phase HPLC (Ascentis C18 column) to analyse the remaining benzylpenicillin. The mobile phase consisted of 10 mM ammonium acetate (pH 4.5 acetic acid): acetonitrile (75:25) with a flow rate of 1 ml/min, UV detection of peaks was at 200 nm,, and the column maintained at 35 °C. The quantity of remaining benzylpenicillin was calculated by comparing the area under the chromatographic curve. Transmission electron microscopy (TEM) To determine the ultrastructure morphology of bacteria after treatment with isolated α-mangostin from the pericarp of GML either alone or in combination with oxacillin, the method of Richards et al. [22] was followed. To investigate the mechanism of action of these agents, the half-MICs of both compounds used alone and Sub-FICs of the combination, were chosen for examination. To confirm the effects of these agents either used singly and in combination on cell size, the cell area from micrographs were analyzed by measuring cell width multiplied by cell length (nm2). The experiment was performed in triplicate, and the cell areas are displayed as mean ± SEM [23]. Immunofluorescence staining and confocal microscopy The disruption of peptidoglycan after exposure to α-mangostin either used singly or in conjunction with oxacillin was performed using immunofluorescence and visualized under a confocal laser scanning microscope following the method of Teethaisong et al. [24]. Shortly, after the FIC index was obtained from checkerboard, the half-MICs value of isolated α-mangostin or oxacillin alone and the 3/4 FIC of this combination that showed synergistic FIC index was chosen for examination. The cells grown without any antibacterial agent were employed as control [25]. Cytoplasmic membrane (CM) permeabilization assays Two methods were used to assess CM permeabilization. Firstly; the CM permeabilization experiment was performed, with some modifications, to confirm results as previously described by Shen et al. and Zhou et al. [26, 27]. Shortly after the FIC index was determined by the checkerboard assay, the half-MIC values for isolated α-mangostin or oxacillin alone, and the 3/4 MIC values for this combination that indicated synergistic FIC index were selected against ORSS to measure CM permeability. This method was performed by measuring the release of UV-absorbing material (Varian’s Cary 100 UV-Vis spectrophotometer, Varian, Inc., California, USA) [21]. Secondly, the α-mangostin-induced permeabilization of the CM of ORSS was determined essentially as recently described [28]. In brief, to assay CM permeabilization, the wells contained 50 μl ONPG plus either half-MIC values for isolated α-mangostin or oxacillin alone and the 3/4 MIC values for this combination that indicated synergistic FIC index were prepared shortly before the experiment. Finally, 50 μl of cell suspension (OD 0.3) was added to the wells to give a final concentration of 100 μg/ml ONPG. After warming to 37 °C the plates were positioned in the plate reader at 37 °C. ONPG uptake and cleavage by β-galactosidase within the cytoplasm was characterized by monitoring absorption over a period of 120 min at 420 nm. Complete permeabilization was induced in the presence of 0.5 μg/ml Nisin as a positive control and wells lacking drugs or isolated α-mangostin test served as a negative control [29, 30]. In vitro cytotoxicity test (MTT assays) The 3T3-L1 preadipocytes were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with high glucose, supplemented with 10 % CBS, 1.5 mg/ml sodium bicarbonate, 100 U/ml penicillin and 100 μg/ml streptomycin until confluent. The cells were maintained at 37 °C in 5 % CO2 and 95 % humidity. The cytotoxic effect of α-mangostin, oxacillin, either alone or in combination on cell proliferation was determined using a tetrazolium dye (MTT) in a colorimetric assay [31]. Briefly, the cells were seeded in a 96-well plate at a density of 5 × 103 cells/well. The cells were allowed to adhere for 48 h and then were treated with various concentrations of three compounds for 24 h. After incubation, the cultured medium was removed, and 0.5 mg/ml of MTT was added. Then, cells were further incubated for 4 h at 37 °C. Formazan crystals formed by viable cells were dissolved in DMSO and absorbance was measured at 540 nm with a microplate spectrophotometer (Benchmark Plus, Bio-Rad, Japan). Statistical analysiss The experiments were carried out in triplicate; data were expressed as mean ± standard error of the mean (SEM). Significant differences in the enzyme assay among each of treated groups at the same time, the cell area of each treated group, CM permeabilization, and MTT assays were analysed by one-way ANOVA followed by Scheffe’s posthoc test. The p < 0.01 was considered as the statistically significant difference. Results and discussion Isolation, purification and identification of α-mangostin The isolated α-mangostin, the percent yield of α-mangostin at 0.016 % (w/w) of dried powder, from the pericarp of GML, was obtained and the chemical structure of α-mangostin from 1H NMR and 13C NMR (Tables 1 and 2) compared to the reference as illustrated in Fig. 1. Also, the results from HPLC chromatograms of this isolated α-mangostin exhibited a major peak of isolated α-mangostin from the pericarp of GML (Fig. 2a) which is practically the same as a major peak of α-mangostin standard (Fig. 2b). The purity (HPLC) of isolated α-mangostin was 98.0 %.Table 1 The 300 MHz 1H NMR (acetone-d6) spectral data of α-mangostin Chemical Shift (δ, ppm) Assignment Chemical Shift (δ, ppm) from reference [15] 13.78 singlet, OH-1 13.72 6.82 singlet, H-5 6.72 6.40 singlet, H-4 6.25 5.27 triplet, H-12, H-17 5.26 4.13 doublet, H-11 4.10 3.80 singlet, 7-OMe 3.78 3.35 doublet, H-16 3.37 2.07 singlet, H-20 1.83 2.05 singlet, H-15 1.82 1.81 singlet, H-14 1.71 1.65 singlet, H-19 1.68 Table 2 The 300 MHz 13C NMR (acetone-d6), spectral data of α-mangostin Chemical Shift (δ, ppm) Assignment Chemical Shift (δ, ppm) from reference [15] 182.81 C-9 181.8 162.92 C-3 161.6 161.40 C-1 160.2 157.39 C-6 155.4 156.23 C-10a 155.2 155.65 C-4a 154.8 144.51 C-7 142.7 138.14 C-8 137.2 131.39 C-13 131.7 124.82 C-17 123.4 123.50 C-12 122.1 112.06 C-8a 111.7 111.00 C-2 109.7 103.63 C-9a 103.1 102.67 C-5 101.6 93.15 C-4 92.4 61.31 7-OCH3 61.2 26.89 C-11 26.3 25.92 C-14 25.7 25.88 C-19 20.7 22.00 C-16 21.3 18.29 C-20 18.1 17.90 C-15 17.7 Fig. 1 Structure of α-mangostin Fig. 2 HPLC chromatograms of α-mangostin: a, Alpha-mangostin (Ext.) = α-mangostin extract from G. mangostana; b, Alpha-mangostin (Std.) = α-mangostin standard MIC and checkerboard determinations The MIC results for isolated α-mangostin, dissolved in 1 % DMSO, from the pericarp of GML, oxacillin, and nisin against all tested S. saprophyticus strains are presented in Table 3. The results indicated that the MICs for isolated α-mangostin, oxacillin, and nisin against all these strains were 8, 128, and 256 μg/ml, respectively. These results indicated that these strains were resistant to oxacillin. Although, the positive control, Staphylococcus aureus ATCC 29213, was susceptible to oxacillin at MIC 0.5 μg/ml [32]. The isolated α-mangostin exhibited some inhibitory effect against these strains. These results are in agreement with the studies of Chomnawang et al. and Iinuma et al. who reported antibacterial activity of bioactive compounds from the pericarp of GML extracts against MRSA, Staphylococcus epidermidis and Propionibacterium strains [13, 33]. Furthermore, previous findings found that the Garcinia mangostana extract exhibited MIC values of 0.039 mg/ml against both Propionibacterium acnes and Staphylococcus epidermidis [34]. The FIC indices for isolated α-mangostin plus oxacillin against all tested S. saprophyticus strains were 0.37. These results indicated that these combinations demonstrated synergistic activity against these strains [7, 18]. These findings suggest that isolated α-mangostin from the pericarp of GML extract not only has some antibacterial activity of their own against these strains but also have the ability to reverse the resistance of such bacterial strains by synergy with oxacillin.Table 3 Minimum inhibitory concentration (MIC) of oxacillin, α-mangostin, and nisin alone and in combination Strains MIC (μg/ml) FIC (μg/ml) FIC index OXA AMT NIS OXA + AMT OXA + AMT S. saprophyticus DMST 27055 128 R 8 256 16 + 2 0.37 SI S. saprophyticus DMST 27058 128 R 8 256 16 + 2 0.37 SI S. saprophyticus DMST 4236 128 R 8 256 16 + 2 0.37 SI S. saprophyticus DMST 4672 128 R 8 256 16 + 2 0.37 SI S. saprophyticus DMST 5057 128 R 8 256 16 + 2 0.37 SI S. saprophyticus DMST 8034 128 R 8 256 16 + 2 0.37 SI S. aureus ATCC 29213P 0.5 S 4 1 N/D N/D S. aureus ATCC 29213 P, was used as a positive control S Susceptible, R Resistant, SI Synergistic interaction, N/D Not determine OXA Oxacillin, AMT α-mangostin, NIS Nisin The MICs are presented as the median values measured from three independent experiments; each experiment was performed in triplicate Kill curve assays The results for the separate and combined effects of isolated α-mangostin from the pericarp of GML and oxacillin on viable counts of ORSS are presented in Fig. 3. The control showed no reduction in the counts of cfu from control inoculum. The viable counts for the cells treated with isolated α-mangostin at 4 μg/ml were rather lower than that of oxacillin at 64 μg/ml (between 1 and 24 h). Clearly, the combination of 2 μg/ml isolated α-mangostin and 16 μg/ml oxacillin greatly decreased the cell count to 1 × 103 cfu/ml after 4 h to 24 h. These results confirmed the checkerboard assay results, which indicated synergistic activity that the combination produced a decrease of ≥ 2 log10 cfu/ml, compared with oxacillin treatment alone [35]. These results are consistent with those of Eumkeb et al. that galangin plus amoxicillin exhibited synergistic activity against penicillin-resistant S. aureus strains at an FIC index of 0.05 [9]. Apart from this, previous findings reported that a synergistic effect using flavonoids and oxacillin against vancomycin-intermediate S. aureus showed 87.5 % synergism with FIC indices between 0.0417 - 0.1333 [36].Fig. 3 The effect of oxacillin, α-mangostin alone or in combination on the viable counts of ORSS-27055. CON = control (drugs free); OXA(64) = 64 μg/ml oxacillin; AMT(4) = 4 μg/ml α-mangostin; OXA(16) + AMT(2) = 16 μg/ml oxacillin plus 2 μg/ml α-mangostin. The values plotted are the means of 4 observations, and the vertical bars indicate the standard errors of the means Enzyme assays The enzyme assay results for the isolated α-mangostin treatment revealed that the levels of benzylpenicillin were significantly higher compared with the controls (p < 0.01). The levels of benzylpenicillin depended on the isolated α-mangostin concentration in a dose-dependent manner (p < 0.01) (Fig. 4). Previous findings found that 82 % of Staphylococcus saprophyticus strains produced β-lactamase [37]. Furthermore, Hirano and Bayer demonstrated that in vivo efficacy of ampicillin plus sulbactam could inhibit oxacillin-resistant Staphylococcus aureus, which commonly produce β-lactamase [38]. In addition, previous findings reported that α-mangostin was found to be active against membrane enzymes of S. mutans UA159 [39]. Furthermore, These results are in agreement with those of Denny et al. and Zhao et al. that galangin and epigallocatechin gallate, both sharing benzene fused rings condensed with pyran rings (Chromone or 1,4-benzopyrone) similar to α-mangostin, inhibited metallo-β-lactamase and penicillinase, respectively [40, 41]. These findings suggest that the isolated α-mangostin from the pericarp of GML extract could be able to inhibit β-lactamase activity. These inhibitory results may be as a consequence of the isolated α-mangostin form the complex with β-lactamase type IV resulting in deactivation of the β-lactamase activity [9]. Also, the increased of benzylpenicillin was observed in a very short time interval (in minutes; Fig. 9). This result correlates well with the rapid killing of a membrane-targeting antimicrobial. Thus, reduction of the β-lactamase activity could be a secondary effect after the membrane is disrupted [42].Fig. 4 The inhibitory activity of α-mangostin against β-lactamase in hydrolysing benzylpenicillin. β-lactamase used from E. cloacae; CON = control (no testing agent); AMT(1) = 1 μg/ml α-mangostin. The graph shows the remaining benzylpenicillin at the same time. Means sharing the same superscript are not significantly different from each other (Scheffe’s test, p < 0.01) Fig. 5 Ultrathin sections of log phase oxacillin-resistant S. saprophyticus DMST 27055 grown in CAMHB containing: a, control (drug-free); b, oxacillin at 64 μg/ml; c, α-mangostin at 4 μg/ml; d, oxacillin at 12 μg/ml plus α-mangostin at 1.5 μg/ml; (Magnification; a, 4000×, bar = 1 μm; b, 6000×, bar = 0.5 μm; c, 10,000×, bar = 0.5 μm; d, 10,000×, bar = 0.3 μm; Inset magnification; a, c, 38,000×; b, 29,000×; d, 10,000×; bar; a, b, c, 200 nm; d, 500 nm) Fig. 6 The effect of either oxacillin or α-mangostin on average cross-section of ORSS-27055 cell areas. CON = control (drugs free); OXA(64) = 64 μg/ml oxacillin; AMT(4) = 4 μg/ml α-mangostin; OXA(12) + AMT(1.5) = 12 μg/ml oxacillin plus 1.5 μg/ml α-mangostin. The mean ± SEM for three replicates are illustrated. Means sharing the same superscript are not significantly different from each other (Scheffe’s test, p < 0.01) Fig. 7 Schematic representation of the results of immunofluorescence and a confocal laser scanning microscope; Samples of oxacillin-resistant S. saprophyticus DMST 27055 after treatment for 4 h with oxacillin, α-mangostin either alone or in combination. CON = control (drugs free); OXA(64) = 64 μg/ml oxacillin; AMT(4) = 4 μg/ml α-mangostin; OXA(12) + AMT(1.5) = 12 μg/ml oxacillin plus 1.5 μg/ml α-mangostin. The cells were stained for DNA with DAPI (blue, DNA) and labelled for peptidoglycan (green, PGC) using respective antibodies. DNA in all groups was localized in the central of the cell and surrounded by a peptidoglycan layer (MERGED). Scale bar = 1 μm Fig. 8 Effects of 260 nm absorbing material (DNA, RNA, and metabolites) in the ORSS-27055 supernatants. These cells were treated with oxacillin, α-mangostin either alone or in combination. CON = control (drugs free); OXA(64) = 64 μg/ml oxacillin; AMT(4) = 4 μg/ml α-mangostin; OXA(12) + AMT(1.5) = 12 μg/ml oxacillin plus 1.5 μg/ml α-mangostin; NIS(0.5) = 0.50 μg/ml nisin. Nisin at 0.50 μg/ml was used as a positive control, and untreated cells were used as a negative control. The mean ± SEM for three replicates are illustrated. Means sharing the same superscript at the same time are not significantly different (Scheffe’s test, p < 0.01) Fig. 9 Time-dependency of oxacillin plus α-mangostin-induced permeabilisation of the cytoplasmic membrane of ORSS-27055. The kinetics of α-mangostin-mediated ONPG passage across the cytoplasmic membrane are monitored as a change in the optical density (OD). The concentration of substrates was 100 μg/ml ONPG. CON = control (drugs free); OXA(64) = 64 μg/ml oxacillin; AMT(4) = 4 μg/ml α-mangostin; OXA(12) + AMT(1.5) = 12 μg/ml oxacillin plus 1.5 μg/ml α-mangostin; NIS(0.5) = 0.50 μg/ml nisin. Nisin at 0.50 μg/ml was used as a positive control, and untreated cells were used as a negative control. Each treatment performed two times in triplicate. The graph shows OD420nm of each treatment at the same time. Means sharing the same superscript at the same time are not significantly different from each other (Scheffe’s test, p < 0.01) TEM The transmission electron micrographs of cells from the log phase of growth of ORSS in the presence of isolated α-mangostin, oxacillin either alone and in combination are presented in Fig. 5. The untreated cells appeared normal, peptidoglycan, and cytoplasmic membranes were clearly seen intact. And the morphology of the cells looked normal (Fig. 5a). The ORSS cells treated with oxacillin are displayed in Fig. 5b. This result showed some disruption to both peptidoglycan and cytoplasmic membrane. The average cross-sectional cell areas of these cells were slightly smaller than the controls, but there was no significant difference (p > 0.01) (Fig. 6). Although, the micrograph of these cells after exposure to isolated α-mangostin alone is shown in Fig. 5c. Clear damage to peptidoglycan and the cytoplasmic membrane was evident. The average cell areas of these cells were significantly bigger than those of controls (p < 0.01) (Fig. 6). Fig. 5d reveals the isolated α-mangostin plus oxacillin-treated cells. These cells displayed the greatest damage to the peptidoglycan and cytoplasmic membrane resulting in leakage of intracellular materials and overall morphological changes. Clearly, these average cell areas were significantly bigger than the control (p < 0.01) (Fig. 6). These results provide evidence that isolated α-mangostin shows stronger activity than oxacillin against this strain at these concentrations. These findings agree with previous findings where the combination of ceftazidime plus galangin led to damage of the cell ultrastructures, the integrity of cell walls and increase in cell size of ceftazidime-resistant S. aureus [9]. Furthermore, this current study shows similarity to that of the work of Koh et al. who reported that α-mangostin caused significant morphological effects to S. aureus (MRSA) including wall damage and cell lysis [42]. Immunofluorescence staining and confocal microscopy The peptidoglycan and DNA-labelled ORSS clearly showed intact coccus-shape and no damage was observed in untreated control cells by confocal laser scanning images (Fig. 7). The cells treated with isolated α-mangostin or oxacillin alone displayed minor peptidoglycan damage and DNA leakage. The combination of these agents caused considerable peptidoglycan damage and DNA leakage compared to controls. The merger of peptidoglycan and DNA images are also shown. These results are in substantial agreement with the TEM study and support a preliminary mechanism of action of this combination being targeted at the peptidoglycan structure. CM permeabilization The CM permeability was measured by examining the release of UV-absorbing materials at 260 nm (Fig. 8). After treatment, ORSS cells with isolated α-mangostin, nisin, and the isolated α-mangostin plus oxacillin combination could induce the release of 260 nm absorbing materials at significantly higher levels compared with the control or oxacillin alone (p < 0.01). The CM permeabilising ability was ranked as follows isolated α-mangostin plus oxacillin > nisin > isolated α-mangostin > oxacillin > control (p < 0.01). These results suggested that the synergistic activity of isolated α-mangostin plus oxacillin resulted in increased cytoplasmic membrane permeability of DNA, RNA, and cellular metabolites [26, 27]. The α-mangostin-induced CM permeabilisation of the ORSS by ONPG uptake results are shown in Fig. 9. The ORSS cytoplasmic membrane was permeabilised much more rapidly by the isolated α-mangostin plus oxacillin compared to other groups. Nisin, which is highly active against the outer membrane, showed CM permeability significantly lower than isolated α-mangostin and oxacillin combination (p < 0.01). These results are consistent with the results of CM permeabilisation with the UV-absorbing material at 260 nm (Fig. 8). In the same way, previous studies found that galangin, which shares the chromone (or 1,4-benzopyrone) structure with α-mangostin, caused CM permeabilisation of S. aureus resulting in potassium loss [9, 43]. Our findings lend support to previous findings that α‐mangostin rapidly disrupted the integrity of the cytoplasmic membrane of MRSA cells, leading to losing of intracellular components in a concentration-dependent manner [42]. The CM permeability results provide evidence that one of the important mechanism of action of isolated α-mangostin is disruption of the cytoplasmic membrane. This disruption in turn leads to deactivation of the β-lactamase activity. Plant-derived antibacterial compounds have weaker antibacterial activity compared to that of synthetic antibiotics. Therefore, synergistic paradigms by combining the conventional antibiotic with phytochemical compounds is proven in several studies to be an effective avenue to treat infectious diseases caused by drug-resistant bacteria [7–9]. Synergistic interaction combats drug-resistant bacteria by achieving multiple synergistic drug targets, interacting with drug-resistant mechanisms of bacteria, and neutralising and eliminating adverse effects [7]. Previous studies found that ceftazidime had synergistic activity with baicalein, luteolin or quercetin, which shared Chromone structure, against Streptococcus pyogenase [23, 44]. In the same way, Eumkeb and co-workers reported that certain β-lactam drugs plus galangin, quercetin or baicalein, which also shared chromone structure, showed synergistic activity against penicillin-resistant S. aureus [9]. Besides, Rukayadi et al. found that Panduratin A, which possesses a benzene ring and an isoprenyl group, displayed an MIC of 1 μg/ml for staphylococcal clinical isolates and generally was more potent than commonly used antimicrobials [45]. Also, molecular dynamic simulations revealed that isoprenyl groups of α-mangostin, which occupy chromone structure and isoprenyl groups, played an important role in penetrating the lipid bilayer of the MRSA membrane resulting in membrane breakdown and increased permeability [42]. These findings provide evidence that the benzene ring and the isoprenyl group of both panduratin A and α-mangostin play a significant role in inhibiting the growth of MRSA strains by s direct interactions with the bacterial membrane [42, 45]. In vitro cytotoxic test (MTT assays) The results of the MTT assays are shown in Fig. 10. The α-mangostin or oxacillin alone initially exhibited cytotoxicity against 3T3-L1 preadipocytes at concentrations 128 and 1024 μg/ml, respectively, which is 16 and 8 times higher than the MICs of α-mangostin and oxacillin against S. saprophyticus respectively (Fig. 10a and b). Moreover, Fig. 10c revealed the combination of α-mangostin and oxacillin at concentrations of 2 and 16 μg/ml, which showed the synergistic effect of antibacterial activity against this strain, had no cytotoxic effect towards 3T3-L1 preadipocytes after 24 h of exposure. Moreover, the four times higher than FICI value, 64 μg/ml oxacillin plus 8 μg/ml α-mangostin, still did not show the cytotoxic effect on 3T3-L1 preadipocytes. Our findings provide evidence that the MICs and higher dosages of agents have not shown a cytotoxic effect on this cell line. Importantly, a greatly desired property of antibacterial compounds is the selective inhibition against bacterial with less cytotoxic effect to normal cells for avoiding side effects to healthy tissues [46]. The previous studies reported that the IC50 of α-mangostin on the MRC-5 cell line is at the concentration of 7.5 μM [47] and 50 μM on 3T3-L1 preadipocyte cell line [48]. These findings imply that α-mangostin used alone or in combination may be useful in developing a novel adjunct phytopharmaceutical to oxacillin for the treatment of ORSS.Fig. 10 The effect of oxacillin, α-mangostin either alone or in combination on 3T3-L1 preadipocytes. 3T3-L1 preadipocytes were treated with: a. oxacillin (0-2048 μg/ml); b. α-mangostin (0-128 μg/ml); and c. oxacillin plus α-mangostin (8 + 1 to 512 + 64 μg/ml), for 24 h. Results are expressed as percentages of cell viability as compared with untreated controls (n = 8). The mean ± SEM for three replicates are illustrated. Means sharing the same superscript are not significantly different from each other (Scheffe’s test, p < 0.01) Conclusions In conclusion, our findings provide evidence that isolated α-mangostin from the pericarp of GML alone has not only some activity against ORSS but also possesses the synergistic activity with oxacillin against this strain. The chromone (or 1,4-benzopyrone) structure and isoprenyl groups of α-mangostin could play an important role in inhibiting this strain. This synergistic activity of isolated α-mangostin plus oxacillin may involve three modes of action of this xanthone. Firstly, potential effects of cytoplasmic membrane disruption and increases permeability. Secondly, inhibition of β-lactamase activity. Finally, peptidoglycan damage. Our findings provide evidence that isolated α-mangostin from the pericarp of GML has a sufficient margin of safety for therapeutic use. Isolated α-mangostin provides potential to develop a useful of novel adjunct phytopharmaceutical to oxacillin for the treatment of ORSS. Future studies should address toxicity tests in animals and humans. Abbreviations CAMHBCation-adjusted Meuller-Hinton broth CMCytoplasmic membrane FICFraction inhibitory concentration FICIFraction inhibitory concentration index GMLGarcinia mangostana L. MHAMeuller-Hinton agar MHBMeuller-Hinton broth MICMinimum inhibitory concentration ONPGO-nitrophenol-β-D-galactoside ORSSOxacillin-resistant Staphylococcus saprophyticus ORSS-20755Oxacillin-resistant Staphylococcus saprophyticus DMST 20755 TEMTransmission electron microscopy. Acknowledgement The authors are highly appreciative of the One Research One Grant (OROG) scholarship from Suranaree University of Technology for assistance in research funds support. Funding The authors are indebted and gratefully to the Thailand Research Fund and The Crystal Biotechnology Co., Ltd. for assistance in research fund support through The Research and Researchers for Industries Ph.D. scholarships (Grant No. PHD58I0015 Code 5712035). Commenting on the report’s findings is the responsibility of the grant recipient, the Thailand Research Fund and the Crystal Biotechnology Co., Ltd. Availability of data and material The raw datasets supporting the conclusions of this article can be made available by emailing the corresponding author. Authors’ contributions SP performed the experiments and wrote the report. MC and GS operated some laboratories, analysed data and gave comments. BD did the cytotoxicity test and interpreted data. GH supervised, suggested, revised the research and polished the manuscript. GE designed the project, supervised the experiments and wrote 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 Not applicable. ==== Refs References 1. Huttner A Harbarth S Carlet J Cosgrove S Goossens H Holmes A Antimicrobial resistance: a global view from the 2013 World Healthcare-Associated Infections Forum Antimicrob Resist Infect Control 2013 2 31 10.1186/2047-2994-2-31 24237856 2. Rukachaisirikul V Arunpanichlert J Sukpondma Y Phongpaichit S Sakayaroj J Metabolites from the endophytic fungi Botryosphaeria rhodina PSU-M35 and PSU-M114 Tetrahedron 2009 65 10590 5 10.1016/j.tet.2009.10.084 3. Var SK Hadi R Khardori NM Evaluation of regional antibiograms to monitor antimicrobial resistance in Hampton Roads, Virginia Ann Clin Microbiol Antimicrob 2015 14 22 10.1186/s12941-015-0080-6 25890362 4. Widerstrom M Wistrom J Ferry S Karlsson C Monsen T Molecular epidemiology of Staphylococcus saprophyticus isolated from women with uncomplicated community-acquired urinary tract infection J Clin Microbiol 2007 45 1561 4 10.1128/JCM.02071-06 17344356 5. Higashide M Kuroda M Omura CT Kumano M Ohkawa S Ichimura S Methicillin-resistant Staphylococcus saprophyticus isolates carrying staphylococcal cassette chromosome mec have emerged in urogenital tract infections Antimicrob Agents Chemother 2008 52 2061 8 10.1128/AAC.01150-07 18362191 6. Hemaiswarya S Kruthiventi AK Doble M Synergism between natural products and antibiotics against infectious diseases Phytomedicine 2008 15 639 52 10.1016/j.phymed.2008.06.008 18599280 7. Wagner H Ulrich-Merzenich G Synergy research: approaching a new generation of phytopharmaceuticals Phytomedicine 2009 16 97 110 10.1016/j.phymed.2008.12.018 19211237 8. Eumkeb G Chukrathok S Synergistic activity and mechanism of action of ceftazidime and apigenin combination against ceftazidime-resistant Enterobacter cloacae Phytomedicine 2013 20 262 9 10.1016/j.phymed.2012.10.008 23218402 9. Eumkeb G Sakdarat S Siriwong S Reversing beta-lactam antibiotic resistance of Staphylococcus aureus with galangin from Alpinia officinarum Hance and synergism with ceftazidime Phytomedicine 2010 18 40 5 10.1016/j.phymed.2010.09.003 21036573 10. Cui J Hu W Cai Z Liu Y Li S Tao W New medicinal properties of mangostins: analgesic activity and pharmacological characterization of active ingredients from the fruit hull of Garcinia mangostana L Pharmacol Biochem Behav 2010 95 166 72 10.1016/j.pbb.2009.12.021 20064550 11. Suksamrarn S Komutiban O Ratananukul P Chimnoi N Lartpornmatulee N Suksamrarn A Cytotoxic prenylated xanthones from the young fruit of Garcinia mangostana Chem Pharm Bull (Tokyo) 2006 54 301 5 10.1248/cpb.54.301 16508181 12. Chairungsrilerd N Takeuchi K Ohizumi Y Nozoe S Ohta T Mangostanol, a prenyl xanthone from Garcinia mangostana Phytochemistry 1996 43 1099 102 10.1016/S0031-9422(96)00410-4 13. Chomnawang MT Surassmo S Wongsariya K Bunyapraphatsara N Antibacterial activity of Thai medicinal plants against methicillin-resistant Staphylococcus aureus Fitoterapia 2009 80 102 4 10.1016/j.fitote.2008.10.007 19022354 14. Sakagami Y Iinuma M Piyasena KG Dharmaratne HR Antibacterial activity of alpha-mangostin against vancomycin resistant Enterococci (VRE) and synergism with antibiotics Phytomedicine 2005 12 203 8 10.1016/j.phymed.2003.09.012 15830842 15. Ee GC Daud S Taufiq-Yap YH Ismail NH Rahmani M Xanthones from Garcinia mangostana (Guttiferae) Nat Prod Res 2006 20 1067 73 10.1080/14786410500463114 17127660 16. Liu IX Durham DG Richards RME Baicalin synergy with β-lactam antibiotics against methicillin-resistant Staphylococcus aureus and other β-lactam-resistant strains of S. aureus J Pharm Pharmacol 2000 52 361 6 10.1211/0022357001773922 10757427 17. Clinical Laboratory Standards Institute Matthew AW Franklin RC William AC Micheal ND George ME David WH Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically Clinical and Laboratory Standards Institute document M07-A9 2012 Pennsylvania Clinical and Laboratory Standards Institute 16 34 18. Marques MB Brookings ES Moser SA Sonke PB Waites KB Comparative in vitro antimicrobial susceptibilities of nosocomial isolates of Acinetobacter baumannii and synergistic activities of nine antimicrobial combinations Antimicrob Agents Chemother 1997 41 881 5 9145838 19. Richards RM Xing JZ Gregory DW Marshall D An electronmicroscope study of the effect of sulphadiazine and trimethoprim on Enterobacter cloacae J Med Microbiol 1993 38 64 8 10.1099/00222615-38-1-64 8418297 20. Mun SH Joung DK Kim YS Kang OH Kim SB Seo YS Synergistic antibacterial effect of curcumin against methicillin-resistant Staphylococcus aureus Phytomedicine 2013 20 714 8 10.1016/j.phymed.2013.02.006 23537748 21. Eumkeb G Siriwong S Thumanu K Synergistic activity of luteolin and amoxicillin combination against amoxicillin-resistant Escherichia coli and mode of action J Photochem Photobiol B 2012 117 247 53 10.1016/j.jphotobiol.2012.10.006 23159507 22. Richards R Xing J Gregory D Marshall D Mechanism of sulphadiazine enhancement of trimethoprim activity against sulphadiazine-resistant Enterococcus faecalis J Antimicrob Chemother 1995 36 607 18 10.1093/jac/36.4.607 8591935 23. Siriwong S Thumanu K Hengpratom T Eumkeb G Synergy and mode of action of ceftazidime plus quercetin or luteolin on Streptococcus pyogenes Evid Based Complement Alternat Med 2015 2015 12 10.1155/2015/759459 24. Teethaisong Y Autarkool N Sirichaiwetchakoon K Krubphachaya P Kupittayanant S Eumkeb G Synergistic activity and mechanism of action of Stephania suberosa Forman extract and ampicillin combination against ampicillin-resistant Staphylococcus aureus J Biomed Sci 2014 21 90 10.1186/s12929-014-0090-2 25208614 25. Tocheva EI Matson EG Morris DM Moussavi F Leadbetter JR Jensen GJ Peptidoglycan remodeling and conversion of an inner membrane into an outer membrane during sporulation Cell 2011 146 799 812 10.1016/j.cell.2011.07.029 21884938 26. Shen L Liu D Li M Jin F Din M Parnell LD Mechanism of action of recombinant acc-royalisin from royal jelly of Asian honeybee against gram-positive bacteria PLoS One 2012 7 e47194 10.1371/journal.pone.0047194 23056609 27. Zhou K Zhou W Li P Liu G Zhang J Dai Y Mode of action of pentocin 31-1: An antilisteria bacteriocin produced by Lactobacillus pentosus from Chinese traditional ham Food Control 2008 19 817 22 10.1016/j.foodcont.2007.08.008 28. Eriksson M Nielsen PE Good L Cell permeabilization and uptake of antisense peptide-peptide nucleic acid (PNA) into Escherichia coli J Biol Chem 2002 277 7144 7 10.1074/jbc.M106624200 11739379 29. Junkes C Wessolowski A Farnaud S Evans RW Good L Bienert M The interaction of arginine- and tryptophan-rich cyclic hexapeptides with Escherichia coli membranes J Pept Sci 2008 14 535 43 10.1002/psc.940 17985396 30. Junkes C Harvey RD Bruce KD Dolling R Bagheri M Dathe M Cyclic antimicrobial R-, W-rich peptides: the role of peptide structure and E. coli outer and inner membranes in activity and the mode of action Eur Biophys J 2011 40 515 28 10.1007/s00249-011-0671-x 21286704 31. Chun SC Jee SY Lee SG Park SJ Lee JR Kim SC Anti-inflammatory activity of the methanol extract of moutan cortex in LPS-activated Raw264.7 cells Evid Based Complement Alternat Med 2007 4 327 33 10.1093/ecam/nel093 17965763 32. Clinical and Laboratory Standards Institute Performance Standards for Antimicrobial Susceptibility Testing; Twenty-Third Informational Supplement. CLSI document M100-S23 2013 23 Wayne, Pennsylvania Clinical and Laboratory Standards Institute 33. Iinuma M Tosa H Tanaka T Asai F Kobayashi Y Shimano R Antibacterial activity of xanthones from guttiferaeous plants against methicillin-resistant Staphylococcus aureus J Pharm Pharmacol 1996 48 861 5 10.1111/j.2042-7158.1996.tb03988.x 8887739 34. Chomnawang MT Surassmo S Nukoolkarn VS Gritsanapan W Antimicrobial effects of Thai medicinal plants against acne-inducing bacteria J Ethnopharmacol 2005 101 330 3 10.1016/j.jep.2005.04.038 16009519 35. Eliopoulos GM Moellering RC Lorian VS Antimicrobial combinations Antibiotic in Laboratory Medicine 1996 4 Baltimore Williams and Wilkins 330 96 36. Basri DF Zin NM Bakar NS Rahmat F Mohtar M Synergistic effects of phytochemicals and oxacillin on laboratory passage-derived vancomycin -intermediate Staphylococcus aureus strain J Med Sci 2008 8 131 6 10.3923/jms.2008.131.136 37. Latham RH Zeleznik D Minshew BH Schoenknecht FD Stamm WE Staphylococcus saprophyticus beta-lactamase production and disk diffusion susceptibility testing for three beta-lactam antimicrobial agents Antimicrob Agents Chemother 1984 26 670 2 10.1128/AAC.26.5.670 6335018 38. Hirano L Bayer A Beta-Lactam-beta-lactamase-inhibitor combinations are active in experimental endocarditis caused by beta-lactamase-producing oxacillin-resistant staphylococci Antimicrob Agents Chemother 1991 35 685 90 10.1128/AAC.35.4.685 2069374 39. Nguyen PT Marquis RE Antimicrobial actions of alpha-mangostin against oral streptococci Can J Microbiol 2011 57 217 25 10.1139/W10-122 21358763 40. Denny BJ Lambert PA West PWJ The flavonoid galangin inhibits the L1 metallo-β-lactamase from Stenotrophomonas maltophilia FEMS Microbiol Lett 2002 208 21 4 11934488 41. Zhao WH Hu ZQ Hara Y Shimamura T Inhibition of penicillinase by epigallocatechin gallate resulting in restoration of antibacterial activity of penicillin against penicillinase-producing Staphylococcus aureus Antimicrob Agents Chemother 2002 46 2266 8 10.1128/AAC.46.7.2266-2268.2002 12069986 42. Koh JJ Qiu S Zou H Lakshminarayanan R Li J Zhou X Rapid bactericidal action of alpha-mangostin against MRSA as an outcome of membrane targeting Biochim Biophys Acta 1828 2013 834 44 43. Cushnie TP Lamb AJ Detection of galangin-induced cytoplasmic membrane damage in Staphylococcus aureus by measuring potassium loss J Ethnopharmacol 2005 101 243 8 10.1016/j.jep.2005.04.014 15985350 44. Siriwong S Pimchan T Naknarong W Eumkeb G Mode of Action and Synergy of Ceftazidime and Baicalein against Streptococcus pyogenes Trop J Pharm Res 2015 14 641 8 10.4314/tjpr.v14i4.12 45. Rukayadi Y Lee K Han S Yong D Hwang JK In vitro activities of panduratin A against clinical Staphylococcus strains Antimicrob Agents Chemother 2009 53 4529 32 10.1128/AAC.00624-09 19651906 46. Hou Z Zhou Y Li J Zhang X Shi X Xue X Selective in vivo and in vitro activities of 3,3'-4-nitrobenzylidene-bis-4-hydroxycoumarin against methicillin-resistant Staphylococcus aureus by inhibition of DNA polymerase III Sci Rep 2015 5 13637 10.1038/srep13637 26323712 47. Al-Massarani SM El Gamal AA Al-Musayeib NM Mothana RA Basudan OA Al-Rehaily AJ Phytochemical, antimicrobial and antiprotozoal evaluation of Garcinia mangostana pericarp and alpha-mangostin, its major xanthone derivative Molecules 2013 18 10599 608 10.3390/molecules180910599 24002136 48. Quan X Wang Y Ma X Liang Y Tian W Ma Q α-Mangostin induces apoptosis and suppresses differentiation of 3T3-L1 cells via inhibiting fatty acid synthase PLoS One 2012 7 e33376 10.1371/journal.pone.0033376 22428036

PMC005xxxxxx/PMC5002193.txt

==== Front J Cardiovasc Magn ResonJ Cardiovasc Magn ResonJournal of Cardiovascular Magnetic Resonance1097-66471532-429XBioMed Central London 27410.1186/s12968-016-0274-xResearchPhysicians’ professional identities: a roadmap to understanding “value” in cardiovascular imaging Keller Eric J. eric.keller@northwestern.edu 1Vogelzang Robert L. vogelzang@northwestern.edu 1Freed Benjamin H. bfreed@nm.org 2Carr James C. jcarr@northwestern.edu 1Collins Jeremy D. (312) 695-2422collins@fsm.northwestern.edu 11 Department of Radiology, Northwestern University Feinberg School of Medicine, 737 N. Michigan Ave Suite 1600, Chicago, IL 60611 USA 2 Department of Medicine-Cardiology, Northwestern University Feinberg School of Medicine, Chicago, IL USA 26 8 2016 26 8 2016 2016 18 1 526 5 2016 12 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 Quality improvement efforts in cardiovascular imaging have been challenged by limited adoption of initiatives and policies. In order to better understand this limitation and inform future efforts, the range clinical values related to cardiovascular imaging at a large academic hospital was characterized. Materials and methods 15 Northwestern Medicine physicians from internal medicine, cardiology, emergency medicine, cardiac/vascular surgery, and radiology were interviewed about their use of cardiovascular imaging and imaging guidelines. Interview transcripts were systemically analyzed according to constructivist grounded theory and combined with 56 previous interviews with interventional radiologists, interventional cardiologists, gynecologists, and vascular surgeons to develop a model describing specialty-specific values. This model was applied to the 15 pilot interviews focused on cardiovascular imaging, highlighting specialty specific differences in values and practice patterns. Transcripts were also reviewed independently by a cardiologist and 2 radiologists followed by a group discussion to assess reproducibility and achieve a consensus regarding the results. Results Differences in perceived value of cardiovascular imaging and use of guidelines among physicians were well explained by three value-associated identity categories (managers, diagnosticians, and fixers) that were further differentiated along three axes (broad v. focused-thinkers, complex v. definitive-answer-seekers, and public visibility). Conclusions Quality improvement in cardiovascular imaging may be limited by a lack of understanding and incorporation of the complexity of medical culture into ongoing initiatives. Both individually and during policy development, it is important to first understand the complexity of stakeholders’ diverse perceptions of “value,” “quality,” and “appropriateness.” Keywords Quality improvementPerceived valueProfessional identityCardiovascular imaginghttp://dx.doi.org/10.13039/100006098Radiological Society of North AmericaResearch Medical Student Grant #1521Keller Eric J. issue-copyright-statement© The Author(s) 2016 ==== Body Background Cardiovascular imaging has been identified by the National Quality Forum and Institute of Medicine as an important area for quality improvement, both economically and socially [1, 2]. Past initiatives in this area of healthcare, like others, have yielded some promising results but remain challenged by a lack of sustainability and physician support. For example, a recent meta-analysis [3] found that physician audit, feedback, and education about the American College of Cardiology’s (ACC) Appropriate Use Criteria (AUC) significantly reduced inappropriate use of cardiac imaging. However, other studies that sought to educate and provide feedback about AUC failed to impact behavior, finding that autonomous professionals “don’t want to be told what to do” [4, 5]. Even those studies reporting initial success have found “inappropriate” scan rates to return to pre-intervention rates in as little as one year [6, 7]. When such interventions fail, it can be easy to assume that other physicians do not care or are acting selfishly. Consider Andy Slavitt’s comment regarding the abrupt end of the Center of Medicaid and Medicare Services’ (CMS) Meaningful Use initiative, “we have to get the hearts and minds of physicians back. I think we’ve lost them” [8]. The dominant approach has been to “educate” and/or financially incentivize (through threat or reward) behaviors that groups of experts have defined as appropriate or cost-effective [9]. However, this approach assumes that most physicians share common definitions of appropriateness and value and are primarily driven by economic gains. Since American medicine lacks much of a central authority beyond the government’s involvement in health insurance [10], medical authority is spread across a complex array of professional societies, guidelines, and training experiences, creating and propagating distinct truths and values [11]. Additionally, although physicians value reimbursement, the highest paid physician is not necessarily valued or respected most by his/her colleagues. Physicians are driven by autonomy, mastery, and a sense of purpose, and if a system incentivizes actions that undermine these values, it can “sap” physicians of their internal drive and lead to poor policy adoption [12–14]. Simply put, drivers of professional behavior and perceptions of value are complex [15], and without a better understanding and appreciation of this complexity, even well-intentioned interventions can fail to resonate with those they aim to affect. In other words, it is not a lack of values or interests that limits physicians’ adoption of guidelines and interventions but the range of values and experiences that are present. Humans’ tendency to form social groups and distinct values is a well described concept in psychology called social identity theory [16]. Past work has illustrated that medical specialties (and any academc discipline) exhibit distinct vocabulary, relics, idols, and thus, cultures that can cause miscommunication and tension [17–20]. However, much of healthcare policy and guideline development has relied on a monocultural view of physicians. This may be due to the fact that the American Medical Association (AMA) and allopathic medicine in general have long emphasized common physician values, objective scientific evidence, and self-regulation [21], leading many to falsely assume that all physicians hold the same values and interpretations of evidence despite perceived personality differences. Few would disagree that an internist is different from a surgeon, but exactly how these professional groups differ in terms of clinical values and decision making remains largely unexplored and underappreciated in healthcare. Cardiovascular imaging is particularly vulnerable to cultural differences between specialty groups because a significant portion of cardiovascular imaging is performed and interpreted by non-radiologists and the decision to order a particular imaging test may or may not be made by the same speciality responsible for performing and interpreting that imaging. Although surveys and cost-utilization comparisons are useful and easier to collect and analyze, previous anthropologic studies exploring the complexity of clinical decision making and values have required a more sensitive qualitative method called “grounded theory (GT).” GT is a well-validated method for systemically exploring poorly understood social processes through the simultaneous generation and analysis of qualitative data that has been used since the 1960s to understand a wide array of patient and/or physician perceptions [22–26]. However, some applications of GT can be quite time-intensive and laborious, limiting their clinical feasibility. Thus, we used what is called Constructivist GT (C-GT) to facilitate a preliminary investigation of differences in the perceived value of CV imaging between specialties. C-GT expedites traditional GT by not approaching the investigation with a completely blank slate (e.g. knowing we are most interested in perceptions of value) and omitting some analysis steps after main results become apparent [27, 28]. Methods Research strategy In order to efficiently illustrate value differences, we selected 3 physicians in each of 5 specialties involved in cardiovascular (CV) imaging [internal medicine (IM), cardiology, radiology, emergency medicine (EM), and cardiac/vascular surgery] at a single institution (Northwestern Medicine). Often C-GT requires sample sizes of at least 6–12 to capture the majority of important concepts in each group [29]; however, this pilot investigation sought to merely illustrate the range of values rather than definitively characterize them. Semi-structured interview design Interviews were conducted by a medical student with 2 years experience conducting semi-stractured interviews with physicians (EJK). This approach facilitated a conversational tone and reduced filtered answers due to the interview style and interviewer’s status as a non-threatening member of the medical community [30]. Interviews were introduced as a pilot study to understand how different specialties “use” CV imaging. Specifically, “use” meant any interaction with imaging as part of their professional role, including but not limited to ordering, interpreting, and discussing results with patients. This general word was used to not limit interviewees’ descriptions of their interactions with CV imaging. Physicians were first asked to describe their professional role and use of CV imaging through typical patient interactions. Follow up questions, such as simply “why” or “can you give me an example,” were used to gather specific details and a richer understanding of their clinical reasoning. The interviewer also summarized concepts back to interviewees to confirm understanding, differentiate more vs less important concepts, and establish rapport. Interviewees were then asked more sensitive questions about how their approach compared to others within and outside their field, their use/opinions of CV imaging guidelines, and possible solutions to any expressed concerns with further follow up questions. Finally, physcians were asked what they valued about their specialty and if there was anything else they felt was important that was not discussed. Data analysis All interviews were transcribed verbatim and systematically analyzed according to C-GT [27], using NVivo 10 (QSR International). Key concepts were identified by considering interviewees’ emphases and frequencies of ideas. For example, an initial EM physician interview yielded concepts such as undifferentiated patients, considering worse case scenarios, gatekeeper, efficiency/triage, and imaging to avoid malpractice concerns. Interview concepts from these 15 interviews were compared within and across specialty groups along with 56 other interviews conducted with interventional radiologists, interventional cardiologists, gynecologists, and vascular surgeons about their choice of specialty and approaches to patient care. These additional interviews were conducted in the same style by the same interviewer during the last 2 years. Similar concepts were present in each specialty group but distinct from others. For example, concepts such as undifferentiated patients and gatekeeper were dominant throughout all 3 EM interviews but not present in any of the 68 other physicians interviews we have conducted to date. In light of this, interviews were reassessed to characterize common themes that differentiated physicians’ values and clinical reasoning. Three dominant value-associated identity categories (managers, diagnosticians, and fixers) and three differentiating axes (broad v. focused-thinkers, complex v. definitive-answer-seekers, and public visibility) emerged from comparing and contrasting the values and themes expressed by each specialty. This coding structure was then applied to previous and remaining interviews to test how well it could explain differences in interviewees’ use of CV imaging and guidelines. Transcripts were also reviewed idependently by a cardiologist (BHF) and 2 radiologists (JDC, RLV) followed by a group discussion to assess reproducibility and achieve a consensus regarding the results. Results Participants All but 1/15 pilot physician interviewees were men in practice between 3 and 41 years. With the addition of the 56 previous physicians interviews, 15/71 (21 %) were women; the practice experience was 1–41 years, with 16/71 (23 %) private practice physicians and 22/71 (31 %) physicians from practices outside Northwestern in CA, AR, OH, WI, and NC. Although the identified professional identity groups and values have been grouped below by specialty, the interviews revealed differing degrees of variation within professional groups. In general, smaller, more-focused groups such as interventional radiologists tended to exhibit less variation in their expressed values and opinions than broader groups such as “obstetricians/gynecologists.” Characterizing and comparing the degrees of intra-specialty value variation regarding CV imaging would require additional interviews. Value-associated identity categories “Manager” physicians (e.g. internists and cardiologists) tended to value thinking broadly about many different factors contributing to patients’ symptoms to develop differential diagnoses. Patients’ stories and clinical context were particularly important, which they use to “manage” patients or conditions over relatively long clinical relationships. “I'm probably like a lot of internists in that we enjoyed the story, hearing the story and trying to use that to make a diagnosis and implement treatment… [We] recognize that you can’t just separate systems, a person’s physical health is connected in many ways across a lot of the systems as well as their psychological and social elements….” In regards to CV imaging, managers tended to view imaging as one piece of information that needs to be considered in light of and incorporated into the clinical story. Internists described using EKGs and stress tests along with histories and physical exam findings to understand best how to manage patients acutely or chronically. Thus, the value of the imaging was simply part of the larger clinical puzzle, “maybe 10 %.” Although some internists feel comfortable ordering cardiac CTs or MRIs, they preferred cardiologists as both their CV consultants and imagers when patients were high risk or needed advanced imaging:“…there has been debate on the front lines of internal medicine that when you send a patient for cardiovascular imaging who’s the best person to get that result from. There tends to be a fairly strong consensus that they would prefer a cardiologist to offer their opinion than a radiologist… I have a sense of what a cardiologist may do or the interest they may take and the training they’ve had versus a radiologist who I expect has probably had a lot less clinical experience with someone with heart disease and probably a lot less exposure to cardiovascular investigation in their training.” –Internist Medical subspecialties such as cardiology shared much of the internist identity: “[As a cardiologist] you are an internist.” Although cardiology is a more focused field, it was defined more in terms of clinical knowledge than anatomy or a technical skill set, e.g., cardiologists described themselves as managing CV diseases/patients rather than fixing hearts. Nevertheless, cardiologists described themselves as valuing procedures and imaging more than other internists, making their use of CV imaging more focused and definitive: they more often described imaging as providing definitive answers and playing a larger role in clinical care. Cardiologists particularly valued advanced imaging such as echocardiography or cinegraphic CMR that let them better appreciate the structure-function relationships that piqued their interest in cardiology as trainees. Their roles as CV imagers varied, but they tended to highly value access to and collaboration with radiologists for imaging interpretation: “My interactions with the radiologists are great. I love their insight… from what I hear it’s mutual.” Nevertheless, it was noted that this is not always the case and can cause tension.“Diagnostician” identities (e.g. radiology and EM) tended to enjoy “knowing something about everything” and getting to solve diagnostic puzzles during relatively short interactions. However, this makes them particularly dependent on other specialties, and so they often felt that they must continually prove their worth to other physicians: “…this concession on [other clinicans’] part is regarded only as a loan of a talent to be developed by us and then returned, is witnessed by the alacrity with which the [medical specialist] takes over the application of x-ray and radium the moment he can make such a move remunerative.” –Radiologist (1920) [31] For radiologists, the value of CV imaging was a definitive answer to the clinical problem at hand, quantifying and characterizing diameters, motions, and intensities. They also had a particular appreciation for adjustments that could make modalities faster, safer, or more objective, which they suspected other specialties also valued about them. EM physicians described themselves as unique in seeing “undifferentiated patients” and thinking in terms of worst case scenarios and efficiency for patients waiting to be seen. Thus, the value of CV imaging was defined in terms of triaging patients. EKGs, CXRs, and sometimes coronary CTAs were valued for efficiently determining if patients should be sent home or admitted to cardiology for further workup. Overall, EM physicians seemed to have a lower threshold for imaging, which they attributed to malpractice concerns as gatekeepers between an organization pushing them to lower costs and patients who expect a 0 % miss rate. They also felt that many general practitioners now rely on them as consultants, “…we’ve become the master clinicians in diagnosis;” yet, like other diagnosticians, EM physicians felt underappreciated by many specialists: “No other specialist will ever tell you this but what they’re really looking for is an emergency medicine consult.”“Fixer” physicians (e.g. cardiac/vascular surgeons) valued tangible outcomes, “doing” technical tasks, and innovation. “I like to see immediate results. I like to work with my hands.” Because of this, they defined their roles primarily by their procedures and/or anatomic regions: “I like to operate on the heart. It’s what defines cardiac surgery….” Due to this reliance on technical skill and ability, value was primarily defined in terms of technical success of fixing specific problems. For example, CV imaging was valued primarily for determining whether or not a procedure was indicated, planning procedures, and assessing the success of procedures. They tended to trust radiologists to provide more consistent measurements and generate official reports but preferred to use 3D reconstructions and ultrasound/fluoroscopy themselves for surgical planning and intraoperative guidance, respectively. Broad v. Focused-thinkers Beyond the three categories, physicians’ professional identities were further divided along three key axes, affecting their clinical values and decision making (Fig. 1). Broad-thinkers (e.g. IM, radiology, EM, and cardiologists) valued knowing about many different areas of medicine and casting a broad net whereas focused-thinkers (e.g. surgeons) prioritized narrowing in on a single issue they could control and address. Interestingly breadth of knowledge seemed to garner less respect among colleagues from other specialties than depth. In terms of imaging, focused-thinkers tended to view CV imaging as a tangible means to an ends, whereas broad-thinkers described imaging as an investigation that could lead to a larger number of valuable endpoints.Fig. 1 Key axes of physicians’ professional identities and values Complex v. Definitive-answer-seekers Physicians that pursue complex answers (e.g. internists, some cardiologists) viewed clinical decision making as more complex and nuanced, often fluid and changing with many caveats and few absolutes. These specialists stressed the importance of clinical context and the interconnections between various aspects of patients’ lives and organ systems. Thus, imaging findings must be accompanied with clinical context to have substantial meaning and rarely offer a definitive conclusion alone.“I see a person with a heart murmur who also says they’re short of breath. The first question is whether the shortness of breath related to the heart murmur? Or is the heart murmur just incidental to why they have this shortness of breath? …until you understand all the elements of why they have those symptoms and what their cardiac status is you can’t even bring [valve repair] up.” –Cardiologist (2016) Conversely, definitive-answer-seekers (e.g. radiologists, surgeons, and some cardiologists) tended to draw hard lines while describing their clinical reasoning and use of CV imaging and guidelines. A procedure or test was indicated or not; outcomes were superior or inferior. Not everything was absolute, but too much “gray area” or subjectivity was undesirable. Complex-answer-seekers were far more comfortable with subjectivity and qualitative imaging interpretation whereas surgeons and radiologists prioritized numbers and measurements. Public visibility The final key axis was public visibility, differentiating internists and surgeons from radiologists. In general, this axis paralleled the amount of patient interaction and thus the likelihood of forming clinical relationships with patients. “My radiologist” is rarely used. Interestingly, this axis also seemed to explain a degree of perceived value. For example, there was a clear sense that a cardiologist who can interpret imaging was likely to be valued more by healthcare administration than a radiologist who can counsel patients about cardiovascular disease. Cardiologists’ roles in managing cardiovascular disease has a higher degree of public visibility than radiologists’ roles in interpreting imaging, which attracts wider appreciation for the complexity of this skill. Consider that a radiologist counselling a patient about his/her cardiovascular disease would be met with more skepticism than a cardiologist interpreting a patients’ cardiac MRI. This axis also differentiated EM physicians from radiologists. EM shares the diagnostician need to prove themselves to other physicians, but EM physicians do not have to prove their worth publicly:“Yeah, one of the funny things about being an ER doc is that if you’re at a cocktail party, everybody thinks that it’s the coolest thing you can do with your life, and if you’re in a room full of other physicians, everybody thinks that you’re a total jackass.” –EM physician (2016) See Table 1 for summary of identity characteristics and views of CV imaging.Table 1 Summary of specialty differences Specialty Identity characteristics Identity descriptions Value of CV imaging Internal Medicine ◊ Manager ◊ Very broad thinker ◊ Complex answers ◊ High public visibility ◊ Patient narrative ◊ Relationship variety ◊ Connecting with patients ◊ Thinking broadly ◊ One piece of the clinical puzzle ◊ “Maybe 10 % of information” ◊ Mainly EKG, stress imaging ◊ Prefer cardiology for further imaging and evaluation Cardiology ◊ Manager ◊ Broad thinker ◊ Complex/Definitive answers ◊ High public visibility ◊ Structure-function relationships ◊ Procedures/Imaging ◊ Prevention ◊ Managing patients ◊ Diagnosis and management ◊ Prioritize structure-function relationship, e.g., echo, cinegraphic CMR ◊ Collaborative imaging interpretation with radiology Emergency Medicine ◊ Diagnostician ◊ Broad thinker ◊ Definitive answers ◊ Moderate public visibility ◊ Efficiency ◊ Undifferentiated patients ◊ Gatekeepers ◊ Variety of patients ◊ Triage patients / Rule out worst case scenario ◊ Mainly EKG, CXR, echo, maybe coronary CT ◊ Refer to cardiology for further testing Radiology ◊ Diagnostician ◊ Broad thinker ◊ Definitive answers ◊ Low public visibility ◊ Technology ◊ Innovation ◊ Knowing something about everything ◊ Consultant ◊ Provide objective answer to clinical question and make patient better ◊ Minimal “gray area” ◊ Right imaging for the right patient ◊ Collaborative imaging interpretation with cardiology Vascular/Cardiac Surgery ◊ Fixer ◊ Focused thinker ◊ Definitive answers ◊ Moderate public visibility ◊ Working with hands ◊ Technical work with immediate outcome ◊ Ability to do/fix something ◊ To determine appropriateness of surgery, surgical planning, surgical follow up ◊ Prioritize 3D CTA or intraoperative ultrasound/fluoroscopy ◊ Radiology does official report but we also interpret imaging Use of imaging guidelines Identity differences also seemed to predict perceptions of imaging guidelines as well as the content and syntax of guidelines themselves. Radiologists tended to view ACR’s appropriateness criteria positively and felt their use of imaging was used to create those guidelines rather than the other way around. However, EM physicians and internists found these criteria less helpful for triaging patients and guiding management, respectively. They instead preferred guidelines from the ACP or ACC and imaging reports from cardiologists. When asked why, both groups mentioned that their time was limited and they did not get a clear sense of how to incorporate the criteria into their clinical thinking. It was not so much that they viewed other specialty guidelines as incorrect but that these guidelines were not as relatable to their professional roles and goals. Cardiologists and surgeons felt they were familiar with ACR’s criteria but cardiologists said they would prioritize the ACC if recommendations differed and surgeons felt these criteria often did not apply to their use of imaging. Again, these perceptions appear to be a reflection of relatability and trust of one’s group. Even in comparing guidelines themselves one can find the same professional value differences. For example, ACP guidelines tend to emphasize the importance of the clinical history and physicial prior to tests which build upon the clinical story and narrow differentials. Conversely, histories and physical exam findings are rarely mentioned in ACR guidelines that instead focus on choosing the “best test” for specific questions. Thus, it is not surprising that internists who seek complex answers would not find ACR guidelines as relatable, and even cardiologists who find these guidelines more relatable would prioritize ACC guidelines as the product of their mentors, past co-residents/fellows, and evaluators. What obscured the perception of these specialty value differences was the idea of “objective” medical evidence. Clinicians across specialties seemed to assume their trusted organization had correctly evaluated and presented all available evidence, implying differing guidelines must have overlooked or misinterpreted evidence. Discussion Understanding physicians’ distinct professional identities and values can not only help physicians individually but can facilitate policy/guideline development that better resonates with diverse senses of value, fostering wider adoption and support. Our 71 physician interviews over the last two years have revealed remarkably consistent language, values, and opinions within each specialty while differing considerably from those of others. This led us to develop a model to describe these differences which we applied to our 15 most recent pilot interviews, focusing on the perceived value of CV imaging across 5 specialties. This model requires futher refinement and validation in a larger, multicenter study, but our preliminary data suggests this approach may serve as a powerful roadmap to collect data for policy development in CV imaging as well as other areas of imaging. On an individual level, practicing physicians often encounter cultural differences that can quickly be perceived only as selfishly motivated turf wars [32, 33]. Common examples include fixers feeling that managers waste time excessively forming differentials or recommending less effective interventions and managers criticizing fixers for neglecting important clinical context, treating patients as events, and over-prioritizing their own procedures to maximize profits. It can be easy to conclude that we should try to abolish such differences, but dividing into social groups is a natural human tendency [16] and competition can be a positive source of innovation and work ethic. As put by one radiologist, “friction creates energy… We need people to think differently. We need people to challenge others.” Better collaboration exists not where there are less identity differences but where individuals are more aware of such differences and better respond to them. For example, a radiologist may need to adjust his/her report not only for the clinical question at hand but the specialist asking the question. In terms of policy development, casting a broad net and working inductively from a richer understanding of stakeholders’ unique and common values seems critical for complex concepts such as “quality” and “appropriateness.” Our results suggest that previous descriptions of “major domains” for imaging quality improvement [34] may fail to appreciate the complexity of medical culture. Common descriptions involve a flow of care events from patient selection (using patient preferences and appropriateness criteria) to results communication (Fig. 2a). Our interviews yielded a similar flow but a complex referral pattern for patient selection (Fig. 2b). This suggests specialty-specific initiatives may be necessary to affect behavior. For example, having cardiologists speak to internists about imaging practices is likely to be more effective than sending radiologists to ‘educate’ them. Quality CV imaing for EM may require addressing malpractice concerns and look different than quality imaging in a cardiology clinic. Additionally, descriptions of the “right” approach to patient selection and results communication have seemed largely physician-centric and specialty-specific. Thus, there is considerable need to better understand how different specialists and patients define value and quality related to these two key steps of CV imaging. In other words, there is a need in healthcare for better specialty role definition and a collective understanding that the right course of action for diagnostician can and probably should differ from that of a manager without either being wrong.Fig. 2 Dimensions of care framework for evaluating quality of cardiovascular imaging (a) proposed previous. Reprinted with permission from the Methodist DeBakey Cardiovascular Journal. Patient selection referral pattern for CV imaging among 15 physicians at Northwestern Memorial Hospital (b) Finally, it is worth noting that our identity categories and axes were not only internally valid, but seem to have persisted since each specialty’s formation. After advancements in anesthesia and antisepsis in the late 19th century, surgical techniques expanded rapidly and there became a clear difference between seasoned and amateur surgeons. Thus, the American College of Surgeons was founded in the early 20th century and worked to limit hospital privileges to only a small group of surgeons they certified, which served as a model for other specialty boards [10, 35]. Because of their emphasis on technical skills and abilities, boundaries could be more sharply drawn, making fixers particularly prone to inter-specialty competition. As new technologies such as imaging became available, their value centered around how it allowed fixers to better perform the procedures and problem solving that defined them. Conversely, there was not a clear initial need for specialization of managers as groups defined more by there breadth of clinical knowledge [10]. Since a cardiologists’ clinical knowledge is built upon his/her internist identity, it is difficult to separate the two. This is likely why manager groups have less clearly defined borders and are content being subspecialties. Thus, the value of imaging and reports has relied on its integration into and support of clinical understanding and management. Finally, diagnosticians evolved in parallel with multidisciplinary hospitals and advances in medical technology and remain closely tied to the developments that gave them birth. Soon after the discovery of X-rays in 1895, roentgenologists sought to be recognized by other clinicians not only as someone to facilitate use of this new technology but also provide an interpretation or treatment as clinicians. Ever since, radiologists’ values have remained rooted in using technology to provide unparalleled clinical information that could not have obtained without their interpretation or guidance:“…findings should be interpreted always in conjunction with the laboratory and clinical findings… [A Radiologist is] not a marker of pretty pictures.” -Radiologist (1912) [36] “…a real radiologist has to be more than a photographer. He should be a very versatile general practitioner specializing in radiology… A great part of the general medical profession, however, does not realize this and some physicians feel that a six weeks’ course will make them fairly familiar with x-ray work.” –Radiologist (1930) [37] “The ultimate objective is to find out what’s wrong with a patient… it is not just a simple report as you say, it’s deciding what is the most appropriate way of making that diagnosis …[radiologists] have to be familiar with the full spectrum of disease…” –Radiologist (2016) Conclusions If we are to develop imaging guidelines and practices that resonate with providers’ and patients’ diverse senses of value and quality, it seems necessary to first understand the range of perceptions in stakeholders’ own words. Such an understanding requires a qualitative method sensitive enough to capture this complexity, providing a critical roadmap for future healthcare quality initiatives. Abbreviations ACCAmerican college of cardiology ACPAmerican college of physicians ACRAmerican college of radiology AMAAmerican medical association AUCAppropriate use criteria C-GTConstructivist grounded theory CMSCenter of medicaid and medicare services CVCardiovascular EMEmergency medicine Acknowledgements None Funding Author EJK received a Radiological Society of North America (RSNA) Research & Education Foundation’s Research Medical Student Grant (#1521) to support interview transcription costs. Availability of data and materials Interview transcripts analyzed for this investigation are prohibited from being shared publically according to the research protocol approved by the Institutional Review Board. This restriction was added to protect the identity of physician participants. Authors’ contibutions Authors EJK, JDC, RLV, and JCC contributed to the study design. Data collection was performed by EJK. Data analysis was performed by EJK, JDC, BHF, and RLV. All authors contributed significantly to manuscript preparation. All authors read and approved the manuscript. Competing interests No authors had any competing interests relevant to this investigation. Ethical approval and consent to participate This investigation was reviewed by a Northwestern University Institutional Review Board and deemed exempt from full review (STU00202378). Data were also incorporated from a separate approved study (STU00105347), for which informed consent was obtained and documented from all participants, and an additional exempt study (STU00201208). No consent was required for the exempt investigations. ==== Refs References 1. National Academies of Sciences, Engineering, and Medicine. Improving Diagnosis in Health Care. Washington, DC: The National Academies Press; 2015. doi:10.17226/21794. 2. National Quality Forum. National Voluntary Consensus Standards for Imaging Efficiency: A Consensus Report. Washington, DC. Available at http://www.qualityforum.org/publications/2012/01/national_voluntary_consensus_standards_for_imaging_efficiency__a_consensus_report.Aspx. 3. Chaudhuri D Montgomery A Gulenchyn K Mitchell M Joseph P Effectiveness of quality improvement interventions at reducing inappropriate cardiac imaging: a systematic review and meta-analysis Circ Cardiovasc Qual Outcomes 2016 9 1 7 13 10.1161/CIRCOUTCOMES.115.001836 26733588 4. Gibbons RJ Askew JW Hodge D Kaping B Carryer DJ Miller T Appropriate use criteria for stress single-photon emission computed tomography sestamibi studies: a quality improvement project Circulation 2011 123 5 499 503 10.1161/CIRCULATIONAHA.110.975995 21262995 5. Willens HJ Nelson K Hendel RC Appropriate use criteria for stress echocardiography: impact of updated criteria on appropriateness ratings, correlation with pre-authorization guidelines, and effect of temporal trends and an educational initiative on utilization JACC Cardiovasc Imaging 2013 6 3 297 309 10.1016/j.jcmg.2012.11.009 23433927 6. Bhatia RS Dudzinski DM Milford CE Picard MH Weiner RB Educational intervention to reduce inappropriate transthoracic echocardiograms: the need for sustained intervention Echocardiography 2014 31 8 916 923 10.1111/echo.12505 24447139 7. Bhatia RS Milford CE Picard MH Weiner RB An educational intervention reduces the rate of inappropriate echocardiograms on an inpatient medical service JACC Cardiovasc Imaging 2013 6 5 545 555 10.1016/j.jcmg.2013.01.010 23582360 8. Miliard M. Meaningful use will likely end in 2016, CMS chief Andy Slavitt says. HealthcareITNews.com, 2016. Available at http://www.healthcareitnews.com/news/meaningful-use-will-likely-end-2016-cms-chief-andy-slavitt-says. 9. Panzer RJ Gitomer RS Greene WH Webster PR Landry KR Riccobono CA Increasing demands for quality measurement JAMA 2013 310 18 1971 1980 10.1001/jama.2013.282047 24219953 10. Weisz G. Divide and conquer: a comparative history of medical specialization. Oxford; New York: Oxford University Press: 2006. xxx, 359 p. p. 11. Winchester DE Wolinsky D Beyth RJ Shaw LJ Discordance between appropriate use criteria for nuclear myocardial perfusion imaging from different specialty societies: A potential concern for health policy JAMA Cardiology 2016 1 2 207 210 10.1001/jamacardio.2016.0030 27437893 12. De Brantes F, Eccleston S. Improving Incentive to Free Motivation. Princeton: Health Care Incentives Improvement Institute & Robert Wood Johnson Foundation, 2013. Available at http://www.rwjf.org/en/library/research/2013/08/improving-incentives-to-free-motivation.html. 13. Holsinger JW Jr Beaton B Physician professionalism for a new century Clin Anat 2006 19 5 473 479 10.1002/ca.20274 16506233 14. Zuger A Dissatisfaction with medical practice N Engl J Med 2004 350 1 69 75 10.1056/NEJMsr031703 14702431 15. Gunderman RB Boland GW Value in radiology Radiology 2009 253 3 597 599 10.1148/radiol.2533090741 19952023 16. Burford B Group processes in medical education: learning from social identity theory Med Educ 2012 46 2 143 152 10.1111/j.1365-2923.2011.04099.x 22239328 17. Oh H. Your Money or Your Life: How Doctors Learn the Business of Health. Sociology. Volume Ph.D.: UCLA; 2014. p. 149 18. Serra H Medical technocracies in liver transplantation: drawing boundaries in medical practices Health 2010 14 2 162 177 20164164 19. Zetka JR Jr Establishing specialty jurisdictions in medicine: the case of American obstetrics and gynaecology Sociol Health Illn 2011 33 6 837 852 10.1111/j.1467-9566.2011.01341.x 21426362 20. Becher T. Academic tribes and territories : intellectual enquiry and the cultures of disciplines. Milton Keynes England; Bristol, PA, USA: Society for Research into Higher Education : Open University Press: 1989. xiii, 200 p. p. 21. Baker RB The American Ethics Revolution. The American Medical Ethics Revolution: How the AMA's Code of Ethics Has Transformed Physicians' Relationships to Patients, Professionals, and Society 1999 Baltimore, MD Johns Hopkins University Press 17 51 22. Glaser BG, Strauss AL. Awareness of dying. Chicago: Aldine Pub. Co; 1965. xi, 305 p. p. 23. Cook DA Holmboe ES Sorensen KJ Berger RA Wilkinson JM Getting maintenance of certification to work: a grounded theory study of physicians' perceptions JAMA Internal Med 2015 175 1 35 42 10.1001/jamainternmed.2014.5437 25365596 24. Cook DA Sorensen KJ Wilkinson JM Berger RA Barriers and decisions when answering clinical questions at the point of care: a grounded theory study JAMA Internal Med 2013 173 21 1962 1969 10.1001/jamainternmed.2013.10103 23979118 25. Ellingson LL The roles of companion in geriatric patient - interdisciplinary oncology team interactions J Aging Studies 2002 16 361 382 10.1016/S0890-4065(02)00071-3 26. Ghant MS Sengoba KS Recht H Cameron KA Lawson AK Marsh EE Beyond the physical: a qualitative assessment of the burden of symptomatic uterine fibroids on women's emotional and psychosocial health J Psychosom Res 2015 78 5 499 503 10.1016/j.jpsychores.2014.12.016 25725565 27. Charmaz K. Constructing grounded theory. London; Thousand Oaks, Calif: Sage Publications: 2006. xiii, 208 p. p. 28. Kenny M Fourie R Tracing the history of grounded theory methodology: from formation to fragmentation Qual Rep 2014 19 103 1 9 29. Greg Guest AB Laura J How many interviews are enough?: An experiment with data saturation and variability Field Methods 2006 18 59 10.1177/1525822X05279903 30. Spradley JP. The ethnographic interview. New York: Holt, Rinehart and Winston: 1979. vii, 247 p. p. 31. Watkins WW. The anomalous relation of roentgenology to medical practice. Am J Roentgenol. 1920;7. 32. Fiol CM Pratt MG O'Connor EJ Managing intractable identity conflicts Acad Manage Rev 2009 34 1 32 55 10.5465/AMR.2009.35713276 33. Jordan B Birth in four cultures 1983 Montréal Eden Press 144 148 34. Douglas PS Chen J Gillam L Achieving quality in cardiovascular imaging II: proceedings from the second American College of Cardiology -- Duke University medical center think tank on quality in cardiovascular imaging JACC Cardiovasc Imaging 2009 2 2 231 240 10.1016/j.jcmg.2008.11.012 19356561 35. Stevens R. American medicine and the public interest. Berkeley: University of California Press: 1998. xxxv, 578 p. p. 36. Baetjer FH. American roentgen ray society: Thirteenth annual meeting, held at niagara falls, n. y., sept. 11-14, 1912. J Am Med Assoc 1912;LIX(13):1209-1210. 37. Hodges FM THe section on radiology J Am Med Assoc 1930 95 12 833 834 10.1001/jama.1930.02720120001001

PMC005xxxxxx/PMC5002194.txt

==== Front BMC BiotechnolBMC BiotechnolBMC Biotechnology1472-6750BioMed Central London 29610.1186/s12896-016-0296-3Research ArticleCharacterization of human papillomavirus type 16 pseudovirus containing histones Kim Hyoung Jin jinco3@naver.com Kwag Hye-Lim hyelimi6018@nate.com Kim Hong-Jin +82 2 820 5613hongjink@cau.ac.kr Laboratory of Virology, College of Pharmacy, Chung-Ang University, 84 Heukseok-Ro, Dongjak-Gu, Seoul 06974 South Korea 27 8 2016 27 8 2016 2016 16 1 635 4 2016 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 Pseudoviruses (PsVs) that encapsidate a reporter plasmid DNA have been used as surrogates for native human papillomavirus (HPV), whose continuous production is technically difficult. HPV PsVs have been designed to form capsids made up of the major capsid protein L1 and the minor capsid proteins L2. HPV PsVs have been produced in 293TT cells transfected with plasmid expressing L1 and L2 protein and plasmid containing the reporter gene. Several studies have suggested that naturally occurring HPV virions contain cellular histones, and histones have also been identified in mature HPV PsVs. However, the effect of the histones on the properties of the PsVs has not been investigated. Using heparin chromatography, we separated mature HPV type 16 PsVs into three fractions (I, II, and III) according to their heparin-binding affinities. Results The amounts of cellular histone and cellular nucleotides per PsV were found to increase in the order fraction I, II and III. It appeared that PsVs in fraction I contains just small amount of cellular histone in Western blot analysis. The proportions of the three fractions in PsV preparations were 83.4, 7.5, and 9.1 % for fraction I, II, and III PsVs, respectively. In the electron microscope PsVs in fraction I appeared to have a more condensed structure than those in fractions II and III. Under the electron microscope fraction II and III PsVs appeared to be covered by substantial amounts of cellular histone while there was no visible histone covering PsVs of fraction I. Also the levels of reporter gene expression in infections of fraction II and III PsVs to 293TT cells were significantly lower than those in infections of fraction I PsV, and fraction II and III particles had significantly reduced immunogenicity. Conclusions Our findings suggest that the involvement of large amounts of cellular histones during PsV formation interferes with the structural integrity of the PsVs and affects their immunogenicity. The fraction I particle therefore has the most suitable characteristics for use as an HPV PsV. Electronic supplementary material The online version of this article (doi:10.1186/s12896-016-0296-3) contains supplementary material, which is available to authorized users. Keywords Human papillomavirusPseudovirionHeparin chromatographyHistonehttp://dx.doi.org/10.13039/501100003725National Research Foundation of KoreaNRF-2015R1D1A1A01057370Kim Hong-Jin issue-copyright-statement© The Author(s) 2016 ==== Body Background Pseudoviruses (PsV) are synthetic viruses that can include genetic material such as DNA and RNA, and are closely related to the structures and characteristics of its native viruses, but lack characteristics shown in the authentic viruses such as capability for replication [1]. PsV systems permit the continuous production of virus particles mimicking naturally occurring particles, and provide high-throughput systems for evaluating anti-viral agents and vaccine candidates. Recently, various types of PsV systems have been developed. The PsV system for human immunodeficiency virus (HIV) uses the TZM-bI cell line, a genetically modified HeLa cell line expressing receptors for HIV and the firefly luciferase reporter gene under the control of the HIV long-terminal repeat [2]. PsVs for Merkel cell polyoma virus (MCV) have been generated by co-transfection with the VP1 and VP2 genes of MCV strain 339 and a Green fluorescent protein (GFP) reporter plasmid of 293TT cells engineered to stably express the cDNA of Simian virus 40 (SV40) T antigen [3]. Similar systems have been used for the PsVs of polyoma virus JCPyV [4], enterovirus 71 [5] and human papillomavirus (HPV) [6]. The HPV PsV system has evolved along with the success of the commercial vaccine against HPV. Currently the HPV PsV system is the most straightforward and widely used system in the HPV research field because the property of HPV PsV is similar to that of native HPV virion. HPV is a non-enveloped DNA virus that infects cutaneous and mucosal epithelial tissues. Most cases of cervical cancer are caused by infection with high-risk HPV types [7, 8]. The HPV capsid is composed of 72 pentamers arranged on a T = 7 icosahedral lattice, each containing an L1 capsomeres [9, 10]. The capsid also includes the minor capsid protein L2 [11]. It has been suggested that the center of each capsomere is occluded with an L2 protein, and the L2 protein is located within the capsid [12, 13]. Some of the known roles of the L2 protein include facilitating capsid assembly, enhancing infectivity, and encapsidation of the HPV genome [14, 15]. Continuous production of the native HPV virion is difficult because its replication is strictly controlled by the cell cycle of the host cell [16, 17]. For this reason, synthetic HPV particles such as virus-like particles (VLPs), PsVs and quasivirions (QVs) have been substituted for native HPV virions in studies investigating aspects of infection, transmission, immunogenicity, and viral structure [18]. VLPs are composed of the L1 protein alone or L1 and L2 proteins [19–21], whereas the pseudovirus capsid contains both L1 and L2 proteins, and encapsidates reporter plasmid DNA [6]. QVs also contain L1 and L2 proteins, and encapsulate full-length HPV genomes [18, 22]. Capsid structure is thought to be stabilized by intermolecular disulfide bonds between capsomeres: an analysis of recombinant HPV16 VLPs suggested that a critical intercapsomeric disulfide bridge occurs between Cys428 and Cys175 [9, 10, 14]. In the case of synthetic HPV virions such as PsVs, and unlike in other viruses, a long period of maturation (>24 h at 37 °C) is required because the formation of disulfide bonds is slow [14]. The HPV genome is approximately 8 kb in length and is replicated in the nucleus [18]. The genome of native HPV virions is thought to have a chromatin-like structure that includes host histones [23], and cellular histones are also thought to be present in the mature HPV PsVs [13]. The presence of cellular histones has been found to be associated with reduced infectivity of HPV virions and of infectious PsVs. However, the properties of the histone-carrying PsVs have not been investigated in detail. It is generally agreed that differences in the heparin binding affinities of different viruses that attach to the cell via heparin sulfate are responsible for differences in their properties, such as virulence and pathogenesis [24–29]. Since cellular histones bind strongly to heparin [30], we hypothesized that the histone-containing HPV PsVs would bind heparin more strongly than those without histones. In the present work we identified three types of HPV type 16 (HPV16) PsVs according to their heparin-binding affinities, and compared their structures, infectivity and immunogenicity, in order to identify the specific characteristics of histone-carrying PsVs. Results Separation of HPV16 PsVs by heparin chromatography Mature HPV16 PsVs were produced in 293TT cells by co-transfection of p16sheLL (expressing L1 and L2 proteins) and pYSEAP (expressing SEAP) and purified by size-exclusion chromatography (SEC, Fig. 1), and the PsVs were subdivided into fractions I, II, and III according to their heparin-binding affinities (Additional file 1: Figure S1 and Additional file 2: Figure S2): the heparin chromatography condition facilitates separation of PsVs containing small amounts of cellular histone from those with large amounts of histone. SEC demonstrated that cellular histones are principally involved in the production of HPV PsV (Fig. 1). It is known that correctly folded HPV capsids and HPV PsVs can bind to heparin-bound resin in 0.325 M NaCl at pH 7.2 [31, 32]. In the present work 0.65 M NaCl at pH 7.2 was used for the binding studies because under these conditions HPV PsVs bind only when they contain substantial amounts of histone (Additional file 1: Figure S1 and Additional file 2: Figure S2). Thus the flow-through fraction under these conditions was designated fraction I PsV. The PsVs bound to the heparin were eluted successively with buffer containing 0.8 M NaCl and buffer containing 1 M NaCl, and the eluted PsVs were designated fraction II and III PsVs, respectively (Additional file 1: Figure S1). The detail sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) banding patterns of the fractions from the heparin chromatography can be seen in Additional file 2: Figure S2.Fig. 1 Purification of mature HPV16 PsVs by SEC. a SDS-PAGE and Western blots for human histone H3 of SEC fractions of a 293TT cell lysate (control experiment). b SDS-PAGE and Western blots for detecting L1, L2 and human histone H3 of SEC fractions of a lysates of 293TT cells transfected with p16sheLL and pYSEAP. c SDS-PAGE and Western blots for detecting L1 protein and human histone H3. Control refers to a mixture of fractions 3 – 8 in panel a; PsV refers to a mixture of fractions 3 – 8 in panel b. M is a protein marker. PsVs eluted early (in fractions 3 to 8, red box in panel b), whereas most cellular proteins eluted later (fractions 10 to 16), indicating that the PsVs were larger than most of the latter (b). Histone H3 co-eluted with L1 protein in fractions 4 to 6 (b) and there was no histone H3 in the corresponding fractions of the 293TT cell control (a). Fractions 3–8 from the 293TT cell control were combined, as were fractions 3–8 from the 293TT cells producing HPV16 PsVs (red boxes in a and b); they were then analyzed by SDS-PAGE and Western blotting (c) Comparison of histone and cellular DNA contents To compare the histone H3 and H2B contents of the three PsV types, identical amounts of PsVs (based on L1 protein content) were loaded (Fig. 2a). It was found that the virus-associated contents of histone protein increased in the order fractions I, II and III (Fig. 2a). PsVs in fraction I appeared to contain small amounts of histones H3 and H2B, which could be seen when the antibody reaction and film exposure times were extended (Fig. 2b). It was confirmed that the amounts of cellular DNAs less than 1000 bp contained in the three PsVs paralleled the amounts of histones (Fig. 2c). The L1 protein content of each fraction of PsV was determined by sandwich enzyme-linked immunosorbent assay (ELISA) (Fig. 2d), and the proportions of PsVs from fraction I, II, and III inferred from this were 83.4, 7.5, and 9.1 %, respectively (Fig. 2d). Therefore, PsV in fraction I, which contains only a small amount of cellular histone, appears to be the predominant type.Fig. 2 Analysis of histones, cellular DNA contents and proportions of HPV16 PsVs from fraction I, II and III. a shows SDS-PAGE and Western blots comparing the L1, histone H3 and histone H2B contents of the three types of PsVs. b shows SDS-PAGE and Western blots for detecting L1, histone H3 and histone H2B in type I PsVs. b is a repeat of panel A but with extended film exposure for Western blots. c is a result of agarose gel showing cellular DNA contents of PsVs from fraction I, II and III. d is result showing the proportions of PsVs from fraction I, II and III. To calculate the proportion of each type, their L1 content was determined by sandwich ELISA. Data represent the mean ± SD of four independent experiments Comparison of the structural integrity of HPV16 PsVs from fraction I, II and III The PsVs were negatively stained with phosphotungstic acid and observed by transmission electron microscopy (TEM) (Fig. 3 and Additional file 3: Figure S3). Fraction I PsVs appeared to be highly ordered and condensed and to be slightly smaller than those from fractions II and III (Fig. 3). Fraction II PsVs were also highly ordered but their capsomers could be seen more clearly than those of fraction I particles, indicating that they were more loosely connected. Fraction III PsVs stained strongly and appeared to consist of blackened particles (Fig. 3). Interestingly, the fraction III PsVs had a fuzzy appearance (Fig. 3 and Additional file 3: Figure S3), which we believe was due to the large amounts of associated cellular histones and DNA. When the PsVs were stained with immunogold to detect human histone H3, the surfaces of fractions II and III particles appeared to be covered by this histone (Fig. 4). The TEM images indicated that fraction I PsVs were smaller and had a more condensed structure than fractions II and III PsVs.Fig. 3 TEM analysis of HPV16 PsVs of fractions I, II and III. The presented images were representatives from triple observations. Magnification is 234,000× (bar 100 nm) Fig. 4 TEM analysis of HPV16 PsVs from fraction I, II and III stained with immunogold. Human histone H3 protein was detected with rabbit anti-human histone H3, together with 5 nm gold-labeled anti-rabbit IgG as described in Methods. Magnification is 234,000× (bar 100 nm) Expression of reporter gene and reporter gene contents of HPV16 PsVs from fraction I, II and III The expression levels of reporter gene, secreted alkaline phosphatase (SEAP), in the different PsV fractions were compared (Fig. 5a). To use identical amounts of PsVs from three heparin chromatography fractions, the L1 contents were determined by sandwich ELISAs and confirmed by Western blots prior to measure the SEAP expressions. 293TT cells were plated in 96-well tissue culture plates, and each type of PsV was serially diluted from 25 ng/mL to 0.39 ng/mL and incubated for 72 h at 37 °C (Fig. 5a). Moreover, the reporter gene contents of PsVs from three fractions were determined by quantitative real-time polymerase chain reaction (PCR) (Fig. 5b). The SEAP expression levels were significantly lower in fraction II and III PsVs than in fraction I (Fig. 5a) whereas the amount of the pYSEAP reporter plasmid associated with fraction II particles appeared to be similar to that associated with fraction I particles (Fig. 5b). The amount of the pYSEAP reporter plasmid associated with fraction III particles was substantially lower (Fig. 5b). Agarose gel analysis (Additional file 4: Figure S4) supported the result of the quantitative real-time PCR in Fig. 5b. These results indicate that substantial amounts of cellular histones are detrimental for formation of PsV with high expression level of SEAP.Fig. 5 Quantification of SEAP and pYSEAP. a Relative SEAP expressions of PsV particles. Each type of PsV was serially diluted from 25 to 0.39 ng/mL (based on the L1 protein concentrations), and SEAP expression levels were determined as described in Methods. b Relative pYSEAP amounts of PsV particles. The values for fraction I PsVs in panel b were set at 100 %. a shows the mean ± SD of duplicate samples, and b presents the mean ± SD of four independent experiments. **, p < 0.01 The immunogenicity of HPV16 PsVs from fraction I, II and III Mice were immunized three times with HPV16 PsVs, and neutralizing antibody titers were measured (Fig. 6). Fraction II and III PsVs elicited lower levels of neutralizing antibody (Fig. 6) than fraction I PsV. In addition, the mice immunized with PsVs from fraction III gave only weak lymphoproliferative responses upon stimulation with HPV16 L1 VLPs (Additional file 5: Figure S5).Fig. 6 Antibody responses following three immunizations with 50 ng of PsVs per dose. Mice were immunized with PsVs (fraction I, II and III) without adjuvant. Neutralizing antibody titers against HPV16 PsVs in mouse sera were determined as described in Methods. Horizontal bars correspond to median values. PBS, n = 7; type I, n = 7; type II, n = 7; type III, n = 7 Discussion Our results indicated that condensed structure, a substantial amount of a plasmid containing the reporter gene and a low content of cellular histones are important for the formation of infectious and immunogenic HPV16 PsVs. Of the HPV16 PsVs, 83 % appeared to be fraction I PsVs, which fulfil the requirements for highly infectious and immunogenic PsVs (Fig. 2d), whereas about 17 % of the PsVs (particles from fraction II and III) were found to have lower infectivity and immunogenicity (Fig. 2d). Therefore, it appears that most HPV16 PsV undergo assembly and maturation with high fidelity. PsVs are known to undergo assembly within the cell, and the resulting particles are in an immature state with their capsomers only loosely connected to each other immediately after cell lysis [14]. These immature PsVs can be condensed and stabilized by a maturation process that requires incubation at 37 °C overnight [14]. It has been suggested that there are substantial losses of virus titer and of encapsidated reporter gene plasmids during the purification of immature PsV [14]. Recently, Cardone and colleagues analyzed the structures of HPV16 PsVs by time-lapse cryo-electron microscopy [33]. They reported that the diameter of the immature PsVs shrank by ∼ 5 % in the maturation process during which the interaction surfaces between adjacent capsomers were consolidated. As shown in Fig. 3 PsVs from fraction I are slightly smaller and more condensed than those from fraction II or III. In addition the capsomers of fraction II particles were less tightly connected. Infections with particles in fraction II and III led to the much lower expression of SEAP than infections with particles in fraction I (Fig. 5a). Therefore, the integrities of fraction II and III particles are similar to those of immature PsVs while the integrity of fraction I particles correspond to that of mature PsVs. SEAP expression was significantly reduced in PsVs from fraction II (Fig. 5a) as was the elicitation of neutralizing antibodies (Fig. 6) although they harbor substantial levels of reporter gene plasmids (Fig. 5b) and have relatively satisfactory architectural structures (Fig. 3). The position of the L2 protein is known to be important for encapsidation of the reporter gene plasmid [13, 14]. However, we found that the three PsV types did not differ in terms of their L2 protein content (Additional file 6: Figure S6). In addition, ELISAs using monoclonal antibodies H16.V5 and H16.E70, which have been widely used for detecting neutralizing epitopes of the HPV16 capsid, indicated that the level of neutralizing epitopes on the surface of fraction II particles was similar to the level on the surface of fraction I particles (Additional file 7, Figure S7). Therefore, it seems that the structural robustness of fraction II PsVs is not enough for efficient delivery of the reporter gene into target cells, and its neutralizing epitopes are not recognized correctly by cells of the immune system. PsVs in fraction II were found to be covered by substantial amounts of cellular histones and cellular nucleotides (Fig. 4) and it seems likely that these interfere with binding of the PsVs to cell receptors and their correct recognition by immune cells. Previous reports have indicated that viral DNAs containing histone, and naked DNAs without histone, coexist in native HPV preparations, and it was suggested that the DNAs of bovine papillomavirus (BPV) and HPV have a chromatin-like structure composed of nucleosomes about 8.0 nm in diameter, interconnected by naked DNA filaments [23]. Histone-free HPV DNA has been found in native HPV virions [23], and there are no histones in purified BPV [34]. Therefore, the suggestion has been made that the complex of HPV DNA with nucleoproteins is a natural intermediate during virion assembly. The histone-containing fractions in our study were derived from HPV16 PsVs rather than native HPV virions. Our results indicated that the presence of cellular histones was detrimental to the formation of infectious PsV. However, whether the binding of histone to the surface of HPV16 PsVs shown in Fig. 4 is inherent to the HPV virion remains to be clarified. The suggested model of the assembly of HPV capsid is that the C-terminal proximal H4 helix of each L1 molecule in a capsomere invades a neighboring L1 molecule in an adjacent capsomere and forms an intermolecular disulfide bond [10]. L1 protein has a DNA binding domain at its C-terminus [35]. Therefore, there are two kinds of explanation for the effects of cellular histone. First, since immature HPV16 PsVs have greater space between neighboring capsomers (Fig. 3, see particles from fraction II) they may expose more of their DNA binding domains and so allow more binding of cellular DNA plus histones. Alternatively, the presence of cellular histones and cellular DNAs during assembly of the capsid may hinder condensation of the capsids. Conclusions Our results suggest that a low content of cellular histone and cellular DNA of HPV16 PsV and a condensed structure are important for the formation of high quality HPV16 PsVs. We believe that the content of cellular histone is a critical parameter for assessing the quality and stage of maturation of HPV16 PsV. Also, the heparin chromatography technique described here appears to be a useful tool for separating PsV particles of high quality from the total pool of mature PsVs. Methods Production and purification of mature PsVs HPV16 PsVs were produced as described by Buck and colleagues (standard protocol) [14, 36]. 293TT cells were transfected with p16sheLL (expressing L1 and L2 proteins) and pYSEAP (expressing SEAP) using Lipofectamine 2000 (Invitrogen, USA). The transfected cells were cultured for 72 h then harvested and lysed. The plasmids were kindly provided by Dr. J. T. Schiller (NIH, Bethesda, USA). To prepare a mature PsV stock, the lysate was treated with 0.1 % Benzonase (Sigma, USA) and 0.1 % Plasmid Safe (Epicentre, USA) DNase, and incubated for 24 h at 37 °C. The NaCl concentration of the lysate was adjusted to 0.8 M and clarified by centrifugation at 12,000 × g for 10 min at 4 °C. The PsVs in the clarified lysate were purified by SEC as follows: the clarified lysate (0.5 mL) was loaded onto a column (Tricon 10/300, 1 × 32 cm, GE Healthcare, USA) packed with Superose-6 resin (GE Healthcare, USA). The column was equilibrated with working buffer [phosphate buffered saline (PBS) + 0.52 M NaCl + 0.01 % Tween 80 pH 7.2, final NaCl concentration 0.65 M] prior to loading the sample, and the SEC was performed at a flow rate of 0.3 ml/min. Twenty fractions (0.9 mL each) were collected and analyzed by SDS-PAGE and Western blotting. Separation of PsVs by heparin chromatography To separate fraction I, II and III PsVs, we modified our previous protocol for heparin chromatography [31]. A 9 cm poly-prep column (Bio-Rad, USA) was packed with 0.1 ml heparin fast-flow resin (HiPrep™ Heparin FF, GE Healthcare, USA) and equilibrated with binding buffer [PBS + 0.52 M NaCl + 0.01 % Tween 80 pH 7.2, final NaCl concentration 0.65 M]. The PsV-containing fractions from the SEC were pooled and loaded onto the heparin resin, and the capsids eluted without being bound (flow-through and wash), eluted with 0.8 M NaCl, and eluted with 1 M NaCl were considered to be fraction I, II, and III PsVs, respectively. The flow-through and eluted PsV fractions were monitored by SDS-PAGE and Western blots. SDS-PAGE and Western blotting SDS-PAGE was performed according to the Laemmli’s protocol using a Mini-PROTEAN® Tetra Cell (Bio-Rad, USA). The protein bands on SDS-PAGE gels were visualized by silver staining. To detect the L1 protein by Western blotting, rabbit anti-HPV16 L1 serum and HRP-conjugated goat anti-rabbit IgG polyclonal antibody (Pierce, USA) were used [37]. Histone H3 was detected using rabbit anti-human histone H3 (sc-10809, Santa Cruz Biotechnology, USA) or rabbit anti-human histone H3 (ab1791, Abcam, USA), while histone H2B was detected with rabbit anti-human H2B (ab61250, Abcam, USA). HRP-conjugated anti-rabbit immunoglobulin G (IgG) (Bethyl, USA) was used as secondary antibody for detecting histones. Analysis of cellular DNA The L1 protein content of each PsV type was determined by Western blotting and SDS-PAGE. 1 μg of L1 protein of each type was used for analyzing the content of cellular DNA. The DNA was extracted by precipitation with phenol-chloroform-isoamyl alcohol mixture (Sigma, USA), washed with 70 % ethanol and analyzed on 0.9 % agarose gels with ethidium bromide staining. Measurements of L1 protein amounts in HPV16 PsVs from fraction I, II, and III To assess yields of PsVs from fraction I, II and III, their L1 protein content was determined by sandwich ELISA as previously described [32]. Purified HPV16 L1 VLPs were used as a standard. The amounts of L1 protein were confirmed by SDS-PAGE and Western blotting. TEM analysis Purified PsVs (5 μg/mL) from fraction I, II and III were absorbed onto carbon-coated grids and negatively stained with 2 % phosphotungstic acid. To detect histone H3 on the surface of the PsVs, the latter were absorbed onto a carbon-coated grid, and incubated with rabbit anti-histone H3 (ab1791, Abcam, USA) followed by 5 nm gold-labeled goat anti-rabbit IgG polyclonal antibody (ab27235, Abcam, USA). Thereafter the PsVs were fixed with 1 % glutaraldehyde (Sigma, USA) and stained with 2 % uranyl acetate. Electron microscopy was performed on a TEM200CX at a final magnification of 234,000× [31]. Determination of the level of SEAP expression per PsV particle To determine SEAP expression levels of PsVs from fraction I, II and III, the L1 protein content of each type was determined as described above and confirmed by Western blotting. 293TT cells were plated in 96-well tissue culture plates at a density of 3 × 104 cells/well for 4 h prior to PsV infection, and each type of PsV was serially diluted from 25 ng/mL to 0.39 ng/mL (based on the L1 protein concentrations) and incubated for 72 h at 37 °C. SEAP content was measured at 405 nm [31]. Quantitative real-time PCR for pYSEAP in PsVs To measure pYSEAP content per particle, equal amounts of PsVs from fraction I, II and III were used (200 ng based on L1 content). pYSEAP was extracted from the PsVs using a PCR extraction kit (Real Biotech Corporation, Taiwan). Prior to the extraction, the PsV stocks were treated with proteinase K (Qiagen, Germany) at 56 °C for 20 min to release encapsulated pYSEAP, and spiked with pcDNA3.1+ (Invitrogen, USA) as a reference. The primers for detecting pYSEAP were 5′-TTT AAC CAG TGC AAC ACG ACA CGC-3′ (sense) and 5′-TCC CAC TGA CTT CCC TGC TTT CTT-3′ (antisense), and those for detecting pcDNA3.1+ were 5′-ATA CGG GAT AAT ACC GCG CCA CAT-3′ (sense) and 5′-TGC ACG AGT GGG TTA CAT CGA ACT-3′ (antisense). Real-time PCR was performed using a QuantiTect SYBR Green PCR Kit (Qiagen, Germany) on an iCycler real-time PCR machine (Bio-Rad, USA). The crossing threshold (Ct) value of pYSEAP was normalized to that of pcDNA3.1+. Mouse immunization Six-week-old female BALB/c mice (Orient Bio, South Korea) were assigned to four groups of seven mice each: PBS and PsVs from fraction I, II and III. The mice were subcutaneously immunized three times at two-week intervals with PsVs without adjuvant. The control group received 100 μL of PBS per dose, and the others received 50 ng of type I, II, or III PsVs per dose. Mouse sera were collected 10 days after the third immunization. Titrations of anti-HPV16 neutralizing antibodies from mice immunized with HPV16 PsVs The neutralizing assays using HPV16 PsVs were performed as previously described [6, 38] The neutralization titer of a mouse serum was defined as the reciprocal of the highest dilution that caused a reduction of at least 50 % in SEAP activity [39]. Statistical analysis The statistical significance of differences between experimental groups was determined by two-tailed Student’s t-tests. Additional files Additional file 1: Figure S1. Schematic diagram of heparin chromatography to separate HPV16 PsVs into fraction I, II and III. (DOCX 75 kb) Additional file 2: Figure S2. SDS-PAGE analysis of heparin chromatography fractions containing HPV16 PsVs. Mature HPV16 PsVs purified by SEC were further separated by heparin chromatography. The binding buffer for heparin chromatography contains 0.65 M NaCl. LS, FT and W refer to loading sample, flow-through, and wash, respectively. M indicates a protein marker. The PsVs in the flow-through and wash fractions were designated fraction I. PsVs in fraction II and III were eluted from the heparin resin by successive additions of 0.8 and 1 M NaCl. (DOCX 197 kb) Additional file 3: Figure S3. TEM analysis of HPV16 PsVs from fraction I, II and III. Magnification x 234,000 (bars are 50 or 100 nm). (DOCX 1656 kb) Additional file 4: Figure S4. PCRs for reporter genes in HPV16 PsVs from fractions I, II and III. A shows the pYSEAP contents of PsVs from fraction I, II and III. B confirms the specificity of the PCR for the relevant region of the pYSEAP construct. Buffer only and pcDNA3.1 are negative controls. To measure pYSEAP content per particle, in panel A, identical amounts of PsVs from fraction I, II, and III were used (20 ng based on L1 content). 25, 30 or 38 cycles of PCR were carried out. Thirty cycles of PCR were performed for panel B. The same primer set used in the quantitative real time PCR for amplifying the 80 bp region of pYSEAP was used (see the quantitative real time PCR section in Methods). Amplicons were analyzed on a 1.5 % agarose gel. (DOCX 406 kb) Additional file 5: Figure S5. Lymphoproliferative responses following four immunizations with HPV16 PsVs from fraction I, II, or III. The mice were immunized four times with 50 ng of PsVs per dose at 2-week intervals. Mouse splenocytes were obtained 5 days after the fourth immunization. Mouse splenocytes were labeled with carboxyfluorescein succinimidyl ester (CFSE), stimulated with purified HPV16 L1 VLPs, and cultured for 4 days. The splenocytes were stained with allophycocyanin (APC)-conjugated anti-CD4 antibody (eBioscience, USA) and examined with a FACSCalibur flow cytometer (BD Bioscience, USA). To count CD4+ cells, the cells were gated according to forward and side scatter, and the upper-left segment of each graph was counted on FITC and APC scatter plots. Panel A shows the flow cytometry results for three individual mice. The value in panel B represents the mean ± SEM (n = 3). (DOCX 171 kb) Additional file 6: Figure S6. L2 protein contents of HPV16 PsVs from fraction I, II and III. To detect the L1 and L2 proteins, 100 or 200 ng of PsVs was loaded per well (based on the L1 amount). The L2 protein was detected by Western blotting using the anti-RG1-4MAP mouse serum (1:500 dilution) with HRP conjugated goat anti-mouse IgG antibody (Bethyl Laboratories). The proteins on SDS-PAGE gels were visualized by silver staining. The quantities of the L2 protein per PsV particle were similar for all three types of PsV. (DOCX 153 kb) Additional file 7: Figure S7. Reacitivities of HPV16 PsVs from fraction I, II and III with H16.V5 or H16.E70 MAbs. The reactivities of fraction I PsVs were set at 100 %. The values are the mean ± SD of quadriplicate assays. (DOCX 43 kb) Abbreviations ELISAEnzyme-linked immunosorbent assay GFPGreen fluorescent protein HIVHuman immunodeficiency virus HPVHuman papillomavirus HPV16Human papillomavirus type 16 IgGImmunoglobulin G MCVMerkel cell polyoma virus PsVPseudovirus QVQuasivirions SDS-PAGESodium dodecyl sulfate-polyacrylamide gel electrophoresis SEAPSecreted alkaline phosphatase SECSize-exclusion chromatography SV40Simian virus 40 TEMTransmission electron microscopy VLPVirus-like particle Acknowledgements We thank Yingji Jin (College of Pharmacy, Chung-Ang University) and So Young Kim (College of Pharmacy, Chung-Ang University, present affiliation: National Institute of Food and Drug Safety Evaluation) for help conducting the experiments. We thank Dr. J. T. Schiller (NIH, Bethesda, USA) for kindly providing the plasmids p16sheLL and pYSEAP and Dr. N. D. Christensen (Pennsylvania State University College of Medicine, USA) for kindly providing the Mabs (H16.V5 and H16.E70). Funding This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2015R1D1A1A01057370). Availability of supporting data The dataset for supporting the results of this article is included within the article as additional files. Authors’ contributions HJK participated in design of the study, performed purifications and fractionations of PsVs, qRT-PCR analysis of PsVs, quantification of PsVs, DNA analysis of PsVs, investigation of immune responses by PsV, TEM analysis of PsVs, investigation of antigenicity of PsVs and written the manuscript. HLK performed mouse experiments for investigating the immune responses by PsV, cell culture for producing PsVs, PsV-based neutralizing assays and flow cytometry analysis. All authors have read and approved the final manuscript. H-JK written the manuscript and is responsible for the integrity of the work as a whole. 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 carried out in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Chung-Ang University. The animal study protocol was approved by the IACUC. Mice were divided into four groups (seven mice each), and housed in cage containing wood shavings with 12 h day-night cycle in a temperature (22 ± 2 °C). Mice were anesthetized intraperitoneally with 10 μl of 4:1 mixture of Zoletil 50 (Virbac, France) and Rompun (Bayer Animal Health, Germany) before collecting blood or spleen. ==== Refs References 1. Lund PE Hunt RC Gottesman MM Kimchi-Sarfaty C Pseudovirions as vehicles for the delivery of siRNA Pharm Res Dordr 2010 27 400 20 10.1007/s11095-009-0012-2 2. Schultz A Koch S Fuss M Mazzotta AS Sarzotti-Kelsoe M Ozaki DA Montefiori DC von Briesen H Zimmermann H Meyerhans A An automated HIV-1 env-pseudotyped virus production for global HIV vaccine trials PLoS One 2012 7 e51715 10.1371/journal.pone.0051715 23300558 3. Pastrana DV Tolstov YL Becker JC Moore PS Chang Y Buck CB Quantitation of human seroresponsiveness to Merkel cell polyomavirus PLoS Pathog 2009 5 e1000578 10.1371/journal.ppat.1000578 19750217 4. Gee GV O’Hara BA Derdowski A Atwood WJ Pseudovirus mimics cell entry and trafficking of the human polyomavirus JCPyV Virus Res 2013 178 281 6 10.1016/j.virusres.2013.09.030 24100235 5. Wu X Mao QY Yao X Chen P Chen XM Shao J Gao F Yu X Zhu FC Li RC Development and evaluation of a pseudovirus-luciferase assay for rapid and quantitative detection of neutralizing antibodies against enterovirus 71 PLoS One 2013 8 e64116 10.1371/journal.pone.0064116 23755115 6. Buck CB Pastrana DV Lowy DR Schiller JT Efficient intracellular assembly of papillomaviral vectors J Virol 2004 78 751 7 10.1128/JVI.78.2.751-757.2004 14694107 7. Walboomers JM Jacobs MV Manos MM Bosch FX Kummer JA Shah KV Snijders PJ Peto J Meijer CJ Munoz N Human papillomavirus is a necessary cause of invasive cervical cancer worldwide J Pathol 1999 189 12 9 10.1002/(SICI)1096-9896(199909)189:1<12::AID-PATH431>3.0.CO;2-F 10451482 8. Bosch FX Lorincz A Munoz N Meijer CJ Shah KV The causal relation between human papillomavirus and cervical cancer J Clin Pathol 2002 55 244 65 10.1136/jcp.55.4.244 11919208 9. Chen XS Garcea RL Goldberg I Casini G Harrison SC Structure of small virus-like particles assembled from the L1 protein of human papillomavirus 16 Mol Cell 2000 5 557 67 10.1016/S1097-2765(00)80449-9 10882140 10. Modis Y Trus BL Harrison SC Atomic model of the papillomavirus capsid Embo J 2002 21 4754 62 10.1093/emboj/cdf494 12234916 11. Doorbar J Gallimore PH Identification of proteins encoded by the L1 and L2 open reading frames of human papillomavirus 1a J Virol 1987 61 2793 9 3039170 12. Trus BL Roden RB Greenstone HL Vrhel M Schiller JT Booy FP Novel structural features of bovine papillomavirus capsid revealed by a three-dimensional reconstruction to 9 A resolution Nat Struct Biol 1997 4 413 20 10.1038/nsb0597-413 9145113 13. Buck CB Cheng N Thompson CD Lowy DR Steven AC Schiller JT Trus BL Arrangement of L2 within the papillomavirus capsid J Virol 2008 82 5190 7 10.1128/JVI.02726-07 18367526 14. Buck CB Thompson CD Pang YY Lowy DR Schiller JT Maturation of papillomavirus capsids J Virol 2005 79 2839 46 10.1128/JVI.79.5.2839-2846.2005 15709003 15. Holmgren SC Patterson NA Ozbun MA Lambert PF The minor capsid protein L2 contributes to two steps in the human papillomavirus type 31 life cycle J Virol 2005 79 3938 48 10.1128/JVI.79.7.3938-3948.2005 15767396 16. Meyers C Frattini MG Hudson JB Laimins LA Biosynthesis of human papillomavirus from a continuous cell line upon epithelial differentiation Science 1992 257 971 3 10.1126/science.1323879 1323879 17. Ozbun MA Meyers C Characterization of late gene transcripts expressed during vegetative replication of human papillomavirus type 31b J Virol 1997 71 5161 72 9188583 18. Conway MJ Meyers C Replication and assembly of human papillomaviruses J Dent Res 2009 88 307 17 10.1177/0022034509333446 19407149 19. Kirnbauer R Booy F Cheng N Lowy DR Schiller JT Papillomavirus L1 major capsid protein self-assembles into virus-like particles that are highly immunogenic Proc Natl Acad Sci U S A 1992 89 12180 4 10.1073/pnas.89.24.12180 1334560 20. Buck CB Day PM Trus BL The papillomavirus major capsid protein L1 Virology 2013 445 169 74 10.1016/j.virol.2013.05.038 23800545 21. Kirnbauer R Taub J Greenstone H Roden R Durst M Gissmann L Lowy DR Schiller JT Efficient self-assembly of human papillomavirus type 16 L1 and L1-L2 into virus-like particles J Virol 1993 67 6929 36 8230414 22. Pyeon D Lambert PF Ahlquist P Production of infectious human papillomavirus independently of viral replication and epithelial cell differentiation Proc Natl Acad Sci U S A 2005 102 9311 6 10.1073/pnas.0504020102 15958530 23. Favre M Breitburd F Croissant O Orth G Chromatin-like structures obtained after alkaline disruption of bovine and human papillomaviruses J Virol 1977 21 1205 9 191643 24. Schneider-Schaulies J Cellular receptors for viruses: links to tropism and pathogenesis J Gen Virol 2000 81 1413 29 10.1099/0022-1317-81-6-1413 10811925 25. Lerch TF Xie Q Chapman MS The structure of adeno-associated virus serotype 3B (AAV-3B): insights into receptor binding and immune evasion Virology 2010 403 26 36 10.1016/j.virol.2010.03.027 20444480 26. Shukla D Liu J Blaiklock P Shworak NW Bai X Esko JD Cohen GH Eisenberg RJ Rosenberg RD Spear PG A novel role for 3-O-sulfated heparan sulfate in herpes simplex virus 1 entry Cell 1999 99 13 22 10.1016/S0092-8674(00)80058-6 10520990 27. Kern A Schmidt K Leder C Muller OJ Wobus CE Bettinger K Von der Lieth CW King JA Kleinschmidt JA Identification of a heparin-binding motif on adeno-associated virus type 2 capsids J Virol 2003 77 11072 81 10.1128/JVI.77.20.11072-11081.2003 14512555 28. Bear JS Byrnes AP Griffin DE Heparin-binding and patterns of virulence for two recombinant strains of Sindbis virus Virology 2006 347 183 90 10.1016/j.virol.2005.11.034 16380143 29. Chen Y Maguire T Hileman RE Fromm JR Esko JD Linhardt RJ Marks RM Dengue virus infectivity depends on envelope protein binding to target cell heparan sulfate Nat Med 1997 3 866 71 10.1038/nm0897-866 9256277 30. Pal PK Starr T Gertler MM Neutralization of heparin by histone and its subfractions Thromb Res 1983 31 69 79 10.1016/0049-3848(83)90008-7 6612698 31. Kim SY Kim HJ Kim H-J Simple and convenient chromatography-based methods for purifying the pseudovirus of human papillomavirus type 58 Protein Expr Purif 2011 76 103 8 10.1016/j.pep.2010.11.003 21059392 32. Kim HJ Kim SY Lim SJ Kim JY Lee SJ Kim H-J One-step chromatographic purification of human papillomavirus type 16 L1 protein from Saccharomyces cerevisiae Protein Expr Purif 2010 70 68 74 10.1016/j.pep.2009.08.005 19686852 33. Cardone G Moyer AL Cheng N Thompson CD Dvoretzky I Lowy DR Schiller JT Steven AC Buck CB Trus BL Maturation of the human papillomavirus 16 capsid MBio 2014 5 e01104 14 25096873 34. Larsen PM Storgaard L Fey SJ Proteins present in bovine papillomavirus particles J Virol 1987 61 3596 601 2822965 35. Li M Cripe TP Estes PA Lyon MK Rose RC Garcea RL Expression of the human papillomavirus type 11 L1 capsid protein in Escherichia coli : characterization of protein domains involved in DNA binding and capsid assembly J Virol 1997 71 2988 95 9060658 36. Laboratory of Cellular Oncology, National Cancer Institute. Production of Papillomaviral Vectors (Pseudoviruses). [http://home.ccr.cancer.gov/Lco/pseudovirusproduction.htm]. Accessed June 2015. 37. Kim HJ Jin Y Kim H-J The concentration of carbon source in the medium affects the quality of virus-like particles of human papillomavirus type 16 produced in Saccharomyces cerevisiae PLoS One 2014 9 e94467 10.1371/journal.pone.0094467 24714383 38 Laboratory of Cellular Oncology, National Cancer Institute. Papillomavirus Neutralization Assay. [http://home.ccr.cancer.gov/Lco/neutralizationassay.htm]. 39 Pastrana DV Buck CB Pang YY Thompson CD Castle PE FitzGerald PC Kruger Kjaer S Lowy DR Schiller JT Reactivity of human sera in a sensitive, high-throughput pseudovirus-based papillomavirus neutralization assay for HPV16 and HPV18 Virology 2004 321 205 16 10.1016/j.virol.2003.12.027 15051381

PMC005xxxxxx/PMC5002195.txt

==== Front World J Surg OncolWorld J Surg OncolWorld Journal of Surgical Oncology1477-7819BioMed Central London 99310.1186/s12957-016-0993-3Case ReportA rare atypical rapidly involuting congenital hemangioma combined with vascular malformation in the upper limb Lu Hui 86-0571-87236121hitman1982@hotmail.com 1Chen Qiang chenqiang888@hotmail.com 1Shen Hui shenhui33@sina.com 1Ye Ganmin yeganmin@126.com 21 Department of Hand Surgery, The First Affiliated Hospital, College of Medicine, Zhejiang University, #79 Qingchun Road, Hangzhou, Zhejiang Province 310003 People’s Republic of China 2 Departments of Orthopedics, Sanmen People’s Hospital, #171 Renmin Road, Taizhou, Zhejiang Province 317100 People’s Republic of China 26 8 2016 26 8 2016 2016 14 1 2293 3 2016 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 Rapidly involuting congenital hemangioma is a congenital soft tissue tumor, which is difficult to excise completely and rather prone to recur. This atypical tumor combined with capillary-lymphatic-venous malformation was not reported in the literature. Case presentation We report an atypical case of a 16-year-old teenager who was born with a mass in his right upper limb. Since there is a recurrence after excision for several times and had a serious impact on daily life, we chose amputation. After the surgery, the patient gained a functional recovery. Two years after the surgery, he had no tumor recurrence. Conclusions For this rare tumor with repeated recurrences and significant impact on daily life, we suggest performing amputation at the early stage. Keywords Rapidly involuting congenital hemangiomaAmputationRecurrenceCapillary-lymphatic-venous malformationZhejiang traditional Chinese medicine Research Program(the grant number 2016ZA124Lu Hui Sanmen Social Development of Science and Technology Projects15302Ye Ganmin issue-copyright-statement© The Author(s) 2016 ==== Body Background Rapidly involuting congenital hemangioma (RICH) is an uncommon, often high-flow vascular tumor that presents at birth. It is generally considered as a congenital condition [1]. Because it often invades the surrounding tissues, it is difficult to excise the tumor completely and it easily has a local recurrence [2–4]. However, surgery is still a conventional treatment for this disease. Aggressive partial resections will easily stimulate the development of the tumor, so we are supposed to avoid it in the future. Case presentation A 16-year-old teenager who was born with a mass in his right hand, which was diagnosed as hemangioma. When he was 5 months old, the tumor was partially excised in Shanghai. The mass developed increasingly after the surgery and presented diffuse growth from the distal to proximal upper limb. The patient was treated with sclerotherapy at the age of three in Zhengzhou but the treatment failed. He was treated with excision again and abdominal pedicle flap surgery at the age of seven in Beijing. One year later, his fingers were separated. And the tumor had been partially excised for several times in Wenzhou before the patient came to our hospital. The patient’s right upper limb could not move upward. He was almost helpless in his daily life and not able to go to school. The physical examination showed multiple scars on his right upper limb and a huge mass in his right hand (Fig. 1). There was no range of motion in his right wrist and right elbow. The right shoulder muscles had atrophied. Laboratory studies revealed that hemoglobin, white cell count, and platelet count are within normal range. Erythrocyte sedimentation rate (ESR), high-sensitivity c-reactive protein, and tumor biological markers were normal. Considering the multiple recurrences of the tumor, the poor quality of his life, and the resistance to surgery, we chose amputation. Considering the patient’s need of prosthetics, the amputation was performed on the middle of the upper limb under general anesthesia. The vascular tumors and thrombosis were visible during the operation (Fig. 2). Gross examination demonstrated segments of the skeletal muscle, containing ill-defined vascular lesions, and partial thrombosis. Pathology showed that skin chronic inflammation was involved with hyperpigmentation below the basal layer and with thin-walled vessels and thin sinusoidal vascular channels (Fig. 3). Immunohistochemistry shows D2-40 (positive) and Glut-1 (negative). The diagnosis of atypical rapidly involuting congenital hemangioma (RICH) combined with capillary-lymphatic-venous malformation (CLVM) was made (ISSVA classification [5]) according to the clinical and histopathologic manifestations. The patient could raise his shoulder and use prosthetic limbs after surgery; the ability of self-care was obviously improved. He could go to school after the surgery. There was no evidence of recurrence at the 2-year follow-up (Fig. 4).Fig. 1 It shows that the patient had a mass on the right upper limb with movement limitation before the amputation Fig. 2 The photograph during the operation, showing segments of the skeletal muscle (black arrow), containing ill-defined vascular lesions and the partial thrombosis (red arrow) Fig. 3 Pathology showed: skin chronic inflammation was involved with hyperpigmentation below the basal layer and thin-walled vessels (hematoxylin-eosin stain, original magnification ×100) Fig. 4 It shows that the patient was able to perform daily activities Discussion In this case, we had difficulty in deciding whether the tumor should be one type within the ISSVA classification. Infantile hemangioma is the commonest tumor in infants with progressive proliferation in the first year of life and slow regression in the next few years. There are also few other rare congenital hemangiomas (RICH, NICH, PICH) with different clinical courses [6–8]. Clinical course and intraoperative view suggest rare locally aggressive vascular tumor (papillary intralymphatic angioendothelioma [9]), lymphatic malformation with fat tissue component, and hemorrhage or mixed vascular malformation: lymphatic and venous (LM-VM) [10]. Venous malformations are the second malformations in occurrence after capillary malformations. None of them fits to describe the case. Klippel-Trenaunay syndrome also should be considered where two vascular malformations and soft tissue hypertrophy is present, but this patient is lack of port-wine stain and varicose veins [11, 12]. The test of Glut-1 [13] was negative; we can exclude infantile hemangioma. This patient was characterized by locally aggressive nature and the test of D2-40 was positive, tend to diagnose with atypical RICH combined with VM [6]. Treatment choice is difficult for this rare case. En bloc resection of the tumor may be the best treatment to prevent recurrence. But due to the nerves, tendon, ligament, and essential structures surrounding the tumor, it will be easy to cause the physical disability if the tumor is excised completely. Patient in this case involved all tissues and growth is progressive after partial resections. The main purpose of the surgery is to release the local symptoms such as persistent pain and numbness, increasing size of the mass, and functional impairment. The patient in this case had several aggressive surgeries, which not only stimulated the growth of the tumor but also resulted in unnecessary injuries in other parts such as the donor site. Therefore, it needs to avoid similar crisis in the future. A prosthesis was fitted after the amputation, the patient could raise his shoulder, and his ability of self-care was obviously improved. Sclerotherapy, chemoembolization, and embolotherapy [14] are treatment options for very vascularized tumors or arterio-venous malformations. Currently, therapy with Sirolimus is considered in severe cases with enhancing results. Conclusions For this rare tumor with repeated recurrences and significant impact on the daily life, we suggest performing an amputation at the early stage. Abbreviations RICHRapidly involuting congenital hemangioma LM-VMLymphatic and venous CLVMCapillary-lymphatic-venous malformation ESRErythrocyte sedimentation rate Acknowledgements First and foremost, I would like to show my deepest gratitude to my colleagues Dr Qiang Chen, Dr Hui Shen, and Dr Ganmin Ye who have provided me with valuable assistance in every stage of writing this paper. Meanwhile, I also appreciate Zhejiang Traditional Chinese Medicine Research Program (the grant number 2016ZA124) and Sanmen Social Development of Science and Technology Projects (the grant number 15302) for sponsoring our research. Last but not the least, I would like to thank all my friends, especially my lovely wife for her encouragement and support. Funding Zhejiang traditional Chinese medicine Research Program (the grant number 2016ZA124) and Sanmen Social Development of Science and Technology Projects (the grant number 15302) supported the work. Availability of data and materials The dataset supporting the conclusions of this article is included within the article. Authors’ contributions HL drafted the manuscript. QC, HS, and GY participated in the design of the study and performed the statistical analysis. HL conceived of the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication All the authors express the consent for publication on World Journal of Surgical Oncology. Ethics approval and consent to participate These study protocols were approved by the Ethics Committee of Medical Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University. Consent Written informed consent was obtained from the patient 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. Abate MV Davanzo R Bibalo C Zennaro F Berti I RICH (rapidly involuting congenital hemangioma): not only a definition of wealth J Pediatr 2012 161 365 365.e1 10.1016/j.jpeds.2012.02.036 22497907 2. Sur A Manraj H Lavoie PM Lim K Courtemanche D Brooks P Albersheim S Multiple successful angioembolizations for refractory cardiac failure in a preterm with rapidly involuting congenital hemangioma AJP Rep 2016 6 e99 e103 10.1055/s-0035-1570342 26929881 3. Andreu-Barasoain M Naz E Diaz M Lopez-Estebaranz JL Rapidly involuting congenital hemangioma associated with transient anemia and thrombocytopenia Int J Dermatol 2013 52 1025 1026 10.1111/j.1365-4632.2011.05091.x 23489172 4. Scalise R Bolton J Gibbs NF Rapidly involuting congenital hemangioma (RICH): a brief case report [J] Dermatol Online J 2014 20 11 16 5. Dasgupta R Fishman SJ ISSVA classification Semin Pediatr Surg 2014 23 158 161 10.1053/j.sempedsurg.2014.06.016 25241091 6. Wassef M Blei F Adams D Alomari A Baselga E Berenstein A Burrows P Frieden IJ Garzon MC Lopez-Gutierrez JC Lord DJ Mitchel S Powell J Prendiville J Vikkula M Board I Scientific C Vascular anomalies classification: recommendations from the International Society for the Study of Vascular Anomalies Pediatrics 2015 136 e203 214 10.1542/peds.2014-3673 26055853 7. Bruder E Perez-Atayde AR Jundt G Alomari AI Rischewski J Fishman SJ Mulliken JB Kozakewich HP Vascular lesions of bone in children, adolescents, and young adults. A clinicopathologic reappraisal and application of the ISSVA classification Virchows Arch 2009 454 161 179 10.1007/s00428-008-0709-3 19107514 8. Galambos C Nodit L Identification of lymphatic endothelium in pediatric vascular tumors and malformations Pediatr Dev Pathol 2005 8 181 189 10.1007/s10024-004-8104-9 15719202 9. Kugler A, Koelblinger P, Zelger B, Ahlgrimm-Siess V, Laimer M. Papillary intralymphatic angioendothelioma (PILA), also referred to as Dabska tumour, in an 83-year-old woman. J Eur Acad Dermatol Venereol. 2015. doi:10.1111/jdv.13300. (PMID:26333144) 10. Al-Adnani M Williams S Rampling D Ashworth M Malone M Sebire NJ Histopathological reporting of paediatric cutaneous vascular anomalies in relation to proposed multidisciplinary classification system J Clin Pathol 2006 59 1278 1282 10.1136/jcp.2006.038240 16751300 11. Franz RW Prok A Klippel-Trenaunay syndrome: treatment of lower extremity pain with a spinal cord stimulator Vascular 2009 17 293 295 10.2310/6670.2009.00021 19769812 12. Sharma D Lamba S Pandita A Shastri S Klippel-trenaunay syndrome—a very rare and interesting syndrome Clin Med Insights Circ Respir Pulm Med 2015 9 1 4 25861232 13. Leon-Villapalos J Wolfe K Kangesu L GLUT-1: an extra diagnostic tool to differentiate between haemangiomas and vascular malformations Br J Plast Surg 2005 58 348 352 10.1016/j.bjps.2004.05.029 15780229 14. Fergusson IL Haemangiomata of skeletal muscle Br J Surg 1972 59 634 637 10.1002/bjs.1800590815 5069204

PMC005xxxxxx/PMC5002196.txt

==== Front BMC Fam PractBMC Fam PractBMC Family Practice1471-2296BioMed Central London 51010.1186/s12875-016-0510-3ErratumErratum to: Towards improving diagnosis of memory loss in general practice: TIMeLi diagnostic test accuracy study protocol Creavin Sam T. Sam.Creavin@bristol.ac.uk 1Cullum Sarah J. 1Haworth Judy 2Wye Lesley 1Bayer Antony 3Fish Mark 4Purdy Sarah 1Ben-Shlomo Yoav 11 School of Social and Community Medicine, University of Bristol, 39 Whatley Road, Bristol, BS8 2PS, UK 2 North Bristol NHS Trust Southmead Hospital, Southmead Road, Westbury-on-Trym, Bristol, BS10 5NB UK 3 Division of Population Medicine, Cardiff University School of Medicine University Hospital Llandough, Penarth, CF64 2XX, UK 4 Department of Neurology, Musgrove Park Hospital, Taunton, Somerset, TA1 5DA, UK 27 8 2016 27 8 2016 2016 17 1 11927 7 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.issue-copyright-statement© The Author(s) 2016 ==== Body Erratum Unfortunately, the original version of this article [1] mistakenly did not include an Acknowledgements section. The Acknowledgements have been included in full in this erratum. Acknowledgements This paper presents independent research part-funded by the National Institute for Health Research School for Primary Care Research (NIHR SPCR). The views expressed are those of the author(s) and not necessarily those of the NIHR, the NHS or the Department of Health. The online version of the original article can be found under doi:10.1186/s12875-016-0475-2. ==== Refs References 1. Creavin ST Cullum SJ Haworth J Wye L Bayer A Fish M Towards improving diagnosis of memory loss in general practice: TIMeLi diagnostic test accuracy study protocol BMC Fam Pract. 2016 17 79 10.1186/s12875-016-0475-2 27430736

PMC005xxxxxx/PMC5002197.txt

==== Front J Ovarian ResJ Ovarian ResJournal of Ovarian Research1757-2215BioMed Central London 25910.1186/s13048-016-0259-2ResearchGenomic alterations in neuroendocrine cancers of the ovary Yaghmour George gyaghmour@WESTCLINIC.com 12Prouet Philippe (901) 448-5814pprouet@uthsc.edu 3Wiedower Eric ewiedower@WESTCLINIC.com 12Jamy Omer Hassan omerjamy@gmail.com 3Feldman Rebecca rfeldman@carisls.com 4Chandler Jason C jachandler@WESTCLINIC.com 12Pandey Manjari mpandey@WESTCLINIC.com 12Martin Mike G mmartin@WESTCLINIC.com 121 The West Cancer Center, 1588 Union Ave., Memphis, TN 38104 USA 2 Department of Hematology & Oncology, The University of Tennessee Health Science Center, 956 Court Ave., Suite H310A, Memphis, TN 38163 USA 3 Department of Internal Medicine, The University of Tennessee Health Science Center, 956 Court Ave., Suite H314, Memphis, TN 38163 USA 4 Caris Life Sciences, 4750 S. 44th Place, Phoenix, AZ 85040 USA 26 8 2016 26 8 2016 2016 9 1 521 6 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 As we have previously reported, small cell carcinoma of the ovary (SCCO) is a rare, aggressive form of ovarian cancer associated with poor outcomes. In an effort to identify new treatment options, we utilized comprehensive genomic profiling to assess the potential for novel therapies in SCCO. Methods Patients with SCCO, SCCO-HT (hypercalcemic type), neuroendocrine tumors of the ovary (NET-O), and small cell carcinoma of the lung (SCLC) profiled by Caris Life Sciences between 2007–2015 were identified. Tumors were assessed with up to 21 IHC stains, in situ hybridization of cMET, EGFR, HER2 and PIK3CA, and next-generation sequencing (NGS) as well as Sanger sequencing of selected genes. Results Forty-six patients with SCCO (10 SCCO, 18 SCCO-HT, 18 NET-O) were identified as well as 58 patients with SCLC for comparison. Patients with SCCO and SCCO-HT were younger (median 42 years [range 12–75] and 26 years [range 8–40], respectively) than patients with NET-O 62 [range 13–76] or SCLC 66 [range 36–86]. SCCO patients were more likely to be metastatic (70 %) than SCCO-HT (50 %) or NET-O (33 %) patients, but at a similar rate to SCLC patients (65 %). PD1 expression varied across tumor type with SCCO (100 %), SCCO-HT (60 %), NET-O (33 %) vs SCLC (42 %). PDL1 expression also varied with SCCO (50 %), SCCO-HT (20 %), NET-O (33 %) and SCLC (0 %). No amplifications were identified in cMET, EGFR, or HER2 and only 1 was found in PIK3CA (NET-O). Actionable mutations were rare with 1 patient with SCCO having a BRCA2 mutation and 1 patient with NET-O having a PIK3CA mutation. No other actionable mutations were identified. Conclusions No recurrent actionable mutations or rearrangements were identified using this platform in SCCO. IHC patterns may help guide the use of chemotherapy in these rare tumors. Keywords Gynecologic malignanciesSmall cell carcinoma of ovaryGenomic profilingActionable mutationChemotherapyissue-copyright-statement© The Author(s) 2016 ==== Body Background Small cell carcinoma of the ovary (SCCO) is a very rare and aggressive form of ovarian cancer that carries a poor prognosis despite early median age at diagnosis. SCCO presents challenges for diagnosis, prediction of outcomes, and overall treatment strategies [1, 2]. Due to limited data, there are no defined treatment protocols [3]. Molecular mechanisms underlying development and progression of SCCO remain poorly understood. Inactivation in SMARCA4, which codes for BRG1, in small cell carcinoma of the ovary hypercalcemic type (SCCO-HT) has been described in 75-100 % of cases. SMARCA4 is a member of the SWI/SNF chromatin-remodeling gene complex, which has been shown to be mutated in several different cancers. SMARCA4 inactivation appears to be a specific diagnostic marker in SCCOHT [4–7]. This inactivation, however, does not define any treatment strategies that would be defined with actionable abnormalities. Profiling of rare tumors using the Caris Molecular Intelligence platform has yielded potential targets in various tumor types. Burzawa, et al. analyzed on 78 samples of small cell cancer of the cervix using Caris Molecular Intelligence (53 samples) and a 50-gene next-generation sequencing (NGS) platform (25 samples). They showed elevated expression of TOPO2A and TOPO1 implying sensitivity to etoposide and topotecan, respectively. They also described potentially targetable mutations in Akt1, KRAS, PIK3CA, and TP53. One of the patients who had a KRAS mutation showed a complete response to MEK inhibitor therapy ongoing for more than 12 months [8]. PIK3CA mutations were identified as potential targets in small call carcinoma of the breast in another study using the Caris Molecular Intelligence platform [9]. Caris Molecular Intelligence is a multi-platform tumor profiling service that includes gene sequencing (next-generation sequencing [NGS] and Sanger), protein expression analysis (immunohistochemistry [IHC]), and gene copy number and translocation analysis (chromogenic or fluorescence in situ hybridization [CISH or FISH]). As patients living with these malignancies have few therapeutic options, we hypothesized that Caris Molecular Intelligence may provide clinically relevant information for women with SCCO [10]. Methods After obtaining Institutional Review Board approval from the University of Tennessee we queried the Caris Life Sciences database for patients profiled by Caris Molecular Intelligence using the keywords “small cell” to search in the clinical history and diagnosis fields for cases of SCCO, SCCO-HT (hypercalcemic type), neuroendocrine tumors of the ovary (NET-O) from 2007–2015. Comparative data was pulled for SCLC from April 2015- September 2015. Caris registry is a proprietary database to which the investigators have access for research purposes, which collates and categorizes molecular alterations and cancer types across patients who have had cancer molecular testing done throughout The United States using a multi-platform tumor profiling service that includes gene sequencing with NGS and Sanger, protein expression analysis with IHC, and gene copy number and translocation analysis with CISH or FISH. Clinicopathologic characteristics of patients with SCCO, SCCO-HT, NET-O and SCLC patients were identified, including age and whether metastatic disease was present at the time of profiling. IHC analysis was performed using commercially available detection kits and automated staining techniques (Benchmark XT, Ventana, Tucson, AZ; and AutostainerLink 48, Dako, Carpinteria, CA). Tumors were assessed with up to 25 IHC stains (ALK, AR, BCRP, c-KIT, ER, PR, cMET, EGFR, HER2, IGF1R, PTEN, PD-1, PDGFR, PD-L1, ERCC1, TS, MGMT, RRM1, TLE3, TUBB3, SPARC, TOP2A, TOPO1, MRP1, PGP). Gene copy number alterations of cMET, EGFR, HER2, PIK3CA, and TOP2A were analyzed by DNA ISH using (FISH and/or CISH probes as part of automated staining techniques (Benchmark XT, Ventana, Tucson, AZ) and automated imaging systems (BioView, Billerica, MA). The ratio of gene to pericentromeric regions of chromosome 7 (EGFR, cMET), 17 (HER2, TOP2A) and 3 (PIK3CA) were used to determine increases in gene copy number. Tumors also underwent NGS analysis, which is a form of parallel sequencing that greatly enhances the efficiency of identifying both somatic and germline mutations [11]. NGS sequencing was performed on genomic DNA isolated from tumor tissue using the Illumina MiSeq platform. PCR products were bi-directionally sequenced using the BigDye Terminator v1.1 chemistry, analyzed using the 3730 DNA Analyzer (Applied Biosystems, Grand Island, NY). Sequence traces were analyzed using Mutation Surveyor software v3.25 (Soft Genetics, State College, PA). NGS and Sanger sequencing of a 47-gene panel (ABL, AKT, ALK, APC, ATM, BRAF, BRCA1, BRCA2, CDH1, cKIT, cMET, CSF1R, CTNNB1, EGFR, ERBB4, FBXW7, FGFR1, FGFR2, FLT3, GNA11, GNAQ, GNAS, HER2, HNF1A, HRAS, IDH1, JAK2, JAK3, KDR, KRAS, MLH1, MPL, NOTCH1, NPM1, NRAS, PDGFRA, PIK3CA, PTEN, PTPN11, RB, RET, SMAD4, SMARCB1, SMO, STK11, TP53, VHL) was evaluated for somatic mutations. Coding changes were accessed for pathogenicity using PolyPhen-2 [10, 12]. Retrospective analysis of biomarker frequency distributions was attained using standard descriptive statistics evaluating the incidence of the aforementioned genomic alterations in these tumors by Caris Molecular Intelligence profiling. The two-tail Fisher’s exact test analyzed whether frequencies differed by subgroup, specifically to determine if any non-random associations between SCCO and SCLC existed. A p-value <0.05 was considered statistically significant and all p-values were 2-sided. Results Forty-six unique patients with SCCO were identified that had profiling between 2007–2015. These included 10 patients with SCCO with median age of 42 (range 12–75), 18 patients with “small cell/hypercalcemic” (SCCO-HT) with median age of 26 (range 8–40), and 18 patients with “neuroendocrine” (NET-O) with median age of 62 (range 13–76). 70 % of SCCO patients, 50 % of SCCO-HT patients, and 33 % NET-O patients had metastatic disease at presentation. For comparison, 58 patients with SCLC were identified. The median age was 66 (range 36–86), and 65 % had metastatic disease at presentation (Table 1).Table 1 Median age and % Metastatic disease at presentation of SCCO, SCCO-HT, NET-O, and SCLC Small Cell Cancer Type Median Age %Metastatic SCCO (n = 10) 42 (12–75) 70 % SCCO-HT (n = 18) 26 (8–40) 50 % NET-O (n = 18) 62 (13–76) 33 % SCLC (n = 58) 66 (36–86) 65 % By IHC, 100 % of SCCO patients (2/2), 60 % of SCCO-HT patients (3/5), and 33 % of NET-O (1/3) subtypes expressed PD1, while 42 % of SCLC patients (22/53) expressed PD1. Moreover, 50 % of SCCO patients (1/2), 20 % of SCCO-HT patients (1/5), and 33 % of NET-O patients (1/3) expressed PDL1, while 0 patients with SCLC (0/54) expressed PDL1. (SCCO 1/2 vs. SCLC 0/54 [p-value = 0.036]). In addition, by IHC, 100 % of SCCO patients (7/7), 92 % of SCCO-HT patients (11/12), and 69 % of NET-O patients (11/16) expressed TOP2A, while 95 % of patients with SCLC (52/55) expressed TOP2A. (NET-O 11/16 vs. SCLC 52/55 [p-value = 0.012]). Also 8 % (1/12) of SCCO-HT patients, 33 % (4/12) of NET-O, and 50 % (1/2) SCCO patients expressed ERCC1. Moreover, 10 % (1/10) of SCCO patients, 38 % (6/16) of SCCO-HT patients, and 59 % (10/17) of NET-O patients expressed MGMT, while 13 % of SCLC patients (7/55) expressed MGMT. (NET-O 10/17 vs. SCLC 7/55 [p-value =0.0003]). Also, 88 % (7/8) of SCCO patients expressed RRM1. SCCO and SCLC patients had similar patterns of other IHC expression (Fig. 1).Fig. 1 Biomarker Differences btw Rare Ovarian Cancers by IHC No amplifications were identified in cMET, EGFR, or HER2 by ISH in any neuroendocrine cancer of the ovary. PIK3CA was amplified by ISH in NET-O patients in 33 % (1/3) of patients (Table 2). NGS and Sanger sequencing of a 47-gene panel revealed TP53 mutations in 25 % of SCCO (1/4) patients and BRCA2 mutations in 50 % of SCCO (1/2) patients. PIK3CA mutations were seen in 16 % (1/6) of sequenced NET-O patients, but no SCLC patients. No other actionable mutations were identified (Fig. 2).Table 2 Amplifications detected by in situ hybrization cMET EGFR HER2 PIK3CA SCCO 0/4 0/2 0/9 n/a SCCO-HT 0/9 0/2 0/13 n/a NET-O 0/10 0/1 0/15 1/3 Fig. 2 Sequencing results by NGS/Sanger Discussion In contrast to other tumor types, Caris Molecular Intelligence did not identify recurrent actionable mutations or alterations in the subtypes of SCCO. IHC patterns suggested potential sensitivities to chemotherapy. Alternative methods of genomic interrogation, such as total exon sequencing, may discover actionable targets. Neuroendocrine cancers of the ovary were similar when profiled by Caris Molecular Intelligence to SCLC. IHC testing for biomarker differences between these rare tumor subtypes showed high rates of TOP2A (100 % of SCCO patients, 92 % of SCCO-HT patients and 69 % of NET-O patients) in all 3 subtypes of neuroendocrine cancers of the ovary, implicating the use of topoisomerase inhibitors as rationale treatment options. Topoisomerase inhibitors such as pegylated liposomal doxorubicin have been commonly used as second line therapy in platinum resistant ovarian cancer [13]. Thibault et al. showed that inhibition of topoisomerase II using polyamine vectorized inhibitor (F14512) significantly inhibited tumor growth in ovarian cells and constitutes a potential new therapy for platinum resistant ovarian cancer [14]. Platinum-based therapies are the current standard of care for ovarian cancer. However, drug resistance plays an important role in determining overall survival. Expression of ERCC1, a strategic marker of nucleotide excision repair, has been associated with resistance to platinum based chemotherapy [15]. SCCO-HT exhibited the lowest rate of ERCC1 in our study, implicating the use of platinum based regimens for this sub-type. MGMT, a DNA repair enzyme, plays a crucial role in mediating resistance to alkylating agents in various tumors, especially malignant gliomas. Epigenetic silencing of this gene has resulted in increased chemo-sensitivity by compromising DNA repair mechanisms. Temozolomide, an alkylating agent, is currently being tested as a single agent for ovarian cancer but has the potential for second line therapy in combination with methoxyamine for patients that have failed platinum and paclitaxel chemotherapy [16, 17]. Our data showed that SCCO exhibited the lowest rate of MGMT, implicating alkylating agents such as temozolomide as a potential agent to be used. Elevated levels of RRM1 gene have been associated with poor outcomes in advanced NSCLC treated with gemcitabine [18, 19]. Several studies have been conducted to explore the potential benefit of gemcitabine in ovarian cancer. Ferrandina et al. looked at both RRM1 and RRM2 in primary ovarian cancer and observed shorter OS with higher RRM2 expression [20]. Our cases of SCCO had high expression of RRM1 but NET-O exhibited the lowest rate of RRM1 expression, implicating gemcitabine as a useful agent for NET-O. Blocking the PD1/PDL1 interaction is being studied in several malignancies with good responses. Our subset of SCCO had a high expression of PD1 and PDL1 and the therapeutic potential of this pathway needs to be explored. Mutations detected by hot spot sequencing panel were infrequent events and amplification by ISH were rare as well. Rare ovarian cancers would benefit from expanded mutation profiling to identify additional potential targets as well as confirm the presence of SMARCA4 in SCCO-HT. Our study had some limitations. This was a retrospective analysis of a limited number of cases. Due to the rarity of this disease only 46 cases were identified and statistical analysis could not be performed. The lack of information such as patient demographics, treatment modalities and outcome prevented us from determining clinical correlation. Conclusion Neuroendocrine cancers of the ovary are rare malignancies. With no recurrent actionable mutations, treatment remains a challenge. Future clinical trials looking at standard of care chemotherapy using platinum based chemotherapeutics and comparing that to specific chemotherapy regimens based off of IHC mutations may prove to be of significance in treating this rare disease. Our study showed that all SCCO types showed high expression of TOP2A, therefore, trials comparing platinum based therapies to topoisomerase II inhibitors such as doxorubicin or etoposide in patients who have high expression of ERCC1, which is associated with platinum resistance, may lead to a change in our approach to this subset of diseases. Also, there is potential for clinical trials comparing standard chemotherapy to PD1/PDL1 inhibition in SCCO as our subset of SCCO patients showed increased expression of PD1 and PDL1. Novel treatment options can be used on a case-by-case basis depending on the genomic profile. Prospective studies need to be conducted for better results, but given the rarity of this disease, that may prove to be difficult. Abbreviations CISHChromogenic in situ hybridization FISHFluorescence in situ hybridization IHCImmunohistochemistry NET-ONeuroendocrine tumor of the ovary NGSNext generation sequencing NSCLCNon-small cell lung carcinoma SCCOSmall cell carcinoma of the ovary SCCO-HTSmall cell carcinoma of the ovary–hypercalcemic type SCLCSmall cell lung carcinoma Acknowledgements Not Applicable. Funding Not Applicable. Availability of data and materials Uploaded as supplemental material to Journal of Ovarian Research. Author contributions GY, PP, EW, and OHJ wrote the paper. RF provided data panel and assisted with research development. JCC and MP analyzed results and reviewed and edited the paper. MGM developed the research, analyzed results, reviewed and edited the paper, and was senior author. All authors read and approved the final manuscript. Competing interests Rebecca Feldman works for Caris Life Sciences. Consent for publication Not Applicable. Ethics approval and consent to participate We gained ethics approval for this research project through The University of Tennessee Health Science Center IRB in Memphis, Tennessee. Reference number is 15-04172-XM. ==== Refs References 1. Bahri M Lahmar R Ben Salah H Kallel N Ben Amar M Daoud J Small cell carcinoma of the ovary Cancer Radiother 2014 18 3 198 200 10.1016/j.canrad.2014.01.002 24637019 2. Harrison ML Hoskins P du Bois A Quinn M Rustin GJ Ledermann JA Small cell of the ovary, hypercalcemic type -- analysis of combined experience and recommendation for management. A GCIG study Gynecol Oncol 2006 100 2 233 8 10.1016/j.ygyno.2005.10.024 16321429 3. Jamy O Yaghmour G Hare F Martin MG Population-based analysis of the clinical features of primary small cell carcinoma of the ovary Anticancer Res 2015 35 5 3091 5 25964600 4. Karanian-Philippe M Velasco V Longy M Floquet A Arnould L Coindre JM SMARCA4 (BRG1) loss of expression is a useful marker for the diagnosis of ovarian small cell carcinoma of the hypercalcemic type (ovarian rhabdoid tumor): a comprehensive analysis of 116 rare gynecologic tumors, 9 soft tissue tumors, and 9 melanomas Am J Surg Pathol 2015 39 9 1197 205 10.1097/PAS.0000000000000475 26135561 5. Witkowski L Carrot-Zhang J Albrecht S Fahiminiya S Hamel N Tomiak E Germline and somatic SMARCA4 mutations characterize small cell carcinoma of the ovary, hypercalcemic type Nat Genet 2014 46 5 438 43 10.1038/ng.2931 24658002 6. Jelinic P Mueller JJ Olvera N Dao F Scott SN Shah R Recurrent SMARCA4 mutations in small cell carcinoma of the ovary Nat Genet 2014 46 5 424 6 10.1038/ng.2922 24658004 7. Ramos P Karnezis AN Craig DW Sekulic A Russell ML Hendricks WP Small cell carcinoma of the ovary, hypercalcemic type, displays frequent inactivating germline and somatic mutations in SMARCA4 Nat Genet 2014 46 5 427 9 10.1038/ng.2928 24658001 8. Jennifer K. Burzawa SZM, Sandeep K. Reddy, Igor A. Astsaturov, Robert L. Coleman, Jubilee Brown, Michael M. Frumovitz. Evaluation of biomarker alterations in small cell cervical cancer identifies therapeutic options. 2015. 9. McCullar B, Pandey M, Hare F, Patel K, Stein M, Yaghmour G, Feldman R, Chandler J, Martin M,. Genomic landscape of small cell carcinoma of the breast compared to small cell carcinoma of the lung. Breast Cancer Treatment and Research, Article in press. 10. Hahn AW Giri S Patel D Sluder H Vanderwalde A Martin MG Next-generation sequencing and in silico analysis facilitate prolonged response to pazopanib in a patient with metastatic urothelial carcinoma of the renal pelvis J Natl Compr Canc Netw 2015 13 10 1181 5 26483058 11. Arsenic R Treue D Lehmann A Hummel M Dietel M Denkert C Budczies J Comparison of targeted next-generation sequencing and Sanger sequencing for the detection of PIK3CA mutations in breast cancer BMC Clin Pathol 2015 15 20 10.1186/s12907-015-0020-6 26587011 12. PolyPhen-2: Prediction of Functional Effects of Human nsSNPs. Available at: http://genetics.bwh.harvard.edu/pph2/. Acccessed September 2015. 13. Erriquez J Becco P Olivero M Ponzone R Maggiorotto F Ferrero A TOP2A gene copy gain predicts response of epithelial ovarian cancers to pegylated liposomal doxorubicin: TOP2A as marker of response to PLD in ovarian cancer Gynecol Oncol 2015 138 3 627 33 10.1016/j.ygyno.2015.06.025 26100858 14. Thibault B Clement E Zorza G Meignan S Delord JP Couderc B F14512, a polyamine-vectorized inhibitor of topoisomerase II, exhibits a marked anti-tumor activity in ovarian cancer Cancer Lett 2016 370 1 10 8 10.1016/j.canlet.2015.09.006 26404751 15. Steffensen KD Waldstrom M Jakobsen A The relationship of platinum resistance and ERCC1 protein expression in epithelial ovarian cancer Int J Gynecol Cancer 2009 19 5 820 5 10.1111/IGC.0b013e3181a12e09 19574766 16. Fishel ML He Y Smith ML Kelley MR Manipulation of base excision repair to sensitize ovarian cancer cells to alkylating agent temozolomide Clin Cancer Res 2007 13 1 260 7 10.1158/1078-0432.CCR-06-1920 17200364 17. Dullea A, Marignol L. MGMT testing allows for personalised therapy in the temozolomide era. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine. 2015. 18. Sun S Shi W Wu Z Zhang G Yang BO Jiao S Prognostic significance of the mRNA expression of ERCC1, RRM1, TUBB3, and TYMS in patients with non-small cell lung cancer Exp Ther Med 2015 10 3 937 41 26622418 19. Ulker M Duman BB Sahin B Gumurdulu D ERCC1 and RRM1 as a predictive parameter for non-small cell lung, ovarian or pancreas cancer treated with cisplatin and/or gemcitabine Contemp Oncol 2015 19 3 207 13 20. Ferrandina G Mey V Nannizzi S Ricciardi S Petrillo M Ferlini C Expression of nucleoside transporters, deoxycitidine kinase, ribonucleotide reductase regulatory subunits, and gemcitabine catabolic enzymes in primary ovarian cancer Cancer Chemother Pharmacol 2010 65 4 679 86 10.1007/s00280-009-1073-y 19639316

PMC005xxxxxx/PMC5002198.txt

==== Front BMC Med Res MethodolBMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 21010.1186/s12874-016-0210-7Research ArticleAdapting the nominal group technique for priority setting of evidence-practice gaps in implementation science Rankin Nicole M. +61 2 9562 5324nicole.rankin@ctc.usyd.edu.au 1McGregor Deborah d.mcgregor@sydney.edu.au 12Butow Phyllis N. phyllis.butow@sydney.edu.au 34White Kate kate.white@sydney.edu.au 5Phillips Jane L. jane.phillips@uts.edu.au 6Young Jane M. jane.young@sydney.edu.au 78Pearson Sallie A. sallie.pearson@unsw.edu.au 9York Sarah sarah.york@sydney.edu.au 1Shaw Tim tim.shaw@sydney.edu.au 121 Sydney Catalyst Translational Cancer Research Cente, The University of Sydney, Level 6, 119-143 Missenden Road, Camperdown, NSW 2050 Australia 2 Faculty of Health Sciences, The University of Sydney, Sydney, Australia 3 Psycho-Oncology Co-operative Research Group, School of Psychology, The University of Sydney, Sydney, Australia 4 Centre for Medical Psychology & Evidence-based Decision-making, The University of Sydney, Sydney, Australia 5 Cancer Nursing Research Unit (CNRU), Sydney Nursing School, Sydney Local Health District and The University of Sydney, Sydney, Australia 6 Faculty of Health, University of Technology Sydney, Sydney, Australia 7 Sydney School of Public Health, The University of Sydney, Sydney, Australia 8 RPA Institute of Academic Surgery, Sydney Local Health District, NSW Ministry of Health, Sydney, Australia 9 Medicines Policy Research Unit, Centre for Big Data Research in Health, University of New South Wales, Sydney, Australia 26 8 2016 26 8 2016 2016 16 1 11014 7 2015 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 There are a variety of methods for priority setting in health research but few studies have addressed how to prioritise the gaps that exist between research evidence and clinical practice. This study aimed to build a suite of robust, evidence based techniques and tools for use in implementation science projects. We applied the priority setting methodology in lung cancer care as an example. Methods We reviewed existing techniques and tools for priority setting in health research and the criteria used to prioritise items. An expert interdisciplinary consensus group comprised of health service, cancer and nursing researchers iteratively reviewed and adapted the techniques and tools. We tested these on evidence-practice gaps identified for lung cancer. The tools were pilot tested and finalised. A brief process evaluation was conducted. Results We based our priority setting on the Nominal Group Technique (NGT). The adapted tools included a matrix for individuals to privately rate priority gaps; the same matrix was used for group discussion and reaching consensus. An investment exercise was used to validate allocation of priorities across the gaps. We describe the NGT process, criteria and tool adaptations and process evaluation results. Conclusions The modified NGT process, criteria and tools contribute to building a suite of methods that can be applied in prioritising evidence-practice gaps. These methods could be adapted for other health settings within the broader context of implementation science projects. Electronic supplementary material The online version of this article (doi:10.1186/s12874-016-0210-7) contains supplementary material, which is available to authorized users. Keywords Health prioritiesImplementation scienceMethodologyHealth services researchLung neoplasmshttp://dx.doi.org/10.13039/501100001171Cancer Institute NSW (AU)Sydney Catalyst TCRCRankin Nicole M. issue-copyright-statement© The Author(s) 2016 ==== Body Background Setting priorities in health services delivery and research is highly relevant in an era of scarce resources. In many settings, health researchers and clinicians participate in collaborative efforts to determine how best to implement evidence into routine clinical practice. Priority setting offers an important opportunity for collaborators to determine what evidence is relevant to implement, how best to do this, and where the greatest gains can be made in changing clinical practice [1]. There is a need to determine the most useful methods for eliciting priorities. Methods for priority setting in service delivery are already well developed [2, 3] but little attention has focused on how these methods could be usefully adapted for implementation science projects. Few studies have addressed methods that specifically prioritise existing gaps between research evidence and clinical practice [1]. To date, only one cluster randomised controlled trial (RCT) in setting priorities for healthcare improvement in community settings has been published [4, 5]. The patient involvement intervention for chronic diseases was based on the Nominal Group Technique (NGT) and trialled in community workshops to set priorities with participants (professionals alone in the control group or professional and patients in the intervention group) [5]. The study found that the NGT intervention was effective and patient involvement significantly shifted priorities towards quality of care factors. The NGT is an interpretive approach that engages stakeholders in interactive discussion to generate priorities [6, 7]. NGT has demonstrated validity, emphasizes considering all participants’ views equally and enables consensus on highly complex issues [8]. Consensus methods such as the NGT seek to overcome group or committee decision making that can be dominated by individuals or coalitions who have a vested interest in a specific outcome [9]. NGT is frequently used for problem identification; in helping to generate appropriate research questions; for development of solutions; and establishing priorities for action [8]. In cancer research, the NGT and modified Delphi techniques have been used to determine priorities in pancreatic cancer [10], haematological cancers [11, 12], colorectal cancer [13], colorectal cancer care coordination [14], psycho-oncology [15], cancer genomics [16] and with consumers in UK treatment centres [17]. To our knowledge, no studies have focused on methods for setting priorities about identified evidence-practice gaps in lung cancer care. Furthermore, there are few guides or resources specific to priority setting in ‘evidence to practice’ research [2, 18]. Our group had separately developed a set of seven evidence-practice gaps in lung cancer on the basis of a scoping review of the literature [19] and examining local data sources such as the Clinical Cancer Registry of New South Wales (NSW), Australia. Our primary aim was to test the relevance of these identified gaps with lung cancer health professionals using existing evidence-based approaches to priority setting (Additional file 1). Our second aim was to systematically develop a suite of robust, evidence based techniques, criteria and tools for use in implementation science projects. This paper describes the outcomes of the second aim. We show how the selected technique is relevant to implementation science, as most theoretical models and frameworks include pre-implementation engagement of stakeholders as a fundamental step in change processes [20–22]. The resulting tools and process were subsequently used in three focus groups with multidisciplinary teams (n = 42 participants) to fulfil our primary aim. Data outcomes from the focus groups are reported in a separate paper [23]. Methods Selecting priority setting techniques We conducted a literature search of journal articles and reports that specifically described priority-setting techniques in health research. We conducted three search strategies in Medline, PsychInfo and PubMed databases to:identify relevant health priority setting approaches or techniques that had been used in focus or small group settings; identify original research studies conducted in oncology settings, where oncology health professionals and/or consumers were the target audience for priority setting; and, locate relevant documents in the grey literature (i.e. reports and publications that may not be peer-reviewed) and web-based sources about group decision making tools used in health. This last search included checking the Cochrane Library (including the publication list of the Cochrane Agenda and Priority Setting Methods Group). For all three search strategies, the inclusion criteria were those articles or sources published between January 2008 and December 2012 and published in English. The exclusion criteria were any studies that had a primary aim of conducting a comparative benefit or cost analyses of health services; we also excluded commentaries or letters to editors. Selecting criteria The team conducted a literature search to identify and review existing criteria for rating evidence-practice gaps. We used the same databases and inclusion and exclusion criteria as described in the previous section. There are many and varied criteria used to assess priorities, which are generally categorised into three dimensions: public health benefit, feasibility and cost [18]. We focused on the first two dimensions and criteria options were collated from a range of sources [24, 25]. Cost criteria were not included as we had already excluded comparative benefit or cost analyses of health services in the previous section. Consensus approach to selecting relevant technique and criteria A consensus process was undertaken to select: a) the most relevant technique; and, b) criteria to use in the rating tools. We formed an expert interdisciplinary consensus group comprising of cancer clinician-researchers and academic experts from multiple disciplines including nursing, psychology, health services research, epidemiology and clinical oncology. Results from the literature reviews were presented. To select the most relevant technique, the consensus group considered the relative advantages, disadvantages and/or the applicability of each one. They considered how optimal and practical each technique was in terms of equity of participation encouraged, time required to enact and the group size, stakeholder mix of disciplines and level of expertise required. To select the criteria for use in the rating tools, the consensus group focused on whether there was a clear definition for each criteria and how easily participants would be able to rate each gap against the criteria. The group discussed the criteria options until consensus was reached. Pilot testing of the process and tools A pilot was conducted with key stakeholders (n = 7, including two medical oncologists and 5 health service researchers) to test the tools and the priority setting process. Minor adjustments were made to the content and formatting of the tools based on participant feedback. Participants indicated a maximum time of two hours should be allocated for the priority setting process. The pilot test confirmed that the priority setting technique and criteria were acceptable and feasible within this time-frame. Process and consensus evaluation We asked the three lung cancer multidisciplinary teams (n = 42) to complete a brief 8-item survey to evaluate the process and provide feedback about whether consensus had been achieved. Results The results are described in the following sections: selecting the technique; selecting the criteria; the resulting tools and the modified NGT process. Selecting the technique: outcomes of the literature search A total of 155 journal articles were identified from the first two search strategies; the abstract for each article was checked for relevance. Twenty-two articles relevant to the review were identified, which included 14 describing frameworks or techniques and eight that were original research articles in oncology. In the third strategy, five reports and two e-bulletins/newsletters from website sources were judged as relevant (see Fig. 1).Fig. 1 Literature review strategy for priority setting techniques The following techniques were identified: the Nominal Group Technique (NGT) [2]; the Delphi technique [10]; the Six-Steps of Listening for Direction [18]; the Combined Matrix Approach (3D CAM) [26] and Dotmocracy [27]. Through consensus, it was agreed that the NGT would be most relevant to modify for the following reasons. It can engage clinicians and researchers in face-to-face structured priority setting groups. This approach allows information to be collected from health professionals working in ‘front line’ clinical areas [8]. NGT is time efficient and cost effective, as priorities can be elicited in a short time period (e.g. 2 h) and requires few resources, which appeals to busy clinicians and resource-limited research budgets [28]. The technique requires no preparation on the part of participants and yet their expertise is of fundamental importance in considering the issues for prioritization. Another advantage is that the results generated can be immediately shared with the group. This provides ample opportunities for further reflection and clarification before the priority setting group session concludes [28]. This is particularly relevant in settings where membership of clinical teams frequently changes. Selecting criteria: outcomes of the literature review and consensus Using this same strategy described above, we identified seven journal articles and reports that listed relevant criteria options and three guides to consensus processes for selecting criteria [2, 18, 22]. Criteria options considered by the investigator team included those listed in: a) the Institute of Medicine (IOM) quality criteria (safety, effectiveness, patient-centered, timely, efficiency and equity) [29]; b) the 3D CAM criteria (disease burden, determinants of health, present level of knowledge, effectiveness and resource flows) [26], and c) the behavioural matrix in the Precede-Proceed model criteria (importance and changeable) [30, 31]. Through the consensus process, four criteria were selected against which the gaps could be rated. These were: 1) relevance to the local setting; 2) magnitude of the issue; 3) burden of suffering; and 4); amenable to change. Criteria 1-3 (relevance, magnitude and burden) were grouped under the heading of ‘significance’, so that results could later be collated and summarised more efficiently. The resulting tools Tool 1: the priority setting matrix This tool is used to elicit individual responses from participants, rating the gaps against the criteria using a matrix scoring sheet (Table 1). We adapted this tool from the 3D CAM [26]. The priority setting matrix places the evidence practice gaps in the left hand column and the criteria on the right. The matrix tool enables participants to privately rate each gap on a Likert scale ranging from 5 (most) to 1 (least) according to the four criteria. Likert scales have been used successfully in the past to rate priority items [32]. It should be noted that evidence practice gaps statements should be independent from each other and without multiple clauses.Table 1 Evidence-practice gaps priority setting matrix Least Most 1 2 3 4 5 Criteria: significance Criteria Evidence-practice gaps Relevance to local setting Magnitude of the gap (size) Burden of suffering (severity) Amenable to change 1. Not all people with lung cancer receive timely diagnosis and referral for treatment 2. People with potentially curable lung cancer who will benefit from active treatment do not always receive it 3. People with advanced lung cancer who will benefit from palliative treatment do not always receive it 4. People with lung cancer who are of an older age or with co-morbidities who may benefit from active treatment do not always receive treatment 5. People with lung cancer who would benefit from review at a multidisciplinary team meeting are not always reviewed 6. People with lung cancer have high levels of psychosocial needs which are not always being met 7. Not all people with lung cancer who would benefit from early referral to palliative care services are offered this option 8. Locally identified gap: Instructions for use: Please rate each evidence-practice gap in the left hand column according to the criteria in the right hand columns. Please rate each one according to a scale from 1 to 5 The same matrix tool can be used to elicit group responses for the purpose of discussing each gap and reaching consensus, and subsequently giving a rating to each one [23]. A ‘Dotmocracy’ approach is used to highlight common themes and reach consensus [27]. Gold and silver stars are allocated by each participant to reflect their highest priorities. Through facilitated discussion, the group is asked to discuss their individual ratings in a small group with each person nominating their top two (first ‘gold’ and second ‘silver’) gaps per criteria. This component is the small group trigger for discussion between participants of justifications for their ratings. Using the Dotmocracy approach can help small groups to determine if participant priorities align and to respond as appropriate until consensus is achieved. Tool 2: validating priority selection through an investment exercise Each participant receives $100 fake ‘dollars’ to spend across the gaps. This seeks to allow participants to make a global decision on resource allocation and also confirms whether the gaps are reflected in how participants would most like to invest in each gap. This acts as a cross check with the priorities selected in the matrix. The modified NGT process Table 2 shows the original NGT steps and the modified NGT steps and a brief description for each.Table 2 Comparison of original NGT steps and the modified process and brief descriptions Original NGT Modified NGT for priority setting Steps Brief description Steps Brief description Step 1: Generating ideas Moderator directs participants to write their ideas in brief phases or statements Step 1: Describe identified evidence practice gaps Presentation about the evidence-practice gap literature review, with a brief summary for each gap provided Step 2: Present local data/information about the gaps Presentation about national, jurisdictional and local data gathered to support the gaps Step 2: Recording idea Round robin feedback session to concisely record each idea Step 3: Elicit feedback and record additional gaps identified by participants Elicit feedback about relevance and appropriateness of evidence-practice gaps in the local service setting. Opportunity for participants to nominate additional local gaps Step 3: Clarify, rank ideas Participants express relative importance of each idea Step 4: Individuals vote privately to prioritise the ideas, using moderator-created criteria Participants privately rate each gap Step 4: Individuals vote privately to prioritise gaps, using moderator-created criteria Participants privately rate each gap using Likert scale on the matrix tool Step 5: Each participant selects the five most important items from the prioritised list Each participant ranks top five ideas, with the highest receiving 5 and lowest 1 Step 5: Each participant selects the two most important gaps from the prioritised list Each participant ranks top two gaps, with the highest receiving 2 and the lowest receiving 1 Step 6: Moderator creates tally sheet The most highly rated ideas are the most favoured actions Step 6: Focus group participants discuss ratings and moderator uses matrix tool as a tally sheet In focus groups, participants share their ratings, speaking in turn to list their top two gaps and provide any clarification for their choices. Responses are recorded by a group facilitator on the matrix sheet in ‘Dotmocracy’ style Step 7: Whole group consensus Small groups reform back into a larger group to review and discuss the gaps and resolve any differences to reach consensus Step 8. Investment exercise Each participant asked to spend 100 fictitious dollars across each gap. Dollars are tallied and feedback provided to whole group The modified NGT process developed for this project was subsequently used in focus groups and is described here as an exemplar. (Please note that the results of the focus groups are described in a separate paper [23]):Step 1: Describe identified evidence practice gaps In this step, an investigator team member (facilitator) gave a presentation about the approach used to review evidence about gaps in lung cancer care, read out a brief summary for each gaps and answered any participant questions. Step 2: Present local data/information about gaps This step included a presentation about national, jurisdictional and local data gathered to support the gaps. Where available, this was tailored to the local site to include analysis of local health district or hospital data collected in registries or administrative databases. This step was an adaptation of Lomas’ second step in the ‘listening model’ [6]. Step 3: Elicit feedback and include additional gaps identified by stakeholders Participants were asked to provide feedback about the seven gaps and whether they were relevant in the local hospital setting and surrounding community. We provided an opportunity for participants to nominate any additional gaps and this was considered as elicitation and inclusion of ‘practice-based evidence’. Additional gaps were only included if consensus was reached within the group that the gap should be added. Step 4: Individuals vote privately to prioritise gaps, using moderator-created criteria The criteria were described by the facilitator and participants had an opportunity to seek any clarification before privately rating each gap. Participants were then asked to rate each gap for across the criteria listed above without sharing these with others. This was considered an important step as this private rating of items allowed participants to make their own considered judgements prior to reaching group consensus. This step also ensured that every member of the group had read and considered each gap in detail. Step 5: Each participant selects the two most important gaps from the prioritised list. At the conclusion of step 4, each participant was asked to select the two most important gaps from on the matrix tool for the two criteria of ‘significance’ and ‘amenable to change’, according to their highest scores. This step replicates the original NGT step 5 but is modified by selecting only two most important items instead of five. Step 6: Focus group participants discuss ratings and moderator uses matrix tool as a tally In this step, participants broke into small groups (maximum of seven participants in each) and shared their ratings for each gap. We pre-assigned participants into small groups that consisted of a mix of professional groups. Each group was asked to nominate a spokesperson to report back to the larger group and provide feedback about the choices made. To achieve equity within small groups, each person spoke in turn about their top two gaps and provided any clarification for their choices. All responses were recorded by a group facilitator on an A4/US letter-sized printed sheet of the matrix tool in ‘Dotmocracy’ style [27], assigning a gold and silver star to the highest priority gaps. Priorities were not weighted across the chosen top two gaps. Step 7: Whole group consensus At the conclusion of discussion, small groups reformed back into a larger group, compared ratings for differences and anomalies and resolved these to reach group consensus. The moderator led discussion about the gaps against the criteria of significance and amenability to change. Step 8. Investment exercise In the final step, participants were asked to spend $100 fake dollars across each gap, configured as five $20 notes. Eight A3/US Ledger-sized envelopes were placed on a table (one for each gap, including the eighth local gap) and participants were asked to spend their money in any way they might choose. Participants were able to configure the dollars in any way so long as all $100 was allocated (e.g. all $100 in one envelope or any configuration of $20 allocations across the envelopes). At the conclusion of the task, the investigator team tallied the dollars in each envelope and then reported back to the group about the proportions (total dollar values and per cent) spent on each gap. Participants’ feedback in the lung cancer exemplar As a result of testing the methodology with the three lung cancer teams, we observed the following.Tools: The matrix tool was easy to complete and the Likert scale enabled participants to examine how they had comparatively rated each gap. By first completing the matrix tool privately, individual participants had an opportunity to focus on the evidence gaps and consider their responses according to the criteria. Using the same matrix for the group consensus process facilitated discussion in the small group setting. The investment tool of $100 fake dollars allocated across the gaps was a useful way to double check priorities after reaching consensus. The research team also observed that the investment exercise was a fun and engaging way to end the group session, with participants interested in the outcome and whether it was reflective of whole group consensus. Criteria: We found that the ‘amenable to change’ criteria was particularly meaningful for participants, as it allowed them to consider whether their current health service was capable of making change in the ‘gap’. Participation was a useful exercise for reflecting on current practice and what would be desirable to change in the future. Process evaluation: Thirty (of 42) (71 %) participants completed the evaluation survey. Results are shown in Table 3. There was strong agreement from participants that the priority setting process enabled full participation (93 %); was clearly communicated by the project team (86 %) and had adequate resources and time allocated (83 %). We asked participants if consensus of priorities had been reached via the process: 73 % responded ‘definitely yes’. Three participants provided a comment about the process: ‘fantastic brainstorming session; a very useful exercise; this was very novel and engaging’.Table 3 Process and consensus evaluation for three focus group participants (n = 30) Definitely yes Somewhat Definitely no Don’t know Process N % N % N % N % Was the priority setting group scheduled and conducted in a way that enabled you to fully participate? 28 93 2 7 - - - - Did the project team clearly communicate information associated with the priority setting process? 25 83 5 17 - - - - Were adequate resources and time allocated to properly completing the priority setting process? 24 80 5 17 1 3 - - Would you be willing to participate in future priority setting activities? 22 73 7 23 - - 1 3 Was this priority setting process beneficial in terms of identifying gaps in your local area? 21 70 8 27 - - 1 3 Consensus - - Do you agree with the priorities that have been identified via this process? 22 73 7 23 - - 1 3 In your opinion, are the priorities selected during this process representative of the broader views of cancer care stakeholders in your local area? 20 67 10 33 - - - - Will participation in this priority setting process lead to any changes within your organisation? 6 20 16 54 1 3 7 23 Discussion This study aimed to describe the methods for selecting and adapting a priority setting technique, selecting criteria and developing tools for implementation science projects. We sought to modify the NGT and relevant tools to allow the identified gaps to be rated according to agreed criteria. We believe this is the first study to demonstrate how to modify the NGT in setting priorities for evidence practice gaps within the context of implementation science projects. The selected technique, criteria, tools and process have drawn on evidence based approaches and we believe these will be highly relevant in other health care settings. A similar process was reported by Buckley and colleagues that assessed the effects of a research prioritization process to influence the research agenda in urinary incontinence [1]. Their team also noted that consensus methods for priority settings are still being developed and questions remain about whether this will impact on research priorities. We consider that there is significant value in prioritising gaps in research evidence and where change should be targeted in clinical settings. Gaining stakeholder engagement is key in building effective relationships for adoption of intervention strategies [21]. The advantages of the NGT made it the most relevant choice for the context of this study and we noted many similar advantages to those reported by Harvey and Holmes in priority setting about pregnancy triaging in emergency departments [8]. We selected it over other techniques including a modified Delphi approach, which has been used in previous cancer research [10, 14, 16], as we sought to reach consensus about priority gaps in a workshop setting. The selected criteria enabled our team to seek participants’ views about what gaps were significant and amenable to change in their local setting, drawing on Green and Kreutzer’s behaviour matrix from the Precede-Proceed model [30]. It is worth reflecting on the modifications made to the NGT process. We modified the first two steps of NGT (generate and recording ideas) to become a presentation of an evidence-practice gaps summary, as the investigator team had already undertaken a significant amount of work to identify relevant gaps. By doing so, we sought to address the documented barrier of clinicians having insufficient time to read and assimilate research evidence [33]. Step 2 focused on presenting data at the jurisdictional (NSW) and local levels (local health district or hospital level) to support the selection of gaps. This adaptation draws on Lomas’ step of identifying and assembling data needed to help inform participants’ discussion [6] and we consider that this is an important component in appealing to a clinical audience whose priorities are likely to focus on the technical aspects of disease management [4] and epidemiological data. Step 3 provided participants with an opportunity to discuss the gaps and elicit any additional gaps for the local setting. This step allowed for group discussion and generation of new gaps of relevance. We replicated Steps 4 and 5 of the original NGT model to be a private ranking test of items, however, in Step 5 we reduced the number of top gaps from five to two. This reduction was due to time constraints, and with having eight gaps to assess, we considered this a reasonable adjustment to make to the process. An alternative is to maintain original NGT step to rank the 5 top items or to rank items on a Likert scale. Our modifications include additional opportunities for group discussion (Steps 6 and 7). In a subsequent study to test success factors about coordinated cancer in care [34], we found in Step 6 that rating items using the top three priorities as scores (3 = top rated item, 2 = second highest item, 1 = third item) was an easier method for reporting results where there were multiple break out groups and reaching whole group consensus. The addition of the investment exercise in Step 8 was helpful way of reflecting on the group’s priorities. In the event that this step reveals different priorities, we recommend considering a return to the group consensus process (Step 7) to determine with the large group how best to resolve divergent responses. Facilitators who are faced with a lack of consensus amongst participants should focus on the provision of an ‘equal voice’ for all participants, rather than expecting a specific outcome. While reaching consensus may be a lengthy process and contentious, there is limited evidence to suggest that gaining stakeholder buy-in results in increased likelihood of success. This is preferable in health settings to ‘majority rules’ where dominant views overrule the minority, creating ‘winners and losers’ thus potential disenfranchising participants and leading to poorer outcomes. Study limitations and future directions The limitations of this study are that we modified the NGT process and tools with a specific purpose of testing evidence-practice gaps in lung cancer with multidisciplinary teams of clinicians working across three hospital settings. Whilst we have used a robust and evidence-based approach to reviewing existing priority setting techniques, it is by no means the only option. Researchers will seek to adapt techniques, criteria, tools and processes to suit their own research aims and local context. However, given the paucity of tools available to researchers working in implementation science [35], we consider that the methods described in this paper make a significant enhancement to the knowledge base. Since conducting this study, we have been approached by three research teams to use this methodology in different health settings. We have utilised this methodology in a project to define success factors for coordinated care in cancer [34] on behalf of the Cancer Institute NSW. Future directions include opportunities to replicate the process with consumers, or with combined groups of health professionals and consumers, similar to the cluster RCT conducted by Boivin and colleagues [4]. Conclusions In conclusion, this study shows that existing priority setting techniques and tools can be successfully adapted for use in implementation science projects. The modified NGT process and tools can be used by collaborations of researchers and clinicians who seek to identify priorities for research within a particular health condition. The modified NGT approach is practical, user-friendly and easy to conduct in small groups of clinicians. The study provides a methodology for how to prioritize evidence-practice gaps in other health settings, within the broader context of implementation science projects. Additional file Additional file 1: An exemplar of evidence-practice gaps in lung cancer. (DOCX 14 kb) Abbreviations RCTRandomised controlled trial NGTNominal group technique NSWNew South Wales 3D CAMCombined matrix approach IOMInstitute of medicine Acknowledgements We would like to thank the lung cancer health professionals who participated in this project and the lead clinicians and managers at clinical sites, Professor David Barnes, Dr Emily Stone, Dr Rob Zielinski and Ms Ruth Jones, for facilitating the study and enabling access to multidisciplinary lung cancer teams. Funding This project was funded by Sydney Catalyst Translational Cancer Research Centre flagship funding grant (2012_CIA_Shaw). Sydney Catalyst is funded by the Cancer Institute NSW through a Translational Cancer Research Centre Program Grant. Availability of data and materials The evaluation dataset generated and analysed during the current study is not publicly available due to the ethics approval condition to confidentially store data. However, these data are available from the corresponding author on request. Authors’ contributions NR, DM and TS drafted the methodological approach and all authors participated in its refinement. DM and NR conducted the literature reviews. NR prepared the manuscript. DM, PB and TS provided critical review of manuscript drafts. All authors and participated in revising the submitted manuscript and approved the final version. Competing interests The authors declare that they have no competing interests. Consent for publication All participants gave written informed consent prior to participating in the evaluation component of this study. Ethics approval and consent to participate Ethics approval was obtained from the University of Sydney Human Research Ethics Committee prior to study commencement (Protocol No. 2013/528). ==== Refs References 1. Buckley BS Grant AM Glazener CMA Case study: A patient–clinician collaboration that identified and prioritized evidence gaps and stimulated research development J Clin Epidemiol 2013 66 5 483 489 10.1016/j.jclinepi.2011.03.016 21816575 2. Campbell S. Deliberative Priority Setting: a CIHR KT module; Ottawa, ON. Canadian Institutes of Health Research;2010. 3. Smith N Mitton C Cornelissen E Gibson J Peacock S Using evaluation theory in priority setting and resource allocation J Health Organ Manag 2012 26 4-5 655 671 10.1108/14777261211256963 23115910 4. Boivin A Lehoux P Lacombe R Burgers J Grol R Involving patients in setting priorities for healthcare improvement: a cluster randomized trial Implement Sci 2014 9 24 10.1186/1748-5908-9-24 24555508 5. Boivin A Lehoux P Lacombe R Lacasse A Burgers J Grol R Target for improvement: a cluster randomised trial of public involvement in quality-indicator prioritisation (intervention development and study protocol) Implement Sci 2011 6 45 10.1186/1748-5908-6-45 21554691 6. Lomas J Fulop N Gagnon D Allen P On being a good listener: setting priorities for applied health services research Milbank Q 2003 81 3 363 388 10.1111/1468-0009.t01-1-00060 12941000 7. Delbecq AL Van de Ven AH Gustafson DH Group Techniques for Program Planning: a Guide to Nominal Group and Delphi Processes 1975 Dallas Scott-Foresman 8. Harvey N Holmes CA Nominal group technique: An effective method for obtaining group consensus Int J Nurs Pract 2012 18 188 194 10.1111/j.1440-172X.2012.02017.x 22435983 9. Jones J Hunter D Consensus methods for medical and health services research BMJ 1995 311 7001 376 380 10.1136/bmj.311.7001.376 7640549 10. Robotin MC Jones SC Biankin AV Defining research priorities for pancreatic cancer in Australia: results of a consensus development process Cancer Causes Control 2010 21 5 729 736 10.1007/s10552-010-9501-1 20072807 11. Clinton-McHarg T Paul C Sanson-Fisher R D’Este C Williamson A Determining research priorities for young people with haematological cancer: a value-weighting approach Eur J Cancer 2010 46 18 3263 3270 10.1016/j.ejca.2010.06.013 20634057 12. Paul C Sanson-Fisher R Douglas H Clinton-McHarg T Williamson A Cutting the research pie: a value-weighting approach to explore perceptions about psychosocial research priorities for adults with haematological cancers Eur J Cancer Care 2010 20 3 345 353 10.1111/j.1365-2354.2010.01188.x 13. Bajracharya SM An assessment of the perceived barriers and strategies to promoting early detection of colorectal cancer: a practitioners’ perspective Int Q Community Health Educ 2006 26 1 23 44 10.2190/13R7-5177-G833-8V85 17686712 14. Young JM Masya LM Solomon M Shepherd HL Identifying indicators of colorectal cancer care coordination: a Delphi study Color Dis 2013 16 17 27 10.1111/codi.12399 15. Rankin NM Butow PN Price MA Evans A Views of psycho-oncology health professionals on priority psycho-oncology research questions Support Care Cancer 2011 19 1133 1141 10.1007/s00520-010-0922-x 20535503 16. Esmail LC Roth J Rangarao S Getting our priorities straight: a novel framework for stakeholder-informed prioritization of cancer genomics research Genet Med 2013 15 2 115 122 10.1038/gim.2012.103 23037935 17. Corner J Wright D Hopkinson J Gunaratnam Y McDonald JW Foster C The research priorities of patients attending UK cancer treatment centres: findings from a modified nominal group study Br J Cancer 2007 96 6 875 881 10.1038/sj.bjc.6603662 17342090 18. Viergever RF Olifson S Ghaffar A Terry RF A checklist for health research priority setting: nine common themes of good practice Health Res Policy Syst 2010 8 9 10.1186/1478-4505-8-36 20370906 19. Rankin N, Shaw T, McGregor D et al. The relevance of implementation science to psycho-oncology: Evidence practice gaps in lung cancer. Psycho-Oncology. 2013;22(S3): 46. 20. Grol R, Wensing M, Eccles M, Davis D. Improving patient care: The implementation of change in health care. West Sussex, UK: Wiley Blackwell; 2013. 21. Graham ID Logan J Harrison MB Lost in Knowledge Translation: Time for a Map? J Contin Educ Health Prof 2006 26 13 24 10.1002/chp.47 16557505 22. Smith N Mitton C Peacock S Cornelissen E MacLeod S Identifying research priorities for health care priority setting: a collaborative effort between managers and researchers BMC Health Serv Res 2009 9 9 10.1186/1472-6963-9-165 19144138 23. McGregor D Rankin N Butow P Closing evidence-practice gaps in lung cancer: Results from priority setting in the clinical context Asia Pac J Clin Oncol 2016 27230524 24. Berra S Sanchez E Pons JMV Tebe C Alonso J Aymerich M Setting priorities in clinical and health services research: Properties of an adapted and updated method Int J Technol Assess Health Care 2010 26 2 217 224 10.1017/S0266462310000012 20392327 25. Ghaffar A Setting research priorities by applying the combined approach matrix Indian J Med Res 2009 129 4 368 375 19535830 26. Global Forum for Health Research. The 3D Combined Approach Matrix: An improved tool for setting priorities in research for health. Geneva, SUI: Global Forum for Health Research; 2009. 27. Public Health Ontario, Health Nexus Canada. Priority Setting – Four Methods for Getting to What’s Important. Ottawa, ON: Ontario Health Promotion E-Bulletin; 2010. 28. Potter M Gordon S Hamer P The Nominal Group Technique: A useful consensus methodology in physiotherapy research N Z J Physiother 2004 32 3 126 130 29. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the TwentyfirstCentury. Washington DC. National Academies Press; 2001. 30. Green LW Kreuter MW Health Program Planning: An Educational and Ecological Approach. 4th Edition ed 2005 McGraw-Hill Higher Education New York 31. Phillips JL Rolley JX Davidson PM Developing Targeted Health Service Interventions Using the PRECEDE-PROCEED Model: Two Australian Case Studies Nurs Res Pract 2012 2012 279431 22852076 32. Hermens R Ouwens M Vonk-Okhuijsen SY Development of quality indicators for diagnosis and treatment of patients with non-small cell lung cancer: A first step toward implementing a multidisciplinary, evidence-based guideline Lung Cancer 2006 54 1 117 124 10.1016/j.lungcan.2006.07.001 16920220 33. Grol R Grimshaw J From best evidence to best practice: effective implementation of change in patients’ care Lancet 2003 362 9391 1225 1230 10.1016/S0140-6736(03)14546-1 14568747 34. Shaw T York S Rankin NM McGregor D Hawkey A Success Factors in Coordinated Cancer Care. A report commissioned by the Cancer Institute NSW 2014 Sydney The University of Sydney 35. Grimshaw J Eccles M Thomas R Toward evidence-based quality improvement. Evidence (and its limitations) of the effectiveness of guideline dissemination and implementation strategies 1966-1998 J Gen Intern Med 2006 21 Suppl 2 S14 S20 16637955

PMC005xxxxxx/PMC5002199.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 357210.1186/s12889-016-3572-1Research ArticleProcess evaluation of a tailored mobile health intervention aiming to reduce fatigue in airline pilots van Drongelen Alwin a.vandrongelen@tno.nl 12Boot Cécile R. L. +31 20 444 9680crl.boot@vumc.nl 13Hlobil Hynek hhlobil.health@klm.com 12Smid Tjabe t.smid@vumc.nl 12van der Beek Allard J. a.vanderbeek@vumc.nl 131 Department of Public and Occupational Health, EMGO Institute for Health and Care Research, VU University Medical Center, PO Box 7057, 1007 MB Amsterdam, The Netherlands 2 KLM Health Services, Schiphol Airport, Schiphol, The Netherlands 3 Body@Work TNO VUmc, Research Center on Physical Activity, Work and Health, VU University Medical Center, Amsterdam, The Netherlands 26 8 2016 26 8 2016 2016 16 1 89425 1 2016 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 MORE Energy is a mobile health intervention which aims to reduce fatigue and improve health in airline pilots. The primary objective of this process evaluation was to assess the reach, dose delivered, compliance, fidelity, barriers and facilitators, and satisfaction of the intervention. The second objective was to investigate the associations of adherence to the intervention with compliance and with participant satisfaction. Thirdly, we investigated differences between the subgroups within the target population. Methods The intervention consisted of a smartphone application, supported by a website. It provided advice on optimal light exposure, sleep, nutrition, and physical activity, tailored to flight and personal characteristics. The reach of the intervention was determined by comparing the intervention group participants and the airline pilots who did not participate. The dose delivered was defined as the total number of participants that was sent an instruction email. Objective compliance was measured through the Control Management System of the application. To determine the fidelity, an extensive log was kept throughout the intervention period. Subjective compliance, satisfaction, barriers, facilitators, and adherence were assessed using online questionnaires. Associations between the extent to which the participants applied the advice in daily life (adherence), compliance, and satisfaction were analysed as well. Finally, outcomes of participants of different age groups and haul types were compared. Results A total of 2222 pilots were made aware of the study. From this group, 502 pilots met the inclusion criteria and did agree to participate. The reach of the study proved to be 22 % and the dose delivered was 99 %. The included pilots were randomized into the intervention group (n = 251) or the control group (n = 251). Of the intervention group participants, 81 % consulted any advice, while 17 % did this during four weeks or more. Fidelity was 67 %. The participants rated the intervention with a 6.4 (SD 1.6). Adherence was not associated with compliance, but was associated with satisfaction (p ≤ 0.001). Pilots of 35 to 45 year old were significantly more interested in advice regarding physical activity than their colleagues, and short-haul pilots were more interested in advice regarding nutrition compared to long-haul pilots. Conclusions The MORE Energy intervention was well received, resulting in an adequate reach and a high dose delivered. The compliance and satisfaction scores indicate that engagement and functionality should be enhanced, and the content and applicability of the advices should be improved to appeal all subgroups of the target population. Trial registration Nederlands Trial Register NTR2722. Registered 27 January 2011. Electronic supplementary material The online version of this article (doi:10.1186/s12889-016-3572-1) contains supplementary material, which is available to authorized users. Keywords Work schedule tolerancePrimary preventionTelemedicineImplementationMobile healthProcess evaluationVU University Medical Centerissue-copyright-statement© The Author(s) 2016 ==== Body Background Due to disruption of the sleep wake pattern and the circadian rhythm, fatigue is inevitable in occupations where individuals are required to work when they normally would be asleep [1]. In the aviation industry, fatigue management strategies have been developed to minimize the health effects of these irregular working hours [2, 3]. Education for flight crew members is an important component of these strategies, for which several educational programs have been developed. Although some of these programs have been studied, the effects and the optimal way to transfer the relevant knowledge remains largely unclear [4–8]. Literature on computerized health education shows that the content of advice should be tailored to the individual needs and should be applicable for all subgroups within a target population [9, 10]. Additionally, when translating the relevant flight crew related knowledge into practical advice, variables such as flight direction, flight duration, and number of time zones crossed should be taken into account [7]. Based on this knowledge, the MORE Energy intervention, aiming to reduce fatigue and improve health in airline pilots through easy obtainable and tailored advice, was developed [11]. The intervention provided participants with evidence-based and relevant fatigue-related advice using a mobile application (app), supported by a website with background information. The usage of mobile health (mobile phone technologies in health care and public health) has expanded rapidly during the last decade [12]. Additionally, because the use of smartphones and tablets increased enormously, apps showed to have great potential for promoting health behaviour [13]. Evidence for these effects of mobile apps is still limited, however [14–17]. In addition, Blackman et al. [18] showed that mobile health studies scarcely report about key implementation factors, while this information is necessary to get more insight in the strength and weaknesses of the implementation of the intervention and to facilitate the interpretation of the results [19–21]. We, therefore, performed a process evaluation alongside our randomized controlled trial. The primary objective of this process evaluation was to assess the reach, dose delivered, compliance (dose received), fidelity, barriers and facilitators (context), and satisfaction of the MORE Energy intervention. The second objective was to investigate whether the MORE Energy intervention was associated with an improvement in relevant behavior of the participants by exploring the association between compliance to the intervention and the extent to which the pilots adhered, i.e. applied the advice in daily life. We also investigated if adherence was associated with the satisfaction of the pilots. The third objective was to investigate how the intervention suited the different subgroups within the target population. Therefore, outcome differences between pilots of the different age groups and haul types were analysed for the process evaluation items compliance and satisfaction. Methods This process evaluation was carried out alongside the randomized controlled trial on the effectiveness of the MORE Energy intervention that aimed to reduce fatigue in airline pilots. The Medical Ethics Committee of the VU University Medical Center (Amsterdam, the Netherlands) assessed the study design and procedures, but according to Dutch law, this study proved to be exempt from a medical ethical review. Participants The study population consisted of the pilots of a large internationally operating airline company. The pilots could participate in the study if they were not on sick leave for more than four weeks at the moment of recruitment and if they owned a smartphone or tablet with iOS (iPhone/iPad Operating System) or an Android operating system. After inclusion, the participants were equally randomized into an intervention group, and a control group which received a minimal intervention. Because this process evaluation addresses the MORE Energy intervention, we focus on the participants of the intervention group only. MORE Energy intervention The MORE Energy intervention was developed systematically. First, a literature study was performed in order to gain insight in the latest scientific knowledge about optimal behaviour regarding disruption of the circadian rhythm and fatigue in flight crew. Next, focus groups were held to find out what medium and implementation strategy should be used to optimise compliance to the intervention. The focus groups made clear that the intervention should be easy available, appealing, and to be used by pilots of all ages and job types. Further, the advices should be made flight schedule specific and applicable for both short and long-haul pilots. To match the intervention with the legislation and the policy of the airline company, interviews with key management stakeholders were held as well. Based on the focus groups and interviews, it was decided to develop a mobile application to transfer the advices to the target population. After the development of the MORE Energy app, it was extensively pre-tested by both pilots and researchers. Based on the results of this first evaluation, the intervention was optimised where necessary. The MORE Energy app contained advices on optimal light exposure, sleep, nutrition, and physical activity, tailored to relevant flight (e.g. flight direction, departure and return time, number of time zones crossed) as well as to personal characteristics (e.g. job type, chronotype). The users could choose to consult background information in the glossary menu and the app guided them to a website to read more, or to view and listen to video and audio files concerning the different topics. Participants were encouraged to consider the advices on the app by means of two types of reminders: timed alerts (when the user did not use the app for longer than three weeks) and geofencing alerts (when the user arrived somewhere outside of the Netherlands, with a maximum of one alert per four days). Screenshots of the MORE Energy app can be seen in Additional file 1. Further details on the development, content and effect evaluation of the intervention have been published elsewhere [11, 22]. Data collection The process evaluation items were taken from the Steckler & Linnan framework [20]: reach, dose delivered, compliance (dose received), fidelity, barriers and facilitators (context), and satisfaction. Adherence, the extent to which the participants applied the advices in daily life, was measured as well. Table 1 presents an overview of the different items and the accompanying collection and processing of the data. The airline company provided data about the gender, age, job type, and haul type of all potential participants.Table 1 Overview of the different process evaluation items Items Definition Resources Reach Information on the number of participants (%) and their demographics, compared to the non-participants. Information on all potential participants provided by the airline company. Dose delivered Total amount of intervention material provided to the participants (%). The number of participants that was sent an email with instructions and login details. Compliance (dose received) Measured consultation of the tailored advice. Objective: user authentication through the CMS (app) and Google Analytics (website). Subjective: online questionnaire. Fidelity Information on all changes, updates, and revisions that happened with the app during the intervention period. Calculated as the weighted average of the percentage of weeks the different components of the intervention were delivered as intended. Log. Satisfaction Participants’ appreciation of the intervention and their opinion on its effectiveness (1–10). Online questionnaire. Barriers and facilitators (context) Barriers and facilitators of the intervention, experienced by both the researchers and the participants. Researchers: log. Participants: online questionnaire. Adherence The extent to which participants applied the MORE Energy advices in daily life. Online questionnaire. Reach Reach is defined as the proportion and representativeness of the intervention group participants in the study, compared to the total group of potential participants [19]. Reach was determined by comparing the following characteristics between the intervention group participants and the airline pilots that did not participate: gender, age, job type, and haul type. Dose delivered Dose delivered is considered as the total amount of intervention material provided to the participants. In this study, the dose delivered was defined as the total number of participants that was sent an email containing instructions and login details to access the intervention material. Compliance (dose received) Compliance is the dose that is received, and refers to the extent to which participants actively engaged with the intervention. In our study, it was objectively measured through the Control Management System (CMS) of the application. This system stored the number of advices per week requested by each participant through user authentication. Likewise, we used a web-analytic tool (Google Analytics) to register and store the total number of page views per participant to the website of the project. The registered number of app advices of four participants proved to be more than 200. Because this was most certainly due to malfunctioning of the CMS, the registered data of these participants was excluded from the objective compliance analyses. The participants were also asked how often they had consulted the advices during the intervention (almost always, sometimes, only a few times, or never) in the online questionnaire at six months after baseline. Further, participants were asked which type of advice they had predominately used (advice regarding preparation for departure, regarding layover, or regarding the return home) and which topics they had consulted the most (exposure to light, sleep, nutrition, or physical activity). Fidelity Fidelity is defined as the extent to which the intervention program was implemented as planned, representing the quality and the integrity of the implementation [19]. Therefore, all changes, updates, and revisions of the app and website that occurred during the intervention period were kept in a log. Fidelity was calculated as the weighted average of the percentage of weeks of the total intervention period that the different components of the intervention were delivered as intended. As the advice delivered through the app was considered the main component of the intervention, this was given the largest weight, whereas the remaining four components were weighted equally:Access (installation, login, offline functionality): 15 % Backend (synchronisation of content and flight schedules): 15 % Advice (tailoring algorithm and glossary): 40 % Reminders (functioning of push alerts): 15 % Website with background information (access, functionality): 15 % Satisfaction The satisfaction with the intervention was assessed through the online questionnaire at six months after baseline. First, the participants were asked to give an overall grade for MORE Energy (1 to 10). Next, they were asked to rate four statements about the usability of the intervention on a 5-point Likert scale ranging from ‘disagree’ to ‘agree’. Additionally, participants were asked if they would recommend the MORE Energy application to their colleagues, and to appreciate the effectiveness of the intervention through rating three statements on perceived effectiveness on the 5-point Likert scale. Barriers and facilitators (context) Context refers to “the larger physical, social, and political environment that either directly or indirectly affects an intervention program” [20]. Possible context factors that affected the intervention were registered in a log. We also asked the participants which barriers or facilitators they had experienced. First, participants were asked if they would recommend the MORE Energy application to colleagues who did not have access to it yet. If they answered ‘no’, they were asked why they held that opinion. Next, participants were asked what their reasons were not to consult the advices more often (content already known, no need for further consultation, technical problems, lack of usability, or another reason). Adherence The extent to which participants applied the MORE Energy advices in daily life was assessed through asking the participants to rate the statement “After reading the advices, I actually applied them as well.” on a 5-point Likert scale ranging from 1 ‘disagree’ to 5 ‘agree’. Data analysis Regarding the first objective, descriptive analyses were performed. Differences (gender, age, job type, and haul type) between participants and non-participants were analysed with t-tests for independent samples and Chi-square tests. For the second objective, associations between compliance and the extent to which the participants applied the advice in daily life (adherence) were analysed by calculating Spearman's (rho) correlation coefficients. Regarding the objective compliance, participants were divided into four groups of equal size related to the amount of compliance. Furthermore, a linear regression analysis was performed to explore the association between the level of adherence (independent variable) and satisfaction with the intervention (dependent variable). Participants who indicated not to have applied the advices in daily life, or who had a neutral opinion towards this question, were used as one reference category in the analysis. To answer the third objective, outcomes for the different age groups (<35, ≥35- < 45, ≥45) and haul types (short-haul vs. long-haul) on compliance and satisfaction were analysed with t-tests for independent samples, Chi-square tests and one-way ANOVAs. Participants who flew to intercontinental destinations were considered long-haul pilots, while participants who only flew to European destinations were considered short-haul pilots. A two-tailed significance level of p < 0.05 was considered to be statistically significant in all analyses. Analyses were conducted with the Statistical Package for Social Sciences (SPSS), version 20.0. Results Reach A total of 2222 potential participants were made aware of the project by means of a publicity campaign through the airline company, after which they received an email with a link to the baseline questionnaire. From this group, 522 (23 %) pilots agreed to participate. A total of 20 pilots (<1 %) did not meet the inclusion criteria. Of the remaining 502 pilots, 251 were randomized into the intervention group, and 251 into the control group. At the end of the intervention period, 148 (59 %) of the participants of the intervention group completed the process evaluation questionnaire. Baseline characteristics of the participants and the non-participants are shown in Table 2. Between the intervention and control group, no significant differences were present. The intervention group participants proved to be significantly younger than the non-participants, and the percentage of females was higher.Table 2 Reach of the MORE Energy intervention; characteristics of the participants and the non-participants Reach characteristics Category Intervention group (n = 251) Control group (n = 251) Non-participants (n = 1720) Age in years mean (SD) 41.0 (8.0)* 40,7 (8,7) 42.5 (8.3) Age group n (%) <35 60 (23.9)* 66 (26.3) 313 (18.2) ≥35- < 45 98 (39.0)* 89 (35.5) 652 (37.9) ≥45 93 (37.1)* 96 (38.2) 755 (43.9) Female n (%) 21 (8.4)* 13 (5,2) 65 (3.8) Job type n (%) Captain 111 (44.2) 113 (45,0) 750 (43.6) First Officer 97 (38.7) 96 (38,2) 706 (41.0) Second Officer 43 (17.1) 42 (16,7) 264 (15.3) Haul type n (%) Long-haul 179 (71.3) 179 (71,3) 1287 (74.8) Short-haul 72 (28.7) 72 (28,7) 433 (25.2) * Significant difference between the participants and non-participants (p < 0.05) Dose delivered We sent emails containing the login details and instructions to 251 participants. One email was bounced because the email address proved not to exist. After three months, one other participant reported to have not received the instruction email. The dose delivered therefore was determined to be 99 % (249/251). Compliance (dose received) Objective compliance It was registered that during the intervention period, 54 of the 251 participants never consulted any advice on the app. Two of these participants indicated that they wanted to drop out of the study. Five of them only consulted the website with the background information. The remaining 47 participants either did not download or use the app or website after receiving the instructions, did not receive or read the email with the instructions, or only used the glossary section of the app (consultation of the glossary was not registered by the CMS). During the six month intervention period, 68 (27 %) participants consulted the advices during one week only, 54 (22 %) consulted them during two weeks, and 32 (13 %) consulted them during three weeks. A total of 43 (17 %) participants consulted the advices on the app during four weeks or more. In total, 1677 advices were requested. The mean number of requested advices per participant was 6.8 (SD 14.0), while the median was 3 advices per participant. If the data of the four participants with outliers was included, the mean number of requested advices would have been 12.6 (SD 55.3). The CMS registered that the advices regarding the preparation for departure from home were requested the most (49 %). The advices concerned with time spent during layover (23 %) and the advices about the preparation for the return flight and arrival home (27 %) were consulted less often. In total, 32 (13 %) participants went to the website with background information. The mean number of page views of these participants was 9.2 (SD 8.6). It was determined that 27 (11 %) participants used both the app and the website. Most of the participants did use the app but never logged on to the website (68 %). Subjective compliance Of the 148 participants that answered the process evaluation questionnaire, 62 (42 %) indicated they had never really used the MORE Energy advices, 39 participants (26 %) indicated that they had used the advices a few times, and 46 (31 %) reported to have used the advices occasionally. One participant reported to have used the advices before and during every flight. Next, of the 86 participants that had used the advices, most of them indicated that they had consulted the advices regarding layovers (70 %). Next, 49 % indicated to have consulted the advices regarding the return home, and 36 % indicated to have consulted the advices concerned with preparation before departure. Further, most of these participants pointed out that they consulted the advices concerned with sleeping behaviour (62 %) and nutrition (72 %). The advices regarding exposure to light (28 %) and physical activity (23 %) were consulted less often. Fidelity Through the six month intervention period the implementation of the intervention was predominantly affected by the following (a detailed explanation of all bugs can be found in Additional file 2)After two weeks the second version of app became available for iOS: after downloading the update, participants could no longer consult all types of advices and iPhone5 users could not get access to the updated version of the app. After two months the third version of app became available for iOS to solve the problems above:∘ This version could not be installed automatically. Participants had to delete the previous version of the app before being able to install the new one. ∘ After installation of third version, reminder alerts malfunctioned. Researchers found out this problem had most probably occurred in the Android app from the start of the intervention as well. The five components of the intervention were affected to a different extent. Table 3 shows the number of weeks the components could be delivered as intended, and the calculation of the different fidelity scores. The weighted total fidelity score of the intervention was 67 %.Table 3 Fidelity score calculation for the 26-week intervention period Component Weight factor Weeks delivered as intended Fidelity score App Access 15 % 16 62 % Backend 15 % 25 96 % Advice 40 % 19 73 % Reminders 15 % 0 0 % Website Access 15 % 25 96 % Total fidelity score 67 % Satisfaction Figure 1 shows the satisfaction of the participants who completed the process evaluation questions and indicated that they had used the advices of the MORE Energy app. The participants were satisfied with an app as the medium to transfer the advices (>75 % agreed) but less than half of the participants found that the advices given were easy to apply in their daily life (46 %). 56 % of the participants indicated that the MORE Energy advices were useful.Fig. 1 Perceived satisfaction with the MORE Energy smartphone application More than 65 % of the participants that used the advices thought that the MORE Energy intervention could prevent fatigue and improve fitness. Further, 78 % indicated that they would recommend the MORE Energy advices to their colleagues (data not shown). On average, the participants rated the MORE Energy intervention with a 6.4 (SD 1.6). Barriers and facilitators (context) The 86 pilots that indicated to have used the advices were asked what their reasons were not to consult the advices more often. The most selected reason was that the content of the advice was already known to them (58 %). Furthermore, 44 % of the participants indicated that they did not need to consult the advices anymore after a few times. Less selected reasons were technical problems with the app (12 %) and lack of usability (7 %). Out of the additional reasons reported by 24 participants, three main themes could be composed. First, nine participants reported that the advices could not be applied in daily life because that would conflict with their social obligations, both at home and during duty. Predominantly participants with young children and short-haul pilots pointed out this problem. Four participants indicated that they had simply forgotten to consult the app, or that they did not receive an alert to remind them. Similar barriers were addressed by the 19 participants who reported that they would not recommend the MORE Energy app to their colleagues; the advices were too common or not innovative enough (n = 5), the advices were not applicable in daily life (n = 4), the content was already known (n = 3), or the app had too much technical problems (n = 2). Association between compliance, adherence, and satisfaction Of the participants indicating to have used the advices during the intervention period, 47 (55 %) somewhat agreed, and 14 (16 %) agreed with the adherence statement that they applied the advices after consulting them. Further, 25 (29 %) participants somewhat disagreed or had a ‘neutral’ opinion towards the statement. Spearman’s rho correlation between the objectively and subjectively measured compliance and adherence was 0.04 (p = 0.71) and 0.30 (p = 0.004), respectively. Table 4 shows that participants who indicated to have applied the advices after consulting them rated the intervention significantly higher compared to the participants of the reference category. The participants who somewhat agreed with the adherence statement, rated the intervention significantly higher as well (p ≤ 0.001).Table 4 Linear regression analysis results for adherence and satisfaction with the intervention β p-value 95%CI Lower Upper Agree (n = 14) 2.42 0.000 1.41 3.42 Somewhat agree (n = 47) 1.40 0.000 0.68 2.11 Neutral or disagree (n = 25) Reference 95%CI 95 % Confidence Interval Subgroup differences Compliance The objective compliance of the subgroups was comparable with compliance in the total group of participants. As can be seen in Table 5, differences between the groups were small. Because layover advices were not available for short-haul pilots, their registered number of consultations of this type of advice was close to zero.Table 5 Compliance scores within the subgroups Compliance Subgroup Objective (n = 247) Subjective (n = 86) Mean number of advices (SD) Median Type of advice (% of advices) % using the advices Type of advice (% of users) Subject of advice (% of users) Home During layover Return Never A few times Sometime/always Home During layover Return Light Sleep Nutr. Phys. act. <35 5.65 (10.76) 2.00 51 % 25 % 25 % 53 % 26 % 21 % 28 % 78 % 56 % 17 % 78 % 83 % 11 %* ≥35- < 45 6.34 (13.21) 3.00 52 % 18 % 29 % 39 % 30 % 32 % 43 % 63 % 46 % 34 % 57 % 80 % 43 %* ≥45 7.98 (16.36) 2.00 46 % 28 % 27 % 38 % 23 % 40 % 33 % 73 % 48 % 27 % 58 % 58 % 9 %* Short-haul 7.42 (14.34) 2.00 64 % 1 % 35 % 44 % 28 % 28 % 50 % 68 % 36 % 32 % 59 % 91 %* 36 % Long-haul 6.53 (13.84) 3.00 42 % 34 % 24 % 41 % 26 % 33 % 31 % 70 % 53 % 27 % 63 % 66 %* 19 % * Significant difference (p < 0.05) Results on subjective compliance show that 53 % of the pilots younger than 35 years indicated to have never consulted the advices. Forty percent of the oldest group of pilots indicated that they used the advices sometimes or always. All age groups indicated they consulted the advices about the time spent during layover the most (63 to 78 %), followed by advices concerned with the return flight (46 to 56 %), and the advices before departure from home (28 to 43 %). With regard to the content of the advices, the youngest group of pilots was more concerned with advice regarding sleep (78 % vs. 57/58 %, NS), while the oldest group was less concerned with advice regarding nutrition compared to their colleagues (58 % vs. 80/83 %, p = 0.06). Further, significantly more 35 to 45 year old pilots were interested in advice regarding physical activity than their younger and older colleagues (43 % vs. 9/11 %, p = 0.002). Although layover advice was not available for short-haul pilots, subjective compliance results show that both groups indicated to have consulted this type of advice the most. Next, 50 % of the short-haul pilots consulted the advices with regard to departure from home, while 53 % of the long-haul pilots consulted the advices regarding the return flight. Advice regarding nutrition was consulted significantly more by short-haul compared to long-haul pilots (91 % vs. 66 %, p = 0.02). Satisfaction No significant differences in satisfaction were present between the subgroups (Table 6). The youngest group of pilots showed the lowest percentage that agreed that the app was accessible and usable (61 %). However, this group showed the highest percentage that reported the advice to be easy to apply (56 %) and that indicated to have learned from the intervention (44 %). Of the oldest group of pilots, 40 % indicated that the advices were easy to apply, and 33 % indicated that they learned a lot. Still, 85 % of the oldest group of pilots would recommend the intervention to their colleagues.Table 6 Satisfaction scores within the subgroups Subgroup Rating Satisfaction (agree) Mean grade (SD) Pleasant medium Accesible and usable Pleasant language Useful advices Easy to apply Learned a lot Prevents fatigue Recommend to colleagues <35 6.2 (1.5) 83 % 61 % 83 % 61 % 56 % 44 % 61 % 72 % ≥35- < 45 6.5 (1.5) 67 % 73 % 73 % 45 % 45 % 36 % 63 % 74 % ≥45 6.5 (1.9) 83 % 73 % 77 % 63 % 40 % 33 % 62 % 85 % Short-haul 6.1 (1.9) 75 % 75 % 85 % 40 % 55 % 30 % 70 % 70 % Long-haul 6.5 (1.5) 77 % 69 % 74 % 61 % 43 % 39 % 70 % 87 % Comparing the two haul types, it can be seen that 55 % of the short-haul pilots reported that the advices were easy to apply, compared to 43 % of the long-haul pilots. However, 40 % of the short-haul pilots considered the advices also useful, compared to 61 % of the long-haul pilots (p = 0.06). Further, 70 % of the short-haul pilots would recommend the advices to their colleagues, while 87 % of the long-haul pilots would do that (p = 0.06). Discussion Main findings The primary objective of this process evaluation was to assess the reach, dose delivered, compliance, fidelity, context, and satisfaction of the MORE Energy intervention. The reach among the source population was 22 %. This percentage is quite high compared to the 1.5 to 8 % reach published in other mobile health studies [16, 23, 24], and compared to more conventional studies promoting health behaviour at worksites [25]. The participating pilots were significantly younger compared to the non-participants, possibly because younger pilots are more familiar with mobile apps. Female pilots were also overrepresented, possibly caused by the fact that women tend to exhibit more active information-seeking behaviour and are more likely to participate in scientific studies than men in general [26, 27]. It was shown that the dose delivered and initial compliance was high. Of the participants, only 19 % never used any advice on either the app or the website. It is even possible that some of these participants never received the email containing the instructions and login details, because of for instance a strict junk-email filter. However, during the intervention period, only one participant reported not to have received the instruction email. The compliance during the whole intervention period was rather low with 17 % of the participants consulting the advices on the app for more than four weeks during six months. Technical problems with some components of the app, as could be seen in the calculation of the fidelity score, might have contributed to this. The fidelity score of 67 % is difficult to interpret since we are the first mobile health study to calculate such a score. Moreover, a mobile health intervention review showed that only 13 % of the published mobile health studies on physical activity promotion reported fidelity information whatsoever [18]. Our results showed a distinction between the objectively and subjectively measured compliance regarding the type of advices. Despite the registered data displayed that the advice concerned with departure from home was consulted most often, the participants themselves indicated that they used the advice concerned with layovers the most. It might be possible that the participants thought that they were asked when they had used the advice most often. This would also explain the finding that short-haul pilots indicated to have used the layover advice the most although that type of advice was not available for them; they might have predominately considered the advice regarding their next flight at the end of a duty day (i.e. during their layover). Despite the moderate compliance scores, a majority of the users (65 %) was convinced that the intervention was able to fulfill its purpose, preventing fatigue and improving health of pilots. Also, 78 % of the participants would recommend the intervention to their colleagues. In this perspective, the 6.4 (range 1 to 10) appreciation score for the MORE Energy intervention as a whole, is somewhat low. On the other hand, a majority of the pilots (54 %) did not agree with the statement that the advices were easy to apply in daily life. This lack of applicability of the flight schedule specific advices was also shown in the barriers and facilitators section. Participants indicated that applying the advices would conflict with their social life at home (e.g. young children) or during duty (e.g. habits during layovers). Furthermore, the correlation between registered compliance and adherence proved to be very low (r = 0.041). This might indicate that although participants were interested to see what advice the app would provide regarding their upcoming flight schedule, the content did not pursue them to change their behavior. However, once applied, the advices might have been useful: participants that did indicate to have applied the advices in daily life rated the intervention significantly higher compared to the participants who did not apply them. Another objective of this process evaluation was to assess whether there was a difference in compliance and satisfaction outcomes between participants of the two haul types and the three age groups present in the population. The results showed that the differences in both objective and subjective compliance between the age groups were small. The oldest participants tended to be the most critical regarding the applicability of the advices. Possibly, these pilots were unwilling to give up the patterns and habits which they developed throughout their career. This was already mentioned during the focus group interviews before the development of the intervention: it would be hard to alter the (social) patterns of the more experienced colleagues. The apparent contradicting high percentage of the oldest group of pilots (85 %) that would recommend MORE Energy to their colleagues may be in accordance: these pilots do not need the advice for themselves but they think it might be useful for their less experienced colleagues. Regarding the participants of the two haul types, no significant differences in objective compliance existed either. The relatively large number of short-haul pilots that participated in the study did find the advices easy to apply, but found them not very useful. Consequently, short-haul pilots tended to be less satisfied with the intervention compared to the long-haul pilots. Although the operations of the two haul types differ substantially, short-haul schedules can trigger fatigue as much as intercontinental schedules [28]. However, most of the scientific knowledge and practical advice available concerns the disruption of the circadian rhythm and is mostly applicable for long-haul pilots. Consequently, the content of our specific advice regarding short-haul schedules was less extensive. The short-haul pilots participating in the test phase of the intervention noticed this already. Although we reconsidered and extended the short-haul advices afterwards, the outcomes of our process evaluation indicate that this elaboration was probably not sufficient. Strength and limitations This process evaluation is one of the first evaluating an mobile health intervention promoting health behaviour. One of the major strengths is that we used a combination of objective and self-reported data to evaluate the implementation of the MORE Energy intervention. The compliance with both the app and the website was registered through user authentication, which is more reliable than self-reported information [29]. Nevertheless, most probably due to malfunctioning of the CMS, data of four participants had to be excluded from the analyses. We did use self-reported data as well, to gather more detailed information on subject specific compliance, adherence and appreciation. We achieved a reach of 22 %. Although this is quite high compared to participation rates published in other mobile health studies, the generalizability of the results might be hampered due to selection effects. It is possible that the non-participating pilots did not possess an iOS or Android smartphone or tablet, or were not sufficiently familiar with mobile apps. The comparison of the participants with the non-participants also showed that participants were significantly younger, and that a relatively large number of female pilots participated. Another limitation of this study is that only 59 % of the participants filled out the process evaluation questionnaire. Therefore, selection bias may have occurred, which challenges the reliability and validity of the outcomes. Another limitation of the study is that we could not measure all activities of the users of the mobile app. The number and types of requested advices were registered, but consultation of the background information and time spent on the app was not. The MORE Energy intervention aimed to improve relevant behaviour among airline pilots. Although we asked participants whether they had applied the advices in daily life, their actual change in behaviour could not be objectively measured. Therefore, to maximize the interpretation of our results, it would have been useful if the app had used more built-in features of smartphones to measure behaviour (e.g. timing of sleep using the motion sensor of the mobile device) [13]. Implications for research and practice The reach of this study shows that airline pilots are willing to participate in an intervention using an app to promote health behaviour. Although guidelines have been proposed to improve the way the outcomes of mobile health interventions are reported [21], qualitative process evaluations publications are necessary as well in order to provide more information about the implementation and working mechanisms of these kind of interventions [18]. High compliance is important for the success of any intervention intending to modify behaviour, but especially so in web-based interventions since there is no direct contact with the participants [14]. Because we found that compliance dropped since the start of the intervention, similar studies should put more effort into keeping participants involved throughout the intervention period. Although technical problems were not mentioned as main reasons for non-compliance, time consuming update installations and the resulting loss in functionality, might have been of influence. One of the updates during the intervention period caused the malfunctioning of the reminder alerts. This might have led to a decrease in compliance since well-timed and adequate prompts can be effectively used in mobile health intervention studies. Other possibilities that could keep users engaged involve altering and updating the content of the intervention material, providing personal feedback, and introducing goal setting [12]. Our process evaluation gives insight into the different aspects involved in the implementation of the MORE Energy intervention, and will help to improve the interpretation of the results of the trial. The outcomes of the different items showed which parts of the intervention should be improved before offering the intervention to all flight crew throughout the airline company. First, the app should pursue users more to change relevant behavior, despite the consequences for their social responsibilities. Further, the content of the advice should be better applicable for both experienced and inexperienced employees, involved in both short and long-haul schedules. The airline company involved can assist the subsequent implementation of the improved app by integrating it with flight crew scheduling, by giving sustained attention to the topic within the present flight crew members, and by introducing the tool to newly hired employees [7]. After the improvement of the MORE Energy intervention, it might be transformed into a ‘white label’ tool in order to make it possible to adapt and implement it as a fatigue management tool for flight crew members or shift workers in other companies as well. Conclusions The process evaluation of MORE Energy showed that this mobile health intervention was well received, resulting in an adequate reach and a high dose delivered. Although more than 80 % of the participants did use the intervention, most of them were not compliant throughout the intervention period. The intervention could not be delivered as intended and perceived satisfaction was moderate. Further, the combination of compliance and satisfaction scores indicates that the content and applicability of the advices should be improved to appeal all subgroups first, before making the app available for the other flight crew members within the airline company. After this improvement, MORE Energy might be adapted and implemented as a fatigue management tool for other employees and companies as well. Additional files Additional file 1: Screenshots of the mobile application used during the MORE Energy intervention. (PDF 1072 kb) Additional file 2: Detailed explanation of all bugs that occured during the intervention period. (DOCX 16 kb) Abbreviations 95%CI95 % Confidence Interval AppApplication CMSControl Management System iOSOperating System for iPhone or iPad SDStandard Deviation Acknowledgements None. Funding This study is part of a through VU University Medical Center funded PhD program. Availability of data and materials The airline involved does not prefer to make the data publicly available. Authors’ contributions AvD analysed the data and wrote the manuscript. CB, HH, TS, and AvdB helped to design the study and gave essential feedback on all versions of the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interest. Consent for publication Not applicable Ethics approval and consent to participate The Medical Ethics Committee of the VU University Medical Center (Amsterdam, the Netherlands) assessed the study design and procedures, and decided that, according to Dutch law, this study proved to be exempt from a medical ethical review. ==== Refs References 1. Signal TL Ratieta D Gander PH Flight crew fatigue management in a more flexible regulatory environment: an overview of the New Zealand aviation industry Chronobiol Int 2008 25 2–3 373 388 10.1080/07420520802118202 18484369 2. Gander P Hartley L Powell D Cabon P Hitchcock E Mills A Popkin S Fatigue risk management: Organizational factors at the regulatory and industry/company level Accid Anal Prev 2011 43 2 573 590 10.1016/j.aap.2009.11.007 21130218 3. Cabon P Deharvengt S Grau JY Maille N Berechet I Mollard R Research and guidelines for implementing Fatigue Risk Management Systems for the French regional airlines Accid Anal Prev 2012 45 41 44 10.1016/j.aap.2011.09.024 22239930 4. Hauck EL Avers KB Banks JO Blackwell LV Evaluation of a Fatigue Countermeasures Training Program for Flight Attendants 2011 Oklahoma City Civil Aerospace Medical Institute, Federal Aviation Administration 5. Cabon P Mollard R Coblentz A Fouillot JP Speyer JJ Recommendations for alertness and sleep management of aircrews operating longhaul flights Med Aeronaut Spat 1995 34 134 119 128 6. Rosekind MR Gregory KB Mallis MM Alertness management in aviation operations: Enhancing performance and sleep Aviat Space Environ Med 2006 77 12 1256 1265 10.3357/ASEM.1879.2006 17183922 7. Avers KE Hauck EL Blackwell LV Nesthus TE Federal Aviation Administration: Flight Attendant Fatigue 2009 Oklahoma City Civil Aerospace Medical Institute, Federal Aviation Administration 8. Flower DJC Alertness management in long-haul flying Transp Res Part F Traffic Psychol Behav 2001 4 1 39 48 10.1016/S1369-8478(01)00012-2 9. Kroeze W Werkman A Brug J A systematic review of randomized trials on the effectiveness of computer-tailored education on physical activity and dietary behaviors Ann Behav Med 2006 31 3 205 223 10.1207/s15324796abm3103_2 16700634 10. De Vries H Brug J Computer-tailored interventions motivating people to adopt health promoting behaviours: introduction to a new approach Patient Educ Couns 1999 36 2 99 105 10.1016/S0738-3991(98)00127-X 10223015 11. Van Drongelen A Van Der Beek AJ Hlobil H Smid T Boot CRL Development and evaluation of an intervention aiming to reduce fatigue in airline pilots: design of a randomised controlled trial BMC Public Health 2013 13 1 776 10.1186/1471-2458-13-776 23971514 12. Free C Phillips G Galli L Watson L Felix L Edwards P Patel V Haines A The efffectiveness of mobile-health technology-based health behaviour change or disease management interventions for health care consumers: a systematic review PLoS Med 2013 10 1 e1001362 10.1371/journal.pmed.1001362 23349621 13. Fiordelli M Diviani N Schulz PJ Mapping mHealth research: a decade of evolution J Med Internet Res 2013 15 5 e95 10.2196/jmir.2430 23697600 14. Van Den Berg MH Schoones JW Vlieland TPMV Internet-based physical activity interventions: a systematic review of the literature J Med Internet Res 2007 9 3 e26 10.2196/jmir.9.3.e26 17942388 15. Buchholz SW Wilbur J Ingram D Fogg L Physical activity text messaging interventions in adults: a systematic review Worldviews Evid Based Nurs 2013 10 3 163 173 10.1111/wvn.12002 23746267 16. Kirwan M Duncan MJ Vandelanotte C Mummery WK Using smartphone technology to monitor physical activity in the 10,000 Steps program: a matched case-control trial J Med Internet Res 2012 14 2 e55 10.2196/jmir.1950 22522112 17. Fanning J Mullen SP McAuley E Increasing physical activity with mobile devices: a meta-analysis J Med Internet Res 2012 14 6 e161 10.2196/jmir.2171 23171838 18. Blackman KC Zoellner J Berrey LM Alexander R Fanning J Hill JL Estabrooks PA Assessing the internal and external validity of mobile health physical activity promotion interventions: a systematic literature review using the RE-AIM framework J Med Internet Res 2013 15 10 e224 10.2196/jmir.2745 24095951 19. Saunders RP Evans MH Joshi P Developing a process-evaluation plan for assessing health promotion program implementation: a how-to guide Health Promot Pract 2005 6 2 134 147 10.1177/1524839904273387 15855283 20. Steckler AE Linnan LE Process evaluation for public health interventions and research 2002 San Francisco Jossey-Bass 21. Eysenbach G CONSORT-EHEALTH: Improving and standardizing evaluation reports of web-based and mobile health interventions J Med Internet Res 2011 13 4 e126 10.2196/jmir.1923 22209829 22. van Drongelen A Boot CRL Hlobil H Twisk JW Smid T van der Beek AJ Evaluation of an mHealth intervention aiming to improve health-related behavior and sleep and reduce fatigue among airline pilots Scand J Work Environ Health 2014 40 6 557 568 10.5271/sjweh.3447 25121620 23. Quinn CC Shardell MD Terrin ML Barr EA Ballew SH Gruber-Baldini AL Cluster-randomized trial of a mobile phone personalized behavioral intervention for blood glucose control Diabetes Care 2011 34 9 1934 1942 10.2337/dc11-0366 21788632 24. Ryan D, Price D, Musgrave SD, Malhotra S, Lee AJ, Ayansina D, Sheikh A, Tarassenko L, Pagliari C, Pinnock H. Clinical and cost effectiveness of mobile phone supported self-monitoring of asthma: multicentre randomised controlled trial. Br Med J. 2012;344:e1756. 25. Kwak L Kremers SP van Baak MA Brug J Participation rates in worksite-based intervention studies: health promotion context as a crucial quality criterion Health Promot Int 2006 21 1 66 69 10.1093/heapro/dai033 16339773 26. Eysenbach G, Wyatt J. Using the Internet for surveys and health research. J Med Internet Res. 2002;4(2):e13. 27. Galea S Tracy M Participation rates in epidemiologic studies Ann Epidemiol 2007 17 9 643 653 10.1016/j.annepidem.2007.03.013 17553702 28. Powell D Spencer MB Holland D Broadbent E Petrie KJ Pilot fatigue in short-haul operations: Effects of number of sectors, duty length, and time of day Aviat Space Environ Med 2007 78 7 698 701 17679568 29. Resnicow K Strecher V Couper M Chua H Little R Nair V Polk TA Atienza AA Methodologic and design issues in patient-centered e-health research Am J Prev Med 2010 38 1 98 102 10.1016/j.amepre.2009.09.034 20117564

PMC005xxxxxx/PMC5002200.txt

==== Front BMC GenomicsBMC GenomicsBMC Genomics1471-2164BioMed Central London 302210.1186/s12864-016-3022-6Research ArticleTranscriptome responses to temperature, water availability and photoperiod are conserved among mature trees of two divergent Douglas-fir provenances from a coastal and an interior habitat Hess Moritz ssehztirom@gmail.com 125Wildhagen Henning henning.wildhagen@forst.uni-goettingen.de 14Junker Laura Verena laura.junker@utoronto.ca 13Ensminger Ingo +1 905 569 4599ingo.ensminger@utoronto.ca 131 Forest Research Institute of Baden-Württemberg (FVA), Wonnhaldestrasse 4, D-79100 Freiburg i. Brsg., Germany 2 Institute for Biology III, Faculty of Biology, Albert Ludwigs University Freiburg, Schänzlestrasse 1, D-79104 Freiburg i. Brsg., Germany 3 Department of Biology, Graduate Programs in Cell & Systems Biology and Ecology & Evolutionary Biology, University of Toronto, 3359 Mississauga Road, Mississauga, ON L5L 1C6 Canada 4 Present Address: Department of Forest Botany and Tree Physiology, Büsgen-Institute, Georg-August-University Göttingen, Büsgenweg 2, D-37077 Göttingen, Germany 5 Present Address: Institute of Medical Biometry, Epidemiology and Informatics (IMBEI), University Medical Center Mainz, Obere Zahlbacher Strasse 69, 55131 Mainz, Germany 26 8 2016 26 8 2016 2016 17 1 68224 9 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 Local adaptation and phenotypic plasticity are important components of plant responses to variations in environmental conditions. While local adaptation has been widely studied in trees, little is known about plasticity of gene expression in adult trees in response to ever changing environmental conditions in natural habitats. Here we investigate plasticity of gene expression in needle tissue between two Douglas-fir provenances represented by 25 adult trees using deep RNA sequencing (RNA-Seq). Results Using linear mixed models we investigated the effect of temperature, soil water availability and photoperiod on the abundance of 59189 detected transcripts. Expression of more than 80 % of all identified transcripts revealed a response to variations in environmental conditions in the field. GO term overrepresentation analysis revealed gene expression responses to temperature, soil water availability and photoperiod that are highly conserved among many plant taxa. However, expression differences between the two Douglas-fir provenances were rather small compared to the expression differences observed between individual trees. Although the effect of environment on global transcript expression was high, the observed genotype by environment (GxE) interaction of gene expression was surprisingly low, since only 21 of all detected transcripts showed a GxE interaction. Conclusions The majority of the transcriptome responses in plant leaf tissue is driven by variations in environmental conditions. The small variation between individuals and populations suggests strong conservation of this response within Douglas-fir. Therefore we conclude that plastic transcriptome responses to variations in environmental conditions are only weakly affected by local adaptation in Douglas-fir. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-3022-6) contains supplementary material, which is available to authorized users. Keywords Phenotypic plasticityTranscriptomeLocal adaptationPseudotsuga menziesiiGene expressionDroughtPhotoperiodHeatNatural environmentRNA-Seqhttp://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftEN 829/4-1Ensminger Ingo NSERCissue-copyright-statement© The Author(s) 2016 ==== Body Background Plants continuously experience variations in environmental conditions on short (e.g. minutes to days) and long time scales (e.g. weeks or growing season). Short term responses to dynamic environments require strict physiological regulation and are known as phenotypic plasticity. On evolutionary timescales, responses to a specific environment may result in adaptation to local environment through genetic differentiation of populations, which is known as local adaptation [1]. As a consequence of local adaptation, plant populations frequently show the best growth performance or fitness in [2] or next to their habitat of origin [3] and outperform populations from distant habitats [4]. Potential genomic targets of adaptation to climate have been identified by linking single nucleotide polymorphisms (SNP) to local climate (e.g. in Arabidopsis thaliana [5], Pinus taeda [6], Picea mariana [7]) or by identification of associations among SNPs and traits that are known to co-vary with climatic clines, e.g. bud set and cold resistance (Picea sitchensis [8]), carbon isotope discrimination (Pinus taeda [9]) or cold hardiness (Pseudotsuga menziesii [10]). With respect to the anticipated rapidly changing climate [11], forest trees, which have long generation times, need to adjust their metabolism in response to changing abiotic factors [12]. Global transcriptome analysis has been extensively used in studies with highly controlled conditions to characterize plasticity and diversity of gene expression metabolism in response to abiotic factors among different populations (e.g. in Arabidopsis thaliana [13–15], Populus [16], Helianthus annuus [17], Pinus spp. [18] or Picea spp. [18, 19]. Plants do evolve in complex natural environments, and controlled conditions in a laboratory or greenhouse facility rarely represent the ever-changing complex conditions experienced by plants in natural environments [20]. Few studies investigated genome-wide gene expression responses to abiotic stimuli in natural environments [21, 19, 22–25] and studies in perennial, woody plants are exceptionally rare but see e.g. [19] on Picea sitchensis, [22] on Vitis vinifera and [25] on Populus euphratica. Douglas-fir is a commercially important Pinaceae tree species originating in western North America. Two subspecies, Pseudotsuga menziesii var. menziesii (Coastal Douglas-fir) and Pseudotsuga menziesii var. glauca (Interior Douglas-fir) diverged about 1 million years ago [26, 27] and cover a wide natural range with contrasting environmental conditions along the pacific coast and the Rocky Mountains, respectively. Local adaptation of Douglas-fir populations growing in defined geographical areas (provenances) has been shown in several studies [10, 28]. These characteristics make Douglas-fir an ideal model organism to study the effect of local adaptation on transcriptome responses to environmental conditions. In this study we aimed to 1) identify transcriptome dynamics in field-grown adult Douglas-fir trees in response to temperature, water availability and photoperiod, 2) use overrepresentation analysis to reveal common functional themes in gene sets that respond to environmental factors, and 3) evaluate differences in transcriptome dynamics between provenances. For this purpose we compared 50-year-old trees of two divergent Douglas-fir provenances originating from contrasting environments in British Columbia, Canada at two contrasting common garden field-sites in Southern Germany. To our knowledge, this is the first genome wide assessment of the effect of abiotic environmental factors on the transcriptome responses of heterogeneous, locally adapted populations of mature trees, grown in contrasting natural environments. Results Sequencing, alignment, quantification and functional annotation We quantified transcript expression in 25 50-year-old Douglas-fir trees, growing at two common gardens (Wiesloch and Schluchsee) in southwestern Germany, during the growing season of 2010. 12 of these trees were from provenance Cameron Lake (LA), 13 were from provenance Salmon Arm (AR). A total of 75 RNA extracts from needle samples collected at noon in May, June, July and September at both field sites were sequenced on the Illumina HiSeq2000 (Fig. 1a). Reads were aligned to the set of 176753 non-redundant Douglas-fir putative unique transcripts (PUT) (Fig. 1b). The alignment yielded on average 33 million aligned reads (Mreads) per sequencing library. After excluding low abundant PUTs we detected 59189 PUTs (~34 %) of the 176753 PUTs present in the non-redundant set. Around 40 k PUTs were functionally annotated by alignment to the NCBI plant RefSeq data base (Table 1). GO annotations were identified for 34 k PUTs using BLAST2GO. 6330 PLAZA gene families were identified in the set of all detected PUTs.Fig. 1 Analysis Pipeline. The sequence libraries were aligned (a) to a non-redundant set of two Douglas-fir PUT sets (b). After log-transforming the count data (c), linear random effect models were used to (d) detect sources of variation within the data. Differential expression between genotype and environment was investigated by linear mixed models (LMM) (e). Nested models were compared by an F-test with Kenward-Roger approximation [102]. Additionally the difference in AIC (dAIC) and R2 (dR2 ) was calculated. PUTs are only considered as responding to an environmental factor if the corresponding regressor increases the R2 by more than 0.2. Overrepresentation of GO categories within the differentially expressed PUTs was investigated (f) Table 1 Annotation statistics of detected PUTs Cluster type Number of PUTs Unique Annotations PUTs detected with 1 RPM in at least 4 Libraries 59189 hit in RefSeq 38248 21218 hit in Picea glauca ORFs 45867 14544 hit in Pinus taeda ORFs 42602 13842 hit in Vitis vinifera 38841 11263 hit in Oryza sativa 37408 11220 hit in Arabidopsis thaliana 34356 10762 B2GO annotation 34375 9062 PLAZA gene family 41175 6330 RefSeq: NCBI plant RefSeq peptide data base, Picea glauca ORFs: Picea glauca full length ESTs, Pinus taeda ORFs: de novo assembled ESTs which have been used to annotate the Pinus taeda genome, Vitis vinifera: Vitis vinifera peptide data base (PLAZA), Oryza sativa: Oryza sativa peptide data base (PLAZA), Arabidopsis thaliana: TAIR10 peptide data base, B2GO: GO annotations inferred by the BLAST2GO pipeline Variance components contributing to PUT expression variation In a first step we used a linear variance components model to estimate the contribution of the environment and the genotype to the variation in PUT expression. Expression variation driven by environment was investigated between the eight sampling time points across the two field sites (DATE) and between two field-sites (SITE). Expression variation driven by genotype was investigated between provenances (PROVENANCE) and between individual trees (TREE). The log-transformed number of reads that aligned to each PUT was used as the dependent variable in 59189 linear random effects models (Fig. 1d). For 25 % of all PUTs, at least 40 % of the variation in PUT expression was explained by TREE followed by DATE (30 %) and PROVENANCE (≥12 %) (Fig. 2). The least contribution to the variation in PUT expression was attributed to the field site (SITE; ≥7 %).Fig. 2 Variance components of biological and environmental factors. Proportion of total expression variation attributable to provenances (PROVENANCE), field sites (SITE), individual trees (TREE) sampling time point (DATE) for all detected 59189 PUTs. A linear random effects model that comprises all above mentioned random effects was fit to all PUTs using the R package lme4. Variance components for each random effect were extracted and divided by the sum of all variance components and the residual variance Detection of differentially expressed PUTs between provenances and in response to environmental variation The effect of environment (E), genotype (G) represented by PROVENANCE, or their interaction (GxE) on PUT expression was further investigated using linear mixed models (see Fig. 1e). This analysis identified 1764 PUTs that were differentially expressed between the two provenances (PROVENANCE), 39614 differentially expressed PUTs between sampling time-points (DATE), and 21 PUTs that were affected by the interaction of provenance and environment (FDR <0.01, F-test with Kenward-Roger approximation) (Fig. 3).Fig. 3 Differentially expressed PUTs between provenances and sampling time-points. Number of PUTs that are differentially expressed (FDR <0.01, F-test with Kenward-Roger approximation) between sampling time-points (DATE) or provenances and number of PUTs with significant interaction of provenances and sampling time-points Assuming that DATE integrates the effects of temperature, day length and water availability on PUT expression, we further analysed the direct effects of these environmental factors on PUT expression using a second set of linear mixed models (Fig. 1e). These models comprised proxies for temperature, day length and water availability. Since absolute temperature data were correlated with day length, temperature was detrended by subtracting the four weak running average from air temperature on the day of sampling resulting in the new variable TEMPERATURE. Day length was represented as deviation of photo period from the length of the day at solstice (DAYLENGTH) and total available soil water (TAW) was modeled using the forest water model WBS3 (Fig. 4a-c). We inspected correlation of all variables (Additional file 1: Figure S1) and observed no strong correlation between TEMPERATURE, DAYLENGTH and TAW. However SITE was strongly correlated with TAW (Additional file 1: Figure S1) and since the variance components analysis revealed that not much expression variance can be attributed to differences between the field sites (Fig. 2) SITE was omitted from this second set of models.Fig. 4 Effects of environmental factors on PUT expression. X-Y plots show soil water availability (a) temperature (b) and day length (c) during the course of the year. The corresponding bar plots show the values of the respective regressors TAW, TEMPERATURE and DAYLENGTH that were used in the linear models. Red and blue lines indicate the sampling time point in the common gardens in Wiesloch and Schluchsee, respectively. Curves were smoothed using loess transformation. Temperature shown in the bar plot (b) was centered to a 4 week running average while day length (c) was centered to the maximum day length of the year. Day length in May was transformed to positive values to mimic the directed developmental processes occurring during a growing period. d Standardized expression of PUTs that are significantly responding to TAW, TEMPERATURE or DAYLENGTH in Salmon Arm (magenta) and Cameron Lake (cyan) relative to the environmental parameters TAW (Fig. 4a) was generally lower at the field site Wiesloch during all sampling time-points (DATE) and low TAW was often accompanied by higher temperature (Fig. 4b). The highest temperatures at both sites were recorded in July, when temperature exceeded the four week average by 7 °C in Wiesloch. DAYLENGTH in September did contrast with May, June and July, with day length in September being 3–4 h shorter than during May, June and July (Fig. 4c). Further analysis of the PUTs that responded to TAW, TEMPERATURE or DAYLENGTH revealed the expression profile in response to each of the environmental factors (Fig. 4d). The relationship of expression levels of PUTs to any of the three environmental factors indicated a highly coordinated response over the entire range of TAW, TEMPERATURE or DAYLENGTH (Fig. 4d). The expression profile of PUTs with a positive response clearly contrasted with the expression profile of PUTs with a negative response to each of the environmental factors. Increases in TAW corresponded with an increased expression of 2119 PUTs (FDR <0.01, F-test with Kenward-Roger approximation, dR2 >0.2) and a decreased expression of 2047 PUTs. When TEMPERATURE increased, we observed increased expression of 1466 PUTs versus 1771 PUTs showing the opposite response. An increase in DAYLENGTH corresponded with increased expression of 2234 PUTs compared to 2581 PUTs with decreased expression. The specific response to environmental factors (Fig. 4d) was also reflected by only 446 PUTs out of the more than 12 k differentially expressed PUTs that were effected by more than one environmental factor (Fig. 5a).Fig. 5 Variation in expression response to environment among provenances. a Number of PUTs that are differentially expressed between provenances or responding to TEMPERATURE, DAYLENGTH or TAW. b Expression variation among provenances of PUTs that respond to environmental factors or are differentially expressed between provenances compared to all detected PUTs. c Expression variation among trees of PUTs that respond to environmental factors or are differentially expressed between provenances compared to all detected PUTs. The red line in b and c indicates the median of expression variation in all detected PUTs. Numbers at the whiskers of each box plot indicate the number of PUTs within a category In addition to the low number of PUTs with provenance by environment interaction, only 224 PUTs out of the 1764 PUTs that were differentially expressed between provenances were also differentially expressed in response to at least one of the three environmental factors included in our analysis (Fig. 5a). This set of 224 PUTs included 101 PUTs differentially expressed between provenances and in response to DAYLENGTH, 67 PUTs differentially expressed between provenances and in response to TAW, and 62 PUTs differentially expressed between provenance and in response to TEMPERATURE. The low number of PUTs that responded to climate and were differentially expressed between provenances coincided with reduced expression variation of these PUTs among provenances (Fig. 5b) and trees (Fig. 5c). This effect is particularly pronounced for PUTs that responded to TEMPERATURE. Gene Ontology categories overrepresented in differentially expressed PUTs An overrepresentation analysis was performed to identify Gene Ontology (GO) categories in the differentially expressed PUTs that showed a response to environmental conditions or showed differences between the provenances. Using Fisher’s exact test (P-value < 0.01, minimum 10 differentially expressed PUTs within a GO category) we identified 63 GO categories that were overrepresented in higher expressed PUTs when TAW was low, in contrast to 83 GO categories that were overrepresented when TAW was high. For TEMPERATURE we identified 54 GO categories that were overrepresented in higher expressed PUTs when TEMPERATURE was high and 143 GO categories that were overrepresented when TEMPERATURE was low. For DAYLENGTH we identified 101 GO categories that were overrepresented in higher expressed PUTs when DAYLENGTH was longer, compared to 80 GO categories that were overrepresented when DAYLENGTH was shorter. A comparison of GO ontologies between the two provenances revealed 117 GO categories that were overrepresented in PUTs that were stronger expressed in provenance Cameron Lake compared to one single GO category that was overrepresented in PUTs that were stronger expressed in provenance Salmon Arm (Additional file 2: Table S1). For each regressor and direction of expression we chose the 20 most overrepresented GO categories and investigated the biological functions of PUTs within each GO category (Fig. 6). We investigated the best hit in the Arabidopsis thaliana peptide data base (TAIR10) and in the NCBI RefSeq data base of PUTs that we observed within the overrepresented GO categories. The best hits in the TAIR10 data base as well as in the NCBI RefSeq data base are listed in Additional file 3: Table S2. Hits were ranked according to their absolute log10 transformed P-value from the test of differential expression. All functional descriptions in the next sections that are not followed by a citation were retrieved from the TAIR data base.Fig. 6 GO term overrepresentation analysis of PUTs that respond to environment or are differentially expressed between provenances. The top 20 GO categories that were significantly overrepresented within PUTs that respond to TEMPERATURE, TAW, DAYLENGTH or PROVENANCE are shown. Redundant GO categories have been removed using REVIGO [103] with default parameters. Numbers in columns indicate the amount of overrepresentation which is the number of significant PUTs within a GO category divided by the number of expected PUTs within a GO category and are only shown for significant overrepresentation of a GO category (Fisher’s Exact Test, P < 0.01). The p-value of Fisher’s exact test is indicated by color intensity. Red color indicates high overrepresentation of a GO category within PUTs that are positively correlated with TEMPERATURE and DAYLENGTH, negatively correlated with TAW or stronger expressed in Cameron Lake (PROVENANCE). Blue color indicates high overrepresentation of GO categories within PUTs that are negatively correlated with TEMPERATURE and DAYLENGTH, positively correlated with TAW or stronger expressed in Salmon Arm (PROVENANCE). The dendrogram indicates the similarity of GO categories in terms of shared PUTs among GO categories, expressed by Cohen’s Kappa. Numbers behind GO categories indicate the hierarchy of the GO graph, while more basal terms have smaller numbers Biological functions of differentially expressed PUTs that respond to TAW GO categories related to starch metabolism, e.g. “starch biosynthetic process”, “starch metabolic process”, and nucleic acid metabolism, e.g. “RNA metabolic process”, “nucleic acid metabolic process”, “gene expression”, “regulation of cellular macromolecule compound metabolic process” and “nucleobase-containing compound metabolic process” (Fig. 6; rows 44, 45, 6, 7, 8, 9, 12; Additional file 2: Table S1 - TAW_Down) were highly overrepresented in PUTs that were higher expressed when TAW was low (P-value Fisher’s exact test < 1e-04). PUTs within the GO categories related to nucleic acid metabolism included homologs of the protein kinase WEE1 (AT1G02970; Additional file 3: Table S2 - TAW_Down; rank 1) which negatively regulates the entry into mitosis [29], the CDC2 related kinase AFC1 (AT3G53570; rank 6), the WUSCHEL-related homeobox gene family member WOX13 (AT4G35550; rank 13), the apurinic endonuclease-redox protein ARP (AT2G41460; rank 15) which is involved in DNA repair, the splicing factor RSZ22 (AT2G24590; rank 24), the NAC domain transcription factor ANAC002 (AT1G01720; rank 28) whose mRNA levels increase in response to abscisic acid in Arabidopsis thaliana or the zinc-finger proteins CCCH20 and CCCH49 (AT2G19810; ranks 33, 187) that are involved in RNA binding in the response to osmotic stress [30]. Within the GO categories related to starch metabolism we identified homologs of the phosphoglucomutase PGMP (AT5G51820; rank 2), which controls photosynthetic carbon flow, the beta amylase BAM3 (AT4G17090; ranks 20, 34), the starch branching enzyme SBE2.2 (AT5G03650; rank 21), or the fructose 1–6 bisphosphatase FBP (AT1G43670; rank 41). GO categories related to intracellular transport, e.g. “Golgi vesicle transport”, “vesicle mediated transport or “intracellular transport” (Fig. 6; rows 73, 74, 77; Additional file 2: Table S1 - TAW Up), were highly overrepresented in PUTs that were higher expressed when TAW was high (P-value Fisher’s exact test < 1e-08). Within these categories we detected homologs of the vacuolar sorting receptor VSR6 (AT1G30900; Additional file 3: Table S2 - TAW Down; ranks 4, 8, 487), the MATE efflux protein (AT1G51340; rank 5), the membrane trafficking proteins SYTA and SYTB (AT2G20990 and AT2G20080; rank 6 and rank 14), the UDP glucose / UDP galactose transporter UTr7 (AT4G31600; rank 11), or the sucrose transporter SUC3 (AT2G02860; ranks 19, 181, 430, 590). Biological functions of differentially expressed PUTs that respond to TEMPERATURE GO categories related to oxidative and heat stress, e.g. “response to heat”, “response to hydrogen peroxide” (Fig. 6; rows 54, 55; Additional file 2: Table S1 - TEMPERATURE _Up) were highly overrepresented (P-value Fisher’s exact test < 1e-16) in PUTs that were stronger expressed when TEMPERATURE was high. PUTs within these categories were homologs to heat shock proteins (Additional file 3: Table S2 - Temperature Up; e.g. ranks 13, 16, 17, 22, 27, 31). Other proteins identified are involved in the acclimation to heat, e.g FKBP62 which is engaged in thermotolerance (AT3G25230; ranks 5,6) [31] or the multiprotein bridging factor MBF1C (AT3G24500; rank 138) which interacts as a co-factor with bZip transcription factors [32]. GO categories related to control of gene expression, e.g. “regulation of cellular macromolecule biosynthetic process”, “negative regulation of gene expression” (Fig. 6; rows 9, 43) were also overrepresented. These categories comprised homologs of ethylene responsive transcription factors with ERF/AP2 domain (AT1G19210; ranks 116, 204, 281 - AT5G21960; ranks 119, 161, 178, 181 - AT1G74930; rank 271 - AT5G11590; rank 258) which are involved in various stress responses in plants [33]. We also identified homologs of HSF2A (AT2G26150; rank 24) which is an important heat shock transcription factor [34]. In addition to genes directly involved in control of gene expression, we also observed homologs of genes which are involved in chromatin modification [35], such as the histone methyltransferase SUVH4 (AT5G13960; rank 14) and OTLD1 (AT2G27350; rank 155) [36]. GO categories “photosynthesis, light reaction”, “photosynthesis” and “generation of precursor metabolites and energy” (Fig. 6; rows 17, 18, 19) were highly overrepresented (P-value Fisher’s exact test < 1e-06) in PUTs showing a negative response to high TEMPERATURE. Within these categories, we observed homologs of the triose phosphate/3-phosphoglycerate translocator APE2 (AT5G46110; Additional file 3: Table S2 - “TEMPERATURE_ Down”; rank 2), the large subunit of ADP-glucose pyrophosphorylase ADG2 (AT5G19220; rank 28) which catalyzes the first rate limiting step in starch biosynthesis, the cytosolic malate dehydrogenase c-NAD-MDH1 (AT1G04410; ranks 22,100) or the chloroplast protein phosphatase TAP38/PPH1 (AT4G27800; ranks 62, 241) which dephosphorylates the light harvesting complex of photosystem II [37]. Biological functions of differentially expressed PUTs in response to DAYLENGTH GO categories related to signalling and regulation, e.g. “regulation of cellular process”, “response to blue light”, “regulation of response to stimulus”, “signal transduction”, “response to hormone” (Fig. 6; rows 2, 51, 66, 69, 70) were highly overrepresented (P-value Fisher’s exact test < 1e-05) in PUTs that were higher expressed under long day conditions. PUTs within these categories were homologous to the floral homeotic protein APETALA2 (AT4G36920; Additional file 3: Table S2 - “DAYLENGTH Up”; ranks 3,4), the homeobox leucine zipper proteins HB1 (AT3G01470; rank 9), HB5 (AT5G65310; rank 1) and HB6 (AT2G22430; rank 2), the flavonoid 3 hydroxylase 2 CYP75B1 (AT5G07990; ranks 11,177, 247), the lipoxygenase LOX5 (AT3G22400; ranks 14, 116), the myb family transcription factor MYB33 (AT5G06100; rank 26, 43) or the inositol triphosphate 5 phosphatase 2 5PTASE2 (AT4G18010; ranks 12, 29, 33) which regulates growth in seedlings [38]. Within the above mentioned GO categories we also observed homologs of members of the flowering control network which is tightly coupled to sensing changes in the duration of photoperiod [39]. Examples are GIGANTEA (GI; AT1G22770; ranks 311, 317, 417), flowering locus t (FT; AT1G65480; ranks 185, 309), phytochrome B (AT2G18790; rank 327) or leafy (LFY; AT5G61850; rank 103). GO categories “ribosome biogenesis” and “translation” (Fig. 6; rows 14, 41) were highly overrepresented (P-value Fisher’s exact test < 1e-09) in PUTs that were higher expressed when daylength decreased. However the PUTs that were most strongly differentially expressed in these conditions were found within the weakly overrepresented GO categories related to gene expression, e.g. “nucleic acid metabolic process” or “gene expression”, and pyruvate metabolism, e.g. “pyruvate metabolic process” (Fig. 6; rows 7, 8, 25). PUTs within these GO categories were homologous to the small hydrophobic protein RCI2A (AT3G05880; Additional file 3: Table S2 - “DAYLENGTH_Down”; rank 2), the histidine kinase phytochrome PHYE (AT4G18130; rank 4), the GATA transcription factor GATA12 (AT5G25830; rank 5), a member of the DREB subfamily A2 (AT5G05410; rank 9), the enolase ENO1 (AT1G74030; ranks 13, 14), the thylakoid protein PSB29 (PSB29, AT2G20890; rank 24) [40] or the phosphatidylglycerol phosphate synthase PGP1 (AT2G39290; rank 42) which is involved in cold acclimation [41]. In addition we observed many NAC domain containing transcription factors (e.g. AT4G29230; rank 27 – AT1G01720; rank 34 – AT4G29230; rank 51). PUTs with homology to RCI2A were also found in GO category “osmotic stress” which was overrepresented as well (Additional file 2: Table S1 - DAYLENGTH Down). Within this category we identified homologs of the cold regulated proteins COR314 and COR413 (AT1G29390; rank 6 – AT1G29395; rank 10), the alcohol dehydrogenases ADH and ADH2 (AT1G77120; ranks 11, 15 – AT5G43940; rank 16), RCI2B (AT3G05890; rank 37) or the NADPH dependent aldo-keto reductase AKR4C9 (AT2G37770; rank 40). Biological functions of PUTs differentially expressed between provenances GO categories related to pigment biosynthesis or photosynthesis, e.g. “cofactor metabolic process”, “isoprenoid metabolic process”, “photosynthesis” or “plastid organization” (Fig. 6; rows 28, 21,18, 16) were highly overrepresented (P-value Fisher’s exact test < 1e-06) in PUTs that were higher expressed in Cameron Lake than in Salmon Arms. Upregulated PUTs observed within the GO categories related to photosynthesis included homologs of the blue light receptor NPL1 (AT5G58140; Additional file 3: Table S2 - “PROVENANCE_Up”; rank 8) which mediates stomatal opening and chloroplast movement [42], the protease DEG1 (AT3G27925; rank 9) which is targeted to the chloroplast to repair damages of the photo system [43], the chlorophyll A oxigenase CAO (AT1G44446; rank 40) which enhances photosystem efficiency by increasing the antenna size of photosystems [44], or the small subunit of RUBISCO (AT1G67090; rank 223). PUTs within GO categories related to pigment biosynthesis were homologous to the phytoene synthase PSY (AT5G17230; rank 12), the NADPH thioredoxin reductase NTRC (AT2G41680; rank 16), the ferrochelatase 2 FC-II (AT2G30390; ranks 35,155), or the zeaxanthin epoxidase ABA1 (AT5G67030; rank 79) which converts the photoprotective xanthophyll zeaxanthin into antheraxanthin and violaxanthin [45]. Only the GO category “phenylpropanoid metabolism” was overrepresented in PUTs that were higher expressed in Salmon Arm (Fig. 6; row 48). Within this GO category we identified PUTs that were homologous to the most basal enzymes of the phenylpropanoid pathway, e.g. the 4-coumarate ligase 4CL (AT1G20510; Additional file 3: Table S2 - “PROVENANCE_Down”; ranks 1–5), the cinnamate-4 hydroxylase C4H (AT2G30490; rank 17), the O-methyltransferase OMT1 (AT5G05170; rank 12) or the chalcone synthase TT4 (AT5G13930; rank 7). Discussion A large part of the Douglas-fir transcriptome responds to variations in environmental conditions in the field We have investigated transcript expression in needles of adult Douglas-fir trees growing under natural field conditions. We assessed differences in transcriptome dynamics in response to variations in environmental conditions but also variations in transcript abundance among individual trees and among two differentially adapted provenances. The alignment of our RNA-Seq data to our non-redundant Douglas-fir PUT set using the unigene catalogues of [46, 47] allowed identification of 59189 expressed PUTs. These 60 k PUTs correspond to 14539 unique hits in the Picea glauca gene catalogue [48]. We assume that these unique hits correspond to an equivalent number of unique gene loci which is in concordance with the number of expressed genes in Pinaceae needle tissue [49]. Despite stringent cut-off values (FDR <0.01, dR2 >0.2), many Picea homologs were differentially expressed in response to variations in environmental conditions: 15 % for TAW, 10 % for TEMPERATURE, 16 % for DAYLENGTH and 82 % for DATE. A recent gene expression study conducted in Pinaceae detected 5794 of 14691 (FDR <0.01) ortholog sequences among Picea and Pinus to be differentially expressed in response to environmental conditions in a growth chamber [18], this being in the range of our estimates. The most comprehensive transcriptome analysis performed in Oryza sativa under field conditions estimated that 43 % of all expressed genes respond to temperature, radiation and other macroenvironmental factors [50]. Richards et al. [23] detected variation in expression over time in almost all genes in two Arabidopsis accessions that were grown in a natural environment. This indicates that a large part of expressed genes in leaf tissue responds to variations in natural environmental conditions. PUTs that are differentially expressed in response to TEMPERATURE and DAYLENGTH reveal homology to genes controlled by heat stress and photoperiod High temperature at the field sites clearly shaped gene expression in Douglas-fir needles. We observed stronger expression of PUTs that are related to heat shock proteins and other heat stress related proteins like MBF1 when TEMPERATURE was high, e.g. the ascorbate peroxidase APX2 or HSFA2 [34]. In addition we observed stronger expression of PUTs related to ERF/AP2 family transcription factors that are also well known to be stronger expressed in response to stress [51]. In contrast, PUTs related to photosynthetic activity, sugar and energy metabolism were weaker expressed when TEMPERATURE was high. Notable examples are PUTs related to the triosephosphate/3-phosphoglycerate translocator APE2 which is a key component in transporting assimilated carbon from the chloroplast into the cytosol, or the ADP-glucose pyrophosphorylase ADG2 which is important in starch biosynthesis. Reduced expression of genes involved in photosynthesis in response to temperature stress has already been described by [52] in Arabidopsis and by [53] in Arabidopsis, Populus and Glycine. Taken together, the observed gene expression pattern suggests a highly conserved response to temperature in several herbaceous plants and trees such as poplar, and based on our data also in Douglas-fir. Day length controlled more than 16 % of the transcriptome. Major GO categories that were overrepresented in PUTs that were more abundant when day length was long were “meristeme development” or “response to hormone stimulus” (Fig. 6; rows 8–11). Upon deeper analysis, we observed many members of the gene network that controls flowering in angiosperms [39]. Exposure to short day length resulted in a complex response of the transcriptome. PUTs that were higher expressed when DAYLENGTH was low displayed homology to proteins involved in cold acclimation like RCI2 -A and -B [54], ADH [55] or the NADPH dependent aldo-keto reductase AKR4C9 [56]. Preparation for cold acclimation is associated with osmotic stress [57]. Thus, it is not unexpected that we observed PUTs with homology to DREB2 [58] or NAC domain containing transcription factors that are known to be induced by dehydration stress [51]. The most comprehensive study of cold acclimation in a natural environment has been conducted in Picea sitchensis seedlings where the authors contrasted gene expression during October, November and December with late summer gene expression during August [19]. Several transcripts with homology to proteins that are known to be involved in the adaptation to cold were identified in this study. Although their sampling time-points did only partially overlap with ours, over 40 % of the best hits in the Arabidopsis thaliana peptide data base observed in [19] were also observed in the differentially expressed PUTs in response to DAYLENGTH. This consistent pattern is remarkable as we used a different measurement method, used strict filter criteria to classify PUTs as significant, and used adult individuals of a different conifer species. Examples for overlapping annotations are GI or RCI2-A. Interestingly, neither [19] nor our study detected stronger expression of CBF/DREB transcription factors which are known to be essential for acclimation to low temperatures in Arabidopsis thaliana [59, 60] when day length decreases. Reduced TAW induced expression of transcripts related to starch metabolism, a conserved response to reduced water availability in plants In response to low TAW, GO categories related to nucleic acid and starch metabolism were overrepresented, while GO categories related to intracellular transport were overrepresented when TAW was high (Fig. 6; Rows 6–12, 44, 45 and 73–81 respectively). Although no GO categories related to osmotic stress were significantly overrepresented we observed upregulation of individual PUTs that are involved in osmoregulation. We observed homologs of the osmosensor HK1 (AT2G17820; rank 179) which is higher abundant in Arabidopsis when osmolarity is especially high or low [61], or to aquaporins which are involved in handling osmotic stress [62]. In particular the aquaporins TIP1 (AT2G36830; rank 140), PIP2.8 (AT2G16850; rank 234), and PIP2.2 (AT2G37170; rank 279) were identified. Compared to the response to reduced day length, fewer specific indicators for osmotic stress were identified. However, we identified several homologs representing general stress responses when TAW was low. We observed PUTs with homology to the AP2/ERF domain containing transcription factors ERF-1 (AT4G17500; Additional file 3: Table S2 - TAW_Down; rank 70), DEAR2 (AT5G67190; rank 146) and EBP (AT3G16770; rank 224). AP2/ERF domain containing transcription factors are involved in general stress responses but are also involved in osmotic stress [51]. In addition we observed homologs of the copper/zinc superoxide dismutase CSD1 (AT1G08830; rank 40) which is known to be expressed in drought stressed plants [63], the NADPH dependent thioredoxin reductase NTRC (AT2G41680; rank 46) and the protochlorophyllide oxidoreductase PORA (T5G54190; rank 81) which protect the chloroplast against oxidative damage [64, 65]. These annotations are indicators for a certain amount of stress that the trees were confronted with during June and July at the site Wiesloch when TAW was low. Transcription factors that are typically not known to be involved in the response to reduced water availability were also identified in the PUTs that were stronger abundant when water availability was low. Examples are the helix-loop-helix protein CIB (AT1G26260; rank 77), the transcriptional repressor MYB4 (AT4G38620; ranks 124, 132, 329, 334) which is involved in the response to UV-B [66] or SIG5 (AT5G24120; rank 171) which is expressed in response to high light [67]. In contrast to the responses to heat and changes in photoperiod, gene expression responses to decreases in water availability are less uniform among different experiments conducted in the same species and organ [68]. This is because the experimental manipulation of water stress is far more difficult to control than the manipulation of photoperiod and temperature. Nevertheless, an increase in starch metabolism under conditions of water shortage was also observed under conditions of mild drought stress in Arabidopsis thaliana [14]. A recent meta-analysis also revealed that enhanced expression in response to drought is conserved among species including Oryza sativa, Arabidopsis thaliana, Triticum aestivum or Glycine max [69]. Pinheiro and Chaves [68] and Prasch and Sonnewald [52] did also report downregulation of genes related to intracellular transport in Arabidopsis thaliana which is consistent with our data. We also observed overrepresentation of the GO category “cell growth” in PUTs that were higher expressed when TAW was high (Additional file 2: Table S1 - TAW_Up). Thus, our results indicate reduced cell growth and proliferation when water availability is low. This has also been reported for Arabidopsis thaliana by [70] and [14]. Indicators for the reduced growth are homologs of WEE1 which negatively regulates the entry into mitosis [29] that were higher expressed when TAW was low. In contrast to the response to high TEMPERATURE, no GO categories related to photosynthesis were overrepresented in PUTs that were weaker expressed when TAW was low. This indicates no effect of low water availability on photosynthetic related gene expression. This phenomenon has also been reported by [14] in Arabidopsis thaliana. The two provenances Cameron Lake and Salmon Arm differ in constitutive expression of transcripts related to photosynthesis The effect of environment on the global transcript expression was high, nonetheless expression differences between the provenances were rather small, since only 1764 PUTs were differentially expressed between the provenances. The weak expression differences between the provenances were not only indicated by the low number of differentially expressed PUTs, but also by the generally small variation in PUT expression levels among provenances (Fig. 5b) and a high variation in expression among trees (Fig. 5c). These findings are consistent with reports by [19] who detected only about 900 differentially expressed transcripts between provenances from contrasting habitats in Picea sitchensis, or by [71] who reported weak genetic variation in the metabolite abundance among multiple Douglas-fir families. Despite their small number, the transcripts that are differentially expressed between the two provenances are likely responsible for differences in adaptive traits. For example, PUTs that were observed in higher abundance in Cameron Lake were related to photosynthesis. Examples are homologs of the serine/threonine kinase NPL1 (also known as PHOT2), which can act as a blue light photoreceptor and is involved in controlling of stomatal opening [72]. Overexpression of NPL1 resulted in enhanced photosynthetic activity and growth in Arabidopsis thaliana [42]. Other examples are the protease DEG1, which is targeted to the chloroplast to repair damages of the photosystems [43] or the chlorophyll A oxygenase CAO which enhances the efficiency of the photosystems by increasing the antenna size of photosystems [44]. The higher expression of these genes might translate into a generally higher photosynthetic activity in Cameron Lake. Moreover, it has been frequently reported from common garden experiments that water use efficiency (WUE) is higher in coastal than in interior Dougals fir [73, 74] and [75]. This is a counterintuitive observation, since one would expect that interior Douglas fir provenances from higher altitudes would generally be better adapted to episodic water limitations, which is often associated with increased WUE. Aitken et al. [74] and Zhang et al. [75] suggested that the location of the experimental sites must have an influence on WUE. However, the observation that NPL1/PHOT2, which controls stomatal opening and hence WUE, is higher expressed in the coastal provenance Cameron Lake indicates that higher WUE in a coastal Douglas fir provenance might result from an increased ability for regulating stomatal behavior and mediating higher WUE. Interactions of provenance and environment The effect of the interaction among provenance and DATE which represents the genotype by environment interaction (GxE), was surprisingly low with only 21 PUTs showing a response (Fig. 3, Additional file 4: Figure S2). Only three of these PUTs were homologous to Arabidopsis thaliana genes: a calmodulin binding protein (AT2G26190), an Armadillo repeat protein (AT4G34940) and a subunit of the cytochrome oxidase COX1 (ATMG01360). Due to the small extent of the GxE effect we did not further investigate interactions among provenances and the individual environmental regressors (TAW, TEMPERATURE and DAYLENGTH) because we expected those interactions to be even weaker. Weak GxE effects on the transcriptome dynamics were also observed in field-grown seedlings of Arabidopsis thaliana from contrasting habitats [23]. In addition, there is also evidence that transcript levels of only a small number of genes are influenced by eQTL x environment interactions in Arabidopsis thaliana [76]. PUTs that responded to environment (TAW, DAYLENGHT, TEMPERATURE) showed a lower variation of PUT abundance among provenances (Fig. 5b) and trees (Fig. 5c) compared to all detected PUTs. Thus, the small GxE effect in our data from field-grown adult Douglas-fir trees suggests that plastic transcriptome responses to variations in environmental conditions are strongly conserved both at the tree and provenance level. These findings are supported by results from [18] who observed that 74 % of the genes that respond to variations in environmental conditions in a growth chamber experiment also displayed conserved expression patterns in Picea and Pinus despite the large divergence time of both species (> 140 million years). Most notably, in our experiments the expression of PUTs that responded to TEMPERATURE varied less among provenances and trees compared to all other environmental factors (Fig. 5b, c). Previous experiments performed in growth chambers comparing Glycine max, Arabidopsis thaliana and seedlings of Populus trichocarpa revealed that transcriptome dynamics in response to high temperature are conserved across these angiosperm species [53]. Nevertheless, the observation that there is such a small variation in the response to high temperature among mature field grown trees and provenances of a conifer species is striking and highly relevant for foresters for adapting forests to climate change. Overall, the small GxE effects indicate that local adaptation has a rather small effect on the ability of Douglas-fir trees to modulate gene expression and their ability to deal with novel climates. It seems that differences in gene expression between the provenances (factor PROVENANCE), indicated by 1764 differentially expressed PUTs and overrepresentation of GO categories including photosynthesis and secondary metabolism, by far outweigh the importance of GxE effects in the two provenances included in our study. Nevertheless we only investigated a rather small number of genotypes, and only over one growing season. Thus, it would be important to assess if the genes that contributed here to the plastic response to environment between the two provenances will be also involved among larger groups of populations and larger temporal scales. Conclusions Whole transcriptome responses to natural and highly variable environmental conditions were studied in adult Douglas-fir trees representing two populations from contrasting habitats. We investigated the correlation of transcript abundance with regressors that represented high temperature, photoperiod and water availability. Functional annotation and overrepresentation analysis revealed that Douglas-fir transcript regulation was similar to other species, indicating the high conservation of transcript expression in response to environmental cues. Thus our data set represents a rich repository of validated transcriptional responses to the main abiotic parameters of a natural and highly variable environment. Almost no transcripts with divergent plastic expression responses between the provenances were observed. In addition, the transcripts that responded to environmental cues varied less among trees and among provenances compared with expression variations in the transcriptome. In contrast, we observed a substantial constitutive differentiation in gene expression activity related to photosynthesis and secondary metabolism between both provenances. Therefore we assume that local adaptation in Pseudotsuga menziesii is unlikely to be driven by divergent transcriptional short-term responses to environment among populations, suggesting that local adaptation is not reflected in short-term responses and instead determines long term physiological and metabolic processes. Methods Experimental design and plant material Two Douglas-fir provenances originating from the western pacific coast of North America, Cameron Lake (LA) and westwards from the Rocky Mountains, Salmon Arm (AR) were investigated. While Cameron Lake represents a coastal Douglas-fir originating from Vancouver Island, Salmon Arm originates from an inland hybridization zone of coastal and interior Douglas-fir. The origins are in relative proximity (~ 1000 km) and vary by one degree in latitude. The origins differ in elevation (AR: 650 m, LA: 210 m), mean annual temperature (AR: 7.8 °C, LA: 10 °C) and most importantly in mean annual precipitation (AR: 500 mm, LA: 1475 mm). Recent studies using SNP [77] and microsatellite markers [78] revealed a clear genetic differentiation of these two provenances. For this study we used 50-year-old Dougls-fir trees from two common garden experiments near Schluchsee (S; 47°84′ N, 8°11′ E) and Wiesloch (W; 49°30′ N, 5°53′ E) in south-western Germany. The trees were planted during the International Douglas-fir provenance trial of 1958 [79]. The two sites differ in annual precipitation (S: 1345 mm, W: 660 mm) as well as in elevation (S: 1050 m, W: 105 m above sea level) and annual mean temperature (S: 6.1 °C, W: 9.9 °C). A detailed description of the provenances and the two field sites can be found in [78, 80]. At each field site needle samples were taken repeatedly from 8 trees per provenance and on four different dates during the 2010 growing season (in Schluchsee on May 27, June 30, July 28, September 15; in Wiesloch on May 12, June 16, July 14, September 8; these eight sampling dates reflect the levels of the factor DATE that was used for modeling of gene expression, see below). This resulted initially in a total of 128 samples. Subsequent quality testing in the lab revealed variation in RNA quality. Excluding samples with RNA that was not suitable for RNA sequencing resulted in a final number of 75 needle samples that were used to generate libraries for RNA sequencing. A detailed overview listing all trees and samples included in this study is provided in Additional file 5: Table S3. Needle samples were collected from the upper sun exposed southern canopy (~ 3 m below the top, at a height of about 25–30 m). Previous year’s needles (2009) were collected around noon (12:30–15:00) and immediately frozen in liquid nitrogen upon collection. Library preparation and Illumina mRNA sequencing After homogenization in liquid nitrogen using mortar and pestle, total RNA was extracted using a CTAB based extraction method, modified after [81]. After precipitation and resuspending the RNA, a DNase I digestion was performed, using Qiagens RNase-Free DNase Set (Cat. no. 79254). Afterwards, the DNase was removed using silica columns (Qiagen RNEasy MinElute, Cat. no. 74204). The integrity of the total RNA was checked on a 2100 Bio-analyzer (Agilent, CA, USA) using the RNA 6000 nano assay and the Plant total RNA protocol. The purity of total RNA was checked on a Nanodrop ND-1000 (Thermo Scientific, Bremen, Germany). From each sample, one deep mRNA sequencing library was prepared, using the TruSeq RNA Sample Preparation Kit v2 starting from 4 μg total RNA (Illumina, CA, USA). The libraries were prepared and sequenced on two Illumina 100 bp paired end (PE) flow cells on the Illumina HISEQ 2000 at the Genome Quebec Innovation Centre in Montreal, Canada. Based on earlier investigations [82], we aimed at an effective sequencing depth of 20 million aligned reads per sample. Alignment to joint Douglas-fir PUT set We merged the recently published putatively unique transcript (PUT) sets for Douglas-fir [46, 47] to create a non-redundant set of PUTs using CD-HIT-EST (Version 4.6) [83], a tool for fast clustering of nucleotide and protein sequences. PUTs that were entirely covered by longer PUTs and have 99 % sequence identity with the longer PUT were removed. Finally all PUTs smaller than 200 bp were discarded. The 100 bp paired-end (PE) reads from the Illumina sequencing were aligned to the joint Douglas-fir PUT-set with Bowtie2 (Version 2.1.0) [84] using the global alignment mode and otherwise default settings. The alignment was parsed and analyzed with a custom Python script which builds on the HTSeq python library [85]. All hits below a threshold sequence identity of 95 % and alignment length of 80 were discarded. When multiple best hits existed, i.e. hits with the same sequence identity and the same alignment length, one of these hits was randomly selected by Bowtie2. The PUT set contains potential splice variants. Splice variants that are anticipated to emerge from a common genomic locus are grouped in one isogroup and sequence stretches may appear multiple times in the PUT set. Thus, uniqueness of a hit was not considered. Subsequently, reads that aligned to a PUT were counted and converted to counts per PUT using a custom Python script. Only PUTs with more than 1 read per million aligned reads (RPM) in at least four libraries were retained. Functional annotation of PUTs To functionally annotate the detected PUTs, we did a BLASTX search in the Arabidopsis thaliana peptide data base (TAIR 10) and the NCBI Plant RefSeq peptide data base (date of download: May-8-2013) (BLASTX, NCBI BLAST+ suite, Version 2.2.24+, E < 1e-3, sequence identity >40 % ). For Gene Ontology (GO) [86] annotation, the results of the BLASTX search against the NCBI plant RefSeq proteins (RefSeq) were fed into the Blast2GO pipeline [87]. Functional annotation is also described in [82]. The detected PUTs were also aligned to a high quality mostly full length Picea glauca EST set [48], to a set of transcripts of Pinus taeda which has been used to annotate the Pinus taeda genome [88] and to Vitis vinifera and Oryza sativa peptides stored in the PLAZA data base (Version 2.5) [89] using RAPSEARCH [90]. RAPSEARCH is similar to BLAST but uses a reduced amino acid alphabet to increase processing speed and the identification of seeds in the query sequences. It is thus 20 to 90 times faster than BLAST at the drawback of a slightly reduced sensitivity [90]. The reduced sensitivity results in missing alignments with high E-values. We retained alignments identified by RAPSEARCH with E-value < 1e-5. The expressed PUTs were assigned to gene families stored in the PLAZA data base using the TRAPID functional annotation pipeline [91]. Data transformation and exploratory analysis Normalization for sequencing library size and variance stabilizing transformation Differences in library size between the deep sequencing libraries were corrected using the method implemented in the Bioconductor [92] package DESeq [93] which is one of the most robust methods to correct for library size [94]. Log transformation of the number of aligned reads per PUT (Fig. 1c) resulted in normally distributed data for more than 95 % of PUTs (Kolmogorov-Smirnov test, P > 0.05). Estimation of variance components attributed to individual tree, provenance and environment To estimate the amount of variance in PUT abundance that is attributed to each individual tree, provenance or environment, we estimated the respective variance components using a linear random effects model. We fit a model that included the individual tree (TREE), the common garden (SITE), the provenance (PROVENANCE) and the sampling time-point (DATE) to the transformed count data of each detected PUT using the R [95] package lme4 [96] and restricted maximum likelihood (REML). Variance components for each random effect were extracted and divided against the sum of all variance components and the residual variance. Linear modelling of PUT abundance Description of environmental parameters and genetic factors as regressors for the linear regression models Total available soil water (TAW), air temperature (TEMPERATURE) and day length (DAYLENGTH) were used as environmental parameters in our analysis. TAW was obtained from the forest water model WBS3 [97] which uses mean daily temperature and daily precipitation as the meteorological input parameters, combined with latitude, soil type, plant cover, slope and slope aspect. Mean daily temperature and mean daily precipitation were measured at two weather stations, one operated by “Deutscher Wetterdienst” (DWD) close to the field site Wiesloch and one privately operated close to the field site Schluchsee. TEMPERATURE in our analysis represents the average air temperature of the sampling day recorded between 10:00 am and 2:00 pm centered to the four week running average of air temperature. Centering was performed to detrend the temperature which is correlated with day length and to identify time-points of above-average temperature. Day length (DAYLENGTH) on the day of sampling was centered to the length of the longest day of the year (solstice on June 21). To account for the season and the direction of an increasing versus a decreasing day length before and after solstice, day length before June 21 was assigned a positive value and after June 21 a negative value. The field site and the provenance were encoded as dummy variables (SITE and PROVENANCE, respectively). Detection of differential expression Differential PUT expression was investigated using linear mixed models. Models were fit to the log-transformed count data using the function lmer in lme4 [96]. We compared nested models using an F-test with Kenward-Roger approximation implemented in the R-package pbkrtest. P-values from the F-test were adjusted for multiple testing using the Benjamini-Hochberg procedure. In addition to the F-test the dAIC and dR2 were calculated as the difference in Akaike information criterion (AIC) and coefficient of determination (R2) of the model containing the regressor compared to the model without it (Fig. 1e). The R2 for the fixed effects part of the model was calculated according to [98]. Rmarg2=σfix2σfix2+σrand2+σerr2 (σ2fix = variance attributed to fixed effects, σ2rand = variance attributed to random effects, σ2err = residual variance). To test for differential expression between provenances and sampling time-points (DATE) including interaction of both we set up and compared four models (Fig. 1e). Each of the models contained TREE as random intercept. After testing for the interaction effect it was removed from the model and the effects of DATE and PROVENANCE were tested. Differential expression in response to the environmental parameters represented by TEMPERATURE, DAYLENGTH and TAW was investigated by removing a single environmental parameter from the model TAW + TEMPERATURE + DAYLENGTH + PROVENANCE and comparing both models. PUTs are only considered to respond to an environmental parameter if the corresponding regressor increases the R2 by more than 0.2, in addition to an FDR smaller than 0.01. Multicollinearity (variance inflation factor >10) did not allow for introducing water availability (TAW) and common garden (SITE) simultaneously into the model (Additional file 1: Figure S1), thus we did not include SITE. We did not consider interactions among the environmental parameters to avoid over fitting due to the limited number of sampling points (n = 8). Identification of Gene Ontology (GO) category overrepresentation GO categories were defined as significant, if they are overrepresented in PUTs that respond to a regressor (p < 0.01, Fisher exact test). Furthermore, we required significant GO categories to contain more than ten PUTs that responded to a regressor. Overrepresentation analyses were conducted using the Bioconductor package topGO [99]. The PUTs investigated in our analysis are assembled EST sequences and therefore contain various isoforms of a gene product as well as incompletely assembled sequences, although overrepresentation analyses by GO categories assume a single gene locus. Therefore, we reduced the risk of artificially inflated abundance of annotations by randomly selecting one PUT from all PUTs that shared a best hit in the NCBI plant RefSeq data base. We repeated the overrepresentation analysis 100 times and averaged the results because different PUTs with the same best BLAST hit might not equally respond to a regressor. For visualization of the results of the GO analysis, GO categories were grouped according to the amount of overlapping PUTs among two GO categories. This overlap was identified based on the Kappa statistic (Cohen’s Kappa). κ=Omn−Amn1−Amn Clustering of genes based on the Kappa statistic has been described by [100]. In brief the number of shared PUTs among two GO categories is expressed by the Kappa statistic calculated from presence absence matrices (0 / 1) where rows correspond to GO categories and columns correspond to PUTs. Omn corresponds to the co-occurrence of PUTs in GO categories m and n while Amn represents the co-occurrence of PUTs in GO categories m and n expected by chance. In contrast to [100] who employed a heuristic fuzzy partition algorithm, we identified clusters of GO categories which shared PUTs by hierarchical clustering on distance matrices created from the Kappa scores (method “euclidean”). As basal GO terms do generally carry a low information content [101], we selected only GO categories with at least five parental terms. Additional files Additional file 1: Figure S1. Relationship among environmental regressors. Color indicates the absolute value of Pearson’s correlation coefficient calculated for each regressor pair. TAW = Total available soil water, SITE = common garden, TEMPERATURE. DAYLENGTH = interaction of TEMPERATURE and DAYLENGTH, TAW. DAYLENGTH = interaction of TAW and DAYLENGTH, TAW. TEMPERATURE = interaction among TAW and TEMPERATURE. (PDF 333 kb) Additional file 2: Table S1. Results of the overrepresentation analysis (P-value Fisher’s exact test <0.01). P-value Fisher’s exact test = P-value from the overrepresentation analysis, Expected = expected frequency of the GO category, given the abundance of the GO category in the set of all detected PUTs, Significant = observed frequency of GO category. Results are ranked based on the P-value of the overrepresentation test (Fisher’s exact test). (XLS 105 kb) Additional file 3: Table S2. Functional annotation of differentially expressed PUTs within overrepresented GO categories. P-values from the test of differential expression were averaged over all PUTs (PUTs) with the same best hit in the RefSeq data base. Annotations are ranked (Rank) according to the absolute log10 transformed P-value (p_Value_log) from the test of differential expression. Accession = RefSeq ID of the best hit in the RefSeq data base, GI = NCBI Entrez ID of the best hit in the RefSeq data base, Ara = best hit in the Arabidopsis thaliana peptide data base (TAIR10), Annot = annotation inferred by the Blast2GO pipeline, Ara ID = Arabidopsis thaliana gene id, GO Reference = GO annotation shown in Fig. 6, PUTs = significant PUTs, separated by “--”, BLAST TAIR10 = e-value of the best hit in the TAIR10 peptide database inferred by a BLASTX search for the corresponding PUT, separated by “--”, BLAST RefSeq = e-value of the best hit in the RefSeq database inferred by a BLASTX search for the corresponding PUT, separated by “--”. (XLSX 738 kb) Additional file 4: Figure S2. Expression of the 21 PUTs with significant GxE interaction effect (FDR <0.01). Expression is shown for each time-point on each common garden. Number of aligned reads were normalized for sequencing depth and log10 transformed. Whiskers of boxplots extend to 1.5 times the interquartile range, dots represent outliers. Cameron Lake = cyan color, Salmon Arm = magenta. (PDF 50 kb) Additional file 5: Table S3. Detailed information about the sample characteristics. (XLS 36 kb) Abbreviations AICAkaike information criterion ARSalmon Arm dAICdelta AIC dR2delta R2 DWDDeutscher Wetterdienst EEnvironment FDRFalse discovery rate GGenotype GOGene ontology GxEGene by environment k1000 LACameron Lake LMMLinear mixed model MreadMega reads PEPaired end PUTPutative unique transcript REMLRestricted maximum likelihood RPMReads per million aligned reads SSchluchsee TAWTotal available soil water TMMTrimmed mean of M-values WWiesloch WUEWater use efficiency Acknowledgements This project is part of the collaborative project ‘DougAdapt’ with funding from the German Research Foundation to IE (DFG-project EN 829/4-1). The article processing charge was funded by the German Research Foundation (DFG) and the Albert Ludwigs University Freiburg in the funding programme Open Access Publishing. The authors are grateful to Anita Kleiber and Anna-Maria Weisser for technical assistance in the field and in the lab. The authors would like to thank Wolfgang Hess for valuable comments and discussions. Availability of supporting data The data set supporting the results of this article is available in the NCBI gene expression omnibus (GEO; Accession number GSE73420). Authors’ contributions MH processed and analyzed data, discussed data and drafted the manuscript. HW supervised research and contributed to data analysis, discussion and writing of the manuscript. LJ modelled TAW and contributed to writing of the manuscript. IE designed and supervised research, contributed to data analysis, discussion and writing of the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Kawecki TJ Ebert D Conceptual issues in local adaptation Ecol Lett 2004 7 1225 1241 10.1111/j.1461-0248.2004.00684.x 2. Savolainen O Lascoux M Merilä J Ecological genomics of local adaptation Nat Rev Genet 2013 14 807 820 10.1038/nrg3522 24136507 3. Fournier-Level A Korte A Cooper MD Nordborg M Schmitt J Wilczek AM A map of local adaptation in Arabidopsis thaliana Science 2011 334 86 9 10.1126/science.1209271 21980109 4. Leimu R Fischer M A meta-analysis of local adaptation in plants PLoS One 2008 3 e4010 10.1371/journal.pone.0004010 19104660 5. Hancock AM Brachi B Faure N Horton MW Jarymowycz LB Sperone FG Toomajian C Roux F Bergelson J Adaptation to climate across the Arabidopsis thaliana genome Science 2011 334 83 6 10.1126/science.1209244 21980108 6. Eckert AJ Bower AD González-Martínez SC Wegrzyn JL Coop G Neale DB Back to nature: ecological genomics of loblolly pine (Pinus taeda, Pinaceae) Mol Ecol 2010 19 3789 805 10.1111/j.1365-294X.2010.04698.x 20723060 7. Prunier J Laroche J Beaulieu J Bousquet J Scanning the genome for gene SNPs related to climate adaptation and estimating selection at the molecular level in boreal black spruce Mol Ecol 2011 20 1702 16 10.1111/j.1365-294X.2011.05045.x 21375634 8. Holliday JA Ritland K Aitken SN Widespread, ecologically relevant genetic markers developed from association mapping of climate-related traits in Sitka spruce (Picea sitchensis) New Phytol 2010 188 501 14 10.1111/j.1469-8137.2010.03380.x 20663060 9. Gonzalez-Martinez SC Huber D Ersoz E Davis JM Neale DB Association genetics in Pinus taeda L. II. Carbon isotope discrimination Heredity 2008 101 19 26 10.1038/hdy.2008.21 18478029 10. Eckert AJ Bower AD Wegrzyn JL Pande B Jermstad KD Krutovsky KV St Clair JB Neale DB Association genetics of coastal Douglas fir (Pseudotsuga menziesii var. menziesii, Pinaceae). I. Cold-hardiness related traits Genetics 2009 182 1289 302 10.1534/genetics.109.102350 19487566 11. Solomon S, editor. Climate change 2007-the physical science basis: Working group I contribution to the fourth assessment report of the IPCC. Cambridge University Press; 2007. 12. Franks SJ Weber JJ Aitken SN Evolutionary and plastic responses to climate change in terrestrial plant populations Evol Appl 2014 7 123 39 10.1111/eva.12112 24454552 13. Van Leeuwen H Kliebenstein DJ West MA Kim K van Poecke R Katagiri F Michelmore RW Doerge RW Clair DAS Natural variation among Arabidopsis thaliana accessions for transcriptome response to exogenous salicylic acid Plant Cell Online 2007 19 2099 2110 10.1105/tpc.107.050641 14. Des Marais DL McKay JK Richards JH Sen S Wayne T Juenger TE Physiological genomics of response to soil drying in diverse Arabidopsis accessions Plant Cell 2012 24 893 914 10.1105/tpc.112.096180 22408074 15. Hannah MA Wiese D Freund S Fiehn O Heyer AG Hincha DK Natural genetic variation of freezing tolerance in Arabidopsis Plant Physiol 2006 142 98 112 10.1104/pp.106.081141 16844837 16. Hamanishi ET Raj S Wilkins O Thomas BR Mansfield SD Plant AL Campbell MM Intraspecific variation in the Populus balsamifera drought transcriptome Plant Cell Environ 2010 33 1742 55 10.1111/j.1365-3040.2010.02179.x 20525001 17. Rengel D Arribat S Maury P Martin-Magniette M-L Hourlier T Laporte M Varès D Carrère S Grieu P Balzergue S Gouzy J Vincourt P Langlade NB A gene-phenotype network based on genetic variability for drought responses reveals key physiological processes in controlled and natural environments PLoS One 2012 7 e45249 10.1371/journal.pone.0045249 23056196 18. Yeaman S Hodgins KA Suren H Nurkowski KA Rieseberg LH Holliday JA Aitken SN Conservation and divergence of gene expression plasticity following c. 140 million years of evolution in lodgepole pine (Pinus contorta) and interior spruce (Picea glauca × Picea engelmannii) New Phytol 2014 203 578 91 10.1111/nph.12819 24750196 19. Holliday JA Ralph SG White R Bohlmann J Aitken SN Global monitoring of autumn gene expression within and among phenotypically divergent populations of Sitka spruce (Picea sitchensis) New Phytol 2008 178 103 22 10.1111/j.1469-8137.2007.02346.x 18194148 20. Anderson JT Wagner MR Rushworth CA Prasad KVSK Mitchell-Olds T The evolution of quantitative traits in complex environments Heredity 2014 112 4 12 10.1038/hdy.2013.33 23612691 21. Harfouche A Meilan R Altman A Molecular and physiological responses to abiotic stress in forest trees and their relevance to tree improvement Tree Physiol 2014 34 1181 1198 10.1093/treephys/tpu012 24695726 22. Dal Santo S Tornielli GB Zenoni S Fasoli M Farina L Anesi A Guzzo F Delledonne M Pezzotti M The plasticity of the grapevine berry transcriptome Genome Biol 2013 14 r54 10.1186/gb-2013-14-6-r54 23759170 23. Richards CL Rosas U Banta J Bhambhra N Purugganan MD Genome-wide patterns of Arabidopsis gene expression in nature PLoS Genet 2012 8 e1002662 10.1371/journal.pgen.1002662 22532807 24. Travers SE Tang Z Caragea D Garrett KA Hulbert SH Leach JE Bai J Saleh A Knapp AK Fay PA Nippert J Schnable PS Smith MD Variation in gene expression of Andropogon gerardii in response to altered environmental conditions associated with climate change J Ecol 2010 98 374 383 10.1111/j.1365-2745.2009.01618.x 25. Brosché M Vinocur B Alatalo ER Lamminmäki A Teichmann T Ottow EA Djilianov D Afif D Bogeat-Triboulot M-B Altman A Polle A Dreyer E Rudd S Paulin L Auvinen P Kangasjärvi J Gene expression and metabolite profiling of Populus euphratica growing in the Negev desert Genome Biol 2005 6 R101 10.1186/gb-2005-6-12-r101 16356264 26. Gugger PF Sugita S Cavender-Bares J Phylogeography of Douglas-fir based on mitochondrial and chloroplast DNA sequences: testing hypotheses from the fossil record Mol Ecol 2010 19 1877 97 10.1111/j.1365-294X.2010.04622.x 20374486 27. Wei X-X Beaulieu J Khasa DP Vargas-Hernández J López-Upton J Jaquish B Bousquet J Range-wide chloroplast and mitochondrial DNA imprints reveal multiple lineages and complex biogeographic history for Douglas-fir Tree Genet Genomes 2011 7 1025 1040 10.1007/s11295-011-0392-4 28. St Clair JB Mandel NL Vance-Borland KW Genecology of Douglas fir in western Oregon and Washington Ann Bot 2005 96 1199 1214 10.1093/aob/mci278 16246849 29. De Schutter K Joubès J Cools T Verkest A Corellou F Babiychuk E Van Der Schueren E Beeckman T Kushnir S Inze D Arabidopsis WEE1 kinase controls cell cycle arrest in response to activation of the DNA integrity checkpoint Plant Cell 2007 19 211 225 10.1105/tpc.106.045047 17209125 30. Wang D Guo Y Wu C Yang G Li Y Zheng C Genome-wide analysis of CCCH zinc finger family in Arabidopsis and rice BMC Genomics 2008 9 44 10.1186/1471-2164-9-44 18221561 31. Meiri D Breiman A Arabidopsis ROF1 (FKBP62) modulates thermotolerance by interacting with HSP90. 1 and affecting the accumulation of HsfA2-regulated sHSPs Plant J 2009 59 387 399 10.1111/j.1365-313X.2009.03878.x 19366428 32. Suzuki N Sejima H Tam R Schlauch K Mittler R Identification of the MBF1 heat-response regulon of Arabidopsis thaliana Plant J 2011 66 844 851 10.1111/j.1365-313X.2011.04550.x 21457365 33. Shinozaki K Yamaguchi-Shinozaki K Gene networks involved in drought stress response and tolerance J Exp Bot 2007 58 221 7 10.1093/jxb/erl164 17075077 34. Kotak S Larkindale J Lee U von Koskull-Döring P Vierling E Scharf K-D Complexity of the heat stress response in plants Curr Opin Plant Biol 2007 10 310 6 10.1016/j.pbi.2007.04.011 17482504 35. Campi M D’Andrea L Emiliani J Casati P Participation of chromatin-remodeling proteins in the repair of ultraviolet-B-damaged DNA Plant Physiol 2012 158 981 995 10.1104/pp.111.191452 22170978 36. Krichevsky A Zaltsman A Lacroix B Citovsky V Involvement of KDM1C histone demethylase–OTLD1 otubain-like histone deubiquitinase complexes in plant gene repression Proc Natl Acad Sci 2011 108 11157 11162 10.1073/pnas.1014030108 21690391 37. Pribil M Pesaresi P Hertle A Barbato R Leister D Role of plastid protein phosphatase TAP38 in LHCII dephosphorylation and thylakoid electron flow PLoS Biol 2010 8 274 10.1371/journal.pbio.1000288 38. Gunesekera B Torabinejad J Robinson J Gillaspy GE Inositol polyphosphate 5-phosphatases 1 and 2 are required for regulating seedling growth Plant Physiol 2007 143 1408 1417 10.1104/pp.106.089474 17237190 39. Turck F Fornara F Coupland G Regulation and identity of florigen: FLOWERING LOCUS T moves center stage Annu Rev Plant Biol 2008 59 573 594 10.1146/annurev.arplant.59.032607.092755 18444908 40. Keren N Ohkawa H Welsh EA Liberton M Pakrasi HB Psb29, a conserved 22-kD protein, functions in the biogenesis of photosystem II complexes in Synechocystis and Arabidopsis Plant Cell 2005 17 2768 2781 10.1105/tpc.105.035048 16155179 41. Xu C Härtel H Wada H Hagio M Yu B Eakin C Benning C The pgp1 mutant locus of Arabidopsis encodes a phosphatidylglycerolphosphate synthase with impaired activity Plant Physiol 2002 129 594 604 10.1104/pp.002725 12068104 42. Wang Y Noguchi K Ono N Inoue S Terashima I Kinoshita T Overexpression of plasma membrane H + -ATPase in guard cells promotes light-induced stomatal opening and enhances plant growth Proc Natl Acad Sci 2014 111 533 538 10.1073/pnas.1305438111 24367097 43. Kapri-Pardes E Naveh L Adam Z The thylakoid lumen protease Deg1 is involved in the repair of photosystem II from photoinhibition in Arabidopsis Plant Cell 2007 19 1039 1047 10.1105/tpc.106.046573 17351117 44. Tanaka R Koshino Y Sawa S Ishiguro S Okada K Tanaka A Overexpression of chlorophyllide a oxygenase (CAO) enlarges the antenna size of photosystem II in Arabidopsis thaliana Plant J 2001 26 365 373 10.1046/j.1365-313X.2001.2641034.x 11439124 45. Barrero JM Piqueras P González-Guzmán M Serrano R Rodríguez PL Ponce MR Micol JL A mutational analysis of the ABA1 gene of Arabidopsis thaliana highlights the involvement of ABA in vegetative development J Exp Bot 2005 56 2071 2083 10.1093/jxb/eri206 15983017 46. Müller T Ensminger I Schmid KJ A catalogue of putative unique transcripts from Douglas-fir (Pseudotsuga menziesii) based on 454 transcriptome sequencing of genetically diverse, drought stressed seedlings BMC Genomics 2012 13 673 10.1186/1471-2164-13-673 23190494 47. Howe GT Yu J Knaus B Cronn R Kolpak S Dolan P Lorenz WW Dean JF A SNP resource for Douglas-fir: de novo transcriptome assembly and SNP detection and validation BMC Genomics 2013 14 137 10.1186/1471-2164-14-137 23445355 48. Rigault P Boyle B Lepage P Cooke JEK Bousquet J MacKay JJ A white spruce gene catalog for conifer genome analyses Plant Physiol 2011 157 September 14 28 10.1104/pp.111.179663 21730200 49. Raherison E Rigault P Caron S Poulin P-L Boyle B Verta J-P Giguère I Bomal C Bohlmann J Mackay J Transcriptome profiling in conifers and the PiceaGenExpress database show patterns of diversification within gene families and interspecific conservation in vascular gene expression BMC Genomics 2012 13 434 10.1186/1471-2164-13-434 22931377 50. Nagano AJ Sato Y Mihara M Antonio BA Motoyama R Itoh H Nagamura Y Izawa T Deciphering and prediction of transcriptome dynamics under fluctuating field conditions Cell 2012 151 1358 69 10.1016/j.cell.2012.10.048 23217716 51. Yamaguchi-Shinozaki K Shinozaki K Transcriptional regulatory networks in cellular responses and tolerance to dehydration and cold stresses Annu Rev Plant Biol 2006 57 781 803 10.1146/annurev.arplant.57.032905.105444 16669782 52. Prasch CM Sonnewald U Simultaneous application of heat, drought, and virus to Arabidopsis plants reveals significant shifts in signaling networks Plant Physiol 2013 162 1849 66 10.1104/pp.113.221044 23753177 53. Weston DJ Karve AA Gunter LE Jawdy SS Yang X Allen SM Wullschleger SD Comparative physiology and transcriptional networks underlying the heat shock response in Populus trichocarpa, Arabidopsis thaliana and Glycine max Plant Cell Environ 2011 34 1488 506 10.1111/j.1365-3040.2011.02347.x 21554326 54. Medina J Catalá R Salinas J Developmental and Stress Regulation of RCI2A andRCI2B, Two Cold-Inducible Genes of Arabidopsis Encoding Highly Conserved Hydrophobic Proteins Plant Physiol 2001 125 1655 1666 10.1104/pp.125.4.1655 11299347 55. Jarillo JA Leyva A Salinas J Martínez-Zapater JM Low temperature induces the accumulation of alcohol dehydrogenase mRNA in Arabidopsis thaliana, a chilling-tolerant plant Plant Physiol 1993 101 833 837 12231733 56. Éva C Zelenyánszki H Tömösközi-Farkas R Tamás L Transgenic barley expressing the Arabidopsis AKR4C9 aldo-keto reductase enzyme exhibits enhanced freezing tolerance and regenerative capacity South Afr J Bot 2014 93 179 184 10.1016/j.sajb.2014.04.010 57. Welling A Palva ET Molecular control of cold acclimation in trees Physiol Plant 2006 127 167 181 10.1111/j.1399-3054.2006.00672.x 58. Liu Q Kasuga M Sakuma Y Abe H Miura S Yamaguchi-Shinozaki K Shinozaki K Two transcription factors, DREB1 and DREB2, with an EREBP/AP2 DNA binding domain separate two cellular signal transduction pathways in drought-and low-temperature-responsive gene expression, respectively, in Arabidopsis Plant Cell 1998 10 1391 1406 10.1105/tpc.10.8.1391 9707537 59. Stockinger EJ Gilmour SJ Thomashow MF Arabidopsis thaliana CBF1 encodes an AP2 domain-containing transcriptional activator that binds to the C-repeat/DRE, a cis-acting DNA regulatory element that stimulates transcription in response to low temperature and water deficit Proc Natl Acad Sci 1997 94 1035 1040 10.1073/pnas.94.3.1035 9023378 60. Jaglo-Ottosen KR Gilmour SJ Zarka DG Schabenberger O Thomashow MF Arabidopsis CBF1 overexpression induces COR genes and enhances freezing tolerance Science 1998 280 104 106 10.1126/science.280.5360.104 9525853 61. Urao T Yakubov B Yamaguchi-Shinozaki K Shinozaki K Stress-responsive expression of genes for two-component response regulator-like proteins in Arabidopsis thaliana FEBS Lett 1998 427 175 178 10.1016/S0014-5793(98)00418-9 9607306 62. Maurel C Verdoucq L Luu D-T Santoni V Plant aquaporins: membrane channels with multiple integrated functions Annu Rev Plant Biol 2008 59 595 624 10.1146/annurev.arplant.59.032607.092734 18444909 63. Bray EA Classification of genes differentially expressed during water-deficit stress in Arabidopsis thaliana: An analysis using microarray and differential expression data Ann Bot 2002 89 803 811 10.1093/aob/mcf104 12102506 64. Sperling U Cleve B Frick G Apel K Armstrong GA Overexpression of light-dependent PORA or PORB in plants depleted of endogenous POR by far-red light enhances seedling survival in white light and protects against photooxidative damage Plant J 1997 12 649 658 10.1046/j.1365-313X.1997.d01-11.x 9351249 65. Perez-Ruiz JM Spinola MC Kirchsteiger K Moreno J Sahrawy M Cejudo FJ Rice NTRC is a high-efficiency redox system for chloroplast protection against oxidative damage Plant Cell Online 2006 18 2356 2368 10.1105/tpc.106.041541 66. Jin H Cominelli E Bailey P Parr A Mehrtens F Jones J Tonelli C Weisshaar B Martin C Transcriptional repression by AtMYB4 controls production of UV-protecting sunscreens in Arabidopsis EMBO J 2000 19 6150 6161 10.1093/emboj/19.22.6150 11080161 67. Nagashima A Hanaoka M Shikanai T Fujiwara M Kanamaru K Takahashi H Tanaka K The multiple-stress responsive plastid sigma factor, SIG5, directs activation of the psbD blue light-responsive promoter (BLRP) in Arabidopsis thaliana Plant Cell Physiol 2004 45 357 368 10.1093/pcp/pch050 15111710 68. Pinheiro C Chaves MM Photosynthesis and drought: can we make metabolic connections from available data? J Exp Bot 2011 62 869 82 10.1093/jxb/erq340 21172816 69. Shaar-Moshe L Hübner S Peleg Z Identification of conserved drought-adaptive genes using a cross-species meta-analysis approach BMC Plant Biol 2015 15 111 10.1186/s12870-015-0493-6 25935420 70. Juenger TE Sen S Bray E Stahl E Wayne T Mckay J Richards JH Exploring genetic and expression differences between physiologically extreme ecotypes: comparative genomic hybridization and gene expression studies of Kas-1 and Tsu-1 accessions of Arabidopsis thaliana Plant Cell Environ 2010 33 1268 1284 10.1111/j.1365-3040.2010.02146.x 20302603 71. Robinson AR Ukrainetz NK Kang K-Y Mansfield SD Metabolite profiling of Douglas-fir (Pseudotsuga menziesii) field trials reveals strong environmental and weak genetic variation New Phytol 2007 174 762 773 10.1111/j.1469-8137.2007.02046.x 17504460 72. Jarillo JA Gabrys H Capel J Alonso JM Ecker JR Cashmore AR Phototropin-related NPL1 controls chloroplast relocation induced by blue light Nature 2001 410 952 954 10.1038/35073622 11309623 73. Zhang J Marshall J Variation in carbon isotope discrimination and photosynthetic gas exchange among populations of Pseudotsuga menziesii and Pinus ponderosa in different environments Funct Ecol 1995 9 402 412 10.2307/2390003 74. Aitken SN Kavanagh KL Yoder BJ Genetic variation in seedling water-use efficiency as estimated by carbon isotope ratios and its relationship to sapling growth in Douglas-fir Genet 1995 2 199 206 75. Zhang J Marshall JD Jaquish BC Genetic differentiation in carbon isotope discrimination and gas exchange in Pseudotsuga menziesii Oecologia 1993 93 80 87 10.1007/BF00321195 76. Lowry DB Logan TL Santuari L Hardtke CS Richards JH DeRose-Wilson LJ McKay JK Sen S Juenger TE Expression quantitative trait locus mapping across water availability environments reveals contrasting associations with genomic features in Arabidopsis Plant Cell 2013 25 3266 79 10.1105/tpc.113.115352 24045022 77. Müller T Freund F Wildhagen H Schmid KJ Targeted re-sequencing of five Douglas-fir provenances reveals population structure and putative target genes of positive selection Tree Genet Genomes 2015 11 1 17 10.1007/s11295-014-0804-3 78. Neophytou C Weisser A-M Landwehr D Šeho M Kohnle U Ensminger I Wildhagen H Assessing the relationship between height growth and molecular genetic variation in Douglas-fir (Pseudotsuga menziesii ) provenances Eur J Forest Res 2016 79. Hermann RK Lavender DP Douglas-fir planted forests New For 1999 17 53 70 10.1023/A:1006581028080 80. Kohnle U Hein S Sorensen FC Weiskittel AR Effects of seed source origin on bark thickness of Douglas-fir ( Pseudotsuga menziesii ) growing in southwestern Germany Can J For Res 2012 42 382 399 10.1139/x11-191 81. Chang S Puryear J Cairney J A simple and efficient method for isolating RNA from pine trees Plant Mol Biol Report 1993 11 113 116 10.1007/BF02670468 82. Hess M Wildhagen H Ensminger I Suitability of Illumina deep mRNA sequencing for reliable gene expression profiling in a non-model conifer species (Pseudotsuga menziesii) Tree Genet Genomes 2013 9 1513 1527 10.1007/s11295-013-0656-2 83. Li W Godzik A Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences Bioinformatics 2006 22 1658 1659 10.1093/bioinformatics/btl158 16731699 84. Langmead B Salzberg SL Fast gapped-read alignment with Bowtie 2 Nat Methods 2012 9 357 9 10.1038/nmeth.1923 22388286 85. Anders S, Pyl PT, Huber W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics. 2014;btu638. 86. Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Harris MA Hill DP Issel-Tarver L Kasarskis A Lewis S Matese JC Richardson JE Ringwald M Rubin GM Sherlock G Gene ontology: tool for the unification of biology. The Gene Ontology Consortium Nat Genet 2000 25 25 9 10.1038/75556 10802651 87. Götz S García-Gómez JM Terol J Williams TD Nagaraj SH Nueda MJ Robles M Talón M Dopazo J Conesa A High-throughput functional annotation and data mining with the Blast2GO suite Nucleic Acids Res 2008 36 3420 35 10.1093/nar/gkn176 18445632 88. Wegrzyn JL Liechty JD Stevens KA Wu L-S Loopstra CA Vasquez-Gross HA Dougherty WM Lin BY Zieve JJ Martinez-Garcia PJ Holt C Yandell M Zimin AV Yorke JA Crepeau MW Puiu D Salzberg SL De Jong PJ Mockaitis K Main D Langley CH Neale DB Unique Features of the Loblolly Pine (Pinus taeda L.) Megagenome Revealed Through Sequence Annotation Genetics 2014 196 891 909 10.1534/genetics.113.159996 24653211 89. Proost S Van Bel M Sterck L Billiau K Van Parys T Van de Peer Y Vandepoele K PLAZA: a comparative genomics resource to study gene and genome evolution in plants Plant Cell 2009 21 3718 31 10.1105/tpc.109.071506 20040540 90. Ye Y Choi J-H Tang H RAPSearch: a fast protein similarity search tool for short reads BMC Bioinformatics 2011 12 159 10.1186/1471-2105-12-159 21575167 91. Van Bel M Proost S Van Neste C Deforce D Van de Peer Y Vandepoele K TRAPID: an efficient online tool for the functional and comparative analysis of de novo RNA-Seq transcriptomes Genome Biol 2013 14 R134 10.1186/gb-2013-14-12-r134 24330842 92. 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 R80 10.1186/gb-2004-5-10-r80 15461798 93. Anders S Huber W Differential expression analysis for sequence count data Genome Biol 2010 11 R106 10.1186/gb-2010-11-10-r106 20979621 94. Dillies M-A, Rau A, Aubert J, Hennequet-Antier C, Jeanmougin M, Servant N, Keime C, Marot G, Castel D, Estelle J, Guernec G, Jagla B, Jouneau L, Laloë D, Le Gall C, Schaëffer B, Le Crom S, Guedj M, Jaffrézic F. A comprehensive evaluation of normalization methods for Illumina high-throughput RNA sequencing data analysis. Brief Bioinform. 2012. 95. Dillies MA, Rau A, Aubert J, Hennequet-Antier C, Jeanmougin M, Servant N, Keime C, Marot G, Castel D, Estelle J, Guernec G. A comprehensive evaluation of normalization methods for Illumina high-throughput RNA sequencing data analysis. Brief Bioinform. 2013;14(6):671-83. 96. Bates D Maechler M Bolker BM Walker S lme4: Linear Mixed-Effects Models Using Eigen and S4 2014 97. Keitel C Matzarakis A Rennenberg H Gessler A Carbon isotopic composition and oxygen isotopic enrichment in phloem and total leaf organic matter of European beech (Fagus sylvatica L.) along a climate gradient Plant Cell Environ 2006 29 1492 1507 10.1111/j.1365-3040.2006.01520.x 16898013 98. Nakagawa S Schielzeth H A general and simple method for obtaining R 2 from generalized linear mixed-effects models Methods Ecol Evol 2013 4 133 142 10.1111/j.2041-210x.2012.00261.x 99. Alexa A Rahnenführer J Lengauer T Improved scoring of functional groups from gene expression data by decorrelating GO graph structure Bioinformatics 2006 22 1600 1607 10.1093/bioinformatics/btl140 16606683 100. Huang DW Sherman BT Tan Q Collins JR Alvord WG Roayaei J Stephens R Baseler MW Lane HC Lempicki RA The DAVID Gene Functional Classification Tool: a novel biological module-centric algorithm to functionally analyze large gene lists Genome Biol 2007 8 R183 10.1186/gb-2007-8-9-r183 17784955 101. Mistry M Pavlidis P Gene Ontology term overlap as a measure of gene functional similarity BMC Bioinformatics 2008 9 327 10.1186/1471-2105-9-327 18680592 102. Kenward MG Roger JH Small sample inference for fixed effects from restricted maximum likelihood Biometrics 1997 53 3 983 997 10.2307/2533558 9333350 103. Supek F Bošnjak M Škunca N Šmuc T REVIGO summarizes and visualizes long lists of gene ontology terms PLoS One 2011 6 e21800 10.1371/journal.pone.0021800 21789182

PMC005xxxxxx/PMC5002201.txt

==== Front BMC Vet ResBMC Vet. ResBMC Veterinary Research1746-6148BioMed Central London 79510.1186/s12917-016-0795-4Research ArticleNew insight into dolphin morbillivirus phylogeny and epidemiology in the northeast Atlantic: opportunistic study in cetaceans stranded along the Portuguese and Galician coasts Bento Maria Carolina Rocha de Medeiros (+351) 21 365 2800mcarolinabento@fmv.ulisboa.pt 1Eira Catarina Isabel Costa Simões catarina.eira@ua.pt 23Vingada José Vitor jvv@bio.uminho.pt 34Marçalo Ana Luisa amarcalo@gmail.com 23Ferreira Marisa Cláudia Teixeira mctferreira@socpvs.org 35Fernandez Alfredo Lopez cemma@arrakis.es 26Tavares Luís Manuel Morgado ltavares@fmv.ulisboa.pt 1Duarte Ana Isabel Simões Pereira anaduarte@fmv.ulisboa.pt 11 Centre for Interdisciplinary Research in Animal Health, Faculty of Veterinary Medicine, University of Lisbon, 1300-477 Lisbon, Portugal 2 Department of Biology and CESAM, University of Aveiro, 3810-193 Aveiro, Portugal 3 Portuguese Wildlife Society, Department of Biology, Minho University, 4710-057 Braga, Portugal 4 Department of Biology and CESAM, Minho University, 4710-057 Braga, Portugal 5 Department of Biology and CBMA, Minho University, 4710-057 Braga, Portugal 6 Coordinadora para o Estudo dos Mamíferos Mariños, 36380 Gondomar, Pontevedra Spain 26 8 2016 26 8 2016 2016 12 1 17620 11 2015 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 Screening Atlantic cetacean populations for Cetacean Morbillivirus (CeMV) is essential to understand the epidemiology of the disease. In Europe, Portugal and Spain have the highest cetacean stranding rates, mostly due to the vast extension of coastline. Morbillivirus infection has been associated with high morbidity and mortality in cetaceans, especially in outbreaks reported in the Mediterranean Sea. However, scarce information is available regarding this disease in cetaceans from the North-East Atlantic populations. The presence of CeMV genomic RNA was investigated by reverse transcription-quantitative PCR in samples from 279 specimens stranded along the Portuguese and Galician coastlines collected between 2004 and 2015. Results A total of sixteen animals (n = 16/279, 5.7 %) were positive. The highest prevalence of DMV was registered in striped dolphins (Stenella coeruleoalba) (n = 14/69; 20.3 %), slightly higher in those collected in Galicia (n = 8/33; 24.2 %) than in Portugal (n = 6/36; 16.7 %). Conclusions Phylogenetic analysis revealed that, despite the low genetic distances between samples, the high posterior probability (PP) values obtained strongly support the separation of the Portuguese and Galician sequences in an independent branch, separately from samples from the Mediterranean and the Canary Islands. Furthermore, evidence suggests an endemic rather than an epidemic situation in the striped dolphin populations from Portugal and Galicia, since no outbreaks have been detected and positive samples have been detected annually since 2007, indicating that this virus is actively circulating in these populations and reaching prevalence values as high as 24 % among the Galician samples tested. Electronic supplementary material The online version of this article (doi:10.1186/s12917-016-0795-4) contains supplementary material, which is available to authorized users. Keywords Cetacean morbillivirusDolphin morbillivirusStriped dolphinsEastern Atlantichttp://dx.doi.org/10.13039/501100001871Fundação para a Ciência e a TecnologiaCetSenti RECI/AAG-GLO/0470/2012UID/AMB/50017/2013SFRH/BD/30240/2006SFRH/BPD/82407/2011SFRH/BPD/64889/2009Eira Catarina Isabel Costa Simões Marçalo Ana Luisa Ferreira Marisa Cláudia Teixeira Fernandez Alfredo Lopez Life+ MarproLife09 NAT/PT/000038http://dx.doi.org/10.13039/501100007048EEA GrantsPT0039issue-copyright-statement© The Author(s) 2016 ==== Body Background Morbillivirus infection affects mainly the upper respiratory tract, central nervous system and the immune system [1, 2] and has been identified as a cause of death and stranding in marine mammals [3]. In odontecetes, infection has been associated with high mortality rates occurring during disease outbreaks in different parts of the world [4]. The mortality rate in striped dolphins from the Mediterranean Sea in the beginning of the nineties was the highest recorded so far [2, 5, 6]. Further studies are needed to deepen the knowledge about this disease. An integrated approach taking into consideration epidemiological and environmental parameters should provide a better picture of the ecology and evolution of Cetacean Morbillivirus (CeMV) in free-ranging cetaceans [2]. CeMV includes three well characterized viral strains [7]: porpoise morbillivirus (PMV), dolphin morbillivirus (DMV) and pilot whale morbillivirus (PWMV); three novel cetacean morbillivirus strains were recently reported [8–10], adding to the genetic diversity of these viruses. Morbilliviruses affecting cetaceans have been described in the last decades [2] after the initial detection of viral antigens in these species in the late eighties. The first evidence of morbillivirus infection in cetaceans occurred in 1988 during a PMV outbreak, when the viral antigen was detected in harbour porpoises (Phocoena phocoena) stranded in Ireland [11]. In the early nineties, dolphin morbillivirus (DMV) was isolated from striped dolphins from the Mediterranean [12, 13]; in 2000 PWMV was first described in a long-finned pilot whale (Globicephala melas) from the US coast [14] and later, in 2011, from a short-finned pilot whale (Globicephala macrorynchus) in the Canary Islands [15]. Due to the virus pathogenic impact on cetacean populations, further information about morbillivirus infection in cetaceans worldwide is relevant to understand its epidemiology in these animals. Studying infectious diseases in these species is important, especially considering that additional non-infectious aggressions, mainly due to human activities, render these populations even more susceptible to disease. An annual average of 200 stranded cetaceans were registered between 2010 and 2012, considering the Algarve and the Northern region of the Portuguese continental coast [16] and fisheries bycatch was identified as the most significant cause of death. To this date, no molecular data was published on morbillivirus infection in animals stranded in Portugal or northern Spain. In 2014, dolphin morbillivirus infection was reported in a retrospective study affecting striped dolphins and a common dolphin from the Canary Islands [17], causing non-suppurative meningoencephalitis. Also, a fatal systemic morbillivirus infection was detected in a bottlenose dolphin stranded in 2005 in the Canary Islands [18]. It was suggested that DMV was not endemic in harbour porpoises and common dolphins (Delphinus delphis) from the NE Atlantic (British Isles) in the period 1996–1999 [19], as low antibodies titres were detected in animals from Spain and the North Sea. DMV infection apparently did not persist as an endemic infection in Mediterranean striped dolphins after the 1990–92 epidemic [19]. Both epidemics in the Mediterranean Sea (1990–92 and 2006–07) started near the Gibraltar Strait [20] and it has been suggested that DMV-infected cetaceans may have entered the Strait of Gibraltar and infected striped dolphins, the most common cetacean at the time [7, 21]. Pilot whales had been already proposed as reservoirs in 1995 [22, 23]. Later, in 2006 several long-finned pilot whales were found stranded along the coast of the Alborean Sea,and morbillivirus infection was detected [24]. In this epidemic, deaths were first detected close to the Gibraltar Strait and spread further into the Mediterranean Sea. Recently described sequences found in striped dolphins from the Canary Islands show high identity with sequences from the Mediterranean outbreaks, indicating the possible circulation of viruses between the Atlantic and the Mediterranean [17]. The role of other cetacean species as reservoirs needs to be further assessed. The objective of the present study was to clarify not only the prevalence of DMV in cetacean populations from the eastern Atlantic, but also to investigate the relationship between the dolphin morbillivirus circulating in the eastern Atlantic and elsewhere in the world, especially in the Mediterranean. Methods Sample collection Stranded cetaceans were collected by the Sociedade Portuguesa de Vida Selvagem (SPVS) in Northern Portugal and the Algarve within the Marine Animal Stranding Network, managed by the Instituto para a Conservação da Natureza e Florestas (ICNF) and in Galicia by the Coordinadora para o Estudo dos Mamiferos Mariños (CEMMA). Permission was issued by the National Authority (ICNF) to SPVS technicians to collect wildlife samples within the national territory according to laws n.140/99, n.49/2005, n.156-A/2013, and n.316/89. Also, SPVS is a registered CITES scientific research institution (code PT009). CEMMA holds a permit from the Conselleria de Medio Ambiente, Territorio e Infraestruturas de Xunta de Galicia (Spain) to collect and maintain cetacean samples according to law 42/2007 and law 9/2001. The animals were assigned a decomposition code (1 to 5) according to already established protocols [25]. Animals with a score ranging from to 1 to 3 (fresh to moderate decomposition) were surveyed in the present study. During necropsy, tissue samples were collected from 279 cetaceans: brain, lung, pulmonary lymph node, mesenteric lymph node, spleen, kidney and liver, whenever possible. For animals from Galicia the only available sample was the lung. Samples collected in Portugal were stored in vials with RNAlater® at −20 °C and samples collected in Galicia were frozen at −20 °C. All samples were kept in the marine animals’ tissue banks (MATBs) of SPVS and CEMMA. Samples from different species were collected between 2004 and 2015: common dolphins (DD), striped dolphins (SC), bottlenose dolphins (Tursiops truncatus; TT), long-finned pilot whales (Globicephala melas; GM), Pigmy Sperm Whale (Kogia breviceps; KB), True’s Beaked Whale (Mesoplodon mirus; MMi) and Fin whale (Balaenoptera physalus; BP) (Table 1). Samples were identified with a code composed by the species identification (e.g., DD, SC, TT), a number attributed to each stranding, and the year of stranding. From cetaceans stranded in the Portuguese coastline 91 animals from 2011, 56 from 2012, 33 from 2013, 49 from 2014 and 7 from 2015 were tested. From Galicia, a total of 33 lung samples from striped dolphins were tested. Available tissue samples from 10 animals stranded in Portugal from previous years were also included in this study (6 striped dolphins and 4 pilot whales from 2004 to 2009).Table 1 Number of stranded cetaceans tested for DMV per year 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Total Portugal Common dolphin (DD) – – – – – – – 84 45 29 29 6 193 Striped dolphin (SC) 1 – 1 4 – – – 5 6 3 16 – 36 Pilot whale (GM) – – – – 2 2 – – 1 – – – 5 Bottlenose dolphin (TT) – – – – – – – 1 2 1 3 – 7 True’s beaked whale (MMi) – – – – – – – 1 – – – – 1 Pigmy sperm whale (KB) – – – – – – – – 1 – 1 – 2 Fin whale (BP) – – – – – – – – 1 – – 1 2 Galicia Striped dolphin (SC) 2 1 1 2 3 6 4 2 3 5 4 33 Total RNA extraction Total RNA was extracted from a pool of tissue homogenates using RNeasy mini kit (Qiagen, GmbH, Germany), according to the manufacturer’s instructions. The pool included, whenever possible: lung, brain, pulmonary lymph node and mesenteric lymph node. Total RNA quantification and purity was determined using a Nanodrop 2000C spectrophotometer (ThermoScientific, USA) and stored at −80 °C until used. Detection of dolphin morbillivirus genomic RNA by reverse transcription-quantitative PCR (RT-qPCR) The detection of viral RNA for the DMV strain of CeMV was performed by RT-qPCR in a StepOnePlus thermocycler (Applied Biosystems), using primers (Stabvida genomics lab, Portugal) and probe (Eurogentec), targeting the N gene of DMV, as previously described [26] (Table 2). A previously detected positive sample for DMV was used as a positive control of the PCR reaction. Negative reaction controls were always included.Table 2 Primers and probe set used in RT-qPCR assays 5′ Fluorophore 3′ Quencher Sequences (5′–3′) Amplicon size Annealing (°C) DMV-N-FP – – TGCCAGTACTCCAGGGAACATCCTTC 173 60 [26] DMV-N-RP – – TTGGGTCGTCAGTGTTGTCGGACCGTT 173 60 DMV-N-probe Cy3 BHQ1 A + CA + CCAAA + AGGGA + CA – 60 One step RT-qPCR assays were performed using 100 ng of the template RNA, in a total reaction volume of 20 μL containing: 10 μL of 1-step qPCR-ROX Mix (2×); 1 μL of RT enhancer; 0,2 μL of Verso Enzyme Mix (Verso 1-Step qRT-PCR ROX kit, ThermoScientific®); 0.4 μM of each primer and 0.25 μM of probe. For positive samples, total RNA was extracted individually for each of the available organs and the infection was evaluated individually in the different organs. The amplified DMV fragment was cloned into a plasmid vector (Pgem Teasy – Promega) and serial tenfold dilutions of the recombinant plasmid DNA were used to construct the standard curve (Fig. 1). The results showed a high correlation (R2 = 0.997) with a calculated efficiency of 81 %. The primers and probe could detect viral RNA copies down to 102, and the limit of detection was 224 copies.Fig. 1 Standard curve and equation for the determination of the efficiency of the RT-qPCR for the molecular detection of DMV. The N gene fragment obtained in the RT-qPCR reaction was cloned into a plasmid vector (Pgem Teasy – Promega) and serial tenfold dilutions of the recombinant plasmid DNA were amplified by qPCR in duplicate reactions and used to construct the standard curve. Y axis represents the mean CT values obtained from the duplicates and X axis represents the LOG10 of calculated copy numbers (ranging from 2.24E + 08 to 224 copies). Calculated efficiency of 81 % was determined using the formula: Efficiency = 10(-1/slope) x (-1). Results showed a high correlation (R2 = 0.997) Conventional PCR for amplification of DMV genes Additional sequences were amplified from the positive samples by conventional reverse transcription-PCR (RT-PCR) using previously described primers (Table 3) purchased from Stabvida genomics lab (Portugal). Primers were used in different combinations, targeting different genomic regions (Fig. 2).Table 3 Primers used in conventional PCR assays Primer Target gene Sequence (5′–3′) (sense) Tm Genome position Reference CeMV-He1 H CRTTGATACTYGTGGGTGTG (+) 59 7194–7213 [15] CeMV-He2 H TGTTAACTTCTGGGGCATCC (−) 59 7407–7426 DMVFu-F F GGCACCATAATTAGCCAGGA (+) 51 6483–6502 DMVFu-R F GCCCAGATTTGTGCCTACAT (−) 51 6655–6674 DMV-C P ATGTTTATGATCACAGCGGT (+) 51 2132–2151 [35] DMV-P2 P ATTGGGTTGCACCACTTGTC (−) 51 2541–2560 NgeneF N CCHAGRATYGCTGAAATGATHTGTGA (+) 48 849–874 [14] NgeneR N AACTTGTTCTGRATWGAGTTYTC (−) 48 1056–1078 Fig. 2 Schematic representation of the primers used to amplifiy different genomic regions by conventional RT-PCR. Schematic representation of the DMV genome and location of the primers used in the conventional RT-PCR reactions, targeting: the N gene (NgeneF and NgeneR) to amplify a fragment of 229 basepairs (bps); the P gene (DMV-C and DMV-P2) for a fragment of 428 bps; the F gene (DMVFu-F and DMVFu-R) (191 bps) and for a fragment of 232 bps from the H gene (CeMVHe1 and CeMVHe2) The obtained amplicons were used to perform a phylogenetic analysis of the DMV sequences, along with sequences retrieved from NCBI for the same genes. L and M genes were not targeted in the conventional RT-PCR since very few sequences were available at the NCBI database. The amplicons were directly sequenced by Sanger sequencing at Stabvida, Portugal and the specificity of the nucleotide sequences was compared by Blast analysis http://blast.ncbi.nlm.nih.gov/Blast.cgi with CeMV sequences available in the GenBank. Phylogenetic analysis The nucleotide sequences of the Portuguese and Galician sequence datasets available in the GenBank (National Center for Biotechnology Information) repository,with the following accession numbers KP835987; KP835991; KP835995; KP835999; KP836003; KP835986; KP835990; KP835994; KP835997; KP836002; KP836006; KP835985; KP835989; KP835993; KP835996; KP836001; KP836005; KP835984; KP835988; KP835992; KP835998; KP836000; KP836004; KP835983; KT878649; KT878650; KT878651; KT878652; KT878653; KT878654; KT878655; KT878656; KT878657; KT878658; KT878659; KT878660; KT878661, were compared with the available CeMV sequences and outgroup taxa (Canine Distemper Virus [CDV], Phocine Distemper Virus [PDV] and Measles Virus [MV]), retrieved from GenBank (Table 4), according to their primary structure similarity using the multiple alignment ClustalW program [27].Table 4 Accession number for GenBank sequences used to the phylogenetic analysis and corresponding description Complete genomes and common sequences NC_0014981 Measles Virus AY649446 Canine Distemper Virus KC802221 Phocine Distemper Virus AJ608288 Dolphin Morbillivirus complete genome HQ829973 Striped dolphin 2007 SP (Med) HQ829972 Long-finned pilot whale 2007 SP (Med) Gene N X84739 Porpoise 1988 IRL AF200818 Long-finned pilot whale 1999 USA FJ842380 Short-finned pilot whale 1996 SP (Can Isl) Gene P KF695110 Bottlenose dolphin 2005 SP (Can Isl) JX195718 Longman’s beaked whale 2010 USA EU039963 Long-finned pilot whale 2007 SP KF650727 Porpoise 1990 NL EF451565 White-beaked dolphin 2007 GM AF200817 Long-finned pilot whale 1999 USA AF333347 Pigmy sperm whale 2001 TW KJ139451 Striped dolphin 2002 SP (Can Isl) KJ139452 Striped dolphin 2007 SP (Can Isl) JN210891 Striped dolphin 2011 SP (Med) KF711855 Guiana dolphin 2010 BR KJ139454 Striped dolphin 2011 SP (Can Isl) KC572861 Striped dolphin 2012 SP (Med) KJ139453 Striped dolphin 2009 SP (Can Isl) KR337460 Fin whale 2013 IT KR704575 Longman’s beaked whale 2013 NC KC888945 White-beaked dolphin 2011 NL Gene F AJ224704 Striped dolphin 90’s SP Z30086 DMV 1994 FJ842382 Short-finned pilot whale 1996 SP Gene H FJ648457 Porpoise MV 1988 IRL AJ224705 Striped dolphin 90’s SP Z36978 DMV 1994 FJ842382 Short-finned pilot whale 1996 SP (Can Isl) SP Spain, (MED) Mediterranean, (Can Isl) Canary Islands, IRL Ireland, USA United States of America, NL Netherlands, GM Germany, TW Taiwan, BR Brazil, IT Italy, NC New Caledonia Six sets of alignments were considered for the phylogenetic analysis: nucleotide sequence alignments for genes N, P, F and H, composed by sequences of 218, 342, 449 and 316 base pairs, respectively; concatenated sequence of amino acids (540 aa) and nucleotides (1446 bps). Due to heterogeneity of the available DMV sequences it was not feasible to maintain the same set of DMV sequences in the alignment for each gene. In the concatenated alignment only the sequences with all partial genomic regions were included. The multiple sequence alignments were manually corrected with Jalview, Version 2.0.1 [28] removing long internal gaps and unmatched ends to maximize genetic similarities and phylogenetic trees were inferred by Bayesian methods (MrBayes v.3.2.1) [29, 30]. For the Bayesian analysis a Markov chain Monte Carlo (mcmc) simulation technique was carried out to approximate the PP of trees [30]. The evolutionary GTR (nucleotides) and LG (amino acids) models were selected with gamma-distributed rate variation across sites and a proportion of invariable sites (rates = invgamma). The analysis was initiated using a random tree from the dataset with four chains running simultaneously for 20 × 106 generations, sampling every 100 generations. The first 25 % trees were discarded and a majority rule consensus tree was generated from the remaining trees. The graphical representation and edition of the phylogenetic tree were performed with FigTree v1.3.1. Only support values equal or greater than 0.70 of PP are shown in the trees. Statistical analysis Chi-square test of association was performed to assess if the difference in prevalence was statistically significant between different species (DD and SC) and between animals from different origins (Portugal and Galicia). For this analysis an online website for statistical computation was used url: http://vassarstats.net/. A confidence interval (CI) of 95 % (for a p value ≤0.05) was considered for all the statistical analysis. Results A total of 16 DMV positive cetaceans were identified by RT-qPCR, representing a prevalence of 5.7 % (IC95 %: 3.42;9.32). With respect to the Portuguese coastline, 8 positive animals were detected, including 6 striped dolphins (SC) and 2 common dolphins (DD) [SC/15/2007, SC/257/2011, SC/221/2012, DD/302/2012, SC/11/2013, DD/191/2013, SC/193/2014 and SC/290/2014]. In Galicia, 8 positive striped dolphins were detected [SC/21/2007, SC/24/2008, SC/31/2009, SC/42/2010, SC/49/2011, SC/51/2012, SC/53/2012 and SC/55/2012]. Among all cetacean species, striped dolphins (n = 69) revealed a significantly higher DMV prevalence reaching 20.3 % (IC95 %: 11.92; 32.02), whereas common dolphins (n = 139) recorded a prevalence of 1.0 % (IC95 %: 0.18; 4.09) (P value 0.00). Positive striped dolphins were detected every year (from 2007 to 2014) while positive common dolphins were only detected in 2012 and 2013. The DMV prevalence in striped dolphins stranded in Galicia was 24.2 % (IC95 %: 11.74; 42.63) whereas in Portugal the DMV prevalence was 16.7 % (IC95 %: 6.97; 33.47). From the positive animals stranded along the Portuguese coastline, each organ included in the tissue pool was tested individually for viral RNA. Two animals tested positive in all available organs; four were positive for viral RNA only in the brain and one animal tested positive in the lung, and in the pulmonary and mesenteric lymph node (Table 5). Lung was the only available sample to test in samples from Galicia.Table 5 Mapping of DMV infection in the available organs in Portuguese samples Tested organs are shown in grey and positive organs are marked with an (X) * only lung samples were available For samples SC/15/2007, SC/257/2011, SC/53/2012, SC/55/2012, DD/302/2012, SC/290/2014, SC/11/2013, SC/31/2009, SC/51/2012, SC/21/2007, SC/221/2012 and DD/191/2013 longer genomic regions were amplified by one step RT-conventional PCR with the primers described previously (Table 3). For samples SC/24/2008, SC/42/2010, SC/49/2011 and SC/193/2014 no fragments were amplified by conventional RT-PCR. For the nucleotide sequences of each genomic region, phylogenetic trees were inferred by Bayesian methods. All trees exhibited a similar sequence topology, supported by robust PP values, regardless of the total number of sequences in each tree. In the tree of the concatenated nucleotide sequences (Fig. 3) the CeMV sequences were distributed in three main branches supported by high PP values. Portuguese and Galician samples from 2011, 2012 and 2013 were included in one branch; sequences from the Mediterranean from 2007, early nineties (AJ608288) and the Portuguese sequence SC/15/2007 in another branch. The only PWMV included in this tree is isolated in a third branch. The tree of the amino acid concatenated sequences presented a similar pattern (Fig. 4), although with a rearrangement within the older sequences branch ([SC/15/2007, GM/2007, Med, SC/2007/Med]; [SC/1990/Med]).Fig. 3 Phylogenetic tree for the concatenated nucleotide sequences. Phylogenetic tree generated with concatenated nucleotide sequences alignment, inferred by Bayesian methods. Sequences for the outgroup taxa were retrieved from NCBI for Pilot Whale Morbillivirus (PWMV [accession numbers FJ842380 and FJ842382]); Phocine Distemper Virus (PDV [accession number KC802221]); Canine Distemper Virus (CDV [accession number AY649446]) and Measles Virus (MV [accession number NC001498]). Three DMV sequences were also retrieved from NCBI: one isolate from a pilot whale [HQ829972] and one from a striped dolphin [HQ829973], both from 2007; one isolate from 1990, also from a striped dolphin [AJ608288]. Sequences obtained for animals Sc/257/2011 (KP835983; KP835984; KP835985; KP835986), Dd/191/2013 (KP836003; KP836004; KP836005; KP836006), Sc/53/2012 (KP835991; KP835992; KP835993; KP835994), Dd/302/2012 (KP835999; KP836000; KP836001; KP836002) and Sc/55/2012 (KP835987; KP835988; KP835989; KP835990), Sc/15/2007 (KP835995; KP835996; KP835997; KP835998) were also included in this tree Fig. 4 Phylogenetic tree for the concatenated amino acid sequences. Phylogenetic tree generated with the concatenated amino acid sequences alignment, inferred by Bayesian methods. Sequences for the outgroup taxa were retrieved from NCBI for Pilot Whale Morbillivirus (PWMV [accession numbers FJ842380 and FJ842382]); Phocine Distemper Virus (PDV [accession number KC802221]); Canine Distemper Virus (CDV [accession number AY649446]) and Measles Virus (MV [accession number NC001498]). Three DMV sequences were also retrieved from NCBI: one isolate from a pilot whale [HQ829972] and one from a striped dolphin [HQ829973], both from 2007; one isolate from 1990, also from a striped dolphin [AJ608288]. Sequences obtained for animals Sc/257/2011 (KP835983; KP835984; KP835985; KP835986), Dd/191/2013 (KP836003; KP836004; KP836005; KP836006), Sc/53/2012 (KP835991; KP835992; KP835993; KP835994), Dd/302/2012 (KP835999; KP836000; KP836001; KP836002) and Sc/55/2012 (KP835987; KP835988; KP835989; KP835990), Sc/15/2007 (KP835995; KP835996; KP835997; KP835998) were also included in this tree In the nucleotide tree for the F gene (Additional file 1) additional available sequences from the early nineties were included. Samples collected in the Atlantic during the 2011-2013 period clustered in the same branch; samples from the nineties clustered in a separate branch, and samples from the Mediterranean from 2007 clustered in a third branch, together with the sample SC/15/2007, similarly to the distribution of the concatenated trees. The PWMV was included in a unique branch. All branches were supported with a high PP values. For the H gene nucleotide tree (Additional file 2), a higher number of sequences were included. A set of sequences (9) from Portugal and Galicia ranging from 2009 to 2014 clustered in the same branch, supported by a PP value of 0.98. The SC/15/2007 sequence still clustered with Mediterranean samples from 2007 and samples from the early nineties were grouped in a separate branch. The new sequence for PMV included in this tree, branches out from the DMV samples, similarly to the PWMV sequence (PP value of 0.9). The nucleotide tree for the P gene (Fig. 5) contained the higher number of sequences (35). One sequence from a guiana dolphin (Sotalia guianensis) collected in 2010 in Brazil appeared to be a distinct strain from the already characterized strains of CeMV (PMV, PWMV and DMV. The two PWMV samples clustered in the same branch and the only PMV included in the tree was isolated from all the other sequences. All these strains were supported by high PP values. The DMV sequences included in this tree were all similar, including sequences from distinct geographic origins, such as Germany, Taiwan or the Mediterranean. Two sequences obtained from white-beaked dolphins in Germany and the Netherlands in different years (2007 and 2011 respectively) clustered together with a PP value of 0.99. Sequence AJ608288 from a striped dolphin collected in 1990 in the Mediterranean and sequence AF333347 from a pigmy sperm whale from Taiwan collected in 2001 also clustered together (PP 0.79). Samples from the Canary Islands collected in 2005, 2007 and 2009 clustered with samples from the Mediterranean (2007 and 2011), one sample from New Caledonia and one sample from Portugal (SC/15/2007). The remaining Portuguese and Galician samples clustered in the same clade with two samples with a different origin (KJ139454, Canary Islands and KC572861, Mediterranean).Fig. 5 Phylogenetic tree for the P gene nucleotidic sequences. Phylogenetic tree generated with the aligned sequences for the P gene, inferred by Bayesian methods. Sequences for the outgroup taxa were retrieved from NCBI for Pilot Whale Morbillivirus (PWMV [accession numbers FJ842380 and AF200817]); Porpoise Morbillivirus (PMV [accession number FJ650727]); Phocine Distemper Virus (PDV [accession number KC802221]); Canine Distemper Virus (CDV [accession number AY649446]) and Measles Virus (MV [accession number NC001498]). Two recently described sequences of CeMV (JX195718 and KF711855) were also included, along with: one isolate of DMV from 2002 collected in the Canary Islands (KJ139451), one collected in Taiwan in 2001 (AF333347), five sequences from 2007 collected in Spain and Germany (EF451565, HQ829972, HQ829973, EU039963, KJ139452), one sequence from the 90’s (AJ608288), one from 2005 (KF695110) and one from 2011 collected in the Mediterranean (JN210891). Sequences obtained for the P gene of Portuguese and Galician isolates were also included for animals Sc/290/2014 (KT878659), Sc/11/2013 (KT878656), Sc/31/2009 (KT878660), Dd/191/2013 (KP836003), Sc/55/2012 (KP835987), Sc/51/2012 (KT878661), Dd/302/2012 (KP835999), Sc/21/2007 (KT878657), Sc/221/2012 (KT878658), Sc/257/2011 (KP835983), Sc/53/2012 (KP835991), and Sc/15/2007 (KP835995) The N gene nucleotide tree (Additional file 3) showed a dislocation of sequences between branches. One branch included Atlantic samples from 2011 to 2013 grouped with the SC/15/2007 sequence and with sequences from the Mediterranean (1990 and 2007); Atlantic sequences also from 2012 to 2014, were grouped separately. The remaining PWMV and PMV sequences appeared as two different outgroups. Discussion In this study we surveyed 279 animals and our results indicate a higher prevalence of DMV among stranded striped dolphins (20.6 %) when compared to stranded common dolphins (1 %) from the Atlantic based populations. Similar results had been previously described in the Mediterranean during the 1990–92 and 2006–08 CeMV breakouts, when striped dolphins presented higher death and stranding rates than other species [2, 31, 32]. Several theories have been hypothesized for this higher mortality rate amongst striped dolphins in the Mediterranean: they were the most numerous species in the Mediterranean and serological studies suggested that, prior to the 2006–08 outbreak, antibody levels were low in this population rendering them more susceptible to the CeMV infection [19]; also, the fact that they are highly gregarious and tend to live in large pods could contribute to the spread of CeMV infection [33]; high polychlorinated biphenyl (PCB) levels were also detected in the affected animals, leading to the hypothesis that an impaired immune system might have facilitated the infection by CeMV; finally, genetic susceptibility as a result of inbreeding in the Mediterranean population [33], which had already been reported as relatively isolated from the Atlantic populations [34]. Prevalence among striped dolphins from Galicia was 24.2 % while prevalence in striped dolphins stranded in Portugal was 16.7 %. Although this difference was not statistically significant, it is important to highlight that prevalence among striped dolphin samples from Galicia was probably underestimated since only lung samples were tested. Samples from the Portuguese coastline allowed testing several organs and antigen was only detected in brain samples of four individuals out of the 6 positive striped dolphins. It is therefore possible that the prevalence in striped dolphins from Galicia is being strongly underestimated. Previous studies from the Atlantic based populations were performed in the western part of the Atlantic, along the USA coast, and bottlenose dolphins were the most affected cetaceans in that area. In the Canary Islands a retrospective study was published in 2014 and 6 animals were positive for CeMV (5 striped dolphins and 1 common dolphin) [17]. In this study striped dolphins seem to be the most affected species sampled from the East Atlantic. In four animals it was not possible to amplify viral genomic fragments by conventional RT-PCR. These samples recorded high CT values in the RT-qPCR, corresponding to a low target copy number (ranging from 105 to 943 copies), which would present a downside using a less sensitive conventional assay. Also, three of the four samples were collected in animals from Galicia originally stored at −20 °C, which may possibly imply RNA degradation hampering the amplification of longer genomic fragments, by conventional RT-PCR. The genetic distances between samples were low among all sequences included in the phylogenetic trees. Nonetheless, PP values were high and consistent in all trees particularly in the DNA concatenated tree, adding robustness to the phylogenetic arrangement. In the phylogenetic trees for the concatenated nucleotides the grouping of viral sequences followed a temporal arrangement, with samples collected since 2007 forming different clades. When a higher number of sequences was added to the trees (P gene tree) a phylogeographic arrangement becomes clear: all samples from Portugal and Galicia cluster together (with isolates ranging from 2007 to 2014), further away from the samples from the Mediterranean. The only exception seems to be the sequence from the animal SC/15/2007,clustering with samples from the Mediterranean, as well as with samples from the Canary Islands. Even samples from animals stranded in the south of Portugal (Algarve), such as SC/11/2013, clustered separately from samples obtained in the Mediterranean. This suggests that these populations may be relatively isolated from each other, which is supported by previous findings by other authors [34]. It is worth noticing that only one sample from a striped dolphin collected in the Canary Islands clusters closer to the Portuguese and Galician samples. All the other samples from the Canary Islands are closer to Mediterranean samples. Positive samples for DMV antigen were detected annually since 2007 to 2013, showing that the virus is circulating in cetacean populations from the Atlantic off the coast of Portugal and northern Spain and both striped dolphins and common dolphins were found to be positive to viral infection. The infection was mapped in the available organs and positive lung samples were detected without association to higher mortality or stranding rates. Further studies would be necessary to determine if these animals had an acute, sub-acute or chronic infection and if the DMV infection was the cause of death. Animals DD/191/2013, DD/302/2012, SC/21/2007, SC/51/2012, SC/53/2012, SC/11/2013 and SC/257/2011 stranded alive and were in general emaciated and with high parasite loads, suggesting a sub-acute or chronic systemic infection. Histological and immunohistochemical studies should be performed to further characterize the necropsy findings. Four animals (SC) were positive only in brain samples, which might imply the development of chronic localized encephalitis after a systemic infection. The two common dolphins positive for viral antigen (DD/302/2012 and DD/191/2013) were both alive at the time of stranding and presented high parasite loads and poor body condition. Animals DD/302/2012 and SC/221/2012 are also positive for cetacean gamma herpesvirus (unpublished observations Bento, C.) with viral antigen detected systemically. Unlike Mediterranean populations of striped dolphin [2], morbillivirus infection seems to be endemic in the population of striped dolphins from the Atlantic. This correlates to the serological survey conducted in 2011 in which 21.6 % (n = 37) of the analysed cetaceans cross reacted with Canine Distemper Virus antigen in a commercially available ELISA kit (unpublished observations Bento, C.). To date, the harbour porpoise was reported as the most affected species with morbillivirus infection in the north-eastern Atlantic, although infection is probably not endemic considering porpoises’ solitary behaviour [2]. Large populations are needed to maintain morbillivirus infections as endemic [19] and although striped dolphin abundance has increased over the last years in the Portuguese Continental coast it is still a rather small population if compared to the common dolphin population (Araújo, H. personal communication). Notwithstanding, evidence suggests an endemic situation rather than an epidemic, since no outbreaks have been detected in the striped dolphin population of the Atlantic. Moreover, positive samples have been detected annually since 2007, indicating that this virus is actively circulating in this population reaching prevalence values as high as 24 % in the Galician samples. In 1999, dolphins stranded along the Atlantic coast of Spain had low antibody titres for CeMV. Considering the results obtained in this study, further serological studies are needed to deepen the knowledge about the epidemiology of this disease in striped dolphins. Unlike striped dolphins, the prevalence of stranded common dolphins positive for viral antigen is much lower (1 %). The difference in CeMV prevalence between stranded common and striped dolphins needs to be fully assessed and further studies are needed to clarify the virus impact on cetacean populations and why do striped dolphins appear to be more susceptible to DMV infection. New approaches should be considered: viral enrichment and random amplification techniques associated with next generation sequencing could contribute to deepen the knowledge on this virus and its interaction with other pathogens. Surveys are a unique tool to provide information on viral epidemiology, especially in free-ranging cetaceans. Conclusion Our results suggest that DMV infection is endemic in striped dolphin populations of the Eastern Atlantic. Since it was first reported in cetaceans in the early nineties subtle but consistent changes in the reported viral sequences suggest that the Atlantic and the Mediterranean populations are relatively isolated from each other, as suggested by other authors. The prevalence of infection in stranded common dolphins is very low when compared to striped dolphins, and our results are in agreement with previous reports that point to a higher susceptibility of striped dolphins to CeMV. Reasons for differences in susceptibility to this viral infection in different species should be further investigated and serological surveys should also be performed to assess their protection level towards CeMV infection. Additional files Additional file 1: Phylogenetic tree for the F gene nucleotidic sequences. Phylogenetic tree generated with the aligned sequences for the F gene, inferred by Bayesian methods. Sequences for the outgroup taxa were retrieved from NCBI for Pilot Whale Morbillivirus (PWMV [accession number FJ842382]); Phocine Distemper Virus (PDV [accession number KC802221]); Canine Distemper Virus (CDV [accession number AY649446]) and Measles Virus (MV [accession number NC001498]). Three DMV sequences from the 90’s were retrieved from NCBI and included in the trees (Z30086; AJ224704 and AJ608288) along with two sequences from 2007 (HQ829972 and HQ829973). Sequences obtained for the F gene of Portuguese and Galician isolates were also included for animals Sc/257/2011 (KP835986), Dd/302/2012 (KP836002), Dd/191/2013 (KP836006), Sc/53/2012 (KP835994), Sc/55/2012 (KP835990) and Sc/15/2007 (KP835997). (TIF 5123 kb) Additional file 2: Phylogenetic tree for the H gene nucleotidic sequences. Phylogenetic tree generated with the aligned sequences for the H gene, inferred by Bayesian methods. Sequences for the outgroup taxa were retrieved from NCBI for Pilot Whale Morbillivirus (PWMV [accession number FJ842382]); Porpoise Morbillivirus (PMV [accession number FJ648457]); Phocine Distemper Virus (PDV [accession number KC802221]); Canine Distemper Virus (CDV [accession number AY649446]) and Measles Virus (MV [accession number NC001498]). Three DMV sequences from the 90’s were retrieved from NCBI and included in the trees (Z36778; AJ224705 and AJ608288) along with two sequences from 2007 (HQ829972 and HQ829973). Sequences obtained for the H gene of Portuguese and Galician isolates were also included for animals Sc/31/2009 (KT878652), Sc/290/2014 (KT878651), Sc/221/2012 (KT878650), Sc/55/2012 (KP835989), Sc/11/2013 (KT878649), Dd/302/2012 (KP836001), Dd/191/2013 (KP836005), Sc/257/2011 (KP835985), Sc/53/2012 (KP835993), and Sc/15/2007 (KP835996). (TIF 3936 kb) Additional file: 3 Phylogenetic tree for the N gene nucleotidic sequences. Phylogenetic tree generated with the aligned sequences for the N gene, inferred by Bayesian methods. Sequences for the outgroup taxa were retrieved from NCBI for Pilot Whale Morbillivirus (PWMV [accession number FJ842380]); Porpoise Morbillivirus (PMV [accession number X84739]); Phocine Distemper Virus (PDV [accession number KC802221]); Canine Distemper Virus (CDV [accession number AY649446]) and Measles Virus (MV [accession number NC001498]). Two sequences from 2007 (HQ829972 and HQ829973) and one from the 90’s (AJ608288) were also included in this tree. Sequences obtained for the N gene of Portuguese and Galician isolates were also included for animals Sc/15/2007 (KP835998), Dd/302/2012 (KP836000), Sc/257/2011 (KP835984), Dd/191/2013 (KP836004), Sc/53/2012 (KP835992), Sc/55/2012 (KP835988), Sc/11/2013 (KT878653), Sc/290/2014 (KT878655) and Sc/221/2012 (KT878654). (TIF 4079 kb) Abbreviations BPBalaenoptera acutorostrata CDVCanine distemper virus CEMMACoordinadora para o estudo dos mamiferos mariños CeMVCetacean morbillivirus DDDelphinus delphis DMVDolphin morbillivirus GMGlobicephala melas ICNFInstituto de conservação da natureza e florestas KBKogia breviceps MATBsMarine animals tissue banks MMiMesoplodon mirus MVMealses virus PDVPhocine distemper virus PDV-1Phocine distemper virus -1 PMVPorpoise morbillivirus PPPosterior probability PWMVPilot whale morbillivirus RT-PCRReverse transcription PCR RT-qPCRReverse transcription quantitative PCR SCStenella coeruleoalba SPVSSociedade portuguesa de vida selvagem TTTursiops truncatus Acknowledgements The authors thank SPVS and CEMMA for assistance with data and sample collection. The authors further thank to CIISA at FMV-Ulisboa, where the laboratorial work was developed. We also recognize Margarida Duarte (PhD) for the revision of the manuscript from Instituto Nacional de Investigação Agrária e Veterinária, I.P., Laboratório de Virologia and Isabel Marques, from the Bioinformatic Unit of Instituto Gulbenkian de Ciência, Portugal, for the phylogeny analysis. Funding Sample collection was partially supported by SafeSea (EEAGrants PT0039), MarPro (Life09 NAT/PT/000038 co-funded by the EU) and CetSenti RECI/AAG-GLO/0470/2012 (FCOMP-01-0124-FEDER-027472), FCT/MCTES (PIDDAC) and FEDER - COMPETE (POFC). C. Bento was supported by PhD grant from Project CetSenti (RECI/AAG-GLO/0470/2012) and CIISA (UID/CVT/00276/2013), C. Eira was supported by the Portuguese Science Foundation (FCT) through CESAM (UID/AMB/50017/2013),A. Marçao, M. Ferreira and A. Lopez were supported by FCT grants (SFRH/BPD/64889/2009, SFRH/BD/30240/2006 and SFRH/BPD/82407/2011, respectively). Availability of data and materials The datasets supporting the conclusions of this article are available in the GenBank (National Center for Biotechnology Information) repository in http://ncbi.nlm.nih.gov. Access numbers for the sequences: KP835987; KP835991; KP835995; KP835999; KP836003; KP835986; KP835990; KP835994; KP835997; KP836002; KP836006; KP835985; KP835989; KP835993; KP835996; KP836001; KP836005; KP835984; KP835988; KP835992; KP835998; KP836000; KP836004; KP835983; KT878649; KT878650; KT878651; KT878652; KT878653; KT878654; KT878655; KT878656; KT878657; KT878658; KT878659; KT878660; KT878661. Phylogenetic data was submitted to TreeBase (submission number 19466) and is available at: http://purl.org/phylo/treebase/phylows/study/TB2:S19466. Authors’ contributions CB, CE, AD were responsible for the conception and the study design and actively participated in the analysis and data interpretation. CE, MF, AL, AM, LT and JV actively participated in the sample collection and data interpretation. CB, CE and AD were also involved in the drafting and revision of the article. 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 Not applicable. ==== Refs References 1. Beineke A Siebert U Wohlsein P Baumgärtner W Immunology of whales and dolphins Vet Immunol Immunopathol 2010 133 81 94 10.1016/j.vetimm.2009.06.019 19700205 2. Van Bressem MF Duignan P Banyard A Barbieri M Colegrove K De Guise S Di Guardo G Dobson A Domingo M Fauquier D Fernandez A Goldstein T Grenfell B Groch K Gulland F Jensen B Jepson P Hall A Kuiken T Mazzariol S Morris S Nielsen O Raga J Rowles T Saliki J Sierra E Stephens N Stone B Tomo I Wang J Cetacean morbillivirus: current knowledge and future directions Viruses 2014 6 5145 81 10.3390/v6125145 25533660 3. Domingo M Visa J Pumarola M Marco AJ Ferrer L Rabanal R Kennedy S Pathologic and immunocytochemical studies of morbillivirus infection in striped dolphins (Stenella coeruleoalba) Vet Pathol 1992 29 1 10 10.1177/030098589202900101 1557861 4. Van Bressem MF Van Waerebeek K Raga JA A review of virus infections of cetaceans and the potential impact of morbilliviruses, poxviruses and papillomaviruses on host population dynamics Dis Aquat Organ 1999 38 53 65 10.3354/dao038053 10590929 5. Aguilar A Raga J The striped dolphin epizootic in the Mediterranean Sea Ambio 1993 22 524 8 6. Forcada J Aguilar A Hammond PS Pastor X Aguilar R Distribution and numbers of striped dolphins in the western mediterranean sea after the 1990 epizootic outbreak Mar Mamm Sci 1994 10 137 50 10.1111/j.1748-7692.1994.tb00256.x 7. Van Bressem MF Raga JA Guardo G Jepson P Duignan P Barrett T César M Santos DO Moreno I Siciliano S Aguilar A Emerging and recurring diseases in cetaceans worldwide and the role of environmental stressors Dis Aquat Organ 2009 86 143 57 10.3354/dao02101 19902843 8. Groch KR Colosio AC Marcondes MCC Zucca D Díaz-Delgado J Niemeyer C Marigo J Brandão PE Fernández A Catão-Dias JL Novel cetacean morbillivirus in Guiana Dolphin, Brazil Emerg Infect Dis 2014 20 511 3 10.3201/eid2003.131557 24565559 9. Stephens N Duignan PJ Wang J Bingham J Finn H Bejder L Patterson IAP Holyoake C Cetacean morbillivirus in coastal indo-pacific bottlenose dolphins, Western Australia Emerg Infect Dis 2014 20 666 70 10.3201/eid2004.131714 24656203 10. West KL, Sanchez S, Rotstein D, Robertson KM, Dennison S, Levine G, Davis N, Schofield D, Potter CW, Jensen B. A Longman’s beaked whale (Indopacetus pacificus) strands in Maui, Hawaii, with first case of morbillivirus in the central Pacific. Mar Mamm Sci. 2012;767–76. 11. Kennedy S Smyth JA Cush PF McCullough SJ Allan GM McQuaid S Viral distemper now found in porpoises Nature 1988 336 21 10.1038/336021a0 3185717 12. Van Bressem MF Visser I Van de Bildt M Teppema J Raga J Osterhaus A Morbillivirus infection in Mediterranean striped dolphins (Stenella coeruleoalba) Vet Rec 1991 129 471 2 10.1136/vr.129.21.471 1763470 13. Domingo M Ferrer L Pumarola M Marco A Plana J Kennedy S McAliskey M Rima BK Morbillivirus in dolphins Nature 1990 348 21 10.1038/348021a0 2234055 14. Taubenberger JK Tsai MM Atkin TJ Fanning TG Krafft AE Moeller RB Kodsi SE Mense MG Lipscomb TP Molecular genetic evidence of a novel morbillivirus in a long-finned pilot whale (Globicephalus melas) Emerg Infect Dis 2000 6 42 5 10.3201/eid0601.000107 10653568 15. Bellière EN Esperón F Fernández A Arbelo M Muñoz MJ Sánchez-Vizcaíno JM Phylogenetic analysis of a new Cetacean morbillivirus from a short-finned pilot whale stranded in the Canary Islands Res Vet Sci 2011 90 324 8 10.1016/j.rvsc.2010.05.038 20576281 16. Vingada J Marçalo A Ferreira M Eira C Henriques A Miodonsky J Oliveira N Marujo D Almeida A Barros N Oliveira I Monteiro S Araújo H Santos J: Capítulo I Interações Entre as Espécies-Alvo E as Pescas. Anexo Ao Relatório Intercalar Do Projecto LIFE MarPro NAT/PT/00038 2012 17. Sierra E Sanchez S Saliki JT Blas-Machado U Arbelo M Zucca D Fernandez A Retrospective study of etiologic agents associated with nonsuppurative meningoencephalitis in stranded cetaceans in the canary islands J Clin Microbiol 2014 52 2390 7 10.1128/JCM.02906-13 24759718 18. Sierra E Zucca D Arbelo M García-Álvarez N Andrada M Déniz S Fernández A Fatal systemic morbillivirus infection in bottlenose dolphin, Canary Islands, Spain Emerg Infect Dis 2014 20 269 71 10.3201/eid2002.131463 24447792 19. Van Bressem MF Waerebeek KV Jepson PD Raga JA Duignan PJ Nielsen O Di Beneditto AP Siciliano S Ramos R Kant W Peddemors V Kinoshita R Ross PS López-Fernandez A Evans K Crespo E Barrett T An insight into the epidemiology of dolphin morbillivirus worldwide Vet Microbiol 2001 81 287 304 10.1016/S0378-1135(01)00368-6 11390111 20. Raga J Banyard A Domingo M Corteyn M Van Bressem M Fernández M Aznar F Barrett T Morbillivirus Emerg Infect Dis 2008 14 471 3 10.3201/eid1403.071230 18325265 21. Di Sciara GN Venturino MC Zanardelli M Bearzi G Borsani FJ Cavalloni B Cetaceans in the central Mediterranean Sea: distribution and sighting frequencies Bolletino di Zool 1993 60 131 8 10.1080/11250009309355800 22. Duignan PJ House C Geraci JR Duffy N Rima BK Walsh MT Early G St Aubin DJ Sadove S Koopman H Morbillivirus infection in cetaceans of the western Atlantic Vet Microbiol 1995 44 241 9 10.1016/0378-1135(95)00017-5 8588318 23. Duignan PJ House C Geraci JR Early G Copland HG Walsh MT Bossart GD Cray C Sadove S Aubin DJST Moore M Morbillivirus infection in two species of pilot whale (Globicephala sp.) from the western Atlantic Mar Mamm Sci 1995 11 150 62 10.1111/j.1748-7692.1995.tb00514.x 24. Fernández A Esperón F Herraéz P de Los Monteros AE Clavel C Bernabé A Sánchez-Vizcaino JM Verborgh P DeStephanis R Toledano F Bayón A Morbillivirus and pilot whale deaths, Mediterranean Sea Emerg Infect Dis 2008 14 792 4 10.3201/eid1405.070948 18439363 25. Geraci JR Lounsbury VJ Marine mammals ashore: a field guide for strandings 1993 Galveston Texas A&M Sea Grant College Program 26. Grant RJ Banyard AC Barrett T Saliki JT Romero CH Real-time RT-PCR assays for the rapid and differential detection of dolphin and porpoise morbilliviruses J Virol Methods 2009 156 117 23 10.1016/j.jviromet.2008.11.008 19084557 27. Thompson JD Higgins DG Gibson TJ CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice Nucleic Acids Res 1994 22 4673 80 10.1093/nar/22.22.4673 7984417 28. Waterhouse AM Procter JB Martin DM Clamp M Barton GJ Jalview Version 2-a multiple sequence alignment editor and analysis workbench Bioinformatics 2009 25 1189 91 10.1093/bioinformatics/btp033 19151095 29. Huelsenbeck JP Ronquist F MRBAYES: Bayesian inference of phylogenetic trees Bioinformatics 2001 17 754 5 10.1093/bioinformatics/17.8.754 11524383 30. Ronquist F Huelsenbeck JP MrBayes 3: Bayesian phylogenetic inference under mixed models Bioinformatics 2003 19 1572 4 10.1093/bioinformatics/btg180 12912839 31. Rubio-Guerri C Melero M Esperón F Bellière EN Arbelo M Crespo JL Sierra E García-Párraga D Sánchez-Vizcaíno JM Unusual striped dolphin mass mortality episode related to cetacean morbillivirus in the Spanish Mediterranean Sea BMC Vet Res 2013 9 106 10.1186/1746-6148-9-106 23702190 32. Bellière EN Esperón F Sánchez-Vizcaíno JM Genetic comparison among dolphin morbillivirus in the 1990–1992 and 2006–2008 Mediterranean outbreaks Infect Genet Evol 2011 11 1913 20 10.1016/j.meegid.2011.08.018 21888991 33. Valsecchi E Amos W Raga JA Podestà M Sherwin W The effects of inbreeding on mortality during a morbillivirus outbreak in the Mediterranean striped dolphin (Stenella coeruleoalba) Anim Conserv 2004 7 139 46 10.1017/S1367943004001325 34. Bourret VJR Macé MRJM Crouau-Roy B Genetic variation and population structure of western Mediterranean and northern Atlantic Stenella coeruleoalba populations inferred from microsatellite data J Mar Biol Assoc UK 2007 87 265 10.1017/S0025315407054859 35. Barrett T Visser IK Mamaev L Goatley L Van Bressem MF Osterhaus ADME Dolphin and porpoise morbillivirus are genetically distinct from phocine distemper virus Virology 1993 193 1010 2 10.1006/viro.1993.1217 8460473

PMC005xxxxxx/PMC5002202.txt

==== Front BMC Med GenomicsBMC Med GenomicsBMC Medical Genomics1755-8794BioMed Central London 21610.1186/s12920-016-0216-3DebateExplorations to improve the completeness of exome sequencing Du Chen chen.du@alumni.uni-heidelberg.de Pusey Barbara N. barbara.pusey@nih.gov Adams Christopher J. christopher.adams@nih.gov Lau C. Christopher laucc@mail.nih.gov Bone William P. wpbone06@gmail.com Gahl William A. gahlw@mail.nih.gov Markello Thomas C. markellot@mail.nih.gov Adams David R. dadams1@mail.nih.gov NIH Undiagnosed Diseases Program, Common Fund, National Institutes of Health, National Human Genome Research Institute, Bethesda, MD USA 27 8 2016 27 8 2016 2016 9 1 5630 11 2015 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.Background Exome sequencing has advanced to clinical practice and proven useful for obtaining molecular diagnoses in rare diseases. In approximately 75 % of cases, however, a clinical exome study does not produce a definitive molecular diagnosis. These residual cases comprise a new diagnostic challenge for the genetics community. The Undiagnosed Diseases Program of the National Institutes of Health routinely utilizes exome sequencing for refractory clinical cases. Our preliminary data suggest that disease-causing variants may be missed by current standard-of-care clinical exome analysis. Such false negatives reflect limitations in experimental design, technical performance, and data analysis. Results We present examples from our datasets to quantify the analytical performance associated with current practices, and explore strategies to improve the completeness of data analysis. In particular, we focus on patient ascertainment, exome capture, inclusion of intronic variants, and evaluation of medium-sized structural variants. Conclusions The strategies we present may recover previously-missed, disease causing variants in second-pass exome analysis. Understanding the limitations of the current clinical exome search space provides a rational basis to improve methods for disease variant detection using genome-scale sequencing techniques. Electronic supplementary material The online version of this article (doi:10.1186/s12920-016-0216-3) contains supplementary material, which is available to authorized users. Keywords Clinical exome sequencingAnalytical qualityPerformance enhancementClinical genomicsRare diseasesCompleteness problemFalse negative resultshttp://dx.doi.org/http://dx.doi.org/10.13039/100000002National Institutes of Healthissue-copyright-statement© The Author(s) 2016 ==== Body Background The Undiagnosed Diseases Program (UDP) of the National Institutes of Health (NIH) was established in 2008 to evaluate patients who were undiagnosed despite an extensive medical workup [1–3]. Besides thorough clinical phenotyping by multiple specialists, the UDP has been utilizing exome sequencing and SNP array analysis when a genetic etiology is suspected. The feasibility of using exome sequencing to identify new disease genes was first demonstrated in 2009 [4, 5] and has ever since contributed to the discovery of many Mendelian disease genes [6]. Exome sequencing studies have become an increasingly routine clinical approach with a reported 25 % molecular diagnostic rate [7, 8]. In the past few years, we observed that an increasing percentage of pediatric patients referred to the UDP has already been studied with a clinical exome. Hence, we employed an extended set of analytic approaches to identify disease-causing variants beyond those detected by current clinical exome analysis. Several specific features of an exome analysis pipeline contribute to the final sensitivity and specificity of the overall test. They include the design of the analysis, e.g., included family members, limitations of the underlying sequencing technology, and a number of specific analytic parameters that affect variant filtration and prioritization. These include segregation rules, application of allele frequency cutoffs derived from control populations, and predictions of deleteriousness [9, 10]. The choice of analytic parameters depends on the experimental design and scope of testing, with an overall goal of optimizing the final list of prioritized variants to be subjected to manual curation. Stringent parameters, focusing only on coding sequences in known disease genes and common Mendelian inheritance models, may filter out a “true” variant and create false negative results. This approach is commonly applied in clinical exome analysis where clinical interpretability is prioritized over new gene discovery. Conversely, relaxed filtration settings feature enhanced sensitivity but generate a number of false positive variants that increase the work associated with final curation. This approach is generally more suitable to a research-level analysis, where variants in genes not yet associated with disease may be of intrinsic interest. Our current goal is to optimize research-level exome analysis for single, small-pedigree families. We quantify the consequences of widening the final search space using several analytic techniques. They include sequencing additional family members; evaluation of minimal exome coverage; reducing the number of false negative results by considering variants in non-coding regions; and searching for medium-sized indels missed by the standard genotyping modules of current analytical pipelines. Each of these potential second-pass procedures can be employed when standard exome analyses fail to provide a satisfactory explanation for the patient’s clinical features. Results Patients A growing number of pediatric patients referred to the UDP present with prior unrevealing clinical exome studies (Fig. 1a), which necessitates the development of novel diagnostic strategies to improve experimental power. Cohort-based studies, which have been successful in detecting disease genes by exome sequencing [4, 5], are not an appropriate tool to employ for our subjects given their high degree of phenotypic diversity as demonstrated by comparison of Human Phenotype Ontology terms [11, 12] that were used to characterize each affected patient (Fig. 1b). Additionally, most of our cases are limited to a nuclear family that is too small to achieve the LOD score thresholds used in linkage analyses (Fig. 1c). These factors have focused our attention on the general problem of maximizing the information that can be extracted from small-family, n = 1 cases [13].Fig. 1 Characteristics of the UDP patient cohort. a Percentage of pediatric cases (n = 11-28 per quarter year) with prior inconclusive exome sequencing that applied to the UDP. b Number of HPO terms that are shared by a given number of patients (total n = 350 affected individuals). The top five HPO terms that were used in more than 50 patients were spasticity, global developmental delay, gastroesophageal reflux, seizures and short stature. c Family structures of nuclear families in the UDP cohort, n = 329 Ascertainment of family members Our study participants have undergone extensive medical testing before being accepted into our program. As a result, many easily-identifiable diagnoses, and interpretable DNA sequence variants, have been previously excluded in patients of the UDP cohort, in particular when patients had undergone prior clinical exome testing. Therefore, our patients require an exome sequencing strategy optimized for agnostic testing, going beyond a clinical routine analysis pipeline, which is optimized for interpretable results and identifying variants in known disease genes. A typical exome analysis pipeline includes an unsupervised variant filtration component. This may be followed by manual BAM file inspection, manual bioinformatic curation and expert clinical evaluation. Practice varies with regard to the number of family members included in exome sequencing and analysis. The value of family trios is increasingly recognized [8, 11]; sequencing of all available siblings has been a common practice in the UDP [14]. To study the effect of included family members on exome analysis, we performed a standard variant filtration analysis on 45 families, while varying the family composition. These families included the proband, both unaffected parents and at least two additional siblings (at least a quintet in total). Taking different numbers of family members into account (Fig. 2), we filtered variants for segregation with disease, population frequency and transcript effects, but did not filter for known disease genes, in silico predictions of deleteriousness, or family-based linkage. The difference in the number of variants returned for manual evaluation was most striking between a singleton and a trio analysis, with about 1,126 (range 886-1521) vs. 117 (range 59-265) variants passing automated filtration, similar to the 10-fold reduction reported previously [11]. A large part of this effect is attributable to a decrease in the number of heterozygous variants. In a trio, such variants could be assessed for de novo occurrence in light of parental genotypes and/or phased correctly as compound heterozygous variant pairs. The number of false positive homozygous recessive variants was greatly reduced by the awareness of parental genotypes. Additional number of siblings beyond a trio provided a further reduction in the average number of variants down to 88 (range 22-209), 69 (range 8-171) and 54 (range 11-109) for quartets, quintets and sextets, respectively. These reductions were obtained even without mapping recombination sites [15], by which siblings can exclude genomic regions from consideration and therefore remove large numbers of false positive variants.Fig. 2 Effects of family members on the number of variants returned by computerized filtration. Variants of 36 quintets and 9 sextet families were analyzed with an increasing number of family members. Bars show average number of variants that passed a basic filtration algorithm for segregation with disease, population frequency and transcript effects, based upon various inheritance models In our research setting, the increased filtration power of additional informative meioses in family members balanced against the effort required to obtain the correct affected status, and the financial cost of sequencing and data processing. We justified this approach as being worthwhile in order to reduce the number of false positive variants when analyzing individual families (n = 1), so that time and effort could instead be used to reduce the extent of false negative results and therefore to improve completeness of analysis. Exome coverage Capture and enrichment of the exome for medical sequencing limits most of the analysis to the protein coding sequences –about 1-2 % of all 3.2 billion base positions in the human genome– where the majority of previously documented disease-causing variants occur [16]. To assess loss of data due to lack in exome capture, we selected a subset of exomes that included 54 probands of the UDP cohort. Each member of this subset was sequenced under the same conditions at one sequencing center, and aligned through the same pipeline. When coverage is based on the targeted regions as determined by the capture kit (TruSeq, about 61 Mb), we observed a mean coverage of 76 reads and a median coverage of 57 reads. 85 % of targeted positions were covered more than 20×, a coverage frequently used as a minimal read depth requirement for confident genotyping of both alleles. Since capture kits differ in their target regions, we evaluated the coverage of all exonic regions defined by CCDS (about 31 Mb), as a measure of desired capture. Here we observed a mean coverage of 82 reads (range 44-130) and a median coverage of 61 reads (range 32-110), with an average of 88 % (range 74-94) of exonic positions covered more than 20× (Fig. 3a, b).Fig. 3 Potential false negative space due to lack of coverage. a Mean and median coverage based on exonic CCDS positions (total about 31.9 Mb) in 54 exomes of probands used for coverage analysis. b Fraction of exonic positions as a function of coverage. c-f Coverage was based on the minimum read depth that occurred in CCDS exons, grouped by exons (total about 190,000), first exons (total about 20,600) or genes (total about 18,400). g-i The minimum number of mapped reads at each position of known HGMD variants (classes “DM” and “DM?”) was determined by SAMtools and grouped by variant positions (total about 118,900) or genes (total about 4,000) Next we examined coverage based on CCDS exons instead of individual positions and grouped all exons into categories based on the minimum number of reads occurring in a given exon. This analysis showed that an average of 2.2 % of exons (range 1.6-2.8) had no coverage at all, 2.1 % (range 1.6-2.8) of exons had partially no coverage, 6 % (range 2-17) showed a minimum coverage of fewer than 10 reads, and 12 % (range 2-32) fewer than 20 reads (Fig. 3c). In total, an average of 10 % of exons (range 6-21) had a minimum coverage of fewer than 10 reads and 22 % (range 8-53) of exons had a read depth of fewer than 20 reads in at least one position (Fig. 3d). Of these low coverage exons, an average of more than 25 % (range 24-30) were in genes known to harbor disease-causing or likely disease-causing variants (HGMD classes “DM” and “DM?”). In addition, coverage below 10 reads and below 20 reads was observed in 27 % (range 19-38) and 42 % (range 26-63) of first exons, respectively (Fig. 3e), indicating a notable contribution of first exons to the low coverage regions of exon sequencing, as generally recognized. When analyzed by entire genes (total number n = 18.351 for females, n = 18.409 for males), an average of 56 % (range 39-78) and 76 % (range 51-94) of genes had a minimum coverage below a depth of 10 and 20 reads, respectively (Fig. 3f), suggesting that low coverage occurred across the entire length of genes. To estimate the impact of insufficient coverage on the diagnosis of known diseases, we queried the read depth at positions of known disease-causing variants listed in HGMD (n = 118.861 positions for females, n = 118.949 positions for males). This subset of positions showed a higher coverage with only 5 % (range 3-11) of known disease-causing variant positions covered less than 10×, and 11 % (range 4-31) were covered with fewer than 20 reads (Fig. 3g-i). Low coverage variant positions of a read depth of less than 10× occurred in about 25 % (range 17-38) of all HGMD genes and in 38 % (range 22-62) based on a coverage of less than 20 × . In a next step we analyzed the consistency of coverage across proband samples. Of 64,818 autosomal exons that showed read depths below 10 reads, 8199 exons were affected by low coverage in all 54 probands, of which 4406 exons were within the target sequence of the capture kit used (list of exons in Additional file 1: Table S1). When analyzed by genes, 1304 genes were covered with fewer than 10 reads in all 54 probands in more than 25 % of their exons. In these genes, all low coverage exons in 841 genes were within the targeted region of the capture kit used (list of genes in Additional file 1: Table S2). Coverage below 10 reads in all exons occurred in 409 genes, of which 277 genes consisted of a single exon only. While it is possible to characterize the number of variants in known disease genes that are missed due to incomplete capture, we hypothesize that variants in genes not yet associated with disease may be missed due to lack of coverage of one of the two alleles, especially if strict requirements for read depth are applied during data analysis. Non-exonic variants The majority of reported disease-causing variants reside in coding regions or canonical splice sites at exon/intron boundaries [16] but there are instances where deep intronic variants have been associated with disease [17–19] and genome sequencing has revealed evidence of selective pressure on intergenic and intronic regions, suggesting functional conservation [20]. Targeted re-analysis of candidate loci for non-coding variants has been successfully attempted [21]. However, non-coding variants are also detected as a result of off-target capture, and may be of high quality [22]. Therefore we quantified the analytical potential of targeted and off-target non-exonic variants, while exploring strategies to optimize the less-favorable signal to noise characteristics. We analyzed variants called in 54 probands and their parents in non-exonic locations, defined as outside the UCSC exon regions (Fig. 4). On average, about 156,000 non-exonic variants were called in each proband (range from about 121,000 to 252,000). The median distance to the nearest exon was 169 bases (range 142-298) and the mean was 15,000 bases (range 11,600-25,700), with a maximal distance observed of 1.7 million bases on average. About 70 % of variants were within a distance of 500 bases from the nearest exon boundary (Fig. 4a). The number of non-exonic variants dropped consistently at a distance of about 300-500 bases from the nearest exon (2.47 to 2.69 on the logarithmic scale in Fig. 4b). Given an average DNA libraries size of 280 bp, this cutoff appears to correspond to the maximal distance between a non-coding variant and targeted exonic sequence that occurred on either end of the same DNA fragment. This suggests that most non-exonic variants were sequenced as flanking regions of intended capture, rather than captured due to sequence similarity or presented as artifacts due to misalignment.Fig. 4 Prioritization of non-exonic variants. Non-exonic variants of 54 probands and their parents were evaluated. a The distance of non-exonic variants to the nearest exon was determined based on UCSC exon regions and the average number of variants was grouped by distance. b The average number of variants observed at a given distance to the nearest exon plotted over distance on a logarithmic scale. c Minor allele frequencies estimated by a founder population comprised of 106 parents, excluding each proband’s own parents. Number of variants was averaged and graphed per allele frequency group. d CADD scores [23] were calculated and the average number of non-coding variants were grouped by raw CADD score. e Fraction of variants as a function of Phred-scaled CADD scores. f Average number of non-coding variants remaining for consideration after each filtration or prioritization step To filter the large number of non-coding variants, we estimated the minor allele frequencies based on a founder population comprised of 106 parents. We found that only 6 % of the variants seen in each proband were absent from the founder population, which excluded each proband’s own parents (Fig. 4c). Of variants present in the founder population, only 6 % were rare at a 1 % lower confidence interval limit of the estimated minor allele frequency. Overall, variant filtration based on allele frequency reduced the number of unique and rare non-exonic variants to 19,000 per proband on average (range from 7,500 – 89,000). Prediction of deleteriousness is particularly difficult for non-coding variants since most prediction tools are limited to non-synonymous codon changes or canonical splice site positions. We tested the use of CADD scores [23] on our non-coding variant set, since this approach allows scoring of all SNVs and CNVs, not limited to coding variants. Raw CADD scores below zero were returned for 47 % of all non-coding variants, indicating that these variants were not different from known benign variation (Fig. 4d). When we analyzed Phred-scaled CADD scores, we observed a mean score of 3.82, median of 2.84, a minimum score of 0.001 and a maximum of 36.5 on average. Only 7 % of variants obtained a score higher than 10 and only 1 % higher than 15 (Fig. 4e), indicating that a very low number of non-coding variants were actually predicted deleterious by CADD scores. Therefore, CADD scores could be an approach to highlight potentially interesting variants within the vast pool of mostly benign, but poorly annotated non-coding variants. More focused tools to predict splice site changes using multiple different algorithms and that can be incorporated into an automated computational pipeline should be able to prioritize additional variants for consideration. Taking all information together, filtration based on allele frequency reduced the number of variants from initially 156,000 non-exonic variants per proband to 19,000 (Fig. 4f). Using a conservative approach to predict benign variants by negative raw CADD scores reduced the number of remaining variants to 9,400. Applying prediction of deleteriousness, an average of 1,200 variants were predicted damaging at a Phred-scaled CADD score of 10. When more stringent filters were applied, only 285 variants on average obtained a CADD score higher than 15 and only 40 variants on average scored higher than 20. Depending on the desired stringency, we found that these filtration strategies generated a tractable number of additions to a routine second pass analysis. Medium-sized indel calling The limitations of genome-scale data analysis to identify structural variants is another known cause of false negative results. Standard variant callers typically identify indels up to about 50 bases. Supplementing exome diagnostics with SNP chip or array-CGH data is known to detect indels larger than a few kilobases genome wide. Therefore, a large range of medium-sized indels from 50 bases to a few kilobases remain unaccounted for in subsequent variant evaluation pipelines, contributing to incomplete analyses. Many attempts have been made to address this issue by calling indels from exome sequencing data with additional methods [24–28], and such efforts have recently been implemented in large scale exome research studies [11]. While some programs examine read depth against a reference population, Pindel [29] extracts unmapped reads from BAM files and analyzes soft clipped bases of read pairs for evidence of medium-sized structural variation. We used Pindel to quantify the extent of incomplete analysis resulting from missed indels in a cohort of 54 probands and their parents. Pindel detected on the order of 33,000 structural variants per proband on average (range about 22,000 to 54,000), of up to +/- 16,000 bases in size (Fig. 5a). Half of these indels occurred within the target region defined by the capture kit and 50 bases of flanking regions (Fig. 5b). On average, 63 % of all indels and 61 % of indels within target regions were also detected by the variant caller and were therefore redundant (Fig. 5d). A breakdown by size showed that deletions of fewer than 50 bases and insertions of fewer than 10 bases were identified by both Pindel and the genotype caller in 56 % to 76 % of variants (Fig. 5c). The overlap of called variants dropped rapidly beyond deletions larger than 100 bases and insertions larger than 50 bases.Fig. 5 Calling medium sized indels with Pindel. Pindel [29] was used to detect structural variants in exome sequencing data of n = 54 probands. a Average number of variants by size detected by Pindel within the target region +/- 50 bases or in all regions. b Percentage of variants within or outside the target region. c Comparison of Pindel variants to the number of indels called by the variant caller broken down by variant size. d Percentage of Pindel variants within or outside the target regions +/- 50 bases that were also called by the variant caller. e Estimated allele frequencies based on a founder population of 106 parents. f Phasing of variants detected in the probands considering all variants or only variants that were unique to Pindel and that were rare. g Average number of Pindel variants after each step of filtration left for continued evaluation We also used Pindel to call indels in the probands’ parents and used their variants to estimate allele frequencies. This analysis revealed that only 5 % of indels in the probands were absent from the control population (Fig. 5e). Of the indels that were present in the founder population (excluding each proband’s own parents), 8 % were rare at a 1 % lower confidence interval limit of the estimated minor allele frequency. Thus, filtering all Pindel variants by allele frequency for unique or rare variants reduced the number of variants to about 4,200 on average (range from 1,500 – 11,700). When we analyzed phasing of all Pindel variants of the proband, about 34 % of all variants appeared to be inherited by one of the parents, 46 % occurred in both parents and 20 % could not be associated with an inheritance pattern (Fig. 5f). As expected, in a subset of variants that are only detected by Pindel and that are rare in or absent from the founder population, the percentage of variants that are present in both parents is greatly reduced (12 %), and enriched for variants not detected in the founder population (68 %). Combining all information gained on the indels called by Pindel (Fig. 5g), we started with an average of 33,000 Pindel variants per proband. An average of 12,200 variants were detected by Pindel only and not by the standard variant caller, of which an average of 2,600 variants were rare or unique indels in the proband by filtration based on founder frequencies (of which 1,200 variants map within target regions). This appears to be the lower limit that purely frequency-based bioinformatics tools can reach based on this cohort size, before variants are evaluated for other parameters such as Mendelian inheritance models and prediction of deleteriousness. With increased numbers and matched ethnicity of individuals in the control population, the power of filtration of ancient and benign variants based on allele frequency will improve and optimize computerized prioritization of indels when included in a second-pass analysis pipeline. Discussion In most clinical situations when exome sequencing is ordered as a diagnostic test, it is necessary and sufficient to limit the search parameters to well-defined areas of known disease genes and predictable protein changes. However, if such an attempt fails to reveal a molecular diagnosis and an agnostic research effort is made to discover potential new disease genes, non-Mendelian inheritance patterns, multi-gene conditions or unusual pathogenicity, it is desirable to broaden the analytical range to more obscure territories. Widening search parameters to improve completeness of analysis by including non-exonic variants and medium sized indels may include the “true” variant that may be missed otherwise, but at the expense of increased noise. To handle the additional load of variants, we used the genotype data of 106 unaffected parents as a control cohort, under the assumption that the probands do not share the same rare condition. Although the power of filtration against large control populations, such as 1000 Genomes [30, 31], ESP (NHLBI GO Exome Sequencing Project, URL: http://evs.gs.washington.edu/EVS/) or the ExAC database (Exome Aggregation Consortium, URL: http://exac.broadinstitute.org), is well recognized, we were impressed by the power of filtration that a relatively small cohort of 106 individuals sequenced under the same technical conditions, could provide for analyzing non-exonic variants and medium-sized indels. The use of genome sequencing information released with phase 3 of the 1000 Genomes project will further improve our ability to interpret the biological significance of SNVs and CNVs in non-exonic regions, such as in introns, promoter regions, UTRs, enhancers, intergenic regions and in transcribed non-coding DNA. This is especially important when the search parameters exceed coding regions and canonical splice sites, since most tools to predict deleteriousness are typically limited to these territories. We characterized the lack of coverage in our exome sequencing studies, which is a generally recognized contributor of false negative results. The application of genome sequencing is known to provide a more homogenous coverage and therefore represents a true advance in the attempt to analyze all genes [32, 33]. However, even this technology lacks coverage at difficult to sequence, possibly non-mappable genomic regions [34]. Other technical advances, such as low amplification technologies (so called “PCR-free chemistry”), which minimizes allele dropout, is only available with larger DNA sample acquisition and whole genome sequencing chemistry. In current exome sequencing, even highly medically important sets of genes, such as the 56 genes recommended by the American College of Medical Genetics and Genomics for clinical testing, can lack full coverage [35, 36]. Spiking in extra baits enhances the capture of known disease genes. Although coverage has improved with newer capture kits [32], this approach appears inefficient when trying to fully capture all known genes [37], including those not yet associated with disease. In addition to improved coverage, genome sequencing has also proven useful in the detection of small and large structural variants [38, 39], which may be missed by the standard variant calling pipelines used in exome sequencing. Advances in long read sequencing are hoped to provide optimized identification of structural variants and also allow detection of disease-causing tandem repeats [40–42]. In the future, it will be desirable to include other approaches to variant discovery into an analysis pipeline, beyond classical models of SNVs and CNVs that alter protein function. Examples include changes in DNA methylation [43, 44], disease-causing mobile elements [45] and the spatial organization of the genome [46–48], for example the disruption of long-range enhancer function or perturbation of insulators of topologically associating domains. Once a research based exome sequencing experiment returns promising variants or novel disease gene candidates, demonstration of causality remains a challenge [10], especially given the typical n = 1 family situation. Apart from functional studies to elucidate molecular mechanisms, identification of other similarly affected families will be key to validate research findings. Given the limited number of patients seen at a single clinical site, sharing data of both exome sequencing as well as clinical features, especially in standardized form using HPO terms that enable computerized phenotype matching, will become increasingly important [49]. Additionally, comparison of standardized phenotype information to model organisms and integration of pathway analyses have the potential to identify new disease genes for research studies [50–52]. Conclusions We explored strategies to improve completeness of analysis in a research setting and demonstrated that the number of variants added is tractable when the search parameters are expanded to include non-coding variants and medium-size indels. Future applications of these approaches will establish to what degree these additional efforts contribute to the number of solved cases, in light of feasibility and efficiency, relative to the number of cases with inconclusive clinical exomes referred for research studies. Methods Patients Patients of the UDP were enrolled in clinical protocol 76-HG-0238, “Diagnosis and Treatment of Patients with Inborn Errors of Metabolism and Other Genetic Disorders”, approved by the Institutional Review Board of the National Human Genome Research Institute (NHGRI), and gave written informed consent. Human Phenotype Ontology terms (HPO) were used for standardization as part of clinical phenotyping [12, 53]. The medical records of pediatric UDP applicants reviewed between 2013 and 2015 were screened to determine if exome or genome sequencing had been performed prior to evaluation by the UDP. To estimate family structures, a manually curated pedigree file containing all UDP participants who underwent molecular testing by SNP chip and/or exome sequencing was analyzed. Exome sequencing and analysis DNA was extracted from whole blood using the FlexStar system (Autogen). Libraries of ~280 bp and paired-end index adapters were prepared according to Illumina’s TruSeq V1 or V2 method and sequenced at the NIH Intramural Sequencing center (NISC) on a HiSeq2000 sequencer (Illumina) using 101-bp paired-end reads. Alternatively, 275-325 bp DNA libraries were constructed using KAPA library preparation kit (KAPA), captured with SeqCap EZ Exome plus UTR Lib capture kit (Nimblegen) and sequenced at NISC on a HiSeq2500 (Illumina) using 126-bp paired-end reads. Other exomes were captured using TruSeq kit (Illumina) and sequenced at Axeq (Rockville, MD). For all exome sequencing experiments, short reads were aligned to human reference genome GRCh37 using an in-house developed pipeline based on Novoalign (Novocraft Technologies). Variants were called with HaplotypeCaller and GenotypeGVCFs [54–56]. Annotations utilized snpEff [57] and a combination of internal cohort statistics and publically available data sources (NHLBI GO Exome Sequencing Project (ESP), URL: http://evs.gs.washington.edu/EVS/, 1000Genomes [30]). Basic, computerized variant filtration was used to analyze the effects of family members on the number of variants returned. In brief, rare, non-synonymous, start-gain/loss, stop-gain/loss, frameshift, canonical splice site variants and intronic variants (up to 20 bases from splice sites) were evaluated under homozygous recessive, compound heterozygous, X-linked and de novo dominant disease models in families of European descent. A cohort of 54 cases, comprising of the proband and both parents, that were sequenced under the same conditions at NISC using TruSeqV2 capture was used for analysis of low coverage regions, non-coding variants, and medium-sized indels. Coverage analysis Positions within target regions, exons as annotated by the Consensus Coding Sequence project (CCDS) [58–60] or variant positions annotated as disease-causing in the Human Gene Mutation Database (release 2014-1, classes “DM” and “DM?”) [16] were queried for read depth using SAMtools in 54 probands of the UDP [61, 62]. For per base coverage analysis, read depth at each unique position was considered. For exon- and gene-based analysis, the minimum read depth that occurred in a given exon was determined and used to group into categories. Non-exonic variants Non-coding variants outside the regions defined by UCSC exons (annotated in hg19 assembly by the University of California Santa Cruz [63]) were analyzed, while left-normalizing multiallelic variants. The parents of the 54 probands served as a control population, excluding parents of the proband who was examined (n = 106 individuals, 212 alleles for autosomal variants, 159 alleles for X-chromosomal variants, and 54 alleles for Y-chromosomal variants). Absence of an alternative allele in the founders was treated as presence of a reference allele. To hypothesize that a variant was too frequently present in the control population to be causative for a rare disease, a cutoff of 6 or more variant alleles was used for autosomal alleles, which corresponds to a minor allele frequency (MAF) of 0.0283 (95 % confidence interval (CI) of 0.013 to 0.0604 using population proportion interval estimation [64]). For X-chromosomal variants, a cutoff at 5 or more alleles was used (MAF of 0.0316, 95 % CI of 0.0136 to 0.0719) and 2 or more variant alleles for Y-chromosomal variants (MAF of 0.037, 95 % CI of 0.0102 to 0.1253). Variants in the probands were annotated with raw and Phred-scaled CADD v1.3 scores [23]. A negative raw score was used to assume that a variant was benign. Phred-scaled CADD scores were used to predict deleteriousness. Medium-sized structural variants in exome sequencing Pindel [29] was used to detect medium-sized structural variants compared to GRCh37 assembly in 54 probands and their parents. The presence of supporting reads was interpreted as a heterozygous variant allele. Structural variants were compared to those called during pipeline genotyping, applying left-normalization of multiallelic variants. Parental data were used to determine the phase of variants in the probands. The parents of 54 probands served as a control population, excluding the parents of the proband who was examined. Under the assumption that all indels identified by Pindel are in heterozygous state, we applied the thresholds as stated above. Additional file Additional file 1: Table S1. Coverage of exons in 54 probands. Table S2. Genes with low coverage exons in all 54 probands. (XLSX 9613 kb) Abbreviations UDPUndiagnosed diseases program NIHNational Institutes of Health NHGRINational Human Genome Research Institute HPOHuman phenotype ontology terms NISCNIH intramural sequencing center CCDSConsensus coding sequence project UCSCUniversity of California Santa Cruz MAFMinor allele frequency CADDCombined annotation dependent depletion HGMDHuman gene mutation database SNVSingle nucleotide variant CNVCopy number variant SNPSingle nucleotide polymorphism Acknowledgements We thank our patients and their families for their participation in the UDP, and we appreciate the contributions of our dedicated clinical team members. Special thanks to the bioinformatics team. Funding This research is supported by the Intramural Research Program of the National Human Genome Research Institute and the Common Fund of the National Institutes of Health. Availability of data and materials All genomic data is in the process of being submitted to dbGaP. Once the submission is complete, the authors would be happy to supply a specific dbGaP Project ID on request. Authors’ contributions CD, BNP, CCL, TCM and DRM designed the study. CD, BNP, CJA and WPB wrote code for data analysis. CD, BNP and DRA interpreted data. CD, WAG, TCM and DRA wrote the manuscript. All authors reviewed and approved the final manuscript. Competing interests The authors declare that they have no competing interest. Consent for publication Not applicable since this manuscript does not contain any individual person’s data. Ethics approval and consent to participate Patients of the UDP were enrolled in clinical protocol 76-HG-0238, “Diagnosis and Treatment of Patients with Inborn Errors of Metabolism and Other Genetic Disorders”, approved by the Institutional Review Board of the National Human Genome Research Institute (NHGRI), and gave written informed consent. ==== Refs References 1. Gahl WA Boerkoel CF Boehm M The NIH Undiagnosed Diseases Program: bonding scientists and clinicians Dis.se models & mechanisms 2012 5 1 3 5 10.1242/dmm.009258 2. Gahl WA Markello TC Toro C Fajardo KF Sincan M Gill F The National Institutes of Health Undiagnosed Diseases Program: insights into rare diseases Genetics in medicine : official journal of the American College of Medical Genetics 2012 14 1 51 9 10.1038/gim.0b013e318232a005 22237431 3. Gahl WA Tifft CJ The NIH Undiagnosed Diseases Program: lessons learned Jama 2011 305 18 1904 5 10.1001/jama.2011.613 21558523 4. Ng SB Turner EH Robertson PD Flygare SD Bigham AW Lee C Targeted capture and massively parallel sequencing of 12 human exomes Nature 2009 461 7261 272 6 10.1038/nature08250 19684571 5. Ng SB Bigham AW Buckingham KJ Hannibal MC McMillin MJ Gildersleeve HI Exome sequencing identifies MLL2 mutations as a cause of Kabuki syndrome Nat. Genet. 2010 42 9 790 3 10.1038/ng.646 20711175 6. Chong JX Buckingham KJ Jhangiani SN Boehm C Sobreira N Smith JD The Genetic Basis of Mendelian Phenotypes: Discoveries, Challenges, and Opportunities Am. J. Hum. Genet. 2015 97 2 199 215 10.1016/j.ajhg.2015.06.009 26166479 7. Yang Y Muzny DM Reid JG Bainbridge MN Willis A Ward PA Clinical whole-exome sequencing for the diagnosis of mendelian disorders N. Engl. J. Med. 2013 369 16 1502 11 10.1056/NEJMoa1306555 24088041 8. Farwell KD Shahmirzadi L El-Khechen D Powis Z Chao EC Tippin Davis B Enhanced utility of family-centered diagnostic exome sequencing with inheritance model-based analysis: results from 500 unselected families with undiagnosed genetic conditions Genet. Med. 2015 17 7 578 86 10.1038/gim.2014.154 25356970 9. Markello TC Han T Carlson-Donohoe H Ahaghotu C Harper U Jones M Recombination mapping using Boolean logic and high-density SNP genotyping for exome sequence filtering Mol. Genet. Metab. 2012 105 3 382 9 10.1016/j.ymgme.2011.12.014 22264778 10. MacArthur DG Manolio TA Dimmock DP Rehm HL Shendure J Abecasis GR Guidelines for investigating causality of sequence variants in human disease Nature 2014 508 7497 469 76 10.1038/nature13127 24759409 11. Wright CF Fitzgerald TW Jones WD Clayton S McRae JF van Kogelenberg M Genetic diagnosis of developmental disorders in the DDD study: a scalable analysis of genome-wide research data Lancet 2015 385 9975 1305 14 10.1016/S0140-6736(14)61705-0 25529582 12. Robinson PN Kohler S Bauer S Seelow D Horn D Mundlos S The Human Phenotype Ontology: a tool for annotating and analyzing human hereditary disease Am. J. Hum. Genet. 2008 83 5 610 5 10.1016/j.ajhg.2008.09.017 18950739 13. Zhi D Chen R Statistical guidance for experimental design and data analysis of mutation detection in rare monogenic mendelian diseases by exome sequencing PloS one 2012 7 2 e31358 10.1371/journal.pone.0031358 22348076 14. Adams DR Sincan M Fuentes Fajardo K Mullikin JC Pierson TM Toro C Analysis of DNA sequence variants detected by high-throughput sequencing Human mutation 2012 33 4 599 608 10.1002/humu.22035 22290882 15. Roach JC Glusman G Smit AF Huff CD Hubley R Shannon PT Analysis of genetic inheritance in a family quartet by whole-genome sequencing Science 2010 328 5978 636 9 10.1126/science.1186802 20220176 16. Stenson PD Mort M Ball EV Shaw K Phillips A Cooper DN The Human Gene Mutation Database: building a comprehensive mutation repository for clinical and molecular genetics, diagnostic testing and personalized genomic medicine Hum. Genet. 2014 133 1 1 9 10.1007/s00439-013-1358-4 24077912 17. Agrawal A Hamvas A Cole FS Wambach JA Wegner D Coghill C An intronic ABCA3 mutation that is responsible for respiratory disease Pediatr. Res. 2012 71 6 633 7 10.1038/pr.2012.21 22337229 18. Harland M Mistry S Bishop DT Bishop JA A deep intronic mutation in CDKN2A is associated with disease in a subset of melanoma pedigrees Hum. Mol. Genet. 2001 10 23 2679 86 10.1093/hmg/10.23.2679 11726555 19. Vache C Besnard T le Berre P Garcia-Garcia G Baux D Larrieu L Usher syndrome type 2 caused by activation of an USH2A pseudoexon: implications for diagnosis and therapy Hum. Mutat. 2012 33 1 104 8 10.1002/humu.21634 22009552 20. Yu F Lu J Liu X Gazave E Chang D Raj S Population genomic analysis of 962 whole genome sequences of humans reveals natural selection in non-coding regions PloS One 2015 10 3 e0121644 10.1371/journal.pone.0121644 25807536 21. Mele C Lemaire M Iatropoulos P Piras R Bresin E Bettoni S Characterization of a New DGKE Intronic Mutation in Genetically Unsolved Cases of Familial Atypical Hemolytic Uremic Syndrome Clin. J. Am. Soc. Nephrol.: CJASN 2015 10 6 1011 9 10.2215/CJN.08520814 25854283 22. Guo Y Long J He J Li CI Cai Q Shu XO Exome sequencing generates high quality data in non-target regions BMC Genomics 2012 13 194 10.1186/1471-2164-13-194 22607156 23. Kircher M Witten DM Jain P O'Roak BJ Cooper GM Shendure J A general framework for estimating the relative pathogenicity of human genetic variants Nat. Genet. 2014 46 3 310 5 10.1038/ng.2892 24487276 24. de Ligt J Boone PM Pfundt R Vissers LE Richmond T Geoghegan J Detection of clinically relevant copy number variants with whole-exome sequencing Hum. Mutat. 2013 34 10 1439 48 10.1002/humu.22387 23893877 25. Fromer M Moran JL Chambert K Banks E Bergen SE Ruderfer DM Discovery and statistical genotyping of copy-number variation from whole-exome sequencing depth Am. J. Hum. Genet. 2012 91 4 597 607 10.1016/j.ajhg.2012.08.005 23040492 26. Plagnol V Curtis J Epstein M Mok KY Stebbings E Grigoriadou S A robust model for read count data in exome sequencing experiments and implications for copy number variant calling Bioinformatics 2012 28 21 2747 54 10.1093/bioinformatics/bts526 22942019 27. Le Scouarnec S Gribble SM Characterising chromosome rearrangements: recent technical advances in molecular cytogenetics Heredity 2012 108 1 75 85 10.1038/hdy.2011.100 22086080 28. Challis D Antunes L Garrison E Banks E Evani US Muzny D The distribution and mutagenesis of short coding INDELs from 1,128 whole exomes BMC Genomics. 2015 16 143 10.1186/s12864-015-1333-7 25765891 29. Ye K Schulz MH Long Q Apweiler R Ning Z Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short reads Bioinformatics 2009 25 21 2865 71 10.1093/bioinformatics/btp394 19561018 30. Abecasis GR Altshuler D Auton A Brooks LD Durbin RM Gibbs RA A map of human genome variation from population-scale sequencing Nature 2010 467 7319 1061 73 10.1038/nature09534 20981092 31. Abecasis GR Auton A Brooks LD DePristo MA Durbin RM Handsaker RE An integrated map of genetic variation from 1,092 human genomes Nature 2012 491 7422 56 65 10.1038/nature11632 23128226 32. Lelieveld SH Spielmann M Mundlos S Veltman JA Gilissen C Comparison of Exome and Genome Sequencing Technologies for the Complete Capture of Protein-Coding Regions Human Mutation 2015 36 8 815 22 10.1002/humu.22813 25973577 33. Meienberg J Zerjavic K Keller I Okoniewski M Patrignani A Ludin K New insights into the performance of human whole-exome capture platforms Nucleic Acids Res. 2015 43 11 e76 10.1093/nar/gkv216 25820422 34. Gudbjartsson DF Helgason H Gudjonsson SA Zink F Oddson A Gylfason A Large-scale whole-genome sequencing of the Icelandic population Nat. Genet. 2015 47 5 435 44 10.1038/ng.3247 25807286 35. Green RC Berg JS Grody WW Kalia SS Korf BR Martin CL ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing Genet. Med. 2013 15 7 565 74 10.1038/gim.2013.73 23788249 36. Park JY Clark P Londin E Sponziello M Kricka LJ Fortina P Clinical exome performance for reporting secondary genetic findings Clin. Chem. 2015 61 1 213 20 10.1373/clinchem.2014.231456 25414276 37. Meynert AM Ansari M FitzPatrick DR Taylor MS Variant detection sensitivity and biases in whole genome and exome sequencing BMC Bioinformatics. 2014 15 247 10.1186/1471-2105-15-247 25038816 38. Brand H Pillalamarri V Collins RL Eggert S O'Dushlaine C Braaten EB Cryptic and complex chromosomal aberrations in early-onset neuropsychiatric disorders Am. J. Hum. Genet. 2014 95 4 454 61 10.1016/j.ajhg.2014.09.005 25279985 39. Gilissen C Hehir-Kwa JY Thung DT van de Vorst M van Bon BW Willemsen MH Genome sequencing identifies major causes of severe intellectual disability Nature 2014 511 7509 344 7 10.1038/nature13394 24896178 40. Usdin K The biological effects of simple tandem repeats: lessons from the repeat expansion diseases Genome research 2008 18 7 1011 9 10.1101/gr.070409.107 18593815 41. Zavodna M Bagshaw A Brauning R Gemmell NJ The accuracy, feasibility and challenges of sequencing short tandem repeats using next-generation sequencing platforms PloS One 2014 9 12 10.1371/journal.pone.0113862 25436869 42. English AC Salerno WJ Hampton OA Gonzaga-Jauregui C Ambreth S Ritter DI Assessing structural variation in a personal genome-towards a human reference diploid genome BMC Genomics. 2015 16 286 10.1186/s12864-015-1479-3 25886820 43. Lv J Liu H Su J Wu X Liu H Li B DiseaseMeth: a human disease methylation database Nucleic Acids Res. 2012 40 Database issue D1030 5 10.1093/nar/gkr1169 22135302 44. Robertson KD DNA methylation and human disease Nat. Rev. Genet. 2005 6 8 597 610 10.1038/nrg1655 16136652 45. Solyom S Kazazian HH Jr Mobile elements in the human genome: implications for disease Genome Med 2012 4 2 12 10.1186/gm311 22364178 46. Bickmore WA The spatial organization of the human genome Annu. Rev. Genomics Hum. Genet. 2013 14 67 84 10.1146/annurev-genom-091212-153515 23875797 47. Dixon JR Selvaraj S Yue F Kim A Li Y Shen Y Topological domains in mammalian genomes identified by analysis of chromatin interactions Nature 2012 485 7398 376 80 10.1038/nature11082 22495300 48. Lupianez DG Kraft K Heinrich V Krawitz P Brancati F Klopocki E Disruptions of topological chromatin domains cause pathogenic rewiring of gene-enhancer interactions Cell 2015 161 5 1012 25 10.1016/j.cell.2015.04.004 25959774 49. Sobreira N, Schiettecatte F, Valle D, Hamosh A. GeneMatcher: A Matching Tool for Connecting Investigators with an Interest in the Same Gene. Human mutation. 2015. doi:10.1002/humu.22844. 50. Robinson PN Kohler S Oellrich A Wang K Mungall CJ Lewis SE Improved exome prioritization of disease genes through cross-species phenotype comparison Genome Res 2014 24 2 340 8 10.1101/gr.160325.113 24162188 51. Buske OJ, Girdea M, Dumitriu S, Gallinger B, Hartley T, Trang H et al. PhenomeCentral: A Portal for Phenotypic and Genotypic Matchmaking of Patients with Rare Genetic Diseases. Human Mutat. 2015. doi:10.1002/humu.22851. 52. Bone WP Washington NL Buske OJ Adams DR Davis J Draper D Computational evaluation of exome sequence data using human and model organism phenotypes improves diagnostic efficiency Genetics in medicine : official journal of the American College of Medical Genetics 2015 53. Kohler S Doelken SC Mungall CJ Bauer S Firth HV Bailleul-Forestier I The Human Phenotype Ontology project: linking molecular biology and disease through phenotype data Nucleic Acids Res 2014 42 Database issue D966 74 10.1093/nar/gkt1026 24217912 54. McKenna A Hanna M Banks E Sivachenko A Cibulskis K Kernytsky A The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data Genome Res 2010 20 9 1297 303 10.1101/gr.107524.110 20644199 55. DePristo MA Banks E Poplin R Garimella KV Maguire JR Hartl C A framework for variation discovery and genotyping using next-generation DNA sequencing data Nat. Genet. 2011 43 5 491 8 10.1038/ng.806 21478889 56. Van der Auwera GA, Carneiro MO, Hartl C, Poplin R, Del Angel G, Levy-Moonshine A et al. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Current protocols in bioinformatics / editoral board, Andreas D Baxevanis [et al.]. 2013;11(1110):11 0 1- 0 33. doi:10.1002/0471250953.bi1110s43. 57. Cingolani P Platts A le Wang L Coon M Nguyen T Wang L A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3 Fly 2012 6 2 80 92 10.4161/fly.19695 22728672 58. Farrell CM O'Leary NA Harte RA Loveland JE Wilming LG Wallin C Current status and new features of the Consensus Coding Sequence database Nucleic Acids Res. 2014 42 Database issue D865 72 10.1093/nar/gkt1059 24217909 59. Pruitt KD Harrow J Harte RA Wallin C Diekhans M Maglott DR The consensus coding sequence (CCDS) project: Identifying a common protein-coding gene set for the human and mouse genomes Genome research 2009 19 7 1316 23 10.1101/gr.080531.108 19498102 60. Harte RA, Farrell CM, Loveland JE, Suner MM, Wilming L, Aken B et al. Tracking and coordinating an international curation effort for the CCDS Project. Database : the journal of biological databases and curation. 2012;2012:bas008. doi:10.1093/database/bas008. 61. Li H Handsaker B Wysoker A Fennell T Ruan J Homer N The Sequence Alignment/Map format and SAMtools Bioinformatics 2009 25 16 2078 9 10.1093/bioinformatics/btp352 19505943 62. Li H A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data Bioinformatics 2011 27 21 2987 93 10.1093/bioinformatics/btr509 21903627 63. Karolchik D Hinrichs AS Furey TS Roskin KM Sugnet CW Haussler D The UCSC Table Browser data retrieval tool Nucleic Acids Res 2004 32 Database issue D493 6 10.1093/nar/gkh103 14681465 64. Newcombe RG Interval estimation for the difference between independent proportions: comparison of eleven methods Stat. Med. 1998 17 8 873 90 10.1002/(SICI)1097-0258(19980430)17:8<873::AID-SIM779>3.0.CO;2-I 9595617