Datasets:

keerthanapacific

/

sample4

like 0

555567

Ethmoidal osteoid osteoma with orbital and intracranial extension – a case report

Background Osteoid osteoma is a benign bone neoplasm which is seen in the long bones of appendicular skeleton. It is rarely seen in the cranium. Skull base osteoid osteoma is extremely rare and has been anecdotally reported. Case presentation The authors report a case of a large osteoid osteoma of the ethmoid with intraorbital and intracranial extension in a 33 year old male patient. He presented with loss of vision in the left eye. The intra-cranial extension was excised through a single burr-hole fronto-orbitotomy. The ethmoid and orbital portions were approached and excised through a Weber-Ferguson incision and inferior orbitotomy. Radical excision of the tumor could thus be achieved through a craniofacial approach. Conclusion Although benign and rare, skull base osteoid osteoma can present with neurological deficit due to its mass effect and involvement of vital structures. A multispeciality team approach is advisable in such cases if radical excision is to be achieved. A craniofacial approach made radical single stage excision of this multicompartmental osteoid osteoma possible with an uneventful postoperative period.

Background Osteoid osteoma is a rare benign osteoblastic lesion usually involving the long bones of the lower limbs. Cranial involvement has been mainly localised to the skull vault. Osteoid osteoma of the skull base is a rare entity [ 1 ]. Surgical management of skull base osteoid osteoma may be challenging due to its proximity to vital structures, access and hard consistency. This report deals with a case of an ethmoidal osteoid osteoma invading the adjacent orbit and anterior cranial fossa and the team approach employed to achieve radical excision. The radiological findings and the surgical procedure employed are presented. Case presentation Clinical presentation A 33 year old male presented with decreased vision in the left eye and left sided headache of 3 months duration. Examination revealed left eye blindness with primary optic atrophy and no other neurological deficit. Imaging revealed a bony tumor in the left ethmoid sinus invading the left orbit and compressing the left optic nerve. Intracranial extension into the anterior cranial fossa on the left side was noted (Figure 1 ). Core biopsy of the mass showed an osteoid osteoma. Surgical management A multispeciality team approach was devised to achieve radical excision of the tumor. Bicoronal scalp flap and pericranial flaps were raised separately. A single burr hole left fronto orbital bone flap was raised including the orbital roof and left zygomaticofrontal process (Figure 2 ). Dura was retracted and the bony hard whitish tumor visualised. This was excised using the high speed drill (Figure 3 ). Weber Ferguson incision was used to access the orbital portion of the tumor. Medially the tumor could be felt in the orbit but retraction of the globe was difficult. Hence inferior orbitotomy was done by removing the lower and lateral orbital margins. The intraorbital contents could now be retracted laterally and the tumor visualized (Figure 4 ). The tumor was then detached from the ethmoid sinus and the intraorbital extension excised. The ethmoidal portion was drilled and radical excision achieved (Figure 5 ). Dural tears were covered with temporalis fascia and glue. The ethmoidal sinus was packed with free temporalis muscle graft. Vascularised pericranial graft was used to cover the anterior skull base. The frontal sinus was exenterated and packed with gelfoam. The bone flap and orbital margins were replaced. Postoperative period and follow up Post-operatively patient had a frozen left eye (possibly due to retraction of the orbital contents) with no improvement in the left eye vision in the immediate postoperative period. His postoperative period was otherwise uneventful. Histopathological examination revealed an osteoid osteoma (Figure 6 ). On follow up after 12 months patient was disease free (Figure 7 ). Discussion Osteoid osteoma is a benign osteoblastic lesion and constitutes 1% of all bone tumors and 11% of benign bone lesions [ 1 ]. It is usually seen in the second and third decades and a male preponderance has been noted. It can occur throughout the skeleton but the long bones of the lower extremities and the vertebrae are most commonly affected. They are usually metaphyseal but may be epiphyseal occasionally. It is frequently localized to the cortex (85%) but may also occur in spongiosa (13%) and subperiosteal region (2%) [ 2 ]. Cranial cases are seen to generally arise from the skull vault. Skull base osteoid osteomas are extremely rare and occur in the frontal or ethmoidal sinuses [ 1 , 3 ]. It usually presents with sharply localized pain and tenderness especially at night. In our case the osteoid osteoma was seen to originate from the ethmoid sinus and pain was not a presenting feature. The radiological diagnosis rests on Computerised Tomography and isotope bone scan [ 1 ]. Radiographically osteoid osteoma appears as a radio opaque lesion with a nidus which has a radiolucent centre surrounded by dense sclerosis [ 2 ]. This may at times be mistaken for Garre's osteomyelitis. Occasionally the nidus may have a radio opaque centre with a surrounding radiolucent area. In our patient no definite nidus could be visualized probably due to the large size and unusual location. The treatment generally consists of en bloc resection or curettage of the tumor. Recurrence rate after incomplete resection may be upto 10% [ 1 ]. If asymptomatic and small, the lesion may be left alone and observed. However a rare complication of a large pneumocephalus has been reported by Ferlito et al from a frontoethmoidal osteoid osteoma [ 4 ]. Our patient presented with loss of vision due to a large osteoid osteoma of the ethmoid invading the left orbit. Anterior skull base lesions have been approached through a frontoorbitotomy which is usually removed as two separate parts. We have found that a single burr hole frontoorbitotomy flap gives excellent exposure to the anterior skull base without excessive retraction on the brain and is also cosmetically superior. Combining the craniotomy with a Weber Ferguson incision and orbitotomy made a single stage radical excision possible. Histologically the nidus is sharply delineated from the surrounding variably thick layer of dense bone. The nidus is composed of more or less calcified osteoid lined by plump osteoblasts within a highly vascularised connective tissue stroma [ 2 ] (Figure 6 ). It does not invade the adjacent tissue. No malignant transformation has been reported [ 1 ]. A differential diagnosis of benign osteoblastoma may be entertained. However, in the case of the latter, active osteoblasts are more numerous and the stroma is richly vascularized and extravasated blood with large number of multinucleated giant cell macrophages are noted [ 5 ]. Several authors have stressed the fact that the two are identical histologically and the differentiation between them if any can only be on the basis of size [ 5 ]. In our patient although the tumor was large, none of the above mentioned histological features of a benign osteoblastoma could be noted. Conclusion Osteoid osteoma of the skull base is rare and anecdotally reported. Radical excision is difficult especially if the tumor involves major blood vessels and cranial nerves. The surgical team constituted the neurosurgeon, surgical oncologist and plastic surgeon. A craniofacial approach made radical single stage excision of this multicompartmental osteoid osteoma possible with an uneventful postoperative period. Competing interests All the authors of the article "ETHMOIDAL OSTEOID OSTEOMA WITH ORBITAL AND INTRACRANIAL EXTENSION -CASE REPORT" hereby declare that there are no competing interests – financial and non financial. Authors' contributions SBP performed the craniotomy, excised the intracranial extension of the tumor, assisted in excising the orbital portion, drilled the ethmoidal portion and drafted the manuscript. KH excised the orbital portion and helped draft the manuscript. MSV and US performed the orbitotomy, helped in tumor excision and closure. DJ did the histopathological examination and helped in drafting the manuscript. All authors have read and approved of the manuscript . Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555567.xml

545953

Gene expression in the brain and kidney of rainbow trout in response to handling stress

Background Microarray technologies are rapidly becoming available for new species including teleost fishes. We constructed a rainbow trout cDNA microarray targeted at the identification of genes which are differentially expressed in response to environmental stressors. This platform included clones from normalized and subtracted libraries and genes selected through functional annotation. Present study focused on time-course comparisons of stress responses in the brain and kidney and the identification of a set of genes which are diagnostic for stress response. Results Fish were stressed with handling and samples were collected 1, 3 and 5 days after the first exposure. Gene expression profiles were analysed in terms of Gene Ontology categories. Stress affected different functional groups of genes in the tissues studied. Mitochondria, extracellular matrix and endopeptidases (especially collagenases) were the major targets in kidney. Stress response in brain was characterized with dramatic temporal alterations. Metal ion binding proteins, glycolytic enzymes and motor proteins were induced transiently, whereas expression of genes involved in stress and immune response, cell proliferation and growth, signal transduction and apoptosis, protein biosynthesis and folding changed in a reciprocal fashion. Despite dramatic difference between tissues and time-points, we were able to identify a group of 48 genes that showed strong correlation of expression profiles (Pearson r > |0.65|) in 35 microarray experiments being regulated by stress. We evaluated performance of the clone sets used for preparation of microarray. Overall, the number of differentially expressed genes was markedly higher in EST than in genes selected through Gene Ontology annotations, however 63% of stress-responsive genes were from this group. Conclusions 1. Stress responses in fish brain and kidney are different in function and time-course. 2. Identification of stress-regulated genes provides the possibility for measuring stress responses in various conditions and further search for the functionally related genes.

Background Until recently multiple gene expression profiling was applied almost exclusively to human and a few model organisms. At present cDNA microarrays are being constructed for new species including teleost fishes [ 1 - 6 ]. Since EST sequencing projects are carried out with a large number of species, continuous development of new platforms can be expected in the future. We designed a salmonid fish cDNA microarray primarily to characterize responses to stress, toxicity and pathogens. This paper focuses on time-course comparisons of stress responses in rainbow trout and the usage of functional annotation to conduct analyses of gene expression data. Functional annotation of genes, especially Gene Ontology [ 7 ] is increasingly being used for analyses and interpretation of microarray results [ 8 - 13 ]. We applied Gene Ontology in several modes to facilitate implementation of our research tasks. Furthermore, experimental results generated guidelines for the development of specialized microarrays. Well designed platforms are expected to ensure identification of differentially expressed genes while containing representative coverage from important functional groups. Custom made microarrays include clones from cDNA libraries and/or selected genes, which have advantages and drawbacks. Indiscriminant spotting of EST may result in under representation of many functional classes. On the other hand selection of genes fully relies on annotations and hypotheses, which can be misleading and limit possibilities for nontrivial findings. We used clones from normalized and subtracted cDNA libraries as well as genes selected by the functional categories of Gene Ontology for inclusion onto a microarray targeted at characterizing transcriptome responses to environmental stressors. Designing a new platform requires balancing a large number of genes versus multiple replications of spots, which enhances statistical analyses of data. The rainbow trout microarray was prepared by spotting of relatively small number of genes (1300) in 6 replicates. We show that multiple replications combined with the dye-swap design of hybridization [ 14 , 15 ] allows for accurate detection of relatively small alterations in expression levels, which is important for the functional interpretation of results. Stress is closely associated with many diverse issues in fish biology and environmental research (reviewed in [ 16 ]). Stress is generally defined as the reaction to external forces and abnormal conditions that tend to disturb an organism's homeostasis. To illustrate the major trends in the studies of stress in fish, we performed a computer-assisted analysis of Medline abstracts covering this area (Table 1 ). Salmonids have been studied more extensively than any other fish species. Research has focused on various biotic and abiotic factors including toxicity, environmental parameters (oxygen, temperature, salinity, acidosis), diseases, social interactions (crowding, aggressiveness) and farming manipulations. Analysis of Medline abstracts indicated physiological processes, cellular structure and selected proteins that have been the major foci of previous fish stress studies. This provided an outline for interpretation of our results. We analyzed the effects of stress on the transcriptome in the brain and kidney, which are considered important target tissues along with muscle, blood cells, liver and epithelia. We report a profound difference of stress response in these tissues and the identification of a diagnostic set of genes. Table 1 Thematic associations in studies of fish stress. Computer-assisted analysis of 11129 Medline abstracts was performed as described in Methods. Terms that were over-represented in the abstracts (exact Fisher's test, P < 0.05) are ranked by the numbers of occurrence. Category Terms (counts/1000 abstracts) Species Salmonids (126.4), carp (68.9), eels (67.0), catfish (37.7), tilapia (38.7) Stressors Toxicity (440.6), temperature (178.3), oxygen (91.5), confinement (52.8), salinity (46.2), hypoxia (54.7), diseases (20.8), crowding (23.6), acidosis (17.9), aggressiveness (11.3) Messengers Cortisol (208.5), catecholamines (159.4), steroids (92.5) Tissues Muscle (197.2), blood cells (152.9), pituitary (119.8), liver (123.6), epithelia (96.2), brain (90.6), kidney (89.6), heart (51.9), skin (42.5) Cellular structures Cytosol (42.5), collagen (17.0), cytoskeleton (15.1), microsome (15.1), microtubule (14.2), lysosomes (13.2), peroxisome (4.7) Oxidative stress Glutathion (167.9), oxidant (93.4), antioxidant (90.6), peroxide (66.0), radical (55.7), superoxide (40.6), catalase (35.8), redox (18.9) Other processes Immunity (91.5), secretion (80.2), metabolism (74), transport (56.6), defense (52.8), necrosis (28.3), apoptosis (18.9), phosphorylation (15.1), proteolysis (7.5) Metabolites Ion (987.7), iron (215.1), glucose (141.5), lactate (67.9), lipid (74.5), zinc (51.9), phospholipid (11.3), triglyceride (11.3), lipopolysaccharide (9.4) Proteins Enzymes (180.2), heat-shock proteins (84.0), hemoglobin (37.7), metallothionein (37.7), transferase (32.1), phosphatase (26.4), chaperones (21.7), glutathion-S-transferase (17.0), transaminase (17.0), Na/K-ATPase (17.0), aminotransferase (8.5), mitogen-activated kinases (4.7) Results 1 Design of cDNA microarray The rainbow trout cDNA microarray was composed of EST and selected genes. The cDNA libraries were prepared from tissues of stressed fish using suppression subtractive hybridization, SSH [ 17 ] and a modification of the cap-finder method [ 18 ] supplemented with enzymatic normalization [ 19 ]. We sequenced 2000 clones and redundancy of the subtracted libraries was markedly greater than that of the normalized (306% and 134% respectively). In addition to EST we selected rainbow trout transcripts from the normalized multi-tissue cDNA library [ 20 ] based on their assignment to functional categories of Gene Ontology (stress and defense response, regulation of cell cycle, signal transduction, chaperone activity and apoptosis). The selected genes substantially improved the coverage of many functional classes (Table 2 ), though the number of differentially expressed genes in this group was markedly inferior to EST (Figure 1 ). Subtraction cloning enriched genes that showed strong alteration of expression at response to stress (p < 0.01 or lower, Figure 1A ), however the SSH clone set did not provide any advantage when microarray was used for the related research tasks (Figure 1B ). Table 2 Presentation of the Gene Ontology functional categories in the microarray. Table shows the numbers and frequncies of genes in the clone sets that were used for spotting (SSH – subtracted libraries, EST – normalized libraries). Gene Ontology classes N on slide SSH EST Selected Response to external stimulus 147 11 (0.07) 48 (0.11) 88 (0.31) Response to stress 145 7 (0.04) 30 (0.07) 108 (0.38) Defense response 105 6 (0.04) 34 (0.08) 65 (0.23) Humoral immune response 42 3 (0.02) 13 (0.03) 26 (0.09) Apoptosis 79 6 (0.04) 10 (0.02) 63 (0.22) Cell communication 139 11 (0.07) 45 (0.11) 83 (0.29) Cell proliferation 82 8 (0.05) 23 (0.05) 51 (0.18) Cell cycle 62 2 (0.01) 17 (0.04) 43 (0.15) Signal transduction 114 5 (0.03) 32 (0.07) 77 (0.27) Receptor activity 49 3 (0.02) 18 (0.04) 28 (0.10) Intracellular signaling cascade 49 3 (0.02) 15 (0.04) 31 (0.11) DNA metabolism 47 5 (0.03) 15 (0.04) 27 (0.09) Transcription 67 9 (0.05) 21 (0.05) 37 (0.13) Chaperone activity 41 4 (0.02) 12 (0.03) 25 (0.09) Figure 1 Performance of the clone sets used for preparation of the microarray. Figure shows frequencies of genes that were differentially expressed in at least 5 samples at different p-values (Student's t-test). A: this study (stress response), B: related experiments (exposure to aquatic contaminants [34], response to stress, cortisol and combination of these treatments, challenge with bacterial antigens, M74 disease). SSH – subtracted cDNA libraries, EST – normalized libraries, Select – genes chosen by the Gene Ontology functional categories. 2 Stress response in the brain and kidney of rainbow trout 2.1 Differentially expressed genes Fish were stressed with netting and samples were collected 1, 3 and 5 days after the first exposure. We used plasma cortisol as a stress marker [ 21 ]. The hormone levels increased 7.6-fold after 1 day and did not change significantly to the end of experiment (Figure 2 ). Figure 2 Plasma cortisol levels. The data are mean ± SE (n = 4). Difference between the control and stressed fish is significant (Student's t-test, p < 0.05). Microarray results were submitted to GEO ( GSM22355 ). Two genes were up-regulated in both tissues at all time-points (Figure 3 ). One is a putative homolog to the mammalian N-myc regulated genes, which are induced with steroid hormones in the brain [ 22 ] and kidney [ 23 ]. Mitochondrial ADP, ATP carrier can be implicated to both normal functions and cell death [ 24 ]. Metallothionein-IL, a classical stress marker was induced to the end of experiment and a similar profile was seen in midkine precursor (growth factor), histone H1.0 and B-cell translocation protein 1. In kidney we observed consistent up-regulation of genes related to energy metabolism, such as mitochondrial proteins (cytochromes b and c, cytochrome oxidases), enzymes (glyceraldehyde 3-phosphate dehydrogenase, fructose-bisphosphate aldolase, serine-pyruvate aminotransferase) and similar profiles were seen in two heat shock proteins and two signal transducers (cytohesin binding protein and GRB2-adaptor). The repressed genes were related to actin binding (coronin and profilin) and immune response (meprin, immunoglobulin epsilon receptor, thymosin and lysozyme). Figure 3 Examples of differentially expressed genes . Pooled RNA from 4 fish was hybridized in dye-swap experiments to two microarrays on which each gene was printed 6 times (total of 12 replicates). Differential expression was analysed with Student's t-test (P < 0.01); the expression ratio is coded with color scale. Rapid alteration of gene expression was a remarkable feature of stress response in the brain. Only one gene, aquaporin, was up-regulated for the duration of the experiment. Water channel aquaporin plays a key role in water homeostasis being implicated in various physiological processes and pathological conditions [ 25 ]. A panel of genes which showed markedly increased expression after 1 day was also suppressed after 5 days. Surprisingly, this group included mainly genes that are predominantly expressed in skeletal or cardiac muscle (myosin light chain 1 and 2, skeletal and cardiac isoforms, myosin heavy chain, troponin I, T and C) or are involved in regulation of muscle contraction (parvalbumin alpha and sarcoplasmic reticulum calcium ATPase). An opposite tendency was shown by a large group of genes however the magnitude of expression changes was smaller. We analysed 5 differentially expressed genes with qPCR and the results were in close concordance with the microarray data (not shown). 2.2 Functional classes The search for enriched Gene Ontology functional categories in the lists of differentially expressed genes found almost no overlap between the tissues (Table 3 ). In the brain stress affected binding and transport of metal ions, especially calcium and manganese, chaperones and heat shock proteins, cytoskeleton and microtubules and a number of signaling pathways; whereas, mitochondrion, extracellular structures and peptidases appeared the primary targets in the kidney. Table 3 Enrichment of Gene Ontology categories in the lists of differentially expressed genes. Analysis with exact Fisher's test, (p < 0.05) was made using the composition of microarray as a reference. The numbers of differentially expressed genes and genes on the microarray are in parentheses. Brain Kidney Intracellular signaling cascade (19/47) Mitochondrion (19/71) RAS protein signal transduction (6/9) Electron transporters (13/43) GTPase mediated signal transduction (11/16) Extracellular (19/70) Chaperones (16/40) Endopeptidases (8/22) Heat shock proteins (8/16) Metallopeptidases (7/12) Metal ion binding (31/80) Zinc ion binding (8/24) Carriers (15/37) Potential-driven transporters (7/9) Calcium ion binding (20/41) Magnesium ion binding (8/14) Cytoskeleton (27/76) Myofibril (16/16) Microtubule-based process (6/6) Comparison of the differentially expressed genes by the Gene Ontology categories suggested coordinated regulation of various cellular functions in the brain. Early stress response was marked with transient induction of the cytoskeleton proteins and similar profiles were observed in the metal binding proteins and enzymes of carbohydrate metabolism (Figure 4 ). An opposite expression pattern was shown by a large group of genes involved in stress and immune response, regulation of growth and cell cycle, apoptosis, signal transduction and cell to cell signaling. This was in parallel with enhancement of transcription and translation, ubiquitin-dependent protein catabolism and protein folding. In the kidney the temporal alterations were much weaker. Expression of metal binding proteins increased slowly in parallel with peptidases. Strong induction of collagenases coincided with decrease of collagen expression. At the same time a number of metabolic functions were suppressed (oxidative phosphorylation and oxidoreductase activity, amine metabolism and RNA binding). Figure 4 Time-course of stress response in the brain and kidney . Differentially expressed genes were grouped by the Gene Ontology categories and mean log (expression ratios) were analysed with Student's t-test. Panel presents examples of categories that showed significant difference between the time points (p < 0.05). The values are coded with color scale. 3 Stress-responsive genes Microarray design included genes from functional categories which were expected to be affected by stress (Table 2 ). Overall observations of differences in gene expression from this group in response to handling stress were minimal; however, this could be accounted for by its heterogeneity. Therefore we searched for the subgroups of genes with correlated expression profiles within the functional classes using results of 35 microarray experiments conducted by our laboratory. Both factorial and cluster analyses revealed 9 defense response genes that showed tightly coordinated expression being induced with stress. We continued search using the consensus profile of this subgroup and found 47 positively and 1 negatively correlated genes (Pearson r > |0.65|). Of these 29 were identified by the protein products (Figure 5A ), 19 being from the set of selected clones. Expression of the stress-responsive genes changed significantly in several experiments including this study (Figure 5B ). They were up-regulated in kidney with stress and injection of cortisol, combination of these treatments showed an additive effect (Figure 5C ). These genes also responded to the model water contaminants, being induced with low and medium and repressed with high doses (Figure 5D ). Figure 5 Expression of stress-responsive genes . A : Experiments. 1–6 : response to handling stress, this study. Kidney, 1 day (1), 3 days (2) and 5 days (3); brain, 1 day (4), 3 days (5) and 5 days (6). 7–12 : response to handling stress and exogenous cortisol in kidney. Cortisol, 1 day (7) and 3 days (8); stress, 1 day (9) and 3 days (10), combination of stress and injection of cortisol, 1 day (11) and 3 days (12). 13–20 : exposure of yolk sac fry to model contaminants [34]. β-naphthoflavone, low (13) and high (14) dose; cadmium, low (15) and high (16) dose; carbon tetrachloride, low (17) and high (18) dose; pyrene, low (19) and high (20) dose. 21–22: response of yolk sac fry to transportation stress, rainbow trout (21) and Atlantic salmon (22). Ranks are coded with color scale; correlation coefficients (Pearson r) with the mean expression profile are indicated. B-C: the mean ranks ± SE of the stress-responsive genes in 3 experiments. A : this study; B – response to handling stress and injection of cortisol in kidney; C – exposure of yolk sac fry to β-naphthoflavone, cadmium and pyrene at low, medium and high doses. Discussion I Stress response in rainbow trout Our study aimed at comparison of time-course of stress response in rainbow trout brain and kidney and finding of a diagnostic set of genes. These tasks were implemented with an aid of Gene Ontology annotation, which was used in several modes. The most straightforward and commonly used approach is counting of Gene Ontology classes in the lists of differentially expressed genes. Statistical inference of enrichment and depletion is made with Z-score of hypergeometric distribution, exact Fisher's test or its modifications. Such analyses helped us to interpret differences of stress responses in the brain and kidney (Table 3 ). In the brain handling stress mainly affected expression of transcripts for structural proteins (especially cytoskeleton), signal transduction, and binding of metal ions, whereas mitochondria, extracellular structures and peptidases appeared the key targets in the kidney. Computer-assisted analysis of Medline abstracts suggested that most of these themes have not been addressed in the studies of fish stress (Table 1 ). Searches of the enriched Gene Ontology categories associated with differentially expressed transcripts is useful for rapid screens of microarray data; however, it presumes coordinated expression of functionally related genes. This assumption is not valid for many classes, especially large and heterogenous groups, such as stress and defense response. Because the gene composition of microarray is used as a reference, uneven presentation of functional categories can distort the results. Finally, this analysis does not take into account direction and magnitude of differential expression. To overcome these problems, enrichment of Gene Ontology classes is analysed in groups of genes with similar expression profiles revealed with cluster or factorial analyses. In this study we preferred straight comparison of Gene Ontology classes by the mean log expression ratios which helped in interpretation of the time-course of stress response. In the kidney temporal alterations were relatively weak though significant. Expression of peptidases (especially collagenases) increased steadily, which implied possible degradation of tissue with prolonged stress. We could expect abrupt fluctuations in the rainbow trout brain, since transient induction and up-regulation of gene expression was observed in response to cold in the brain of channel catfish [ 4 ]. In our study most differentially expressed genes fell into two groups with distinct temporal profiles which showed remarkable coherence of the functional classes. Early phase was associated with dramatic up-regulation of structural and metal binding proteins, which were repressed in later phases. Expression of genes involved in stress and defense response, apoptosis and signal transduction, cell cycle and growth changed in a reciprocal fashion. Activation of metal binding proteins could be accounted for the role of ions (particularly calcium) in multiple pathways of gene expression regulation in the brain [ 26 ]. Motor proteins of cytoskeleton play key roles in the transport of vesicles and the establishment and rearrangement of neuronal networks [ 27 - 29 ] which also could be implicated to the stress response in fish. However, in mammals these functions are associated with non-muscle isoforms and therefore differential expression of the sarcomeric proteins was unexpected. Additional experiments confirmed induction of these proteins at early phase of stress response. Previously we observed high activity of skeletal α-actin and myosin light chain 2 promoters in the neural tissues of rainbow trout embryos [ 30 ]. Sequencing of salmonid fish cDNA libraries provided evidence for the brain expression of sarcomeric proteins, but their role remains fully unknown. At present there is sparse evidence for differential expression of structural muscle proteins in the mammalian brain. For example regulation of troponin I with dextromethorphan (antagonist of excitatory amino acid receptors) was reported in the rat hyppocampus and cortex [ 31 ]. Grouping of individual differentially expressed genes by the functional classes reduced noise and enhanced cluster and factorial analyses. This helped to identify stress-responsive genes that showed correlated expression in 35 microarray experiments (22 experiments are shown in Figure 5A ). Association with stress is well established for most of these proteins and some are used as stress markers. The list of enriched Gene Ontology categories (stress, defense and humoral immune response, signal transduction and response to oxidative stress, p < 0.05) suggested biological relevance of this group. Computer analysis of Medline abstracts (Table 1 ) showed that immunity and metabolism of reactive oxidative species are prioritized in studies of fish stress and these functional categories were enriched in the list of stress-responsive genes. Thus Gene Ontology provided a useful starting point for search of functionally related genes and results of these analyses can be used further for the revision of annotations. 2 Construction of microarrays Results of our experiments helped to evaluate the strategy used in construction of the rainbow trout microarray. Researchers developing microarrays for new species are commonly choosing between specially selected genes and clones from normalized and subtracted cDNA libraries. We used SSH, which is at present probably the most popular method of subtraction. Though proven efficient in many studies, this method has a number of drawbacks. Subtraction requires re-association of tester and formation of double-stranded DNA, hence many rare transcripts are not cloned and variations in concentrations of cDNA and hybridization conditions may have strong impacts on library composition. High redundancy is a common feature of the SSH libraries. Apart from these problems, rapid alterations of gene expression observed in this study and many other microarray experiments make the advantages of subtractive cloning ambiguous. Subtraction achieves enrichment of the transcripts, which are over or under represented in the test sample. In many cases one sample will not provide coverage of differentially expressed genes for the whole series, whereas pooling of samples may reduce fluctuations. Furthermore, we observed relatively high ratio of differentially expressed genes among the clones from the unsubtracted cDNA libraries, which are easier for construction and much less redundant. The advantage of subtractive cloning becomes negligible when microarrays are used for different, though related research tasks (Fig. 1B ). At present selection of genes for microarrays is facilitated with advances of functional annotation. This helped us to improve presentation of many functional categories (Table 1 ) and enhanced interpretation of results. Most of the selected genes did not show differential expression in our studies, however 63% of stress-responsive genes were from this group. In our view, this finding is a strong argument for utilizing Gene Ontology in the development of specialized platforms. Given the limited number of spots on slides, microarray design requires a careful balance between the number of genes and replication of spots. Apparent advantages of genome-wide platforms are compromised with the problems associated with identifying significantly differentially expressed genes. We preferred combination of multiple spotting and dye-swap normalization, which ensured robust normalization and accurate detection of differential expression at low ratios. Coordinated expression of functionally related genes suggested biological relevance of relatively small alterations in the transcription levels. Selection of differentially expressed genes by the cutoff values would result in loss of valuable information in our experiments. For instance, most of the stress-responsive genes showed small or moderate expression changes, the identification of this group would not be likely without multiple replications. Conclusions 1. Combination of EST and selected genes appears a reasonable way for construction of cDNA microarrays. Multiple replications of spots and dye swap design of hybridization ensure robust normalization and high power of statistical analyses. Finding of differential expression at small ratios is essential for the functional interpretation of microarray data. 2. Stress response in fish brain and kidney is different both by the target functions and time-course. In brain slow progression of adaptive response was preceded with dramatic transient induction of motor and metal ion binding proteins. Prolonged stress was likely to result in slow degradation of extracellular matrix in kidney. 3. Finding of stress-responsive genes provides possibility for measurement of stress in various conditions and search for the functionally related genes. Methods 1. Computer-assisted analysis of Medline abstracts Search of Medline was made with queries: "fishes AND stress" (1060 abstracts) and "fishes NOT stress" (10069 abstracts). Abstracts were split into separate words and a list of non-redundant terms was composed. The numbers of abstracts including each term were estimated. The terms were ranked by the Z-scores of hypergeometric distribution and enrichment was analysed with exact Fisher's test (p < 0.05). 2. Experiments with fish, exposure and sampling One year old rainbow trout were stressed with netting for 2 min, this treatment was repeated once a day for a duration of 5 days. Fish were killed with over-dose of anaesthetic (MS-222) and blood was taken from the caudal vein. The kidneys and brains were snap-frozen in liquid nitrogen. Plasma cortisol was determined with RIA using Orion Spectra Cortisol kit. 3. Preparation of microarrays RNA was extracted with Trizol reagent (Invitrogen) and mRNA was purified with Dynabeads kit (Dynal). SSH cloning was performed as described [ 17 ]. For preparation of normalized libraries, synthesis of cDNA with PowerScript reverse transcriptase (Clontech) was primed with oligonucleotides including EcoRI and NotI sites: 5'-ACGAGGC GAATTC ACAGAGAGGAG(T)VN-3', 5'-GAGAGAGAGTGGT GCGGCCGC GGTGTATGGGG-3'). Double-stranded cDNA was generated using Advantage DNA polymerase mix (Clontech) and PCR primers: 5'-ACGAGGC GAATTC ACAGAGAGGAG-3' and 5'-GAGAGTGGT GCGGCCGC GGTGTA-3'. The PCR products were purified with QIAquick kit (Qiagen), precipitated with ethanol and dissolved to 1 μg/μl in hybridization buffer (1 M NaCl, 50 mM HEPES (pH 8.3), 1 mM EDTA). DNA was denaturated for 5 min at 94°C. Following re-association at 72°C for 16 hours, DNA was ethanol precipitated and digested with 150 U of exonuclease III (MBI Fermentas) for 15 min at 37°C. This treatment eliminates re-associated double-stranded DNA [ 19 ]. Single-stranded DNA was PCR amplified, size separated with agarose gel electrophoresis and cloned into pGEM ® -11Zf (+) (Promega). Normalized and subtracted cDNA libraries were prepared from the stressed fish (whole fry, brain, kidney and spleen of 1-year old fish). A number of clones were from the rainbow trout and Baltic salmon cDNA libraries constructed in University of Turku. The sequences were analysed with stand-alone blastn and blastx [ 32 ] Microarray incldued 315 genes selected by the Gene Ontology functional categories. Of these, 282 clones were from the normalized multi-tissue library [ 20 ] and the rest were produced with RT PCR. The cDNA inserts were amplified with PCR using universal primers and purified with Millipore Montage PCR96 Cleanup Kit. DNA was spotted onto poly-(L) lysine-coated slides and each clone was printed in 6 replicates. 4. Microarray analyses Total RNA was extracted with Trizol reagent (Invitrogen) and 4 individuals were pooled in each sample. Stressed fish was compared with time-matched control. Labeling with Cy3- and Cy5-dCTP (Amersham Pharmacia) was made using SuperScript III (Invitrogen) and oligo(dT) primer; cDNA was purified with Microcon YM30 (Millipore). We used a dye swap experimental design [ 14 , 15 ] and each sample was hybridized to two microarrays. For the first slide, test and control cDNA were labeled with Cy5 and Cy3 respectively, and for the second array dye assignments were reversed. The slides were pretreated with 1% BSA, fraction V, 5 x SSC, 0.1% SDS (30 min at 50°C) and washed with 2 x SSC (3 min) and 0.2 x SSC (3 min) and hybridized overnight in cocktail containing 1.3 x Denhardt's, 3 x SSC 0.3% SDS, 0.67 μg/μl polyadenylate and 1.4 μg/μl yeast tRNA. All chemicals were from Sigma-Aldrich. Scanning was performed with ScanArray 5000 and images were processed with QuantArray (GSI Luminomics). The measurements in spots were filtered by criteria I/B ≥ 3 and ( I - B )/( S I + S B ) ≥ 0.6 , where I and B are the mean signal and background intensities and S I , S B are the standard deviations. After subtraction of mean background, lowess normalization [ 33 ] was performed. Differential expression was analysed with Student's t-test (p < 0.01) and the genes were ranked by the log(p-level). 5. Quantitative RT PCR Primers (Table 4 ) were designed to amplify 194–305 b fragments. RNA was processed with Rnase-free Dnase (Promega). Synthesis of cDNA with Superscript III reverse transcriptase (Invitrogen) was primed with oligo(dT). Analyses were carried out using Dynamo SYBR Green kit (Finnzymes) and ABI Prism 7700 (Amersham-Pharmacia). Table 4 Primers used for qPCR. Gene Sequence GRB2-related adaptor protein 2 Forward 5'-GCCAGAGCACCCCAGGAGAT-3' Reverse 5'-GGCTGAGAGGATGGGGCTGA-3' Collagenase type IV Forward 5'-AACATCAGAAACGCCCTCAT-3' Reverse 5'-TGGTGGTAGTGGTAGTGGAC-3' Troponin T Forward 5'-TGGGAAGAAGGAAACTGAGA-3' Reverse 5'-CTCTTACGCAGGGTTGTGAC-3' 40S ribosomal protein S12 Forward 5'-AGACCGCACTCATCCACGAC-3' Reverse 5'-CCACTTTACGGGGTTTTCCT-3' EST1 Forward 5'-CGGAGAAGGAGAACCCACAG-3' Reverse 5'-CCCTCAAACAAGCAAAGTG-3' EST2 Forward 5'-GCAAATGACAGCCCTCTTAG-3' Reverse 5'-AGCAGGTTTTCATCAAGGA-3' Author's contributions AK designed microarray, carried out experiments with fish and data analyses and drafted the manuscript. HK conducted the microarray analyses. PP developed software for annotation of genes and analyses of Medline abstracts. CR constructed the multi-tissue cDNA library and provided the selected genes. SA developed software for management of microarray data and performed the statistical analyses. HM coordinated research. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545953.xml

514886

A Developmental Role for Fatty Acids in Eukaryotes

Health food stores have long hawked fish oil capsules as a cure-all for everything from migraines to heart disease. And though such claims are often weak on scientific evidence, fish oil, it turns out, is no snake oil. A recent review of scientific studies concludes that omega-3 fatty acids can indeed protect against heart disease, and the American Heart Association now recommends fish oil capsules for patients with coronary heart disease. Fatty acids come in hundreds of varieties, distinguished primarily by their structure, which in turn determines their physiological role. Unlike proteins or genes—which are polymers made up of amino acids and nucleotides, respectively—fatty acids are a large group of compounds containing long chains of carbon and hydrogen atoms with a carboxylate group (acid) attached at the end. It is this asymmetrical chemical configuration that gives fatty acids their unique properties. Fatty acid diversity comes from variations in the length of the carbon chain and in the number of double bonds between carbons. Fatty acids with one or more double bonds are called unsaturated fatty acids. Fatty acids play an essential role in metabolism, providing the cell with a concentrated source of energy, and form the structural foundation of the cell membrane, where they are most conspicuous and perhaps best understood. Long-chain (unbranched) fatty acids, which run ten to 22 carbons long, are the most common fatty acids in animal cells and the most studied. One much less understood class of fatty acids—the monomethyl branched-chain fatty acids (mmBCFAs)—has been found in organisms from bacteria to humans, but its role remains obscure. In this issue of PLoS Biology , Marina Kniazeva et al. explore the origin and function of mmBCFAs in the worm Caenorhabditis elegans and find that these relatively obscure fatty acids play a crucial role in growth and development. mmBCFAs are abundant in diverse genera of bacteria, which use a supply of branched-chain amino acids and enzymes to assemble the fatty acid chains. mmBCFA biosynthesis has been characterized in bacteria, but not in eukaryotes. (Worms, and humans, are eukaryotes; our cells have nuclei.) Here, Kniazeva et al. identified worm genes that are homologous to the gene that codes for an enzyme called elongase in another eukaryote, yeast. Elongases are enzymes that extend the length of fatty acid chains by two carbons. To see what kind of fatty acid molecules the homologous worm genes were synthesizing, the authors used a technique called RNA interference (RNAi) to “silence” the genes' expression in the worms. Surprisingly, two of the eight inhibited genes had a specific effect on branched-chain fatty acid levels: elo-5 and elo-6 . Inhibiting elo-5 function had deleterious effects on the growth and development of the worms. The progeny of worms treated as embryos with RNAi for elo-5 stopped growing at the first larval stage, while the progeny of worms treated at later stages developed to adulthood but got progressively sicker and showed reproductive problems. These defects were corrected when the researchers fed the mmBCFAs directly to the worms, indicating that these mmBCFAs are essential for normal larval growth and development. Given the widespread distribution of mmBCFAs in organisms as diverse as bacteria and humans, it's perhaps not too surprising that they regulate essential physiological functions during animal development. It's still not clear, however, what all the components of the fatty acid manufacturing machinery are or how an organism monitors production levels. And though it's still an open question as to how these ubiquitous molecules function in mammals, the fact that they have been conserved throughout evolution underscores their importance—and suggests they may play a similar role.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514886.xml

524361

Generation of competent bone marrow-derived antigen presenting cells from the deer mouse (Peromyscus maniculatus)

Background Human infections with Sin Nombre virus (SNV) and related New World hantaviruses often lead to hantavirus cardiopulmonary syndrome (HCPS), a sometimes fatal illness. Lungs of patients who die from HCPS exhibit cytokine-producing mononuclear infiltrates and pronounced pulmonary inflammation. Deer mice ( Peromyscus maniculatus ) are the principal natural hosts of SNV, in which the virus establishes life-long persistence without conspicuous pathology. Little is known about the mechanisms SNV employs to evade the immune response of deer mice, and experimental examination of this question has been difficult because of a lack of methodologies for examining such responses during infection. One such deficiency is our inability to characterize T cell responses because susceptible syngeneic deer mice are not available. Results To solve this problem, we have developed an in vitro method of expanding and generating competent antigen presenting cells (APC) from deer mouse bone marrow using commercially-available house mouse ( Mus musculus ) granulocyte-macrophage colony stimulating factor. These cells are capable of processing and presenting soluble protein to antigen-specific autologous helper T cells in vitro. Inclusion of antigen-specific deer mouse antibody augments T cell stimulation, presumably through Fc receptor-mediated endocytosis. Conclusions The use of these APC has allowed us to dramatically expand deer mouse helper T cells in culture and should permit extensive characterization of T cell epitopes. Considering the evolutionary divergence between deer mice and house mice, it is probable that this method will be useful to other investigators using unconventional models of rodent-borne diseases.

Background Hantaviruses (family Bunyaviridae ) are rodent-borne and can cause hemorrhagic fever with renal syndrome (HFRS 4 ) or hantavirus cardiopulmonary syndrome (HCPS) [ 1 ]. While HFRS is usually associated with Eurasian hantaviruses, HCPS is caused by any of several recently described New World hantaviruses [ 2 - 4 ]. In North America, the great majority of HCPS cases have occurred in the western United States and Canada and were caused by Sin Nombre virus (SNV). Patients afflicted with HCPS exhibit pronounced pulmonary inflammation due to capillary leak syndrome, with the consequent hypotension often leading to rapid decline and death [ 5 ]. Virus is found in the lungs of infected humans, but without discernible cytopathology, and mononuclear infiltrates are observed that produce proinflammatory cytokines, including IL-2, IL-4, IFN-γ, TNF and lymphotoxin (LT) [ 6 ], suggesting that HCPS is an immunopathologic response to the virus. To date, more than 370 infections with hantavirus have been documented in the United States, with a 36% fatality rate. Deer mice ( Peromyscus maniculatus ) are the principal reservoir host of SNV [ 4 , 7 ]. As is usual with some natural hosts, SNV infection of deer mice does not result in discernible pathology [ 8 ]. Infection parallels that of humans, with virus infecting capillary endothelial cells in many tissues, including the lungs, but without conspicuous cytopathology. However, in contrast to human HCPS, no pulmonary inflammation, capillary leakage, or mononuclear infiltrates are observed, and most, if not all, deer mice remain persistently infected for the remainder of their lives [ 9 ]. Deer mice are among the most common mammals in North America, found from the subarctic to central Mexico, except for the Atlantic seaboard and the southeast United States where other peromyscine species predominate [ 10 , 11 ]. Serosurveys of natural rodent populations suggest that hantavirus infections occur throughout the range of deer mice [ 7 , 12 - 15 ], which thus poses a potential threat to individuals who are in contact with these rodents. In addition, deer mice and other peromyscine rodents have been shown to harbor other human pathogens [ 16 - 28 ]. Very little is known about the mechanism by which the deer mouse immune system engages SNV because few reagents and methodologies have been developed and no susceptible inbred deer mice are available. The only useful immunological data that can be acquired is by use of serology; infected deer mice produce a neutralizing IgG response that is inadequate to clear the virus [ 8 , 9 , 29 ]. We previously cloned several deer mouse cytokine genes [ 30 - 32 ], but quantitative assays for the detection of the expression of these genes have not been developed. These limitations have made it difficult to determine what immunological events occur that impair an effective immune response without pathology. In some viral infections, persistence has been shown to occur because of impairment of helper and cytotoxic T cell responses, antigen presenting cell (APC) function, and development of APC from bone marrow progenitors [ 33 - 37 ]. Currently, none of these functions can be evaluated in deer mice. Recent advances in hematopoietic stem cell research have identified an important role for granulocyte-macrophage colony stimulating factor (GM-CSF) in the expansion and maturation of bone marrow cells into competent APC [ 38 - 40 ]. We previously cloned a partial cDNA representing deer mouse GM-CSF and found that one of its receptor-binding domains is nearly identical to that of the common laboratory house mouse ( Mus musculus ) [ 30 ]. This led us to hypothesize that house mouse GM-CSF, which is commercially-available, might be useful in expanding and differentiating deer mouse bone marrow cells into competent APC. If so, then it should be possible to generate large pools of APC from individual deer mice that could be aliquotted and frozen for use in long-term T cell cultures, which would preclude the necessity for inbred deer mice. We present evidence that such cells can be propagated in vitro and that they are capable of processing antigen and stimulating antigen-sensitized autologous T cells. This technique could provide sufficient APC, such that conventional T cell cloning and peptide-mapping experiments could be performed. In addition, because deer mice (New World rodents) and house mice (Old World rodents) are divergent by 25 to 50 million years [ 41 ], it is possible that this approach may be useful to investigators using other unconventional rodent models of infectious diseases. Results Cloning of the 5' end of deer mouse GM-CSF We used RACE to obtain the complete 5' end of GM-CSF. This sequence was translated using the default translation table within MacVector. The polypeptide is predicted to have a 25 residue signal peptide based upon orthologous sequences from other species [ 42 - 45 ] (Figure 1 ). The receptor-binding domains of deer mouse and house mouse GM-CSF share 13/15 identical residues. This region forms the α-helix (helix A) that binds with high affinity to β chain subunit of the GM-CSF receptor complex that is shared with the IL-3 and IL-5 receptors [ 46 - 49 ]. Figure 1 Amino acid alignment of deer mouse (DM), cotton rat (CR), house mouse (HM) and human (HU) GM-CSF. Polypeptides were aligned with the clustal algorithm in Macvector. Conserved (light shading) and identical (dark shading) amino acids are enclosed in boxes. The 25-residue signal peptide is enclosed in box A. Helix A, which binds to the β chain subunit of the GM-CSF receptor, is enclosed in box B. Deer mouse and house mouse GM-CSF share 13 of 15 identical residues in this domain. Morphologic characteristics of bone marrow-derived APC Deer mouse bone marrow cultures contained mostly cells that appeared dead or dying after 24 hours in culture with GM-CSF. However, at 48 hours clusters of cells were apparent, while control wells without GM-CSF had fewer live cells than at 24 hours. By day 3, adherent stromal cell foci were conspicuous, while semiadherent and nonadherent cells were more evident and these became the prominent cells for the duration of culture. Day 12 bone marrow cells incubated for an additional 48 hours were large, ranging from 12 to 18 μm in diameter, and possessed macropinocytic vesicles and processes (Figure 2A ). Although the method that was employed selects for DC in the house mouse [ 39 ], the deer mouse cells appeared to resemble macrophages, with abundant cytoplasmic vesicles, rather than dendritic cells, with lamellipodia or characteristic long processes extending from the cell. (S. K. Chapes, pers. comm.). However, the cells' exact definition will not be complete until better phenotypic characterization is possible. Treatment of the cells with recombinant TNF diminished the macropinocytic vesicles (Figure 2B ). Figure 2 Morphologic characteristics of deer mouse bone marrow-derived APC. Day 14 bone marrow cells cultured in GM-CSF were processed by cytospin and stained with Wright's stain. The cells exhibited conspicuous cytoplasmic vesicles and small processes (A). Cells collected on day 12 and incubated for 48 hours with 20 ng/ml of hmTNF displayed less conspicuous cytoplasmic vesicles (B). Proliferation of deer mouse cells to house mouse GM-CSF and human IL-2 Day 8 deer mouse bone marrow cells were cultured with various concentrations of house mouse GM-CSF. Two days later, proliferation was assessed (Figure 3A ) and maximal proliferation was observed at about 0.5 ng/ml of GM-CSF. Deer mouse splenocytes proliferated in response to human IL-2 (Figure 3B ). In this experiment, deer mouse splenocytes were cultured with a suboptimal concentration of PHA and various concentrations of recombinant human IL-2. Maximal proliferation occurred at 20 U/ml of IL-2. In another proliferation assay, in vitro deer mouse T cells that were collected 8 days after stimulation with APC and antigen exhibited slightly greater proliferation to IL-2 (data not shown). Figure 3 Proliferation of deer mouse cells to recombinant cytokines. (A) After 8 days of incubation with GM-CSF, deer mouse bone marrow cells were washed and then cultured with dilutions of GM-CSF in duplicate for 48 hours, then proliferation assessed by MTS assay. The data are representative of four deer mice. (B) To assess proliferative capacity of deer mouse T cells to human IL-2, splenocytes were cultured with a suboptimal dose of PHA (2 μg/ml) and dilutions of recombinant human IL-2 in duplicate for 48 hours, and proliferation assessed by MTS assay. The data are representative of two deer mice. Expression of MHC class II I-Eβ and TCRβC by deer mouse cells propagated in vitro BM-APC and T cells were examined for the expression of orthologous I-Eβ and TCRβC, respectively, by RT-PCR (Figure 4 ). For I-Eβ, primers were designed from previously published deer mouse sequences [ 50 ], while primers for TCRβC were those that are described in this work. In each instance, products of the expected sizes were amplified. The amplified BM-APC product was cloned, sequenced, and verified to be I-Eβ. Figure 4 RT-PCR of TCRβC and MHC class II I-Eβ in deer mouse cells. Total RNA was extracted from T cells and bone marrow-derived APC. Expression of the constant β chain of the TCR by the T cells and class II I-Eβ chain by the APC were detected by RT-PCR. β-Actin primers were used as controls for each sample. BM-APC induce antigen-specific proliferation of autologous T cells Deer mice were immunized with keyhole limpet hemacyanin (KLH), and 10 days later the lymph nodes, spleens and bone marrow were processed for in vitro expansion of polyclonal T cells (lymph node cells), while the bone marrow cells and splenocytes were frozen. Sera were tested for antibodies to KLH by ELISA and in each deer mouse tested the titer was greater than or equal to 8,000 (data not shown). For recall proliferation, KLH, in vitro-propagated T cells (14 days with IL-2) and BM-APC (14 days with GM-CSF) or freshly thawed splenocytes were cultured together for 72 hours, and proliferation was assessed by MTS assay (Figure 5 ). For each deer mouse, the BM-APC were between 10× to 20× more efficient at stimulating antigen-specific helper T cell proliferative responses. While each T cell line exhibited a 50% maximal stimulation in the presence of about 10 to 20 μg/ml antigen with splenocytes, 50% maximal stimulation was usually near 1 to 2 μg/ml antigen with BM-APC. In parallel experiments, cultures incubated with 20 ng/ml house mouse TNF did not exhibit noticeably different proliferative responses compared to control cultures (data not shown), despite morphological evidence suggesting an effect on macropinocytosis (Figure 2 ). Figure 5 BM-APC stimulate helper T cell proliferation. Deer mice were immunized with 20 μg of KLH subcutaneously and 10 days later the lymph nodes, bone marrow and splenocytes were retrieved for expansion of helper T cells and BM-APC. After expansion of these cells in culture, proliferation assays were performed comparing mitomycin-C-treated autologous splenocytes (SC) and BM-APC (BMC) for their capacity to stimulate T cells. In each instance, the BM-APC were more effective at stimulating T cell responses. The data are of two deer mice (DM212 and DM213). Antibody augments BM-APC stimulation of T cells Deer mouse antiserum raised against KLH and incubated with antigen for one hour prior to addition of cells increased the sensitivity of T cell proliferation (Figure 6 ), suggesting that Fc receptors are present on the surface of the BM-APC. The presence of antibody increased the 50% maximal T cell proliferation from one deer mouse (DM223) about 10-fold, from about 500 ng/ml without antibody to 50 ng/ml with antibody. Similarly, for another deer mouse (DM224) the presence of antibody was substantially more effective at inducing T cell proliferation, increasing 50% max 50-fold (from 50 ng/ml to about 1 ng/ml). Notably, the T cell response in DM223 was also less vigorous (50% max 500 ng/ml) without antibody compared to DM224 (50% max 50 ng/ml). The proliferative responses of six other deer mouse T cell lines were similar to those of these deer mice (data not shown). Figure 6 Antigen-specific antibody augments BM-APC-induced T cell proliferation. T cell proliferation responses from deer mice 223 and 224 were assessed as described in Figure 5 using BM-APC. KLH-specific antiserum or normal deer mouse serum were diluted 1:2,000 in DMM-5 and incubated with dilutions of KLH for 1 hour in 96-well plates at room temperature. BM-APC and T cells were added to the wells and incubated 72 hours, and proliferation was assessed by MTS assay. Discussion To our knowledge, no previous efforts have been made to develop long-term cultures of T cells from unconventional laboratory rodents. The principal reason for this is that highly inbred strains, required for conventional long-term T cell work, are not available from rodents not routinely used in laboratory work. At least for deer mice, we have developed a method of fulfilling this need by using commercially-available house mouse GM-CSF. This cytokine apparently binds to the GM-CSF receptor on deer mouse cells such that it generates competent APC from the bone marrow. These cells are capable of processing and presenting soluble antigen to autologous antigen-specific helper T cells. Our initial suspicions that house mouse GM-CSF might bind to deer mouse GM-CSF receptor was the result of previous work [ 30 ] in which we cloned a partial cDNA of deer mouse GM-CSF, including most of its A helix that is involved in binding to the β chain subunit of the receptor. We used 5' RACE to obtain the complete N-terminus and found that all but two of the residues from helix A are identical between the two species. Subsequent experiments demonstrated that GM-CSF induces proliferation of deer mouse bone marrow cells, and since GM-CSF is routinely used to generate APC from the bone marrow we hypothesized that it would do so with deer mouse bone marrow. We used a method that has been shown to generate dendritic cells in house mice; however, the cells obtained from deer mouse bone marrow more closely resembled macrophages rather than DC. These cells contained many large macropinocytic vesicles, but conspicuous dendrites typical of DC were not observed. Microscopically, these cells also appeared sensitive to TNF, which decreased macropinocytosis, but it had no effect on the capacity of these cells to present antigen to T cells as has been reported for human DC derived from blood mononuclear cells [ 51 ]. TNF can induce a physiologic change in the APC from an active pinocytotic cell into one that becomes highly efficient at MHC class II antigen presentation, thus facilitating transition from the innate phase to the adaptive phase of the immune response. It is unclear why TNF treatment does not augment T cell proliferative responses with deer mouse APC, but it may be that species-specific differences in GM-CSF and TNF signaling occur that account for these disparities in APC development and behavior. It is also possible that TNF does not induce complete maturation of the cells into highly efficient APC, as has been reported for some DC [ 52 ]. It is currently impossible to phenotype these cells because no antibodies specific to deer mouse APC subpopulations, such as CD markers, are available, nor have genes for these markers been cloned, despite many attempts (unpublished observations), that might facilitate identification of these cells. Regardless, the cells are highly efficient at processing and presenting antigen, and inclusion of antigen-specific antibodies augments these functions. Since the deer mice are outbred, this method requires the immunization and collection of cells from individual animals (Figure 7 ). These cells are derived from lymph nodes (T cell source), splenocytes (APC source) and bone marrow (APC source). Most of the recovered cells can be propagated in vitro and/or aliquotted and stored frozen so that viable cells can be used as necessary to propagate and characterize helper T cell lines. We routinely recover 10 7 bone marrow cells from a deer mouse, which is sufficient for freezing 5 vials at 2 × 10 6 cells each. Each vial is used to seed a 100 mm bacterial Petri dish, which produces about 10 7 BM-APC at 14 days of culture. For deer mice, the most significant limitation for cells is from the spleen. Although deer mice are slightly smaller than BALB/c mice, their spleens are disproportionately small (unpublished observations). We routinely recover 7 × 10 6 splenocytes from a deer mouse, while BALB/c house mice usually provide 10-fold more. Because of this limitation, we have begun to use BM-APC to propagate T cells. This method involves culturing of bone marrow cells with GM-CSF for 10 days, then freezing aliquots of 10 6 cells. Three days before T cell restimulation, the 10-day BM-APC are thawed and cultured with GM-CSF, then used for restimulation with fresh antigen in one well of a 24-well plate. The T cells are fed fresh IL-2 DMM-5 at two-day intervals for expansion. Figure 7 Schematic overview of the process for culturing autologous deer mouse BM-APC and T cells. Ten days post immunization, lymph nodes, spleens and bone marrow are harvested from euthanized deer mice. The lymph node cells are cultured with antigen for four days, while the splenocytes and bone marrow cells are frozen at -70°C. On day 4, the blasting lymph node T cells are recovered and cultured with fresh antigen and thawed splenocytes (without mitomycin-C treatment). Simultaneously, bone marrow cells are thawed and cultured with GM-CSF. Expansion of the T cells is performed with huIL-2 and the BM-APC with GM-CSF for 14 days. These cells are then used for proliferation experiments. We have used this method to establish nine T cell lines, six specific for KLH and three specific for SNV nucleocapsid antigen (data not shown). Based upon typical cell yields, it should be possible to assay several thousand wells on 96-well plates, which we estimate to be sufficient for many T cell activities, including cloning, peptide epitope mapping, TCR variable gene segment usage, and cytokine profiling. We believe the methods described in this work will allow the characterization of antigen presentation and T cell responses in infected deer mice. Many viruses impair pathways involved in APC and T cell functions so that they can evade a sterilizing immune response. With hantaviruses and their rodent hosts, millions of years of evolution have presumably allowed a coadaptation of the viruses and host immune responses such that pathology does not occur and the virus is not eliminated. It is possible that hantaviruses possess some as yet unidentified mechanism for suppressing an aggressive inflammatory immune response in rodent hosts, which is ineffective in human infections and often leads to inflammatory immunopathology. Because of the substantial evolutionary divergence of deer mice and house mice (about 25–50 million years) [ 41 ], it is likely that this method could be used with many divergent rodent species, and thus useful for examining APC and T cell responses in a variety of rodent systems, including natural hosts and animal models of disease. A limitation of this approach is that BM-APC lose their ability to proliferate in the presence of GM-CSF after four to six weeks, similar to what has been observed with house mouse bone marrow-derived cells [ 53 , 54 ]. Since a finite number of bone marrow cells can be harvested, this limitation prevents indefinite propagation of deer mouse T cells. It is possible that competent BM-APC might be propagated indefinitely by the introduction of oncogenes, such as transforming retroviruses [ 55 ]. Alternatively, other house mouse or human hematopoietic cytokines that are commonly used to propagate bone marrow progenitor cells may bind to deer mouse receptors. For example, human or house mouse Flt3 ligand is active on both human and house mouse cells, suggesting that one or both would have an effect on deer mouse cells as well. In this manner, large numbers of progenitor cells might be produced in vitro for storage and then thawed, as needed, for culturing in GM-CSF to produce functional APC. Conclusions We have developed a method for generating large numbers of competent antigen presenting cells from deer mouse bone marrow using house mouse GM-CSF. This method resulted in the production of antigen-specific T cell lines from outbred deer mice. Inclusion of antigen-specific antibody in cultures augments T cell proliferation, suggesting the APC express Fc receptors. This method will allow characterization of APC and T cells in deer mice and may be extended to other rodent species that are important in infectious disease research. Methods Deer mice The deer mice used in these experiments were from a colony of animals established with deer mice trapped in western Colorado [ 30 ]. All procedures were approved by the Mesa State College Institutional Animal Care and Use Committee and in accordance with the Animal Welfare Act. Cloning of the deer mouse GM-CSF 5' cDNA The 5' end of the deer mouse GM-CSF cDNA was obtained by using RACE. Briefly, a primer (Table 1 ) was designed from a partial cDNA of deer mouse GM-CSF [ 30 ] and used to amplify the 5' end according to manufacturer's directions (SMART RACE, BD CLONTech, Palo Alto, CA) using con A-activated spleen cell culture cDNA. The fragment was cloned into pGEM-T Easy (Promega, Madison, WI) and sequenced using the Big Dye Terminator sequencing kit (Applied Biosystems, Foster City, CA) and an ABI 310 DNA Analyzer, and the sequence was deposited into Genbank (AY247762). The signal peptide and receptor-binding domain was identified by comparison to house mouse GM-CSF [ 49 ]. GM-CSF polypeptides from the house mouse (X02333), cotton rat ( Sigmodon hispidus , AAL55394) and human (NP_000749) were aligned using MacVector's (Accelrys, San Diego, CA) clustal algorithm. Table 1 Primers used in this work 1 Gene Forward Reverse Size (bp) I-Eβ GTC ATT TCT ACA ACG GGA CG TCT CCG CTG CAC AAT GAA GC 242 TCRβC AGG ACC TGA GCA AGG TGA GC GCA CAG CAT ACA GGG TGG CC 474 β-Actin ATG TAC GTA GCC ATC CAG GC TCT TGC TCG AAG TCT AGG GC 283 GM-CSF 5' RACE N/A GTT GCC CCG TAG GCC CTT CTC ATA TAA CT 273 1 All sequences are listed 5' to 3'. Cloning of the deer mouse T cell receptor β constant domain cDNA Total RNA from activated splenocytes was reverse-transcribed using an oligo-dT primer and Superscript II (Invitrogen, Carlsbad, CA). TCRβC cDNA sequences from house mouse, rat and human were aligned with MacVector. PCR primers (Table 1 ) were designed from highly conserved regions within the alignment. PCR was performed on activated splenocytes with 95°C for 30 sec, 58°C for 30 sec, and 72°C for 1 min for 35 cycles. The amplified fragment was cloned and sequenced as described above, and deposited into Genbank (AY307417). Immunization of deer mice Deer mice were bilaterally immunized subcutaneously at the base of the tail with 20 μg of KLH (Sigma Chemical Co, St. Louis, MO) emulsified in CFA (Sigma). Ten days later, draining lymph nodes, spleens and bone marrow were recovered for in vitro experiments. For production of high-titer KLH antiserum, deer mice were immunized i.p. with 20 μg of KLH emulsified in CFA and boosted with 20 μg in IFA one month later. Sera were collected 7 days after boosting. Processing of tissues from immunized deer mice Immunized deer mice were euthanized by cervical dislocation and the draining lymph nodes, spleens and bone marrow from individual animals were separately collected in Hank's balanced salt solution for processing. The lymph nodes served as a source of antigen-specific T cells, while the splenocytes were treated with ammonium chloride (Cambrex Bioproducts, Walkersville, MD) to lyse RBCs, then frozen in 10% DMSO/5% FBS deer mouse medium (DMM-5: 5% FBS, RPMI-1640 supplemented to 315 mOsm [with 2.5 ml of 5 M NaCl/L], 2.5 μg/ml Fungizone, 100 U/ml penicillin, 100 μg/ml streptomycin, 50 μg/ml gentamicin, 50 μM β-ME, 10 mM HEPES, 2 mM L-glutamine) in aliquots for use as autologous APC for additional rounds of in vitro T cell stimulation. The bone marrow cells were collected from tibiae and femurs, washed twice in DMM-5, and aliquotted at 2 × 10 6 cells per vial in 1 ml of 10% DMSO/DMM-5 and stored at -70°C. Bone marrow culture Bone marrow-derived APC (BM-APC) were generated with modification of a previously described method for dendritic cells (DC) [ 39 ]. One vial of bone marrow cells (2 × 10 6 ) was quick-thawed in a 37°C water bath and cultured without washing in 100 mm bacterial Petri dishes in DMM-10 containing 20 ng/ml recombinant house mouse GM-CSF (R&D Systems, Minneapolis, MN) at 37°C under 7% CO 2 . Fresh GM-CSF/DMM-10 was provided on days 3, 6, 8, 10, and 12 for the generation of APC. Cells were collected with a scraper for use in experiments. Cells were processed by cytospin and stained with Wright's stain for morphological examination. Assessment of T cell sensitivity to human IL-2 Deer mouse splenocytes depleted of RBC by ammonium chloride treatment were incubated with a suboptimal dose of PHA (2 μg/ml) (Sigma) in DMM-5 and recombinant human IL-2 (R&D Systems) for 48 hours. Proliferation was determined by MTS assay (Cell Titer-96 AQ, Promega). The means and standard deviations of duplicate samples were calculated, with the mean of cells without IL-2 subtracted from sample means. Assessment of T cell receptor and MHC class II expression Total RNA was extracted from 14 day T cell and BM-APC cultures (Versagene RNA, Gentra Systems, Minneapolis, MN) and converted into cDNA. Class II expression of the bone marrow cells was assessed by PCR using a forward primer from exon 2 and a reverse primer that overlaps the boundaries of exons 2 and 3 of deer mouse I-Eβ (Table 1 [ 50 ]). The amplified fragment was cloned and sequenced as described above. T cells were defined by PCR amplification of the TCRβC chain using the primers listed above (forward, exon 1; reverse, exon 2). β-Actin expression was assessed for each population. ELISA serology Sera were collected at euthanasia by cardiac puncture. Plates were coated with 5 μg/ml KLH in PBS overnight and washed 5× with wash buffer (PBS-0.1% TWEEN-20). Plates were then blocked with blocking buffer (5% nonfat powdered milk in wash buffer) for 1 hour at room temperature. The sera and remaining reagents were diluted in blocking buffer. Sera were incubated in duplicate for 2 hours at room temperature, followed by goat anti- P. leucopus IgG (H&L) (KPL, Gaithersburg, MD) for 1 hour, then horse anti-goat IgG-HRP conjugate for 1 hour (Vector, Burlingame, CA). ABTS substrate (Sigma) was incubated for 15 min, and plates were read at 414 nm. Means were calculated with the background (1:100 normal deer mouse serum) subtracted. In vitro helper T cell expansion In vitro stimulation of helper T cells was performed essentially as described elsewhere [ 56 , 57 ]. Lymph nodes from immunized deer mice were made into single-cell suspensions by gently disrupting the capsule between the ends of sterile frosted microscope slides. The cells were washed twice in HBSS and plated at 5 × 10 6 cells per well (24 well plate) with 20 μg/ml KLH in DMM-5. After a 4 day incubation, the lymph node cells were collected and washed twice in DMM-5. The number of recovered lymph node cells varied between animals, but between 2 × 10 5 and 5 × 10 5 cells were recovered and plated with fresh antigen and 3 × 10 6 thawed autologous splenocytes in DMM-5 in 24 well plates in 1 ml of DMM-5. At 2 day intervals, cultures were fed by removing 750 μl of media and replacing it with DMM-5 containing 20 U/ml of recombinant human IL-2. When cultures were greater than 80% confluent, cells were passaged 1:2 into additional wells or into T25 tissue culture flasks. This process was continued for 14 days to expand T cells. Cultures were also restimulated at two week intervals with fresh mitomycin-C treated autologous splenocytes as above, or 2 × 10 6 BM-APC to continue propagation of T cell lines. Functional assessment of bone marrow-derived APC BM-APC were examined for functional capacity to process and present KLH to sensitized autologous deer mouse T cells expanded in culture. For these experiments, mitomycin-C treated BM-APC (10 4 ) or splenocytes (2 × 10 5 ) were used to stimulate 10 5 T cells in the presence of KLH in 96-well plates. In other experiments, the effects of house mouse tumor necrosis factor (20 ng/ml) were evaluated on APC morphology and capacity to stimulate T cell proliferation. Lastly, antisera to KLH were produced in deer mice and used to assess the capacity of the BM-APC to capture and process antigen for presentation to T cells. In these experiments, antiserum or normal deer mouse serum were diluted to 1:2,000, a saturating dilution in ELISA as described above, with KLH and incubated for 1 hour at room temperature. BM-APC and T cells were then added and the cultures incubated for 72 hours prior to determining proliferative responses by MTS. Means and standard deviations were calculated from duplicate samples. List of abbreviations HFRS, hemorrhagic fever with renal syndrome; HCPS, hantavirus cardiopulmonary syndrome; APC, antigen presenting cell; SNV, Sin Nombre virus; GM-CSF, house mouse GM-CSF; DC, dendritic cell; CFA, complete Freund's adjuvant; RT-PCR, reverse transcription polymerase chain reaction; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium Authors' contributions BD conducted bone marrow cell culture work. DGW and TAC performed RT-PCR experiments. JP cloned and sequenced the TCRβ cDNA. RMF cloned and sequenced the MHC class II cDNA. TS immunized deer mice and generated T cell lines.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524361.xml

548278

CASCAD: a database of annotated candidate single nucleotide polymorphisms associated with expressed sequences

Background With the recent progress made in large-scale genome sequencing projects a vast amount of novel data is becoming available. A comparative sequence analysis, exploiting sequence information from various resources, can be used to uncover hidden information, such as genetic variation. Although there are enormous amounts of SNPs for a wide variety of organisms submitted to NCBI dbSNP and annotated in most genome assembly viewers like Ensembl and the UCSC Genome Browser, these platforms do not easily allow for extensive annotation and incorporation of experimental data supporting the polymorphism. However, such information is very important for selecting the most promising and useful candidate polymorphisms for use in experimental setups. Description The CASCAD database is designed for presentation and query of candidate SNPs that are retrieved by in silico mining of high-throughput sequencing data. Currently, the database provides collections of laboratory rat ( Rattus norvegicus ) and zebrafish ( Danio rerio ) candidate SNPs. The database stores detailed information about raw data supporting the candidate, extensive annotation and links to external databases (e.g. GenBank, Ensembl, UniGene, and LocusLink), verification information, and predictions of a potential effect for non-synonymous polymorphisms in coding regions. The CASCAD website allows search based on an arbitrary combination of 27 different parameters related to characteristics like candidate SNP quality, genomic localization, and sequence data source or strain. In addition, the database can be queried with any custom nucleotide sequences of interest. The interface is crosslinked to other public databases and tightly coupled with primer design and local genome assembly interfaces in order to facilitate experimental verification of candidates. Conclusions The CASCAD database discloses detailed information on rat and zebrafish candidate SNPs, including the raw data underlying its discovery. An advanced web-based search interface allows universal access to the database content and allows various queries supporting many types of research utilizing single nucleotide polymorphisms.

Background Single nucleotide polymorphisms (SNPs) are the most common form of genetic variation within species. As a result, SNPs are now becoming the most popular type of marker in genetic association and mapping studies. SNPs are also most likely to be the molecular basis for the majority of phenotypic variation in (outbred) populations. In particular, SNPs in regulatory and protein-coding regions can have an effect on gene expression levels and protein activity, respectively. The phenotypic differences observed between selected (sub) strains in model organisms may be the result of specific (combinations of) natural occurring polymorphisms. Hence, a comprehensive inventory of SNPs, including extensive annotation will be extremely valuable in the search for functional polymorphisms. There is often a vast unexplored potential in large sequence datasets that have been collected for other purposes, for example, EST and whole genome sequencing (WGS) projects. In an effort to address these two issues, we have developed an in silico candidate SNP mining pipeline that uses all publicly available sequence data for a specific organism, and designed a database, CASCAD (CAscad SNP CAndidates Database), that allows storage of a wide variety of primary source data, cross-annotation to other databases, and analysis parameters for SNPs associated with expressed sequences. Construction and content We applied the SNP discovery pipeline to both rat [ 1 ] and zebrafish (unpublished results) and identified about 33,000 and 52,000 high-quality candidates, respectively, that were extensively annotated and stored in the CASCAD database (Table 1 ). The database includes detailed primary information on which the discovery of the candidate SNPs was based. This information, including sequence quality information (Phred score), the number of supporting reads for every nucleotide observed at the SNP position, and expected alignment lengths, was found to be very valuable for filtering for predicted variants that have the highest likelihood to be experimentally confirmed. Two verification experiments for the rat resulted in confirmation rate estimates of 59% (68 candidate SNPs in 10 different laboratory rat strains) and 50.3 % (340 candidates in 5 laboratory and 2 wild rat isolates) [ 1 ]. A set of 139 CASCAD entries tested in 7 zebrafish isolates confirmed 67.6% of them as true polymorphisms (unpublished data). The success rate values obtained are likely to be underestimates since only limited number of isolates/samples were used and we were unable to include exactly the same isolates that were used for generating the primary data (e.g. EST sequencing). Table 1 Input data (number of sequence reads) for the CASCAD pipeline and number of predicted candidate SNPs. Rattus norvegicus Danio rerio Input data mRNA 25, 634 3, 366 EST 244, 518 283, 572 WGS 19, 813, 313 11, 588, 394 Candidate SNPs predicted 33, 305 51, 769 synonymous 3, 842 9, 111 non-synonymous 3, 708 6, 217 nonsense 162 158 We designed a web-based interface with Perl scripts communicating to a MySQL database, and displaying HTML pages through Apache server running on SuSE Linux. The interface provides simple, advanced (Figure 1 ), and sequence-based search forms. Parameters that can be used in a search include strain information, different formats of sequence identifiers (e.g. GenBank, UniGene, LocusLink accessions or gene symbol), map positions (genetic and physical), and a wide variety of SNP characteristics, such as experimental evidence, a likelihood score for verification as deduced from extensive verification experiments [ 1 ], and information regarding restriction sites that have been affected, facilitating the design of RFLP based assays. To this end CASCAD is tightly linked to primer design [ 2 ] and local genome assembly [ 3 ] interfaces, enabling a fast, reliable, and universal primer design for a chosen SNP candidate even when there is no assembled genome sequence available. To facilitate the retrieval of candidate SNPs with higher verification rates, represented by more reliable and common SNP candidates, we implemented the possibility to restrict searches to entries characterized by location at hypervariable CpG site, by minimal basecalling quality (Phred score), sequence match size, and/or minimal number of supporting reads for either allele. Figure 1 CASCAD advanced search form. In addition to primary sequence data analysis, the effect of all SNPs on protein coding capacity was evaluated and non-synonymous SNPs were categorized in classes reflecting the severity of the polymorphism using a BLOSUM-based score. The predicted missense SNPs were analyzed by SIFT [ 4 ] and Polyphen [ 5 ] programs that utilize not only substitution information but also phylogenetic conservation and structural protein information to predict a potential effect of the polymorphism on protein function. Query results are summarized on the SNP details page (Figure 2 ), listing the SNP characteristics and including active links to other databases and resources, such as dbSNP, Ensembl, UniGene, and LocusLink. More detailed information regarding raw data underlying the candidate SNP (links to the original sequence files and a full nucleotide and protein alignment) can be obtained by clicking on the observed nearly exact hits between nucleotide sequences. Moreover, statistics on the data that support the SNP (number of occurrences for every nucleotide at SNP position, range of Phred basecalling quality scores) are provided. Figure 2 SNP details page Utility and discussion For many applications, it is important to be able to distinguish between SNP candidates by their characteristics, as they may be predictive for verification success rate or carry biologically relevant information. Non-confirmed candidate polymorphisms may represent variants uncommon for a given population, but also sequencing errors (all types of sequences), RNA editing events and reverse transcriptase errors (EST reads). In order to minimize the contribution of false positives, one can exclude polymorphisms based on a single read for either allele, as is common for many in silico discovery pipelines [ 6 ]. Although this is a valid approach when selecting SNPs for population or association genetics, one could inadvertently discard many rare variants that may be associated with phenotypic variation, for example by affecting protein structure or function. Information on such polymorphisms can be very useful when mapping disease or QTL alleles. We have developed our database to fulfill the needs of any particular SNP application by providing control over every parameter we used in the polymorphism discovery step. Applications of the CASCAD database include queries for potentially deleterious SNPs in a specific genomic region of interest, for example a QTL interval, design of SNP-based mapping panels using either RFLP or any other technology, and identification of informative SNPs for fine-mapping. Custom sequences can be provided to search for known SNPs in any sequence of interest. In addition, the CASCAD pipeline [ 1 ] can be used to build a candidate SNP database for any model organism of interest for which sufficient sequencing data is available. Conclusions The main purpose of CASCAD database is to provide flexible access to candidate single nucleotide polymorphisms, which were predicted using a computational approach from publicly available sequence data of the rat and zebrafish. The resulting database is crosslinked to most common public databases and can be queried for SNPs using accession numbers, sequence context, SNP characteristics, but also using parameters specific to the SNP discovery process, allowing stringent or relaxed conditions suitable for different types of applications. Availability and requirements The database is freely accessible through the website . Programs, scripts, MySQL database dumps, and instructions for setting up a species-specific SNP database can be obtained from the authors upon request. Authors' contributions VG designed and implemented the CASCAD database. EB tested database and interface. EC provided supervision and guidance for the project.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548278.xml

423156

Taking the Stem Cell Debate to the Public

In response to the Blackburn and Rowley essay on the President's Council on Bioethics, several thought-provoking opinions on ethical challenges in biomedical research are expressed by prominent stakeholders

In their essay in the April 2004 issue of PLoS Biology , Elizabeth Blackburn and Janet Rowley (2004) , two distinguished cellular biologists and members of the President's Council on Bioethics, strongly question the scientific foundation of two reports from the Council ( President's Council on Bioethics 2003 , 2004 ). The Council on Bioethics was formed by executive order “to advise the President on bioethical issues that may emerge as a consequence of advances in biomedical science and technology.” An open discussion between ethicists and scientists is critical to the advisory system. The recent administrative dismissal of Dr. Blackburn from the Council is very alarming. By stacking the deck with conservative opinions, and not accurately discussing the scientific issues, the Bioethics Council has become irrelevant to the scientific community and presents a jaundiced view to the public. Stem cell research and its applications have the potential to revolutionize human health care. Recent polls show support for embryonic stem cell research, even with conservative voters. The public, as the major benefactor of biomedical research and the target population of beneficial clinical advances, has the right to a fact-based discussion of the science regarding stem cells. It is therefore time that the debate on stem cell research, with its risks and benefits, be taken to the public. A debate on stem cell research restricted to the President's Council on Bioethics is a disservice to the public. Nearly three decades ago, the advent of recombinant DNA technology and in vitro fertilization (IVF) techniques, raised similar concerns regarding research. Contrary to apprehensive expectations, recombinant DNA technology has boosted enormous advances in the health care and pharmaceutical industry. IVF evolved to be a widely accepted, safe medical procedure, with over one million healthy babies born by IVF and related treatments. Similarly, once stem cells are successfully used in the clinic, most of today's political and ethical issues will evaporate. The International Society for Stem Cell Research (ISSCR), a society whose membership encompasses the bulk of the stem cell research brain trust, holds the position that research on both adult and embryonic stem cells will guarantee the fastest progress in scientific discovery and clinical advances. The ISSCR also strongly opposes reproductive cloning and supports the National Academy of Science's proposal to develop voluntary guidelines to encourage responsible practices in human embryonic stem cell research. One of the original recommendations of the President's Council on Bioethics was a four-year moratorium on stem cell research. The purpose of this moratorium was theoretically to open a large, national discourse on the topic of stem cell research, a debate intended to bring all sides into thoughtful reflection on the issue. To that end, the ISSCR has repeatedly and consistently offered an open forum for all sides in the debate at our conferences, and has carefully offered invitations to join our society and to speak at our annual meeting to members of the President's Council, including colleagues whose opposition to stem cell research has been clear. None have accepted. Dr. Kass, in particular, has received several direct appeals but has turned down every such opportunity to make his case to the researchers who arguably are his discourse partners, from whom he could learn much, and whom he should be actively engaged in teaching. It is tragic that voices of dissent and debate are stilled, for it is this very quality of open debate that is at the heart of both the scientific method and an ethically directed American democracy—surely a goal that we all share.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC423156.xml

539330

Relaxed Molecular Clock Provides Evidence for Long-Distance Dispersal of Nothofagus (Southern Beech)

Nothofagus (southern beech), with an 80-million-year-old fossil record, has become iconic as a plant genus whose ancient Gondwanan relationships reach back into the Cretaceous era. Closely associated with Wegener's theory of “Kontinentaldrift”, Nothofagus has been regarded as the “key genus in plant biogeography”. This paradigm has the New Zealand species as passengers on a Moa's Ark that rafted away from other landmasses following the breakup of Gondwana. An alternative explanation for the current transoceanic distribution of species seems almost inconceivable given that Nothofagus seeds are generally thought to be poorly suited for dispersal across large distances or oceans. Here we test the Moa's Ark hypothesis using relaxed molecular clock methods in the analysis of a 7.2-kb fragment of the chloroplast genome. Our analyses provide the first unequivocal molecular clock evidence that, whilst some Nothofagus transoceanic distributions are consistent with vicariance, trans-Tasman Sea distributions can only be explained by long-distance dispersal. Thus, our analyses support the interpretation of an absence of Lophozonia and Fuscospora pollen types in the New Zealand Cretaceous fossil record as evidence for Tertiary dispersals of Nothofagus to New Zealand. Our findings contradict those from recent cladistic analyses of biogeographic data that have concluded transoceanic Nothofagus distributions can only be explained by vicariance events and subsequent extinction. They indicate that the biogeographic history of Nothofagus is more complex than envisaged under opposing polarised views expressed in the ongoing controversy over the relevance of dispersal and vicariance for explaining plant biodiversity. They provide motivation and justification for developing more complex hypotheses that seek to explain the origins of Southern Hemisphere biota.

Introduction An important principle of evolutionary inference is that explanations for the past require an understanding of mechanisms and processes applicable in the present [ 1 ]. It is perhaps this sticking point more than any other that has polarised views over the relative importance of vicariance and dispersal for explaining extant plant biodiversity. In 1915, Alfred Wegener put forward a testable hypothesis and mechanism that could explain the transoceanic distribution of animal and plant species. In the 21st century, with many DNA studies now implicating the importance of long-distance dispersal for explaining plant biodiversity [ 2 , 3 , 4 , 5 ], it is disconcerting that there is currently a very poor understanding of the mechanisms of transoceanic dispersal (but see [ 6 , 7 , 8 , 9 , 10 ]). Indeed, the inference that the seeds of extant Nothofagus species are not suited for dispersal across large distances has played a major role in motivating the hypothesis that transoceanic distributions of Nothofagus ( Figure 1 ) can only be explained by vicariance [ 11 , 12 , 13 , 14 , 15 ]. This hypothesis posits that following the Cretaceous breakup of Gondwana, Nothofagus rafted and evolved in situ upon different Southern Hemisphere lands. Whilst very attractive, this hypothesis fits somewhat uncomfortably with the findings from analyses of morphological and molecular data. In particular, whilst earlier molecular data have been insufficient for rigorous molecular clock analyses, their interpretation has favoured hypotheses of transoceanic dispersal [ 16 , 17 , 18 ]. Figure 1 Southern Hemisphere Maps and Present-Day Nothofagus Distribution (A) Transoceanic distribution of Nothofagus subspecies Lophozonia and Fuscospora and South American species N. nitida (subgenus Nothofagus ). Map adapted from Swenson et al. [ 43 ]. ASE, Australia; NCA, New Caledonia; NGU, New Guinea; NZE, New Zealand; SAM, South America; TAS, Tasmania. (B) Relationship of Australia, New Zealand, and South America 65 Myr and 35 Myr before present, reconstructed from http://www.odsn.de/ (link “Plate Tectonic Reconstructions”). Based on the sequence of Gondwana breakup, a hypothesis of vicariance most parsimoniously predicts that Australian Nothofagus species should be most closely related to South American species. This follows since South America and Australia were connected via Antarctica until approximately 35 million years (Myr) ago ( Figure 1 ). In contrast, New Zealand is thought to have separated from Australia 80 Myr ago [ 19 , 20 ]. Thus to explain the close relationship between Australian and New Zealand species by vicariance, it is necessary to argue that extinction of Australian and/or closely related South American species has occurred [ 12 ]. Whilst this explanation is ad hoc, the fossil record does provide evidence for numerous Nothofagus extinctions in Australia, South America, and New Zealand [ 21 , 22 , 23 ]. However, the fossil record has also been interpreted as indicating multiple events of transoceanic dispersal of Nothofagus from Australia to New Zealand. Whilst the extinct “ancestral” Nothofagus pollen type occurred in New Zealand prior to the breakup of Gondwana, Fuscospora pollen first appeared in New Zealand during the Palaeocene (65 Myr ago) and Lophozonia pollen first appeared during the late Eocene (50 Myr ago; [ 24 ]). Sixty-five Myr ago the Tasman Sea had already reached its present-day size [ 19 , 20 ]. Hence it is possible that extant New Zealand Nothofagus subgenera did not have the opportunity to reach New Zealand via overland migration. Hill [ 25 ] has also described the species Nothofagus cethanica, which first appeared in Oligocene macrofossils from Tasmania. This species shares unique features with extant N. fusca and N. truncata from New Zealand and may share a sister relationship with these species explained by trans-Tasman Sea dispersal [ 26 ]. A contribution to the debate over the relative importance of vicariance and dispersal can be made by estimating the divergence times of extant species. However, DNA sequences of insufficient length have prevented robust molecular clock analyses from being undertaken. For this reason, we report the sequencing of a 7.2-kb chloroplast genome fragment covering the gene regions ( trnL–trnF and atpB–psaI; see Table 1 for accession numbers) for 11 species of three Nothofagus subgenera ( Lophozonia, Fuscospora, and Nothofagus ). Our aim has been to date divergence of extant species in the subgenera Lophozonia and Fuscospora. We have carried out relaxed molecular clock analyses using the methods of Sanderson [ 27 , 28 ] and Thorne et al. [ 29 ]. Our findings are that, whilst vicariance is likely to explain some transoceanic relationships amongst Nothofagus species, phylogenetic relationships between trans-Tasman species in both Lophozonia and Fuscospora can only be explained by mid- to late-Tertiary transoceanic dispersal. Table 1 Origin of Nothofagus Samples and Sequence Accession Numbers Results Figure 2 shows an optimal maximum-likelihood reconstruction of phylogenetic relationships for chloroplast DNA sequences (7.2-kb comprising the atpB–psaI region and the trnL–trnF region; 7,269 nucleotide sites) for Nothofagus (subgenera or pollen groups (a) Lophozonia, (b) Fuscospora, and (c) Nothofagus ) and outgroup Castanea sativa (not shown). In a sensitivity analysis of 60 substitution models, the tree shown in Figure 2 was always recovered with very little difference in branch lengths regardless of the substitution model used. Of all substitution models evaluated, K81uf+G was identified as the best fitting one for the data based on hierarchical likelihood ratio tests and the Akaike Information Criterion. This substitution model and also the F84+ Γ 8 model were used for further analyses. The latter was included because the Bayesian relaxed molecular clock (BRMC) approach as implemented in the program MULTIDIVTIME (see Materials and Methods ) only allows the use of the JC and the F84 models. Thus analysis with the F84+ Γ 8 model allowed a comparison of date estimates to be obtained using different relaxed molecular clock methods. All nodes of the optimal ML tree recovered in the sensitivity analysis received nonparametric bootstrap support greater than 97%, with the only exception being the grouping of N. cunninghamii with N. moorei, which received 72% support. Figure 2 ML Tree Indicating Evolutionary Relationships for Nothofagus Species Based on the atpB–psaI and trnL–trnF Region of the Chloroplast Genome (7,269 bp) Divergence dates (in Myr) were obtained with an F84+ Γ 8 substitution model using the BRMC approach of Thorne et al. [ 29 ]. For the dates indicated, the age of the root node and that of F/N1 were constrained to 70–80 Myr; L2 was also constrained in accordance with fossil data [ 26 ] at 20 Myr. Violet numbers show bootstrap values. The pollen grains represent the first appearance of the respective pollen type in the New Zealand fossil record. Plio, Pliocene; Oligo, Oligocene; Palaeo, Palaeocene; Ma, Maastrichian; Campan, Campanian. L1–L4, Lophozonia 1–4; F1–F2, Fuscospora 1–2; F/N1, Fuscospora / Nothofagus 1. Divergence times for the nodes in this tree ( Figure 2 ) were estimated using the penalized likelihood (PL) method [ 27 ] and BRMC method [ 29 , 30 , 31 ]. For these analyses, a period of 70–80 Myr was used to calibrate the divergence between the three fossil pollen groups representing subgenera Lophozonia, Nothofagus, and Fuscospora. These four pollen groups all first appeared in the fossil record approximately 75 Myr ago [ 32 ]. A second constraint of a minimum of 20 Myr for the divergence of N. cunninghamii and N. moorei was also used. This constraint was based on observations reported by Hill [ 26 ] that 20-Myr-old fossils intermediate between N. moorei and N. cunninghamii were recorded from Tasmania and that fossils closely resembling N. moorei were also present at that time. The inferred ages for the remaining nodes of the tree, obtained under the F84+ Γ 8 model of substitution are given in Table 2 and graphically illustrated on Figure 2 . The variance on these estimates was low and the values were little influenced by the choice of substitution model ( Table 3 ). The robustness of the estimates to calibration error was tested by constraining the divergence of Australian and New Zealand sister taxa to 65 Myr (the time before present when the Tasman Sea reached its present position; thus this date provided us with a lower bound for divergence times of trans-Tasman Nothofagus disjunctions that might be explained by vicariance). Constraining these two nodes in this way produced unrealistic age estimates for all basal nodes. For example, using the BRMC method, which additionally required a prior expectation to be specified for the age of the root node (which we specified at 75 Myr—the time of appearance of all four extant pollen types), we estimated a more likely age for the root node at 191 Myr. For the PL approach, which does not require specification of a prior, we estimated the age of the root node at 634 Myr. Other basal nodes in both the Fuscospora and Lophozonia lineages were also much older than reasonably expected (see Table 2 ). Table 2 Estimated Divergence Dates and Standard Deviations (in Brackets) of Different Nothofagus Clades The numbers in bold are all the nodes that were estimated without constraints Dates are based on different calibration dates and estimation approaches and are given in Myr before present a Node fixed b Node constrained DOI: 10.1371/journal.pbio.0030014.t002 Table 3 Variation of Estimated Divergence Times (in Myr) under 60 Symmetrical Models of DNA Substitution Dates estimated using PL approach a Node constrained Discussion Our findings from molecular clock analyses using five independent calibrations (for four nodes), suggest that the sister relationships of the Australasian (Australia and New Zealand) species within both Lophozonia and Fuscospora lineages are too young to be explained by continental drift (as indicated by the inferred ages of nodes F1 and L3). Transoceanic dispersal appears the most likely explanation for the trans-Tasman sister relationships indicated in Figures 1 and 2 . In contrast, the age inferred for node F2, using both relaxed clock methods is compatible with a hypothesis of continental drift as an explanation for the sister relationship between South American and Australasian Fuscospora lineages. The age for node L4, which separates Australasian and South American Lophozonia, may also be consistent with vicariance. The BRMC method dates it at 34 Myr before present. However, the PL method estimates this node to be only 25 Myr old, an age too recent to be consistent with vicariance. Thus we regard our results for node L4 as equivocal. Nevertheless, southern beeches are likely to have been present in Antarctica 25 myr ago [ 33 ], and thus long-distance dispersal across the young southern ocean between South America and Australia via Antarctica may be conceivable. The robustness of our phylogenetic inferences has been investigated by varying the substitution model (60 symmetric models were used), estimating the variance of age estimates, and evaluating the influence of calibrations on divergence times. With the exception of the root node, the PL method consistently gave more recent age estimates than did the BRMC method. Both methods showed sensitivity to the number of calibration points used, a finding consistent with recent observations on the performance of relaxed molecular clock methods [ 34 ]. In general, the date estimates produced by the BRMC approach were more consistent with the fossil record [ 26 ]. A relevant question is whether or not additional calibration points could make date estimates older and thus change our conclusion of trans-Tasman dispersal. We suggest that this may be unlikely, given the observation that constraining a minimum age for trans-Tasman sister species to 65 Myr leads to greatly inflated and unrealistic age estimates for all basal nodes. Hence to explain this finding we would need to invoke a further hypothesis of a dramatic and independent slowing in the rate of evolution in Lophozonia, Fuscospora, and Nothofagus lineages. Thus the hypothesis that present-day distribution patterns of Nothofagus can be explained by continental drift following the breakup of Gondwana and subsequent extinction of some species [ 24 ] can be rejected on the basis of the divergence dates that we have estimated. These dates also indicate that present-day Nothofagus species in New Zealand are not the direct descendants of the Fuscospora and Lophozonia southern beeches that reached New Zealand in the Palaeocene and Eocene eras, respectively [ 24 ]. This finding highlights the caution that needs to be taken when interpreting fossil evidence for the apparent first appearance of extant taxa. Fossils that identify specific evolutionary lineages may not necessarily indicate the origins for extant taxa or suggest a continuous presence for these taxa. Similar concerns follow from the findings of molecular analyses for Ascarina and Laurelia in New Zealand [ 2 , 4 ]. The strength of our molecular analyses highlights the importance of future research into potential mechanisms of long-distance dispersal, and in particular reinvestigation of the transoceanic dispersal properties of Nothofagus seeds. For the reasons that we outline in our introduction, it seems likely that only once the mechanisms of long-distance dispersal are understood will hypotheses based on DNA divergence time estimates be truly convincing. DNA sequence analyses have also suggested that long-distance dispersal and continental drift are both important for explaining distributions of the conifer Agathis (Araucariaceae) in the South Pacific [ 35 ]. Although the molecular evidence for Agathis is not as strong as it is for Nothofagus, the findings from the molecular studies on these genera highlight the importance of considering more complex hypotheses of relationship in debates concerning the relative importance of dispersal and vicariance. Materials and Methods Sequence data Chloroplast DNA sequences (7.2 kb comprising the atpB–psaI region and the trnL–trnF region) were determined for each of 11 accessions of Nothofagus (subgenera or pollen groups Lophozonia, Fuscospora, and Nothofagus ) sampled in South America, Australia, and New Zealand (see Table 1 ). These genome regions were also determined for C. sativa (an outgroup taxon from Fagaceae) and aligned using progressive multiple-sequence alignment: ClustalX version 1.81 [ 36 ]. This resulted in an unambiguous alignment of 7,269 nucleotide sites. Data are missing for approximately 250 bp of the atpB gene and atpB – rbcL intergene region of Nothofagus. Tree building Phylogenetic analyses were conducted with PAUP* version 4.0b10 [ 37 ] under the ML criterion. A model sensitivity test was conducted, investigating a range of 60 symmetrical models of DNA substitution corresponding to the 56 implemented in MODELTEST version 3.06 [ 38 ] ( http://darwin.uvigo.es/software/modeltest.html ) plus F84, F84+I, F84+Γ 8 , and F84+I+Γ 8 . ML parameters of these models were estimated by PAUP* following the approach used in MODELTEST. These parameters were then used to conduct 60 individual ML heuristic searches in PAUP* with tree bisection-reconnection branch swapping and a neighbour-joining starting tree. ML bootstrap proportions were obtained after 100 replications, using the same search strategy and ML parameters as for the analysis of the original dataset. Molecular dating: The PL method Divergence dates were obtained using the PL method of Sanderson [ 27 ] as implemented in the program r8s, version 1.60 [ 28 ] ( http://ginger.ucdavis.edu/r8s/ ) with the TN algorithm. The outgroup was excluded using the “prune” command. The degree of autocorrelation within lineages was estimated using cross-validation as suggested by Sanderson [ 27 ], and the correcting smoothing parameter λ defined accordingly. Divergence dates were estimated on the 60 ML phylograms recovered in the phylogenetic model sensitivity analysis. Ages for each node across the 60-ML trees were summarized using the “profile” command. Confidence limits on dating estimates were computed by using nonparametric bootstrapping of the original dataset as suggested by Sanderson and Doyle [ 39 ]. The program SEQBOOT of the PHYLIP 3.6 package [ 40 ] was used to generate 100 bootstrap resampled datasets of 7,269 sites in length. ML branch lengths of the optimal topology were then estimated under the F84+ Γ 8 model for each of the bootstrap resampled datasets using PAUP*. Divergence estimates were then calculated for each of the 100 bootstrap replicates using r8s to obtain standard deviations on each node by the “profile” command and the settings described above. Molecular dating: The BRMC method The BRMC approach was applied using the program MULTIDIVTIME as implemented in the Thornian Time Traveller (T3) package [ 41 ]. First, the program BASEML of the PAML package version 3.13 [ 42 ] ( http://abacus.gene.ucl.ac.uk/software/paml.html ) was used to estimate the ML parameters of the F84+ Γ 8 substitution model, using the ML topology previously identified. Second, the program ESTBNEW ( ftp://abacus.gene.ucl.ac.uk/pub/T3/ ) was used to estimate branch lengths of the ML topology and the corresponding variance–covariance matrix. Finally, the program MULTIDIVTIME was used to run a Markov chain Monte Carlo for estimating mean posterior divergence times on nodes with associated standard deviations from the variance–covariance matrix produced by ESTBNEW. The Markov chain was sampled 10,000 times every 100 cycles after a burn-in stage of 100,000 cycles. We used a 75-Myr (SD = 37.5 Myr) prior [ 32 ] for the expected number of time units between tip and root and a prior of 200 Myr for the highest possible number of time units between tip and root. Other priors for gamma distribution of the rate at root node and the Brownian motion constant describing the rate variation (i.e., the degree of rate autocorrelation along the descending branches of the tree) were derived from the median branch length. As for the PL method, the outgroup was not included in this analysis. Supporting Information Accession Numbers The GenBank ( http://www.ncbi.nlm.nih.gov/ ) accession numbers for the sequences discussed in this paper are given in Table 1 .

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539330.xml

527873

Gastric T-cell lymphoma associated with hemophagocytic syndrome

Background Lymphoma-associated hemophagocytic syndrome (LAHS) occurs in mostly extra nodal non-Hodgkin's lymphoma. LAHS arising from gastrointestinal lymphoma has never been reported. Here we report a case of gastric T-cell lymphoma-associated hemophagocytic syndrome. Case presentation A 51-year-old woman presented with pain, redness of breasts, fever and hematemesis. Hematological examination revealed anemia. Gastroscopy revealed small bleeding ulcers in the stomach and the computed tomography scan showed liver tumor. She underwent total gastrectomy for gastrointestinal bleeding and the histopathology revealed gastric T-cell lymphoma. She continued to bleed from the anastomosis and died on the 8th postoperative day. Autopsy revealed it to be a LAHS. Conclusions If Hemophagocytic syndrome (HPS) occurs in lymphoma of the gastrointestinal tract, bleeding from the primary lesion might be uncontrollable. Early diagnosis and appropriate treatment are needed for long-term survival.

Background Hemophagocytic syndrome (HPS) in adults is characterized by reactive and systemic proliferation of benign histiocytes that phagocytose blood cells [ 1 ]. It is often associated with infections, malignant neoplasms, autoimmune diseases and various immunodeficiencies. Lymphoma-associated hemophagocytic syndrome (LAHS) mostly occurs from extra nodal lymphoma and is known to have a poor prognosis. Here we report a case of LAHS arising from gastric lymphoma with a fulminant clinical course and difficult diagnosis until the time of autopsy. Case presentation A 51-year-old female was admitted on May 9, 1995, because of severe hematemesis. The patient had been treated elsewhere for one month for pain and redness of both breasts and fever (≥ 38°C). There was no generalized lymphadenopathy. On gastroscopic examination multiple small ulcers were observed in the stomach. An abdominal computed tomographic (CT) scan showed liver tumor and a normal spleen. Hematological and biochemical examination at admission showed the following results: RBC 352 × 10 4 /mm 3 , hemoglobin 10.3 g/dl (post transfusion), WBC 4,900/mm 3 , Platelets 51,000/mm 3 , serum albumin 1.5 g/dl, total bilirubin 0.6 mg/dl, AST 691 IU/l, ALT 187 IU/l, LDH 2976 IU/l, fibrinogen 134 mg/dl, FDP 10 μg/ml, and AT-III 40%. Bleeding from the stomach continued and did not stop with conservative treatment; therefore, two days later the patient underwent total gastrectomy and a partial liver resection. Histopathology of the resected specimen showed it to be a gastric lymphoma (pleomorphic medium-large cell type, non-Hodgkin's T-cell lymphoma) with liver metastasis (Fig. 1 ). From first postoperative day (POD), bleeding from the esophagojejunostomy continued; the patient developed disseminated intravascular coagulopathy and died on 8 th postoperative day. Figure 1 Photomicrograph showing medium-large sized atypical lymphoid cells with pleomorphic features in the stomach suggesting a gastric lymphoma (Hematoxylin and Eosin, ×170). On autopsy, malignant lymphoid cell infiltration and hemophagocytosis were observed in the liver, spleen, heart, small bowel, lung, both breasts, kidney, pancreas, uterus, and gastroduodenal lymph nodes (Fig. 2 ). The bone marrow presented hyperplasia and hemophagocytic macrophages but no infiltration by lymphoma cells. Immunohistochemically the neoplastic cells were positive for T-cell marker UCHL1 (CD45RO) and EBV by EBER in situ hybridization. The final diagnosis was EBV-related T-cell LAHS. Figure 2 Photomicrograph of the lymph node at autopsy illustrating histiocytes that show hemophagocytosis of normoblast in a lymph node (Hematoxylin and Eosin, ×200). Discussion HPS is a clinicopathological entity characterized by systemic proliferation of benign hemophagocytic histiocytes, fever, cytopenia, liver dysfunction, hepatosplenomegaly, and coagulopathy [ 1 ]. This syndrome has been observed during the clinical course of a wide variety of disorders, including viral infections and malignant neoplasms. Diagnostic guidelines of Henter et al , [ 2 ] are widely used for the diagnosis of HPS. However, these guidelines are not satisfactory in diagnosing HPS in adults; therefore, a number of studies on adult HPS have used their own criteria [ 1 , 3 , 4 ]. On the other hand for the diagnosis of LAHS, in addition to the clinical features, it is also important to confirm the presence of malignant lymphoid cells histopathologically. Takahashi et al , [ 5 ] has proposed a set of new diagnostic criteria for adult LAHS that has been detailed in Table 1 . Table 1 Diagnostic criteria for adult lymphoma associates hemophagocytic syndrome (LAHS) 1 High fever for more than a week (peak 38.5°C) 2 Anemia (Hb < 9 g/dl) or thrombocytopenia (platelet < 100,000 μ/l) 3 a) LDH ≥ 2 × upper limit b) Hyperferritinemia (≥ 1,000 ng/dl) c) Hepatosplenomegaly on CT, US or MRI d) FDP ≥ 10 μg/ml 4 Hemophagocytosis in bone marrow, spleen or liver 5 No evidence of infection 6 Histopathologically confirmed malignant lymphoma A diagnosis of LAHS requires that all of the above conditions are fulfilled. Of the item 3, at least two of the four sub-items (a~d) should be fulfilled. When item 1 to item 5 are present for 2 weeks and glucocorticoid or γ-globulin therapy is not effective, a diagnosis of probable LAHS can be made and chemotherapy against malignant lymphoma can be started. In Japan, T-cell LAHS accounts for 48.5% of all adult LAHS [ 5 ]. T-cell LAHS mostly occurs in extra nodal, especially nasal, cutaneous, or malignant lymphoma involving liver and spleen. There have been no reports on LAHS from gastric lymphoma. As the diagnosis in the present case was made at autopsy it is not clear as to when the HPS occurred initially. One possibility is the setting of disseminated T-cell lymphoma. This is supported by the patient's fever, which continued for one month, liver dysfunction, and coagulopathy, which existed from the initial stage of the disease, however the bone marrow did not show any lymphoma infiltration. It could also be considered that the hemophagocytic syndrome occurred as a result of the surgery as pancytopenia and hepatosplenomegaly were not observed before the operation and hemophagocytosis was not recognized on histopathological examination in the resected stomach. In T-cell lymphoma, the hemophagocytic syndrome is assumed to be caused by cytokines, especially, tumor necrosis factor-α, and interferon-γ released from neoplastic T-cells [ 4 , 6 ]. Uncontrolled secretion of cytokines may stimulate the proliferation and phagocytic activity of macrophages. It seems likely that hypercytokinemia due to surgical resection might have contribute to the development of HPS in the present case. In our opinion the former is more likely however based on the findings of this case the second hypothesis too cannot be rejected. The poor prognosis of LAHS, especially T-LAHS, is well known. The median survival time from the diagnosis is reported to be 143 and 69 days respectively in Japan [ 5 ]. For LAHS prompt initiation of treatment with multi agent chemotherapy is required to improve the symptoms and survival [ 7 ]. Bone marrow transplantation is considered to be a treatment for chemotherapy-resistant LAHS [ 8 ]. The median survival time of LAHS patients without chemotherapy is only 11 days [ 5 ]. In this case, the initial presentation was mastalgia and hence it took a considerable amount of time to reach a diagnosis. Furthermore, bleeding from the anastomosis continued leading to a rapidly progressive fatal clinical course. In HPS occurring in lymphoma of the gastrointestinal tract uncontrollable bleeding from the primary lesion might occur. Therefore, an earlier diagnosis of HPS should be made by bone marrow aspirates, and appropriate treatments should be started as soon as possible. Surgery if performed, must be performed with utmost caution. Conclusions LAHS could also occur from lymphoma of the gastrointestinal tract. For long-term survival; early diagnosis and appropriate treatment are needed. Surgery if performed without a proper diagnosis could prove fatal. Competing interests The author(s) declare that they have no competing interests. Authors' contributions RF, FH, TY, RD, KO and KH were gastrointestinal surgeons. MO referred this patient to us. KK and MS performed pathological examination and the autopsy. KS was a member of the intensive care team. TH, YY, HN gave us helpful comments about the manuscript

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC527873.xml

423142

Information Transport across a Membrane

From a biochemical perspective, a living cell is a collection of molecules jampacked into a confined space by a flexible barrier, called the plasma membrane. A diverse array of proteins embedded in the plasma membrane act as conduits between the cell interior and its external environment, conveying nutrients, metabolites, and information. The life of a cell—as well as that of any multicellular organism—depends on a cell's ability to communicate with its neighbors, both near and far. One way cells do this is with transmembrane receptors outfitted with both extracellular and intracellular domains that mediate information flow between the cell's external and internal environment. One class of transmembrane receptors, called integrin receptors, specializes in interacting with and binding to other cells and the extracellular matrix, a complex of molecules surrounding cells that provides structural support. By integrating various components of the extracellular matrix, integrins (also known as adhesion receptors), play an important role in such diverse processes as cell differentiation, programmed cell death, wound healing, and metastasis. Association between integrin α and β subunit transmembrane domains Integrins can be regulated by signals within the cell to bind to their ligands with either low or high affinity. While a multitude of integrin ligands have been identified and the general mechanics of both the extracellular and intracellular domains of these receptors are known, exactly how a signal crosses the receptor's transmembrane segment to regulate affinity has remained obscure. Now, Bing-Hao Luo, Timothy Springer, and Junichi Takagi have taken a mutational approach to shed light on the inner workings of the transmembrane segment and to explain how it transmits information. Much of what we know about the function of integrins has come from studying the crystal structures and models obtained from structural analysis. These analyses have generated information not only about the structure and composition of the extracellular and intracellular domains of integrins, but also about the conformational changes that accompany signaling events. Integrins contain a large extracellular domain, a transmembrane segment, and a relatively short intracellular “tail.” Integrins are heterodimers—molecules that contain two subunits composed of different amino acids—made up of an α chain and a β chain. Tight association of the two subunits is associated with an inactive, or low-affinity, state of the extracellular ligand-binding domain. Separation of the intracellular subunits is associated with a dramatic conformational change and activation of the extracellular domain, changing a bent structure with a downward-pointing ligand-binding site into an extended one with an outwardly stretched ligand-binding site. This mechanism differs from most transmembrane signaling molecules, which usually achieve activation through association with their target molecules. To investigate how the transmembrane segment mediates these changes, Luo, Springer, and Takagi systematically replaced amino acids in both the α and β transmembrane domains of the heterodimer with cysteines, creating the potential for binding interactions through a chemical reaction, disulfide bond formation, between the two subunits. By analyzing 120 possible cysteine pairs, the researchers not only confirmed the structure of the transmembrane region as helical but also mapped the proximal amino acid residues between the helices. To understand how the helical transmembrane domains transmit signals, the team introduced activating mutations in the amino acids of the α subunit cytoplasmic tail. Using this approach, they observed the loss of the contact between the subunits, indicating a separation of the transmembrane helices. Furthermore, when disulfide bond formation occurred, linking the transmembrane segments together, activation was suppressed. While previous models had proposed various modes of subunit movements, including hinge- and piston-like models, these results strongly support the notion that lateral separation of the subunits is the driving force behind the signal. As many diseases arise from defects in integrin adhesion, understanding the conformation and mechanism of integrin activation could suggest promising avenues for drug development aimed at correcting such defects.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC423142.xml

548522

Association between fetal growth restriction and polymorphisms at sites -1 and +3 of pituitary growth hormone: a case-control study

Background Fetal growth restriction is associated with significantly increased risks of neonatal death and morbidity and with susceptibility to hypertension, cardiovascular disease and NIDDM later in life. Human birth weight has a substantial genetic component, with at least a quarter of the variation attributable to additive genetic effects. Methods One hundred twenty-five subjects (83 control and 42 case) were selected using stringent inclusion/exclusion criteria. DNA sequencing was used to identify 26 single nucleotide polymorphisms in the pituitary growth hormone gene (GH1) at which all subjects were genotyped. Association with fetal growth restriction was tested by logistic regression for all sites with minor allele frequencies greater than 5%. Results Logistic regression identified significant association with fetal growth restriction of C alleles at sites -1 and +3 (relative to the start of transcription) that are in complete linkage disequilibrium. These alleles are present at higher frequency (6% vs. 0.4%) in fetal growth restricted subjects and are associated with an average reduction in birth weight of 152 g in normal birth weight and 97 g in low birth weight subjects. Conclusions There is suggestive association between fetal growth restriction and the presence of C alleles at sites -1 and +3 of the pituitary growth hormone gene.

Background Fetal growth restriction (FGR) is a major risk factor for illness in the perinatal period and throughout life, with the smallest 7.5 percent of infants accounting for two-thirds of infant deaths [ 1 ]. Term, low birth weight infants are at least five times more likely to die in the first year [ 2 , 3 ] and are second only to premature infants in their rates of morbidity and mortality [ 4 ]. FGR infants have an increased frequency of hypoglycemia, hypothermia, polycythemia, neurodevelopmental deficits, and cerebral palsy [ 5 ]. Later in life, individuals born FGR are at elevated risk of hypertension, cardiovascular disease, and non-insulin dependent diabetes [ 6 ]. For example, FGR increases the risk for adult onset of non-insulin dependent diabetes two- to three-fold [ 7 , 8 ]. The ability to both diagnose and treat FGR early in gestation has enormous potential to reduce childhood and adult illness. It is difficult to distinguish the genetic and environmental components of human birth weight variation, but recent studies support a major genetic component to birth weight variation. Clausson et al.'s [ 9 ] study of the offspring of female dizygotic and monozygotic twins (2,009 twin pairs) estimated heritability for human birth weight of 42%, although the confounding influence of shared environmental effects must be considered. In a study of 3,562 captive macaques that minimized environmental heterogeneity, Ha et al. [ 10 ] estimated a total heritability for birth weight of 51%, with an additive genetic component of 23%. These findings demonstrate that comparatively simple and readily identifiable genetic factors influence birth weight. In concert with this recent research, FGR tends both to cluster in families and to recur in successive generations [ 11 - 14 ]. The five genes of the human growth hormone locus reside within about 45 kb on chromosome 17 [ 15 ]. Pituitary growth hormone (GH1) is by far the most thoroughly studied of the genes and lies at the 5' end of the cluster. The remaining four genes, placental growth hormone (GH2) and three chorionic somatomammotropins (CS1, CS2, and pseudogene CS5 or CSHP1), are expressed only from the placenta. The promoter region of GH1 is unusually polymorphic, with 16 SNPs having been identified in a span of 535 bp [ 16 - 18 ]. Most of these SNPs occur at the comparatively small number of sites that exhibit sequence differences among the five genes of the GH locus, and this has been interpreted as evidence of gene conversion [ 16 , 19 ]. Here we use DNA sequencing to identify and to determine the frequencies of both 12 newly-identified single nucleotide polymorphisms (SNPs) in the promoter and coding region of the GH1 gene and 15 previously reported SNPs [ 16 , 18 ]. Using a case-control design, we identify two SNPs in complete linkage disequilibrium near the start of transcription of the GH1 gene that may predispose to reduced birth weight. Methods Human subjects DNA was extracted from placental tissue from 125 live births (83 normal birth weight, 42 low birth weight) at Baystate Medical Center (Springfield, MA) in a case-control study of the genetic predisposition to fetal growth restriction. All subjects were Caucasian, both Hispanic and non-Hispanic. Classification of newborns as fetal growth restricted (or IUGR, intrauterine growth retarded) followed the definition of the American College of Obstetricians and Gynecologists as those newborns below the 10 th percentile of size for gestational age [ 20 ]. Our cut-off of 2,500 g at term (>37 weeks gestation) corresponds to the lowest 7.5 percentile of all US Caucasian deliveries, or the lowest 10 th percentile of female and lowest 6.5 percentile of male deliveries [ 21 ]. Stringent inclusion/exclusion criteria (Table 1 ) were employed by a placental pathologist (TKB) to reduce nongenetic contributors to birth weight variation. Ethical approval to conduct this study was obtained from the Human Subjects Institutional Review Board of the University of Massachusetts. Table 1 Inclusion and exclusion criteria for the study of fetal growth restriction Inclusion Criteria ≥ 37 weeks gestation (full-term) Mother 17–35 years old Singleton pregnancy Case (fetal growth restricted): <2,500 g birth weight Control: 3,000 to 4,000 g birth weight Exclusion Criteria Karyotypic abnormalities, including confined placental mosaicism Placental abnormalities Birth defects or syndromic conditions (i.e., Silver-Russell Syndrome) Pregnancy complications Preeclampsia Type 1, Type 2 or gestational diabetes Meconium staining Uterine infection Maternal chronic illnesses (i. e., hypertension, AIDS, hepatitis, endocrinological) Known illicit drug use Rh disease Polymerase chain reaction (PCR) and sequencing The region from -624 (relative to the start of transcription; GenBank accession J03071) to +1,726 (197 nucleotides after the termination codon) of GH1 was amplified in two overlapping fragments: -624 to +541 (GHN-1F 5' AGGGACCTGGGGGAGCCCCAGCAA 3', GHN-1R 5' TCACCCCTTCCTGCCACCCCTGAT 3') and +450 to +1,726 (GHN-2F 5' CCATCGTCTGCACCAGCTGGCCTT 3'; GHN-2R 5' GCCCTACAGGTTGTCTTCCCAACT 3'). Approximately 50 ng of DNA was used as a template in a polymerase chain reaction with 30 cycles of 95°C (1 minute), 62°C (1 minute), and 72°C (2 minutes 30 seconds). PCR products were purified from agarose using a QIAquick PCR Purification kit (Qiagen) and sequenced directly with BigDye v2.0 chemistry (Applied Biosystems) and run on either an ABI Prism 377 or 3100 automated DNA sequencer. PCR products were sequenced with the PCR primers and additional internal primers: 5' AAGCACAGCCAATAGATTG 3', -459 to -441; 5' GCACAAGCCCGTCAGTGGCC 3', -108 to -89; 5' GGATTTTAGGGGCGCTTACC 3', +71 to +90; 5' CATCTCCCTGCTGCTCATC 3', +931 to+949; 5' GCGCTTGGGYACTGTTCCCT 3', +1280 to +1299. Single nucleotide polymorphism (SNP) genotyping Sequence traces were aligned and assembled into contigs by the program Polyphred [ 22 ]. Contigs were viewed in either the program Consed [ 23 ] or Sequencher (Gene Codes Corp.) and polymorphisms confirmed visually. Twenty-six polymorphic sites were identified (Table 2 ), including all of the sites identified by Horan et al. [ 18 ] in 154 British military recruits with the exception of site -339 which had a minor allele (deletion of G) frequency of 3.6% in their study. Table 2 Frequency of alleles at 26 single nucleotide polymorphisms in the promoter and coding region of pituitary growth hormone and the nucleotide present at the homologous site in other members of the human GH locus Frequency GH1 paralogs Position* Alleles This Study Horan et al. CS-5 CS-1 GH2 CS-2 Categorization Function** -580 A 0.985 A A A A Constant G 0.015 -476 A 0.012 0.013 A G A G Variant G 0.988 0.987 -360 A 0.972 G G G G Variant G 0.028 -352 G 0.012 T G G G Variant T 0.988 -308 G 0.732 0.753 T C T C Variant T 0.268 0.247 -301 G 0.732 0.753 T T T T Variant T 0.268 0.247 -278 G 0.628 0.601 T A T A Variant NF1 T 0.372 0.399 -168 C 0.024 0.019 T C T C Variant T 0.976 0.981 -75 A 0.900 0.886 G A G A Variant PIT-1 G 0.100 0.114 -57 G 0.687 0.633 G T A T Variant Vitamin D Receptor T 0.313 0.367 -31 G 0.882 0.867 G G - G Variant Vitamin D Receptor - 0.118 0.133 -6 A 0.565 0.588 A G A G Variant Transcription Start G 0.435 0.412 -1 A 0.847 0.932 C T A T Variant Transcription Start C 0.044 0.003 T 0.109 0.065 3 C 0.044 0.003 C G G G Variant Transcription Start G 0.956 0.997 16 A 0.976 0.981 G A A A Variant 5' UTR G 0.024 0.019 25 A 0.980 0.981 C A A A Variant 5' UTR C 0.020 0.019 59 G 0.072 0.049 G G G G Variant 5' UTR T 0.928 0.951 69 A 0.968 G C G G Variant Thr/Ala G 0.032 124 A 0.988 G A G A Variant Intron G 0.012 128 A 0.988 C C T C Variant Intron T 0.012 140 A 0.004 G G G G Constant Intron G 0.996 144 A 0.012 G G G G Constant Intron G 0.988 281 C 0.024 T C C C Variant Intron T 0.976 596 C 0.986 T T T T Variant Intron T 0.014 1070 A 0.004 G G G G Constant Synonymous G 0.996 1169 A 0.331 T T T T Constant Intron T 0.669 * Relative to the start of transcription ** Polymorphisms in known transcription factor binding sites are shown. Site +69 is part of the signal peptide. Statistical analyses The five genes of the human growth hormone locus exhibit high sequence similarity, and the paralogous regions corresponding to the portion of GH1 sequenced in this study (-624 to +1,726) were multiply aligned. Nucleotide positions (Table 2 ) were designated as invariant if all five genes had the same nucleotide or the four paralogs of GH1 were identical and matched the major allele at that site in GH1. This categorization explicitly assumes that only the minor alleles in GH1 are the product of gene conversion and that minor alleles not observed in paralogs of GH1 are the result of unique mutations. The proportion of GH1 polymorphisms at invariant versus variant sites was compared by Fisher's Exact Test to determine if there was an over-representation of polymorphic sites among variant sites. Logistic regression, using FGR status as the outcome, was performed on gestational age and genotypes at sites with a minor allele frequency above 5% (Table 3 ). Sites -301 and -308 (relative to the start of transcription) are in complete linkage disequilibrium and the minor C alleles at sites -1 and +3 are in complete linkage disequilibrium. Therefore, sites -301 and +3 were excluded from logistic regression to avoid multicollinearity. Based on the results from logistic regression, separate ANOVA (Table 4 ) was performed on gestational age and the AA versus AC genotypes at site -1 within low birth weight and within normal birth weight subjects. Empirical p values for the F statistic for the genotypic effect, corrected for multiple comparisons, were determined by 2,000 random permutations of the genotypic data [ 24 ]. All statistical analyses were performed using the Stata program (Stata Corp., College Station, TX). Table 3 Logistic regression on FGR status based on gestational age and SNP genotype for GH1 polymorphisms with minor allele frequency greater than 5% Variable Odds Ratio 95% CI Z Score P-value Gestational Age 0.42 0.26–0.66 -3.75 <0.001 -308 2.66 0.54–13.19 1.20 0.23 -278 1.62 0.27–9.80 0.52 0.60 -75 1.70 0.55–5.23 0.93 0.35 -57 3.34 0.62–18.12 1.40 0.16 -31 0.85 0.27–2.65 -0.29 0.78 -6 2.11 0.66–6.74 1.26 0.21 -1 A/T 0.76 0.23–2.44 -0.47 0.64 -1 A/C 0.10 0.01–0.77 -2.21 0.03 +59 1.48 0.32–6.95 0.50 0.62 +1169 1.05 0.38–2.95 0.10 0.92 Table 4 ANOVA on A/C genotypes at site -1 and gestational age in normal and low birth weight subjects Normal Birth Weight Low Birth Weight Site -1 Gestational Age Site -1 Gestational Age Mean AA 3382.8 2287.9 Mean AC 3230.2 2190.4 ANOVA F 3.75 4.55 3.12 2.25 P-value 0.056* 0.002 0.073* 0.099 * Empirical P value corrected for multiple testing determined by 2,000 random permutations of the genotypic data [24] Results Polymorphism in the GH1 gene Among the 125 subjects sequenced from -624 to +1,726 of the GH1 gene, 26 polymorphic sites were identified (Table 2 ). These included all but one of the 15 sites characterized by Horan et al. (2003). In the region of overlap between the two studies, we failed to detect variation at site -339, where there is a minor allele deletion of a single nucleotide with frequency 3.6% in British army recruits, and identified two additional variants at sites -360 and -352 with minor allele frequencies 2.8% and 1.2%, respectively. Therefore, the discrepancies between the two studies in the identification of SNPs can most likely be ascribed to sampling error. Outside the region surveyed by Horan et al. (2003), we detected 10 additional polymorphisms. All of these had minor allele frequencies ≤ 3.2%, except an intron four polymorphism at site +1169 with a minor allele frequency of 33.1%. In general, polymorphisms in the promoter of GH1 tend to be more densely clustered and exhibit higher minor allele frequencies than in the transcribed region. Evidence of gene conversion An alignment of the region of GH1 sequenced in this study with the paralogous sequences of placental growth hormone and chorionic somatomammotropins indicated that there are 1,979 invariant sites and 293 variant sites (excluding 78 sites that could not be unambiguously aligned), as defined in the methods. Of the invariant sites, 5 are polymorphic in GH1, while 21 of the variant sites are polymorphic. A comparison of the proportion of polymorphic sites at invariant and variant sites by Fisher's Exact Test is highly significant (P <<0.001). This result indicates that there is a strong bias for polymorphisms in GH1 to occur at the minority of sites that exhibit sequence divergence among the paralogous genes. The high correspondence between the sequence of minor alleles in GH1 and nucleotides present in paralogs of GH1 supports previous assertions that the unusually high polymorphism of the GH1 gene is driven by gene conversion [ 16 , 18 ]. A proportion of this bias may be explained by selective constraint, in that sites that are polymorphic within GH1 may be under less selective constraint and thus more free to exhibit sequence divergence among paralogs. However, to entirely explain the bias towards polymorphism at sites of divergence, one must assume that about 1,804 (91%) of the invariant sites are selectively constrained and not free to vary. Given that a substantial proportion (814 bp, ~36%) of the sequence surveyed in this study is composed of introns, this assumption seems unreasonable. Association with fetal growth restriction Logistic regression (Table 3 ) was performed on gestational age and SNPs with minor allele frequencies greater than 5%, excluding sites -301 and +3 because they are in complete linkage disequilibrium with other sites, to identify associations with fetal growth restriction. Gestational age was significant because the FGR subjects exhibit a slightly younger average estimate of gestational age (38.1 vs. 39.1 weeks, t = 4.9, p < 0.0001; gestational ages rounded to nearest week). However, even accounting for the effect of gestational age, the C allele at site -1 was significantly associated with FGR in the combined regression. However, this allele did not retain significance in a regression on only gestational age (p < 0.001) and the A/C polymorphism at site -1 (p = 0.242). Although the A/C polymorphism at site -1 was not significant in a reduced model, we decided to investigate this site for three reasons. First, the C allele at site -1 is in complete linkage disequilibrium with C at site +3. Second, both of these sites are located at the start of transcription, making them good candidates for affecting the level of transcription of GH1. Third, the only paralog that shares C at these sites is CS-5, a pseudogene, consistent with the possibility that C at sites -1 and +3 is disadvantageous. The C allele at both sites exhibits a much higher frequency in low birth weight (6%) versus normal birth weight (0.4%) subjects. Restricting examination only within the normal or low birth weight subjects, the AC genotype is associated with an average reduction of birth weight of 152 g and 97 g in normal and low birth weight subjects, respectively. However, this difference does not achieve significant in an ANOVA on the A/C polymorphism and gestational age (Table 4 ). Discussion Birth weight in humans and other primates exhibits substantial heritability [ 9 - 13 ]. Although a suite of environmental and genetic/karyological insults are known to cause fetal growth restriction, perhaps as many as 40% of FGR cases have no known etiology [ 25 ]. Therefore, identification of the underlying genetic variants that predispose to FGR could have significant medical significance if it allows us to identify early in gestation those pregnancies that are at increased risk of growth retardation. A logical place to begin such a search is among those genes that are known to be major regulators of fetal growth and to exhibit significant differences in circulating protein concentrations between normal and FGR pregnancies, such as the members of the growth hormone-prolactin and insulin-IGF families of hormones, receptors, and binding proteins. Here, we use a stringently selected set of subjects to report suggestive association of SNPs in GH1 with fetal growth restriction. Adjusted for gestational age, the C alleles at sites -1 and +3 of the GH1 gene appear to be associated with reduction in birth weight. The marginal significance of these results may be the result of several factors. First, we examined 125 total subjects, roughly one-third of them FGR, and this number may give inadequate statistical power to identify weak genetic effects. Second, the C allele is low frequency, providing a small number of heterozygotes for the allele and no homozygotes. It is worth noting that among normal birth weight subjects we observed a very similar frequency for the C allele as Horan et al. (2003) did among British army recruits (0.4% vs. 0.3%) but that among the low birth weight subjects the frequency of the C allele is substantially higher (6%). Comparing all the other allele frequencies (Table 2 ) between the two studies of Caucasian populations, no other allele shows such a large magnitude of difference, although the proportional difference in frequency of C at -1 and +3 may be somewhat distorted by sampling error at low allele frequencies. Third, FGR undoubtedly has a heterogeneous genetic etiology, and it may be unlikely to find any genetic variant that accounts for more than a small proportion of cases. If the C allele at -1 and +3 exhibits true association with FGR, it unfortunately may be very difficult to determine if the effect is due to one or both sites because they are in complete linkage disequilibrium. Nevertheless, the presence of both alleles at the start of GH1 transcription provides both variants with a biologically plausible possibility to affect the level of transcription of GH1. Future work should be devoted to examining the effect of these alleles on transcription levels both alone and in combination. Site -1 is somewhat unique among single nucleotide polymorphisms in that three alternate alleles exist at that site (the major allele A and the minor alleles C and T). There are two possible explanations for this observation. First, site -1 could be a hotspot for nucleotide mutation, with transversions from A to C/T occurring often. Second, GH1 may be the recipient of gene conversion events from more than one paralog within the GH locus. Among the four other genes of the GH locus, all three alleles observed at site -1 occur (Table 2 ). Given the predominant effect that gene conversion [ 16 , 18 ] appears to have on the patterns of nucleotide polymorphism in the GH1 gene, the latter explanation may be more plausible. It must be pointed out that the common wisdom is that GH1 plays no role in regulating fetal growth because the GH receptor is expressed fairly late in gestation and because anencephalic infants or those born without a pituitary achieve nearly normal length [ 26 ]. Importantly, we restricted our study to full term deliveries. It is possible that late in gestation, when fetal GH1 is expressed and GHR receptors are present in a wide variety of fetal tissues, pituitary growth hormone begins to have a growth stimulatory role sufficient to account for the 90–150 g difference in birth weight between genotypes for the A/C polymorphism at -1 and +3 that we observed. Conclusions In a stringently selected set of subjects, C alleles at sites -1 and +3 relative to the start of transcription of GH1 have a higher frequency (0.4% vs. 6%) in fetal growth restricted newborns. These two alleles are in complete linkage disequilibrium and their presence is associated with a reduction in birth weight of 152 g in term, normal birth weight subjects and 97 g in term, low birth weight (<2,500 g) subjects. In combination with environmental, behavioural and other genetic factors, these alleles may contribute to fetal growth restriction. Competing interests The author(s) declare that they have no competing interests. Authors' contributions RMA conceived the project and directed its design and execution. CC and RV performed the molecular genetic work and participated in preliminary analyses. TKB performed the subject selection. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548522.xml

544867

Progress, challenges, and responsibilities in retrovirology

In this editorial, Retrovirology's choice for best basic science "retrovirus paper of the year" and a perspective on challenges and responsibilities facing HIV-1 and HTLV-I research are presented.

Progress The beginning of a year provides an occasion to look back upon progress made over the past 52 weeks. With the end of 2004, Retrovirology concluded its first calendar year of publishing. In actual fact, Retrovirology launched as an Open Access journal the final week of February 2004 and has been publishing continuously for a little more than 10 months. Over that period, with the wonderful efforts from my 6 very capable Associate Editors (Monsef Benkirane, Ben Berkhout, Masa Fujii, Mike Lairmore, Andrew Lever, and Mark Wainberg), the journal has thrived. The goal that we set for Retrovirology is to provide a visible forum for retrovirologists so that their works can be read by all in a free and openly accessible manner. What this means is that if you are a human immunodeficiency virus (HIV)-researcher and you had published a paper in Retrovirology , a graduate student in Sri Lanka updating his/her research protocol, an AIDS activist in South Africa looking for the latest information, and even your long-lost high school sweetheart wondering what you have been doing all these years, can all find your work (i.e. through a simple Google or PubMed search) and read your most recent findings. Perhaps more relevant to the enterprise of scientific communication is that numerous academic peers in Eastern Europe, Asia, South America, Africa and elsewhere do not have funds which would permit them to read Cell , Science or Nature . Hence, while some can read your research in Cell , Science , or Nature , all colleagues, rich or poor alike, can read your Retrovirology paper. Are they reading Retrovirology ? You bet! Our monitored statistics tell us that in 2004, the most highly accessed review article [ 1 ] published in Retrovirology was read by over 3,500 individuals while a comparably popular original research article [ 2 ] was read more than 2,400 times. Readers also read Retrovirology articles with great immediacy. Thus, a recent paper by Rana and colleagues [ 3 ] appeared in Retrovirology on December 27 th , 2004; and already by December 31 st , 2004, a short 4 days later, that study had been read 389 times. Just as readers are quick to read our papers, I am equally pleased by our unmatched speed in publishing authors'works. In 2004, based on all papers we published in Retrovirology , the time from submission to publication averaged 40 days. From my personal experience of publishing in other virological journals, this duration is 3 to 4 times faster than our best competitors. Different journals/magazines recognize "Molecule of the Year", "Breakthrough of the Year", or even "Person of the Year". With this editorial, Retrovirology will initiate the annual recognition of the best basic science "retrovirus paper of the year". The Associate Editors and I decided that in 2004, the best basic science retrovirus paper was the work from Joseph Sodroski and colleagues describing the HIV-1 restrictive property of the tripartite motif 5 (TRIM5) protein [ 4 ]. Thus, these researchers characterized in primates a restriction factor, similar to the Friend virus susceptibility factor-1 (Fv1) in mice, which counters the ability of infecting retrovirus to establish a proviral form in target cells. In coming years, I anticipate that Retrovirology Editors will find it fitting to recognize a Retrovirology paper as the "best basic science retrovirus paper" of the preceding year. Challenges and responsibilities I explain to my postdoctoral fellows that challenges are those issues which you think others should solve, while responsibilities are items that you think you should tackle. As a retrovirologist depending on how you regard yourself, pressing problems are either others' challenges or your responsibilities. I study two retroviruses, HIV-1 and human T-cell leukaemia virus type 1 (HTLV-I). The start of a new year offers me a chance to review briefly my personal bias on the important research question that confronts HIV-1 and HTLV-I, respectively. For HIV-1, the "holy grail" remains the development of an effective vaccine against the virus. As we enter 2005, mortality from AIDS is staggering. It is estimated that in 2004, 3.5 million individuals perished worldwide from AIDS; or nearly 10,000 AIDS deaths each day. We can view this number in another way. The recent tsunami in South Asia is estimated to have caused 150,000 fatalities. AIDS in 2004 is then the equivalent of 23 tsunamis. Imagine, unrelentingly tsunami-like casualties every fortnight from people dying from HIV-1! While it is laudable that the World Health Organization has a goal to treat three million HIV-1 positive individuals globally using anti-retroviral (ARV) medicine over the next five years, that approach will unlikely address the full magnitude of the AIDS problem, especially in developing nations. On the other hand, 100 million infants (even those in remote regions of the world) receive basic vaccinations each year. This fact suggests that when an AIDS vaccine does become available, that vaccine could be logistically and practically effective. Separately, statistics from the World Cancer Report for the year 2000 show that 5.3 million men and 4.7 million women developed malignancy, and altogether 6.2 million persons died from cancer worldwide. The American Cancer Society estimates that approximately 553,000 individuals succumb to cancer in the United States each year. I see HTLV-I, the etiological agent for adult T-cell leukaemia (ATL), as perhaps the best retroviral system for retrovirologists to study human cancer. Substantive progress has indeed been made in our understanding as to how HTLV-I transforms cells in tissue culture [ 5 ]. What remains needed is the development of a good non-human primate model to investigate the genesis of adult T-cell leukaemia by the virus in vivo . Opportunities and limitations at mid-career I first started studying viruses in the fall of 1977 at age 19 as a MD-PhD student at the Johns Hopkins school of medicine. For me personally, 2004 marked over a quarter-century of virus-research. At age 46, the unbridled youthful optimism of 19 is now tempered by realizations of physical and career limitations (i.e. some very interesting research problems are going to take longer to resolve than the remaining span of my scientific and physical endeavor). Nonetheless, I am optimistic and hopeful that, despite enormous odds, the opportunity to see a successful AIDS vaccine will come before I leave retrovirology research in 20 or some years. Regarding Retrovirology , I am also optimistic that I started this project at an age that provides ample time to develop this journal into a premier research forum. Let me conclude this writing by thanking all authors, reviewers, Editorial board members, and our wonderful staff at Biomed Central who have contributed to Retrovirology's progress in our first year.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544867.xml

529321

Are Animals As Irrational As Humans?

Animals in the wild are constantly confronted with decisions: Where to nest? Who to mate? Where's the best forage? To explore the mechanisms underlying such decisions, animal behavior studies often incorporate concepts from economic theory. Mainstream models of choice in both economics and biology predict that preferences will be rational, or consistent across contexts, as a result of being motivated by self interest or, in the case of animals, reproductive success. Yet many studies report that when making decisions people often take shortcuts, using cognitive heuristics that may lead to economically irrational decisions, with similar claims now showing up in animal behavior studies. In a new study, Cynthia Schuck-Paim, Lorena Pompilio, and Alex Kacelnik ask whether studies applying economic rationality to animal behavior are controlling for potentially confounding effects inherent in such approaches. The authors suggest that observed “breaches of rationality” may stem from differences in the physiological state of animals “unwittingly imposed” by experimental design rather than from real irrational decisions. Choice studies typically offer subjects a range of choices that include clearly superior and inferior alternatives. While humans can simply hear about the various alternatives and their respective properties, animals must be trained to learn about the different choices. This difference is far from trivial, Schuck-Paim et al. argue, and could well require different interpretations of results in animal and human studies. In fact, economic theory states that optimal choices depend on both the properties of the option and the chooser's state. Training animals to learn of different choices typically involves giving them food rewards, which means that an animal's energetic state—that is, hungry versus sated—will change over a day of training. A bird that's eaten an ounce of birdseed is more likely to opt for an “irrational” option—say, a choice that dispenses little food—than one that's hungry. To examine this theoretical constraint under experimental conditions, Schuck-Paim et al. trained European starlings to choose between two rich food sources (called focal options) and one of two poorer “decoys” in different contexts. One of the focal options offered more food while the other offered a shorter delay between pecking a key and receiving the food, but their amount/delay ratios were equal. The decoys were considered less preferable because their ratio of amount to delay was lower than that of the focal options. But the decoys could potentially confound the results because repeated training to each decoy could sate a bird's appetite to different degrees: although amount/delay was equal among the decoys, their long term energetic consequences differed. European starlings make rational decisions The authors tested for preference between the focal options under three experimental conditions: altering the birds' food intake/energetic state with no decoys; changing the decoy and not controlling for its corresponding energetic contributions; and changing the decoy but controlling for its energetic consequence (by supplemental feeding). Schuck-Paim and collaborators show that the birds' preferences between the focal options differed significantly between treatments, in apparent breach of economic rationality; the preference for the larger reward option over the shorter delay option was much stronger when the trial involved lower accumulated intake than when the accumulated intake was high. Introducing the decoys resulted in an “irrational” preference only when the decoys were allowed to have an effect on food intake, suggesting that the choice resulted from the birds' energetic state rather than from cognitive mechanisms of choice similar to those used to explain irrationality in human subjects. The authors offer an evolutionary and mechanistic explanation for why animal preference might be governed by energetic state, including the possibility that animals are less motivated to focus exclusively on the richest option when they are well fed. But they are careful to disabuse the notion that “state-dependence accounts for all reported inconsistencies in animal choice” or that animals do not employ cognitive mechanisms of choice similar to those of humans. Altogether, Schuck-Paim and co-authors argue, these results warn that studies appropriating ideas from other disciplines can introduce confounding effects. And that researchers would do well to carefully examine the underlying causes of observed animal behaviors when testing ideas formulated in a nonbiological framework.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529321.xml

526760

Improving epidemic malaria planning, preparedness and response in Southern Africa

Malaria is a major public health problem for countries in the Southern Africa Development Community (SADC). While the endemicity of malaria varies enormously across this region, many of the countries have districts that are prone to periodic epidemics, which can be regional in their extent, and to resurgent outbreaks that are much more localized. These epidemics are frequently triggered by climate anomalies and often follow periods of drought. Many parts of Southern Africa have suffered rainfall deficit over the past three years and countries expect to see increased levels of malaria when the rains return to more 'normal' levels. Problems with drug and insecticide resistance are documented widely and the region contains countries with the highest rates of HIV prevalence to be found anywhere in the world. Consequently, many communities are vulnerable to severe disease outcomes should epidemics occur. The SADC countries have adopted the Abuja targets for Roll Back Malaria in Africa, which include improved epidemic detection and response, i.e., that 60% of epidemics will be detected within two weeks of onset, and 60% of epidemics will be responded to within two weeks of detection. The SADC countries recognize that to achieve these targets they need improved information on where and when to look for epidemics. The WHO integrated framework for improved early warning and early detection of malaria epidemics has been recognized as a potentially useful tool for epidemic preparedness and response planning. Following evidence of successful adoption and implementation of this approach in Botswana, the SADC countries, the WHO Southern Africa Inter-Country Programme on Malaria Control, and the SADC Drought Monitoring Centre decided to organize a regional meeting where countries could gather to assess their current control status and community vulnerability, consider changes in epidemic risk, and develop a detailed plan of action for the forthcoming 2004–2005 season. The following is a report on the 1 st Southern African Regional Epidemic Outlook Forum, which was held in Harare, Zimbabwe, 26 th –29 th September, 2004.

Introduction The Southern African region has a long and varied history of malaria control with periodic epidemics occurring [ 1 , 2 ]. These epidemics can be regional in scale, as in 1996 and 1997, or much more focal, affecting specific districts or sub-districts. The countries of the Southern African Development Community are committed to the Abuja Targets for Roll Back Malaria in Africa, and this includes improved detection and response to epidemics [ 1 ]. To meet these targets countries are expected to detect 60% of malaria epidemics within two weeks of onset, and respond to 60% of epidemics within two weeks of their detection. The countries recognize that to achieve these targets they need improved information on where epidemics are most likely to occur, and ideally some indication of when they are likely to occur. The WHO guidelines on the development of Malaria Early Warning Systems (MEWS) for Africa are seen as offering a useful framework for an integrated approach to epidemic preparedness and response planning [ 3 - 5 ]. Experience and evidence of the successful application of this approach within the National Malaria Control Programme in Botswana over the past few years was demonstrated by the national malaria programme manager at the Southern Africa Regional Malaria Planning and Consultation Meeting in Gaborone in July 2004. Other countries in the SADC region considered this approach to provide a useful framework for planning epidemic preparedness and response strategies and, in view of the perceived vulnerability of communities throughout much of the region, called for a regional meeting that could launch the scaling-up of this process to include other epidemic prone countries beyond Botswana. The WHO Southern Africa Inter-Country Programme for Malaria Control (SAMC) responded to this demand and together with SADC's Drought Monitoring Centre (DMC) organized the 1 st Southern African Regional Epidemic Outlook Forum, which was held in Harare, Zimbabwe, 26 th –29 th September, 2004 and hosted by Zimbabwe's Ministry of Health and Child Welfare. Representatives from malaria control services in nine Southern African countries participated in the meeting: Angola, Botswana, Madagascar, Mozambique, Namibia, Swaziland, Tanzania, Zambia, and Zimbabwe. The purpose of the meeting was: to enable malaria control services from epidemic prone countries to gather and review their control programme status and epidemiological trends for the past 3–5 years and identify and map districts they consider to be vulnerable to epidemics; to learn about advances in the science of seasonal climate forecasting and review the implications of the forecast for the forthcoming season; to learn about environmental variables pertinent to epidemic risk and readily available sources of monitoring information; to review methods of early detection using case surveillance data; and, using the WHO framework for MEWS, to develop plans of action for epidemic preparedness and response for the forthcoming season. Discussion The MEWS framework as set out by WHO consists of four components: 1) vulnerability monitoring; 2) seasonal climate forecasting; 3) environmental monitoring; and 4) sentinel case surveillance. This framework is illustrated in Figure 1 . Figure 1 MEWS gathering cumulative evidence for early and focused response (WHO 2004) Vulnerability monitoring There are many factors that increase the vulnerability of a population to malaria epidemics [ 6 , 7 ] and increase the severity of disease outcome should a malaria epidemic occur. Co-infection with other diseases such as HIV-AIDS is a major consideration for Southern African countries. Resistance to therapeutic drugs and insecticides has also been a recent problem throughout much of the region. Drought, food insecurity and associated population movements between areas of differing endemicity combine to make certain populations more vulnerable to epidemics. These factors and consideration of the where and how to get appropriate information were discussed and countries were encouraged to identify measurable indicators and key informants. Seasonal climate forecasting In recent years there have been significant scientific advances in our ability to predict climate on the seasonal timescale [ 8 ]. The skill associated with these predictions varies from region to region, but is generally higher within the tropics. Scientists from the SADC Drought Monitoring Centre and the International Research Institute for Climate Prediction (IRI) joined with meteorologists from Democratic Republic of Congo, Malawi, Namibia, Zimbabwe and the World Meteorological Organization (WMO) to deliver the climate forecast for the forthcoming 2004–2005 season. An overview of climate variability in the SADC region was presented. The inherent issues of probability and uncertainty in climate forecasting were discussed with participants from the malaria control services. A number of myths were exploded and the variables that could or could not be skilfully forecast were reviewed. The malaria control participants gave their views on how communication of the forecast should be improved and made more understandable to the non-climate-specialist. Following a subsequent working session by the climate and meteorological specialists, an outline of additional or alternative forecast indicators was provided. Environmental monitoring The availability of environmental variables pertinent to malaria transmission, such as rainfall, temperatures, humidity, and flooding, were discussed and information on where they could be obtained was provided. The two basic sources of such information are periodic summaries (usually satellite-derived and interpolated estimates) available through the internet from the SADC DMC, the Famine Early Warning Systems Network (FEWS-NET) or the International Research Institute for Climate Prediction, Columbia University (IRI) websites; or directly from national meteorological services' ground-based weather observations. Generally, summary products are available free of charge, whereas the meteorological services may need to charge for raw data. Countries were encouraged to begin dialogue with their national meteorological services and discuss the more specific information requirements and support they may need. Sentinel case surveillance The paramount importance of developing good health information and sentinel surveillance systems was acknowledged. The process of MEWS development is seen as offering opportunities for strengthening integrated health systems surveillance. It is in itself dependent on good epidemiological data for testing and validating the relationships between the component parts. Methods of using indicators for epidemic early detection were discussed. Various indicators such as the mean × 2 standard deviations, the 'normal channel', cumulative-sum and weekly case thresholds have been tried, tested and used in a number of Southern Africa countries, and countries are encouraged to develop and use what is most appropriate and effective for their purposes. However, a number of the countries acknowledged having a poor statistical basis on which to develop and test early warning and detection indicators. Following the formal presentations setting out the MEWS components and epidemiological trends, the discussions centred around the countries' perceived control needs over the coming season and the information requirements for developing appropriate plans of action for epidemic preparedness and response. The countries represented varied markedly in their current levels of endemicity/epidemicity, surveillance and control coverage. Tanzania is for the most part a highly endemic country with an estimated 16–19 million cases per year. Botswana and Swaziland, by contrast, are currently recording cases in the low thousands and hundreds respectively. Zimbabwe's economic situation has recently compromised its control programme, and two of the countries, Mozambique and Angola, are in process of reconstructing their control programmes after recently emerging from major disruption due to long-term conflict situations. However, all of the countries did acknowledge the integrated MEWS approach as offering a useful framework for improving their epidemic planning, preparedness and response capabilities. Based on the the climate forecast for October, November, December, and the extended forecast for January, February, March, which are posted on the DMC website . The participants discussed the difficulties in access to and interpretation of meteorological data. The representatives from the meteorological services expressed a willingness to engage in closer collaboration to address these issues. The participants voiced a clear need to improve the availability of the seasonal climate forecast to the epidemic prone districts. They also highlighted the need for better communication of the forecast to non-climate users. Requests were specifically made for forecasts that are more 'meaningful' to the health sector. In response, the meteorological sector pointed at the necessity to know more specifically what information the health sector requires in order to then meet this need. Forecasts could, for example, be expressed simply as the probability of the coming season being wetter or drier than the previous season, or two, or three, or n seasons; or compared to that of the last epidemic season; or as probabilities of exceeding a given threshold for the season. However, it was stressed that forecasts will always be probabilistic and not deterministic. Moreover, countries were encouraged to refer to forecast updates as the season progresses. The issues of how to communicate better the probabilities and uncertainties associated with seasonal climate forecasts were addressed more closely. While many activities in malaria control are based on probabilistic, uncertain premises (clinical diagnosis and presumptive treatment, for instance), public health professionals are well aware of the limitations of their own indicators. While recognizing the potential value of advance lead-times for planning, they are understandably cautious in basing critical decisions on uncertain information from others, and the health and meteorological sectors probably need to work this through in more local collaborative settings. One additional issue that came out strongly during the meeting was the need for broader cross-border collaboration on epidemic prevention and control as 'true epidemic' prone areas are often based on particular environmental zones rather than administrative boundaries. For example, high rainfall anomalies in Angola may ultimately find their way as increased stream-flow into Botswana and Namibia, and create extensive breeding sites for vectors. Drought, food security, or a range of other factors, may lead to migrations of people across borders from one level of endemicity to another and pose a significant increase in epidemic risk. Development of national epidemic risk maps therefore ought to reflect the situation in neighbouring countries. There are a few examples of cross-border initiatives in the region: Republic of South Africa, Swaziland and Mozambique; Republic of South Africa and Zimbabwe. Both are showing promising results. Conclusions The meeting ended with the presentation of the recommendations, to be followed up within the next twelve months. The majority of the recommendations highlighted the need for stronger collaboration a) within the health sector itself; b) among the health sector and the climate-meteorological sector, and other relevant sectors; and c) among the various countries in the region. The participants committed themselves, with their partners, to developing integrated early warning systems as a decision support tool for improving epidemic preparedness and response planning. They recognized that this will be best achieved by drawing on appropriate scientific and technological advancements (and challenging these where necessary), by conducting operational research, and with the help of technical development and support, strengthen the capacity for improved epidemic preparedness and response in the districts at risk. The successful implementation of MEWS will depend on close cooperation among several partners: National Malaria Control Programmes must work closely with National Meteorological Services, supported by the regional Drought Monitoring Center, WMO and WHO. Opening these channels of communication will allow public health professionals and climate-environmental scientists and practitioners to incorporate more meaningful variables into the seasonal forecasts. In addition, it is necessary to exchange information with institutions dealing with vulnerable populations such as food security and refugee agencies, to develop mutually beneficial mechanisms that ensure easy access and utilization of relevant information for planning or decision-making. It was recognized that by adopting the MEWS approach for malaria control planning the overall health information and surveillance system would be strengthened as other diseases have strong climate and environmental components to their distribution, and further dialogue with Integrated Disease Surveillance and Response services would be useful. Training and capacity building requirements were discussed with WHO-AFRO regarding implementation within the national IDSR activities in the Southern Africa sub-region, and sub-regions elsewhere in Africa. In the final evaluation of the meeting participants, from both health and climate communities, considered that the meeting had provided a very useful overview of the issues and offered a good starting point for them to develop more flexible Plans of Action for Epidemic Preparedness and Response in their countries. It was recommended that a similar meeting be held each year prior to the onset of the rainy season.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526760.xml

519023

Food and nutrient intake in relation to mental wellbeing

Background We studied food consumption and nutrient intake in subjects with depressed mood, anxiety and insomnia as indices of compromised mental wellbeing. Methods The study population consisted of 29,133 male smokers aged 50 to 69 years who entered the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study in 1985–1988. This was a placebo-controlled trial to test whether supplementation with alpha-tocopherol or beta-carotene prevents lung cancer. At baseline 27,111 men completed a diet history questionnaire from which food and alcohol consumption and nutrient intake were calculated. The questionnaire on background and medical history included three symptoms on mental wellbeing, anxiety, depression and insomnia experienced in the past four months. Results Energy intake was higher in men who reported anxiety or depressed mood, and those reporting any such symptoms consumed more alcohol. Subjects reporting anxiety or depressed mood had higher intake of omega-3 fatty acids and omega-6 fatty acids. Conclusions Our findings conflict with the previous reports of beneficial effects of omega-3 fatty acids on mood.

Background Diet has an effect on mood and cognitive function [ 1 ]. There is some evidence that deficiency or supplementation of nutrients can affect not only mood, but also behavioral patterns. A double-blind placebo-controlled trial with 30 patients showed that omega-3 essential fatty acid supplements alleviated symptoms in patients with bipolar disorder [ 2 ]. In a recent double-blind, placebo-controlled trial on 231 young adult prisoners, by comparing the number of their disciplinary offences before and during the supplementation, antisocial behavior was reduced by the supplementation of vitamins, minerals and essential fatty acids [ 3 ]. Vitamin D supplementation during winter was reported to improve mood in a double-blind, placebo-controlled trial on 44 healthy volunteers [ 4 ]. A number of studies have shown that acute tryptophan depletion produces depressive symptoms and results in worsening of mood [ 5 ]. Folic acid deficiency may also correlate with depression, and it has particular effects on mood, cognitive as well as social functioning [ 6 ]. Recently, it have been reported that low levels of dietary folic acid are associated with elevated depressive symptoms in middle-aged men [ 7 ]. In general, a low-fat diet may have negative effects on mood [ 8 ], and altered dietary fat intake can lead to acute behavioral effects such as drowsiness, independent of energy consumption [ 9 ]. A high intake of proteins also seems to increase alertness [ 1 ]. Increased dietary serine and lysine may be linked to the pathogenesis of major depressive disorder [ 10 ]. Apart from specific nutrients or vitamins, certain foods may have an effect on mental wellbeing. Warm milk, for instance, has been traditionally used as self-medication for insomnia. Individuals drinking regular coffee with caffeine have reported to have decreased total sleep time and sleep quality, and increased sleep latency [ 11 ]. It has been reported that people with a high consumption of fish appear to have a lower prevalence of major depressive disorder [ 12 , 13 ]. Recently, it has been also reported that increased fish intake in people without depressive symptoms had no substantial effect on mood [ 14 ]. Depressed subjects tend to consume more carbohydrates in their diets than non-depressed individuals [ 15 ], and they show heightened preference for sweet carbohydrate or fat rich foods during depressive episodes [ 16 ]. High carbohydrate intakes increase brain uptake of tryptophan, which in turn stimulates the synthesis of serotonin [ 1 ]. At present, there are some studies focusing on the use of dietary supplements in individuals with mental disorders, but there is a lack of consistent data concerning the impact of nutrition, diet and eating habits on mental health. Aims We set out to study whether food consumption and intake of nutrients in subjects with depressed mood, anxiety and insomnia differed from those in subjects without any such symptoms. Methods This study was based on the cohort of a randomized, double-blind, placebo-controlled primary prevention trial testing the hypothesis that daily supplementation with α-tocopherol or β-carotene reduces the incidence of lung and other cancers [ 17 ]. The study participants were recruited between 1985 and 1988 from the total male population 50–69 years of age, residing in southwestern Finland (n = 290,406). These men were sent a questionnaire on current smoking status and willingness to participate in the trial. Smokers of at least five cigarettes per day and who were willing to participate were then invited to visit their local study center for further evaluation of their eligibility. A previous cancer diagnosis, current severe angina with exertion, chronic renal insufficiency, cirrhosis of the liver, alcohol dependence, or a disorder limiting participation in the long-term trial, such as mental disorder or physical disability, were reasons for exclusion. A total of 29,133 men were randomly assigned to receive supplements of either α-tocopherol, β-carotene, both, or placebo, in a 2 × 2 factorial design. The ethics review boards of the participating institutions approved the study, and all subjects provided written informed consent prior to randomization. At baseline, subjects completed a questionnaire on their background and medical history, including three questions on mental wellbeing. These items concerned anxiety, depressed mood and insomnia experienced in the past four months. Height and weight were also measured, and a blood sample was drawn for determining total and high-density lipoprotein (HDL) cholesterol concentrations. Diet and alcohol consumption was assessed from a self-administered dietary history questionnaire [ 18 ], which asked the frequency of consumption and the usual portion size of 276 food items during the past year, using a color picture booklet as a guide for portion size. Complete dietary data were available for 27,111 participants. Dietary nutrient data were analyzed by linking the questionnaire data to the food composition database of the National Public Health Institute, Finland. For analysis, we considered three main groups: principal nutrients, specific nutrients selected on the basis of a priori hypotheses, and certain foods. The principal nutrients were energy, carbohydrates, proteins and fats. The hypothesis-based nutrients were omega-3 and omega-6 fatty acids, lysine, serine, tryptophan, and two vitamins, vitamin D and folic acid. Omega-3 fatty acids from fish consist of long-chain fatty acids, while the omega-3 fatty acids in vegetables are shorter-chain molecules. The food items included were fish, milk, meat, vegetables, margarine, coffee and alcohol. We also evaluated the total energy intake. The trial involved three follow-up visits annually. At each follow-up visit the participants were asked whether they had felt anxiety, depression, or insomnia since the preceding visit (Have you felt feelings of depression in last three months? Have you felt feelings of anxiety in last three months? Have you had insomnia in last three months?). To identify subjects who suffered chronically from these symptoms we took into account the symptoms reported throughout the first follow-up year, i.e. at baseline and the three follow-up visits (at baseline, 4 months, 8 months and 12 months). Men reporting anxiety, depression, insomnia, or all these symptoms at all four visits were included in these analyses. Statistics As potential risk factors, baseline age, body-mass index (BMI), energy intake, alcohol consumption, education level, marital status and smoking were entered into regression models as covariates. Dietary factors were adjusted for energy intake in the models [ 19 ]. Results At study entry, 4314 (16%) men reported depressed mood in the past four months, 6498 (24%) feelings of anxiety, and 5550 (21%) insomnia. The mean intake of energy was 1 to 3% greater and consumption of alcohol 30 to 33% greater in subjects with any such symptoms, compared with symptom-free individuals (Table 1 ). Men reporting all three symptoms consumed as much as 47% more alcohol than those without any symptoms. Subjects with insomnia consumed 7% less coffee than symptom-free individuals, whereas those with depressed mood or anxiety consumed only about 2% less coffee (Table 2 ). Table 1 Baseline characteristics of subjects with self-reported depressed mood, anxiety or insomnia, and subjects with all three or none of the symptoms. Depressed mood Anxiety Insomnia All three symptoms No symptoms (n = 4314) (n = 6498) (n = 5550) (n = 1670) (n = 19116) Mean SD Mean SD Mean SD Mean SD Mean SD Age (years) 57.2 4.9 57.0 4.8 57.8 5.1 56.9 4.8 57.8 5.1 Energy (kcal/day) 2877 813 2888 801 2828 818 2886 863 2793 777 Alcohol consumption (g/day) 21.7 26.2 21.5 25.1 22.0 25.4 24.3 28.5 16.5 19.8 BMI (kg/m 2 ) 26.3 3.9 26.2 3.9 26.1 3.9 26.1 3.8 26.3 3.7 Total serum cholesterol (mmol/l) 6.16 1.19 6.22 1.18 6.15 1.19 6.13 1.21 6.26 1.16 Serum HDL-cholesterol (mmol/l) 1.24 0.36 1.26 0.36 1.27 0.37 1.27 0.38 1.23 0.34 Table 2 Baseline daily food consumption and nutrient intake of subjects self-reporting depression, anxiety or insomnia, and all three or none of the symptoms. Depressed mood (n = 4314) Anxiety (n = 6498) Insomnia (n = 5550) All three symptoms (n = 1670) No symptoms (n = 19116) Mean SD Mean SD Mean SD Mean SD Mean SD Fish (g) 39.3 30.2 39.9 30.2 40.1 30.3 40.3 32.9 39.3 29.8 Milk (g) 212 315 203 316 226 321 219 325 220 322 Coffee (ml) 595 374 601 372 567 364 583 382 609 349 Meat (g) 78.0 38.4 80.2 38.8 77.6 38.6 78.0 37.8 78.6 37.2 Vegetables (g) 256 103 264 104 253 104 255 106 263 101 Margarine (g) 11.5 21.1 11.5 20.8 10.7 20.3 11.8 21.3 10.2 19.8 Carbohydrate (g) 308 97.7 309 96.8 300 96.7 304 97.6 303 94.0 Protein (g) 105 30.6 105 30.2 103 31.2 105 31.6 103 28.9 Fat (g) 125 41.8 125 41.8 123 42.2 125 43.4 122 40.6 Sugar (g) 38.5 27.5 38.3 28.0 36.9 26.7 37.7 27.6 38.1 26.5 Lysine (g) 6.42 1.97 6.44 1.95 6.37 2.01 6.44 2.04 6.30 1.86 Serine (g) 4.12 1.31 4.27 1.30 4.22 1.33 4.28 1.35 4.18 1.24 Tryptophan (g) 1.28 0.38 1.29 0.38 1.27 0.39 1.29 0.40 1.26 0.36 Omega-3 fatty acids (total) (g) 2.21 0.93 2.24 0.92 2.16 0.92 2.23 0.97 2.14 0.87 Omega-3 fatty acids (from fish) (g) 0.47 0.28 0.48 0.29 0.48 2.89 0.49 0.30 0.46 0.28 Omega-3 fatty acids (from vegetables) (g) 1.77 0.82 1.79 0.80 1.70 0.80 1.77 0.86 1.70 0.77 Omega-6 fatty acids (g) 10.12 6.82 10.14 6.70 9.70 6.65 10.17 7.05 9.44 6.31 Omega-6/omega-3 ratio 4.47 2.00 4.45 2.01 4.41 2.03 4.50 2.30 4.34 1.85 Folic acid (μg) 342 106 344 105 335 106 340 107 336 103 Vitamin D (μg) 5.59 3.21 5.65 3.18 5.60 3.18 5.72 3.23 5.45 3.08 In subjects with depressed mood, the mean intake of omega-6 fatty acids was 7% greater than in symptom-free subjects. In individuals with anxiety, the mean intake of omega-6 fatty acids was 7% greater and that of omega-3 fatty acids from vegetables 5% greater than in subjects with no symptoms. Intake of fish or omega-3 fatty acids from fish were not associated with anxiety or depressed mood. When the symptoms reported during the first trial follow-up year were taken into analysis, 782 men reported depressed mood, 1237 feelings of anxiety, 1234 insomnia, and 166 men all three symptoms on all four occasions. The mean intake of energy was 7% greater in subjects reporting all three symptoms repeatedly compared with symptom-free individuals. Subjects with insomnia consumed 11% less coffee but 10% more milk than those with no insomnia. Both in subjects with depressed mood and with anxiety, the mean intake of total omega-3 fatty acids was 9% greater and that of omega-3 fatty acids from vegetables 6% greater than in respective symptom-free subjects, whereas the mean intake of omega-6 fatty acids was 6% greater in subjects with depressed mood and 9% greater in subjects with anxiety. Discussion Our subjects reporting anxiety had higher intakes of omega-3 and omega-6 fatty acids, but omega-3 fatty acids from fish were not linked to anxiety. Margarine was the main source of both omega-3 fatty acids from vegetables and omega-6 fatty acids. Subjects with depressed mood also had a higher intake of omega-6 fatty acids. Because 3138 (73%) subjects with depressed mood also had feelings of anxiety, it may be that anxiety is the dominant symptom, and the greater intake of omega-3 and omega-6 fatty acids is primarily related to feelings of anxiety. Previously, it has been suggested that omega-3 fatty acids may alleviate the effects of depressive symptoms but not those of mania [ 20 ]. Recently, we have reported that the low dietary intake of omega-3 fatty acids is not associated with depression [ 21 ]. Our present results show now that individuals suffering from symptoms of depressed mood have higher intakes of omega-6 and omega-3 fatty acids. More investigation is needed to elucidate the specific effects of omega-3 fatty acids on mood. Subjects with any or all of the symptoms consumed more alcohol than the symptom-free subjects. Subjects with all three symptoms consumed most alcohol of all, and they received 6% of their total energy from alcohol, compared with 4% in subjects with no symptoms. Energy from alcohol, however, did not explain the differences in the mean intake of energy between groups. Body-mass index was lower, despite a higher caloric intake, in subjects with any of the symptoms compared with symptom-free subjects. Subjects reporting insomnia drank more milk than symptom-free subjects, but less coffee. Warm milk has long been taken as a self-medication for insomnia, and our finding among those with insomnia accords with this traditional habit. In addition, they avoided consuming large amounts of coffee, which is known to have impact of sleep. We also found that subjects reporting depressed mood consumed more carbohydrates than subjects with no symptoms. This finding is consistent with the attempt by depressed subjects to alleviate the carbohydrate craving associated with symptoms of depression. Tryptophan intake showed no association with mental wellbeing in our study population. Interestingly, a number of negative studies has been published recently, suggesting that the effects of tryptophan depletion on mood are inconsistent [ 22 - 24 ], and the rationale for augmentation has now been challenged [ 25 ]. The intakes of vitamin D and folic acid exceeded the daily recommendations and showed no association with mental wellbeing. Neither did the consumption of fish, milk, meat or vegetables. Limitations There are some limitations in our study. Our study was a cross-sectional study, and it cannot provide causal evidence on the association between the diet and symptoms of depression, anxiety or insomnia. The study participants included only men, aged 50 to 69 years, and all were smokers. Our exclusion criteria limit the generalization of our findings, but the study still provides valid and reliable data on a community-based, homogenous sample of older men. Dietary intake and alcohol consumption were assessed with a validated food use questionnaire to measure the habitual dietary intake over the previous year as completely as possible. For most nutrients, both the reproducibility and the validity of this method are 0.6 to 0.7 [ 18 ]. For example, they are 0.66 and 0.73 for energy intake, 0.88 and 0.85 for alcohol, 0.70 and 0.75 for carbohydrates, and 0.70 and 0.64 for vitamin D, respectively. The assessment of self-reported depression was based on a single item only that might have compromised the specificity, but not sensitivity. For example, two questions only may be as effective as more detailed screening instruments in detecting probable cases of major depression [ 26 ]. One of these questions ("During the past month, have you often been bothered by feeling down, depressed, or hopeless?") is rather similar to the item that we applied for being indicative of depressed mood. Conclusion The scientific examination of relationships between nutrition and mental wellbeing is a relatively new area of study. Most of the studies focus on the use of dietary supplements, which provide more concentrated amounts of specific nutrients than most food sources. There are few data evaluating food consumption and nutrient intake among subjects with compromised mental health. Our main finding was that we did not find any association between omega-3 fatty acids from fish and mental wellbeing. In general, more attention need to be paid to the intake of nutrients in patients suffering from symptoms of depression, anxiety or insomnia. Further studies are needed to clarify complex associations between the diet and mental wellbeing, and to elucidate their mechanisms of action.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC519023.xml

529309

SIVdrl detection in captive mandrills: are mandrills infected with a third strain of simian immunodeficiency virus?

A pol-fragment of simian immunodeficiency virus (SIV) that is highly related to SIVdrl-pol from drill monkeys ( Mandrillus leucophaeus ) was detected in two mandrills ( Mandrillus sphinx ) from Amsterdam Zoo. These captivity-born mandrills had never been in contact with drill monkeys, and were unlikely to be hybrids. Their mitochondrial haplotype suggested that they descended from founder animals in Cameroon or northern Gabon, close to the habitat of the drill. SIVdrl has once before been found in a wild-caught mandrill from the same region, indicating that mandrills are naturally infected with a SIVdrl-like virus. This suggests that mandrills are the first primate species to be infected with three strains of SIV: SIVmnd1, SIVmnd2, and SIVdrl.

Findings To date over 30 strains of simian immunodeficiency virus (SIV) have been isolated from African primate species and sequenced [ 1 ]. Mandrills ( Mandrillus sphinx ) are quite exceptional among African monkeys in that they harbour two distinct SIV strains, designated SIVmnd1 and SIVmnd2, with a separate geographic distribution [ 2 , 3 ] (see also Figure 1A ). SIV infections are mostly non-pathogenic in their natural hosts. SIVmnd1, despite high virus levels in chronically infected mandrills, has only a small effect on the T-cell counts, and primary infection does not induce clinical symptoms [ 4 , 5 ]. However, two cases of immunodeficiency were reported in mandrills after long-term (>18 years) SIV infection [ 6 ]. In 2003, a 20-year old male captive mandrill (mandrill CAS) housed at Artis Zoo (Amsterdam, The Netherlands), suffering from heart failure and poor general condition, was found to be positive for serum SIV antibodies. Inspection of the other three mandrills of his group, a ten-year old female (mandrill REB) and their offspring (mandrills RAF, 3 years old and HAB, 2 months old), showed that the female and one of the offspring (HAB) were also SIV antibody-positive. Figure 1 A) Geographic distribution of the genus Mandrillus , based upon mitochondrial haplotypes (adapted from [10]), and the SIV strains they harbour. B) Phylogenetic tree generated with Kimura-2-parameter distances and the NJ option of the MEGA package , from 267 nt of the mitochondrial cytochrome b gene of the four captive mandrills and reference sequences from GenBank (accession numbers are indicated). C) Phylogenetic tree generated with the NJ option of MEGA, based upon Kimura-2-parameter distances of SIVpol nucleotide fragments from captive mandrills CAS and REB, and reference sequences for SIVdrl, SIVmnd1, and SIVmnd2, respectively. Numbers shown are bootstrap confidence levels (BCL). D) SIV-pol amino acid consensus sequence from captive mandrills compared with homologous sequences from SIVdrl and SIVmnd2, respectively. The translated SIV-Artis sequence is the consensus sequence of 14 PCR clones derived from two animals. The YMDD motif within the catalytic core of the RT enzyme is underlined. Both EDTA-plasma and PBMC (isolated with the OptiPrep system (Nycomed, Oslo, Norway)) was obtained from the animals for further analysis. Our goal was to investigate whether the monkeys were virus carriers, and which strain of SIV they harboured. To detect both SIVmnd1 and SIVmnd2, we designed two nested primers sets based on published pol gene sequences that amplify an RT fragment of 282 nucleotides (Table 1 ). Nucleic acids were isolated from PBMC by a procedure using silica and guanidium thiocyanate [ 7 ]. cDNA was synthesized with the 3'primer and AMV-RT (Roche Diagnostics, Penzberg, Germany). PCR amplifications were performed using the following protocol: denaturation for 5 min at 95°C and amplification for 35 cycles (first PCR) or 25 cycles (second PCR) of 1 min at 95°C, 1 min at 55°C, and 2 min at 72°C, followed by an extension of 10 min at 72°C. Products were cloned with the TOPO TA cloning kit (Invitrogen, San Diego, Calif.). Sequencing of at least four clones per sample was done with the Bigdye Terminator Cycle Sequencing kit and an ABI 377 automated sequencer (both from ABI, Foster City, Calif.), using M13 forward and M13 reverse primers. For species identification, fragments of the mitochondrial 12S and cytochrome B genes were amplified [ 8 ], and sequenced. Table 1 PCR primers used to amplify Mandrillus SIV-pol Primer Sequence ('5→'3) Description Fragment size SIVmnd1A AGATATAGGGGATGCCTATT 5' first primer A-B = 356 nt SIVmnd1B TCTTCCACTTATCTGGGTGT 3' first primer SIVmnd1C AGATTATAGACCCTATACTGC 5'second primer C-D* = 282 nt SIVmnd1D CATCCAATGAAAGGGAGGTTC 3' second primer SIVmnd2A GGACATAGGGGATGCCTATT 5' first primer A-B = 356 nt SIVmnd2B CTGTCCATTTCTTTGGGTGC 3' first primer SIVmnd2C GGACTTTAGAAAGTACACTGC 5'second primer C-D* = 282 nt SIVmnd2D CATCCACTCAAAGGGAGGTTC 3' second primer SIVdrlA GGATGTAGGTGATGCCTATT 5' first primer A-B = 356 nt SIVdrlB CTGTCCACTTCTTTGGATGC 3' first primer SIVdrlC = SIVmnd2C 5'second primer C-D* = 282 nt SIVdrlD CATCCATTCATAAGGAGGATTG 3' second primer * corresponding to nucleotides 2503–2784 of SIVdrl (acc. no. AY159321) Mitochondrial 12S and cytochrome B sequences were identical in all four animals and confirmed that the monkeys were M. sphinx [ 9 , 10 ]. In addition, the cytochrome B sequences were indistinguishable from the recently described northern mandrill haplotype (Figure 1B ) [ 10 ], suggesting that the captive animals descended from founders originating from a locale north of the Ogooué River (see Figure 1A ). SIV-pol fragments could be amplified from PBMC of both adult mandrills CAS and REB with the primer set specific for SIVmnd2, but not with SIVmnd1 specific primers. However, analysis of the cloned fragments showed that these were 96–97% identical to SIVdrl-1FAO (GenBank acc. no. AY159321) isolated from a drill monkey ( Mandrillus leucophaeus ), with a lower sequence identity to SIVmnd2 (± 85% to GenBank acc. no. AF367411), and to SIVmnd1 (<64% to GenBank acc. no. M27470). Although SIVmnd2 and SIVdrl are more closely related to each other than the two SIVmnd strains, SIVdrl has several mismatches with the PCR primers designed for detection of SIVmnd2. This could explain why one seropositive animal was tested as PCR-negative. Therefore, we designed a new primer set that amplifies the same gene fragment based on the SIVdrl sequence (Table 1 ). Reanalysis of the mandrill PBMC samples with this drill-specific primer set again resulted in only two positive samples from the two adult animals: mandrills CAS and REB. Sequence analysis confirmed the high similarity to SIVdrl-1FAO (97%), and a lower similarity to SIVmnd2 (87%). SIV pol fragments from both mandrills were 98–99% similar to each other. Clones obtained from a single animal with the two primer sets were not identical to each other (98–99% identity), suggesting that each PCR amplified a subset of the virus population (Figure 1C ). Mandrills presently inhabit Cameroon, Gabon, and the southwestern part of the Republic of Congo. Two mitochondrial haplotypes are described in this species, separated by the Ogooué River in Gabon [ 10 ]. Interestingly, the distribution of mandrill SIV strains follows approximately the same geographic distribution, with SIVmnd1 being present in the southern part of the mandrill range, and SIVmnd2 in the northern part. Drill monkeys are found in Nigeria and Cameroon separated from the mandrill territory by the Sanaga River (Figure 1A ). Mandrills and drills are currently believed to be non-sympatric, but it is not unlikely that the situation was different in the past. SIVmnd2 and SIVdrl are closely related, and both are equidistant from SIVmnd1. SIVmnd2 is found in northern mandrills, which are closest to the current drill habitat. A wild-caught mandrill from south Cameroon was found to harbour a SIVdrl virus strain [ 3 , 11 ], suggestive of cross-species transmission [ 11 ]. Multiple cross-species transmissions are now believed to obscure the evolution and distribution of SIV strains in African primate species [ 1 ]. SIV cross-species transmissions are ongoing, and African green monkey strains have recently been detected in patas monkeys and baboons [ 1 , 12 , 13 ], species that are found in close proximity to each other. All four mandrills examined here were born in captivity. Male CAS was born in 1983 at the now closed Wassenaar Zoo (The Netherlands), and moved to Artis Zoo in 1986. The female, REB, was born in Budapest Zoo (Hungary), and moved to Artis Zoo when she was 5 years old. Their offspring, RAF and HAB, were both born in Amsterdam. Exposure to drills during their lifetime is unlikely as none of the zoos kept drills ( M. leucophaeus ) at any time. Drills are rare in European zoos, and only the zoos of Nikolaev (Ukraine) and Saarbruecken (Germany) reported keeping both drills and mandrills in a 1992 survey [ 14 ]. So, it is improbable that the Artis mandrills acquired SIVdrl from a recent contact with captive drills. Another way of acquiring a drill SIV strain could be if one of the monkeys was actually a hybrid between a drill and a mandrill. Hybridisation between different species of Cercopithecinae is possible, and offspring is sometimes fertile depending upon the exact species. The genus Mandrillus cannot hybridise in the wild, as the habitats of the two species do not overlap, but it does so in captivity. The morphologic differences between female drills and mandrills are less obvious than those between males and are mainly noticeable in the colouration of the muzzle and the size of the animal. A single hybrid M. sphinx × M. leucophaeus has been reported from Vienna Zoo, Austria, in a 1992 survey [ 14 ], and two hybrid M. leucophaeus × M. sphinx were described from a Wildlife Rescue Centre in Cameroon [ 15 ]. Mitochondrial sequencing as performed in this study cannot alone be used to resolve hybridisation, as it only characterises the mother lineage. The male mandrill is, however, unlikely to be a hybrid as its description fits exactly that of a male mandrill. The female was also listed in the Artis Zoo database records as a non-hybrid. She was registered with the European studbook programme (ESB) for mandrills supervised by the Budapest Zoo, Hungary, and was born from registered mandrill parents. Because SIVdrl was also found in a wild-caught Cameroonian mandrill [ 3 , 11 ], it is plausible that a SIVdrl-like virus is naturally present in mandrills, making them the only primate species that is naturally infected with three strains of SIV. The presence of SIVdrl in one of the two adult captive mandrills occurred probably through transmission from a wild ancestor, and this animal probably infected the other mandrill once joined in Artis Zoo. Sexual and mother-to-child transmission of SIV in mandrills have been reported to be rare [ 16 , 17 ]. Here, in offspring born to a SIV-positive mother SIV could not be detected. The persistence of maternal antibodies could explain why the 2-month old young tested seropositive, although there is a possibility of the animal having a viral copy number below the detection limit of the PCR assay. Transmission of SIV between mandrills and drills could have taken place either by biting or sexual contacts. SIVmnd1 is mainly transmitted between males during aggressive contacts [ 17 ], and might also be transmitted when fighting off other receptive primates species. Sexual transmission could have occurred during interbreeding. Each of these possibilities requires an overlap of habitats, which could have existed in the past. If we assume that SIVdrl(Artis) left Africa at least ten years ago (when the female was born in captivity), its genome conservation is remarkable: only 2 conserved amino acid differences separate the consensus pol-sequence from the SIVdrl reference sequence (Figure 1D ). However, to gain a deeper insight into the characteristics and evolution of this virus strain, a full-length sequence, or at least additional sequences of the gag or env regions, would be required. Only then could it be determined whether the virus carried by the captive mandrills is really SIVdrl or a novel recombinant virus. Several recombinant SIVs have been described in naturally infected primate species (see: [ 1 ]). Unfortunately, further sequence analysis is difficult due to shortage of material as the monkeys were euthanized soon after the SIV antibody test results became known. Competing interests The author(s) declare that they have no competing interests. Authors'contributions ACvdK designed the study, analysed the sequences, and drafted the manuscript. RvdB carried out the PCR assays and performed the cloning and sequencing. MJH did the medical examinations and collected the blood samples. RAG carried out the SIV antibody assays. ADMEO and BB conceived of the study, and participated in its coordination.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529309.xml

555956

Influence of immunomagnetic enrichment on gene expression of tumor cells

Background Metastasis is the leading cause of cancer-related death. Bone marrow (BM) is a frequent site for the settlement of disseminated tumor cells which occurs years before overt metastases signal incurability. Methods Here we describe a new method to assess the initial stage of metastasis development in cancer patients. By immunomagnetic selection with HER2/neu and EpCAM as catcher antigens single disseminated tumor cells can be enriched from BM samples. To examine whether the immunomagnetic enrichment technique may change gene expression in the selected tumor cells, we performed differential expression profiling with the breast cancer cell lines MCF-7 and BT474 as models. The profiles were performed using 1.2 Cancer Arrays (Clontech) containing 1176 cDNAs that can be grouped into different functional categories, such as signal transduction, cell cycle, adhesion, cytoskeleton plasticity, growth factors and others. Results The reproducibility of the gene expression profiling between duplicate cDNA-array experiments was assessed by two independent experiments with MCF-7 breast cancer cells. Scatter blot analysis revealed a good reproducibility of the cDNA array analysis (i.e. less than 10% difference in the gene expression between the arrays). Subsequent comparative cDNA-array analyses of immunobead-selected and unselected MCF-7 and BT474 cancer cells indicated that the antibody incubation during the immunomagnetic selection procedure did not considerably alter the gene expression profile. Conclusion The described method offers an excellent tool for the enrichment of micrometastatic tumor cells in BM without largely changing the gene expression pattern of these cells.

Introduction Solid tumors derived from epithelial organs are the main form of cancer in industrialized countries. The first phase of the metastatic development consists of local tumor cell invasion, followed by tumor cell circulation in the blood and homing to secondary distant organs [ 1 , 2 ]. As indicator organ for early systemic dissemination of epithelial tumor cells to distant sites, BM has played a prominent role [ 3 ]. BM can easily be aspirated from the iliac crest and single metastatic cells are already present in 20–40% of patients with epithelial tumors (e.g., breast, lung or colon carcinomas) years before overt distant metastases occur in the skeleton or other distant organs [ 3 - 5 ]. The molecular description of these cells has been, however, hampered by the low concentration of these cells (e.g., 10 -5 -10 -6 per BM cell). To predict and monitor therapeutic responses the assessment of the gene expression profile of disseminated tumor cells seems to be of utmost importance. However, it is uncertain to which extent incubation with antibodies used for immunomagnetic isolation of these cells might affect their expression profile. We have addressed this aspect, using monoclonal antibodies against two prominent antigens, EpCAM and HER2/neu that are frequently and independently expressed on micrometastatic tumor cells [ 3 ]. Materials and methods Ficoll density gradient centrifugation and immunocytochemistry The enrichment of tumor cells from BM by Ficoll density gradient centrifugation and the immunocytochemical detection of epithelial tumor cells in cytological BM preparations has been described elsewhere in detail [ 5 , 6 ]. Immunomagnetic cell separation and immunocytochemistry of BM samples Two ml of a BM sample usually containing 2 × 10 6 mononuclear cells were washed with Hank's Salt Solution (Biochrom KG, Germany). The pellet was resuspended in 2 ml Hanks and 3.2 × 10 7 (80 μl) CELLection™ and pan-mouse immunomagnetic beads (Dynal, Oslo, Norway) coated with anti-EpCAM (MAb 3B10) and anti-HER2/neu (MAb 7C1) antibodies (Micromet, Munich, Germany) were added. All solutions and cell preparations were kept at 4°C during the whole procedure to avoid nonspecific binding of immunomagnetic beads. After an incubation time of 30 min at 4°C and 20 min at room temperature on a rotating mixer the magnetically labeled cells were isolated in a magnetic particle concentrator and resuspended in 200 μl bead removing buffer (40 mM Tris, 10 mM MgSO 4 , and 1 mM CaCl 2 , pH7.4, prewarmed to room temperature). The immunomagnetic beads were removed by DNase treatment with 15 μl DNase (50 U/μl) at room temperature for 15 min. After separation in a magnetic particle concentrator the supernatant was collected and centrifuged onto glass slides. Tumor cells were identified by immunostaining with monoclonal anti-cytokeratin antibody A45-B/B3 according to the manufacturer's instruction (Micromet, Munich, Germany). Cytokeratins are specific constituents of the epithelial cytoskeleton and they have become the marker antigen of choice for the detection of disseminated epithelial tumor cells in mesenchymal organs such as BM [ 3 , 7 ]. To avoid unspecific binding of the antibody via Fc-receptors present on leukocytes, we used F ab fragments of A45-B/B3 that were directly conjugated to the marker enzyme alkaline phosphatase. Cell culture and antibody incubation MCF-7 cells and BT474 cells were maintained in RPMI (Invitrogen, Karlsruhe, Germany) supplemented with 5 % glutamine (Invitrogen) and 10 % FCS (C-C Pro, Neustadt, Germany). MCF-7 and BT474 cells (ATCC HTB-22 and HTB-20) were allowed to reach a logarithmic growth phase in culture. At 90 % confluency, the cells were incubated for 30 min at 4°C and 20 min at room temperature in Hanks containing 1 μg/ml of anti-EpCAM (MAb 3B10) as well as anti-HER2/neu (MAb 7C1) antibodies according to the immunomagnetic cell separation protocol. In a negative control experiment, cells were suspended in Hanks devoid of antibodies. cDNA probe preparation and hybridization Total RNA was isolated using the peqGold TriFast™ (Peqlab, Erlangen, Germany) according to the manufacturer's instruction. In order to remove genomic DNA contamination, a DNase step was included using the DNA-free™ kit (Ambion, Cambrigeshire, England) according to manufacturer's instructions. RNA was dissolved in RNase-free H 2 O with 1 U/μl RNase inhibitor (SUPERase. IN™, Ambion). 5 μg purified total RNA was used for [α- 33 P] dATP (3000 Ci/mmol, 10 μl; Amersham, Freiburg, Germany) labeled cDNA synthesis as previously described [ 8 ]. The cDNA probe was purified with nucleotide removal columns (Qiagen, Hilden, Germany). The Atlas Human 1.2 Cancer Arrays (Clontech, Heidelberg, Germany) were hybridized according to the manufacturer's protocol. cDNA- array data analysis The membranes were exposed to phosphoimager plates (Raytest Isotopen-Meβgeräte, Straubenhardt, Germany) for 3 days, and plates were scanned with the phosphoimager Fuji Bas (Raytest) at a 100 μm-resolution. The images were analyzed using the Imagene 5.5 software (Biodiscovery, CA, USA). The data of the arrays were normalized on the basis of the genes ubiquitin, HLAC and beta actin. The ratio between antibody-treated and non-treated cells was calculated for each gene. Ratios lower than 0.5 or higher than 2 were considered as differentially expressed if at least one sample showed an expression above 0.5. We performed duplicates of each experiment and created scatter blots with the SPSS software for windows. Results In a recent work [ 9 ] we demonstrated that our immunomagnetic separation works on clinical samples (Figure 1 ) and is superior to the standard Ficoll density centrifugation technique, used in most previous studies on cancer micrometastasis [ 4 , 5 , 7 , 10 ]. Figure 1 CK-positive cells detected after immunomagnetic enrichment of BM from breast cancer patients. One single cell and one 2-cell cluster is shown in a 400× magnification. To test whether the antibody incubation during the immunomagnetic enrichment approach affects gene expression in the selected cells, we applied cDNA-array analysis. We subsequently evaluated whether the immunomagnetic enrichment method affected gene expression in the selected cells. This aspect is of utmost importance for further molecular description of disseminated tumor cells and has not been addressed before. The profiles were performed using 1.2 Cancer Arrays (Clontech) containing 1176 cDNAs that can be grouped into different functional categories, such as signal transduction, cell cycle, adhesion, cytoskeleton plasticity, growth factors and others [ 11 ]. As models, we used the breast cancer cell lines MCF-7 and BT474 (ATCC HTB-22 and HTB-20), because they express heterogeneous levels of the target antigens HER2/neu and EpCAM comparable to micrometastatic breast cancer cells in vivo [ 3 ]. The reproducibility of the gene expression profiling between duplicate cDNA-array experiments was assessed by two independent experiments with MCF-7 breast cancer cells. As shown in Figure 2A , scatter blot analysis revealed a good reproducibility of the cDNA array analysis (i.e. less than 10% difference in the gene expression between the arrays). We plotted the data of the antibody-treated and untreated MCF-7 (B) or BT474 (C) cells two dimensionally in a scatter plot; y-axis represents the data of untreated cells and the x-axis represents the data of cells treated with anti-HER2/neu or anti-EpCAM. For both cell lines, the scatter plots show that expressed genes in antibody-treated versus untreated cells (Figure 2B, C ) was in principal within the range observed in the duplicate experiments with MCF-7 cells (Figure 2A ). However, subtle changes in the expression of individual genes after antibody incubation were observed in particular in BT474 cells. In this cell line 38 genes were strongly differentially expressed (ratio >3) in antibody-treated versus untreated cells (Table 1 ). Most of these genes play a role in extracellular matrix remodeling, signal transduction and replication, as well as repair and transcription. MCF-7 cells showed in this experimental approach 31 differentially expressed genes with a ratio of over 3 (data not shown). Although similar group of genes were affected, only 3 common genes (CDC7, SGI and KIR) were differentially expressed in both cell lines after antibody incubation. Figure 2 Representative scatter blots of breast cancer cells using the 1.2 Cancer Array for expression analysis. (A) untreated MCF-7 breast cancer cells (results of duplicate experiments), (B) antibody-treated versus untreated MCF-7 cells, and (C) antibody-treated versus untreated BT474 breast cancer cells. Table 1 Genes differentially expressed in antibody-treated versus untreated BT474 cell Genes GenBank Accession# Ratio * Extracellular matrix remodeling : COL11 J04177 - 6.2 MMP17 X89576 - 4.7 MMP16 D50477 -4.6 SPARC J03040 5.7 Adhesion: PKD1 U24497 - 7.3 NCAM AF002246 7.3 M-cadherin D83542 3.5 Cytoskeleton plasticity: SPTA1 M61877 7.4 Signal transduction: RGS4 U27768 - 10.1 GAS L13720 - 7.1 BMP1 U50330 - 7.0 FGFR4 L03840 - 6.5 PMEL17 M77348 - 6.5 ETS-1 - 5.9 ERBB2 M95667 - 5.8 TGF-beta X02812 - 5.2 N-ras X02751 - 4.8 HRS D84064 - 4.7 KIR U10550 9.0 BMP6 M60315 7.5 BIN1 U68485 7.4 SGI Y00064 4.1 CDC7 AF015592 3.8 CNTF S72921 3.4 SH3BP2 AF000936 3.1 Apoptosis: CD27BP U82938 6.0 DR5 F016268 4.8 Metabolism: PPAT U00238 6.3 HPRT P00492 5.4 Immune response: MHC class I U65416 8.4 Replication/repair/transcription: CHAF1A U20979 - 12.3 NEK3 Z29067 - 5.6 BTG U72649 8.9 HRC1 M91083 6.2 TOP1 J03250 4.9 CLK1 L29222 4.0 Functionally unclassified PIG7 AF010312 - 4.9 menin U93236 3.1 *Ratio of normalized data from antibody-treated versus untreated BT474 cells as described in the Materials and Methods section. Negative values indicate downregulated and positive values upregulated genes. Discussion Here, we investigated whether an immunomagnetic enrichment procedure for micrometastatic cancer cells present in BM aspirates leads to significant changes in the gene expression pattern of the enriched tumor cells. In order to mimic the biological conditions of a tumor type with frequent BM involvement, we used two breast cancer cell lines (MCF-7 and BT474). Both cell lines expressed the target antigens, EpCAM and HER2/neu, for immunomagnetic separation at different levels [ 12 ] and they were incubated with the anti-EpCAM and anti-HER2/neu antibodies according to the same immunomagnetic enrichment protocol used for the BM samples from cancer patients analyzed recently [ 9 ]. It has been shown by other groups that some of the mAb directed against HER2/neu (e.g., Herceptin R ) can specifically block cell proliferation and affect gene expression in HER2/neu-positive breast cancer cells [ 13 , 14 ]. Furthermore the incubation of human cells with anti-EpCAM-specific mAbs (e.g. KS1/4 mAb) can induce considerable changes in the expression of insulin and glucagons [ 15 ]. However, our present results suggest that the two antibodies against EpCAM and HER2/neu used for the immunomagnetic selection process did not considerably influence the gene expression pattern of the enriched cells, although the HER2/neu- positive cell line showed a slightly increased number of differentially expressed genes. These genes are involved in extracellular matrix remodeling, signal transduction and replication, repair and transcription, and they were either up or downregulated after antibody incubation. For example, HER2/neu gene expression was downregulated after antibody binding, as expected from reports in the literature [ 16 ]. Taken together, we cannot exclude subtle changes in the expression of individual genes after antibody incubation, but we observed no obvious shift in the expression pattern that exceeds the normal variability of duplicate experiment. Thus, we conclude that the immunomagnetic selection protocol described here might be useful for experimental approaches aimed to determine the gene expression profile and genome of disseminated CK-positive cells [ 17 , 18 ]. As we performed our study only on two breast cancer cell lines, larger series of similar experiments with further cancer cell lines as well as with enriched tumor cells from the blood or BM must be investigated to draw firm conclusions. The detection and characterization of micrometastatic cancer cells will provide new insights into the biology of the metastatic process in cancer patients. This will lead to an improved molecular staging of cancer patients and to the identification of new biological targets for adjuvant systemic therapies aimed to eradicate micrometastatic disease before the onset of overt metastasis signals incurability.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555956.xml

544873

The Latin American Social Medicine database

Background Public health practitioners and researchers for many years have been attempting to understand more clearly the links between social conditions and the health of populations. Until recently, most public health professionals in English-speaking countries were unaware that their colleagues in Latin America had developed an entire field of inquiry and practice devoted to making these links more clearly understood. The Latin American Social Medicine (LASM) database finally bridges this previous gap. Description This public health informatics case study describes the key features of a unique information resource intended to improve access to LASM literature and to augment understanding about the social determinants of health. This case study includes both quantitative and qualitative evaluation data. Currently the LASM database at The University of New Mexico brings important information, originally known mostly within professional networks located in Latin American countries to public health professionals worldwide via the Internet. The LASM database uses Spanish, Portuguese, and English language trilingual, structured abstracts to summarize classic and contemporary works. Conclusion This database provides helpful information for public health professionals on the social determinants of health and expands access to LASM.

Background Public health practitioners have long recognized the connections between patients' socioeconomic conditions and their health [ 1 - 8 ]. Yet these practitioners and their empirically oriented researcher colleagues have faced difficulties in establishing the precise linkages between socioeconomic variables and sub-optimal health status. Social medicine is a diverse field that studies these relationships between society (and its socioeconomic conditions) and the health of populations. In Latin America, social medicine consists of a widely respected and influential field of research, teaching and professional practice [ 9 ]. Professionals working in this field seek to identify and to understand better the linkages between socioeconomic conditions and patients' health. Until recently, however, most of the knowledge base in this discipline has remained largely unknown outside Latin America. Language barriers and disincentives to distribute this information more widely are two major reasons for this lack of awareness. Some readers might have first learned about Latin American social medicine (LASM) through recent critical reviews [ 9 ] or through a special issue of the American Journal of Public Health that focused on LASM [ 10 , 11 ]. LASM traces its historic origins to European researchers such as Rudolf Virchow and the belief systems of indigenous cultures. Both the European and indigenous sources of current social medicine practices emphasized the importance of linking social conditions to health status. Contemporary social medicine in Latin America continues to emphasize these linkages between social conditions and the health of populations. Social medicine professionals participate in a wide array of settings in Latin America and represent diverse specialties. Their integration into healthcare systems has varied by era and country [ 12 ]. Construction and content Innovative approaches to disseminating work in LASM have become increasingly available due to Internet technology. The project "Enhanced Access for Latin American Social Medicine" at The University of New Mexico with funding from the U.S. National Library of Medicine seeks to make information on the connections between social conditions and health problems available to a wide audience. The project has sought to bridge the prior information gap primarily through delivering structured abstracts of social medicine publications in Spanish, English, and Portuguese via the Internet on the LASM database beginning in 2001. Other goals of this project include: publishing full text social medicine electronic journals on behalf of medical societies in Latin America; and, maintaining a repository for key classic and contemporary social medicine publications. Structured abstracts are posted in three languages in the LASM database at The University of New Mexico for both the classic and contemporary social medicine literatures. The first phase of this project involved preparing and posting the structured abstracts of 25 landmark books, 50 book chapters, and 100 journal articles from the classic social medicine literature in Spanish, Portuguese, and English. A peer selection committee identified and agreed upon the specific selections of classic books, book chapters, and journal articles to be abstracted for this project. Table 1 lists the members of this committee, representing institutions in Brazil, Chile, Colombia, Cuba, Ecuador, Mexico, the United States, and Venezuela [ 13 ]. Representative examples of some of the classic books [ 14 - 16 ], book chapters [ 17 - 19 ], and articles [ 20 - 22 ] can be found in the list of references following this article. Table 1 Members of the Peer Selection Committee Country Name Institution Brazil Emerson Elias Merhy University of Campinas, São Paulo Chile Alfredo Estrada L. Investigation and Training Group in Social Medicine, Santiago Colombia Saul Franco Agudelo National School of Public Health, Bogotá Cuba Francisco Rojas Ochoa National School of Public Health, Havana Ecuador Jaime Breilh Health & Research Advisory Center, Quito Mexico Ángeles Garduño Autonomous Metropolitan University-Xochimilco, Mexico City Mexico Asa Cristina Laurell Secretariat of Health, Mexico City Mexico Francisco Mercado Martínez University of Guadalajara, Guadalajara Peru Marcos Cueto Peruvian University "Cayetano Heredia," Lima United States Elizabeth Fee National Library of Medicine, Bethesda, MD United States Norman Frankel American Medical Association, Chicago, IL United States Allen Jones American Public Health Association, Washington, DC United States Antonio Ugalde University of Texas, Austin, TX Venezuela Oscar Feo Department of Public Health, Maracay Venezuela Maria Urbaneja Latin American Social Medicine Association and Foreign Ministry of the Venezuelan national government, Caracas The contemporary literature summarized in the LASM database has been drawn primarily from the 12 journals currently or previously published in Latin America. Table 2 lists the 12 journals, with their titles translated into English. These specific journals also have been identified and approved by the peer selection committee. The website that hosts the LASM database provides further information about this committee's members, including their institutional affiliations and areas of research. The peer selection committee consists of experts in social medicine and information technology. The LASM steering committee based at The University of New Mexico meets with the peer selection committee twice a year via online conferencing to decide on selection policies, actual lists of resources slated for inclusion in the LASM database, and administrative matters regarding the project [ 23 ]. Table 2 Current journal subscriptions monitored for noteworthy articles on social medicine Country Title Argentina Cuadernos Médico Sociales (Medico-Social Notebooks) 197?- Argentina Salud Problema y Debate (Health – Problem and Debate) + Brazil *Cadernos de Saúde Pública (Notebooks of Public Health) 1985- Brazil Ciência & Saúde Coletiva (Science and Collective Health) 1996- Brazil Interface (Interface) 1997- Brazil Revista Brasileira de Epidemiología (Brazilian Journal of Epidemiology) 1998- Brazil Saúde e Sociedade (Health and Society) 1992- Brazil Saúde em Debate (Health in Debate) 1976- Cuba Archivos del Ateneo Juan César García (Archives of the Juan César García Circle) 2000- Cuba *Revista Cubana de Medicina Tropical (Cuban Journal of Tropical Medicine) 1996- Cuba Revista Cubana de Salud Pública (Cuban Journal of Public Health) 1975- Mexico Salud Problema (Health – Problem) + Notes: * Indexed by the MEDLINE database from the National Library of Medicine + Start date occurred during 1900s but exact date not known As a pilot, the project has made two social medicine journals available in electronic full text format. Several issues of these journals, Saúde em Debate (Brazil) and SaluCo Bulletin (Cuba), are available from the host website. The host website also posts structured abstracts of important articles from these full text journals. The LASM database encompasses several broad themes within LASM. Subject areas include the history, theories, methodologies, and organizational dimensions of social medicine. Other subjects pertain to institutional analysis, social/critical epidemiology, and strategic planning. Table 3 summarizes the specific topics emanating from these broad subject areas. Readers will recognize that many of the subjects have direct bearing on the health of populations, such as health disparities and managed care. Table 3 Subjects covered by Latin American social medicine Health policy analysis Violence and health Medical education reform Mental health services and mental health policies Primary care research and preventive services Determinants of mental illness in race, ethnicity, social class, or gender Strategic planning Indigenous, complementary and herbal medicine Environmental health Social, environmental and nutritional causes of infant and perinatal mortality Labor and health Economic development, demographic change, and aging Ethnic/racial disparities and health Socioeconomic barriers to cancer prevention Social class disparities and health Social processes of alcohol and drug abuse Gender disparities and health Chronic illnesses Infant and perinatal mortality Urban health International health Managed care The LASM database is a web application developed using the Cold Fusion application server. The data are stored in a Microsoft SQLServer relational database. The data sources are the original Latin American publications, which are summarized in Spanish, Portuguese, and English languages in structured abstract format. The English-language structured abstracts are quality checked by the principal investigator, who reviews the abstracts for substantive content, and then by the librarian investigator who reviews the structured abstracts for final quality assurance purposes. Each record contains fields for author, title (book, book chapter, or article), place of publication, publisher and structured abstracts in each language. Records contain volume, number, and pagination when applicable. All records contain terms the from Medical Subject Headings (MeSH) system, a controlled vocabulary developed and maintained by the National Library of Medicine. Users can search the database on controlled fields of author, title, and MeSH terms in any of the three available languages. Full text abstract searching is also available. In addition to the abstract searching facilities, the LASM database can be browsed alphabetically by title. The browsing interface offers a convenient way to become familiar with the extent and diversity of the LASM literature. Utility and discussion From January 2002 through December 2003, a total of 17, 853 visits were made to the website hosting the LASM database. The largest numbers of visits in descending rank order originated from Brazil, Mexico, Argentina, Spain, and Colombia. A preliminary, qualitative evaluation has been favorable. Following completion of this project, we will conduct and publish a comprehensive summative evaluation. A total of 250 structured abstracts in Spanish, Portuguese, and English had been posted to the LASM database as of June 30, 2004. The host website presents more detailed information about this project, as well as the structured abstracts themselves. The LASM database comprises a dynamic and searchable web application containing structured abstracts in English, Portuguese, and Spanish. It is designed using industry standard web application design principles, but its content is unique to the LASM domain. A database searching design and utility problem unique to the LASM database, and other web applications like it, relates to the problem of web-based multilingual searching. Widely available search engines cannot preprocess a multilingual language translation of search terms (e.g. retrieving results containing the Spanish word "pública" for the English search term "public"). Additionally, search engines do not recognize that the unaccented "publica" (as it might be entered by an English speaker as a search string) might be the same word as the Spanish accented "pública" and therefore will not return the user's expected search result. Key combinations and modifiers that are used to create special and accented characters in word processing programs like Microsoft Word often do not work in browser based search and form fields. Our primary instruction to users for constructing search terms containing special and accented characters (i.e., diacritics) suggests that they use a separate text editor that accepts keyboard modifiers to construct special and accented characters to build the search term and then copy and paste the completed search string into the field. On our search pages we also list common special and accented characters in Spanish and Portuguese that users can copy and paste into browser based search fields. Although there are no other alternatives for entering special and accented characters into browser form fields, both methodologies are somewhat cumbersome. Therefore, to improve the searching utility of the LASM database we also preprocess entered search strings to allow the search engine to perform selected character substitution on the entered search string and attempt to solve the accented character problem from the search engine side. All search strings are first passed through a regular expression routine, which substitutes single character search engine wildcards for the following characters: A, E, I, O, U, a, e, i, o, u, N, C, n, c. This effectively removes potential special and accented character misspellings. To return to the previous example, a user might enter the search string "publica" ("public" in English) in a search field as an attempt to find article titles that contain the Spanish or Portuguese word "pública". Properly spelled, "Pública" uses the accented character "ú" rather than "u" (ASCII 163 rather than ASCII 117). Unfortunately, current search engines are not intelligent enough to infer that "pública" is a match for the search string "publica". Therefore, the literal search for publica (unaccented) will not return any search results that contain "pública" although there are hundreds of instances of the word "pública" in the LASM database. In the case of "publica", the preprocessed search string that is finally submitted to the search engine is "p*bl***". The search engine will return all seven-letter word matches that contain the letters p, b, and l in the first, third, and fourth positions respectively. The net effect of this technique is to under-specify the search result. That is, the search engine may possibly return records that contain other words that happen coincidentally to match the submitted search string. On the other hand, the returned result set can be guaranteed to contain the desired search result. Problems created from under-specifying the search are limited based on experience gained from using this technique. The positional constraints of submitted characters within the search string generally are restrictive enough to prevent most problems. Specifically, within a limited domain search surface like the LASM database, the likelihood of the occurrence of most alternative word matches is very low. The character substitution methodology described here is not perfect, and many other alternative strategies for addressing the multilingual search problem have been explored by LASM technical staff. Most of the alternative strategies considered, however, involved much higher costs in terms of acquiring or developing specialized search engine capabilities or much higher abstract preparation costs. Therefore, we chose the character substitution strategy as a compromise between implementation cost and search utility. Conclusions Internet technology via websites and web browsers has created numerous opportunities for public health colleagues to inform one another and to collaborate across wide geographic space. The LASM database clearly demonstrates the efficacy of the Internet for communicating its informative structured abstracts posted in the Spanish, Portuguese, and English languages. This database provides useful information that would otherwise be unavailable for public health professionals on the social determinants of health. Furthermore, it expands access to LASM through its inclusion of both the classic and contemporary literature. Availability and requirements Anyone with access to the World Wide Web and a web browser can access all structured abstracts in the LASM database . As the data reported above indicate, the LASM database already has been accessed steadily since its initial small-scale publicity began in 2002. We hope that public health readers will utilize the LASM database to improve their research, teaching, and practice. Competing interests Neither the authors nor any support personnel possess competing interests related to the LASM database or to this article's publication. Authors' contributions JE conceived of and wrote the initial version of this article, served as an early collaborator on and helped design the project, was responsible as an investigator for acquiring journals for the Latin American social medicine collection, provided quality assurance editing on structured abstracts, played a major role in the formative evaluation of the project, and coordinated all revisions to this article. HW initiated and designed the project, served as principal investigator, obtained funding, translated and edited abstracts in the English-language, and edited this manuscript. HSB served as co-principal Investigator, helped design this project, obtained funding, and led the effort for the formative evaluation of this project. JT helped administer the project and edited this manuscript. CI coordinated the project, wrote Spanish-language structured abstracts, and provided reference materials for this article. KW and JT developed all programming aspects of the website including the search strategies, managed the web interface and underlying database and contributed the text for portions of this article. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544873.xml

529447

Validity of the Clock Drawing Test in predicting reports of driving problems in the elderly

Background This study examined the use of the Folstein Mini Mental Status Exam (MMSE) and the Clock Drawing Test (CDT) in predicting retrospective reports of driving problems among the elderly. The utility of existing scoring systems for the CDT was also examined. Methods Archival chart records of 325 patients of a geriatric outpatient clinic were reviewed, of which 162 had CDT results (including original clock drawings). T -test, correlation, and regression procedures were used to analyze the data. Results Both CDT and MMSE scores were significantly worse among non-drivers than individuals who were currently or recently driving. Among current or recent drivers, scores on both instruments correlated significantly with the total number of reported accidents or near misses, although the magnitude of the respective correlations was small. Only MMSE scores, however, significantly predicted whether or not any accidents or near misses were reported at all. Neither MMSE nor CDT scores predicted unique variance in the regressions. Conclusions The overall results suggest that both the MMSE and CDT have limited utility as potential indicators of driving problems in the elderly. The demonstrated predictive power for these instruments appears to be redundant, such that both appear to assess general cognitive function versus more specific abilities. Furthermore, the lack of robust prediction suggests that neither are sufficient to serve as stand-alone instruments on which to solely base decisions of driving capacity. Rather, individuals who evidence impairment should be provided a more thorough and comprehensive assessment than can be obtained through screening tools.

Background Assessment of cognitive function pertaining to capacity for safe and independent living among elderly patients is a central responsibility of many geriatric medical clinics and service agencies. Specific concerns pertaining to judgments of driving capacity are also befalling upon the medical profession in primary care settings [ 1 ]. To aid in this task, a number of brief assessment screens are often employed to identify cognitive problems that may be indicative of a range of pragmatic concerns, including driving capacity [ 2 ]. Specifically, results on assessment instruments purported to assess attention, reaction time, and visuospatial abilities are often used to inform clinical judgment about driving capacity in such settings. Two such screening instruments typically used to gauge general cognitive function, and inform questions pertaining to driving capacity specifically, are the Folstein Mini Mental Status Exam (MMSE) [ 3 ] and the Clock Drawing Test (CDT). The MMSE is a widely used cognitive screening tool, due to its brevity, ease of administration, and relative breadth [ 4 ]. Numerous studies over the past 40 years have supported its utility as a valid and reliable indicator of general cognitive function [ 5 ]. The MMSE consists of 30 items comprising subscales assessing orientation, word registration, attention (via a serial sevens or spelling task), word recall, and language. Additionally, a figure copy exercise is included to examine visuospatial abilities. The CDT is hypothesized to assess more specific aspects of planning, organization and visuospatial skill. Directions for completing the CDT involve asking a patient to draw the face of a clock, including the numbers, and then to place the hands to designate a certain time, such as "ten minutes after eleven." Although different scoring templates for the CDT exist, most often code for features such relative size, spacing and placement of numbers or hands, disorganization, perseveration, completeness, and other potential errors that are hypothesized to indicate cognitive impairment [ 6 - 8 ]. In addition to the MMSE, results on the CDT are often used in clinical settings to inform clinical impressions pertaining to whether or not patients are impaired to such an extent that they should not be driving [ 9 ]. Although empirical reviews note that performance on the CDT should only be examined in conjunction with other assessments in this regard, anecdotal evidence also suggests that the CDT is often used as a stand-alone instrument to inform judgments of driving capacity, in both medical and non-medical settings. Despite this apparent widespread use, there appears to be a dearth of research addressing the validity of the CDT for detecting driving impairment. Although a small number of studies exist that suggest CDT scores may relate to driving problems, the size of this literature base coupled with methodological concerns indicate a need for further research. For example, one study [ 10 ] examined the effectiveness of the CDT and MMSE, in addition to the Trail Making Test, Part A [ 11 ] and a visual acuity test, in predicting driving ability as judged by a driving instructor after participants completed a road test. A discriminant function analysis indicated that the set of test scores and participant age correctly identified 80% of drivers judged to be impaired, and 85% of drivers judged not to be impaired, according to driving instructor assessments. The authors reported that the discriminant model did not include the MMSE, however, because it did not add significant discriminatory power. The authors then suggested that the overall battery may be useful as a screening instrument in primary health care settings for detecting potential problems in driving that would warrant further examination. Separate univariate data on the predictive power for each of the separate instruments, however, was not provided. Additionally, the authors incorporated a 4-point scoring system for the CDT that was created for the study and differs from scoring systems used in other studies. Furthermore, given that only the component instruments are typically used in practice as opposed to the more extensive batteries advocated, the unique predictive power of the CDT warrants further investigation. Additional evidence for the potential utility of the CDT in predicting driving behaviors is provided in an examination of neurophysiologic phenomena related to caregiver reports of driving impairment in 79 individuals with Alzheimer's disease [ 12 ]. Single photon emission computerized tomography was incorporated to examine brain function. Additionally, scores on the MMSE, CDT, and caregiver ratings of driving ability were analyzed. CDT scoring was based upon a 5-point system that was constructed for the study. Results indicated that MMSE scores did not significantly differ between individuals based upon driving ability, but that CDT scores were predictive of driving impairment based upon level of impairment and whether participants were instructed to simply copy an existing clock, or construct their own according to specific directions. Furthermore, imaging also indicated that level of driving impairment related positively to changes in cortical function. These authors hypothesized that cognitive tests assessing visuospatial abilities and executive function may thus show greater discriminative power between driving impaired and non-impaired subjects than MMSE scores, which may be impacted to a greater extent by other non-relevant verbal tasks. The validity of the scoring system constructed for the CDT in comparison to other scoring systems, however, was not further explored. A pilot study examining the comparability of simulated driving tests in predicting actual driving problems also suggested that CDT scores may be significant predictors [ 3 ]. A small sample of nine older adults was incorporated, four of whom were classified as cognitively impaired based in part on abnormal CDT and MMSE scores. It was found that simulated driving tasks correlated moderately with actual driving problems across the groups. No data was provided, however, on the extent to which the CDT or MMSE uniquely predicted impairment. Given the typically low rate of follow-up for patients referred for more formal driving assessments, it would be beneficial to further investigate the relations between scores on the CDT and reports of actual driving problems. Furthermore, the predictive power of the CDT alone and in conjunction with other assessment tools in predicting reported driving problems has yet to be fully assessed. Additionally, previous studies examining CDT scores and driving behaviors have employed markedly small sample sizes, warranting future research with greater numbers of participants. Finally, previous studies differ in terms of what, if any, scoring systems were used to score the clock drawings. Thus, further investigation of the comparability of different scoring systems is needed. To address these concerns, this study explored the relations of patient scores on the CDT and MMSE to patient or family reports of driving problems. In so doing, the utility and comparability of three scoring systems for the CDT that are commonly used by researchers and practitioners, namely the Shulman et al. [ 6 ], Sunderland et al. [ 7 ], and Wolf-Klein et al. [ 8 ]systems, were also examined. Specifically, the Shulman system incorporates a 1–6 rating scale, where higher scores indicate higher levels of impairment. Conversely, scores on the Wolf-Klein and Sunderland systems range from 1–10, with lower scores indicating greater levels of impairment. Although specific criteria differ, each system codes for elements pertaining to spacing, organization, and comprehension of the task, among other criteria. Exploratory analyses also were conducted to examine the predictive utility of the CDT and MMSE in predicting whether driving problems, namely accidents or near misses, were reported. Further analyses examined whether linear relationships existed between CDT and MMSE scores and the reported number of such incidents. Finally, regression tests examined whether the CDT and MMSE uniquely predicted the number of reported incidents. Methods IRB approval was obtained for the study, and data was collected from archival records of patients seen over a 10-year period at a geriatric assessment center of a general teaching hospital in the Midwest. The center operated as a full-service outpatient clinic, where new patient assessments included a full medical and psychosocial history. This history included patients' and collateral others' reports of driving behaviors within the past year, including whether patients were currently driving or had recently stopped driving within the past year, and number of driving accidents or near misses. The data was often collected during the initial intake assessment, when both the patient and available family members or caregivers were interviewed by a geriatrician, social worker, and/or a nurse specialist. In addition, the MMSE and CDT were typically administered to patients to assess cognitive functioning. Chart records did not clarify whether the reported driving problems were acknowledged by the patient or caregiver, nor the extent to which any discrepancies existed, but rather only reported the number of incidents. The content of the incidents was also not always documented, but examples that were provided typically included crashes or minor accidents for which the patients were at fault. Nevertheless, despite the subjectivity inherent to such reports, it was the intent of these authors to remain true to the figures documented in the patient charts. Indeed, given that medical professionals typically must rely to some extent on subjective reports of patients or caregivers during intake evaluations to inform initial judgments about patient safety, it was decided that incorporation of such data in the present study would nonetheless be useful. Data was obtained from charts of 325 patients, including 162 original clock drawings that were scored according to the systems provided by Shulman et al. [ 6 ], Sunderland et al. [ 7 ], and Wolf-Klein et al. [ 8 ]. Two advanced students in psychology were trained in each of the three scoring systems, and subsequently scored the clocks independently of each other and blinded to information about driving. MMSE scores, driving status, and reports of driving problems were also coded for subsequent analyses. The initial sample consisted of 81 men and 242 women (gender data was unavailable for 2 individuals). Of these, 287 (88.3%) were Caucasian, 34 (10.5%) were African American, and one individual was Asian American. Ethnicity data was not available for the other three individuals. The mean patient age was 79.75 ( SD = 6.67), and ranging from 58 to 99 years of age. As is typical of many outpatient geriatric populations, there was a range in type and severity of presenting concerns, with some patients reporting relatively few problems and others evidencing diagnoses of vascular dementia, Alzheimer's disease, or depression in addition to other health concerns. Of these, concerns due to cognitive function predominated; approximately 60% of the patient sample was referred to the clinic for evaluation of memory loss, cognitive decline, or dementia. MMSE data was available for 311 patients; of these, 159 also had CDT data sufficient for analysis. Of the 162 charts that had CDT data, only 3 did not also have MMSE data. Results and discussion The raters' corresponding CDT scores for each scoring system correlated above 0.70, suggesting adequate inter-rater correspondence. The corresponding scores for each scoring system were then averaged to create three composite scores for each clock drawing, one for each scoring system. Descriptive data pertaining to scores for the overall sample on the MMSE and CDT is provided in Table 1 . Table 1 Descriptive statistics for overall sample scores on cognitive measures Test N Mean SD Range MMSE 311 21.51 6.15 0–30 CDT (Shulman Score) 162 3.87 1.25 1–6 CDT (Wolf-Klein Score) 162 6.58 2.10 1–10 CDT (Sunderland Score) 162 6.35 2.49 1–10 Initial exploratory t -tests were conducted to examine whether CDT scores and MMSE scores differed between individuals who had been currently or recently driving, versus those who had not been reported to be driving for a more extensive time period. In each case, current and recent drivers evidenced better scores on all of the cognitive measures than individuals who had not been driving. Results for these analyses are provided in Table 2 . Table 2 Mean differences in CDT and MMSE scores based on driving status Variable N Mean SD t df MMSE Score Currently or Recently Driving 114 24.32 4.87 6.41** 305 Not Currently driving 193 19.98 6.18 Shulman Score Currently or Recently Driving 61 3.39 1.28 -3.94** 157 Not Currently driving 98 4.16 1.15 Wolf-Klein Score Currently or Recently Driving 61 7.26 1.83 3.29** 157 Not Currently driving 98 6.18 2.13 Sunderland Score Currently or Recently Driving 61 7.18 2.28 3.28** 157 Not Currently driving 98 5.89 2.49 Note. **p < .01. Further analyses examined whether CDT or MMSE scores predicted the presence of reported driving problems among individuals who had been current or recent drivers. Patients who had not been driving for a longer period of time were excluded from the analyses, since no driving problems would have been reported as a function of not driving. Specifically, t -tests were incorporated to examine whether CDT and MMSE scores differed among individuals for whom driving problems had been reported, versus those with none. Drivers with reported problems evidenced significantly lower MMSE scores, but no significant differences were obtained for CDT scores. Nevertheless, the trends for the overall mean differences on CDT scores, although small, were in the same direction as the findings for the MMSE. Specifically, in every case the CDT scores for each scoring system were worse for drivers with reported problems than those with none. Overall, these results suggest that the presence of driving problems may have been reflective of greater levels of cognitive impairment, although the overall differences reflected in CDT scores were nonetheless very small in magnitude. These results are detailed in Table 3 . Table 3 Mean differences in CDT and MMSE scores based on presence of reported driving problems among current or recent crivers Variable N Mean SD t df MMSE Score Did Report Problems 51 23.08 6.03 -2.44* 112 Did Not Report Problems 63 25.32 3.41 Shulman Score Did Report Problems 27 3.57 1.35 1.03 59 Did Not Report Problems 34 3.23 1.22 Wolf-Klein Score Did Report Problems 27 7.09 2.25 -.64 59 Did Not Report Problems 34 7.40 1.43 Sunderland Score Did Report Problems 27 6.72 2.64 -1.41 59 Did Not Report Problems 34 7.54 1.91 Note. *p < .05. Next, the linear relations for both CDT and MMSE scores in predicted the number of reported problem incidents were examined. Specifically, correlation coefficients were calculated separately for number of reported accidents or near misses, and scores on the CDT and MMSE. Patients who had not been currently or recently driving were excluded from the analysis, since no problems would have been reported if they had not been driving. The number of reported incidents correlated significantly and positively with the level of cognitive impairment as measured by MMSE and CDT scores. Additionally, each of the CDT scoring systems appeared to evidence similar predictive utility, as they correlated highly (above 0.80). Means, standard deviations, and correlations for these variables are provided in Table 4 . Given that not all patient charts necessarily contained all of the requisite MMSE and CDT data, cases that were missing data were excluded from some of the cells. Thus, the n of the resultant cases is included for each cell. Table 4 Means, standard deviations, and correlations for cognitive measures and reported number of problems among current or recent drivers Variable N M SD 1 2 3 4 1. MMSE Score 114 24.32 4.87 2. Shulman Score 61 3.39 1.28 -.45** (59) 3. Wolf-Klein Score 61 7.26 1.83 .50** (59) -.80** (61) 4. Sunderland Score 61 7.15 2.30 .58** (59) -.82** (61) .83** (61) 5. Reported Number of Driving Problems 116 .62 .90 -.27** (110) .23* (57) -.24* (57) -.27* (57) Note. ** p < .01, *p < .05. The N for each cell is provided in parentheses. Total reported number of driving problems ranged from 0–4 for each patient. Finally, hierarchical regression analyses examined whether MMSE or CDT scores uniquely predicted number of reported accidents or near misses. The non-significant R-squared change term in the second step of each regression indicates that neither the MMSE nor set of CDT scores predicted significant incremental variance. Thus, it appears that the variance in reported accidents or near misses predicted by the MMSE and CDT was redundant. Regression results are provided in Table 5 . Table 5 Hierarchical regression analyses predicting number of reported driving problems from CDT and MMSE scores among current or recent drivers ( N = 54) Regression Criterion and Steps R R 2 F df R 2 change F change Reported Number of Accidents or Near Misses Step 1: CDT Scores .25 .06 1.17 3,51 .06 1.17 Step 2: MMSE Score .31 .10 1.37 1,50 .04 1.92 Step 1: MMSE Score .30 .09 5.19* 1,53 .09 5.19* Step 2: CDT Scores .31 .10 1.37 3,50 .01 .18 Note. *p < .05 Conclusions The results of this study suggest that both the MMSE and the CDT appear to have only limited utility in predicting retrospective reports of driving problems among elderly drivers. The finding that MMSE and CDT scores were worse among patients who had not been currently or recently driving may be due to a number of factors, including the possibility that some individuals may have never driven at all before. Nevertheless, it appears likely that many of these individuals probably had been driving in the past, but may have since stopped due to problems related to cognitive impairment. This assertion is supported by the finding that individuals who had been currently or recently driving at the time of the assessment, and who had lower MMSE scores, were more likely to have had reports of accidents or near misses. Although similar mean tests with the CDT were not significant, it is notable that the direction of the obtained differences for each scoring system of the CDT was consistent with the findings of the MMSE. Furthermore, the relatively modest n -sizes within each cell may have limited statistical power. More robust findings were obtained, however, for the correlations examining number of reported accidents or near misses to CDT and MMSE scores. Among individuals who had been currently or recently driving at the time of assessment, greater levels of cognitive impairment as evidenced by MMSE and CDT scores also predicted greater numbers of reported accidents or near misses. This finding held regardless of which CDT scoring system was incorporated, suggesting that each may have equal utility. Finally, the results of the regression analyses appear to indicate that the predictive power of the CDT and MMSE are somewhat redundant, since neither added significant incremental variance to prediction. Although it is possible that the regressions may have had limited power to detect significant incremental differences due to the relatively small sample sizes, in each case the increment to R-squared was small nonetheless. Thus, it appears that both the MMSE and CDT served as gross assessments of general cognitive function, versus more specific cognitive capacities, in predicting reported numbers of accidents or near misses. Although the current results appear to suggest limited predictive utility for the MMSE and CDT in predicting driving problems, an additional cautionary note is in order. The significant predictive power for each instrument as demonstrated by the magnitudes of the correlation coefficients was nevertheless small. Furthermore, significant predictive utility was not obtained for every test in the current research. Additionally, the use of a retrospective design does not necessarily allow for definitive conclusions about predicting instances of future driving problems. Thus, although poor CDT or MMSE scores appear to indicate greater potential for driving problems, the current data do not support the use of the CDT or MMSE alone in making definitive decisions pertaining to driving competence. Rather, the empirical findings of the current research appear to best support the use of the CDT and MMSE solely as their originally intended purpose as screening tools. Thus, scores evidencing impairment on either of these instruments may indicate a need for driver caution, followed by more comprehensive and extensive assessment of driving capacity on which to base decisions regarding safety. As such, the role and utility of these instruments in predicting driving problems may be more fully understood through future research that incorporates a prospective design, along with a more comprehensive assessment of specific and relevant cognitive skills (like psychomotor speed or executive function) and objective assessment of driving abilities (such as can be obtained through simulated or practice driving situations). Competing interests The author(s) declare that they have no competing interests. Authors' contributions ND, AG, and JM conceived of the study purpose and design. ND, EB, and MM collected, entered, and analyzed the data. ND conducted the literature review and critique, and drafted the manuscript. AG, JM, EB, and MM provided comments on the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529447.xml

340127

Wnt Signaling Relies on Nuclear Armadillo

A couple of years ago, a paper was published in a high-profile journal that challenged a long-established model of cell signaling. While researchers in the field mostly greeted the results with skepticism, some went into the lab to investigate the discrepancy. Many elements of this pathway, called the Wnt pathway, have been well characterized. The standard model of Wnt signaling holds that when the Wnt protein binds to its receptor, it initiates a labyrinthine signaling cascade that sends a protein called β-catenin into the cell's nucleus where, together with a protein complex, it initiates transcription. In the absence of this signal, β-catenin binds to an inactivating complex in the cytoplasm and is targeted for degradation. The paper that disputed this view suggested that β-catenin can effect gene expression without entering the nucleus and that it can activate the Wnt pathway while tethered to the cell membrane. Before that paper was published, Nicholas Tolwinski and Eric Wieschaus had shown that β-catenin, also known as Armadillo (Arm) in the fruitfly, is sent into the nucleus in response to Wnt signaling. Upon entering the nucleus, Arm interacts with a second protein complex to activate transcription. Now Tolwinski and Wieschaus have reexamined the function of Arm in the fruitfly and have demonstrated that the pathway “in fact does depend on the nuclear localization of β-catenin.” While their paper was in the final stages of acceptance, the dissenting paper was retracted, after it was learned that the results had been fabricated. Tolwinski and Wieschaus' findings confirm what had already been known about Arm's role in Wnt signaling and also fill in important details about how it works. Multicellular organisms rely on elaborate communication networks of signaling proteins and enzymes to exchange information between cells. The Wnt signaling pathway regulates the expression of a host of different genes during embryogenesis to control body patterning and cell differentiation in organisms from fruitflies to mammals. Miscommunications in this tightly regulated pathway contribute to a variety of developmental defects and cancers. In the developing fruitfly, Wnt signaling is normally restricted to the front of each larval segment, where it produces a smooth surface; the rear of the segments, where Wnt signaling is absent, is hairy. If Wnt signaling is overexpressed, it produces fruitfly larvae with only smooth segments; lack of Wnt signaling produces hairy segments. Using the smooth phenotype as a measure of Wnt signaling, Tolwinski and Wieschaus delved deeper into the role of Arm in this signaling process. These experiments are complicated because Arm functions not just in Wnt signaling, but also in cell adhesion. The trick is to make the endogenous Arm (the version encoded by the fly genome) defective for signaling, while leaving the cell adhesion functions fairly normal. Set against this “background,” an additional Arm protein is expressed that is tethered to the membrane; it still retains the protein domains required for signaling, but it's stuck on the inside of the membrane and can't move into the nucleus. In this background, any signaling in response to Wnt must mean that Arm can signal to the nucleus without actually having to get inside. Tolwinski and Wieschaus prove—again—that this is not the case. They do this by showing that the weak, medium, and strong endogenous Arm mutants—these classifications reflect the severity of the mutations' effects—have different effects on signaling in the presence of the membrane-tethered Arm. It's clear, they argue, that tethered Arm cannot signal on its own and must somehow be helping the weaker mutants signal. To further investigate how the tethered Arm activates the endogenous mutants, Tolwinski and Wieschaus developed two new and “cleaner” Arm mutants that impair Arm's signaling ability but have no effect on its cell adhesion function. The tethered Arm could not produce a completely smooth phenotype with these nonsignaling endogenous mutants. These experiments, the authors conclude, indicate that the tethered form of Arm produces its transcriptional effects by activating the endogenous Arm protein. Normal activation of the pathway liberates Arm proteins from the inactivating complex, which allows them to enter the nucleus and activate transcription. Tethered Arm appears to accomplish this by sequestering the inactivating complex at the cell membrane, preventing it from interfering with endogenous Arm. Even though Tolwinski and Wieschaus started these experiments based on what turned out to be fabricated results, their investigations produced valuable contributions. They not only reaffirm the standard model of Wnt signaling, but reveal important new insights into the workings of a major player in the pathway.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC340127.xml

545210

Predicting Tumor Responses to Gefitinib and Erlotinib

Tyrosine kinases regulate signaling pathways that control cell growth, proliferation, motility, and other critical cellular processes. Mutations in tyrosine kinase genes can lead to abnormal kinase activity, and some tumors become dependent upon this activity for growth and survival. Thus, kinases are attractive targets for anti-cancer drugs. Examples of new kinase inhibitors include gefitinib and erlotinib, which have recently shown promise in treating non-small-cell lung cancer. Unfortunately, gefitinib and erlotinib work only in a subset of patients, and they can have severe side effects, albeit infrequently. So researchers have been trying to find ways to predict who will benefit from therapy with these drugs and who won't. Assessing lung tumors for gene mutations could help guide therapy Following the work of Lynch et al. (N Engl J Med 350: 2129–2139) and Paez et al. (Science 304: 1497–1500), William Pao and colleagues have previously shown that the epidermal growth factor receptor (EGFR), a tyrosine kinase, is often mutated in non-small-cell lung cancers, and that tumors that harbor such mutations are sensitive to gefitinib and erlotinib. In this new study, they focused on a signaling protein called KRAS, which functions downstream of many tyrosine kinases, including EGFR. The KRAS gene is also often mutated in lung cancers, but very few cancers have mutations in both EGFR and the KRAS gene. To find out whether KRAS mutations could help to predict which patients would respond to gefitinib or erlotinib, the researchers looked for mutations in EGFR and KRAS genes in 60 tumors for which sensitivity to either drug was known. They extended their earlier findings that EGFR mutations (which were found in 17 of the tumors) were associated with sensitivity to the kinase inhibitors, and found that tumors that had mutations in KRAS (a total of nine) were refractory (i.e., did not respond) to either drug. These results need to be validated in larger and prospective trials that use standardized mutation detection techniques. If they are confirmed, knowing the mutation status of EGFR and KRAS in tumors could help physicians decide which patients should receive gefitinib and/or erlotinib. As Inoue and Nukiwa state in a Perspective that accompanies the article, “By combining all the factors that relate to response or resistance, patients who will benefit from treatment can hopefully be identified. Undoubtedly we have taken a great step forward in molecular therapy for lung cancer treatment.”

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545210.xml

524177

Identification of hip fracture patients from radiographs using Fourier analysis of the trabecular structure: a cross-sectional study

Background This study presents an analysis of trabecular bone structure in standard radiographs using Fourier transforms and principal components analysis (PCA) to identify contributions to hip fracture risk. Methods Radiographs were obtained from 26 hip fracture patients and 24 controls. They were digitised and five regions of interest (ROI) were identified from the femoral head and neck for analysis. The power spectrum was obtained from the Fourier transform of each region and three profiles were produced; a circular profile and profiles parallel and perpendicular to the preferred orientation of the trabeculae. PCA was used to generate a score from each profile, which we hypothesised could be used to discriminate between the fracture and control groups. The fractal dimension was also calculated for comparison. The area under the receiver operating characteristic curve ( A z ) discriminating the hip fracture cases from controls was calculated for each analysis. Results Texture analysis of standard radiographs using the fast Fourier transform yielded variables that were significantly associated with fracture and not significantly correlated with age, body mass index or femoral neck bone mineral density. The anisotropy of the trabecular structure was important; both the perpendicular and circular profiles were significantly better than the parallel-profile ( P < 0.05). No significant differences resulted from using the various ROI within the proximal femur. For the best three groupings of profile (circular, parallel or perpendicular), method (PCA or fractal) and ROI ( A z = 0.84 – 0.93), there were no significant correlations with femoral neck bone mineral density, age, or body mass index. PCA analysis was found to perform better than fractal analysis ( P = 0.019). Conclusions Both PCA and fractal analysis of the FFT data could discriminate successfully between the fracture and control groups, although PCA was significantly stronger than fractal dimension. This method appears to provide a powerful tool for the assessment of bone structure in vivo with advantages over standard fractal methods.

Background The NIH Consensus Statement defines Osteoporosis as "a skeletal disorder characterised by compromised bone strength predisposing to an increased risk of fracture" [ 1 ]. Bone strength was defined as "the integration of two main features: bone density and bone quality". Currently, clinical diagnosis is based solely on bone mineral density (BMD) in accordance with the World Health Organisation guidelines [ 2 ]. Previous studies, however, have found that trabecular bone structure also plays a significant role in determining bone strength [ 3 - 5 ] with BMD explaining only 60 to 80 % of the variability in mechanical resistance [ 6 ]. Trabecular bone structure is visible on standard pelvic radiographs and many attempts have been made to quantify the quality of the structure and assess its relationship to osteoporosis and BMD. These range from visual scoring systems, such as the Singh index [ 7 ], through to sophisticated computerised methods based on fractals [ 8 - 10 ] and other image processing methods [ 11 - 13 ]. A review of the literature suggests that fractal analysis has been a method of choice in recent years for the analysis of trabecular bone structure in CT scans [ 14 , 15 ], MRI [ 16 ], histology [ 17 ] and radiographs [ 18 - 21 ], although it has not been established categorically that it is preferable to other methods of texture analysis [ 22 , 23 ]. By reducing all the information in the image to one descriptor, the fractal dimension [ 24 ], a large part of the information is lost. The Fourier transform of an image expresses the information in the image in terms of spatial frequencies rather than distances. Various methods can be applied to extract information from the Fourier transform [ 25 ], including the fractal dimension [ 24 ]. However such methods have not been fully exploited for analysing bone structure [ 8 , 26 - 30 ]. In this study we investigate the use of Fourier transforms and Principal Components Analysis to generate a mathematical model of the data which can be used to help classify individuals according to the presence or absence of a hip fracture. Principal component analysis (PCA) [ 31 ] is a data reduction technique that has been applied in many fields of study, including investigation of gene expression [ 32 ], development of an electronic nose [ 33 ] and tracing of the evolutionary changes in fish morphometry [ 34 ]. It describes data in terms of a small number of orthogonal, linearly independent components which contain the majority of the information. PCA has no preconditions, such as relying on the data to fit a normal or fractal distribution, but builds a mathematical model based on the correlations present in the data. An eigenanalysis of the correlation or covariance matrix is used to perform PCA. The resulting components are then selected in order of the amount of variance they account for, enabling an efficient mapping of the data. As the first few components account for the vast majority of the variance in the original data, they can be selected for analysis whilst the remainder are discarded as 'noise'. In this way, the number of variables can be greatly reduced whilst maintaining the information present in the original data. In this pilot study we used these methods to investigate the similarities and differences between trabecular bone structure in fracture and control groups using standard radiographs of the proximal femur. Methods Study data A set of digitised standard pelvic radiographs was available from a previous investigation into the morphology of the proximal femur [ 35 ]. These radiographs were taken from an earlier study [ 36 ], that had examined three groups (osteoporotic, osteoarthritic and control) of age matched, postmenopausal women (30 subjects per group). Subjects with osteoarthritis were excluded from the present study. All patients had undergone a scan of the unfractured hip by dual-energy x-ray absorptiometry (DXA) using a Norland XR-26 scanner (CooperSurgical Inc, Trumbull, CT). The controls had had their left hip scanned. All patients and controls had had a pelvic antero-posterior radiograph recorded within a year of the DXA scan. We used those radiographs and the femoral neck BMD (Neck-BMD) data in the current study. A data set of 50 digitised radiographs was available comprising 26 hip fracture patients (HIP) and 24 controls (CNT). The radiographs were digitised, using a Howtek MultiRAD 850 scanner (Howtek, Hudson, New Hampshire) at a resolution of 584 dpi (44 μm per pixel) and a depth of 12 bits. The age, height and weight of each subject were also recorded. Region selection Five regions of interest (ROIs) were selected relative to the principal trabecular systems in locations known to be related to hip fracture via the Singh index [ 7 ] and BMD analysis [ 37 ]. To ensure reproducibility, their locations were determined in relation to the centre and angle of the narrowest part of the femoral neck and the centre and radius of the femoral head on each image, as shown in Figure 1 . Figure 1 Regions of interest. Displays the five regions of interest, upper femoral head (UH), central femoral head (CH), upper femoral neck (UN), Ward's triangle area (WA) and the lower femoral neck (LN) used for analysis. Points A to G are determined by the femoral head and neck and used to locate the ROIs. Points A and E mark the femoral neck width. Points B, C and D lie at 1/4, 1/2 and 3/4 along this line. Point F is the centre point of the femoral head, point G at 1/2 the radius of the femoral head at an angle of 45 degrees to the neck width, 135 degrees to the neck shaft, shown as a dashed line through point C. Each ROI was 256 × 256 pixels (11.3 mm square), to enable use of the fast Fourier transform, and were selected as follows. The upper region of the head (UH) lies on the upper part of the principal compressive trabeculae, the central region of the head (CH) is at the intersection of the principal compressive and tensile trabeculae, the upper region of the neck (UN) lies on the principal tensile trabeculae, the lower region of the neck (LN) is at the base of the principal compressive trabeculae and finally Ward's triangle (WA) which lies between these structures. The points and regions were identified using a macro written for Image Pro Plus software (version 4.1.0.0, Media Cybernetics, Silver Spring, Maryland). The femoral head was described by a best-fit circle, calculated from a series of manually marked points around the outline of the femoral head. Between 15 and 20 evenly spaced points were used to describe the outline, depending on the size of the head. The radius and centre (marked as F in Figure 1 ) of the femoral head were then taken from this circle. The narrowest part of the neck (neck-width) was determined using two automatic edge traces, marking the upper and lower outlines of the femoral neck. The first point and the direction for each trace were marked manually; the edge of the neck could then be identified automatically by the software. The neck width (A – E in Figure 1 ) was calculated by finding the smallest Euclidean distance between the traces. The centre of the neck was located at the mid-point of this line (point C) and the axis of the femoral neck was taken to be a line perpendicular to this through the centre of the neck (dashed line). The top right corner of the WA region was located at the midpoint of the neck width (point C). Points B and D were placed 25% and 75% of the way along the neck width and used as the midpoints of the UN and LN regions respectively. Point F, the centre of the femoral head marked the centre of the CH region and point G, the centre of the base of the UH region. Point G was placed one half of the femoral head radius above point F, at a 45-degree angle to the neck width (A-E). Region analysis Analysis was performed using Matlab software (version 6.1.0, MathWorks Inc, Natick, Massachusetts). A fast Fourier transform was generated for each ROI and three profiles were generated using data from the power spectrum. Firstly a global or circular profile (CircP) was generated, composed of the magnitude at each spatial frequency averaged across all angles, resulting in a profile with 128 data points. To create this profile, each pixel in the Fourier transform was assigned to the integer spatial frequency that most closely matched its' distance from the zero'th component. The angle of preferred orientation was calculated by finding the angle of the maximum value in the power spectrum for the first 25 spatial frequencies [ 38 ]. The maximum value over this range relates to the dominant texture orientation within the image, the trabecular structure. As data in the frequency domain relate to features in the spatial domain rotated by 90°, the median of the values plus 90° was taken as the angle of preferred orientation for each image. Due to the symmetry of the Fourier power spectrum, angles were only calculated between 0° and 180°, rather than 0° and 360°. Two more profiles were then generated, parallel with (ParP) and perpendicular to (PerP) the angle of preferred orientation. In this case the average value was calculated at each spatial frequency from all points lying within ± 5° of the desired angle (Fig. 2 ). Figure 2 Profile generation. (A) Shows a typical region of interest (contrast enhanced for visualisation) showing the trabecular bone structure, in this case aligned approximately 22° to the vertical. (B) The central section of the FFT (128 × 128 pixels). The horizontal and vertical axes have been marked with a mid-grey tone to indicate that they have been excluded from the angle calculation. The bright strip at the centre (running from top left to bottom right) shows the preferred orientation of the trabeculae. Angles calculated from the Fourier power spectrum correspond to the same angles in the spatial domain, rotated by 90°. (C) The pixels with the maximum values are marked using white squares for the first 25 spatial frequency values of the Fourier power spectrum. The median angle, lying 21.8° from the horizontal is shown by a dashed white line. (D)_The regions used to generate the parallel (shaded black) and perpendicular (shaded white) profiles, based on the orientation of the trabecular structure. Principal component analysis Principal component analysis [ 31 ] was used to model statistically the shape of each set of profiles (parallel, perpendicular and circular). This was performed using an eigenanalysis of the correlation matrix. The eigenvectors then become the principal components and are selected in order, depending on their eigenvalue. The eigenvalues are associated with the components in decreasing order, the largest eigenvalue is associated with the first component and the smallest with the last. In order to choose the number of components for analysis, a scree plot [ 31 , 39 ] was generated by plotting the eigenvalues (representing the proportion of variance described by each component) against the component number (Figure 3 ). In each case, the first few principal components were selected for analysis using the scree test [ 39 ] to find an 'elbow' in the slope of the plot. This is used as a threshold between the components that contained useful information, which were then used as input variables for further analysis, and those that could be attributed to noise. Figure 3 Scree plot. Example of a scree plot from the perpendicular profile. The first component typically accounts for the largest amount of variance. The components are chosen to the left of an 'elbow' in the plot. Here components 1 to 5 are included in the analysis as they lie before the 'elbow' at point6 (eigenvalue = 1.63). Fractal analysis Fractal analysis was performed on each profile using a method similar to the Fourier transform technique described by Majumdar et al [ 40 ]. The average power spectrum of the circular profile was plotted on a log-log scale, three approximately linear regions were defined and the gradient ( slope ) of a straight line fitted to each region was found; slopeA , a 'coarse' slope, where the log of the spatial frequency is less than or equal to 1.0, slopeB a 'medium' slope, where the log of the spatial frequency lies between 1.0 and 1.75 and slopeC , a 'fine' slope where the log of the spatial frequency is above 1.75. The fractal dimension was calculated for each slope using the formula suggested by Majumdar et al [ 40 ] Statistical analysis Stepwise discriminant analysis was used to select principal components that could be combined to build a linear classifier. If the stepwise procedure failed to select any components, the most accurate of the individual components was chosen. The same procedure was used to discriminate between the groups using the fractal dimension. Measurement of the area under the ROC curve was used to compare the classifiers built using the discriminant analysis [ 41 ]. A three way ANOVA was applied in order to determine whether there were significant differences between the performance of classifiers depending on the type of analysis, the profile used or the region analysed. Pearson product moment correlation was applied to examine the relationship with age, BMI and Neck BMD for the strongest classifiers. A one-way ANOVA was used to test for significant differences in the performance of the slopes from each spatial frequency band used in the fractal analysis. T-tests, correlation and ANOVA were performed using SigmaStat (version 2.03, SPSS Science, Chicago). Principal component analysis, discriminant analysis, and measurement of the area under the ROC curve were calculated using SPSS (version 10 SPSS Science, Chicago). Results There were no significant differences between the age, height, weight or body mass index (BMI) of the fracture and control groups (Table 1 ). As expected femoral neck-BMD was significantly lower in the fracture group in comparison to the control group ( P = 0.001). Table 1 Summary of anthropometric variables for the fracture and controls groups. Mean and standard deviation (SD) of the age, height, weight, BMI and BMD of the fracture and control groups. P values were obtained from a two-tailed t-test. Variable Control Group (n = 24) Fracture Group (n = 26) Mean SD Mean SD P Age, years 69.1 6.5 69.2 6.3 0.97 Height, cm 158.6 7.1 157.1 0.4 0.38 Weight, kg 63.4 9.5 61.0 9.0 0.38 Body Mass Index, kg/m 2 25.2 3.2 24.8 4.1 0.72 Femoral neck BMD (g cm -2 ) 0.70 0.11 0.604 0.066 0.001 The Receiver Operating Characteristic (ROC) curve is a plot of True Positive Fraction v False Positive Fraction (or Sensitivity v 1 – Specificity). The area underneath the curve ( A z ) represents the performance of the classifier ranging from a value of 0.5 if it is no better than chance to 1.0 for a perfect discriminator. Table 2 shows A z for PCA analysis by region for the circular, perpendicular and parallel profiles respectively, discriminating fracture and control cases. A wide range of values was observed (overall mean 0.70, standard deviation 0.11). Some were little better than chance ( A z = 0.5) (mostly derived from the parallel profile) and the strongest ones were from the perpendicular profiles. The 5 largest areas under the ROC curve were obtained by PCA of the perpendicular profile of the lower neck, upper and central head regions (Table 3 ) ( A z = 0.93, 0.84 and 0.84 respectively), followed by PCA analysis of the circular profile in the upper head region ( A z = 0.76) and, finally, fractal analysis of the parallel profile in the upper neck region ( A z = 0.75). Femoral neck BMD lay between the third and fourth best texture measures ( A z = 0.79 95% CI = 0.66 – 0.91). Plots of the ROC curves for the strongest combinations of image analysis classifier are shown in Figure 4 . Table 2 Classification accuracy for each region-profile combination. Area under the ROC curve for principal component analysis of each profile by region of the femoral neck. Analysis using three-way ANOVA found that the area under the ROC curve was significantly higher in the perpendicular profile than in the parallel profile. ( P < 0.05) Region Circular (95% CI) Parallel (95% CI) Perpendicular (95% CI) Upper head 0.76 (0.63 – 0.89) 0.57 (0.41 – 0.73) 0.84 (0.73 – 0.95) Central head 0.59 (0.43 – 0.75) 0.56 (0.40 – 0.73) 0.84 (0.72 – 0.95) Upper neck 0.72 (0.58 – 0.86) 0.72 (0.57 – 0.86) 0.67 (0.52 – 0.82) Wards triangle 0.74 (0.61 – 0.88) 0.61 (0.45 – 0.76) 0.71 (0.56 – 0.86) Lower neck 0.71 (0.56 – 0.85) 0.55 (0.39 – 0.71) 0.93 (0.87 – 1.00) Table 3 The best five classifiers: Area under the curve and correlation with BMD, age and BMI. Area under the ROC curve ( A z ) for each of the best 5 classifiers and the correlation with age R age , femoral neck BMD (R BMD ) and body mass index (R BMI ) and associated significance values ( P ). Analysis Profile ROI A z (95% CI) R BMD ( P ) R age ( P ) R BMI ( P ) PCA PerP LN 0.93 (0.87 – 1.00) 0.09 (0.55) 0.14 (0.34) -0.08 (0.58) PCA PerP UH 0.84 (0.73 – 0.95) 0.09 (0.52) -0.17 (0.24) -0.03 (0.86) PCA PerP CH 0.84 (0.72 – 0.95) 0.06 (0.70) 0.27 (0.055) -0.11 (0.46) PCA CircP UH 0.76 (0.63 – 0.89) -0.16 (0.28) -0.15 (0.29) 0.07 (0.62) Fractal ParP UN 0.75 (0.61 – 0.89) -0.30 (0.034) 0.25 (0.081) -0.04 (0.78) Figure 4 Comparison of ROC curves. Comparison of the ROC curves for the strongest classifier from the combination of (A) PCA analysis of the perpendicular profile (Lower neck region), (B) PCA analysis of the circular profile (Upper head region) and (C) Fractal analysis of any profile (Upper neck region). Table 3 also shows the correlations between the top five classifiers with age, BMI and Neck-BMD. No significant correlations were found between any of these classifiers and either age or BMI and, for the top three, there was also no significant correlation with Neck-BMD ( P > 0.05). The fifth placed classifier, fractal analysis of the parallel profile in the upper neck region, was the only one significantly associated with Neck-BMD ( P = 0.034). A three-way analysis of variance was used to examine differences in performance due to the region, profile or type of analysis used. It showed that overall PCA analysis performed significantly better than fractal analysis ( P = 0.019) and that analysis of both the perpendicular and circular profiles performed significantly better than the parallel profile ( P = 0.003 and 0.011 respectively). No significant differences were found between the different regions of the femoral neck ( P = 0.241) (despite the apparently large differences in A z ). The power of this test was 0.69, 0.97 and 0.15 for the investigation of differences due to the method of analysis, type of profile used and the region analysed respectively. Table 4 presents the mean A z for the slope from each of the spatial frequency bands for all regions of interest. This was assessed for each profile individually and also for all the profiles together. A one-way ANOVA was used to test for significant differences in Az between slopes A, B and C. In the individual profiles, slopeA performed significantly better than slopeC for the circular profile ( P = 0.008), however when all the profiles were considered, no significant differences were apparent ( P = 0.26). Table 4 Comparing slopeA , slopeB and slopeC . The average and standard deviation of the area under the ROC curve ( A z ) are presented for each of the slopes used in the fractal analysis for all regions of interest. A significant difference was found between slopeA and slopeC in the circular profile, however when all the profiles were compared, no significant differences were found. SlopeA SlopeB SlopeC P All profiles 0.601 (0.074) 0.598 (0.055) 0.565 (0.067) 0.260 Circular 0.670 (0.072) 0.611 (0.022) 0.531 (0. 026) 0.008 Parallel 0.544 (0.042) 0.620 (0.083) 0.563 (0.037) 0.140 Perpendicular 0.589 (0.047) 0.563 (0.032) 0.600 (0.104) 0.678 Discussion and conclusions In these short series, this study found that texture analysis of standard radiographs using the fast Fourier transform can yield variables that are significantly associated with fracture but not significantly correlated with age, body mass index or Neck-BMD. Both PCA and fractal analysis of the FFT data could be used to discriminate successfully between the groups, although overall PCA was significantly stronger than fractal dimension. The best results from this study were not significantly correlated with femoral neck-BMD, age or BMI, indicating their potential for use as an independent predictor of fracture. The radiographic appearance of bone is known to be affected by factors including the size of the patient. As there was no significant difference in the BMI of the fracture and control groups, it is unlikely that this has influenced the results, however it is an issue that will need addressing in future studies. The PCA method extends a method previously developed for analysis of histological sections [ 26 ]. The use of oriented profiles improved the performance of the analysis by selecting directions in which there was the most information about bone structure i.e. perpendicular to the preferred orientation of the trabeculae. PCA considerably reduces the number of variables required to characterise the image via its power spectrum. For example, in this study, we start with a 256 × 256 pixel ROI (65,536 pixels), the Fourier transform is performed and a profile of 128 spatial frequency values is generated. For each profile, PCA was able to describe over 70 % of the variance present in the data using only 5 components or fewer. Overall, the performance of principal components analysis was significantly stronger than that of fractal analysis ( P < 0.01). One advantage of PCA that may contribute to this finding is the ability to summarise the information present in the dataset with a small number of components via an economical mapping of the variance present in the data. In addition, the property of orthogonality between these components ensures that the variables generated are linearly independent (Fig. 5 ). Benefits can also be found by the use of a model built on the mathematical distributions present in the data, rather than expecting the data to meet a given mathematical property, such as fitting a fractal distribution. Figure 5 Plot of two principal components. Example of a scatterplot of two principal components. For FFT/PCA analysis of the upper head region, principal components 4 and 5 were selected by stepwise analysis and are shown here. They are plotted against each other with fracture and control subjects identified using separate markers. The lack of correlation between the components can be seen ( r = 0.040, P = 0.997). Previous studies using non-fractal analysis of the Fourier power spectrum have focussed on images of the spine or wrist, where the alignment of trabeculae is generally orthogonal [ 28 - 30 ]. In such images, analysis of trabecular orientation can be performed by examining the vertical and horizontal sectors as the trabeculae lie predominantly in these directions. The trabecular structure of the femur is more complicated as the trabeculae are aligned in arcs, so the preferred orientation changes throughout the proximal femur. Analysis parallel to the preferred orientation of the trabeculae was significantly poorer than analysis using either the perpendicular or circular profiles ( P < 0.05). Analysis in the perpendicular direction was strongest overall, although it was not significantly better than the circular profile. This accords with the increasingly anisotropic nature of trabecular bone with aging; bone loss is not evenly distributed but is lost primarily at angles perpendicular and oblique to the preferred orientation of the trabeculae [ 30 ]. This loss heightens the risk of fracture, especially if the impact is from the side, as expected from a typical fall from standing height, as there are fewer trabeculae orientated in this direction to absorb the force of impact. In summary, this paper presents a new method for analysing the structure of trabecular bone from standard radiographs. It demonstrates that the Fourier transform can be used to describe structural information in images which may be related to fracture, independently of BMD. This study is limited by the small size of the data set and further analysis is needed to validate these findings. This should be performed on a similar series of radiographs, consisting of fracture and control subjects scanned at the same resolution. The methods from this study could then be applied directly to this group (without recalculating the PCA) to evaluate whether they were generally applicable. However the success of both this and our previous study, using similar techniques to analyse histological sections, indicates that this may be an effective method with clinical utility for describing bone quality statistically in terms of structural parameters. Competing interests The authors declare that they have no competing interests. Authors' contributions Author JG helped design the study, performed the image and data analysis and drafted the manuscript Author AS collected the data/images used within this study and helped with writing of the paper. Author PU assisted with some of the practical approaches, and the writing of the paper Author DMR designed the initial case control study and assisted with interpretation of the results and writing the paper Author RMA helped with the design of the study, the interpretation of the results and the writing of the paper. All authors read and approved the final manuscript Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524177.xml

524188

Extraabdominal fibromatosis in retroperitoneal space

Background Fibromatosis or desmoid tumor covers a broad spectrum of benign fibrous tissue proliferations. It is characterized by infiltrative growth and a tendency towards recurrence; however, unlike sarcoma, it never metastasizes. Case presentation We report on a case of extraabdominal fibromatosis originating from the retroperitoneal space in a 43-year-old woman. Seven years earlier she had undergone ureterolysis and ureteroureterostomy for ureteral obstruction. Computed tomography revealed a tumor between the iliocostalis and the psoas muscle. Histopathological evaluation revealed uniform proliferation of spindle cells, with a moderate amount of collagen fibers, suggesting extraabdominal fibromatosis (desmoid tumor). The tumor was surgically resected, and since then, the patient has remained asymptomatic without any restrictions of daily living activities and without any signs of tumor recurrence during the two-year follow-up. Conclusions Complete resection is the treatment of choice. Adjuvant therapy using non steroidal anti-inflammatory agents, tamoxifen, interferon, anti-neoplastic agents, and radiotherapy, either alone or in combination finds application for unresectable or recurrent cases.

Background The term "fibromatosis" covers a broad spectrum of benign fibrous tissue proliferations, the biological behavior of which is similar to both benign fibrous lesions and fibrosarcoma. Like fibrosarcoma, fibromatosis is characterized by infiltrative growth and a tendency towards recurrence; however, unlike sarcoma, it never develops metastasis [ 1 ]. Therefore, the most important strategy is to prevent direct invasion into adjacent tissues. Extraabdominal fibromatosis principally originates from the connective tissue of muscles and the overlying fascia or aponeurosis. It may occur in a variety of anatomical locations, including the muscles of the shoulder, the chest wall and back, thigh, and head and neck. However, solitary occurrence is rare in retroperitoneal space [ 1 , 2 ]. Here, we report on a case of extraabdominal fibromatosis in the retroperitoneum. Resection was successfully performed, and the patient has been tumor-free for two years after surgery. Case presentation A 43-year-old woman with a history of schizophrenia since 1982, and a history of hospitalization to help the patient to acquire social communication abilities, at the age of 23 presented with slight pain on her left flank and back. In 1995, she was treated with ureterolysis and ureteroureterostomy because of left-sided ureteral obstruction. Histological evaluation of the biopsy revealed benign fibrous tissue proliferations; however, no further evaluation and surgical excision was planned as her mental state was deteriorating. She was put on regular follow-up with computed tomography (CT) scans. In May 2002, she was referred from the psychiatric hospital to our Department of Urology, as the tumor tended to grow. CT scan with contrast enhancement revealed a tumor located between iliocostalis and psoas muscles in retroperitoneal space. The peripheral part of the tumor was enhanced, while the central part did not. The left paravertebral muscles around the tumor showed atrophy. The medial margin of the tumor was deformed by a left transverse process of the second lumber spine, suggesting invasive behavior (Figure 1 ). Coronal magnetic resonance imaging (MRI) demonstrated the tumor to be located beneath the left kidney. The central part of the tumor was found to be of iso/low intensity in the T1-weighted phase and of heterogeneously high intensity in the T2-weighted phase. The marginal part showed very low intensity in both phases (Figure 2 ). In May 2002, needle biopsy was performed and revealed that the tumor consisted of well-proliferated spindle cells rich in collagen fibers, an observation that was inconsistent with the histological pictures made in 1995. In June 2002, resection was performed using a paraspinal approach. Although the tumor strongly adhered to adjacent tissues, including the urinary tract and peritoneum, it was marginally resected, including paravertebral muscles and part of the spine. As the kidney was less affected by the tumor, ablation posed no problem. On gross examination, the cut surface appeared homogeneously gray and glossy (Figure 3 ). Histologically, a uniform proliferation of spindle cells with a moderate amount of collagen fibers led to a diagnosis of extraabdominal fibromatosis in the retroperitoneal space (Figure 4 ). No adjuvant treatment was given and during the two years of follow-up, the patient has remained asymptomatic, with no restrictions of daily living. There were no clues as to recurrence of the tumor in computed tomography. Figure 1 Contrast enhanced computed tomography showing the tumor location between iliocostalis and psoas muscles in retroperitoneal space. Figure 2 Coronal magnetic resonance imaging demonstrating the tumor beneath the left kidney. Figure 3 Gross appearance of the resceted tumor. The cut surface homogeneously appears gray and glossy. Figure 4 Photomicrograph showing the tumor composed of uniform spindled proliferation with moderate account of collagen fibers (hematoxylin and eosin ×400). Discussion Extraabdominal fibromatosis may occur in a variety of anatomic locations; the principle sites of the involvement are the shoulder, chest wall and back, thigh and head and neck. Origin of extraabdominal fibromatosis from any mesenchymal tissue is now well recognized [ 1 , 2 ]. Several authors have reported retroperitoneal fibromatosis in patients with familial adenomatous polyposis (Gardner syndrome) [ 2 , 4 - 6 ], however, solitary occurrence of fibromatosis is very rarely reported [ 1 , 2 , 7 - 9 ]. Our patient did not have a family history and upper gastrointestinal endoscopy, colonoscopy, or opthalmoscopy were normal suggesting that our patient may be negative for the syndrome. The exact histological origin of the tumor remains to be verified. The findings of CT suggested an origin from paravertebral muscles. Interestingly, this assumption was corroborated by a computed tomography performed in April 2000, which revealed that the previous tumor was located intramuscularly Principally, complete resection is the therapy of choice for this type of tumors [ 10 ]. Adjuvant therapy using non steroidal anti-inflammatory drugs (NSAIDs), tamoxifen, interferon, anti-neoplastic agents, radiation, and a combinations of these, have been reported for cases that are difficult to resect [ 1 ], the exact benefit offered by them is not known due to thin literature. Radiation therapy is accepted as an effective treatment after incomplete resection [ 11 , 12 ]. Recently, preoperative radiotherapy was reported to be useful for the local control [ 13 ]. In our case the tumor detection was delayed because the psychiatric status of our patient which has been unstable for several years. As wide resection of the tumor reduces the risk of recurrence, an early diagnosis is required for this type of tumor, which is difficult as most of these patients are asymptomatic. While the silent area contains several vital organs, extraabdominal fibromatosis should be considered for the differential diagnosis for such a lesion. Competing interests The authors declare that they have no competing interests. Authors' contributions AKik and AKid performed the operation, are responsible for the clinical work and helped with the preparation of the manuscript. TK is the orthopedic consultant and helped with the preparation and editing of the manuscript. TH coordinated and drafted the manuscript. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524188.xml

526774

Mental health first aid training of the public in a rural area: a cluster randomized trial [ISRCTN53887541]

Background A Mental Health First Aid course has been developed which trains members of the public in how to give initial help in mental health crisis situations and to support people developing mental health problems. This course has previously been evaluated in a randomized controlled trial in a workplace setting and found to produce a number of positive effects. However, this was an efficacy trial under relatively ideal conditions. Here we report the results of an effectiveness trial in which the course is given under more typical conditions. Methods The course was taught to members of the public in a large rural area in Australia by staff of an area health service. The 16 Local Government Areas that made up the area were grouped into pairs matched for size, geography and socio-economic level. One of each Local Government Area pair was randomised to receive immediate training while one served as a wait-list control. There were 753 participants in the trial: 416 in the 8 trained areas and 337 in the 8 control areas. Outcomes measured before the course started and 4 months after it ended were knowledge of mental disorders, confidence in providing help, actual help provided, and social distance towards people with mental disorders. The data were analysed taking account of the clustered design and using an intention-to-treat approach. Results Training was found to produce significantly greater recognition of the disorders, increased agreement with health professionals about which interventions are likely to be helpful, decreased social distance, increased confidence in providing help to others, and an increase in help actually provided. There was no change in the number of people with mental health problems that trainees had contact with nor in the percentage advising someone to seek professional help. Conclusions Mental Health First Aid training produces positive changes in knowledge, attitudes and behaviour when the course is given to members of the public by instructors from the local health service.

Background Community surveys have shown that the public in many countries have poor mental health literacy [ 1 ]. Many people cannot recognise mental disorders correctly, they differ from mental health professionals in their beliefs about causes and the most effective treatments, and they have stigmatizing attitudes which hinder recognition and appropriate help-seeking. This lack of mental health literacy limits the uptake of evidence-based treatments and leads to lack of support for people with mental disorders from others in the community. To help improve mental health literacy, a Mental Health First Aid training course has been developed. This course uses the first aid model that has been successfully applied to training members of the public to help in accidents and emergencies [ 2 ]. The Mental Health First Aid course is designed to give skills to provide initial help in mental health crisis situations and for on-going mental health problems. The course teaches a five-step approach to first aid: 1. Assess risk of suicide or harm, 2. Listen non-judgmentally, 3. Give reassurance and information, 4. Encourage person to get appropriate professional help, and 5. Encourage self-help strategies. These steps are applied to depression, anxiety disorders, psychosis and substance use disorders. In addition, participants are given specific instruction on how to help in the following mental health crisis situations: a suicidal person, a person having a panic attack, a person who has experienced a traumatic event, and a psychotic person threatening violence. An initial uncontrolled evaluation of the course involved comparing the first 210 participants at the beginning and end of the course, and at 6 months follow-up [ 3 ]. The course was found to produce improvement in ability to recognize a mental disorder in a case vignette, to change beliefs about treatment to be more like those of health professionals, to decrease social distance from people with mental disorders, to improve confidence in providing help to others, and to increase the amount of help actually provided. The next stage in the evaluation of Mental Health First Aid involved a randomized controlled trial with 301 employees of two large government departments [ 4 ]. Participants were assigned to either receive the course immediately or were placed on a wait-list for 5 months and received the training after the trial was completed. The trial found a number of benefits, including greater confidence in providing help to others, greater likelihood of advising people to seek professional help, improved concordance with health professionals about treatments, and decreased social distance from people with mental disorders. A surprising finding was that the course improved the mental health of the participants themselves, even though they were not recruited to have mental health problems and no therapeutic benefit was promised. The mental health benefits of the course had not been assessed in the earlier uncontrolled trial. This study involved an "efficacy" trial in that it was carried out under fairly ideal conditions which permitted rigorous experimental control. There was only one instructor who was the originator of the Mental Health First Aid course and very experienced, the trial was carried out in a workplace setting where employees were allowed time off to participate, the participants were a relatively well educated group of civil servants, and it was possible to randomly allocate participants relatively easily. In order to evaluate the course under more typical circumstances, we have now carried out a second trial. This "effectiveness" trial involved members of the public in a large rural area of Australia, who were taught by trained Mental Health First Aid instructors from the local health service. As in the previous trial, participants who received training were compared to a wait-list control group. Participants were randomized by Local Government Area clusters rather than individually because (1) there might have been contamination of information provided across allocated groups (2) the wait list group might have been difficult to maintain if others in the same locality were seen to be receiving training, and (3) individual randomization in some small communities may not have produced sufficient numbers to run a course. The reason for basing the trial in a rural area is that people living in rural Australia are less likely to receive general practitioner services for common mental disorders and also have more limited access to specialist mental health services [ 5 , 6 ]. There is therefore a greater need to develop community capacity to support those with mental disorders. Methods The details of this trial have been reported according to the CONSORT statement for cluster randomized trials [ 7 ]. Participants Eligible participants were residents of the catchment area of the New South Wales (Australia) Southern Area Health Service who were over 17 years of age, who volunteered for training in response to publicity, who were available over the period of the trial, and who were willing to receive interviews assessing trial outcomes. Participants had to volunteer as individuals rather than as a group (e.g. a whole workplace). Publicity took the form of talks to community groups, newspaper ads, a press release and radio interviews. Eligible clusters were the 16 Local Government Areas (cities or shires) in the catchment area of the Southern Area Health Service in 2003. This catchment is located in south-east New South Wales, runs approximately 370 km from north to south and approximately 160 km from east to west, and had a population of 194,435 in 2001. The Local Government Areas varied from popular coastal areas to farming communities to rural towns and ranged in population size from less than 5000 to over 50,000. Intervention Participants received a nine-hour Mental Health First Aid course, in three weekly sessions of three hours each. Training was administered in the local area in groups of up to 25 participants, with a minimum of 10 participants per course. As documentation of the intervention, there is a lesson plan for each session and a participants' manual containing material that was given to take away [ 2 ]. All instructors were given training and a teaching kit of lesson plans, videos, books, master copies of handouts and a set of transparencies. Educators received a one-week training program in how to conduct Mental Health First Aid courses and subsequent supervision in running a course. They were trained by Betty Kitchener who devised the Mental Health First Aid course. The course teaches how to help people in the crisis situations of being suicidal, having a panic attack, being exposed to a traumatic event, or in an acute psychotic state. The symptoms, risk factors and evidenced-based treatments (medical, psychological, alternative and self-help) for the mental disorders of anxiety, depressive and substance use and psychotic disorders are also taught. Figure 1 shows the five steps of providing mental health first aid taught in the course. Participants received training either immediately (experimental Local Government Areas) or after 6 months on a wait-list (control Local Government Areas). Figure 1 The five steps in providing mental health first aid. Training was administered by educators who were recruited from the staff of the Southern Area Health Service. Expressions of interest to become Mental Health First Aid instructors were sought from staff of the Area Health Service and associated community organisations. Five Mental Health first Aid instructors were recruited from a pool of 10 applicants for these positions. All the instructors had experience in mental health work and also a background in training, working with communities or health promotion work. A project coordinator with experience in mental health and health promotion (Ms Karen Peterson), who was employed to work on the project half time, also trained as an instructor. The same instructors taught courses in each paired Local Government Area, so that this factor did not differ between the immediate and wait-list Local Government Areas. The coordinator monitored a sample of courses taught during the trial to assess fidelity to the lesson plans. A fidelity checklist of topics that had to be covered was developed for each session. Four of the instructors had all three course sessions checked, while one of the instructors only had two sessions checked. The percentage of topics covered correctly was 100% for four of the instructors and 81% for one of the instructors. Objectives The hypotheses were that individuals trained in Mental Health First Aid, when compared to wait-list controls, would have increased knowledge of mental disorders and their treatments, decreased social distance, increased confidence in providing help, and that they would provide greater help to people experiencing mental health problems. Outcomes Outcomes were measured in January–February of 2003 (the pre-test assessment), the courses were run for the intervention group in March–April of 2003, and outcomes were measured again in July–August 2003 (the follow-up assessment). The wait-list control group received courses in September–October 2003, after the follow-up assessment was completed. All outcomes were measured at the individual level by telephone interview. The interview content was based on the questionnaire used in the uncontrolled trial of Mental Health First Aid [ 3 ]. The pre-test interview covered the following: whether the participant had ever experienced a mental health problem (yes/no), whether a family member had ever experienced a mental health problem (yes/no), the participant's confidence in helping someone (five-point scale from 1. not at all to 5. extremely ), contact in the last six months with anyone with a mental health problem (yes/no), how many people, whether any help offered (yes/no), what type of help (open-ended question), recognition of the problem in a case vignette (randomly assigned to be a case of depression or one of schizophrenia), what participant would do to help if they knew the person in the vignette (this "mental health first aid intention" involved the presence or absence of 8 elements, arrived at by a qualitative analysis of a sample of the responses, and added up to give a scorefrom 0–8), ratings of the likely helpfulness of a range of interventions for the person in the vignette (scored to give a scale of percentage agreement with mental health professionals about treatment [ 3 ]), a social distance scale relating to the person in the vignette [ 8 ], whether the participant had had a problem like the one in the vignette, whether a family member had had a problem like the one in the vignette, participant's reason for doing the course, and sociodemographic characteristics of the participant (age, gender, education, non-English speaking background, aboriginality). The follow-up questionnaire was the same as the pre-test questionnaire except that it omitted the sociodemographic questions. All outcomes were measured by a scripted telephone interview administered by professional interviewers. In order to reduce the length of the interview, participants were individually randomly assigned to receive either a depression vignette or a schizophrenia vignette, with the same questions asked in respect to each vignette. The interviewers were provided with an ID, name and phone number of each participant and knew whether they were giving the first or second interview to the participant. While they were not told whether the participant was in the experimental or control group, information about which group they were assigned to was given at the end of the interview script. As far as was practical given the very different sizes of the Local Government Area pairs, the same interviewers interviewed participants in each pair. Sample size determination For power calculations and sample size determination, a conservative assumption was made that the waitlist control group would show improvements, possibly due to increased awareness of mental health issues, of about 50% of that of the experimental group. This corresponds to effect sizes in the range 0.28–0.31 for changes on scales and in the range 0.02–0.04 for changes in identifying the correct diagnosis. Sample size estimates using nQuery Advisor software [ 9 ] indicated that a sample size of 200 participants in each of the two groups would be sufficient to detect differences with power of at least 80% in 2-sided tests at the 0.05 level. Clustering effects of individuals in 16 Local Government Areas involved design effects of unknown magnitude in the analysis. It was assumed that these would be of the order of 20%, so that a total achieved sample sizes of 250 in each group would be sufficient to detect differences with 80% power. Randomization: Sequence generation Randomization to immediate participation or wait-list was at the level of Local Government Area. The Local Government Areas were matched in pairs to have similar population and social characteristics. The variables used for matching were population size, interior vs coastal location, and an index of population education/occupation. The first listed LGA of each pair was assigned to the immediate or wait-list group at random, using the Random Integers option of Random.org [ 10 ] to generate a 1 or a 2 for each pair. For LGA pairs receiving a 1, the first member of the pair received immediate training, while for those receiving a 2 it was the second member of the pair. Each individual participant was randomly assigned a variable (values of 1 or 2) to determine which case vignette they received during their interviews. This was done using the Random Integers option of Random.org [ 10 ]. Those assigned a 1 received the interview based on a vignette of a person who is depressed and those assigned 2 received a vignette of a person with schizophrenia. Randomization: Allocation concealment Allocation was on the basis of cluster. In other words, the participant's Local Government Area determined whether they received immediate or wait-list training. Participants were not informed about their allocation to immediate or wait-list training until the end of their baseline interview. Randomization: implementation Local Government Areas were matched in pairs and Anthony Jorm assigned these randomly to immediate training or wait-list. Participants were not able to attend a class from outside their own Local Government Area. There was a recruitment period for all Local Government Areas which was organized by the coordinator Karen Peterson. The coordinator and the participants who were recruited were blind to the allocation of the Local Government Area during the recruitment period. Anthony Jorm revealed the allocation to Karen Peterson after the recruitment period ended. Karen Peterson then organized class times either immediately or after a waiting period, depending on the allocation of each Local Government Area in the pair. Randomization: Blinding (masking) At the time of the baseline interview, the participants did not know whether they were in an immediate or wait-list Local Government Area. However, interviewers had information at the end of the interview script telling whether the participant was assigned an immediate class or had to wait. Blinding of participants was not possible at subsequent interviews. Participants knew whether or not they had received training. While interviewers were not told the allocation of the participants in subsequent interviews, this might have become obvious during the interview if participants mentioned whether or not they had done the course. Interviewers were given a scripted interview to minimize any bias in the assessment due to knowledge of allocation. Ethics Ethical approval for the study was given by the Australian National University Human Research Ethics Committee and by the ethics committee of the South Western Sydney Area Health Service. Statistical methods For outcomes measured on a numeric scale, the change from pre-test to follow-up was analysed using linear regression. For binary outcomes, individuals scoring the same at pre-test and at follow-up were not used, and for those who changed, the direction of change was analysed as a binary outcome using logistic regression. Standard errors and p-values were adjusted for the cluster design using the Huber-White "sandwich" variance estimator, treating the 16 LGAs as the clusters. Analyses were corrected for differences between the LGA pairs by including this as an 8-level fixed-effect factor in the regression models. Missing data were imputed using best-subsets regression. All analysis was done using Stata version 8.2 [ 11 ]. Results Recruitment and Participant flow Recruitment of participants took place in October and November of 2002. Figure 2 shows the number of participants and clusters at each stage of the trial. Figure 2 Flow diagram of the number of participants and clusters at each stage of the trial. Baseline data Table 1 shows the characteristics of each group at the cluster and individual level. The two groups appear to be well matched in terms of sociodemographic characteristics and in history of mental health problems in self and family. However, there was a significant difference in reason for doing the course, with more people in the control group doing it for work reasons. Table 1 Baseline characteristics for each group given at the individual and cluster levels. Mental Health First Aid group Control group P-value Local Government Area characteristics at baseline Number 8 8 Population size: 1.0 <5,000 3 3 5,000–9,999 2 1 10,000–19,999 1 2 20,000–29,999 1 0 30,000–39,999 1 2 Number of participants in each area (smallest to largest) 9,17,18,29,30,48,100,165 8,9,12,16,28,50,53,161 Individual participant characteristics at baseline Number 416 337 Mean age (years) 47.14 47.97 0.42 Number (%) men 79 (19.0) 57 (16.9) 0.40 Number (%) with university degree 85 (20.6) 81 (24.1) 0.36 Number (%) aboriginal 11 (2.6) 10 (3.0) 0.40 Number (%) non-English speaking background 5 (1.2) 7 (2.1) 0.12 Reason for doing course: 0.011 Relating to workplace/voluntary work 180 (43.3) 188 (55.8) Relating to family/close friends 56 (13.5) 29 (8.6) Relating to own mental health status 20 (4.8) 10 (3.0) Duty as a citizen 49 (11.8) 44 (13.1) Just interested 111 (26.7) 66 (19.6) Note: P-values are adjusted for clustering by Local Government Area Numbers analyzed The data were analyzed by an intention-to-treat approach, with single imputation used for missing data. As shown in Figure 2 , the number of participants analyzed was the same as the number randomly allocated. Outcomes and estimation Tables 2 and 3 show the changes found for the dichotomous and continuous outcome measures respectively and the P-value of the comparison between the Mental Health First Aid and control group on these changes. From pre-test to follow-up a significantly larger percentage of the Mental Health First Aid group than the control group changed from not reporting experiencing a mental health problem to reporting experiencing one, from incorrectly to correctly diagnosing the case vignette and from reporting not offering help to a person with a mental health problem to reporting offering help. The Mental Health First Aid group changed significantly more than the control group in their agreement with health professional about treatment, in the degree of reduction in reported social distance from the person in the vignette and in their confidence in providing help. Table 2 Changes in dichotomous outcome measures. Outcome Mental Health First Aid group Control group OR (95% CI) P-value Mental health problems in self Pre-test 154 (37%) 118 (35%) Follow-up 172 (41%) 118 (35%) Change (95% CI) 4% (2 to 6) 0% (-3 to 3) 0.548 (0.304, 0.986), P = 0.045 Mental health problems in family Pre-test 233 (56%) 183 (54%) Follow-up 277 (67%) 205 (61%) Change (95% CI) 11% (4 to 17) 7% (2 to 11) 0.575 (0.318, 1.037), P = 0.066 Correct diagnosis of vignette Pre-test 282 (68%) 249 (74%) Follow-up 337 (81%) 255 (76%) Change (95% CI) 13% (8 to 19) 2% (0 to 4) 0.311 (0.250, 0.387), P < 0.001 Help offered to person with mental health problem Pre-test 305 (73%) 256 (76%) Follow-up 340 (82%) 270 (80%) Change (95% CI) 8% (4 to 13) 4% (-2 to 10) 0.602 (0.380, 0.953), P = 0.031 Professional help advised to person with mental health problem Pre-test 81 (19%) 71 (21%) Follow-up 104 (25%) 73 (22%) Change (95% CI) 6% (3 to 8) 1% (-4 to 5) 0.734 (0.452, 1.191), P = 0.21 Note: P-values and confidence intervals are adjusted for clustering by Local Government Area Table 3 Changes in continuous outcome measures. Outcome Mental Health First Aid group Control group Treatment effect (95% CI), P-value Agreement with health professionals about treatment Pre-test mean (SEM) 60.55 (3.89) 69.46 (2.18) Follow-up mean (SEM) 74.74 (1.91) 70.81 (2.27) Change (95% CI) 14.19 (9.53 to 18.85) 1.35 (-6.04 to 8.75) 11.77 (5.98, 17.56), P = 0.001 Social distance Pre-test mean (SEM) 8.13 (0.24) 8.06 (0.13) Follow-up mean (SEM) 7.59 (0.17) 7.90 (0.20) Change (95% CI) -0.53 (-0.99 to -0.08) -0.17 (-0.41 to 0.07) -0.26 (-0.49, -0.03), P = 0.032 Mental health first aid intention Pre-test mean (SEM) 1.81 (0.04) 1.88 (0.04) Follow-up mean (SEM) 1.83 (0.03) 1.85 (0.07) Change (95% CI) 0.02 (-0.11 to 0.15) -0.03 (-0.15 to 0.08) 0.06 (-0.00, 0.12), P = 0.066 Confidence in providing help Pre-test mean (SEM) 3.13 (0.08) 3.17 (0.07) Follow-up mean (SEM) 3.39 (0.05) 3.21 (0.07) Change (95% CI) 0.27 (0.11 to 0.42) 0.04 (-0.02 to 0.11) 0.21 (0.10, 0.33) P = 0.001 Number of people in contact with who had mental health problem Pre-test mean (SEM) 3.97 (0.31) 4.56 (0.20) Follow-up mean (SEM) 3.89 (0.30) 4.34 (0.29) Change (95% CI) -0.08 (-0.64 to 0.49) -0.22 (-0.83 to 0.40) 0.22 (-0.18, 0.63) P = 0.25 Note: Standard errors of the mean (SEM), confidence intervals and P-values are adjusted for clustering by Local Government Area The intraclass correlations for the continuous outcomes were: for agreement with health professionals about treatment, 0.15 (95% confidence interval 0.01, 0.29); for number of people in contact with that had a mental health problem, 0.02 (0, 0.06); for confidence in providing help, 0.03 (0, 0.07); for mental health first aid intention, 0.002 (0, 0.02); and for social distance, 0.04 (0, 0.08). Thus for all but one outcome, the correlation was small, justifying our assumption of a modest design effect. Adverse events Given that an educational intervention was evaluated with a non-clinical sample, there was no justification for a systematic inquiry into adverse events. Informally, no adverse events were reported. Discussion This study has found that the Mental Health First Aid training produced a number of significant changes in participants compared to a wait-list control group. A number of changes related to how people responded to a vignette of a person with either depression or schizophrenia. We found that there was greater recognition of the disorders in a vignettes, increased agreement with health professionals about which interventions are likely to be helpful, decreased social distance towards the people portrayed in the vignettes. These changes were seen equally with both vignettes. There was also a non-significant trend for those in the trained group to have more ideas for how to help the person in the vignette if it had been someone they knew. Other outcomes with significant changes related more directly to the provision of mental health first aid. There was increased confidence in providing help to others and an increase in help actually provided. There was no change in the number of people with mental health problems that trainees had contact with or in the percentage advising someone to seek professional help. One potential concern of Mental Health First Aid training is that it will lead to over-diagnosis of life problems as mental disorders. In previous trials we have found no evidence that the training affects the perception that the participant or their family have mental health problems [ 3 , 4 ]. By contrast, in the present study there was a significant increase in the percentage who perceived themselves as having a mental health problem and a non-significant trend for an increased perception of family members as having mental health problems. However, in absolute terms the changes were not so great as to be a concern and may, in fact, reflect accurate re-labelling. These findings are similar to those of the earlier efficacy trial. However, the courses were taught by instructors who were not the originators of the Mental Health First Aid program under conditions which more closely approximate those that are typical in practice. The findings are therefore more generalizable than those reported previously. While the more typical conditions of this trial are an advantage for generalizability, they produced greater practical difficulties in running the trial. An important weakness was that attendance data on participants were not collected by some of the instructors. We are therefore uncertain what proportion of the participants received the complete training course. A similar problem was determining the adherence of the instructors to the curriculum. We were able to carry out some formal observation of the instructors'adherence to a list of topics covered by the curriculum and found 100% adherence for most of the instructors, but one had only 81% adherence. Another limitation of this study is that we did not directly measure the mental health of participants. In the earlier trial, we unexpectedly found a mental health benefit and this requires replication. The reason that a mental health measure was not included was that we did not have the results of the earlier trial at the time we designed this one. Another factor was the limited time available in the telephone interviews used to assess outcomes. We used an intention-to-treat approach to the data. Whereas many trials use a last observation carried forward approach to handle missing post-test data, we used data imputation by best-subsets regression. This approach is likely to give better estimates than conventional approaches to missing data even when the missing-at-random assumption is not met [ 12 ]. Since this and the earlier trials were started, the Mental Health First Aid course has been extended from 9 to 12 hours on the basis of consistent requests from trainees for a longer course. The longer course does not add new content, but rather extends the time available to deal with each topic. We have yet to evaluate whether this extension adds to the effectiveness of the training. Conclusions A nine-hour Mental Health First Aid training produces positive changes in knowledge, attitudes and behavior when the course is given to members of the public by instructors from the local health service. This finding shows that the effects of the course are generalizable beyond its originators and when run under typical conditions. Competing interests BAK and AFJ were the developers of the Mental Health First Aid course. Authors' contributions AFJ was involved in securing funding for the study, had a major role in the design of the study, co-developed the evaluation questionnaire, contributed to the data analysis and had a major role in writing the manuscript. BAK was involved in securing funding for the study, developed and taught the Mental Health First Aid Instructor course, had a role in the design of the study, co-developed the evaluation questionnaire, organized the outcome assessment and had a minor role in writing the manuscript. ROK was involved in securing funding for the study, had a role in the design of the study, had a major role in planning and managing the trial's implementation in its initial stages, recruited and supervised the study staff, established and maintained organisational support in the Southern Area, and had a role in the writing of the manuscript. KBGD had a major role in the data analysis and a minor role in writing the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526774.xml

529453

Enhanced cell-permeant Cre protein for site-specific recombination in cultured cells

Background Cell-permeant Cre DNA site-specific recombinases provide an easily controlled means to regulate gene structure and function in living cells. Since recombination provides a stable and unambiguous record of protein uptake, the enzyme may also be used for quantitative studies of cis - and trans -acting factors that influence the delivery of proteins into cells. Results In the present study, 11 recombinant fusion proteins were analyzed to characterize sequences and conditions that affect protein uptake and/or activity and to develop more active cell-permeant enzymes. We report that the native enzyme has a low, but intrinsic ability to enter cells. The most active Cre proteins tested contained either an N-terminal 6xHis tag and a nuclear localization sequence from SV40 large T antigen (HNC) or the HIV Tat transduction sequence and a C-terminal 6xHis tag (TCH 6 ). The NLS and 6xHis elements separately enhanced the delivery of the HNC protein into cells; moreover, transduction sequences from fibroblast growth factor 4, HIV Tat or consisting of the (KFF) 3 K sequence were not required for efficient protein transduction and adversely affected enzyme solubility. Transduction of the HNC protein required 10 to 15 min for half-maximum uptake, was greatly decreased at 4°C and was inhibited by serum. Efficient recombination was observed in all cell types tested (a T-cell line, NIH3T3, Cos7, murine ES cells, and primary splenocytes), and did not require localization of the enzyme to the nucleus. Conclusions The effects of different sequences on the delivery and/or activity of Cre in cultured cells could not be predicted in advance. Consequently, the process of developing more active cell-permeant recombinases was largely empirical. The HNC protein, with an excellent combination of activity, solubility and yield, will enhance the use of cell-permeant Cre proteins to regulate gene structure and function in living cells.

Background The Cre recombinase from bacteriophage P1 has been widely used to induce DNA sequence-specific recombination in mammalian cells [ 1 ]. The enzyme, which catalyzes recombination between 34 nucleotide LoxP sequences during P1 genome replication, has been used in a variety of genetic applications to regulate gene structure and function. These include conditional mutagenesis, gene replacement, chromosome engineering, and regulated gene expression [ 2 - 4 ]. However, the use of site-specific recombination in genetic studies is frequently hampered by difficulties expressing the recombinase in cells at the desired time and place [ 5 ]. Moreover, the use of Cre expression vectors is constrained by the fact that prolonged exposure to the enzyme can be lethal to cells [ 4 , 6 , 7 ]. To address these problems, we [ 8 ] and others [ 9 - 12 ] have developed membrane-permeable Cre recombinase proteins that are capable of entering cells by a process of protein transduction. Protein transduction exploits properties of specific protein sequences [termed protein transduction domains (PTDs)] that enhance the delivery of macromolecules – including peptides, proteins, and DNA fragments – into living cells [ 13 - 15 ]. Cell-permeant Cre proteins provide an effective means to regulate gene structure and function in living cells, and Cre-mediated recombination provides a potentially useful reporter system with which to study the process of protein transduction itself. In particular, recombination provides a stable and quantitative record of protein uptake that circumvents problems of distinguishing between internalized and cell-associated proteins [ 16 ]. In our previous report, recombinant fusion proteins bearing the 12 amino acid membrane translocation sequence (MTS) from fibroblast growth factor 4 (FGF-4) were used to deliver enzymatically active Cre proteins directly into mammalian cells. Of the four recombinant proteins tested, an enzyme containing an N-terminal 6xHis affinity tag, a nuclear localization sequence (NLS) from SV40 large T antigen, and the FGF-4 MTS (HNCM), displayed the best combination of yield, solubility, nuclear localization and enzymatic activity within cells. Recombination was observed in greater than 70% of cells treated with 10 μM HNCM for 2 hours. Widespread recombination was also observed in mice following intraperitoneal administration of HNCM, indicating that a wide variety of terminally differentiated cell types can internalize cell-permeant Cre and are competent to undergo site-specific recombination. In the present study, eleven recombinant Cre proteins were prepared in order to evaluate sequences affecting the uptake and/or activity of the enzyme in cultured cells and if possible to develop more active recombinases. Several constructs were designed to compare the activities of different PTDs, including the FGF-4 MTS [ 17 ], sequences from HIV TAT [ 18 ] and a (KFF) 3 K sequence that was previously used to deliver peptide nucleic acid (PNA) conjugates into cells [ 19 ]. Only the Tat sequence promoted the delivery of active Cre into cells, while all three PTD sequences adversely affected the solubility of recombinant proteins containing polyhistidine tags. The contribution of the SV40 large T antigen NLS [ 20 ] was also examined to understand apparent differences in the behaviour of cell-permeant Cre and proteins expressed following gene transfer. Thus, the activity of cell-permeant Cre was enhanced by the SV40 large T antigen NLS [ 8 , 10 ], whereas, the native Cre protein appears to possesses a functional NLS, whose activity was not augmented by the T antigen NLS [ 21 ]. We report that polyhistidine tags (6xHis) frequently used for protein affinity purification [ 22 ] and the large T antigen NLS each separately enhance cellular uptake of enzymatically active Cre, and we describe the development of a cell-permeant Cre recombinase with an excellent combination of activity, solubility and ease of purification. Results Recombinant Cre fusion proteins Eleven recombinant Cre proteins were prepared in order to evaluate sequences affecting the uptake and/or activity of cell-permeant enzymes (Figure 1A ). Native Cre (Cre) corresponds to the protein encoded by the P1 phage genome [ 23 ]. HC and H 6 C have amino-terminal hexahistidine tags consisting of MGSSHHHHHHSSLVPRGSH and MHHHHHH, respectively, while CH 6 is similar to H 6 C except the His tag is on the C-terminus. His-NLS-Cre (HNC) is similar to HC except a nuclear localization sequence (PKKKRKV) from SV40 large T antigen [ 20 ] is positioned between His and Cre. HT 7 N'C is similar to HNC except the His tag contains 11 additional amino acids (MASMTGGQQMG) from the pET28a(+) polylinker and the arginine of the NLS sequence was converted to a lysine. This change, which resulted from an altered PCR primer, is unlikely to affect nuclear localization activity. HNCM, described previously [ 8 ], contains a membrane translocation sequence (MTS) from the leader sequence of FGF-4 positioned at the C-terminus of HNC. Finally, four different sequences, each reported to have protein transduction activity [ 17 , 19 , 24 ], were placed on the amino terminus of CH 6 . These consisted of the FGF-4 MTS (MCH 6 ), an SV40 large T antigen nuclear localization sequence (NCH 6 ), the HIV Tat sequence (TCH 6 ), and a (KFF) 3 K sequence (KCH 6 ). Cre proteins were expressed from pET28a(+) plasmids in E. coli and, except for native Cre, were purified by Ni 2+ affinity chromatography under non-denaturing conditions [ 8 ]. The native enzyme was purified by a combination of hydroxyapatite column chromatography and Sephacryl S-100 HR FPLC. All proteins were expressed at high levels, with yields of purified proteins ranging from 5 to 41 mg/L of E. coli culture (Figure 1 ). All of the enzymes except TCH 6 and KCH 6 could be dialyzed against DMEM or RPMI media and stored at -20°C until use. However, TCH 6 precipitated under these conditions and was dialyzed instead against PBS supplemented with 0.3 M NaCl (0.45 M total NaCl) and 8% glycerol. The pH of the buffer (ranging from 7.5–8.5) had no obvious effect on protein solubility. The KCH 6 protein was insoluble over a range of pH values and salt concentrations (up to 0.8 M), and the protein was not evaluated further. The remaining proteins could be prepared at final concentrations above 1 mg/ml except HNCM, which precipitated out of solution at protein concentrations above 500 μg/ml. The specific activities of the tagged fusion proteins were similar, ranging from 4.3–8.5 × 10 4 U/mg protein, corresponding to 47–92% of the activity of the native enzyme (Figure 1 ). Native Cre recombinase has protein transduction activity that is enhanced by polyhistidine and NLS sequences The uptake and enzymatic activity of Cre proteins was monitored in Tex.loxP.EG cells [ 8 ]. These cells contain a single integrated retrovirus (Fig 2A ) in which the expression of an enhanced green fluorescent protein (EGFP) gene is prevented by a "stop cassette" consisting of multiple polyadenylation sites positioned between two loxP sites. Deletion of the stop cassette by Cre mediated-recombination activates the expression of the EGFP reporter gene. Cells were exposed to Cre, HC, HNC, HT 7 N'C, HNCM, H 6 C, CH 6 , NCH 6 , and TCH 6 for 2 hrs at concentrations ranging from 0 to 8 μM. The cells were then washed extensively with PBS, were cultured for 24 hrs to provide time for EGFP expression, and the percentage of EGFP-expressing cells was determined by flow cytometry. Recombination was also monitored by Southern blot hybridization (Fig 2C ), thus confirming that expression of the EGFP reporter accurately reflected the extent of template recombination. As shown in Figure 2B , all of the proteins tested induced recombination in a concentration-dependent manner. The native enzyme had the lowest activity, inducing recombination in up to 17% of the cells. Uptake and/or activity was increased by polyhistidine tags positioned on either the amino- (HC and H 6 C) or carboxy-terminal (CH 6 ) ends of the enzyme. HC, which contains the 6xHis tag from pET28a(+), and H 6 C and CH 6 , which contain simple 6xHis sequences, induced similar levels of recombination, ranging between 45 and 60% of cells. Activity was further increased by the addition of an SV40 large T antigen nuclear localization signal (HNC, HT 7 N'C and NCH). At low concentrations, TCH 6 was the most active protein tested, but higher levels of the protein were toxic to cells, presumably because of the concentration of NaCl required to maintain protein solubility during preparation of the enzyme. Finally, presence of the FGF-4 MTS sequence proved inhibitory for recombination in cultured cells. HNC, which lacks the MTS sequence, was approximately 8 times more active than HNCM as determined by the concentration of enzyme required to induce recombination in 50% of the cells. Thus, while the FGF-4 MTS has been used to transduce a variety of proteins and peptides into mammalian cells, the sequence suppressed the activity of cell-permeant Cre. This confirms the results of a earlier study, reported while this work was in progress [ 10 ]. Temperature-dependent protein transduction Transduction of cargoes containing the FGF-4 MTS [ 17 ], including the HNCM protein used in this study [ 25 ], is greatly decreased at 4°C as compared to 37°C. By contrast, there have been conflicting reports with regard to the transduction of proteins containing the Tat and Antennapedia PTDs [ 26 - 33 ]. Cre fusion proteins with (HNCM) and without (HNC, HT 7 N'C and HC) the FGF-4 MTS were therefore tested for their ability to enter cells at 4°C (Fig 3A ). Tex.loxP.EG cells were incubated with varying concentrations of the proteins for one hour at either 37°C or 4°C; were washed extensively and were cultured at 37°C to allow time for EGFP expression. In each case, the uptake of the enzyme was inhibited at 4°C, indicating that the inhibitory effects of low temperature are not limited to cargoes containing the FGF-4 MTS. Low levels of recombination observed in cells treated with higher concentrations probably results from cell-associated enzyme that the washing steps fail to remove and that gains entry when cells are later cultured at 37°C. Serum and cell density effects on protein transduction Serum has been reported to inhibit the transduction of cargoes containing the FGF-4 MTS; whereas, transduction of proteins containing the Tat and Antennapedia PTDs appears to be unaffected by serum. The effects of serum on protein transduction were assessed by treating TexloxP.EG cells with either HC (5 μM) or HNC (2 μM) for 1 hour in the presence of varying concentrations of either fetal bovine serum (Figure 3B , FBS) or normal mouse serum (Figure 3B , MS). Recombination induced by both proteins was inhibited by up to 60% and 80% in media containing 10% FBS and MS, respectively. Serum appeared to inhibit protein transduction specifically, since it had no discernable effect on either the stability or activity of the proteins in vitro (data not shown). These results indicate that the inhibition of protein transduction by serum is not limited to cargoes containing the FGF-4 MTS. To assess the effects of cell density on protein transduction, TexloxP.EG cells were seeded at different concentrations in 2 cm 2 culture dishes and were treated with HNC (2 μM) for 1 hour (Figure 3C ). Recombination efficiencies increased by about 40% as the number of number of cells was increased from 10 4 to 6 × 10 4 cells/cm 2 , and then declined sharply at concentrations above 2 × 10 5 cells/cm 2 . Tex.loxP.EG is a T-cell line and is non-adherent; however, the cells settle to the bottom of the culture dish. The optimum density for recombination was similar to the number of cells (.75 × 10 4 cells/cm 2 ) required to cover the culture dish. Uptake and localization of cell-permeant Cre proteins The kinetics of cell-permeant Cre uptake in cultured cells was monitored by examining cells for recombination after exposure to cell-permeant Cre for different lengths of time (Figure 4 ). Cells were treated with HC (5 μM), HNC (3 μM) and HNCM (10 μM) for 5 to 120 mins, were washed extensively and the extent of recombination was monitored 24 hours later by flow cytometry. Thus, the assay measures the time required for extracellular enzyme to become committed to cell entry. The uptake of HC, HNC and HNCM increased with time, reaching half-maximum levels after 15–20 min. Recombination was induced in cells following exposure to HNC and HNCM proteins for less time than to HC, suggesting a direct role for the NLS in promoting cell binding and/or entry. Note that the observed differences (apparent within minutes after exposing cells to Cre) cannot reflect potential differences in nuclear trafficking since a delay in nuclear import of a few minutes would not affect the extent of recombination measured 24 hours later. The uptake and localization of cell-permeant Cre in cultured cells was also monitored in living cells by using proteins labeled with the fluorescent dye, Alexa 488 (Figures 5 and 6 ). The uptake of tagged HC, HNC and HNCM increased with the time as measured by flow cytometry (Figure 5 ). Again, HNC and HNCM appeared to enter cells more rapidly than HC, which lacks an NLS. Note that the magnitude of the fluorescence is less important than the rate of increase (slope) since the proteins were not labelled to the same extent with Alexa 488. The localization of Alexa 488-tagged HNC and HC proteins was monitored in living cells by fluorescence microscopy. The HNC protein was localized to the nuclei of Cos7 cells, but was predominately cytoplasmic in NIH3T3 cells (Fig 6B ). Cre protein was also localized to the nuclei of Tex.loxP.EGcells (data not shown). Recombination in different cell types Several mammalian cell types were analyzed for their ability to undergo Cre-mediated recombination. Clones of Cos7 and NIH3T3 cells containing a single pBABE.lox.stp.EGFP provirus were treated with HNC for two hour and analyzed by Southern blot hybridization (Figure 7 ). Although the concentration of enzyme necessary to achieve nearly complete recombination was approximately 10 times higher in Cos7 cells than in NIH3T3 cells (10 versus 1 μM, Figure 7 ), the recombination efficiency did not correlate with localization of the enzyme to the nucleus (Figure 6 ). Extensive recombination was also observed in murine embryonic stem cells containing a floxed IKK γ gene (Figure 7 ) and in primary splenocytes explanted from ROSA26R mice (Figure 8 ). Efficient recombination was therefore observed in all mammalian cell types examined. Discussion Cell permeant Cre proteins have generated considerable interest as genetic tools to regulate gene structure and function in mammalian cells [ 5 ]. In addition, Cre mediated recombination provides a quantitative reporter for studies on the protein transduction process itself. In the present study, 11 recombinant fusion proteins were analyzed to characterize sequences and conditions that affect protein uptake and/or activity and to develop more active cell-permeant enzymes. We report that the native enzyme has a low, but intrinsic ability to enter cells. Uptake and was enhanced by the addition of a 6xhistidine-tag on either the amino or carboxyl terminal ends of the protein and was enhanced further by the addition of a nuclear localization sequence from SV40 large T antigen or the transduction sequence from the HIV Tat protein. Finally, the hydrophobic membrane translocation sequence (MTS) from fibroblast growth factor 4 (FGF-4) had a net deleterious effect on Cre-mediated recombination in cultured cells. We had hoped that the native or 6xHis-tagged enzymes would lack transduction activity so that the effects of additional sequences on cell entry, nuclear transport and recombination could be compared. However, we found that the native Cre protein has an intrinsic ability to enter cells, thus confirming observations by Will, et al. [ 11 ]. The mechanism by which Cre gains entry into cells remains to be determined. The enzyme may possess a protein transduction domain analogous to those described for HIV Tat, Antennapedia and a growing list of proteins that can enter cells directly, without requiring specific receptor and transporter systems [ 34 ]. The fact that Cre is a basic protein [ 23 ] is potentially significant considering that many basic peptides are capable of entering cells [ 24 , 35 - 38 ]. The transduction activity of the native enzyme hinders quantitative studies of sequences incorporated into recombinant Cre proteins, since structural changes associated with each modification may have varying effects on the intrinsic ability of Cre to enter cells. Even so, the 6xHis sequence appeared to facilitate cell entry, since two different 6xHis sequences enhanced Cre-mediated recombination while having little effect on the activity of the enzyme in vitro . Moreover, 6xHis sequences were active when positioned on either the amino- or carboxyl-terminal ends of the enzyme. L-histidine heptamers have been shown to enter cells, although much less efficiently than arginine homopolymers [ 37 , 38 ]. Positively charged histidine sequences also bind cell surface heparin sulfate proteoglycans [ 39 , 40 ], and thus may enhance uptake as has been suggested for the HIV Tat transduction sequence [ 41 ]. The transduction activity of the 6xHis sequences is potentially significant given their widespread use as affinity tags to purify recombinant fusion proteins [ 22 ]. We [ 8 ] and others [ 10 ] have shown that the activity of cell-permeant Cre fusion proteins in cultured cells can be enhanced by the addition of an SV40 large T antigen nuclear localization sequence (NLS). The NLS has been shown to enhance the activity of Cre expression vectors [ 42 ], presumably by targeting the protein to the nucleus. However, in the present study the NLS enhanced the delivery of Cre fusion proteins into cultured cells as assessed either by cell-based recombination or by uptake of fluorescent Cre proteins. Moreover, nuclear localization did not appear to contribute to cell type-specific differences in the activity of the HNC protein. These results are consistent with the observation that the T antigen NLS, like a number of other basic sequences, has protein transduction activity [ 24 ]. HNC, which consisted of a 6xHis tag and T antigen NLS attached to the amino-terminus of the enzyme, was highly active despite the absence of a canonical protein transduction sequence. The 6xHis tag and NLS sequences separately contributed to the transduction of the enzyme, which was 5–8 times more active in cultured cells than the HisNLSCreMTS (HNCM) protein described in our earlier study [ 8 ]. By virtue of their positive charge, these elements share similarities with basic protein transduction domains such as HIV Tat and Antennapedia. Early studies suggested that the basic PTDs entered cells through an energy-independent process that was relatively unaffected by low (4°C) temperature [ 26 - 30 ]. However, later studies suggest that the positively charged PTDs promote cell uptake by endocytosis possibly by binding negatively charged proteins on the cell surface [ 12 , 31 - 33 , 43 , 44 ]. Similarly, the uptake of Cre fusion proteins is consistent with an endocytic mechanism. In particular, uptake (time required for commitment to entry) was relatively rapid (10 to 15 min for half maximum uptake), was greatly decreased at 4°C, and was inhibited by up to 80% by serum. The inhibition by serum appeared to involve cell binding and/or entry specifically, since serum had no discernable effect on the stability or activity of Cre fusion proteins in vitro (data not shown). Recombination was markedly suppressed at higher cell densities, possibility because binding sites in or on cells sequester the enzyme, thus lowering the effective protein concentration. Alternatively, cell density may suppress protein transduction by reducing cell size and/or available surface area [ 25 ]. In either event, similar observations have been reported for a TatCre protein that lacks a 6xHis tag [ 9 ], illustrating the need to control for cell density when comparing the effects or cell-permeant proteins in different cell populations. Conclusions The effects of different sequences on Cre uptake or activity were difficult to predict in advance, and consequently, the process of developing a more active cell-permeant recombinase was largely empirical. Proteins containing amino-terminal Tat and (KFF) 3 K sequences (TCH 6 and KCH 6 ) were poorly soluble when dialyzed against isotonic media, and the FGF-4 MTS interfered with the activity of the enzyme in cultured cells. Since Cre is probably not unique in this regard, investigators seeking to develop other cell-permeant proteins would be advised to test a variety of sequences positioned at different places on the protein. In the present case, the HNC protein possesses an outstanding combination of activity, solubility and yield that will enhance the use of cell-permeant Cre to regulate gene structure and function in living cells. Methods Cre expression vectors Native Cre and Cre fusion proteins were expressed in E. coli from pET28a (+)-based plasmids (Novagen). A plasmid expressing native Cre (Cre) was constructed by using the Cre #1 and Cre-stop-HindIII primers to amplify Cre coding sequences from MBP-NLS-Cre-MTS [ 8 ] which were cloned between the Nco I and Hind III sites of pET28a (+). HC, His 6 C, HNC, and HT 7 C were similarly constructed by PCR amplification, using primers Cre #2, Cre #3, Cre #4 and Cre #5, respectively, together with Cre-stop-HindIII primer. The amplified DNA fragments were cloned into Nhe I and Hind III sites of pET28a (+). CHis 6 , NCHis 6 , MCHis 6 , TATCHis 6 , and (KFF) 3 KCHis 6 were constructed by using primers Cre #6, Cre #7, Cre #8, Cre #9 and Cre #10, respectively, together with the Cre-XhoI primer, and then cloned into the Nco I and Xho I sites of pET28b(+). Cre #1: AGAGAGCCATGGGCTCCAATTTACTGACCGTACACCAA Cre #2: GTACATGCTAGCTCCAATTTACTGACCGTACACCAA Cre#3: AGAGAGCCATGGGCCATCATCATCATCATCACAGCTCCAATTTACTGACCGTACACCAA Cre#4: GTACATGCTAGCCCAAGAAGAAGAGGAAGGTGCTCCAATTTACTGACCGTACACCAA Cre#5: GTACATGAATTCTCCAATTTACTGACCGTACACCAA Cre#6: AGAGAGCCATGGGCTCCAATTTACTGACCGTACACCAA Cre#7: AGAGAGCCATGGGCCCCAAGAAGAAGAGGAAGGTGTCCAATTTACTGACCGTACACCAA Cre#8: AGAGAGCCATGGGCGCAGCCGTTCTTCTCCCTGTTCTTCTTGCCGCACCCTCCAATTTACTGACCGTACACCAA Cre#9: AGAGAGCCATGGGCGGTCGCAAGAAACGTCGCCAACGTCGCCGTTCCAATTTACTGACCGTACACCAA Cre #10: AGAGAGCCATGGGCAAATTCTTTAAGTTCTTTAAGTTCTTTAAGTCCAATTTACTGACCGTACACCAA Cre-stop-HindIII: GATACGAAGCTTCTACTAATCGCCATCTTCCAGCAGGCGC Cre – XhoI: AGAGAGCTCGAGATCGCCATCTTCCAGCAGGCGCACCATTGCCCCTGT Protein purification Polyhistidine-tagged Cre fusion proteins were expressed in E. coli strain BL21 (DE3) and purified by Ni 2+ affinity chromatography as described previously [ 8 ]. Bacterial cultures (2L) were grown to an A 600 of 0.6–1.0, were induced with 0.4 mM IPTG, and after harvesting the cells were lysed in 40 ml 50 mM Tris (pH 8.0), 50 mM sodium phosphate and 300 mM NaCl. After affinity chromatography, recombinant proteins were dialyzed against DMEM or RPMI media, except TCH 6 and KCH 6 , which precipitated under these conditions. TCH 6 was dialyzed instead against PBS supplemented with 0.3 M NaCl (0.45 M NaCl final) and 8% glycerol. The native recombinase was similarly expressed; however, lysates were applied to a 50 ml hydroxyapatite column and step-eluted with sodium phosphate buffers (pH 8.0) of increasing concentration (50, 100, 200, 300, and 500 μM). The peak fraction (200 μM) was dialyzed against PBS, was concentrated by centrifugal ultra filtration to about 20 mg/ml and was fractionated by Sephacryl S-100 HR gel filtration FPLC. Peak fractions identified by SDS PAGE, were concentrated and dialyzed against PRMI-1640 medium containing 1% streptomycin/penicillin and 2% glycerol. In vitro assays of Cre enzyme activity measured the release of a circular plasmid inserted into a λ phage (Novagen, Madison, WI) by transformation of E. coli [ 8 ]. One unit (U) of enzyme produces 10 4 colonies (equivalent to 2 × 10 6 circular molecules) in a 30 minute reaction containing 200 ng DNA substrate in 50 mM Tris HCl, pH 7.5, 33 mM NaCl and 10 mM MgCl 2 in a total volume of 15 μl. All Cre proteins were stable for at least 6 months at -80°C without significant loss of enzymatic activity. Cell culture and protein transduction Tex.loxp.EG, 3T3.loxp.EG and Cos7.loxp.EG cells were derived from Tex (a murine thymoma line derived from p53-deficient mice), NIH3T3 and Cos7 cells, respectively, following infection with the pBABE.lox.stp.EGFP retrovirus [ 8 ]. Cells were incubated with serum-free RPMI 1640 (Tex.loxp.EG) or DMEM (3T3.loxp.EG and Cos7.loxp.EG) containing Cre fusion proteins for 2 hours, then washed with PBS twice and cultured in normal growth medium at 37°C incubator for 24 hours. Cre mediated recombination, which induces the expression of an enhanced green fluorescence protein (EGFP) in Tex.loxp.EG cells, was measured by flow cytometry using a FACSort instrument (Becton Dickinson). Alternatively, recombination was monitored by Southern blot hybridization [ 8 ]. For experiments involving low temperature protein transduction, the cells and all solutions were maintained at 4°C until after washing with cold PBS, after which the cells were returned to normal medium at 37°C. Preparation of fluorescent HNC, HC, and HNCM proteins Fluorescent HNC, HC, and HNCM proteins were prepared by using the Alexa 488 protein labeling kit (Molecular Probes, catalog number A-10235), according to the manufacturer's instructions. The proteins were dialyzed against PBS, were labeled at a concentration of 2 mg/ml and purified by gel filtration through a Sephadex G50 column. Cells were treated with the fluorescent proteins at a final concentration of 1 μM. The localization of fluorescent Cre proteins was monitored in living NIH3T3 and Cos7 cells by fluorescence microscopy. The cells were cultured to 50–80% confluence in a slide chamber (NUNC), washed with PBS and then incubated with fluorescent Cre proteins at a final concentration of 1 μM for 1.5 hours. The cells were washed three times with PBS, counterstained with 1 μg/ml 4',6-diamidino-2-phenylindole (DAPI) in PBS for 20 minutes, washed three times with PBS and mounted in anti-fade fluorescence mounting medium. The stained cells were photographed with an Olympus BX60 fluorescence microscope using green and blue filters. Ex vivo recombination by HNC recombinase in primary cells Splenocytes from ROSA-26R mice were cultured for one day in RPMI 1640 medium containing 10% FBS and were treated with HNC recombinase in serum free media for 2 hours. The cells were washed, cultured for 24 hours, and stained with a fluorescent β-galactosidase substrate (ImaGene green™, C 12 FDG) according to instructions provided by the manufacturer (Molecular Probes, Inc.) Recombination, which activates the expression of a floxed lacZ gene, was assessed by flow cytometry. Authors' contributions QL developed and characterized most of the recombinant Cre proteins described in this study. DJ provided the HNCM protein and assisted with in vitro assays for Cre activity. KDGA. investigated inhibition of Cre activity by serum. ER supervised the project and drafted the final manuscript. All authors approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529453.xml

514711

A faster way to make GFP-based biosensors: Two new transposons for creating multicolored libraries of fluorescent fusion proteins

Background There are now several ways to generate fluorescent fusion proteins by randomly inserting DNA encoding the Green Fluorescent Protein (GFP) into another protein's coding sequence. These approaches can be used to map regions in a protein that are permissive for GFP insertion or to create novel biosensors. While remarkably useful, the current insertional strategies have two major limitations: (1) they only produce one kind, or color, of fluorescent fusion protein and (2) one half of all GFP insertions within the target coding sequence are in the wrong orientation. Results We have overcome these limitations by incorporating two different fluorescent proteins coding sequences in a single transposon, either in tandem or antiparallel. Our initial tests targeted two mammalian integral membrane proteins: the voltage sensitive motor, Prestin, and an ER ligand gated Ca 2+ channel (IP 3 R). Conclusions These new designs increase the efficiency of random fusion protein generation in one of two ways: (1) by creating two different fusion proteins from each insertion or (2) by being independent of orientation.

Background Biosensors based on GFP-fusion proteins are powerful tools for observing real-time events within living cells. Insertion of GFP within another protein has produced biosensors capable of signaling intracellular events through intrinsic fluorescence changes [ 1 , 2 ], fluorescence resonance energy transfer (FRET) [ 3 , 4 ], and changes in sub-cellular localization [ 5 ]. The difficult task of finding the right insertion site to produce a biosensor can be accelerated by screening libraries of random GFP insertions [ 6 - 8 ]. The insertional strategies described to date, however, are limited in two ways. First, each insertion produces only one kind, or color, of fluorescent fusion protein. Creating the multicolored libraries necessary for co-expression or FRET analyses requires either separate rounds of insertion and screening for each fluorescent protein or additional subcloning to exchange fluorophores. Second, the efficiency of any random approach is limited to a maximum of 1:6 because a fusion protein can only be produced if the GFP coding sequence lands in the correct orientation and reading frame with respect to the target coding sequence. We reasoned that it might be possible to overcome these limitations by placing two different fluorescent protein coding sequences in a single transposon, either in tandem or antiparallel. Here we present the results of our initial tests with these designs. Results and Discussion The mosaic ends (MEs) that define the hyperactive Tn 5 transposon [ 9 ] have two possible open reading frames (ORFs) through them. We used one of these reading frames to construct the Either-Or transposon (<EYOR>, Figure 1A ). <EYOR> carries the sequence encoding the yellow fluorescent protein (YFP) at its 5' end, flanked by two 8 bp restriction sites ( Asc I – 5' and Srf I – 3'). An identical cassette encoding cyan fluorescent protein (CFP) flanked with Asc I and Srf I sites, is positioned in the same orientation at the 3' end of the transposon. Tn 5 transposition, in vitro , is only ~1% efficient [ 10 ], so the kanamycin resistance gene ( Kan R ) was incorporated between the two fluorescent protein cassettes. Since the YFP sequence has no start codon, it should only be translated if it inserts within another protein coding sequence in the correct orientation and reading frame. Plasmids with transposon insertions that are in-frame with respect to the target coding sequence can be rapidly identified by screening for YFP fluorescence in transiently transfected mammalian cells. Each of these clones produces a truncated fusion protein with YFP at the C-terminus. A stop codon is positioned downstream of the Srf I site to prevent translation beyond the YFP coding sequence. These truncated fusion proteins may provide additional information about which parts of the primary sequence contain trafficking signals. Full-length YFP and CFP fusion proteins are then generated in parallel from each clone by digestion with Srf I or Asc I and re-ligation. By producing identical full-length YFP and CFP fusion proteins from each in-frame insertion <EYOR> should double the efficiency with which multicolored fusion protein libraries can be generated. To test the <EYOR> transposon we targeted Prestin, an integral membrane protein expressed in outer hair cells of the cochlea and believed to be the motor responsible for their rapid changes in length in response to fluctuations in membrane voltage [ 11 ]. The 2.2 kb cDNA encoding Prestin was expressed in a 4.8 kb Ampicillin resistant ( Amp R ) CMV expression plasmid, pBNJ12.5. Transposon insertions that disrupt the plasmid origin or Amp R (together ~1.5 kb) are not recovered [ 8 ], so the predicted number of in-frame insertions in Prestin, is ~7%. After transposition, plasmids conferring Amp R and Kan R were isolated by standard mini-prep procedures and transiently expressed in HEK-293 cells in a 96-well format. In a random sample of 192 transposed clones, 32 produced detectable fluorescence. Eighteen of these (~9%) were clearly localized to intracellular membranes. (Figure 2A ). Of the remaining fluorescent proteins 11 appeared to be YFP alone, with evenly distributed fluorescence throughout the entire cell, and 3 were too dim to determine any sub-cellular localization. Sequencing revealed that the 18 proteins targeted to intracellular membranes were truncated Prestin-YFP fusion proteins resulting from in-frame insertions at 12 unique sites (Figure 2D ). One of the YFP-like clones was an in-frame insertion in the intracellular N-terminus of Prestin (at amino acids 28–30) upstream of the first predicted transmembrane domain. The remaining 13 fluorescent proteins resulted from <EYOR> insertions outside the Prestin coding sequence, most of them being clustered just downstream of the CMV promoter. We could not identify an in-frame start codon (AUG) in any of these clones. There is however, an in-frame CUG codon at the 5' end of the Tn 5 ME sequence that may present an alternate translation initiation site in the presence of the strong CMV promoter [ 12 ]. To verify that <EYOR> could be used to generate full-length fusions with either YFP or CFP unique in-frame clones were digested in parallel with Srf I or Asc I and re-ligated. The resulting fusion constructs were transiently expressed in HEK-293 cells and screened for YFP and CFP fluorescence (Figure 2B,2C ). All 13 unique insertion sites produced fluorescent full-length fusions with both CFP and YFP and all were localized to intracellular membranes. The <EYOR> design could be expanded for a wide range of protein tagging applications by replacing the secondary CFP cassette with another open reading frame. With such a transposon, YFP fluorescence would be used as a reporter to rapidly identify random in-frame insertions. Subsequent digestion with Asc I and re-ligation could then generate fusion proteins that might otherwise be difficult to screen for such as epitope tags, protease cleavage sites, or even a new N-terminus complete with a secretory signal peptide. Several groups have reported similar strategies based on multi-domain transposons for the random insertion of small peptide tags [reviewed in:[ 13 ]]. Like <EYOR>, these transposons utilize a primary reporter domain to identify in-frame insertions. Subsequent excision of the reporter domain (and selectable marker) then restores the full-length target coding sequence with an inserted peptide tag. The <EYOR> design is unique, however, in that its overlapping pairs of Asc I and Srf I restriction sites, allow the user to create identical full-length fusion proteins from both the reporter domain and the secondary coding sequence. The second transposon design, the Double-Barrel transposon (<DBT>, Figure 1B ), encodes green and red fluorescent proteins (GFP and DsRed) in opposite orientations. This means that any <DBT> insertion within another protein coding sequence has a 1:3 chance of being in-frame regardless of its orientation. Therefore, <DBT> should double the efficiency of random fusion protein generation, by producing equal numbers of GFP and DsRed fusions. In addition to their antiparallel orientation, the GFP and DsRed coding sequences in <DBT> each use a different relative reading frame through the Tn 5 MEs. As in <EYOR>, GFP fusion proteins are created by insertions after the third nucleotide of a target codon. The DsRed coding sequence, however, has been shifted by 1 nucleotide relative to the Tn 5 MEs. Therefore, DsRed fusion proteins are generated by transposon insertions between the second and third nucleotides. Using different reading frames for GFP and DsRed doubles the total number of insertion sites in the target coding sequence from which fusion proteins could potentially be made. While this does not alter the frequency of in-frame insertions, it does reduce the screening cost of saturating a target clone by increasing the probability of recovering unique in-frame insertions. To test the efficiency of fusion protein generation with the <DBT> transposon, we targeted pCMVI-9, a CMV expression plasmid carrying cDNA encoding the type 1 IP 3 receptor (IP 3 R) [ 14 ]. The IP 3 R is a ligand gated Ca 2+ channel, composed of 4 homomeric subunits, expressed within the membranes of the endoplasmic reticulum. Each IP 3 R subunit is over 2700 amino acids, and creating a full-length fluorescent IP 3 R fusion protein with such a large cDNA presents a formidable challenge for traditional molecular biological techniques. The high ratio of coding sequence to vector makes it an excellent target for transposition, however, with a predicted frequency of in-frame insertions of ~11% per fluorophore for <DBT>. At 24 hrs post-transfection, visual screening for fluorescence of 288 Amp R + Kan R clones in HEK-293 cells identified 44 clones that produced green fluorescent proteins. Of these, 35 displayed a uniform cytoplasmic distribution and exclusion from the nucleus (Figure 3A ), 3 showed fluorescence throughout the entire cell and 6 were too dim to determine any sub-cellular localization. Screening at several time points between 2 and 4 days post-transfection identified 7 clones producing red fluorescent proteins, 2 of which were clearly excluded from the nucleus (Figure 3B ). The remaining 5 red proteins were too dim to determine any sub-cellular localization. Sequencing out of the transposon confirmed that 41 of the clones encoding green proteins (~14%) and 2 clones encoding red proteins represented in-frame insertions. Consistent with the results of the Prestin transposition, all of the proteins with sub-cellular localization different from that seen with GFP alone were the product of in-frame insertions. After digestion and religation, only 20 of the full-length fusion proteins retained detectable levels of fluorescence (18 green, 2 red). These full-length proteins displayed a dramatic shift in their distribution, with clear ER localization (Figure 3C ). We chose GFP and DsRed to build the <DBT> transposon because their coding sequences are so dissimilar. Our concern was that if we chose two similar sequences, CFP and YFP for example, the antiparallel orientation of these coding sequences could produce extensive mRNA hybridization and secondary structure that would inhibit protein translation. It appears however, that DsRed is not well suited for insertion within other proteins. Indeed, DsRed has not been reported as a fusion protein in the middle of another protein, and even N- and C-terminal fusions with DsRed can be problematic [ 15 ], perhaps due to its being an obligate multimer [ 16 ]. Despite the low yield of DsRed fusions, these results demonstrate that <DBT> can be used to simultaneously generate full-length fusion proteins in two different reading frames. As novel fluorescent proteins are isolated from new species, or old ones are altered, this type of bi-directional transposon could potentially double the output of the screening process. Conclusions The transposons described here should greatly accelerate the creation of multicolored libraries of fluorescent fusion proteins. By creating identical full-length YFP and CFP fusion proteins from each in-frame insertion, the <EYOR> transposon not only facilitates the generation potential FRET pairs, it enables the direct comparison of different fluorophores in otherwise identical fusion proteins. The <DBT> design, on the other hand, has the capacity to double both the throughput of fusion protein generation by virtue of its bi-directionality as well as the total output of novel fusion proteins through the simultaneous use of multiple reading frames. Ultimately, the ability to generate large numbers of novel fusions proteins in days rather than months, should shift the limiting rate at which novel fluorescent protein biosensors are identified to functional screening rather than protein design and construction. Methods Plasmids PCR and standard subcloning procedures were used to create the plasmids encoding <EYOR> (pBNJ55b.1), <DBT> (pBNJ38.5) and Prestin (pBNJ12.5) (full sequences in supplementary material). The fluorescent protein coding sequences used were Venus (YFP) [ 17 ], ECFP-N164H (CFP), EGFP (GFP) and DsRed2 (Clontech). The Kan R gene was obtained from pUniV5-His-TOPO (Invitrogen). The construction of the IP3 receptor expression plasmid, pCMVI-9, was previously described [ 14 ]. Tn5 transposition and plasmid isolation Transposons were amplified from their host plasmids via PCR with a single primer complementary to the 19 bp Tn 5 ME (5'-CTGTCTCTTATACACATCT-3') and purified as previously described [ 8 ]. Purified transposon and target concentrations were each quantified against an independent DNA standard using a DynaQuant 200 fluorimeter. The transposition reaction was performed according to manufacturer's recommendations (Epicentre Technologies, Madison, WI) with 200 ng of target DNA and a molar equivalent of purified transposon. Electrocompetent XL-10 Gold E. coli (Stratagene, La Jolla, CA) were transformed with 0.5 μL of the transposition reaction and plated on LB agar with Ampicillin (100 μg/mL) and Kanamycin (50 μg/mL). Parallel plating of the transformation on LB agar with Ampicillin alone was used to establish the transposition efficiency. Transposed plasmids were isolated in a 96-well format from 1.25 mL LB cultures with Eppendorf PerfectPREP-96 Vac Direct Bind miniprep kits on a PerkinElmer MultiPROBE II HT liquid handling robot and eluted in 70 μL of ddH 2 O. Visual screening HEK-293 cells (American Type Culture Collection CRL-1573) were plated 24 hr prior to transfection, in 96-well glass bottom tissue culture plates (NalgeNUNC) at 6 × 10 4 cells in 100 μL of MEM-E with 10% fetal bovine serum. Transfections were performed with ~300 ng of plasmid DNA and 0.3 μL of Lipofectamine 2000™ (Gibco BRL) in a total volume of 50 μL of Opti-MEMI (Gibco BRL) per well. The cells were screened for fluorescence 24 hr after transfection with a 20× objective on a Zeiss inverted microscope with excitation and emission filter sets optimized for CFP/YFP or GFP/DsRed imaging (Omega, Brattleboro, VT). Sequencing and generation of full-length fusion proteins Exact insertion sites were identified for all fluorescent transposed clones by sequencing 5' out of the transposon with a primer complimentary to the <EYOR> YFP coding region (5'-CTGCAGGCCGTAGCC-3') or <DBT> GFP coding region (5'-TGGCCGTTTACGTCGCCGTCCA-3'). To generate full-length fusion proteins, plasmids with unique in-frame insertions were digested and re-ligated. After restriction digestion (100 ng of plasmid DNA and 0.5 U of Asc I or Srf I in 10 μL total volume), 1 μL of the digest reaction (~20 ng DNA) was re-ligated with Fast-Link™ ligase (Epicentre Technologies) for 15 min at room temperature (0.5 mM ATP, 1X Fast-Link™ buffer, 1 U ligase, 7.5 μL total volume). After heat inactivation (70°C for 15 min.), XL-10 Gold E. coli were transformed with 0.5 μL of the ligation reaction and plated on LB agar with Ampicillin. The following day, colonies were co-inoculated in LB with Ampicillin and Ampicillin + Kanamycin to verify loss of the Kan R prior to plasmid isolation. Authors' contributions D.S. conceived of the transposon designs and carried out their construction, performed the transposition and screening for fluorescent fusion proteins, and drafted the manuscript. T.H. participated in the study design, coordination, and analysis. All authors have read and approved the final manuscript. Supplementary Material Additional File 1 Supplementary Material-hughes This is a PDF file containing both plasmid maps and full sequence data for the plasmids used in this study. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514711.xml

374241

Dissection and Design of Yeast Prions

Many proteins can misfold into β-sheet-rich, self-seeding polymers (amyloids). Prions are exceptional among such aggregates in that they are also infectious. In fungi, prions are not pathogenic but rather act as epigenetic regulators of cell physiology, providing a powerful model for studying the mechanism of prion replication. We used prion-forming domains from two budding yeast proteins (Sup35p and New1p) to examine the requirements for prion formation and inheritance. In both proteins, a glutamine/asparagine-rich (Q/N-rich) tract mediates sequence-specific aggregation, while an adjacent motif, the oligopeptide repeat, is required for the replication and stable inheritance of these aggregates. Our findings help to explain why although Q/N-rich proteins are relatively common, few form heritable aggregates: prion inheritance requires both an aggregation sequence responsible for self-seeded growth and an element that permits chaperone-dependent replication of the aggregate. Using this knowledge, we have designed novel artificial prions by fusing the replication element of Sup35p to aggregation-prone sequences from other proteins, including pathogenically expanded polyglutamine.

Introduction The aggregation of misfolded proteins underlies a diverse range of human diseases, including sporadic amyloidoses such as Alzheimer's disease and hereditary neuropathies such as Huntington's disease ( Dobson 1999 ). Prions are a special class of protein aggregates that replicate their conformation and spread infectiously ( Prusiner 1998 ). After the discovery that prion aggregates are responsible for the mammalian transmissible spongiform encephalopathies, several epigenetically heritable traits in fungi were also found to depend on a prion mechanism ( Wickner 1994 ; Uptain and Lindquist 2002 ; Osherovich and Weissman 2004 ). In Saccharomyces cerevisiae and Podospora anserina, prions are transmitted from cell to cell through mating and cell division, resulting in readily assayed phenotypes with a non-Mendelian pattern of inheritance ( Liebman and Derkatch 1999 ). The yeast non-Mendelian factors [ PSI + ] ( Cox 1965 ) and [ URE3 ] ( Lacroute 1971 ), which are prion forms of the translation termination factor Sup35p and the transcriptional activator Ure2p, respectively, have served as useful models for the formation and replication of heritable protein aggregates. Prion forms of Sup35p and Ure2p lead to defects in their respective biochemical activities (translation termination and nitrogen catabolite repression). Mutational analysis has shown the glutamine/asparagine-rich (Q/N-rich) amino-terminal (N) domains of these proteins to be critical for prion behavior ( Ter-Avanesyan et al. 1993 ; Masison and Wickner 1995 ; Patino et al. 1996 ; Paushkin et al. 1996 ; DePace et al. 1998 ). In vitro, these Q/N-rich domains form self-seeding, β-sheet-rich amyloid fibrils similar to those associated with Alzheimer's and Huntington's diseases ( Glover et al. 1997 ; King et al. 1997 ; Taylor et al. 1999 ). The autocatalytic aggregation of yeast prion proteins often shows a high specificity for like molecules; for example, Sup35p N domains from different yeast species form prion aggregates that preferentially interact with molecules of their own kind ( Santoso et al. 2000 ; Chernoff et al. 2000 ; Kushnirov et al. 2000 ; Zadorskii et al. 2000 ; Nakayashiki et al. 2001 ). [ PSI + ] and [ URE3 ] can be eliminated by transient growth in the presence of guanidine hydrochloride (GuHCl), which “cures” cells of prions by inhibiting Hsp104p, a molecular chaperone needed for prion replication ( Chernoff et al. 1995 ; Jung et al. 2002 ; Ness et al. 2002 ). A surprisingly large number of proteins in S. cerevisiae and other eukaryotes have lengthy Q/N-rich tracts ostensibly similar to those found in the prion-forming domains of Sup35p and Ure2p ( Michelitsch and Weissman 2000 ). From among these, we and another group identified two novel proteins, New1p and Rnq1p, with prion-forming domains resembling those of Sup35p and Ure2p ( Santoso et al. 2000 ; Sondheimer and Lindquist 2000 ). When these Q/N-rich domains were fused to green fluorescent protein (GFP) and overexpressed, they formed visible aggregates resembling those of GFP-labeled Sup35p in [ PSI + ] cells. Fusion proteins in which these domains were introduced in place of the Sup35p prion domain could support distinct, self-specific prion states that recapitulated the translation termination defect associated with [ PSI + ]. Rnq1p was later shown to underlie a naturally occurring prion called [ PIN + ], which promotes the aggregation of Q/N-rich proteins such as Sup35p; overexpressed Sup35p forms aggregates and stimulates the appearance of [ PSI + ] only in [ PIN + ] strains ( Derkatch et al. 1997 ; Derkatch et al. 2001 ). Aggregates of the New1p prion domain, whether resulting from overexpression or from a constitutive prion form (termed [ NU + ]), also promoted the aggregation of other Q/N-rich proteins, emulating the effect of [ PIN + ] ( Osherovich and Weissman 2001 ). Many sequences with Q/N content as high as that of Sup35p and Ure2p, including human polyglutamine expansion disease proteins, form visible aggregates when overexpressed in yeast as GFP fusions ( Krobitsch and Lindquist 2000 ; Osherovich and Weissman 2001 ; Meriin et al. 2002 ). However, only a limited number of Q/N-rich sequences are bone fide prion domains capable of propagating these aggregates over multiple cell generations even when expressed at low levels (J. Hood and J.S.W, unpublished data). To understand what distinguishes generic Q/N-rich aggregates from heritable prions, we conducted a detailed dissection of the prion-forming regions of Sup35p and New1p. We found that the prion properties of Sup35p and New1p require the presence of two independent and portable sequence elements within their prion domains. One element mediates the growth of prion aggregates by incorporation of soluble monomers. The second promotes the inheritance of aggregates, generating new heritable “seeds” which can be partitioned between mother and daughter cells during cell division. Results Distinct Regions of the New1p Prion Domain Mediate Prion Growth and Division Sup35p can alternate between a biochemically active, soluble form ([ psi – ]) and an aggregated prion state ([ PSI + ]) with diminished translation termination activity, which can be monitored by nonsense suppression of the mutant ade1–14 allele ( Liebman and Derkatch 1999 ). Whereas [ psi – ] strains form red colonies on yeast extract-peptone-dextrose (YEPD) medium and cannot grow in the absence of adenine, [ PSI + ] strains suppress the premature stop codon in ade1-14 , and thus appear pink or white on YEPD medium and grow on adenine-free medium (a phenotype termed adenine prototrophy, Ade+). The N or prion domain of Sup35p (residues 1-112) is required for [ PSI + ] formation but is dispensable for the translation termination activity of the carboxy-terminal C domain ( Ter-Avanesyan et al. 1993 ). The charged middle domain (M) is not required for prion behavior, but modulates the efficiency of chaperone-dependent prion transmission ( Liu et al. 2002 ; L.Z.O., unpublished data) ( Figure 1 ). Two distinct regions in the N domain have previously been implicated in Sup35p aggregation: a Q/N-rich tract (residues 1–39) ( DePace et al. 1998 ) and an oligopeptide repeat (residues 40–112) that consists of five and a half degenerate repeats of the consensus sequence P/QQGGYQQ/SYN ( Liu and Lindquist 1999 ; Parham et al. 2001 ; Crist et al. 2003 ). Figure 1 Schematic Diagram of Sup35p and New1p Prion domains of both proteins are enlarged in the center, highlighting the Q/N-rich tract of Sup35p (blue), the NYN tripeptide repeat of New1p (purple), and the oligopeptide repeat sequences (orange) found in both proteins. The sequence of the NEW1 oligopetide repeat (residues 50–70) is QQQRNWKQGGNYQQGGYQSYN, while that of the adjacent tripeptide repeat region (residues 71–100) is SNYNNYNNYNNYNNYNNYNNYNKYNGQGYQ. We had earlier identified New1p as an uncharacterized protein with a Sup35p-like N-terminal domain; when fused to the M and C domains of Sup35p, the first 153 residues of New1p (New1 1–153 ) supported a [ PSI + ]-like prion state termed [ NU + ] ( Santoso et al. 2000 ). Sup35p and New1p have regions of clear similarity beyond their high Q/N content ( Figure 1 ). The prion domains of both have Q/N-rich tracts and oligopeptide repeat regions, although their order is reversed. The C-terminal domains of New1p and Sup35p also appear to be related, based on modest homology and the similarity of the translation termination defects in sup35 ( Song and Liebman 1985 ) and new1 mutants (L.Z.O., unpublished data). To understand the sequence requirements for the prion behavior of New1p, we constructed a series of truncated prion domains ( Figure 2 A) and examined their participation in several critical steps of the prion replication cycle. We first asked whether these truncated prion domains could form visible foci when fused to GFP (aggregation). Next, we examined whether such aggregates could convert New1 1–153 into a [ NU + ] prion state (induction). Finally, we fused these constructs to the M and C domains of Sup35p (–M-C), introduced them in place of endogenous SUP35, and assessed whether these proteins could adopt stable prion states (maintenance). Figure 2 Dissection of the New1p Prion Domain Reveals Distinct Regions Responsible for Aggregation and Prion Inheritance (A) Indicated fragments of New1p (left) were expressed as GFP fusions (inducers) in a [ nu – ] [ pin – ] strain, examined by microscopy for GFP aggregation, then plated on SD-ade medium to assess induction of [ NU + ]. The symbol “+” indicates induction frequencies of at least 5%; the symbol “–” indicates no induction. Maintenance was assessed by the ability of an episomal maintainer version of the indicated fragment to support an Ade+ state after overexpression of New1 1–153 -GFP (see Materials and Methods ). The aggregation of New1-GFP fusions (second column) has been previously reported ( Osherovich and Weissman 2001 ). (B) The NYN repeat of New1p induces [ NU + ] and [ NU + ] mini . New1 70–100 -GFP was overexpressed in [ nu – ] and [ nu – ] mini strains ([ pin – ] and [ PIN + ] derivatives of each), along with vector only or New1 1–153 -GFP controls. Averages of three independent trials, representing 600–2000 colonies, are shown for most induction experiments; inductions using New1 70–100 -GFP were conducted twice. Error bars show minimal and maximal observed induction efficiencies. (C) Reversibility of [ NU + ] mini . The [ pin – ] Ade+ convertants obtained in (B) were colony purified on SD-ade medium and confirmed to have lost the inducer plasmid. A stable [ NU + ] mini isolate is shown before and after induction, as well as after GuHCl treatment, along with [ nu – ] and [ NU + ] reference strains. We found that distinct regions within the New1p prion domain are necessary for the induction and maintenance of [ NU + ] ( Figure 2 A). The asparagine-tyrosine-asparagine (NYN) repeat (residues 70–100), which we had earlier shown to be sufficient for aggregation ( Osherovich and Weissman 2001 ), also proved sufficient for induction of [ NU + ]. As with the full-length New1p prion domain, overexpression of the NYN repeat efficiently stimulated the appearance of Ade+ in [ nu – ] cells ( Figure 2 B, left). However, stable prion maintenance required both the NYN repeat and the adjacent oligopeptide repeat. In a strain with this minimized New1p prion domain (residues 50–100), overexpression of the full prion domain or of the NYN repeat alone promoted the appearance of Ade+ colonies ( Figure 2 B, right). The resulting convertants remained Ade+ after loss of the inducer plasmid but reverted to Ade- after transient GuHCl treatment ( Figure 2 C). We conclude that the oligopeptide repeat and the NYN repeat of New1p together are sufficient to support a prion state, termed [ NU + ] mini , which recapitulates the characteristics of [ NU + ]. Dissection of the Sup35p Prion Domain In light of the similarity between New1p and Sup35p prion domains, we asked whether separate regions of Sup35p were involved in the induction and maintenance of [ PSI + ] aggregates ( Figure 3 ). We constructed a series of truncated Sup35p N domains and analyzed their behavior in the aggregation, induction, and maintenance assays described above for [ NU + ]. Additionally, we examined the ability of truncated N domains to decorate preexisting Sup35p aggregates in [ PSI + ] strains. Figure 3 Dissection of the Sup35p Prion Domain At top are schematic diagrams of these experiments; positive outcomes are shown below the arrows. In some cases, similar experiments have been reported by Parham et al. (2001 ) (indicated by “a”) and are repeated here as controls. Aggregation: Plasmid-borne M-GFP fusions of the indicated Sup35p N domain fragments (green) were overexpressed in a [ psi – ] [ PIN + ] strain and examined for fluorescent focus formation. The symbol “+” indicates that 10% or more of cells displayed aggregates. Sup35 1–57 -M-GFP displayed a lower frequency of aggregation (approximately 1%). Induction: Strains from the aggregation experiment were plated onto SD-ade medium and scored for growth to test whether aggregates of truncated protein (green) convert chromosomally encoded protein (blue) to [ PSI + ]. The symbol “+” indicates approximately 5–10% conversion frequency. Consistent with the aggregation experiment, Sup35 1–57 -M-GFP displayed a lower frequency of [ PSI + ] induction (approximately 1%). Decoration: Indicated proteins were expressed as –M-GFP fusions in [ PSI + ] [ PIN + ] cells, which were examined to determine whether GFP-labeled truncations (green) decorate preexisting aggregates of full-length Sup35p (blue). Curiously, Sup35 1–49 -M-GFP in [ PSI + ] cells formed abnormally large “ribbon” aggregates of the kind typically observed during de novo [ PSI + ] induction; furthermore, approximately 10% of the cells reverted to [ psi – ] (indicated by “*”). Thus, this truncation was a potent dominant PNM mutant. Maintenance: A SUP35-deleted [ PSI + ] [ PIN + ] bearing wild-type SUP35 maintainer (blue) was transformed with maintainer plasmids containing the indicated truncation (purple). The wild-type maintainer was lost by counterselection, and the resulting strain was tested for [ PSI + ] by color and growth on SD-ade medium. The Sup35 1–93 mutant displayed an intermediate pink color and grew poorly on SD-ade medium, as previously reported ( Parham et al. 2001 ). Note: King (2001 ) reports that Sup35 1–61 -GFP fusion could decorate [ PSI + ] aggregates in certain strains and could induce [ PSI + ] de novo when overexpressed. We found that the Q/N-rich tract and a small portion of the adjacent oligopeptide repeat are responsible for Sup35p aggregation and de novo [ PSI + ] induction. Deletions within the Q/N-rich tract or of oligopeptide repeat 1 abolished these properties, whereas a construct containing only the Q/N-rich region and the first two oligopeptide repeats (residues 1–64) aggregated and induced [ PSI + ] at levels comparable to the full prion domain, in agreement with King (2001 ). A construct (residues 1–57) with a partial deletion of oligopeptide repeat 2 could still aggregate and induce [ PSI + ], albeit with decreased efficiency. Although a construct lacking oligopeptide repeat 2 entirely (residues 1–49) did not induce [ PSI + ] de novo, this GFP fusion could nonetheless decorate preexisting Sup35p aggregates. Thus, while oligopeptide repeat 2 contributes to the aggregation of Sup35p, the primary determinants of prion induction reside in the amino-terminal Q/N-rich region and oligopeptide repeat 1. In contrast, the rest of the oligopeptide repeat region is needed for stable inheritance of [ PSI + ] aggregates. Constructs that did not form fluorescent foci could not retain [ PSI + ], suggesting that aggregation is a prerequisite for prion maintenance. However, aggregation is not sufficient for prion inheritance, as Sup35p constructs with deletions spanning oligopeptide repeats 3–5 could not support a prion state despite their ability to form aggregates and efficiently induce [ PSI + ]. Only the sixth (incomplete) oligopeptide repeat proved dispensable for [ PSI + ] maintenance, consistent with an earlier report ( Parham et al. 2001 ). The PNM2-1 Mutation in Oligopeptide Repeat 2 Specifically Compromises the Inheritance of [ PSI + ] Our deletion analysis suggested that oligopeptide repeat 2 participated in both the formation and inheritance of Sup35p aggregates. We made use of a point mutation within oligopeptide repeat 2 known as PNM2-1 (G58D) to distinguish between these two functions. PNM2-1 ( P SI N o M ore) shows strong interference with [ PSI + ] in certain strain backgrounds through a poorly understood mechanism ( McCready et al. 1977 ; Doel et al. 1994 ; Kochneva-Pervukhova et al. 1998 ; Derkatch et al. 1999 ). Using both in vivo and in vitro assays, we established that PNM2-1 does not have a defect in aggregation or [ PSI + ] induction. Earlier work indicated that PNM2-1 is capable of seeding [ PSI + ] in vivo ( Kochneva-Pervukhova et al. 1998 ; Derkatch et al. 1999 ; King 2001 ). Consistent with these reports, we found that overexpression of a PNM2-1-GFP fusion in [ psi – ] [ PIN + ] cells with a wild-type SUP35 locus led to both focus formation and [ PSI + ] induction ( Figure 4 A). A previous study of Sup35p polymerization in extracts had suggested that PNM2-1 might interfere with [ PSI + ] through a defect in seeding ( Kochneva-Pervukhova et al. 1998 ). We tested this by examining the rate of seeded polymerization of recombinant PNM2-1 protein. Like wild-type Sup35p, purified PNM2-1 spontaneously formed amyloid fibrils in vitro; this was accelerated by the addition of preformed Sup35p polymer seeds (data not shown). We measured the initial rates of polymerization of wild-type and PNM2-1 protein seeded by preformed wild-type polymers ( Figure 4 B) and by PNM2-1 polymers ( Figure 4 C) using a thioflavin-T–binding assay. We observed that wild-type and PNM2-1 monomers were seeded by wild-type polymers with similar kinetics; likewise, PNM2-1 polymers seeded both wild-type and PNM2-1 monomers equivalently. Thus, the PNM2-1 mutation does not affect polymerization or seeding. Figure 4 PNM2–1 (G58D) Prevents Inheritance But Not Aggregation of Sup35p Prions (A) PNM2-1 protein can seed [ PSI + ]. A Sup35p inducer containing the PNM2-1 (G58D) mutation was overexpressed in [ psi – ] [ PIN + ] cells; shown are cells (inset) with representative fluorescent foci, which were the same in frequency and appearance as cells with a wild-type inducer. Cells overexpressing inducer versions of wild-type Sup35p (SUP), an aggregation-defective N-terminal truncation (Δ1–38), and PNM2-1 were plated and scored for Ade+. Approximately 1000 colonies were counted. (B) PNM2-1 protein polymerization is similar to that of wild-type protein. (C) Preformed PNM2-1 polymers seed wild-type and PNM2-1 monomers with comparable efficiency. Endpoint PNM2-1 polymers were used to seed fresh reactions. (D) PNM2-1 displays a partially dominant, incompletely penetrant defect in [ PSI + ] maintenance. [ psi – ] (1) and [ PSI + ] (2) SUP35::TRP1 pSUP35 controls are shown. [ PSI + ] [ PIN + ] SUP35::TRP1 pSUP35 was transformed with a second maintainer expressing PNM2-1 (3). The wild-type maintainer (pSUP35) was then lost through counterselection (4). Red sectors from (4) were isolated, retransformed with the wild-type maintainer, and allowed to lose the PNM2-1 maintainer (5). (E) Mitotic instability of [ PSI + ] in the PNM2-1 strain. A pink (Ade+) [ PSI + ] [ PIN + ] PNM2-1 isolate was grown to log phase in SD-ade liquid then shifted into nonselective (YEPD) medium. At indicated time points, aliquots were plated onto SD-ade and YEPD media to determine the fraction of [ PSI + ] cells (minimum of 200 colonies counted per time point). Whereas a wild-type control remained [ PSI + ] through the experiment, the PNM2-1 strain rapidly lost [ PSI + ] during logarithmic growth; during stationary phase (18 h and beyond), the percentage of [ PSI + ] PNM2-1 strains remained unchanged (approximately 5%). (F) Propagon count of PNM2-1 vs. wild-type [ PSI + ] strains. The majority of PNM2-1 cells had no [ PSI + ] propagons (i.e., were [ psi – ]). In both strains, a small number of “jackpot” cells contained over 200 propagons; see Cox et al. (2003 ). Instead, the PNM2-1 strain shows a marked defect in the inheritance of [ PSI + ]. When the wild-type SUP35 gene of a [ PSI + ] strain was replaced with PNM2-1, the strain retained the prion on synthetic defined (SD) yeast medium that selected for [ PSI + ] (SD-ade medium) but reverted to [ psi – ] at a high frequency in nonselective YEPD medium, resulting in sectored colonies ( Figure 4 D). We measured the rate of [ PSI + ] loss in a PNM2-1 strain by growing it in YEPD medium and, at various time points, plating aliquots of the culture onto SD-ade medium to determine the fraction of cells that had retained [ PSI + ] ( Figure 4 E). A wild-type strain retained [ PSI + ] in all of the cells throughout the experiment. By contrast, in the PNM2-1 strain the fraction of [ PSI + ] cells decreased rapidly while the cells grew logarithmically, but remained at a constant level when the cells entered stationary phase. These findings indicate that PNM2-1 acts to eliminate [ PSI + ] in dividing cells, consistent with a defect in prion replication. We next used a recently described assay to measure the number of heritable prion seeds (propagons) in a PNM2-1 strain. Here, prion replication is inhibited by GuHCl treatment. As the cells divide, preexisting propagons are diluted but not destroyed. The number of propagons present in a colony arising from a single cell is then evaluated by removing the GuHCl prion replication block after a large number (10 or more) of cell divisions and counting the total number of [ PSI + ] cells in that colony ( Cox et al. 2003 ). Whereas a wild-type strain had a median of 92 ( n = 24) propagons per cell, the PNM2-1 strain had dramatically fewer: 41 of 50 cells had no [ PSI + ] propagons at all (i.e., were [ psi – ]), and among the remaining nine [ PSI + ] cells, the median propagon number was six ( Figure 4 F). Thus, although a PNM2-1 strain can harbor [ PSI + ] prions, a defect in propagon replication causes mitotic instability, demonstrating the importance of oligopeptide repeat 2 in prion replication or segregation. Design of Novel Prion Domains Our data suggested that the formation and inheritance of prions involve distinct regions of Sup35p and New1p prion domains. To assess the interchangeability of these prion domain components, we constructed a chimeric prion domain, termed F, in which the aggregation-determining NYN repeat of New1p was fused to the oligopeptide repeats of Sup35p ( Figure 5 A). While initially soluble and active, a fusion of F and the Sup35p M and C domains (F-M-C) could be converted into an aggregated state, termed [ F + ], after transient overexpression of F-M-GFP. As with [ NU + ], [ F + ] induction did not require [ PIN + ] (data not shown). [ F + ] could be eliminated by GuHCl treatment ( Figure 5 B) and was inherited in a dominant, non-Mendelian manner ( Figure 5 C). As with Sup35p in a [ PSI + ] strain, F-M-C protein in [ F + ] but not in [ f – ] extracts sedimented entirely to the pellet fraction following high-speed centrifugation ( Figure 5 D). Thus, [ F + ] results from a prion state of F-M-C. Figure 5 F, A New1p–Sup35p Chimera, Shows Prion Characteristics of New1p (A) Schematic diagram illustrating the construction of chimera F. (B) Chimera F forms a prion, [ F + ]. The SUP35 gene in a [ psi – ] [ pin – ] strain was replaced with the F-M-C fusion; after transient overexpression of F-M-GFP, approximately 10% of these cells converted from an Ade- ([ f – ]) to an Ade+ ([ F + ]) state. Shown are examples of[ f – ] and [ F + ] strains, before and after GuHCl treatment, along with [ psi – ] and [ PSI + ] controls. (C) Non-Mendelian inheritance of [ F + ]. A diploid made by mating a [ F + ] MAT a strain against an [ f – ] MAT α displayed a [ F + ] phenotype and, when sporulated, produced four [ F + ] meiotic progeny. All 11 tetrads examined showed this 4:0 pattern of inheritance. (D) Sedimentation analysis of F-M-C. Extracts of [ f – ] and [ F + ] strains, along with [ psi – ] and [ PSI + ] controls, were subjected to 50K × g ultracentrifugation for 15 min. Total, supernatant, and pellet fractions were separated by SDS-PAGE, transferred to nitrocellulose, and probed with anti-SUP35NM serum. As with Sup35p, the prion form of F-M-C sediments primarily to the pellet but remains in the supernatant in [ f – ]. (E) F-M-GFP overexpression induces [ NU + ] but not [ PSI + ]. Indicated inducers and maintainers were used in an induction experiment. The symbol “+” indicates approximately 5–10% conversion to Ade+. F induced [ NU + ] at a comparable efficiency to New1 1–153 ; although New1 1–153 overexpression promoted the appearance of Ade+ colonies in the F-M-C strain, these were fewer in number (less than 5%) and reverted to Ade- after restreaking. (F) [ F + ] and [ NU + ] prion proteins interact with each other but not with [ PSI + ]. Episomal “second maintainers” were introduced into the indicated strains, along with an empty vector control. Antisuppression (red) indicates that the second maintainer is soluble, while white/pink indicates coaggregation of the endogenous and episomal maintainers. We next explored the specificity of [ F + ] prion seeding. Overexpression of the Sup35p prion domain did not induce [ F + ]; conversely, F-M-GFP overexpression did not induce [ PSI + ] ( Figure 5 E). However, F-M-GFP readily induced [ NU + ], indicating that mismatched sequences outside of the aggregating region did not prevent cross-interactions between heterologous proteins. Interestingly, overexpression of New1 1–53 -GFP induced Ade+ colonies in the [ f – ] strain, but this adenine prototrophy proved unstable. We also examined the ability of preexisting prion aggregates to recruit different prion-forming proteins using an antisuppression assay ( Santoso et al. 2000 ) ( Figure 5 F). [ PSI + ], [ F + ], and [ NU + ] strains were transformed with Sup35p–, F-M-C– or New1 1–153 -M-C–encoding plasmids; the color of the resulting colonies indicates whether the second maintainer protein is soluble (red) or aggregates as a result of the resident prion (pink/white). Consistent with the induction data, F-M-C and New1 1–153 -M-C were not incorporated into [ PSI + ] aggregates; likewise, Sup35p did not interact with [ F + ] or [ NU + ] aggregates. However, [ F + ] prions recruited New1 1–153 -M-C and, to a lesser extent, [ NU + ] recruited F-M-C. Thus, F and New1p prion domains can cross-interact during de novo induction and at normal levels of expression, indicating that the NYN repeat is sufficient to specify homotypic interaction between two otherwise distinct prion domains. Can a simple aggregation-prone sequence such as polyglu-tamine ( Zoghbi and Orr 2000 ) be turned into a heritable prion by fusion to an oligopeptide repeat? We designed artificial prion domains containing short (Q22) and pathogenically expanded (Q62) polyglutamine tracts, either alone or adjacent to the Sup35p oligopeptide repeat ( Figure 6 A). These domains were fused to –M-GFP and –M-C to create polyglutamine inducers and maintainers, respectively. Q22 inducers did not aggregate upon overexpression, but Q62 inducers (with and without oligopeptide repeats) formed visible foci in [ psi – ] [ PIN + ] cells ( Figure 6 B). Transient overexpression of Q62 inducers had no effect on the Q22 maintainer with the oligopeptide repeat or on the Q62 maintainer lacking the oligopeptide repeat. However, the Q62 maintainer with an oligopeptide repeat supported prion inheritance, converting to a stable Ade+ state following overexpression of the cognate inducer ( Figure 6 C). Several tests confirmed the prion nature of this state, termed [ Q + ]. Like [ PSI + ], [ Q + ] did not require the presence of the inducer plasmid after transient overexpression, was sensitive to GuHCl treatment ( Figure 6 D), and displayed a dominant, non-Mendelian pattern of inheritance ( Figure 6 E). We further tested the specificity of the [ Q + ] state by introducing a plasmid encoding a noncognate second maintainer into a [ Q + ] strain ( Figure 6 F). The Q62 maintainer failed to be incorporated into [ PSI + ] aggregates, causing antisuppression (red); conversely, Sup35p did not enter [ Q + ] aggregates. Figure 6 [ Q + ], a Prion Form of Pathogenically Expanded Polyglutamine (A) Schematic illustrating the construction of polyglutamine-derived prion domains. (Op) indicates the presence of the Sup35p oligopeptide repeats (residues 40–124). (B) Fluorescence micrographs of [ psi – ] [ PIN + ] strains expressing indicated polyglutamine inducers. (C) Polyglutamine-based prion inheritance. Strains with indicated inducers and maintainers were plated onto SD-ade and YEPD media to determine the fraction of Ade+ after a standard induction experiment. Interestingly, Q62 inducer forms aggregates but does not promote Ade+ in the Q62(Op) maintainer strain. Note that Q62(Op) shows a high rate of spontaneous appearance of Ade+. (D) GuHCl sensitivity of the [ Q + ] state. An Ade+ convertant obtained in (C) was restreaked to lose the inducer plasmid, then grown on GuHCl. Shown are plates before and after GuHCl treatment, along with [ psi – ] and [ PSI + ] controls. (E) Dominance and non-Mendelian inheritance of [ Q + ]. See Figure 5 C. (F) [ Q + ] does not interact with Sup35p and vice versa. [ Q + ] and [ PSI + ] strains were transformed with indicated maintainers; mismatches between the maintainer and the chromosomally encoded allele result in antisuppression (red). Discussion A number of epigenetic traits in fungi result from the stable inheritance of self-propagating, infectious protein aggregrates (prions) ( Uptain and Lindquist 2002 ). Prion inheritance requires three sequential events that must keep pace with cell division to preserve the number of heritable prion units, or propagons, per cell ( Osherovich and Weissman 2004 ). First, prion aggregates must grow in size by incorporating newly synthesized protein. Next, these enlarged aggregates must be divided into smaller ones through the action of cellular chaperones ( Kushnirov and Ter-Avanesyan 1998 ; Borchsenius et al. 2001 ; Ness et al. 2002 ; Kryndushkin et al. 2003 ). Finally, these regenerated propagons must be distributed to mother and daughter cells ( Cox et al. 2003 ); for small, cytoplasmic aggregates, this distribution may occur passively by diffusion during cytokinesis. In the present study, we have dissected the prion-forming domains of Sup35p and New1p to discover the sequence elements involved in these steps. We have found that these domains consist largely of modular, interchangeable elements that serve distinct functions of prion growth and division or transmission. Aggregation underlies the growth phase of the prion replication cycle ( Figure 7 A) and occurs through the templated addition of conformationally compatible monomers onto preexisting seeds. Like other amyloids, yeast prions display a high specificity for homotypic aggregation ( Santoso et al. 2000 ; Chernoff et al. 2000 ; Kushnirov et al. 2000 ; Zadorskii et al. 2000 ; Nakayashiki et al. 2001 ). This discrimination arises from differences in the amino acid sequence and the conformation ( Chien and Weissman 2001 ) of the aggregation-promoting Q/N-rich elements found in each yeast prion protein. Aggregation and specificity are dictated by the NYN repeat (residues 70–100) of New1p and by the Q/N-rich amino terminal region (residues 1–57) of Sup35p. Figure 7 Model for Prion Growth and Division (A) During prion growth, polymers seed the incorporation of monomers through interactions between Q/N-rich aggregation sequences (blue). Proteins with noncognate aggregation sequences (red) are excluded. (B) The division phase of prion replication requires the oligopeptide repeats (orange), which may facilitate the action of chaperones such as Hsp104p (scimitar) in breaking the polymer into smaller, heritable units. In contrast, the conserved oligopeptide repeat sequence mediates the division and/or segregation phase of prion replication ( Figure 7 B). In New1p, the NYN repeat alone can aggregate and induce [ NU + ] but requires an adjacent oligopeptide repeat to form a minimal heritable New1p prion, [ NU + ] mini . Similarly, in Sup35p, the Q/N-rich amino terminal region mediates aggregation whereas most of the oligopeptide repeats are needed for the inheritance of [ PSI + ] propagons. Oligopeptide repeats 1 and 2 appear to contribute to both growth and inheritance, consistent with earlier reports that expansion and deletion within this region modulate in vitro polymerization of Sup35p and the appearance of [ PSI + ] in vivo ( Liu and Lindquist 1999 ). However, the two functions can be distinguished by a point mutant in oligopeptide repeat 2 (PNM2-1), which displays a specific defect in [ PSI + ] inheritance despite normal aggregation. Certain [ PSI + ] variants are resistant to the dominant negative effect of PNM2-1 ( Derkatch et al. 1999 ; King 2001 ); this suggests that although oligopeptide repeat 2 is critical for the replication of the [ PSI + ] variant used in our studies, it may be less important for the replication of other Sup35p prion conformations. Many studies have established that prion inheritance requires the action of cellular chaperones such as Hsp104p and Hsp70 proteins (reviewed in Osherovich and Weissman 2002 ), although how these proteins contribute is poorly understood. We propose that oligopeptide repeats turn nonheritable aggregates into prions by facilitating chaperone-mediated division. Oligopeptide repeats may allow the division of aggregates by providing direct binding sites for chaperones or by altering the conformation of the amyloid core to allow chaperone access. An earlier study established that deletion of residues 22–69 of Sup35p (which include parts of both the Q/N tract and the oligopeptide repeat) interferes with both [ PSI + ] induction and chaperone-mediated prion disaggregation ( Borchsenius et al. 2001 ). Unlike the Δ22–69 mutant, the prion replication defect in PNM2-1 could not be corrected by increasing Hsp104p levels (data not shown), arguing that the mitotic instability of PNM2-1 [ PSI + ] is not due solely to inadequate Hsp104p binding. Our findings help to explain why, among many Q/N-rich proteins in yeast, only a small subset form heritable prions. While many Q/N-rich proteins can aggregate when overexpressed ( Sondheimer and Lindquist 2000 ; Derkatch et al. 2001 ; Osherovich and Weissman 2001 ), prion inheritance of such aggregates requires that they be divided and passed on to the next generation. Although the inheritance of Sup35p and New1p prions is mediated by oligopeptide repeats, other sequences could also serve this purpose. Ure2p lacks an oligopeptide repeat; interestingly, many isolates of [ URE3 ] are mitotically unstable in the absence of selection ( Schlumpberger et al. 2001 ). Rnq1p, which underlies [ PIN + ], also lacks a strict oligopeptide repeat, but a region (residues 218–405) within its prion domain has an amino acid content reminiscent of the oligopeptide repeat sequence (i.e., numerous Q, N, S, Y, and G residues) ( Resende et al. 2003 ). Only two other yeast proteins, YDR210W and YBR016W, have clearly recognizable oligopeptide repeats; both proteins also have Q/N-rich regions. YBR016W forms aggregates when overexpressed ( Sondheimer and Lindquist 2000 ), but it is not known whether either protein can maintain a heritable aggregated state. Although the mammalian prion protein PrP contains a sequence resembling the oligopeptide repeat that can functionally replace one of the Sup35p repeats ( Parham et al. 2001 ), it is unclear whether this sequence is important in the replication of the PrP Sc state. The interchangeable nature of prion domain components allowed us to design novel artificial prions. The F chimera, consisting of the aggregation sequence of New1p and the oligopeptide repeat of Sup35p, demonstrates that the growth and specificity of prions is largely determined by the Q/N-rich tract, not by the oligopeptide repeat. Despite a sequence derived primarily from Sup35p, the F chimera behaved like New1p rather than like Sup35p. The [ F + ] prion cross-interacted with New1p but not Sup35p. Like [ NU + ], [ F + ] could be induced in the absence of a prion-promoting (PIN) factor. Finally, [ F + ] could itself act as a PIN factor, as does [ NU + ] (data not shown). Notably, the NYN repeat of New1p functions as an aggregation module apparently without regard to its position within a protein; this sequence induced prions when overexpressed by itself or with oligopeptide repeats at its N-terminal (in New1 1–153 and New1 50–100 ) or C-terminal regions (in the F chimera). These results suggest that aggregation sequences are portable and functionally separable from the oligopeptide repeat, perhaps constituting a structurally discrete amyloid core. Indeed, a peptide derived from the amino-terminal region of Sup35p forms a self-seeding amyloid in vitro ( Balbirnie et al. 2001 ). A simple aggregation-prone sequence, pathogenically expanded glutamine, also supports prion inheritance when adjacent to the oligopeptide repeat, suggesting that prion domains can consist of little more than a generic, aggregating core sequence and an inheritance-promoting element. In addition to illuminating the principles of yeast prion domain architecture, artificial prions with distinct specificity may be useful as controllable epigenetic regulators of protein activity. Such prion “switches” can be turned on and off by transient overexpression and genetic repression; for example, the Q prion domain could be fused to other proteins in order to conditionally and reversibly inactivate them independently of [ PSI + ]. It may also be possible to design additional artificial yeast prion domains whose aggregation is driven by non-Q/N-rich amyloidogenic proteins such as the Aβ peptide that accumulates in Alzheimer's disease ( Koo et al. 1999 ) or the mammalian prion protein PrP ( Cohen and Prusiner 1998 ). Such artificial prions could serve as models for aggregate–chaperone interactions in metazoans and could provide a genetic system for the high-throughput screening of modulators of human aggregation diseases. Materials and Methods Yeast strains and methods Derivatives of W303 ( Osherovich and Weissman 2001 ), with the initial genotypes ade1-14, his3-11,15, leu2-3, trp1-1, and ura3-1, were used throughout unless otherwise noted; all strains were [ PIN + ]. Strain numbers, with indicated genotypic differences, are as follows: YJW 584 [ psi – ] MAT a , YJW 508 [ PSI + ] MAT α , YJW 716 [ nu – ] MAT α sup35 ::TRP1 pRS315SpNew1 1–153 -M-C, YJW 717 [ NU + ] MAT α sup35 ::TRP1 pRS315SpNew1 1–153 -M-C, YJW 844 [ f – ] MAT α sup35 ::F-M-C C.g. HIS3, YJW 881 [ F + ] MAT a sup35 ::F-M-C C.g. HIS3, YJW 867 [ q – ] MAT α sup35 ::Q-M-C C.g. HIS3, YJW 868 [ Q + ] MAT a sup35 ::Q-M-C C.g. HIS3. Maintainer plasmids used in Figure 3 (see plasmid and gene replacement construction, below) were introduced by plasmid shuffling into YJW 716 or YJW 753 ([ PSI + ] MAT a sup35 ::TRP1 pRS316SpSUP35), followed by loss of the maintainer spontaneously or through 5-FOA counterselection. The PNM2-1 strain in Figure 4 was generated in this manner and was subsequently restreaked on SD-ade to select for [ PSI + ]. HIS3-marked oligopeptide repeat truncations and PNM2-1 maintainers were from Parham et al. (2001 ); all other Sup35p and New1p maintainers were marked with LEU2. The [ f – ] strain was generated by “gamma” chromosomal integration of pRS306 F-M-C into the SUP35 locus of YJW 584; excision of the wild-type gene was confirmed by PCR of Ade- colonies arising from subsequent growth on 5-FOA. The [ q – ] strain was made by “omega” chromosomal gene replacement ( Kitada et al. 1995 ) of SUP35 with a C.glabrata HIS3-marked –M-C variant (with or without oligopeptide repeats) into the SUP35 locus of a diploid [ PSI + ] [ PIN + ] strain. After sporulation, gene replacement was confirmed by PCR and by loss of [ PSI + ] in half of the haploid progeny. Yeast culture methods were according to standard procedures ( Sherman 1991 ), but YEPD-medium plates contained 1/4 of the standard amount of yeast extract to accentuate color phenotypes. For prion curing, strains were grown on YEPD medium plus 3 mM GuHCl, then restreaked onto YEPD medium. Plasmid and gene replacement construction The modular SUP35 cloning system described in previous reports was used throughout ( Santoso et al. 2000 ; Osherovich and Weissman 2001 ). All plasmids are derived from Sikorski and Hieter (1989 ); sequence files of all constructs are available as a web supplement ( Data S1 ). Maintainer plasmids are low-copy CEN/ARS (pRS31x series) with the native SUP35 promoter (Sp) driving the expression of the indicated prion domain followed by the M and C domains of Sup35p. Inducer plasmids are high-copy 2μ (pRS42x series) with the inducible CUP1 promoter (Cp) driving the expression of the indicated prion domain followed by the Sup35p M domain and GFP. New1p inducers did not include the Sup35p M domain. For polyglutamine constructs, polyglutamine tracts (22 and 62) were amplified out of the MJDtr constructs used in an earlier study ( Osherovich and Weissman 2001 ). To permit amplification, primers contained sequences homologous to several codons adjacent to the 5′ and 3′ ends of the polyglutamine tracts plus an initiator ATG codon. Thus, the polyglutamine sequences read MAYFEK(Q22/62)DLSG. The resulting PCR fragments were cloned into maintainer and inducer plasmids, which were used as templates for gene replacement PCR (see yeast strains and methods, above). In vivo prion assays For aggregation, inducers were overexpressed by growth of cells in selective medium with 50 μM CuSO 4 until the culture reached stationary phase; cells were then examined by fluorescent microscopy (Zeiss Axiovert, Zeiss, Oberkochen, Germany; Metamorph imaging software, Universal Imaging Corporation, Downingtown, Pennsylvania, United States). Unless otherwise noted, cultures displaying 10% or more cells with foci were scored as positive. For induction, dilutions of the above cultures were plated onto SD-ade and YEPD media to determine percentage of Ade+. In qualitative assessments, strains were scored as positive if 5% or more of plated cells grew on SD-ade medium after 5 d. In [ NU + ] maintenance experiments, strains with indicated maintainers were tested for the ability to support an Ade+ state following New1 1–153 -GFP overexpression. In [ PSI + ] maintenance experiments, strains that began as [ PSI + ] were tested for Ade+ after plasmid shuffle gene replacement with the indicated maintainer. For decoration, a [ PSI + ] [ PIN + ] strain was transformed with the indicated inducers, grown in selective medium with 50 μM CuSO 4 , and examined by fluorescence microscopy during midlogarithmic phase. Propagon counts were performed as described in Cox et al. (2003 ). For the antisuppression assay, indicated strains were transformed with a second, differently marked maintainer plasmid, and color phenotypes were assayed on medium selective for both plasmids. In vitro prion assays Centrifugation was performed as described in Ness et al. (2002 ). Immunoblots were visualized with MT130 anti-Sup35p N-M domain serum.For the polymerization of PNM2-1, the PNM2-1 N and M domains were cloned as 7-histidine fusions into pAED4 and expressed and purified as described in DePace et al. (1998 ). Thioflavin-T binding was conducted as in Chien et al. (2003 ). The slope of early (0–6 min) dye binding was obtained from seeded polymerization reactions conducted in triplicate. To correct for a difference in dye binding between wild-type and PNM2-1 protein, these values were normalized to the end point (90 min) maximum signal for each protein. Monomer concentrations were 2.5μM. Supporting Information Data S1 DNA Sequences of Constructs (30 KB ZIP). Click here for additional data file. Accession Numbers The GenBank accession numbers for the proteins discussed in this paper are Hsp104p (NP_013074), New1p (NP_015098), Rnq1p (NP_09902), Sup35p (NP_010457), Ure2p (NC_014170), YDR210W (NP_010496), and YBR016W (NP_010319).

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC374241.xml

545204

Retinoic Acid and Arsenic for Treating Acute Promyelocytic Leukemia

What were the critical steps in the development of ATRA and arsenic as treatments for APL? Researchers in Shanghai tell the story and look to the future

Acute promyelocytic leukemia (APL) was first identified as a distinct subtype of acute myeloid leukemia in 1957 by Leif Hillestad. It is called M3 in the French–American–British classification, with a variant type referred to as microgranular (M3v in the French–American–British nomenclature) [1] . APL is characterized by three features: (1) the presence of an accumulation of abnormal promyelocytes (see Glossary ) that do not differentiate into mature granulocytes, (2) the occurrence of fibrinogenopenia and disseminated intravascular coagulation that is often worsened by chemotherapy, and (3) the presence of the specific chromosomal translocation t(15;17)(q22;q21) ( Figure 1 ). Glossary Apoptosis: A genetically determined process of cell death in which the cell uses specialized cellular machinery to kill itself and is then eliminated by phagocytosis or by shedding. Caspase: A family of cysteine proteases with aspartate specificity that are essential intracellular death effectors. Disseminated intravascular coagulation: A hemorrhagic disorder that occurs following the uncontrolled activation of clotting factors and fibrinolytic enzymes throughout small blood vessels, resulting in depletion of clotting factors and tissue necrosis and bleeding. Fibrinogenopenia: A decrease in concentration of fibrinogen in the blood. Granulocyte: Terminally differentiated myelocyte or polymorphonuclear white blood cell (as a basophil, eosinophil, or neutrophil) with granule-containing cytoplasm. Ligand-inducible transcription factors: Transcription factors that structurally have domains associated with DNA binding and ligand (hormone) recognition. When binding to its specific ligand, the transcription factor initiates a series of conformational changes and interacts efficiently with its specific DNA response element to recruit components of the transcriptional machinery. Nuclear receptor superfamily: One of the most abundant classes of transcriptional regulators including receptors for steroid hormones (e.g., estrogens, glucocorticoids, and vitamin D3), RAs, thyroid hormones, and so on. These transcription factors regulate diverse functions such as homeostasis, reproduction, development, and metabolism in animals. Promyelocyte: Granule-containing cell in bone marrow that is in an intermediate stage of development between myeloblasts and myelocytes and that gives rise to a granulocyte. Proteasome: Proteolytic complex that degrades cytosolic and nuclear proteins. Sumoylation: Post-translational modification of proteins by the small ubiquitin-like modifier SUMO. Ubiquitin: A chiefly eukaryotic protein that when covalently bound to other cellular proteins marks them for proteolytic degradation. Figure 1 The Three Features of APL The three features of APL are (A) accumulation of abnormal promyelocytes, (B) fibrinogenopenia and disseminated intravascular coagulation, and (C) the chromosomal translocation t(15;17)(q22;q21), the resultant fusion transcripts, and variants. APL accounts for 10%–15% of all cases of acute myeloid leukemia, with several thousand new cases diagnosed worldwide each year. Before the advent of differentiation therapy, APL was treated with anthracycline-based chemotherapy with a complete remission rate of 60%–76% and a 5-year event-free survival rate of 23%–35% [ 1 , 2 ]. Differentiation Therapy: From Hypothesis to Practice Failure to differentiate terminally characterizes most, if not all, cancer cells of every origin. Whether the induction of differentiation could be a treatment strategy for cancers was hotly debated for decades before the advent of differentiation therapy. An important discovery of the early 1970s was that myeloid leukemic cells could be reprogrammed to resume normal differentiation and to become non-dividing mature granulocytes or macrophages as a result of stimulation by various cytokines [ 3 , 4 ]. Based on this discovery, Leo Sachs at the Weizmann Institute of Science, Rehovot, Israel, hypothesized in 1978 that treatment with agents that induce cancer cells to complete differentiation could be a potential therapeutic option for patients with cancer [5] . In the early 1980s, Breitman and colleagues showed that retinoic acid (RA; Figure 2 ), a derivative of vitamin A, could induce terminal differentiation of human promyelocytic leukemic cells in vitro [ 6 , 7 ]. But the first clinical reports of using RA showed conflicting results. Some case reports showed beneficial effects of 13- cis RA in people with refractory or relapsed APL [ 8 , 9 , 10 ], but other reports showed that 13- cis RA was ineffective in treating APL [11] . Figure 2 Isomers of RA Beginning in the early 1980s, the Shanghai Institute of Hematology conducted a series of experiments on differentiation therapy for APL. These experiments showed that all-trans RA (ATRA) could induce terminal differentiation of HL-60, a cell line with promyelocytic features, and fresh leukemic cells from patients with APL. These intriguing results were the impetus for a clinical trial. Twenty-four patients with APL were treated with ATRA (45 to 100 mg/m 2 /day). The result was dramatic: 23 patients (95.8%) went into complete remission (CR) without developing bone marrow hypoplasia or abnormalities of clotting. The remaining one patient achieved CR when chemotherapy was added [12] . Morphological maturation of bone marrow cells was found in all patients studied. These results were later confirmed by many randomized studies in centers around the world. Further trials showed improved rates of CR, a decrease in severe adverse effects, and lengthening of the duration of remission. Table 1 summarizes the CR rates obtained in most large series of patients. Currently, ATRA combined with anthracycline-based chemotherapy can achieve CR in 90%–95% of patients with APL and overall 5-year disease-free survival in up to 75% of patients [13] . Table 1 CR Rate in Patients with APL Treated with ATRA (in Series Including More Than 50 Cases) Chemo, chemotherapy Arsenic: “Ancient Remedy Performs New Tricks” Arsenic is a common, naturally occurring substance that exists in organic and inorganic forms in nature. The organic arsenicals consist of an arsenic atom in its trivalent or pentavalent state linked covalently to a carbon atom. There are three inorganic forms of arsenic: red arsenic (As 4 S 4 , also known as “realgar”), yellow arsenic (As 2 S 3 , also known as “orpiment”), and white arsenic, or arsenic trioxide (As 2 O 3 ) [14] . Arsenic was used to treat chronic myelogenous leukemia in the 18th and 19th centuries, but was discarded as a treatment in the early 20th century because of its toxicity and the advent of radiation and cytotoxic chemotherapy. In the early 1970s, a group from Harbin Medical University in China found that intravenous infusions of Ailing-1, a crude solution composed of 1% arsenic trioxide with a trace amount of mercury chloride, induced CR in two-thirds of patients with APL. There was an impressive 30% survival rate after 10 years [ 15 , 16 ]. Pure arsenic trioxide at 0.16 mg/kg/day for 28–54 days was shown to induce CR in 14 out of 15 (93.3%) patients with relapsed APL [17] . Tetra-arsenic tetra-sulfide was also reported to be effective in APL treatment [18] . Since 1996, a large number of reports have shown that arsenic compounds induce a CR in 85% to 90% of patients with both newly diagnosed and relapsed APL [13] . Furthermore, after CR is achieved by arsenic compounds, a molecular remission (i.e., negative for promyelocytic leukemia RA receptor a [PML-RARa] transcript detected by reverse transcriptase polymerase chain reaction) is obtainable either with arsenic compounds or with ATRA and chemotherapy as consolidation treatment. It seems likely that arsenic compounds appropriately used in post-remission therapy could prevent recurrence and achieve a longer survival time [ 13 , 18 ]. Studies have shown that arsenic trioxide exerts dose-dependent dual effects on APL cells—it induces apoptosis (programmed cell death) preferentially at relatively high concentrations (0.5 × 10 −6 to 2 × 10 −6 M) and induces partial differentiation at low concentrations (0.1 × 10 −6 to 0.5 × 10 −6 M). The rapid modulation and degradation of the PML-RARa oncoprotein by arsenic trioxide could contribute to these two effects [19] . How Do ATRA and Arsenic Work at the Molecular Level? To understand how ATRA and arsenic compounds act at the molecular level in treating APL, it is first important to understand the role of the PML-RARa fusion protein in the pathogenesis of APL. Retinoids that are crucial for normal myeloid differentiation act via RA receptors (RARs) and retinoid X receptors (RXRs). These belong to the steroid/thyroid/retinoid nuclear receptor superfamily of ligand-inducible transcription factors. Both RAR and RXR families consist of three subtypes: α, β, and γ [20] . RARα forms a heterodimer with RXR and binds to RA response element to control the expression of target genes in the presence of physiological concentrations (10 −9 –10 −8 M) of retinoids ( Figure 3A ). Figure 3 Leukemogenic Effects of PML-RARá and Mechanisms of ATRA/Arsenic Trioxide in the Treatment of APL (A) In the absence of RA, RARα/RXR heterodimers recruit the transcription corepressor (CoR), which mediates transcriptional silencing by mechanisms that include direct inhibition of the basal transcription machinery and recruitment of chromatin-modifying enzymes. Chromatin modification includes histone deacetylation, which leads to a compact chromatin structure that impairs the access of transcriptional activators. In the presence of physiological concentrations (10 −9 –10 −8 M) of RA, the transcription corepressor is released and the coactivator is recruited to the RARα/RXR heterodimer, resulting in histone acetylation (AC) and overcoming of the transcription blockage. (B) PML-RARα fusion protein binds to RARα target genes either on its own or with RXR and then recruits corepressors, leading to transcriptional repression and myeloid differentiation inhibition. PML-RARα oncoprotein sequesters the normal RXR and PML, inhibits the PML/P53 apoptotic pathway, and delocalizes PML and other proteins from the nuclear body. PML-RARα also may affect interferon (IFN) and other signal pathways. Abnormalities in protein tyrosine kinases (e.g., FLT3, c-fms) may collaborate with PML-RARα to cause APL. (C) In the presence of pharmacological doses of ATRA or arsenic trioxide, the PML-RARα fusion is degraded in ways that are dependent on caspases and proteasomes. The degradation of PML-RARα may lead to derepression of transcription suppression and restoration of PML nuclear body structure. The blockade of other signaling pathways is also released, and the anti-apoptotic effect of PML-RARα is lost. ATRA also induces cyclic AMP (cAMP), which reverses the silencing of RXR, induces the expression of RA-induced genes and cyclooxygenase 1 (Cox 1), inhibits angiogenesis, and downregulates tissue factor. Subsequently, ATRA induces terminal cell differentiation, while arsenic trioxide induces partial differentiation and/or apoptosis of APL cells. The effects of ATRA and arsenic trioxide are indicated with red and blue arrows, respectively. AF2, the ligand-dependent transcriptional activation domain contained within the C-terminal E domain of RARα; D522, aspartate at residue 522; K160, lysine at residue 160; K490, lysine at residue 490; RARE, retinoic acid response element; SUG-1, a component of proteasome 19S complex that can bind with the activated AF2 domain of RARα. More than 95% of patients with APL have the t(15;17)(q22;q21) translocation. This results in the fusion of the RARα gene on 17q21 and the promyelocytic leukemia (PML) gene on 15q22, which generates a PML-RARaα fusion transcript [ 21 , 22 ]. Variant translocations can also be detected in APL. The PML-RARα chimeric protein acts as a dominant negative mutant over wild-type RARα. The chimeric protein prevents activation of key target genes required for normal myeloid differentiation by sequestering RXR and other RARa cofactors and inhibiting normal RARα functions. The PML-RARα oncoprotein binds to RAR target genes either on its own or with RXR and then recruits histone deacetlyase complexes, which act as repressors of transcription. PML-RARa may affect transcription in other pathways including those in which the transcription factor AP1 and interferon-responsive factors are involved. PML-RARα also binds to promyelocytic leukemia zinc finger (PLZF) protein and potentially affects its functions (e.g., growth suppression and transcription repression; control of developmental programs and differentiation) [20] . In addition, PML-RARα prevents apoptosis through delocalization of PML and other proteins from the nuclear body. Finally, PML-RARα may cooperate with activated mutations in protein tyrosine kinases, such as FLT3 [23] , which confer proliferative and/or survival advantage to hematopoietic stem/progenitor cells. Whether PML-RARα affects the protein tyrosine kinase activity directly or indirectly is unclear. All these interactions of PML-RARα could be involved in the leukemogenesis of APL ( Figure 3B ). ATRA and arsenic trioxide degrade and cleave the PML-RARα oncoprotein. Although we now have a good understanding of the molecular mechanisms underlying ATRA in differentiation therapy for APL, these mechanisms were shown long after the identification of the efficacy of this drug in treating the disease. Now it is well established that pharmacological concentrations of ATRA (10 −7 –10 −6 M) exert their effects through targeting the PML-RARα oncoprotein, triggering both a change in configuration and degradation of the oncoprotein and the activation of transcription, leading to differentiation. Cleavage of the PML-RARα fusion protein by caspases at residue D522 has been shown in APL cells induced to differentiate by ATRA [24] . Further dissecting of the pathways involved in PML-RARα catabolism led to the discovery of ubiquitin/proteasome-mediated degradation of PML-RARα and RARα, which was dependent on the binding of SUG-1 in the AF2 transactivation domain of RARα with 19S proteasome [ 25 , 26 ]. In contrast to ATRA, which targets the RARα moiety of the fusion, arsenic targets the PML moiety of PML-RARα, through a still unclear mechanism, and causes PML to localize to the nuclear matrix and become sumoylated. Sumoylation at K160 is necessary for 11S proteasome recruitment and arsenic-trioxide-induced degradation, whereas sumoylation at K490 is needed for nuclear localization [ 27 , 28 ]. These results provide a striking similarity in the effect of these two otherwise unrelated agents ( Figure 3C ). The final result of treatment with ATRA and arsenic trioxide is the degradation of the PML-RARa oncoprotein, which results in restoration of normal retinoid signaling. RXR and PML sequestration is abrogated, and PML nuclear body is restored. The corepressor is released and the coactivator is recruited and bound with RARα; thus, the transcription of target genes is derepressed. ATRA also induces cyclic AMP, a differentiation enhancer that boosts transcriptional activation, reverses the silencing of the transactivating function of RXR, and restores ATRA-triggered differentiation in mutant ATRA-resistant APL cells [29] . Additionally, ATRA induces the expression of RA-induced genes [30] and cyclooxygenase 1 [31] , inhibits angiogenesis [32] , downregulates the expression of tissue factor [33] , and restores other signal pathways (e.g., the interferon pathway). Consequently, the abnormal promyelocytes differentiate and die through programmed cell death ( Figure 3C ). Combining ATRA and Arsenic: A Cure for APL? Since ATRA and arsenic trioxide degrade the PML-RARa oncoprotein via different pathways, and since studies in animal models have shown synergic effects of both drugs in prolonging survival or even eliminating the disease [ 34 , 35 ], the group at the Shanghai Institute of Hematology hypothesized that the combination of the two drugs might synergize in treating APL. To test this, 61 patients newly diagnosed with APL were randomized into three treatment groups: ATRA, arsenic trioxide, or a combination of the two drugs [36] . Although CR rates in all three groups were high (>90%), the time to achieve CR differed significantly—the time was shortest in patients treated with the combination. The disease burden (as reflected by fold change of PML-RARα transcripts at CR) decreased significantly more with combined therapy than with ATRA or arsenic trioxide monotherapy ( p < 0.01), and this difference persisted after consolidation therapy ( p < 0.05). Notably, all 20 patients in the combination group remained in CR whereas seven of 37 cases treated with monotherapy relapsed ( p < 0.05) after a follow-up of 8–30 months (median, 18 months). It seems that a combination of ATRA and arsenic trioxide for remission and maintenance treatment of APL produces better results than either of the two drugs used alone, in terms of the time required to achieve CR and the length of disease-free survival. We hope that the use of three treatments—ATRA, arsenic trioxide, and chemotherapy—will ultimately make APL a curable human acute myeloid leukemia [36] . Conclusion The story of ATRA in the treatment of APL shows that by targeting the molecules critical to the pathogenesis of certain diseases, cells can be induced to return to normal. Differentiation therapy is therefore a practical method of treating human cancer that has shown consistent effectiveness in trials. The clarification of the underlying molecular abnormalities of APL is an example of the benefits of a close collaboration between bench and bedside, and is necessary for our understanding of the mechanisms of ATRA in differentiation therapy. It is clearly important to elucidate the molecular and cellular basis of a particular cancer if we are to further develop mechanism-based target therapies. The sequencing of the human genome and ongoing functional genomic research are now accelerating the dissection of disease mechanisms and identification of therapeutic targets. This in turn may facilitate the screening of promising treatments. What we learn from developing curative treatment approaches to APL may help to conquer other types of leukemias and cancers.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545204.xml

543463

Modulation of Statin-Activated Shedding of Alzheimer APP Ectodomain by ROCK

Background Statins are widely used cholesterol-lowering drugs that act by inhibiting HMGCoA reductase, the rate-limiting enzyme in cholesterol biosynthesis. Recent evidence suggests that statin use may be associated with a decreased risk for Alzheimer disease, although the mechanisms underlying this apparent risk reduction are poorly understood. One popular hypothesis for statin action is related to the drugs' ability to activate α-secretase-type shedding of the α-secretase-cleaved soluble Alzheimer amyloid precursor protein ectodomain (sAPP α ). Statins also inhibit the isoprenoid pathway, thereby modulating the activities of the Rho family of small GTPases—Rho A, B, and C—as well as the activities of Rac and cdc42. Rho proteins, in turn, exert many of their effects via Rho-associated protein kinases (ROCKs). Several cell-surface molecules are substrates for activated α-secretase-type ectodomain shedding, and regulation of shedding typically occurs via activation of protein kinase C or extracellular-signal-regulated protein kinases, or via inactivation of protein phosphatase 1 or 2A. However, the possibility that these enzymes play a role in statin-stimulated shedding has been excluded, leading us to investigate whether the Rho/ROCK1 protein phosphorylation pathway might be involved. Methods and Findings We found that both atorvastatin and simvastatin stimulated sAPP α shedding from a neuroblastoma cell line via a subcellular mechanism apparently located upstream of endocytosis. A farnesyl transferase inhibitor also increased sAPP α shedding, as did a dominant negative form of ROCK1. Most conclusively, a constitutively active ROCK1 molecule inhibited statin-stimulated sAPP α shedding. Conclusion Together, these data suggest that statins exert their effects on shedding of sAPP α from cultured cells, at least in part, by modulation of the isoprenoid pathway and ROCK1.

Introduction Alzheimer disease is the leading cause of dementia among the elderly and is characterized by accumulation of extracellular and vascular amyloid in the brain [ 1 ]. Amyloid deposits are composed of the amyloid-β peptide (Aβ), a 4-kDa peptide released during “amyloidogenic” proteolytic processing of the Alzheimer Aβ precursor protein (APP) [ 2 ]. APP can also be cleaved by the nonamyloidogenic α-secretases, a disintegrin and metalloproteinase 10 (ADAM-10) and ADAM-17 [ 3 ], in a reaction that is believed to occur primarily on the plasma membrane [ 4 ] and is known as “ectodomain shedding.” α-Secretase-type ectodomain shedding divides the Aβ domain of APP, thereby generating α-secretase-cleaved soluble APP ectodomain (sAPP α ) [ 4 ]. This reaction can be stimulated by activation of protein kinase C (PKC) or extracellular-signal-regulated protein kinases (ERKs) [ 5 , 6 , 7 ] or by inactivation of protein phosphatase 1 or 2A [ 5 ]. Reports from retrospective analyses suggest that the statin class of cholesterol-lowering HMGCoA reductase inhibitors may lower the risk for Alzheimer disease by as much as 70% [ 8 , 9 , 10 , 11 ]. Studies in wild-type guinea pigs and in plaque-forming transgenic mice have demonstrated that chronic statin treatment can attenuate cerebral amyloidosis [ 12 , 13 ], suggesting that statins may exert their risk-reducing effects, at least in part, by modulating APP metabolism. In cell culture, lovastatin and simvastatin decrease the release of Aβ by rat hippocampal neurons [ 12 , 14 ] while activating α-secretase-type ectodomain shedding [ 15 , 16 ]. However, the molecular mechanisms by which statins modulate ectodomain shedding remain to be elucidated [ 17 , 18 ]. Statin effects on APP metabolism are, to some extent, attributable to cholesterol lowering, but statin actions on APP may also involve cholesterol-independent actions [ 19 ]. Reduction in synthesis of mevalonate leads to decreased generation of a number of isoprenoid lipid derivatives. Isoprenoids, such as farnesyl pyrophosphate and geranylgeranyl pyrophosphate, are 15- or 20-carbon lipid moieties. Through the action of farnesyl transferases and type I geranylgeranyl transferases, isoprenoids are attached to the amino acid sequence Cys-Ala-Ala-Xaa (“CAAX”) at the C-terminus of the Rho family of GTPases [ 20 ]. These posttranslational lipid modifications are essential for attachment of the GTPases to the cytosolic face of intracellular vesicles and/or to the cytosolic leaflet of the plasma membrane, thereby specifying subcellular targets for GTPase action(s). Some members of the Rho GTPase family exert their actions through modulation of protein kinase activities. One of the best characterized is Rho-associated protein kinase 1 (ROCK1, also called ROKβ). ROCK1 is a serine/threonine kinase with an apparent mass of 160 kDa that can be activated by either RhoA or RhoB [ 21 , 22 , 23 ]. Structurally, the ROCK1 N-terminus contains the protein kinase domain, while the C-terminus has both a Rho-binding domain and a pleckstrin homology domain, either of which can modulate protein–protein interactions. In the inactive state, the Rho-binding domain and the pleckstrin homology domain form an autoinhibitory loop by binding and blocking the kinase domain at the N-terminus of the molecule. Activation of ROCK1 occurs when a Rho protein binds to the Rho-binding domain, causing a conformational change that opens the kinase domain for the phosphorylation of downstream effectors [ 23 ]. Once activated, ROCK1 phosphorylates several substrates, including myosin light chain phosphatase, LIM kinases (Lin11, Isl1, and Mec3), and ezrin-radixin-moesin proteins [ 24 , 25 , 26 , 27 ]. ROCK1 has recently been implicated in modifying the site of substrate cleavage by APP γ-secretase [ 28 ], perhaps acting via ROCK1-dependent phosphorylation of a component of the γ-secretase enzyme complex. In the current study, we demonstrate that activation of sAPP α shedding from cultured cells by atorvastatin or simvastatin involves isoprenoid-mediated protein phosphorylation. Treatment of cells with a farnesyl transferase inhibitor or expression of a dominant negative (DN) ROCK1 molecule led to enhanced sAPP α shedding, supporting the notion that shedding is modulated by the isoprenoid pathway. Transfection with the cDNA for a constitutively active (CA) ROCK1 molecule led to inhibition of statin-activated sAPP α shedding. These results raise the possibility that the apparent beneficial effect of statins in the prevention of Alzheimer disease could be, at least in part, mediated by isoprenoid modulation of APP metabolism. Methods Reagents The APP C-terminal specific polyclonal antibody 369 [ 29 ] was used to detect full-length APP and its C-terminal fragments. Monoclonal antibody 6E10 against residues 1–16 of human Aβ (Signet, Dedham, Massachusetts, United States) was used to detect human holoAPP or sAPP α . Anti-ROCK1 antibody was purchased from Chemicon (Temecula, California, United States). Streptavidin-antibody HRP-conjugated C-Myc antibody 9E10, mevalonic acid, arachidonic acid, and phenylarsine oxide were purchased from Sigma (St. Louis, Missouri, United States). Atorvastatin was obtained from Pfizer (Groton, Connecticut, United States), and simvastatin was obtained from LKT Labs (St. Paul, Minnesota, United States). N2 supplement was obtained from Gibco (Carlsbad, California, United States). Sulfo-NHS-LC-Biotin was purchased from Pierce (Rockford, Illinois, United States). CA and DN Myc-tagged ROCK1 vectors were generated as previously described [ 30 , 31 ] and were generous gifts from Liqun Luo (Stanford University). Fugene 6 was purchased from Roche (Basel, Switzerland). Farnesyl transferase inhibitor 1 (FTI-1) was obtained from Biomol (Plymouth Meeting, Pennsylvania, United States). Tumor necrosis factor α (TNFα) protease inhibitor 2 and Y-27632 were purchased from Calbiochem (San Diego, California, United States). Protein concentration assay kit was purchased from Biorad (Hercules, California, United States). LIVE/DEAD Viability/Cytotoxicity Assay Kits and Amplex Red Cholesterol Assay Kits were purchased from Molecular Probes (Eugene, Oregon, United States). Culture Methods and Sample Preparation N2a mouse neuroblastoma cells stably transfected with the Swedish mutant form of APP (SweAPP N2a cells; APP695, 595–596 KM/NL) (gift from G. Thinakaran and S. Sisodia, University of Chicago, Chicago, Illinois, United States) were maintained in DMEM, 10% FBS, and 200 μg/ml G418 in the presence of penicillin and streptomycin [ 32 ]. For the 24 h prior to pharmacological treatments, the culture media were changed to N2-supplemented lipid-free medium. In some studies, cells were transfected in N2-supplemented FBS-free medium 48 h before pharmacological treatments. Transfections were carried out using the Fugene reagent, according to the manufacturer's instructions. All treatments were performed in the presence of 1 μM mevalonic acid, unless otherwise specified. Cells were lysed in 1% Triton-X/PBS buffer containing 1 X complete proteinase inhibitor cocktail (Roche), sonicated twice for 30 s, and centrifuged at 5,000 g for 5 min. Protein concentration in the supernatant was determined using the Biorad Protein Assay kit, following the manufacturer's instructions. For the detection of holoAPP and sAPP α , samples were separated in 7.5% polyacrylamide gels, transferred to nitrocellulose, and the proteins detected with either 369 (1:3,000 for holoAPP and C-terminal fragments) or 6E10 (1:1,000 for holoAPP or sAPP α ), followed by incubation of the transfers with appropriate secondary anti-rabbit or anti-mouse antibodies. For the detection of transfected ROCK1 proteins, samples were immunoprecipitated with 2 μg of anti-Myc antibody, separated in 5% polyacrylamide gels, transferred, and the proteins detected with anti-ROCK1 antibody (1:1,000 dilution). Cell-Surface Biotinylation Cells were plated in a 100-mm dish at a concentration of 5 × 10 6 cells/dish. After treatment, media were harvested, and sAPP α levels were evaluated by immunoblotting as described above. Cells were washed twice in PBS and then incubated with Sulfo-NHS-LC-Biotin for 30 min at 4 o C. Biotinylation reactions were terminated by one wash in Tris followed by two washes in PBS. Cells were lysed in 1% Triton-X/PBS buffer containing protease inhibitor cocktail as indicated above. Lysates were immunoprecipitated with 3 μl of whole 369 antibody serum and 30 μl of protein A beads. After washing twice with 1% Triton/PBS, and then twice with PBS, samples were boiled in sample buffer for 3 min, separated in a 7.5% polyacrylamide gel, and transferred to nitrocellulose. The biotinylated proteins were detected using streptavidin HRP polymer (1:10,000 dilution). Viability/Cytotoxicity Assays Cells were plated in an eight-well slide at a concentration of 1 × 10 4 cells/well. After treatments as indicated, LIVE/DEAD assays were performed following the manufacturer's instructions (Molecular Probes). Cholesterol Assays Cholesterol levels in cell lysates were measured using Amplex Red following the manufacturer's instructions (Molecular Probes). We have previously demonstrated that standard doses of either simvastatin or atorvastatin reduce cholesterol levels in N2a cells by 65%–67% [ 16 ]. Quantification and Statistical Analysis Quantification of protein bands was performed using the UVP Bioimaging System, and statistical analysis was performed on paired observations using the Student's t test. Results Atorvastatin Activates sAPP α Shedding at a Subcellular Site Upstream of Endocytosis from the Plasma Membrane We confirmed our previous observation [ 16 ] that atorvastatin produces an increase in sAPP α shedding that is dose-dependent, reaching a maximum effect at 5 μM. The increase in sAPP α shedding is accompanied by a corresponding increase in levels of the nonamyloidogenic APP α-C-terminal fragment (C83; data not shown). In order to refine our localization of the subcellular target of statin action, we evaluated the steady-state levels of cell-surface APP (csAPP) in the absence or presence of statins. After drug treatments, cells were subjected to the surface biotinylation protocol. Cells and media were harvested, and levels of sAPP α , holoAPP, and csAPP were measured. Treatment with atorvastatin increased csAPP by approximately 1.6-fold, similar to the effect of the drug on holoAPP ( Figure 1 ), while sAPP α shedding was increased by approximately 7-fold ( p < 0.05). Since csAPP levels were only slightly raised in the same statin-treated cells in which sAPP α shedding was dramatically increased, we interpret this disparity to indicate that the effector of statin-stimulated shedding is probably intrinsic to the plasma membrane. In other studies, the plasma membrane has been proposed to be, or to contain, the statin target. For example, statins have been proposed to cause co-localization of α-secretase and APP within lipid rafts [ 15 , 33 ]; statins might also induce modification of the structure and activity of a protein in the plasma membrane α-secretase complex, perhaps in an action similar to how statins bind and “lock” cell-surface integrins [ 34 ]. Figure 1 Atorvastatin Activates sAPP α Shedding Out of Proportion to Its Effect on holoAPP or csAPP (A) SweAPP N2a cells were treated with atorvastatin (Atv) for 24 h and then surface biotinylation was performed as described in Methods. Evaluation of csAPP was performed by immunoprecipitation–immunoblot after surface biotinylation, while holoAPP and sAPP α were evaluated by immunoblot as described in Methods. C, control. (B) Graphic representation of data. Y-axis shows effect of treatment (in arbitrary units) divided by effect of untreated control (in arbitrary units); n = 3 independent experiments; *, p < 0.05; **, p < 0.01; Student's t test). A pulse-chase protocol was also used to study post-transcriptional regulation of APP metabolsm by statins (data not shown). This protocol avoids any confound that might arise because of altered APP transcription. Pulse-chase studies were performed using a 10-min pulse with [ 35 S]methionine followed by various chase times from 0 to 120 min. Typical maturation and half-life of mature cellular holoAPP were observed, as was subsequent release of sAPP α [ 5 , 29 ]. In the presence of either atorvastatin or simvastatin, the time course of maturation and release perfectly paralleled that observed in the absence of either drug, except that the fractional content of cellular mature holoAPP was approximately 2-fold greater in the presence of drug (i.e., at 15 or 75 min chase, mature APP in the presence of statin was approximately 310% of the level of immature APP at t = 0 versus a control [vehicle treatment] of 150% of the level of immature APP at t = 0; also, at t = 30 min, the relative percent values for drug versus vehicle were 380% and 200%, respectively). Fold increases in released sAPP α in the same experiments were approximately 3- to 4-fold (2.0 arbitrary units versus 5.5–8.0 arbitrary units at 120 min for atorvastatin and simvastatin, respectively). Secretory maturation toxicity is one possible mechanism for elevated levels of intracellular mature holoAPP and causes retarded conversion of mature holoAPP to sAPP α . This pattern was not observed following statin treatment, excluding maturation toxicity as a mechanism underlying the altered levels of mature cellular holoAPP. Instead, the pattern that we observed raises the possibility that statins, presumably via isoprenoids (given the reversibility with mevalonate), as discussed in the next section, may alter sorting of cellular holoAPP, diverting holoproteins away from terminal degradation in the endosomal/lysosomal pathway and into the constitutive secretory pathway that generates sAPP α . However, the fold effect on reduced intracellular turnover in the endosomal/lysosomal pathway (or sorting out of the endosomal/lysosomal pathway and into the constitutive secretory/shedding pathway) is apparently insufficient to explain the fold effect on sAPP α generation (2-fold for the former, vs 3- to 4-fold for the latter), indicating a contribution from a downstream site in the processing pathway. When these results are taken together with independent work on regulated shedding of transforming growth factor α (TGFα) [ 35 , 36 ], a parsimonious explanation is that an important target for activation of ectodomain shedding is probably located at the plasma membrane or downstream of APP residence at the plasma membrane. The identification of the regulatory components of the ectodomain shedding machinery have been long-sought in other studies employing phorbol esters to stimulate shedding of sAPP α or TGFα [ 4 , 35 , 36 , 37 ]. Munc-13 has recently been implicated as a phorbol target in regulated shedding [ 38 ]. In our opinion, this molecule is rather unlikely to play a major role in shedding regulation, given the specificity of Munc-13 effects for phorbols and the generalization of the regulated shedding phenomenon to include activation by protein phosphatase inhibitors and neurotransmitters. Neither of these would be predicted to act via the phorbol-binding C1 domain of Munc-13. Sisodia and colleagues [ 39 ] demonstrated that arrest of APP endocytosis from the plasma membrane by deletion of its NPXY clathrin-coated vesicle targeting sequence [ 40 , 41 ] can dramatically stimulate sAPP α shedding, presumably by extending the half-life of co-localized α-secretase and APP on the plasma membrane. In order to exclude the possible contribution of altered endocytosis to statin-stimulated shedding, we evaluated the effect of phenylarsine oxide (PO), an inhibitor of endocytosis, on statin-stimulated shedding ( Figure 2 ). Treatment with either atorvastatin, simvastatin, or PO alone increased sAPP α shedding, as expected. Co-treatment of cells with PO plus either atorvastatin or simvastatin caused stimulation of sAPP α shedding to levels greater than the maximal levels of shedding achievable with inhibition of endocytosis using PO alone or with maximal doses of either statin alone ( p < 0.05). The additivity of statin- and PO-stimulated shedding is consistent with the hypothesis that statins act at or near the plasma membrane, prior to internalization of csAPP. Under all circumstances, stimulated sAPP α shedding was completely blocked using TNFα protease inhibitor 2, a standard α-secretase/metalloproteinase inhibitor ( Figure 2 ). We interpret this as an indication that the statin-induced α-cleavage of APP is probably mediated by one of the molecules usually associated with the phenomenon, i.e., ADAM-10 or ADAM-17 [ 16 ]. Figure 2 Simultaneous Treatment of SweAPP N2a Cells with Statins and an Inhibitor of Endocytosis (PO) Yields More sAPP α Shedding Than Does Treatment with Either Statins or PO Alone (A) SweAPP N2a cells were treated for 24 h with atorvastatin (Atv) or simvastatin (Sim) as indicated. Media were then replaced and cells were treated for an additional 20 min with atorvastatin, simvastatin, TNFα protease inhibitor, PO, or combinations, as indicated. Evaluation of sAPP α and holoAPP was performed by Western blot as described in Methods. C, control. (B) Graphic representation of data. Y-axis shows effect of treatment (in arbitrary units) divided by effect of untreated control (in arbitrary units); n = 3 independent experiments; *, p < 0.05 versus control; **, p < 0.01 versus C; #, p < 0.05 versus atorvastatin alone; ##, p < 0.05 versus simvastatin alone; Student's t test). Compounds that Modulate Isoprenoid Levels Activate sAPP α Shedding As discussed above, there is an established relationship between statins and isoprenoid-modulated protein phosphorylation. We therefore tested the effects of FTI-1 on statin-stimulated sAPP α shedding. FTI-1 increased the shedding of sAPP α , but the combination of a statin plus FTI-1 increased sAPP α shedding to levels greater than those achievable by using either compound alone ( Figure 3 ; p < 0.05). In the same experiment, levels of holoAPP were modestly increased but, again, to an extent insufficient to account for the increase in shed sAPP α ( Figure 3 ). Figure 3 Simultaneous Treatment of SweAPP N2a Cells with a Statin and FTI-1 Causes Greater sAPP α Shedding Than Either Drug Alone (A) SweAPP N2a cells were treated for 24 h with atorvastatin (Atv, 5 μM), simvastatin (Sim, 1 μM), FTI-1 (5 μM), or a combination of FTI-1 plus a statin. Levels of sAPP α (top panel) or holoAPP (bottom panel) were evaluated as described in Methods. C, control. (B) Graphic representation of data. Y-axis shows effect of treatment (in arbitrary units) divided by effect of untreated control (in arbitrary units); n = 3 independent experiments; *, p < 0.05 versus control; **, p < 0.05 versus atorvastatin alone; ***, p < 0.05 versus simvastatin alone; Student's t test). To test whether statin-activated shedding might be attributable to a metabolite downstream of HMGCoA reductase, cells were treated with 1 μM simvastatin, 5 μM FTI-1, and a series of concentrations of mevalonic acid. Since FTI-1 acts downstream of HMGCoA reductase, the stimulatory effect of FTI-1 on sAPP α shedding would not be predicted to be modified by mevalonate supplementation. Low doses (<1 μM) of mevalonic acid did not affect statin-induced sAPP α shedding, but complete inhibition of statin-activated sAPP α shedding was achieved with higher doses of mevalonic acid (100 μM). As predicted, the shedding observed following treatment with FTI-1 was not inhibited by any of the concentrations of mevalonic acid tested ( Figure 4 ). These data are consistent with a role for isoprenoids in statin control of APP metabolism in cultured cells. Figure 4 Mevalonic Acid Reverses Statin-Induced, but Not FTI-1-Induced, sAPP α Shedding SweAPP N2a cells were treated for 24 h with simvastatin (Sim, 1 μM), FTI-1 (5 μM), mevalonic acid (0–100 μM), or combinations as indicated. Levels of sAPP α were evaluated by western blot as described in Methods. This figure is representative of the results of two independent experiments. C, control. Expression of ROCK-Related Molecules Modulates sAPP α Shedding in a Bidirectional Manner Since many isoprenoid-mediated Rho effects converge on ROCKs, we next transfected N2a cells with cDNAs encoding either green fluorescent protein (GFP) (control), CA ROCK1, or DN ROCK1 ( Figure 5 ). Simvastatin caused a typical activation of sAPP α shedding from GFP-transfected cells. When CA ROCK1 was introduced, however, shedding of sAPP α from both untreated and simvastatin-treated cells was diminished ( Figure 5 ; p < 0.05 versus GFP control). Conversely, DN ROCK1 alone activated shedding of sAPP α . Cellular levels of holoAPP were not affected by transfection ( Figure 5 ; p < 0.05 versus GFP control). In studies aimed at independent confirmation of the involvement of ROCK activation in sAPP α shedding, we treated SweAPP N2a cells with arachidonic acid, an activator of ROCK. As shown in Figure 6 , arachidonic acid reduced the shedding of sAPP α without altering levels of holoAPP. Based on this series of results, we concluded that both basal and activated sAPP α shedding from cultured cells are controlled by ROCK activity. Figure 5 Structure and Expression of ROCK cDNAs, and Their Effect on Basal and Statin-Stimulated sAPP α Shedding (A) Graphic representation of the ROCK1 constructs. Myc, Myc tag; KD, kinase domain; PH domain, pleckstrin homology domain; RBD, Rho-binding domain. (B) SweAPP N2a cells were transfected with GFP, CA ROCK1, or DN ROCK1 for 48 h. Cells were lysed and levels of expressed ROCK1 protein evaluated by immunoprecipitation–immunoblot as described in Methods. (C) Model for ROCK activity modulation by Rho. (D) SweAPP N2a cells were transfected for 48 h with control (GFP), CA ROCK1, or DN ROCK1 cDNAs. At the end of this incubation, cells were treated for an additional 24 h with simvastatin (Sim, 1 μM). sAPP α and holoAPP were evaluated by immunoblot as described in Methods. (E) Graphic representation of data. Y-axis shows effect of treatment (in arbitrary units) divided by effect of untreated control (in arbitrary units); n = 3 independent experiments; *, p < 0.05 versus GFP; Student's t test). C, control. Figure 6 Arachidonic Acid Inhibits Basal sAPP α Shedding but Has No Effect on holoAPP Levels (A) SweAPP cells were treated for 24 h with arachidonic acid (5 or 50 μM, represented by AA5 and AA50, respectively). Levels of sAPP α were evaluated by immunoblot as described in Methods. C, control. (B) Graphic representation of data. Y-axis shows effect of treatment (in arbitrary units) divided by effect of untreated control (in arbitrary units); n = 6 independent experiments; *, p < 0.05 versus control; Student's t test. In some experiments, cells were treated with Y-27632 (10 nM to 50 μM), a compound that can inhibit ROCKs. Y-27632 showed no effect on basal sAPP α release and blocked statin-activated sAPP α shedding (data not shown). This result was unexpected in light of the effects of DN ROCK1. Given the internally consistent actions of DN ROCK1 and CA ROCK1, as well as the results employing either FTI-1 or arachidonate, we concluded that the Y-27632 result might be due to inhibition by Y-27632 of protein kinases other than ROCK1 [ 23 ]. The possibility was also considered that cytotoxicity of Y-27632 for the central vacuolar pathway might explain the disparity between the effects of DN ROCK1 and those of Y-27632, but neither impairment of intracellular APP maturation nor increased apoptosis as measured by LIVE/DEAD assay were apparent following Y-27632 treatment (data not shown). Ultimately, we were unable to document any explanation for the disparate results of Y-27632 and DN ROCK1. Discussion The isoprenoid pathway involves lipid modification of various members of the Rho family of small GTPases by the addition of either farnesyl or geranylgeranyl moieties [ 20 , 42 ]. Isoprenylation serves to target the GTPases to the proper organelle membrane, where their actions often relate to cytoskeletal dynamics and/or vesicle trafficking [ 20 , 42 ]. ROCKs are important downstream targets of Rho ( Figure 7 ), catalyzing the phosphorylation of effector phosphoprotein substrates [ 23 ]. The foregoing data indicate that statin-induced activation of APP shedding in cultured cells involves the Rho/ROCK pathway. More specifically, the data indicate that ROCK1 activation blocks the effects of statins on APP ectodomain shedding, while ROCK1 blockade alone can mimic the effect of statins on APP shedding. By extension, these data predict that application of statins to neurons might directly or indirectly inhibit ROCK1 activity. Evaluation of this possibility will be the subject of future investigation. Figure 7 Isoprenoid Pathway and Sites of Action of Compounds Used in This Study FPP, farnesyl pyrophosphate; GGPP, geranylgeranyl pyrophosphate. The first evidence that APP might be a substrate for ectodomain shedding was provided by Weidemann et al. [ 43 ] who identified sAPP α in the cerebrospinal fluid and blood. This aspect of APP metabolism bears resemblance to the proteolytic signal transduction pathways involved in processing pro-TGFα [ 35 , 36 ] and Notch [ 44 ]. In the case of Notch, the process is set in motion by the binding of a ligand to the Notch ectodomain, triggering its release (shedding). For APP, intracellular signal transduction appears to be more important [ 2 , 29 , 45 ]. In early studies, the existence of the shed ectodomain of APP was used to deduce the existence of the proteolytic activity designated α-secretase, which has the unusual specificity of cleaving its substrates at a proscribed distance from the extracellular leaflet of the plasma membrane [ 4 , 39 ]. Ultimately, the integral cell-surface metalloproteinases ADAM-10 and ADAM-17/TACE were found to underlie α-secretase-type ectodomain shedding [ 3 , 46 ]. Why is APP a substrate for ectodomain shedding? To answer this question requires contemplation of the physiological function of APP. APP is a type 1 integral protein that is subjected to a host of post-translational processing events, including N- and O-glycosylation, tyrosyl sulfation, phosphorylation, and proteolysis [ 39 , 43 , 47 , 48 ]. Most (60%–80%) newly synthesized APP is subjected to terminal intracellular degradation that generates no discrete fragments [ 5 ]. A smaller fraction of APP molecules (approximately 20% in PC12 cells under basal conditions [ 5 ]) undergoes ectodomain shedding catalyzed by either the α-secretase (nonamyloidogenic) or β-secretase (potentially amyloiodgenic) pathway. When PKC is activated, the stoichiometry of shed sAPP α rises from 2 mol shed per 10 mol synthesized to 4 mol shed per mole synthesized. Most of this shedding is catalyzed by the α-secretase pathway, but a trace amount (<5% [ 49 ]) is catalyzed by β-secretase/β-site APP cleaving enzyme [ 50 ]. sAPP α and sAPP β differ by the inclusion in sAPP α of the first 16 residues of Aβ. Unlike sAPP α , which is generated at the plasma membrane, most sAPP β is probably generated by cleavage within the trans -Golgi network and endocytic pathway vesicles. HoloAPP levels are likely limiting at one or more sorting steps in the late secretory pathway, since activated sAPP α shedding is apparently accompanied by diminished generation of sAPP β [ 51 ]. What is the function of shed sAPP α ? Again, from other molecules, we know that shedding can serve important cellular functions by releasing diffusible ligands from their membrane-bound precursors (e.g., TGFα and TNFα) or by terminating intercellular signaling (e.g., Notch). A popular model holds that sAPP α may function as a neurotrophic and/or neuroprotective factor, and may promote neurite outgrowth [ 52 ]. More recent evidence suggests that released APP derivatives modulate efficacy of neurotransmission at the synapse [ 53 ]. Targeted deletion of APP has not revealed a striking phenotype [ 54 ], presumably because of functional redundancy supplied by APP-like proteins [ 55 ]. Mice with double and triple null mutations in various combinations of APP, APP-like protein 1, and APP-like protein 2 are now being created, in search of evidence for a definitive function for APP. Cao and Sudhof [ 56 ] have recently discovered that the APP C-terminal fragment generated by α- or β-secretase is itself cleaved to release Aβ and an APP intracellular domain (AICD) that diffuses into the nucleus, possibly acting there as a transcription factor. The pathway leading to AICD must be initiated by ectodomain shedding: holoAPP cannot directly give rise to AICD. Therefore, one important function for α- and/or β-secretase processing of APP may be the eventual generation of AICD. Our results suggest that Rho/ROCK signaling provides modulation of basal and stimulated α-secretase activity. It will now be important to dissect pathways upstream of Rho/ROCK signaling in order to identify the intracellular and intercellular events that participate in Rho/ROCK regulation of α-secretase under physiological and pathological conditions. The potential role of cholesterol in α-secretase-mediated shedding was discovered by Bodovitz and Klein [ 57 ] who used β-cyclodextrin to lower cellular cholesterol. Kojro et al. [ 15 ] confirmed this observation, using not only β-cyclodextrin but also lovastatin to lower cellular cholesterol. These investigators proposed that elevated ADAM-10 activity and protein levels contributed to these effects. These basic observations dovetailed with emerging epidemiological evidence that administration of statins might lead to a diminished incidence of Alzheimer disease [ 8 , 9 , 10 , 11 ]. Despite this, however, the association of statins and cholesterol levels with activated α-secretase-mediated shedding of the APP ectodomain was unexpected and not readily explicable by existing knowledge regarding regulation of α-secretase activity. The best characterized regulation of α-secretase processing typically involves protein phosphorylation via PKC [ 5 , 29 ] or ERKs [ 7 ] or protein dephosphorylation by protein phosphatase 1 or 2A [ 29 ]. We recently excluded the possibility that either PKC or ERK plays a role in statin-activated shedding [ 16 ], raising the possibility that other protein phosphorylation signaling pathways might link statins and/or cholesterol to α-secretase activation. Maillet et al. [ 58 ] implicated the Rho pathway in modulation of α-secretase activity while dissecting the activated shedding process that accompanies serotonergic signal transduction. These investigators discovered that Rap1 acts through Rac to modulate α-secretase processing of APP. Soon thereafter, ROCK1 was discovered by Zhou et al. [ 28 ] to modulate a downstream processing step in APP metabolism that involves presenilin/γ-secretase-mediated proteolysis of APP C-terminal fragments C99 and C83. These investigators discovered that activation of ROCK1 may account for how nonsteroidal anti-inflammatory drugs specify the scissile bond within the APP transmembrane domain that is cleaved by presenilin/γ-secretase to generate the C-terminus of Aβ. Based on these reports, we asked whether the Rho/ROCK pathway might play a role in controlling shedding of sAPP α following statin application. CA ROCK1 and DN ROCK1 molecules yielded direct and complementary evidence that ROCK1 was indeed a candidate for modulation of statin-activated α-secretase action. Further, we were able to demonstrate that α-secretase activity could be modulated by molecules further upstream in the isoprenoid pathway (see Figure 7 ). FTI-1, an inhibitor of farnesyl transferase also known as L-744,832 [ 59 ], mimicked and potentiated statin-activated shedding, presumably by blocking transfer of isoprenoid moieties to a Rho protein by farnesyl transferase, and thereby decreasing Rho activity. However, FTI-1 treatment can also increase the level of geranylgeranylated isoforms of certain Rho proteins, e.g., the inhibitory geranylgeranylated RhoB protein [ 60 ]. In further support of a role for isoprenoids, we were able to demonstrate that supplementation of cells with mevalonate abolished statin-activated shedding (see Figure 4 ). Statins block HMGCoA reductase generation of mevalonate from 3-hydroxy-methyl-glutarate (see Figure 7 ). Therefore, the addition of mevalonate would be predicted to antagonize statin action via the isoprenoid pathway, by relieving statin-induced mevalonate deficiency. As predicted by this model, we observed that statin-activated shedding was abolished by adding mevalonate. Taken together, these results suggest the existence of a reciprocal relationship between isoprenoid-mediated Rho/ROCK signaling and sAPP α shedding, i.e., activation of ROCK1 blocks basal and stimulated shedding while ROCK1 inhibition apparently relieves a tonic negative influence exerted on shedding by ROCK1 activity. As in PKC- and ERK-activated shedding, the ROCK1 substrate effector molecule or molecules that regulate proteolysis by ADAMs remain to be identified. The cytoplasmic domains of both APP and ADAM-17 have been evaluated as candidates for important targets of protein phosphorylation during the regulated shedding process, but neither “substrate activation” nor “enzyme activation” appears to explain the phenomenon, i.e., phosphorylation of neither APP nor ADAM-17 dramatically increases the efficiency of α-secretion [ 61 , 62 ], indicating that activation is more indirect. Our data using statins and PO localize the mediator of statin-activated shedding to the plasma membrane, upstream of endocytosis, as appears to be the case for PKC-activated shedding [ 36 , 63 , 64 , 65 ]. Similar conclusions were drawn by Bosenberg and colleagues [ 36 ] who used streptolysin-porated cells and N-ethylmaleimide to demonstrate that reconstitution of activated shedding of TGFα from CHO cells does not require membrane trafficking and apparently occurs on the plasma membrane. These results suggest that a tightly membrane-associated regulatory subunit of the α-secretase complex is likely to be the key phosphoprotein that mediates α-secretase activity as a function of its state of phosphorylation by PKC and perhaps also ERK and ROCK1. The molecular identity of this phosphoprotein remains unknown. α-Secretase activation is a potential therapeutic strategy for modifying cerebral amyloidosis in Alzheimer disease [ 66 ]. This proposal is supported by recent evidence that either genetic modification of ADAM-10/α-secretase activity [ 67 ] or administration of bryostatin, a PKC activator [ 68 ], can modulate levels of brain Aβ in plaque-forming transgenic mice. α-Secretase activation may explain how statins lower the risk for Alzheimer disease [ 69 ], since atorvastatin diminishes Aβ burden in plaque-forming transgenic mice [ 13 ]. If α-secretase stimulation is to be truly viable as a human clinical intervention, it will be essential to assess the possibility that enhanced APP ectodomain shedding might incur mechanism-based toxicity (analogous to the concerns currently surrounding γ-secretase inhibitors). Along this line, extension of this work to other shed proteins will be important to determine the impact of enhanced shedding via ADAM proteinases on other substrates of those proteinases, including Notch, pro-TGFα, pro-TNFα, and CD44 [ 3 ]. Preliminary results from a pilot proof-of-concept using atorvastatin in a human clinical treatment trial are consistent with the proposed beneficial effects of this class of compounds [ 70 ]. Since atorvastatin has low blood–brain barrier permeability [ 71 ], this beneficial effect, if attributable to Aβ lowering, must be due to altered Aβ metabolism in the periphery. Reduction in levels of free Aβ in the circulation has been demonstrated to lead to diminution in brain plaque burden following active or passive immunization [ 72 ]. It is conceivable that, if statins lower circulating Aβ, this effect could secondarily cause diffusion of central nervous system interstitial Aβ down its concentration gradient and into the cerebrospinal fluid and circulation, from which it is cleared. To date, however, this mechanism is not supported by data from human clinical trials, where statin administration has shown no consistent effect on levels of circulating or cerebrospinal fluid Aβ [ 73 ]. The results reported here point to several areas for additional investigation. As described above, the key substrate or substrates linking cytoplasmic protein phosphorylation to intralumenal or cell-surface protelysis remain to be identified. Nonetheless, α-secretase activation has been validated as a viable therapeutic strategy for modulating cerebral amyloidosis [ 67 ]. Identification of the role of the Rho/ROCK pathway in regulating α-secretase provides a new avenue for its therapeutic activation, even though the potential relevance of atorvastatin-mediated ROCK1 inhibition in neurons may not explain the apparent clinical benefits of the drug. Still, if the reported disease-modifying effect of atorvastatin is confirmed in the National Institute on Aging's large, multi-center trial of simvastatin, one or more compounds of this class may be among the first disease-modifying compounds approved by the Food and Drug Administration for slowing the progression of Alzheimer disease. Supporting Information Accession Numbers The GenBank ( http://www.ncbi.nlm.nih.gov/Genbank/ ) accession numbers for the proteins discussed in this paper are ROCK1 (NP_005397) and APP695, 595–596 KM/NL (NP_958817). Patient Summary Background Large-scale studies have found a link between taking cholesterol-lowering drugs called statins and a decreased risk of developing Alzheimer disease. But it is not clear why statins might protect people from getting the disease. The brains of people who have died from Alzheimer disease show remnants of damaged cells called “tangles” as well as “amyloid plaques” in the spaces between the cells. These plaques are mostly made up of collections of a protein called amyloid-beta. It is the buildup of this protein that is thought to cause the brain damage and dementia associated with Alzheimer disease. The protein itself is formed when another, larger protein called APP (Alzheimer amyloid-beta peptide precursor protein) is broken down (or cleaved). There are two ways in which APP can be broken down. “Bad cleavage” releases the toxic amyloid-beta, whereas “good cleavage” destroys it. When researchers gave statins to animals over a long period of time, they found that statins could slow down the formation of amyloid plaques. From the animal experiments, it seemed that statins somehow caused more good cleavage to occur. Why Was This Study Done? This study examined how statins can affect APP cleavage. What Did the Researchers Do? They studied cells to see which of the players known to be involved in APP cleavage were affected by statin. What Did They Find? Statin's ability to promote “good cleavage” of APP involves a molecular pathway called the Rho/ROCK1 pathway. It seems that when ROCK1 is active, less good cleavage takes place. But in the presence of statins, ROCK1 is less active, shifting the balance toward good cleavage. Consistent with this, when the scientists blocked the Rho/ROCK1 pathway, they saw the good cleavage pattern even without statin. What Does This Mean for Patients? Inhibition of the Rho/ROCK1 pathway could explain some of the beneficial effects of statins against Alzheimer disease. And the pathway itself seems worth more research to see whether it might be a good target for new ways to prevent and treat Alzheimer disease. What Are the Limitations of the Study? Statins are likely to influence the risk for Alzheimer disease by several different pathways, and future studies will need to show how important this particular pathway is in the overall picture. Moreover, studies like this one are by necessity done in cells under carefully controlled laboratory conditions and still a long way from the development of safe and effective drugs. More Information Online Factsheet on statins from the Alzheimer's Association: http://www.alz.org/Resources/TopicIndex/statins.asp General information at the Alzheimer's Disease Education and Referral Center at the United States National Institute of Aging: http://www.alzheimers.org/index.html Homepage of Alzheimer's Disease International, an umbrella organization of Alzheimer disease associations around the world: http://www.alz.co.uk/

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC543463.xml

549589

A Person-Centred Approach to Communicating Risk

Standard approaches to communicating risk to patients do not appear to be very effective, argues Alaszewski. We need a new approach that takes patients' own perceptions into account

Doctors and other health professionals play a key role in communicating risk information. They are advisers to patients, especially when patients have to make fateful decisions that can irrevocably change their lives. There is a developing body of literature on the ways in which risk information can be effectively communicated [ 1 , 2 ]. However, much of this literature focuses on the nature of risk information and ways in which the transfer of this information can be improved. It does not fully take into account the complexity of the real world of clinical practice, nor the importance of considering patients as active partners in communication. The Rational Model of Risk Communication Much of the discussion of risk communication is grounded in the rational model of risk communication [ 3 , 4 ]. This model emphasises the role and position of experts such as doctors who have the ability to identify relevant risk knowledge. In the context of medical decision-making this is knowledge about the probable consequences of different courses of action based on scientific research. The role of the doctor is to make such knowledge available so that the patient can then use it to make an informed decision. With the rational model, when there is evidence that patients have not used risk knowledge effectively, then the response of the professional is to consider ways in which risk communication can be improved, such as by improving its presentation or mode of communication. When patients appear to be making irrational or harmful decisions, for example, continuing to smoke, choosing not to vaccinate a child against measles, mumps, and rubella, or not complying with medication, the professional's response is to work harder to convey the risks. Patients actively seek information on risks from many different sources (Illustration: Margaret Shear, Public Library of Science) But the rational model contains two key flaws. One relates to the nature of risk knowledge and the second to the nature of communication [ 5 ]. Within the rational model, risk knowledge is treated as a relatively simple and straightforward matter—in other words, there is a single uncontested source of knowledge that is relatively easy to access. In reality, risk knowledge is often a complex matter. While such knowledge may be produced by scientific research, it can and often is contested. There may be a scientific consensus, for example, that eating beef or having your child vaccinated against measles, mumps, and rubella is relatively safe, but there are often alternative scientific views, sometimes represented by high-profile media “mavericks” who emphasise the potential hazards [ 6 ]. Risk knowledge cannot actually be used directly by patients to inform their decision-making. Scientific research such as in epidemiology generates knowledge about the probability of harmful events occurring within populations . Individual patients need information on their own personal risks. Expert assessments of risk tend to focus on the knowable and measurable components of risk, that is, the extent to which future events are the same as and predictable by the knowledge of past events. Such assessments by definition exclude uncertainty—those aspects that cannot be assessed and measured. Given the speed of social and technological change, it is not clear that the past is an effective guide to the future. As such, there is an increasing awareness of the uncertainty of risk assessment, for example, in relationship to new diseases such as HIV/AIDS or new technologies such as mobile phones or genetically modified foods. The Need for a Person-Centred Approach Within the rational model of risk communication, the emphasis is on the flow of knowledge from the knowledgeable doctor to the uninformed patient. However, communication is a two-way process, and increasingly there is awareness of the active role of patients and the public [ 7 ]. Patients actively seek information, especially when they are aware that they are facing a crucial decision. While they can use traditional sources such as friends and relatives, if they have the skills and resources they can, through media such as the Internet, access highly sophisticated risk knowledge. For example, via the Cochrane Collaboration Web site ( www.cochrane.org ) they can find the latest evidence-based assessments of medical treatments and technologies, or via the Dr. Foster Web site ( www.drfoster.co.uk ) they can find the risks associated with different treatment facilities in the United Kingdom. Many patients access a variety of different sources, so they can clearly compare and evaluate the information provided by each. Patients do give particular credibility to sources that they know, which may include family and friends but also medical advisers with whom they have developed a relationship. They are particularly concerned about the trustworthiness of particular sources. While individuals can use their personal experience to evaluate the trustworthiness of personal sources, such as a particular relative or doctor, they often use contextual information to judge the trustworthiness of impersonal sources [ 8 ]. For example, information provided by a source that has an identifiable commercial interest, such as a company marketing a food product, will be considered as less trustworthy than a source without such an interest, for example, an expert committee of scientists. Patients will actively interpret risk information. If the information is timely and relevant it will tend to be accepted. Patients tend to define relevance in terms of the way they view or frame a situation, and there may be considerable differences between the ways that experts and patients view the same situation. As Zinn notes, the ways in which individuals frame and perceive risk will be highly influenced by their social situation, especially their personal biography [ 9 ]. Individuals may identify and respond to the same risks in very different ways. For example, Ziegeler has shown how the background and social context of individuals who have been diagnosed as having multiple sclerosis influence the ways in which they identify and manage their risks and opportunities [ 9 ]. Features of a Person-Centred Approach Standard approaches to risk communication, whether targeted at groups or individuals, do not appear to be very effective. For example Ruston and Clayton have shown the ways in which women disregard information and conceptually distance themselves from the risk of coronary heart disease—this applies even to those admitted to hospital with the disease [ 10 ]. Coleman has documented the failure of strategies that focus on providing information about the risks of teenage pregnancy to have any marked effect [ 11 ]. If doctors want to communicate effectively, then they need to develop a person-centred approach to risk communication, one that recognises that communication forms part of a relationship and builds upon it. Communication should be a dialogue that develops as the relationship develops, and those involved should have complementary and linked roles. Thus, the initial stage of communication could involve identifying the key issues, that is, those that cause concern for the patient. In this phase the emphasis might be on the patient talking and the doctor listening. If there is a major difference in the ways in which patient and doctor are framing the risk issues, then there might need to be an exchange or negotiation in which both parties adjust their mutual expectations and seek a mutually acceptable definition of what the problem is. If such an exchange does not take place, and if the patient's definitions are disregarded and not acknowledged, there is the danger that the patient will passively acquiesce but treat much of the information provided by the doctor as irrelevant and disregard it. If and when there is agreement, then there is the possibility of discussing the future and the likely consequences of taking different actions (risk communication in its traditional form). During this part of the exchange the emphasis might shift towards the doctor talking more and the patient listening more. There is a transfer of information, but it is a two-way process. The doctor should learn something about the patient's situation, including the risks that the patient is concerned about and the patient's beliefs about the nature of such risks. The patient should learn about the doctor's views of the nature of the risks that the patient is facing and the options for managing such risks. Underpinning the development of an effective relationship is the development of mutual trust. While trust usually takes time to develop, it is possible even during a short but positive exchange, in which mutual respect is shown, for a form of “swift” trust to be developed [ 12 ]. While I have focussed on communication in face-to-face relationship, the same issues and processes can be identified in more impersonal communication, such as the provision of risk information in health promotion campaigns. In such campaigns special mechanisms need to be created for dialogue. For example, Jones has described a project that engages young drug users in Hong Kong by helping them make videos about drug use, and has shown how such techniques can be used to evaluate and improve current health promotion adverts [ 13 ]. Conclusion There are no quick technical fixes for communicating risk information. If health professionals are serious about communicating risk information so that patients and others can make informed choices, they need to recognise that communication is a two-way process, and they need to take time to access patients' accounts and perceptions. Such investment should pay off both in an improved relationship and also in improved concordance with treatments.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549589.xml

514705

Is drug-induced toxicity a good predictor of response to neo-adjuvant chemotherapy in patients with breast cancer? -A prospective clinical study

Background Neo-adjuvant chemotherapy is an integral part of multi-modality approach in the management of locally advanced breast cancer and it is vital to predict the response in order to tailor the regime for a patient. The common final pathway in the tumor cell death is believed to be apoptosis or programmed cell death and chemotherapeutic drugs like other DNA-damaging agents act on rapidly multiplying cells including both the tumor and the normal cells by following the same common final pathway. This could account for both the toxic effects and the response. Absence or decreased apoptosis has been found to be associated with chemo resistance. The change in expression of apoptotic markers (Bcl-2 and Bax proteins) brought about by various chemotherapeutic regimens is being used to identify drug resistance in the tumor cells. A prospective clinical study was conducted to assess whether chemotherapy induced toxic effects could serve as reliable predictors of apoptosis or response to neo-adjuvant chemotherapy in patients with locally advanced breast cancer. Methods 50 cases of locally advanced breast cancer after complete routine and metastatic work up were subjected to trucut biopsy and the tissue evaluated immunohistochemically for apoptotic markers (bcl-2/bax ratio). Three cycles of Neoadjuvant Chemotherapy using FAC regime (5-fluorouracil, adriamycin, cyclophosphamide) were given at three weekly intervals and patients assessed for clinical response as well as toxicity after each cycle. Modified radical mastectomy was performed in all patients three weeks after the last cycle and the specimen were re-evaluated for any change in the bcl-2/bax ratio. The clinical response, immunohistochemical response and the drug-induced toxicity were correlated and compared. Descriptive studies were performed with SPSS version 10 and the significance of response was assessed using paired t-test. Significance of correlation between various variables was assessed using chi-square test and coefficient of correlation calculated by Pearson correlation coefficient. Results There was a statistically significant correlation observed between clinical, immunohistochemical response (bcl-2/bax ratio) and the drug-induced toxicity. Conclusion Responders also had significant toxicity while non-responders did not show significant toxicity following neoadjuvant chemotherapy. The chemotherapy-induced toxicity was observed to be a cost effective and reliable predictor of response to neo-adjuvant chemotherapy.

Background Neo-adjuvant chemotherapy (NACT) is an integral part of multi-modality approach in the management of locally advanced breast cancer (LABC). It is required both for the local control (to ensure microscopically free margins during surgery) and distant or systemic control [ 1 - 5 ]. In the past few years, considerable research has been done on the molecular aspects of breast cancer. The recognition that tumor growth rate is a product of the proliferative activity and the rate of cell death has lead to a reappraisal of traditional views of tumor response and resistance to cytotoxic Drugs [ 2 ]. Apoptosis is a closely regulated form of active cell death defined by characteristic biochemical and morphological criteria. A large number of anti-cancer agents with widely differing modes of action have been demonstrated to induce apoptosis in vitro, suggesting this as a significant final common pathway for exerting their clinical effects. Mechanisms that suppress apoptosis may be important in the development of intrinsic and acquired resistance to cytotoxic drugs [ 3 ]. It was suggested more than 20 years ago that apoptosis might account for much of the spontaneous cell loss (known from kinetic studies) to occur in many tumors and its extent often is enhanced by well-established modalities such as chemotherapy, irradiation and hormone ablation. However, during the past few years, advances in the understanding of the control of apoptosis at the molecular level has extended its potential oncologic significance far beyond the mere provision of a mechanistic explanation for tumor cell deletion. In particular, the discovery that the products of certain proto-oncogenes can regulate apoptosis has opened up exciting avenues for future research [ 4 ]. The protean effects of various neoplastic agents on synthesis of DNA, RNA, and inhibition of synthesis which may or may not lead to cell death, requires only that some critical concentrations of active drug or metabolite be present in a cell. Proliferating normal cells may therefore be subject to the same detrimental effects of chemotherapy experienced by neoplastic tissue and successful chemotherapy dictates that recuperative abilities of normal tissues are greater than those of cancer. The two tissues generally most adversely affected by antineoplastics are the hemopoetic cells of the bone marrow and the epithelium of the aero digestive tract as a result of high growth fractions and short cycling times of the cells. The ability of the cancer patients to perform normal activities and function is also recognized both as a determinant of how well the patient may respond to chemotherapy and an index of the general toxicities of the anti cancer agents. A variety of anti-cancer drugs have been shown to induce extensive apoptosis in rapidly proliferating "normal" cell population and the tumors alike. Thus enhanced apoptosis is also likely to be responsible for many of the adverse effects observed following chemotherapy [ 5 ]. Apoptosis being the common final pathway both for tumoricidal effect and also for the toxic side effects, a significant correlation should therefore exist between the effects and the toxicity produced. Toxic effects could thus serve as reliable indicators of apoptosis Apoptosis is a regulated phenomenon capable of being inhibited and activated. Indeed there is evidence that stimulation of some cells by trophic cytokines or increase in their levels of expression of Bcl-2 proto-ontogeny can greatly increase their resistance to the apoptosis-inducing effects of anticancer drugs. Thus Bcl-2 proto-ontogeny expression may be implicated in the development of resistance of tumors to therapeutic agents and may contribute to tumor growth and perhaps to ontogenesis by allowing the inappropriate survival of cells with DNA abnormalities [ 6 ]. Deregulated expression of the Bcl-2 protein represents the best-known example of a potent blocker of apoptosis. Over expression of Bcl-2 has now been shown to protect a wide variety of cell types from induction of apoptosis by many different anticancer agents. Several homologues of Bcl-2 protein have also been shown to act as inhibitors of apoptosis, including Bcl-Xl and others as apoptotic proteins such as Bax. In vitro data suggest that it is the relative ratios of anti-apoptotic and pro-apoptotic proteins that determine the likelihood of cells to undergo apoptosis in response to chemotherapeutic drugs [ 2 , 7 ] The increasing use of pre-operative chemotherapy (PCT) in breast cancer offers an in vivo test bed to further confirm the clinical relevance of these observations. The clinical response or the absence of response along with the toxic effects observed could well help predict the outcome with a particular chemotherapeutic regime and facilitate planning of an alternate regime for better response [ 5 - 8 ]. It is vital to assess the response to NACT in order to tailor the regime for a particular patient to predict the intrinsic or acquired chemo resistance. DNA-damaging agents such as chemotherapeutic drugs can induce apoptosis and increased resistance to chemotherapeutic agents, which has been found to be associated with decreased capacity to undergo apoptosis [ 5 - 9 ]. Central to this are proteins that modulate apoptosis, including bcl-2 and bax products. The change in expression of apoptotic markers brought about by various chemotherapeutic regimens is being used to identify drug resistance in the tumor cells. Various other biological markers like p-glycoprotein expression have also been used to predict the response to neoadjuvant chemotherapy [ 9 - 11 ]. The clinical response along with complete pathological response (CPR) is still considered a surrogate marker of response against which all other predictive markers are compared. The need to have a reliable and inexpensive predictor of response in a third world scenario can not be over emphasized since majority of patients present relatively late and the resources are limited and scarce [ 2 , 3 ]. Since both the response and drug related toxicity due to chemotherapeutic agents are the result of the same common final pathway of apoptosis the toxicity due to chemotherapy along with clinical response may be utilized as a cost-effective and reliable indicator (predictor) of response to neo-adjuvant chemotherapy. Against this background a prospective study was contemplated with the following aims and objectives: 1. To assess the clinical and immunohistochemical response to NACT in patients with LABC. 2. To correlate immuno-histochemical (apoptotic markers i.e. Bcl-2/Bax ratio) and clinical response to the drug induced toxicity. 3. To ascertain whether the drug induced toxicity could be utilized as a reliable indicator and predictor of response to neoadjuvant chemotherapy. Methods 50 FNAC proven cases of locally advanced breast carcinoma according to AJCC (American Joint Committee On Cancer) classification were included in the study A thorough clinical and ultrasonographic examination (USG) of all the patients including the opposite breast was performed to stage the disease accurately. A core biopsy using a tru-cut needle was performed for immuno-histochemical estimation of the apoptotic markers i.e. base-line Bcl-2/Bax ratio before initiating the chemotherapy. Routine and metastatic work up was done including complete blood examination (total blood count, platelet count), chest radiograph, ECG (Echocardiography when ECG had a positive finding), liver function tests, Bone Scan, USG abdomen, KFT (Kidney function tests). Patients were subjected to three cycles of FAC regime (cyclophosphamide 600 mg/m 2 , adriamycin -50 mg/m2, 5-fluorouracil-600 mg/m2) at an interval of three weeks. Before each cycle the patient was clinically and sonologically examined for the breast tumor size, axillary lymph node status & appearance of systemic metastasis. All patients were given the same antitoxicity treatment according to a standardized unit protocol including adequate hydration and inject able antiemetics before initiating the chemotherapy. The patients were observed for three main toxicities of the FAC regime i.e. vomiting, alopecia and leucopenia. Vomiting with minimum of four episodes on day one and two after chemotherapy was taken as significant. Total alopecia was considered significant. Leucopenia was assessed in terms of the WHO grading of hematological toxicity. All patients were subjected to Patey's modified radical mastectomy three weeks after the last cycle and the specimen were again subjected to immuno-histochemistry to evaluate for any change in the Bcl-2/Bax ratio and for the histological tumor size, margins. Objective clinical response was defined as >50% reduction in the tumor size after completion of three cycles of NACT, as assessed clinically, sonologically and histologically. Immunohistochemical response was taken as decrease in the Bcl-2/Bax ratio. Any increase or no change in this ratio was considered as no response. Immuno-histochemical methods Biopsy specimen was preserved in buffered formalin solution. Five-micron sections were prepared from paraffin embedded blocks on poly-l-lysine coated glass slides. Sections were deparaffinized in xylene for 15 min. and hydrated in alcohol for 15 minutes. Further, incubation was done in 0.3% hydrogen peroxide in methanol solution for 45 min. The slides were washed with citrate buffer and kept in a water bath at 90–95°C for 45 min. for antigen retrieval. Sections were washed with Tris buffer saline (TBS) solution and incubated with blocking antibodies (DAKO monoclonal mouse antihuman Bcl-2 oncoprotein for Bcl-2 expression and polyclonal rabbit antihuman for Bax expression) at 37°C. Sections were washed with TBS solution. Incubation with avidin-biotin complex (ABC) was done at 37°C for one hour and washed with TBS solution. 3,3 Diaminobenzidine tetra hydrochloride solution applied for 3–5 min. Counter-staining with haemotoxylin solution done for 3–5 min. Sections were washed with distilled water, air dried and mounted using DPX mountant. For Bax, positive controls were taken as germinal centers of the lymphoid follicles and normal breast tissue and negative control was taken as the test slide without primary antibody. For Bcl-2, positive controls were the mantle zone of lymphoid follicles and the negative controls were the test slides without primary antibody. The pattern of positive staining for bcl-2 and bax was cytoplasmic, granular. The primary antibodies for bcl-2 and bax were procured from DAKO. Bax-Rabbit Anti-Human code no. A 3533. Bcl-2 Monoclonal Mouse Anti-Human code no. M 0887. Dilution for both was 1: 40. The results were interpreted on the basis of two criteria: (1) Percentage of cells showing immune bodies; <5%: score = 0, 5–25%: score = 1, 25–75%: score = 2, >75%: score = 3 (2) Intensity of staining; mild: score = 1, moderate: score = 2, intense: score = 3. " Since there was a strong correlation between the intensity of staining and percentage of cells showing immune bodies, the percentage of cells showing immune antibodies alone was considered for calculating the bcl-2/bax ratio ". Statistics Descriptive studies were performed with SPSS version 10. The significance of response assessed using paired t-test. Significance of correlation between various variables assessed using chi-square test and coefficient of correlation was calculated by Pearson correlation coefficient. Results 50 cases of locally advanced breast cancer were included in this study. Mean age of the patients was 45.5 years (range: 28 to 71 years) and 53.3% patients were pre-menopausal. Size of the tumor was measured clinically as well as by ultrasound and the patients were subdivided into four groups: <5 cm(0%), 5–8 cm(56.6%), 8–10 cm(26.6%), >10 cm(16.6%). According to the axillary lymph node status the patients were divided into three groups: N0 (0%), N1 (33%), N2 (67%). Objective clinical response was defined as more than 50% reduction in the tumor size after three cycles of neoadjuvant chemotherapy. Immuno-histochemical response was defined as decrease in the Bcl-2/Bax ratio. Clinical response including the reduction in the tumor size and axillary lymph node status was observed in 70% of patients and was found to be statistically significant (p < 0.0001). There were no patients in the No group and 29.4%of the N1 patients were down staged to N0 while70.6% remained N1. In patients with N2 disease 7.7% were down staged to N0 status while 46.2% were downstaged to N1 status and 46.2% did not show any response. Immuno-histochemical response was observed in 60% and was also found to be statistically significant (p = 0.008). Correlation between immuno-histochemical and clinical response was also found to be statistically significant (p < 0.0001) [Table 1 ]. Acute vomiting was observed in 63.3% patients. 81% clinical responders had vomiting (p = 0.002) and 78% immunohistochemical responders also had vomiting which was statistically significant (p = 0.04). Alopecia was observed in 86% clinical responders (p = 0.000) and 94% immuno-histochemical responders (p = 0.000), which was also significant. Leucopoenia was observed in only 14% and 17% of clinical and immuno-histochemical responders respectively and was found to be an insignificant predictor of response in the present study. When multiple toxicities were correlated with the clinical and immuno-histochemical response, 46.7% of patients had both acute vomiting and alopecia. 67% clinical responders (p = 0.001) had both vomiting and alopecia.72% immunohistochemical responders (p = 0.001) had both vomiting and alopecia. A significant positive correlation was observed between the presence of vomiting (r = +0.558), alopecia(r = +0.802) and response to neoadjuvant chemotherapy. A significant negative correlation was observed between the absence of side effects and poor response to neoadjuvant chemotherapy (Table 2 ). Discussion Carcinoma of the breast is the leading cause of cancer in women all over the world and the second most common malignancy in India after carcinoma of the uterine cervix [ 1 ]. No other common epithelial cancer has been so carefully studied and so extensively characterized biologically [ 1 , 2 ]. In developing countries like India rate of locally advanced breast cancer at first diagnosis is estimated to be as high as 25%–30%[ 2 , 5 ]. The treatment of locally advanced breast carcinoma (LABC) has also evolved from primarily local modalities to treatment regimens that combine both systemic and local therapy. The realization that patients with LABC are likely to have undetectable micro metastases at diagnosis has lead to systemic treatment assuming major focus of the multi-modality approach as the studies have confirmed that surgery alone is an inadequate treatment in the management of these patients. Even aggressive surgical techniques have been observed to have a higher incidence of local recurrence in these patients [ 10 , 11 ]. Most importantly surgery does not change the pattern of distant failure in patients who probably have micrometastatic disease at the time of diagnosis [ 10 - 13 ]. Multi-modality therapy that included surgery, radiation therapy, chemotherapy, hormonal therapy has had the greatest impact on survival in patients with LABC [ 10 - 13 ]. Neoadjuvant chemotherapy (NACT) A new approach in the form of neoadjuvant chemotherapy was first reported in the 1970s and was initially utilized to convert unresectable tumors to smaller tumors making them more amenable to local control with either surgery or radiotherapy. An added advantage of this approach was the ability to assess patient's response to treatment both clinically after a defined number of courses of chemotherapy and pathologically after surgical resection. Perez and colleagues reported their results of a pilot study by the South-Eastern Cancer Study Group in 1979 that the primary tumor showed partial regression (>50%)in 65% of patients after two courses of FAC [ 16 ]. NACT has also shown benefits in the operable breast cancers by increasing the chances of breast conservation by up to 90% in some trials [ 10 - 13 ]. The other important advantage of NACT is that it represents an in vivo chemo sensitivity test for assessment of tumor response from which prognostic information can be obtained. It provides an early treatment of the micrometastatic disease, counteracting the increased growth rate possibly determined by the shrinkage of the tumor. The down staging converts an inoperable case amenable to curative resection [ 10 - 13 ]. Apoptosis Introduced by Kerr et al (1972) to describe characteristic morphological changes seen during programmed cell death [ 3 ]. It is defined as a closely regulated form of active cell death defined by characteristic biochemical and morphological criteria [ 3 , 14 , 15 ]. A wide range of anticancer drugs with widely differing modes of action have been demonstrated to induce apoptosis in vitro, suggesting this as a significant common final pathway through which they exert their clinical effect. Further more the mechanisms that suppress apoptosis may be important in the development of acquired resistance to cytotoxic drugs. Apoptosis or programmed cell death plays an important role in the regulation of tissue development, differentiation and homeostasis. It is therefore possible that deregulation of apoptosis contributes to the pathogenesis of cancer [ 3 , 13 - 15 ]. Apoptosis can be differentiated biochemically and morphologically from necrosis by the following criteria [ 16 ]: (1) Chromophin condensation (2) Membrane blebbing (3) Appearance of apoptotic bodies (4) Fragmentation of genomic DNA Certain biochemical and genetic events have been identified that are associated with multiple cell types including mammary epithelium. These include the DNA fragmentation via end nuclease activation and cleavage of intracellular proteins, expression of bcl-2 family members, tumor suppressor gene p-53 directed events, proto-oncogene activation and activation of transmembrane receptor signaling pathways such as tumour necrosis factor [ 4 , 17 - 22 ]. Although little is known about the mechanisms, which regulate apoptosis in epithelial cells, it is conceivable that defects in apoptosis related genes are involved in the pathogenesis of human cancers. The hypothesis is supported by the fact that the tumor suppressor gene product p-53, which is frequently mutated or deleted in breast cancer, is involved in regulating apoptosis [ 23 ]. The heterogeneous nature of breast cancer has resulted in overwhelming interest in search for prognostic markers to identify patients who might benefit most from the therapeutic modalities available. Assessment of apoptosis and individual components of apoptotic pathway are therefore relevant in determining prognosis in a particular patient [ 24 ]. DNA damaging agents such as ionizing radiations and chemotherapeutic drugs also induce apoptosis. Sakakura et al have shown an association between increased resistance to chemotherapeutic agents and decreased capacity to undergo apoptosis [ 25 ]. Central to this response are proteins that modulate apoptosis, including bcl-2 and bax gene products. Bcl-2 is anti-apoptotic in function, whereas bax is proapoptotic and it is the interaction between the two that determines the likelihood of a tumor to undergo cytotoxic drug mediated regression. Therefore any increase in bcl-2 or decrease in bax will push the balance towards chemo resistance and an increase in bax or decrease in bcl-2 will result in increased apoptosis [ 26 - 30 ]. It was observed in a study conducted by Kymionis et al [ 15 ] that increase in the ratio of anti apoptotic protein bcl-2 to pro-apoptotic protein i.e. bax results in markedly enhanced resistance of tumor cell lines to the cytotoxic effects of essentially all currently available chemotherapeutic drugs. In the present study the clinical response in terms of reduction of tumor size and immuno-histochemical response in terms of change of bcl-2/bax ratio correlated significantly with the drug induced toxicity following NACT. Toxicity related to chemotherapeutic agents [ 19 - 23 ] The time course of various toxic manipulations depends on the drug, its dose and frequency of administration, intrinsic characteristics of the tissue of interest and any local circumstances (e.g. radiation therapy, infection, trauma). There are few general rules however like mucosal toxicities of pain, erythema, ulceration etc. occur 3–10 days after the administration of most offending drugs. Bone marrow effects can be manifestated a few days later averaging 7–14 days. The recovery of normal functioning tissues in both cases is well under way 4–5 days after the zenith of the toxic effects. Alopecia may involve the scalp or the whole body can present within 2 weeks of the drug dose or be a progressive cumulative event. Other cumulative toxicities seen only after administration of a certain quantity of drug over some length of time include cardiomyopathy from anthracyclins. Cyclophosphamide An alkylating agent belonging to the nitrogen mustard subgroup. It is inactive as such produces few acute effects and is not locally damaging. It is transformed in to active metabolites like aldophosphamide and phospharimide mustard in the liver, which then produces the wide variety of antitumour action. The prominent side effects are alopecia and cystitis, which are caused by another metabolite acrolein. 5-Fluorouracil It is a pyrimidine antagonist and is converted in the body to the corresponding nucleotide, 5-fluro-2-deoxy-uridine monophosphate, which inhibits thymidine synthetase and blocks the conversion of deoxyuridilic acid to deoxythymidilic acid. Selective failure of DNA synthesis occurs due to non-availability of thymidylate. Flourouracil may itself get incorporated in to nucleic acids and this may contribute to its toxicity. Even resting cells are affected, though multiplying cells are more susceptible like the cells in the gastrointestinal tract (GIT) and bone marrow. Doxorubicin It is an anti tumor antibiotic and is capable of causing breaks in the DNA strands by activating topoisomerase II and generating quinone type free radicals. They have mutagenic potential. Maximum action is exerted at S phase, but toxicity is usually exerted in G2 phase. Cardiotoxicity is a unique side effect. Rapidly multiplying cells are more susceptible therefore it also acts on cells of GIT, bone marrow in addition to the tumor cells. Leucopenia has not been observed to be a frequently encountered chemotherapy induced toxicity using commonly used regimen in most of the studies [ 19 - 24 ]. This was observed in the present study also. Conclusions The rapidly proliferating normal and tumor cells are more susceptible to the action of chemotherapeutic agents, which could explain the significant correlation observed between the effects and the toxicity in the present study. There was a strong correlation observed between the immunohistochemical response (bcl-2/bax ratio), clinical response and drug toxicity. This indirectly indicates a correlation between chemotherapy induced apoptosis and the toxicity and therefore like apoptotic markers, chemotherapy induced toxic effects along with objective clinical response could serve as reliable and cost-effective indicators or predictors of response to NACT in patients with LABC. While many biological markers are in use and many are under trial to tailor the chemotherapy for a particular patient, most of these markers including apoptotic markers or p-glycoprotein etc. are not very frequently available and are expensive for a third world cancer set up. Thus the chemotherapy-induced toxicity along with clinical response may be utilized as a cost effective and reliable predictor of response to NACT in patients with LABC. This would also serve as an intermediate end point in determining drug sensitivity for adjuvant treatment, especially when adjuvant therapy is planned with the same regimen as induction chemotherapy. This can also help in planning an alternative regime in non-responders. Competing interests None declared. Authors' contribution • CM , the principal and the corresponding author was the Supervisor and the Chief surgeon who performed and standardized surgery on the patients and designed the study. • VS participated in the designing of the study, performed the statistical analysis and was the first surgical assistant and Senior Postgraduate in charge of the cases in the study. • JP , Postgraduate surgical resident was the second assistant in charge of the cases and participated in the sequence alignment. • AL, Resident surgery participated in the data processing and statistical analysis. • SS was in charge of the molecular genetic studies at the Tumor biology lab ICMR. • AB participated in the genetic studies and data processing. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514705.xml

555759

Silhouette scores for assessment of SNP genotype clusters

Background High-throughput genotyping of single nucleotide polymorphisms (SNPs) generates large amounts of data. In many SNP genotyping assays, the genotype assignment is based on scatter plots of signals corresponding to the two SNP alleles. In a robust assay the three clusters that define the genotypes are well separated and the distances between the data points within a cluster are short. "Silhouettes" is a graphical aid for interpretation and validation of data clusters that provides a measure of how well a data point was classified when it was assigned to a cluster. Thus "Silhouettes" can potentially be used as a quality measure for SNP genotyping results and for objective comparison of the performance of SNP assays at different circumstances. Results We created a program (ClusterA) for calculating "Silhouette scores", and applied it to assess the quality of SNP genotype clusters obtained by single nucleotide primer extension ("minisequencing") in the Tag-microarray format. A Silhouette score condenses the quality of the genotype assignment for each SNP assay into a single numeric value, which ranges from 1.0, when the genotype assignment is unequivocal, down to -1.0, when the genotype assignment has been arbitrary. In the present study we applied Silhouette scores to compare the performance of four DNA polymerases in our minisequencing system by analyzing 26 SNPs in both DNA polarities in 16 DNA samples. We found Silhouettes to provide a relevant measure for the quality of SNP assays at different reaction conditions, illustrated by the four DNA polymerases here. According to our result, the genotypes can be unequivocally assigned without manual inspection when the Silhouette score for a SNP assay is > 0.65. All four DNA polymerases performed satisfactorily in our Tag-array minisequencing system. Conclusion "Silhouette scores" for assessing the quality of SNP genotyping clusters is convenient for evaluating the quality of SNP genotype assignment, and provides an objective, numeric measure for comparing the performance of SNP assays. The program we created for calculating Silhouette scores is freely available, and can be used for quality assessment of the results from all genotyping systems, where the genotypes are assigned by cluster analysis using scatter plots.

Background High-throughput single nucleotide polymorphism (SNP) genotyping assays generate large amounts of data, which usually is presented as scatter plots of signals corresponding to the two SNP alleles. A robust SNP genotyping assay is characterized by large distances between the three clusters that define the genotypes and small distances between the data points within each cluster. Numeric quality measures for the scatter plots would allow objective and automatic assessment of the success of a SNP assay. "Silhouettes" were introduced in 1987 as a general graphical aid for interpretation and validation of cluster analysis [ 1 ]. In a Silhouettes calculation, the distance from each data point in a cluster to all other data points within the same cluster and to all data points in the closest cluster are determined. Thus Silhouettes provides a measure of how well a data point was classified when it was assigned to a cluster by according to both the tightness of the clusters and the separation between them. This feature renders Silhouettes potentially well suited for assessing cluster quality in SNP genotyping methods. In high-throughput SNP genotyping, Silhouettes could be used for assessing the quality of automatic genotype assignment by alerting the operator if the quality of the genotype clusters fall below a certain limit. During assay development and optimization, Silhouettes could be used to compare the performance of a genotyping assay at different reaction conditions. It could also be applied for comparing the robustness of different SNP genotyping technologies. In this study we created a program (ClusterA) to calculate numeric Silhouettes for assessing the quality of genotype clusters obtained in SNP genotyping assays. We show the utility of Silhouettes and the program by applying it to our "in-house" developed four-color fluorescence minisequencing system for SNP genotyping in a microarray format [ 2 ]. Single nucleotide primer extension ("minisequencing") is the reaction principle underlying several of the commonly used systems for genotyping single nucleotide polymorphisms (SNPs) [ 3 - 8 ]. In minisequencing a DNA polymerase is employed to specifically extend a detection primer designed to anneal directly adjacent to the SNP position in the complementary DNA strand with a single labelled nucleotide analogue. The DNA polymerase is the most important factor that determines the efficiency and specificity of the primer extension reaction, irrespectively of the assay format. We used Silhouettes to compare the performance of three new commercially available DNA polymerases to the ThermoSequenase DNA polymerase, which is routinely used in minisequencing assays in many laboratories, including our own. We found Silhouettes to provide a relevant measure, in addition to signal-to-noise ratios and genotyping success, for selecting the most favourable enzyme for our assay. Results and Discussion We created a program, denoted ClusterA, for calculating numeric "Silhouettes" for clustered data, such as for example the three clusters of signal ratios commonly obtained in SNP genotyping assays. Figure 1 illustrates the Silhouette calculation for one data point in a typical scatter plot obtained in a SNP genotyping assays. A Silhouette close to 1.0 is obtained when the average distance from a data point to the other data points within its own cluster is smaller than the average distances to all data points in the closest cluster. A Silhouette close to zero indicates that the data-point could equally well have been assigned to the neighbouring cluster. A negative Silhouette is obtained when the cluster assignment has been arbitrary, and the data point is actually closer to the neighbouring cluster than to the other data points within its own cluster [ 1 ]. The mean value from the Silhouette calculations for all data points in each cluster yield an "average Silhouette width" for the cluster. Figure 1 Principle for Silhouette scores. Principle for quality assessment of genotyping clusters using Silhouette scores, illustrated for one data point (i). The SNP genotypes have been assigned based on cluster formation in scatter plots with the signal intensity fraction on the x-axis and the logarithm of the signals from both alleles on the y-axis. For each data point (i) in the scatter plot, the Silhouette s(i) is calculated by the formula in the figure, where a(i) is the average distance from i to all data points in the same genotype cluster (green lines), and b(i) is the average distances from i to all data points in the cluster closest to the data point, either b 1 (i) (blue lines) or b 2 (i) (red lines) [1]. Max and min in the formula denote the largest or smallest of the measures in the brackets. The "average silhouette width" is calculated by calculating the mean of all s(i) for each genotype cluster and the "Silhouette score" for the whole scatter plot (SNP assay) is obtained by taking the mean of the average silhouette width for all clusters. Here, we applied ClusterA to calculate "Silhouettes" for comparing the quality of the genotype clusters obtained in our "in-house" Tag-array minisequencing system. For each scatter plot, the mean of the average silhouette widths for the three genotype clusters were used to define a "Silhouette score" for each SNP assay. Thus the Silhouette score condenses the cluster quality for each SNP assay into a single measure that ranges from 1.0 to -1.0. When calculating the Silhouette score, the distance between data points can be measured either in one dimension, for example on the x-axis, or in two dimension using vectors, as illustrated in Figure 1 . In our Tag-array minisequencing system we used distances measured only in one dimension, along the x-axis, where the signal fraction (Signal Allele2 / (Signal Allele1 +Signal Allele2 ) is plotted, since this is the major determinant for genotype assignment in our system. The logarithm of the sum of the signals from both alleles (Signal Allele1 +Signal Allele2 ) plotted on the y-axis is only used to set the cut-off values for failed genotype calls. Figure 2 shows nine examples of SNP genotype clusters that yielded different Silhouette scores. Negative controls and assays with signals below signal cut-off level are not shown in Figure 2 since they are not included in the Silhouette score calculations. Figure 2 Examples of Silhouette scores. Examples of genotype clusters from nine SNP assays, each with the results from 16 samples genotyped in duplicate using Tag-array minisequencing with the calculated Silhouette scores shown in the right hand upper corner of each panel. The blue circles represent homozygotes for allele 2, the red triangles are heterozygotes and the green squares are homozygotes for allele 1. The SNPs are denoted by their dbSNP identification number, and the DNA polarities analyzed are indicated by "cod" or "nc". The examples in panels E, F and G of Figure 2 illustrate how different clustering patterns can yield similar Silhouette scores. Based on the results from the scatter plots used to assign genotypes in this study, our recommendation is to accept the results from SNP assays with Silhouette scores >0.65 and to fail the whole assays if the Silhouette scores is <0.25. Individual genotype calls for assays where the Silhouette score falls between 0.25–0.65 may be accepted or failed after visual inspection. Excluding some of the outliers will then increase the Silhouette score. Our recommendations is in line with Liu et al., who have included silhouette calculations in the complex algorithm used to interpret the data from the Affymetrix 10K HuSNP hybridization microarray [ 9 ]. Here we exemplify the use of Silhouette scores by comparing the performance of the TERMIPol, Therminator, KlenThermase and ThermoSequenase DNA polymerases in the Tag-array minisequencing system [ 2 ]. Twenty-six SNPs were analyzed in both polarities in 16 DNA samples in two independent experiments. As our Tag-array genotyping system utilizes an "array of arrays" format [ 10 ] with 80 subarrays on each microscope slide, we were able to test all four enzymes in all samples on the same slide at exactly the same conditions, to facilitate a fair comparison between the enzymes. Figure 3 shows the distributions of Silhouette scores in these SNP assays. For all enzymes, 75% of the scatter plots (indicated by light blue rectangles in Figure 3 ) yielded silhouette scores above or close to our recommended limit of 0.65. Results from a total of 79 scatter plots/SNP assays are included in Figure 3 and Table 1 . If a SNP assay failed for all samples with one enzyme, the results from this assay were excluded from the whole enzyme comparison. It should also be noted that a non-stringent genotype calling strategy was applied to reveal possible differences between the enzymes both in clustering properties and genotyping results. This is the reason for the very low Silhouette scores for some SNP assays, which normally would be considered as failed. Using 0.65 as cut-off, 70–76% of the SNP assays would have been successful in this study. Figure 3 Distribution of Silhouette scores from minisequencing assays using four DNA polymerases. The Silhouette score is given on the y-axis. Each black diamond represents the Silhouette score for one SNP assay. The light blue rectangular boxes indicate those 75% of the scatter plots that yielded the highest silhouette scores for each enzyme. Quartiles are indicated by the black horizontal lines. Table 1 Silhouette scores, signal to noise ratios and genotyping performance for four DNA polymerases in Tag-array minisequencing 1 Silhouette score 2 S/N 3 Genotype calls 4 Average Median Highest Average Highest Correct Errors n % n % n % n % TERMIPol 0.72 0.78 20 25.3 4.3 11 13.9 2337 98.9 18 0.8 Therminator 0.69 0.79 15 19.0 3.6 7 8.9 2323 98.3 32 1.4 KlenThermase 0.74 0.79 22 27.8 8.0 21 26.6 2346 99.3 10 0.4 ThermoSequenase 0.71 0.82 22 27.8 8.9 40 50.6 2324 98.3 34 1.4 1 Duplicate experiments, each with duplicate SNP assays in both DNA polarities, were performed and the results are composite values from both experiments. 2 The Silhouette scores were calculated as described in Figure 1. The average and the median score for all SNPs are given for each enzyme together with the number of SNP assays (n) and frequency (%) where an enzyme yielded the highest Silhouette score. 3 Signal to noise ratios (S/N) were calculated from each spot by dividing the fluorescence intensity values from the fluorescently labelled ddNTP/ddNTPs corresponding to a true genotype (signal) by the fluorescent intensity value from the other ddNTPs (noise). The average S/N ratios are given together with the number of SNP assays (n) and frequency (%) where an enzyme yielded the highest S/N. 4 Number of genotype calls (n) and call rate (%). The genotype obtained from the majority of the assays was considered to be the correct one. The percentages of the samples not accounted for in the table failed to give genotypes. In the comparison between the enzymes, KlenThermase displayed the highest average Silhouette score, ThermoSequenase had the highest median Silhouette score and also obtained the highest Silhouette score most frequently (Table 1 ). In addition to the Silhouettes scores, that represent a measure of the robustness of a SNP assay, the signal to noise ratios (S/N) and the genotyping success was assessed (Table 1 ). All four enzymes performed satisfactorily in our minisequencing assay taking into account the non-stringent genotyping criteria used. However, performance varied between the evaluated features with high error rates for Therminator and ThermoSequenase. KlenThermase showed the best results over all and, also taking into account the cost, would be the enzyme of choice based on the results from this study. Conclusion We conclude that "Silhouette scores" for assessing the cluster quality is well suited for comparing the performance of SNP assays. Here we used a one-dimensional calculation of the Silhouette scores, by measuring the distances between the data-points along the x-axis only. A two-dimensional Silhouette calculation using vectors should be applied when genotypes are assigned by scatter plots with the fluorescence signals corresponding to the two alleles on the y- and x-axis. Both options are available in the ClusterA program that also calculates mean, variance and F-statistic for the input data set. The program is freely available through our website . We believe that the ClusterA program for calculating Silhouette scores created in the present study is a useful and general tool for any genotyping system, where the genotypes are called by cluster analysis with the aid of scatter plots. Methods DNA samples Genomic DNA was extracted from blood samples from 16 volunteer blood donors using the Wizard genomic DNA purification kit (Promega, Madison, WI). Genotyping procedure Twenty-six SNPs, selected to be located in unique PCR amplicons, were included in the test panel. For information on the single nucleotide polymorphisms and oligonucleotides used, see the Additional file 1 : SNPinformation.pdf. PCR primers were designed and combined in multiplex PCR reactions. Minisequencing primers with 20 bp 5'-Tag sequences were designed for both DNA polarities. The experimental details of the genotyping procedure have been described in detail previously [ 11 ]. In short it included the following steps: The regions containing the sequence variations were amplified in six optimized multiplex PCRs. For each sample the PCR products were pooled and divided into four aliquots, one for each enzyme. The remaining dNTPs and primers from the PCR reaction mixture were removed by treatment with Exonuclease I and shrimp alkaline phosphatase. The cyclic minisequencing reactions were performed in solution as described below, and the extended minisequencing primers were hybridized to microarrays carrying immobilized covalently coupled oligonucleotides (cTags) complementary to the Tag-sequences of the minisequencing primers. The cTags had been immobilized to CodeLinkTM Activated Slides (Amersham Biosciences, Uppsala, Sweden) via their 3'-end NH 2 -groups to form 80 subarrays per slide, each with 60 cTags as duplicate spots. Finally the microarray slides were scanned, and the fluorescent signals were measured. Minisequencing reaction Cyclic minisequencing reactions were performed in solution with 10 nM of each of the 52 tagged minisequencing primers using 0.1 μM ddATP-Texas Red, ddCTP-Tamra and ddGTP-R110 and 0.15 μM ddUTP-Cy5 (Perkin-Elmer Life Sciences, Boston, MA), and 0.064 U/μl of one of the four DNA polymerases in 15μl of 0.02% Triton-X, 4.1 mM MgCl2 and 33.6 mM Tris-HCl pH 9.5. The cyclic extension reactions were performed on a Thermal Cycler PTC-225 (MJ Research, Watertown, MA) with an initial 96°C for 3 min followed by 55 cycles of 95°C and 55°C for 20 s each. The DNA polymerases were; TERMIPol (Solis BioDyne, Tartu, Estonia), Therminator (New England BioLabs Inc., Beverly, MA, USA), KlenThermase (Gene Craft, Lüdinghausen, Germany), or ThermoSequenase (Amersham Biosciences, Uppsala, Sweden). A custom made reaction rack holding the arrayed slides with a silicon grid to give 80 separate reaction chambers was used during capture of the minisequencing reaction products on the Tag-arrays. Data analysis and genotype assignment The fluorescence signals were measured from the microarray slides using a ScanArray Express ® instrument (Perkin-Elmer Life Sciences, Boston, MA). The excitation lasers were: Blue Argon 488 nm for R110; Green HeNe 543.8 nm for Tamra; Yellow HeNe 594 nm for Texas Red and Red HeNe 632.8 nm for Cy5. The fluorescence signal intensities were determined using the QuantArray ® analysis 3.1 software (Perkin-Elmer Life Sciences, Boston, MA). The QuantArray file was exported to the SNPSnapper v4.0 software ) for genotype assignment. Raw data as fluorescence signals and signal ratios are provided as supplementary material, see Additional file 2 : Rawdata.txt. Genotypes were assigned based on scatter plots with the logarithm of the sum of both fluorescence signals (Signal Allele1 +Signal Allele2 ) plotted on the y-axis, and the fluorescence signal fraction, obtained by dividing the fluorescence signals from one allele by the sum of the fluorescence signal from both SNP alleles (Signal Allele2 / (Signal Allele1 +Signal Allele2 ), on the x-axis [ 11 ]. The result file with the assigned genotypes and the corresponding signal ratios were exported as a text file and used to calculate Silhouettes scores using the ClusterA program. ClusterA is implemented in Microsoft Visual Basic 6.0, and can be run on PCs with the Microsoft Windows operating system. The ClusterA program also provides the mean, variance and F-statistic for the input data. Authors' contributions LL planned the experiments, guided the laboratory work and performed the analysis of results, interpreted the data and drafted the manuscript. AA carried out the laboratory work and part of the data analysis and provided input to the manuscript. MJ programmed the ClusterA program and took part in the interpretation of Silhouettes. ACS initiated the study, supervised it, and coordinated the manuscript writing process. All authors have read and approved the final manuscript. Supplementary Material Additional File 1 Lists the dbSNP identification numbers and the sequences of the PCR and minisequencing primers. Click here for file Additional File 2 Includes the raw fluorescence signals and the fluorescence signal intensity ratios for the two experiments as a tab delimited text file. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555759.xml

515369

Genetic Response to Climatic Change: Insights from Ancient DNA and Phylochronology

Understanding how climatic change impacts biological diversity is critical to conservation. Yet despite demonstrated effects of climatic perturbation on geographic ranges and population persistence, surprisingly little is known of the genetic response of species. Even less is known over ecologically long time scales pertinent to understanding the interplay between microevolution and environmental change. Here, we present a study of population variation by directly tracking genetic change and population size in two geographically widespread mammal species (Microtus montanus and Thomomys talpoides) during late-Holocene climatic change. We use ancient DNA to compare two independent estimates of population size (ecological and genetic) and corroborate our results with gene diversity and serial coalescent simulations. Our data and analyses indicate that, with population size decreasing at times of climatic change, some species will exhibit declining gene diversity as expected from simple population genetic models, whereas others will not. While our results could be consistent with selection, independent lines of evidence implicate differences in gene flow, which depends on the life history strategy of species.

Introduction Phylogeography has advanced our understanding of the spatial distribution of genetic diversity within and between species ( Avise 2000 ). However, empirical evidence of temporal change in genetic diversity in a single locality over time has not yet been placed in a population genetic or phylogeographic framework over ecologically long periods of time. In this paper we attempt to determine variation in genetic diversity experienced by populations of two mammalian species in situ and to place that diversity in the context of a changing environment through time. We view this approach as “phylochronology,” or the study of populations in space and time using phylogenetic and population genetic methods. Similar studies have not used such a long temporal record ( Pergams et al. 2003 ), have not considered gene flow ( Lambert et al. 2002 ; Pergams et al. 2003 ), or have used a spatially averaged sample as a proxy for a single locality ( Leonard et al. 2000 ). Our study takes advantage of a continuous, well-sampled mammalian fossil sequence spanning the last 3,000 years (Lamar Cave, Yellowstone National Park, Wyoming, United States). Lamar Cave has an extraordinarily complete representation of the local species in the vicinity, with over 10,000 identified mammalian specimens representing over 80% of the mammal species in the local habitat ( Hadly 1999 ). Late-Holocene climatic change, including the Medieval Warm Period (1,150 to 650 years before present [ybp]) and Little Ice Age (650 to 50 ybp) ( Soon and Baliunas 2003 ), affected the local abundances of common small mammals in a manner consistent with their habitat preferences ( Hadly 1996 ). We focus on two mesic habitat specialists, Microtus montanus (montane vole) and Thomomys talpoides (northern pocket gopher), species that presently are widespread in mountain habitats of western North America. Due to their preferences for wetter habitats, both responded demographically by increasing in relative abundance during wetter climates and declining during warmer climates. M. montanus showed an increase in abundance relative to other common rodents during periods of wet, cool climate in Yellowstone ( Hadly 1996 ). A 40% decline in M. montanus abundance occurred from 2,525 ybp to about 470 ybp during the Medieval Warm Period. Because T. talpoides also demonstrates a preference for mesic montane conditions, shifts in their relative abundance mimic the response seen in Microtus, decreasing by 50% between 2,525 and 470 ybp. In addition, T. talpoides showed a significant reduction in body size during this time ( Hadly 1997 ). These data highlight the influence of climatic change on the population dynamics and phenotypic response of these species, especially during warming events. Although the population responses of T. talpoides and M. montanus are similar, the ways in which they respond to climatic change at the genetic level are predicted to diverge because of differences in dispersal ability and population substructure. Advancement of ancient DNA (aDNA) techniques allows us to investigate directly the impacts of these environmental perturbations on neutral genetic diversity concurrent with these species population responses. We obtained ancient and modern mitochondrial DNA sequences from M. montanus and used previously published data for T. talpoides. Although these two species are broadly similar in body size ( M. montanus, 50–100 g; T. talpoides, 75–150 g) and are principally herbivorous, they differ in their natural history. T. talpoides is characterized by low population densities (1–62 gophers/hectare [ha]), a fossorial mode of life, maximum dispersal distances of a few hundred meters, and fiercely territorial behavior ( Verts and Carraway 1999 ). Populations of T. talpoides from Lamar Cave exhibit very little genetic variation through time but considerable genetic differences between present localities ( Hadly et al. 1998 ). This spatiotemporal pattern suggests that late-Holocene gene flow did not influence modern genetic variation of T. talpoides within localities over relatively short time scales (hundreds to thousands of years) despite the absence of obvious migration barriers. M. montanus achieves higher average population densities (60–186 voles/ha) ( Sullivan et al. 2003 ) than T. talpoides. In addition, genetic studies of closely related species and other arvicoline species (Microtus pennsylvanicus, Microtus longicaudus, and Microtus agrestis) have found little evidence for population subdivision over the scale of hundreds of kilometers ( Plante et al. 1989 ; Conroy and Cook 2000 ; Bjørnstadt and Grenfell 2001 ; Jaarola and Searle 2002 ), as expected from the ability and proclivity of voles to disperse hundreds to thousands of meters, resulting in migration between populations on generational time scales ( Jenkins 1948 ; Lidicker 1985 ). Such demography, however, implies that historical gene flow may be difficult to detect in M. montanus if only genetic data from the modern animals are used. This is contrary to the pattern expected in T. talpoides because this species has high genetic differentiation between extant populations, and past movement between such populations would be relatively easy to ascertain by the historic presence of unique, divergent haplotypes ( Hadly et al. 1998 ). The primary advantage of a phylochronologic approach as opposed to a single time slice for understanding mammalian response is the ability to reveal changes in genetic variation through time. This is in contrast to modern genetic studies that seek to reconstruct demographic history based on inferences from past climate or geologic records (e.g., Storz and Beaumont 2002 ; Lessa et al. 2003 ). Our study separates demographic and genetic response explicitly, allowing us to understand the microevolutionary forces responsible for the differences in species response over a time scale relevant to evolution within species. This approach is particularly powerful when coupled with environmental data so that perturbations may be linked to organismal response. In order to reveal these microevolutionary forces from our serial ancient data, it was necessary to explore the influence of sampling from the fossil record and to determine how variation in stochastic evolutionary forces (gene flow, drift, and mutation rate) might influence the record of gene diversity over time. Thus, we combined four methods of estimating population size, determining statistical significance, and assessing gene diversity through time. (1) We derived independent ecological estimates of population size through time from abundances of M. montanus and T. talpoides fossil specimens and modern population densities. (2) We calculated gene diversity over this 3,000-year period to determine the impact of environmental perturbations on genetic effective population size and gene diversity in M. montanus and T. talpoides. Unlike previous aDNA work ( Consuegra et al. 2002 ; Hofreiter et al. 2002 ; Lambert et al. 2002 ; Orlando et al. 2002 ; Paxinos et al. 2002 ), we used mitochondrial DNA sequences for samples taken through time from a single locality. (3) While there have been advances in the use of nuclear markers for ancient genetic analyses, we also confine our analyses to more easily derived mitochondrial DNA (mtDNA) data, thus limiting our analyses to a single locus, usually seen as a neutral marker within mammalian species ( Moritz et al. 1987 ). Our approach also constrains us to the fossil sample sizes from this locality, which are extremely large for ancient DNA studies, but limited relative to population genetic studies. Thus, we assessed the statistical power of a single locus for our empirical data using a neutral population model. (4) We used a neutral population model and serial coalescent simulations to determine whether our observed genetic data reflect our ecological estimates of population size and to evaluate statistical significance in changes of gene diversity through time. Despite similar population-level responses to climatic change of the late Holocene, we expected differences in gene diversity change for the two species. For T. talpoides, we predicted that changes in genetic variation through time would be dominated by drift, as suggested by the modern life history characteristics of small effective population size, low dispersal, and high amounts of population substructure. Therefore, as the ecological effective population size of T. talpoides declined with warmer climates, we expected genetic variation to decline. For M. montanus, we predicted that changes in genetic variation through time may be influenced more by migration, as suggested by large effective population sizes, high rates of dispersal, and low amounts of population substructure. As a result, past declines in ecological estimates of population size of M. montanus would not necessarily have resulted in a decrease in genetic variation. Results Fossil Abundance Our assessment of population response to climatic change ( Figure 1 ) depends on reconstruction of population size. Fossil relative abundances give a hint of the census size through time while genetic data (gene diversity, Figure 1 ) should yield independent assessments of the effective population size. The relationship between these measures varies, although most studies suggest that the estimate of effective size derived from ecological data is higher than that derived from genetic data (N e_ecol >> N e_gen ) ( Frankham 1996 ; Kalinowski and Waples 2002 ). However, we can convert census size estimates at any point in time into effective size estimates and vice versa. This allows us to compare explicitly ecological and genetic measures of population size. Figure 2 shows N e_ecol estimates based on low-, high-, and moderate-density estimates for T. talpoides ( Figure 2 A) and M. montanus ( Figure 2 B). For M. montanus they range from 218,652 to 436,981 for low-density estimates and from 677,825 to 1,354,650 for high-density estimates. For T. talpoides low-density estimates range from 2,219 to 5,015 and high-density estimates range from 4,586 to 10,488 individuals in the 7-km radius around Lamar Cave. Figure 1 Proportional Population Size and Gene Diversity of M. montanus and T. talpoides (A) Proportional population size (relative abundance) of M. montanus and T. talpoides ( n = 8,589 fossils) by years before present. (B) Gene diversity (H) of M. montanus and T. talpoides by years before present; 95% confidence intervals are shown. Squares indicate M. montanus; triangles indicate T. talpoides. Figure 2 Estimates of N e_gen and N e_ecol (A) T. talpoides and (B) M. montanus through time. Circles and dashed lines show N e_ecol estimates based on low-, high-, and moderate-density estimates. N e_gen estimates ([A] triangles and [B] rectangles) are based on θ S estimates from Arlequin. Standard errors for N e_gen are represented. Genetic Data: M. montanus The genetic evidence we have assembled from M. montanus suggests that the sequences we obtained for this study are target mtDNA. Of the 312 bp we sequenced for M. montanus, 96.5% of all the mutations were third-position codon changes, with first- and second-position mutations accounting for 3.5% and 0%, respectively. These ratios of variation are concordant with expectations for within-species variation and small overall sequence divergences ( Yang and Yoder 1999 ). Nucleotide base composition is similar to that of other Microtus species ( Conroy and Cook 2000 ; Jaarola and Searle 2002 ), with an excess of adenine (31.2%) and a deficit of guanine (15.7%) (χ 2 ; α = 0.71). Most of the mutations are synonymous (97.7%); the transition-to-transversion ratio of the entire data set was 4.1 to 1, which is consistent with expectations for mammalian cytochrome b and evolution in other Microtus species ( Conroy and Cook 2000 ; Jaarola and Searle 2002 ). Fossil and modern transition-to-transversion ratios are similar (3.1 and 4.2, respectively). All M. montanus sequences are reciprocally monophyletic (including M. pennsylvanicus as outgroup taxon) and translated successfully. Together with the frequency distribution of our pairwise differences, the prevalence of silent and third-position codon changes, and the standard of obtaining both forward and reverse fragments of overlapping sequence regions, these data permit us to conclude that the genetic diversity we have sampled represents authentic mitochondrial population variation and is unlikely to be from nuclear copies or pseudogenes. A total of 282 experiments included 47 fossil extractions and 1,644 PCRs. Eighty-eight percent of our aDNA specimens yielded readable sequence data, with no relationship found between success rate and age of the specimen ( R 2 = 0.004, not significant). All but one (out of 121) of the extraction controls were negative. When sequenced, this extraction blank BLASTed similar to Montanus townsendii, a taxon we had never worked on in the facility; this sequence has not since been amplified in the lab, and that extraction was not used further. Out of 87 successfully amplified samples and one sequence obtained from GenBank (AF119280), we identified 17 haplotypes within four haplogroups (A–D) of M. montanus ( Figure 3 A). The distribution of haplotypes within haplogroups suggests that our groups are defined appropriately. Each haplogroup was defined by at least 3% sequence divergence (≥10 bp) from other haplogroups in the 312-bp cytochrome b fragment. The majority of individuals (98.8%) fall within haplogroups A and D, with 84% of the samples within one substitution of the locally ancestral haplotype A ( Figure 3 A). Figure 3 Haplotype Networks for M. montanus and T. talpoides Haplotype networks ( Clement et al. 2000 ) for (A) M. montanus and (B) T. talpoides from Lamar Cave fossils and from modern specimens collected within a 400-km radius of Lamar Cave. Haplogroups for both species are indicated as A–D. Each haplogroup within a species is defined by at least 3% sequence divergence within the cytochrome b fragment. M. montanus haplogroup B is taken from GenBank. Haplogroup C is a sample from outside our 400-km radius (NK5897, Mono County, California; Museum of Southwestern Biology #53376). Light shading shows modern samples; dark shading shows fossil samples; bars indicate substitutions; cytochrome b sequence positions are indicated by number above base designation. Numbers within parentheses indicate sample sizes for each haplotype. The maximum uncorrected sequence divergence for our complete spatial and temporal data set was 4.5%, demonstrated between haplogroups A and B. Given that the highest average rodent divergence rate for cytochrome b is 6% to 10% per million years ( Irwin et al. 1991 ), these haplogroups have been evolving separately for at least 450,000 years. A similar age (422,000 years) is found when using a rate of 2.3% per million years for third-position transversions ( Conroy and Cook 1999 ). The maximum uncorrected sequence divergence of M. montanus from throughout the Lamar Cave temporal sequence was 4.2%. The maximum sequence divergence of 19 modern individuals from populations of this species within Yellowstone National Park and surroundings was 3.8%. Genetic Data: T. talpoides Protocols for T. talpoides are found in Hadly et al. (1998) . A haplotype network of this species shows three haplogroups and a total of eight haplotypes from 76 specimens ( Figure 3 B). For T. talpoides, 98.0% of all the mutations were third-position codon changes, with first- and second-position mutations accounting for 0% and 2.0%, respectively. Nucleotide base composition shows an excess of thymine (34.7%) and a deficit of guanine (11.4%). All mutations were synonymous; the transition-to-transversion ratio of the entire data set was 4.0 to 1. Gene Diversity through Time Estimates of gene diversity, nucleotide diversity, and number of segregating sites differed between M. montanus and T. talpoides ( Table 1 ). The estimates for M. montanus were higher than those for T. talpoides, as predicted by life history traits including higher ecological effective population size and higher dispersal between populations. Table 1 Summary Statistics, Sample Sizes, Sequence Length, and Number of Haplotypes for the Ancient DNA Samples for T. talpoides and M. montanus from Lamar Cave, Wyoming Gene diversity, number of segregating sites, and nucleotide diversity for each time interval were calculated with Arlequin v. 2.000 ( Jaarola and Tegelström 1996 ; Schneider et al. 2000 ) Our raw data on these species show similar relative abundance patterns but disparate trends in gene diversity (see Figure 1 ). These patterns cannot be linked directly to relative abundance because gene diversity estimates depend on true population size as well as sampling. We investigate how both of these parameters impact the observed trend in the following sections. Comparing Ecological and Genetic Estimates of Effective Size Ecological estimates of population size for T. talpoides (N e-tt_ecol ) exhibit the same trend as the genetic estimates (N e-tt_gen ), namely a population size decline of >50% after 1,500 ybp (see Figure 2 A). For T. talpoides, N e-tt_gen is consistently higher than N e-tt_ecol . When compared to those for M. montanus (see Figure 2 B), the estimates of total ecological and genetic effective population size are much lower. Additionally, unlike with M. montanus, the ecological and genetic effective sizes follow similar trends through the entire time period sampled, indicating that T. talpoides is acting as a closed population. For M. montanus, the estimates of effective size derived from the ecological data (N e-mm_ecol ) are lower than those derived from genetic data (N e-mm_gen ) for all time points (see Figure 2 B). While the estimates are not expected to be identical, comparison of their trends is instructive. Both genetic and ecological estimates follow similar trends between 2,525 and 845 ybp, after which the two estimates follow opposite trajectories. While the ecological size decreases by 50%, the genetic estimates show an initial decline of 30%, followed by an increase in population size equivalent to the pre–1,438-ybp level. This demonstrates that although the population is not recovering ecologically, it does recover genetically from population decline between 1,438 and 845 ybp. Effects of Sampling For M. montanus, our observed data were within the 95% confidence intervals for both sets of simulations ( n sample and n = 100; mutation rate = 4% per million years; moderate density values used to calculate abundance) for all time points except for 2,525 ybp, where observed gene diversity was significantly lower than could have been calculated given our sample size ( Figure 4 ). Since the observed gene diversity is lower than expected, we repeated simulations for five additional combinations of mutation rate and abundance (low mutation rate, high abundance; low mutation rate, moderate abundance; low mutation rate, low abundance; moderate mutation rate, low abundance; high mutation rate, low abundance). Results revealed that the observed gene diversity at 2,525 ybp was within the lower fifth percentile of the predicted distribution for two of the five combinations (when both mutation rate and abundance were low and for low mutation rate, moderate abundance). The overlap between the 95% confidence intervals for both sets of simulations suggests that sampling bias does not significantly impact the observed patterns of gene diversity except at 1,438 ybp ( n = 4), suggesting that we do not have sufficient power to detect processes at this time period. Figure 4 Expected and Observed Gene Diversity of M. montanus Boxes represent the 95th, 50th, and fifth percentiles for expected gene diversity of M. montanus given estimates of N e-mm_ecol at 2,525, 1,438, 845, 470, and 166 ybp and the associated sample sizes ( n = 7, 7, 18, 4, and 6) based on the Ewens sampling distribution (assumed mutation rate = 4% per million years per bp for a 312-bp fragment). Bars represent the 95th and fifth percentiles for a sample size of 100 at the same points in time. Diamonds represent observed gene diversity from empirical genetic data. The empirical data for each time unit fall within the expected ranges of gene diversity, except those for 2,525 ybp, which are much too low for the seven samples to detect, suggesting that observed gene diversities are not limited by sample size. Although the gene diversity for T. talpoides is not different given expectations from a closed population, we attempted to determine the statistical limitations of these data. Investigation of the effects of sampling for T. talpoides revealed that given the smaller number of base pairs (64 bp), we do not have enough statistical power to reject the null hypothesis. For every sampling time point, the Ewens distribution predicted that only one haplotype would be present in the genetic samples (unpublished data). As a result, the predicted gene diversity was zero for all time points. Because the observed gene diversity for T. talpoides was higher than predicted we also simulated five combinations of mutation rate and abundance, which could result in a higher predicted diversity (high mutation rate, high abundance; high mutation rate, moderate abundance; high mutation rate, low abundance; moderate mutation rate, high abundance; low mutation rate, high abundance) for 166 ybp ( n = 34). Results for all five combinations predicted presence of a single haplotype. Since we do not have adequate statistical power given the genetic data for T. talpoides, we did not conduct significance tests for this species. However, the observed values of gene diversity are not unexpected from dynamics within a closed population. Significance of Changes in Gene Diversity in M. montanus For eight of the nine combinations of mutation rate and effective size, we could reject the null hypothesis (closed population, no selection, changes in abundance inferred through fossil abundance) based on the observed change in M. montanus gene diversity throughout the entire time series (2,525 to 166 ybp) ( Table 2 ). The expected distribution of change in gene diversity given the null hypothesis and based on moderate M. montanus densities and moderate mutation rate is shown in Figure 5 ( Table 2 shows all combinations), along with the observed change. However, given the observed change in gene diversity specifically between 2,525 and 845 ybp, we were able to reject the null hypothesis for all nine combinations of mutation rate and effective size. These results suggest that M. montanus was not acting as a closed population during this period of time. Because the serial coalescent model presented here does not discriminate between selection and migration, either of these processes could have caused the observed change in gene diversity. Figure 5 Distribution of Change in Gene Diversity for M. montanus between 2,525 and 166 ybp, Based on Serial Coalescent Simulations Sampling is modeled at two points in time. N e-mm_ecol estimates from Figure 2 are used to specify demographic history. Eight of the nine combinations of mutation rate and density allow us to reject the null hypothesis for a closed population ( Table 2 ). This figure illustrates simulation results for moderate density and moderate mutation rate (4% per million years per bp). The probability of the observed change (shown by dashed arrow) is significant ( p = 0.015). Table 2 The Average (over 1,000 Simulations) Expected Change in Gene Diversity for Nine Combinations of Density and Mutation Rate for M. montanus between 2,525 and 166 ybp Significance values for observed change are in parentheses. All combinations other than moderate mutation rate and low density allow rejection of the null hypothesis of a closed population Our results may depend on our assumptions of equilibrium population size prior to 2,525 ybp. We investigated sensitivity to this assumption by modeling a population bottleneck in M. montanus prior to 2,525 ybp. We modeled population reduction to 104,577 (0.75 × N e-mm_ecol2525ybp ), 69,718 (0.5 × N e-mm_ecol2525ybp ), and 34,859 (0.25 × N e-mm_ecol2525ybp ) prior to 2,525 ybp. For a 75% bottleneck prior to 2,525 ybp, we were no longer able to reject the null hypothesis of a closed population. These results indicate that an extreme bottleneck where population size was reduced to 75% or more might result in the observed change in gene diversity. Our simulations reveal that unless an extreme bottleneck happened prior to 2,525 ybp, we can be confident that the observed change in gene diversity is not likely to be from events that occurred immediately prior to our historic data, and thus is due to migration or selection. Coalescent simulations used to investigate the significance of the observed gene diversity value for the modern samples demonstrated that the null hypothesis of past population size change could be rejected at only two of the nine mutation rate and density combinations. These results reveal that given data from only the modern samples, it was not possible to reject the null hypothesis of past population size change in M. montanus. Historic genetic data allow us to discriminate between population processes over millennia much better than do modern data alone. Other Evidence for Migration Independent lines of genetic and demographic evidence also point to the influence of gene flow in M. montanus populations. Estimates of β -diversity (used here to measure haplotypic turnover) from haplotypic data reveal that turnover was highest between 2,525 and 845 ybp ( β 2525–845ybp = 3; β 845–166ybp = 1.5; and β 2525–166ybp = 1.5). Closer examination of the haplotypic distributions demonstrates that five novel haplotypes appeared by 845 ybp, three of which are ≥3.2% different from the most common haplogroup (A) ( Figure 3 A) at 2,525 ybp, further implicating gene flow between 2,525 and 845 ybp. Discussion Our results demonstrate different genetic responses by two species of small mammals to changes in population size driven by climatic change. Fossil abundance data reveal population decline for both T. talpoides and M. montanus between 1,438 and 470 ybp, a period spanning the Medieval Warm Period ( Hadly 1996 ). For T. talpoides, the genetic response is directly related to changes in population size: Decrease in population size results in lowered gene diversity. M. montanus demonstrates the opposite relationship: A decrease in population size (between 1,438 and 166 ybp) results in an increase in gene diversity. We attempted to statistically validate our results by the use of serial coalescent simulations to demonstrate that the change in gene diversity of M. montanus between 2,525 and 845 ybp is significantly different from that expected based on the decrease in ecological estimates of population size. Taken together, these results indicate a departure from conditions of equilibrium (closed population without selection) for M. montanus. Our results have the following possible explanations: (1) the sampling area for fossils changed, (2) the local population size expanded, (3) selection occurred, and/or (4) gene flow occurred. Selection versus Gene Flow Results from all three of our analyses suggest that gene flow could be responsible for the patterns in gene diversity observed in our empirical data. Additionally, recent results of experimental studies of density dispersal dynamics in the root vole, Microtus oeconomus, indicate that migration occurs most frequently in and between low-density patches ( Andreassen and Ims 2001 ). These results indicate that density and dispersal in voles may be inversely related, a finding that is consistent with our results. An alternative explanation is that selection is governing the observed gene diversity patterns. While cytochrome b may not be under intense selection ( Irwin et al. 1991 ), it is linked to other portions of the mitochondrial genome that may be selectively advantageous in particular environments. Using cytochrome b as a marker for the accumulations of adaptations elsewhere on the genome may yield information about the effects of selection on local populations through time. Further exploration is necessary to investigate and identify the presence of locally adapted mtDNA and the rates of evolutionary change necessary to produce the variation in gene diversity we have observed (e.g., Pergams et al. 2003 ). Conclusions Here we demonstrate, using a phylochronologic approach, that it is possible to distinguish the dynamic processes that govern gene diversity over relatively short time scales (hundreds to thousands of years). We have documented environmental change, population response, genetic diversity change, and the correlations between the three. Without serial data, we would capture just a single record of these historic processes: modern genetic diversity. Although it is possible to hypothesize about historic events using modern data, phylochronology affords a unique look into the past and the potential ability to separate cause from effect. In particular, we show that M. montanus has a history recording responses both within populations (fluctuations in population size, possible selection) and between populations (gene flow). Discrepancy between the ecological and genetic estimates of population size and significant changes in haplotypic diversity prior to the Medieval Warm Period implicate increased gene flow into the Lamar Cave M. montanus population. Additionally, the observed haplotypic turnover in the Yellowstone population during this period suggests that as abundance of M. montanus declined through the last 845 years, relatively more individuals carried newly introduced haplotypes. Our results indicate that the presently observed widespread genetic variation across the geographic range in this species arose not because gene flow was equivalent through all populations through time, but because during particular time periods, certain local populations (and/or genotypes) declined while others expanded. Our data show that even with a prolonged ecological population size decline, the genetic diversity of M. montanus was maintained. In contrast, gene flow has not played a significant role in the recent genetic history of T. talpoides. This species, instead, responded more as a closed population over this time. The disparate nature of population response to climatic change of these two species is likely due to differences in demographic dispersal patterns between their populations. Such differences in species demography have resulted in differential genetic response to climatic change, even when ecological response is similar. Thus, genetic response to environmental change can be viewed as “individualistic,” similar to unique adjustments of species ranges ( Root et al. 2003 ). Life history traits such as dispersal ability contribute to the overall gene diversity of species in both space and time. If life history has such a large impact for common species, such differences will be particularly important in understanding how entire communities are affected by global change. Ultimately, knowledge from such analyses will lead to distinct, and perhaps predictable, patterns of species persistence through climatic changes, insights that will prove invaluable to future conservation of biodiversity. Materials and Methods Fossil locality Lamar Cave contains well-stratified, thoroughly radiocarbon-dated deposits, which display high fidelity to the local mammalian community ( Hadly 1996 , 1999 ). The most common animals from Lamar Cave are also the most common in the sagebrush grassland ecosystem in which Lamar Cave is located. Relative abundances are based on the entire data set of 10,597 specimens (except for those in Figure 1 A, which uses the five most common small mammals [ n = 8,589]) and are concordant with expectations of taxonomic diversity in montane mammal communities of western North America ( Hadly and Maurer 2001 ). The cumulative number of bones in Lamar Cave is correlated with time, demonstrating a constant “rain” of bones from the past to the present environment ( R 2 = 0.86; unpublished data). Age assignment The historical Microtus samples encompass 15 of 16 radiocarbon-dated stratigraphic levels from Lamar Cave ( Hadly 1996 ), with a maximum radiocarbon age of 2,860 ± 70 ybp (CAMS-20356). Data from the stratigraphic levels were pooled into five discrete intervals for Lamar Cave representing the last 3,000 years ( Hadly 1996 ). Interval boundaries were based on the stratigraphic pattern of deposition observed during excavation as well as a detailed radiocarbon chronology. Thus each interval represents a biologically significant packet of specimens. Age for each interval was assigned as the midpoint of the span of the calibrated radiocarbon ages ( Hadly 1996 ). Ecological estimates of population size Absolute population sizes in both ancient and modern communities are difficult to estimate. However, relative abundance changes of the small mammals are consistent with the climatically caused changes in habitats and the habitat preferences documented by modern trapping data proximate to the fossil site ( Hadly 1999 , Figure 1 A). By calculating the area preferred by M. montanus and T. talpoides using the geographical information system, we were able to standardize the relationship between taphonomy and population size in these species. This was possible because the collection radius of the fossils from Lamar Cave has been documented to be less than 7 km by using strontium isotopes ( Porder et al. 2003 ). We then estimated ecological effective population size from relative abundance through time, current population density, and the current area of preferred habitat in the collection radius. The current area of M. montanus habitat within the 7-km radius totals 1,992 ha; for T. talpoides it totals 971 ha. High, moderate, and low densities for T. talpoides ( Verts and Carraway 1999 ) and M. montanus ( Sullivan et al. 2003 ) were used in association with the corresponding areas of occupied habitat to estimate current census size. Because the rate of accumulation of bones through time is constant, the current census size was indexed against the percentage of Microtus bones from the uppermost level of Lamar Cave and used to calculate historic census sizes through time. In order to compare ecological estimates of effective population size to genetic estimates, we estimated mitochondrial effective population size assuming that N e /N census = 0.5 ( Storz et al. 2002 ; estimates from small mammals), and given that mitochondrial effective size is N e /4. Our estimates of N e were based entirely on relative fossil abundance and current ecological data and assume a mitochondrial effective size; thus we labelled them N e-tt_ecol (T. talpoides) and N e-mm_ecol (M. montanus). Genetic data from M. montanus From the fossil material, we used the upper first molar (from only one side of the jaw per level) to avoid sampling the same individual multiple times. Some Microtus species are cryptic with respect to these teeth. Genetic diagnosis indicated that these 78 fossil samples are derived from multiple arvicoline species, with a total of 47 M. montanus specimens (see Table 1 ). Fossil samples ( Microtus molariform teeth) ranged from 3.5 to 15.3 mg (average, 8.8 mg). We used two previously described extraction methods on the fossil teeth ( Hadly et al. 2003 ). We obtained a 312-bp fragment of mitochondrial cytochrome b from ancient samples ( n = 47) using the following primers: forward primers (5′–3′), CLETH 37 TAY AAY ATA ATY GAA ACH TGA A (5′ end of cyt b 319 equals Mus musculus 14458), CLETH 37L AYG GMT CTT AYA ACA TAA TCG AAA CAT G (cyt b 311, M. musculus 14450), MMONT 1 CAG TAA TTA CAA AYC TWC TAT CA (cyt b 452, M. musculus 14591), and MMONT 3 AGT GAA TCT GAG GGG GCT TCT CAG TAG A (cyt b 485, M. musculus 14621); reverse primers (5′–3′), ARVIC 08 CAG ATY CAY TCY ACT AGT GTT G (cyt b 473, M. musculus 14612), ARVIC 08L CTC AGA TTC ACT CTA CTA GTG TTG TG (cyt b 471, M. musculus 14610), MMONT 4 TTR TTT GAT CCT GTT TCG TGT AGG AAT A (cyt b 595, M. musculus 14631), and MMONT 2L TTG ACT GTG TAG TAA GGG TGA AAT GGG A (cyt b 653, M. musculus 14792). Attempts were made to amplify the region in two overlapping fragments using CLETH37/ARVIC 08, CLETH37L/ARVIC 08L, and MMONT 1/MMONT 2L. However, the low rate of success for MMONT 1/MMONT 2L (40%) in the fossil samples necessitated the breaking of the second fragment into two overlapping fragments using MMONT 1/MMONT 4 and MMONT 3/MMONT 2L. Ancient DNA samples were run on an ABI PRISM 310 Genetic Analyzer in the post-PCR lab, and modern DNA samples were run on an ABI PRISM 377 Sequencer in a separate sequencing facility (Protein and Nucleotide Facility, Beckman Center, Palo Alto, California, United States). Fragments were sequenced in both directions, primer regions were overlapped, and sequences with any ambiguous sites were rerun until completely resolved in order to provide additional corroboration and eliminate ambiguity. We adhered to strict extraction and amplification protocols ( Hadly et al. 2003 ). The protocol further included (1) independent sequence corroboration of two samples (J. Mountain lab, Anthropological Sciences, Stanford University), (2) processing of modern samples using personnel and reagents in another lab (D. Petrov lab, Biological Sciences, Stanford University) all physically separate from the aDNA facility, (3) monitoring contamination with several extraction and PCR controls, (4) primer design specific to arvicoline species, and (5) no prior or concurrent history of working with murid species in any of the DNA facilities involved. Modern (spatial) genetic sampling was obtained from a variety of sources including modern skins, modern liver tissue, museum skins, and teeth derived from modern raptor pellets. DNA was successfully extracted from 16 modern skins (n = 13 M. montanus; n = 3 other Microtus) and 12 teeth (n = 5 M. montanus; n = 7 other Microtus) from the vicinity (within 10 km) of the fossil site. In addition, DNA was successfully extracted from 51 modern specimens (liver tissue and museum skins) collected within a 400-km radius of Lamar Cave. Of these 51 samples, 22 were derived from museum skins (n = 21 M. montanus; n = 1 M. longicaudus) and 29 from liver tissue (n = 9 M. montanus; n = 18 other Microtus) of specimens trapped in the field. We followed the animal tissue protocol using the Qiagen Dneasy Tissue Kit (Qiagen, Valencia, California, United States) on 6.5 mg of liver samples and 7.5 mg of tooth samples ( n = 3). For museum samples, DNA was extracted from the ventral skin incision (0.5 to 10.3 mg; average, 2.7 mg). We amplified the entire cytochrome b gene (1,143 bp) for some modern skins ( n = 13) and liver tissue ( n = 29) with the following primers: forward primers (5′–3′), MVZ 05 CGA AGC TTG ATA TGA AAA ACC ATC GTT (cyt b −51, M. musculus 14088) and ARVIC 07 AAA GCC ACC CTC ACA CGA TT (cyt b 514, M. musculus 14653); reverse primers (5′–3′) MICRO 06 GGA TTA TTT GAT CCT GTT TCG T (cyt b 602, M. musculus 14741) and VOLE 14 TTT CAT TAC TGG TTT ACA AGA C (cyt b 1170, M. musculus 15309). DNA was extracted from a total of 81 modern Microtus samples, and 79 of those yielded successful amplification. Of these 79 samples, 48 were positively identified as M. montanus . Of the M. montanus samples, 41 specimens were successfully haplotyped. Sequences have been deposited in GenBank ( see Supporting Information ). Genetic data from T. talpoides We have built upon the previously published T. talpoides data set ( Hadly et al. 1998 ) with additional temporal sampling ( n = 3) (see Table 1 ). T. talpoides from Lamar Cave demonstrated remarkable continuity in gene diversity through time, with only three haplotypes present, all of which differ from each other by one synonymous third-position mutation. The majority of the fossil T. talpoides specimens are from haplotype A (83%), which is not found elsewhere in a 400-km radius around Lamar Cave. This constancy in the genetic lineage of T. talpoides within a single locality persists in spite of climatic changes and concurrent significant population and body size changes ( Hadly et al. 1998 ). Data analysis Arlequin v. 2.000 ( Jaarola and Tegelström 1996 ; Schneider et al. 2000 ) was used to calculate gene diversity, number of segregating sites, and nucleotide diversity for each time interval (see Figure 1 B and Table 1 ). Additionally, genetic data were used to estimate the mitochondrial effective population size (N e ) for all points in the past. Assuming a neutral model of molecular evolution, θ S (where S is the number of segregating sites; θ S = 2 N e μ , where μ is the mutation rate for the complete sequence per generation) was used to estimate N e . θ S was preferred over other estimators of θ since θ H is biased for single locus estimates, θ k does not incorporate sequence information, and θ π has higher variance. These estimates are determined entirely from genetic data; thus we label them N e-tt_gen (T. talpoides) and N e-mm_gen (M. montanus). We assumed a range of μ (2% [low], 4% [moderate], or 10% [high] per million years per bp) ( Table 2 ). Additionally, to estimate haplotype turnover we calculated β -diversity (used traditionally in ecology to measure species turnover) using Cody's index ( Cody 1968 ). For β -diversity each haplotype was treated as a “species.” Effects of sampling Given the general limitations of obtaining aDNA sequence data, sample sizes will always present challenges to ancient population genetic studies. To investigate whether our samples are adequate for addressing temporal gene diversity in both species, we used a neutral population model to evaluate expected diversity. Ewens (1972) derived expressions for the sampling distribution characterizing a closed population of size N, a gene with mutation rate μ, and an infinite alleles model given a sample size n. Here we use the Ewens distribution to ascertain the ability of our data to detect variation in values of gene diversity. In addition, this distribution allows us to (1) predict the distribution of expected gene diversity at each point in time (independently) given estimates of N e_ecol , a moderate mutation rate of 4% per million years, typed sequence length (312 bp for M. montanus and 63 bp for T. talpoides ), and sample size ( M. montanus: n 166ybp = 7, n 470ybp = 7, n 845ybp = 18, n 1438ybp = 4, and n 2525ybp = 6; T. talpoides: n 166ybp = 34, n 470ybp = 5, n 845ybp = 29, n 1438ybp = 4, and n 2525ybp = 11) and (2) predict change in the expected gene diversity distribution for a large sample size of 100. A modified version of MONTE CARLO ( Slatkin 1994 , 1996 ) was used to generate possible allele configurations for a given set of parameters (θ, K, and n), and gene diversity was calculated for each configuration. Simulations were repeated 1,000 times for each sampling time point, resulting in the 95% confidence intervals for the distribution of predicted gene diversities given a closed, selectively neutral population through time. Serial coalescent simulations We assessed the significance of the observed changes in gene diversity between time points using simulations. The serial coalescent was used as a framework to model M. montanus evolution during the past 2,525 years. Simulations were repeated 1,000 times to generate genetic data and changes in gene diversity for the null hypothesis. The significance of the observed changes was then inferred by comparing it to the generated null values (we thus used a Monte Carlo significance test). Estimates of N e_ecol were used to set up a null hypothesis specifying population size change through time in a closed population. We assumed that past changes in population size between intervals were due to exponential growth or decline. The effective population size at two time points and the time between the points was used to calculate a growth rate. The estimated growth rates were r 1438–2525ybp = −0.000178, r 845–1438ybp = 0.000732, r 470–845ybp = 0.0004385, and r 166–470ybp = 0.0003212. The null hypothesis thus corresponds to effective population sizes at particular points in the past and exponential growth or decline between those intervals, and represents an ecologically realistic description of the past 2,525 years. We assumed a constant population size prior to 2,525 ybp, as we have no data before this point in time. The coalescent program SIMCOAL ( Excoffier et al. 2000 ) was modified to incorporate temporal sampling (also known as heterochronous sampling): n 1 samples modeled back in time, with n 2 samples added to the genealogy at a time point t 1 generations in the past. The serial coalescent ( Rodrigo and Felsenstein 1999 ; Drummond et al. 2003 ) has been used to estimate parameters such as μ for HIV and ancient mtDNA (most recently using an MCMC approach; Drummond et al. 2002 ; Lambert et al. 2002 ). In this paper, we present what we believe to be its first application as a simulation tool used to predict change in gene diversity for the null model of population size change described above. Running the model 1,000 times provides the expected distribution for change in gene diversity. Using N e-mm_ecol estimates based on high-, moderate-, and low-density estimates for M. montanus (186, 126, and 60 voles/ha; Sullivan et al. 2003 ) and a high, moderate, and low mutation rate ( μ = 10%, 4%, and 2% per million years per bp; sequence length = 312 bp; finite sites mutation model, no rate heterogeneity), we investigated the significance of observed changes in gene diversity at all nine parameter combinations over the entire time span (2,525 to 166 ybp) and between 2,525 and 845 ybp (spanning the Medieval Warm Period). Additionally, we also investigated the relevance of temporal data to our ability to reject the null hypothesis. Significance tests were repeated assuming the observed data were from only the most recent genetic samples. Again, simulations were repeated for all nine combinations of mutation rate and density estimates. Supporting Information Accession Numbers Sequences for the successfully haplotyped M. montanus specimens described in Materials and Methods have been deposited in GenBank under accession numbers AY660606 to AY660629.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC515369.xml

555765

Comparative promoter region analysis powered by CORG

Background Promoters are key players in gene regulation. They receive signals from various sources (e.g. cell surface receptors) and control the level of transcription initiation, which largely determines gene expression. In vertebrates, transcription start sites and surrounding regulatory elements are often poorly defined. To support promoter analysis, we present CORG , a framework for studying upstream regions including untranslated exons (5' UTR). Description The automated annotation of promoter regions integrates information of two kinds. First, statistically significant cross-species conservation within upstream regions of orthologous genes is detected. Pairwise as well as multiple sequence comparisons are computed. Second, binding site descriptions (position-weight matrices) are employed to predict conserved regulatory elements with a novel approach. Assembled EST sequences and verified transcription start sites are incorporated to distinguish exonic from other sequences. As of now, we have included 5 species in our analysis pipeline (man, mouse, rat, fugu and zebrafish). We characterized promoter regions of 16,127 groups of orthologous genes. All data are presented in an intuitive way via our web site. Users are free to export data for single genes or access larger data sets via our DAS server . The benefits of our framework are exemplarily shown in the context of phylogenetic profiling of transcription factor binding sites and detection of microRNAs close to transcription start sites of our gene set. Conclusion The CORG platform is a versatile tool to support analyses of gene regulation in vertebrate promoter regions. Applications for CORG cover a broad range from studying evolution of DNA binding sites and promoter constitution to the discovery of new regulatory sequence elements (e.g. microRNAs and binding sites).

Background Comparative sequence analysis has been a powerful tool in bioinformatics for addressing a variety of issues. Applications range from grouping of sequences (e.g. protein sequences into families) to de novo pattern discovery of functional signatures. Speaking of gene regulation, it has been known for a long time that there is considerable sequence conservation between species in non-coding regions of the genome. A comprehensive explanation of this observation is still elusive. However, sequence conservation within promoter regions of genes often stems from transcription factor binding sites that are under selective pressure (see [ 1 ] for a review and [ 2 ] for a systematic assessment of binding site conservation in man and mouse comparisons). Conserved sequence elements of other types have recently caught much attention. Not all non-coding conserved DNA in the vicinity of a gene's transcription start site necessarily functions at the level of transcriptional regulation. For example, most known methylation-guide snoRNAs are intronencoded and processed from transcripts of housekeeping genes [ 3 ]. A few microRNAs are apparently linked to protein coding genes, most notably mir-10 and mir-196 which are located in the (short) intergenic regions in the Hox gene clusters of vertebrates [ 4 - 7 ]. A second class of conserved sequence elements exert their function as regulatory motifs in the untranslated region (UTR) of the primary transcript or the mature mRNA. The UTRsite database [ 8 ], for example, lists about 30 distinct functional motifs including the Histone 3'UTR stem-loop structure (HSL3) [ 9 ], the Iron Responsive Element (IRE) [ 10 ], the Selenocysteine Insertion Sequences (SECIS) [ 11 ], and the Internal Ribosome Entry Sites (IRES) [ 12 ]. Most of these elements are contained in CORG since short intergenic regions or introns upstream of the translation start site are entirely covered by our definition of an upstream region. Phylogenetic footprinting The CORG framework aims at detecting and describing regulatory elements that are proximal to the transcription start site. In this context, the comparison of upstream regions of orthologous genes is particularly valuable. This concept is called "phylogenetic footprinting" and an overview of this approach can be found in [ 13 ]. Phylogenetic footprinting in a strict sense is carried out on orthologous promoter regions. Local sequence similarities can then be directly interpreted as related regions harboring conserved functional elements. We denote these similarities as Conserved Non-coding Blocks (CNBs). Multi-species sequence conservation Comparative approaches gain power from the inclusion of sequences from more than two species [ 14 ]. Multi-species comparisons help to increase specificity at the expense of intra-species sensitivity since supporting evidence (conservation) stems from many observations. To give an example, Man-mouse-rat comparisons enhance the detection of transcription factor binding sites since the rat genome is more divergent from the mouse genome than anticipated [ 15 ]. A nice property of vertebrate microRNAs is the high degree of sequence conservation which is found in alignments of man, mouse and fish microRNAs [ 16 ]. Both types of comparisons are available in CORG. In CORG, we consider cross-species conservation between promoter regions from 5 vertebrate genomes, namely Homo sapiens , Mus musculus , Rattus norvegicus , Danio rerio and Fugu rubripes . Multiple alignments are built from pairwise CNBs as described in the subsequent section. Construction and content Groups of orthologous genes In this work, we take a gene-centered view of phylogeny. Homology among proteins and thus genes is often concluded on the basis of sequence similarity. The EnsEMBL database [ 17 ] allows to distinguish orthologous from merely homologous genes by taking information on conserved synteny into account. We employed single linkage clustering on the graph of EnsEMBL orthologous gene pairs to define the CORG gene groups. Genomic mapping of validated promoter regions Various recent experimental efforts supply information about the position of transcriptional start sites in the human and mouse genome. Table 1 gives an overview on the resources that were employed in CORG. Table 1 Resources for validated transcription start sites Database name Features Eukaryotic promoter database (EPD) [44] The Eukaryotic promoter database is the smallest in size, but largely consists of manually curated entries. DataBase of Transcriptional Start Sites (DBTSS) [45] The DBTSS contains reliable information on the transcriptional start sites for man and mouse promoters. They exploit the oligo-capping technique to enrich their pool of clones for full-length 5'-to-3' cDNAs H-Invitational Database (H-InvDB) [46] H-InvDB is an international effort to integrate annotation of 41,118 full-length human cDNA clones that are currently available from six high throughput cDNA sequencing projects. FANTOM 2 (RIKEN) [47] The RIKEN consortium presented the FANTOM collection of RIKEN full-length cDNA clones. FANTOM stands for Functional Annotation of Mouse cDNA clones. The Reference Sequence project (RefSeq) [48] The Reference Sequence project aims to provide a comprehensive, integrated, non-redundant set of sequences, including full-length transcripts (mRNA) Some repositories offer genomic coordinates for their start site entries. Existing genomic mapping information was incorporated unless the underlying genome assembly build differed. The remaining data were projected onto the genome with SSAHA (Sequence Search and Alignment by Hashing Algorithm), a rapid near-exact alignment algorithm [ 18 ]. Sequence retrieval The notion of "promoter region" deserves some further explanation in the context of our approach. Typically, though not exclusively, we expect conserved regulatory regions to appear in the vicinity of the transcription start site of a gene. Since we do not know the precise location of the start of transcription for each and every gene, we chose to compare the sequence regions upstream of the start of translation from orthologous genes. If verified transcription start sites are known, we define a sequence window that is large enough to hold both, translation and transcription start sites, plus 5 kB upstream sequence. In case we lack this information, our observations on known transcription start sites indicate that most promoter regions should be captured in a sequence window of 10 kb size (Additional File 1 ). The size of a promoter region may be bounded by the size of the corresponding intergenic region. If an annotated gene happens to lie within the primary sequence window, the promoter region is shortened to exclude exonic sequence. Figure 1 Genomic context of human SRF. This image is displayed after the user selected a gene identifier on the search page. It provides the user with the genomic context of the selected gene. Known and predicted transcription start sites are shown as labelled red dots. Local similarities to homologous regions from other species are shown as connected purple boxes. Blue bars depict all upstream regions as contained in CORG. The structure of the corresponding EnsEMBL transcripts as well as the extent of RefSeq transcripts is show in the bottom track. Detection of pairwise local sequence similarities Significant local sequence similarities (phylogenetic footprints) in two sequences are computed with an implementation of the Waterman-Eggert algorithm. We have already given an account of the algorithm and statistics in [ 19 , 20 ]. The underlying alignment scoring scheme is the general reversible model [ 21 ]: where Q is the transition rate matrix. We left out the elements on the diagonal, which are constrained by the requirement that the sum of all elements in a row equals zero. The π i are the stationary nucleotide frequencies, their sum is constrained to be one. Although the two genomes under consideration are in general not in their stationary state with respect to the substitutional process we take the mean of the two observed nucleotide frequencies, , to be the best estimate of the stationary base composition. From other studies we have further knowledge about the relative rates between transversions, the transition A:T→G:C, and the transition G:C→A:T, which occur in roughly in the ratio 1:3:5 along vertebrate lineages [ 22 ]. These ratios of rates would generate sequences with 40% GC in their stationary state. To accommodate the observed nucleotide frequencies π i we have to allow for deviation from those ratios. We do this by choosing for example α ∝ ( R ( A → T )/ π T + R ( T → A )/ π A )/2, where R ( i → j ) is either 1, 3, or 5 depending on the process under consideration. At the end we scale the matrix Q , such that the PAM distance [ 23 ] of the substitution model equals the observed degree of divergence between the two species under comparison. Since we were mainly interested in highly conserved regulatory elements, we demanded an average similarity level at least as high as the average exon conservation between the species under comparison. The score for aligning two nucleotides i and j is then s ( i , j ) = log( P ( i , j )/( π i π j )) where P ( i , j ) is the probability of finding the pairing of i and j under the above substitution model [ 21 ]. Joining pairwise into multiple alignments All CNBs from pairwise sequence alignments are split up into groups as defined by gene homology. For each group a graph O = ( V , E ) with vertices V and edges E is constructed, which represents the species-internal overlap of CNBs on the genomic coordinate level. Each vertex a ∈ V represents a footprint, which is a pairwise local alignment between two species. An undirected edge is placed between two vertices if the corresponding CNBs have only one species in common and show an overlap of at least 10 bp on the sequence level. In our graph O , cliques of minimal size three are detected with an implementation of the Bron-Kerbosh algorithm [ 24 ]. Only those cliques are selected whose species count is equal to their size. This move prohibits the emergence of multiple alignments by similarity of multiple short CNBs to a single long CNB. Multiple alignments are then computed based on all cliques that meet the outlined criteria. We chose to employ the multiple alignment method of [ 25 ] who applies partial order graphs (POG) to the multiple alignment problem. Partial order graphs belong to the class of directed acyclic graphs (DAGs). A DAG is a graph consisting of a set of nodes N and edges E , which are one-way edges and form no cycles. The multiple alignment problem is then reduced to to subsequent alignment steps of individual sequences to a growing multiple alignment graph. If the sequences to be aligned share substantial sequence similarity, the number of bifurcation points within the POG stays low and allows rapid computation of the multiple alignment. Alignment results are subsequently trimmed to encompass the leftmost and rightmost ungapped block of at least 6 nucleotides. Annotation of promoter regions Exon detection with assembled EST clusters Promoter regions in CORG always extend upstream from the most downstream coding start (ATG). As a consequence, promoter regions may contain exons that are not translated. Our way of detecting such exons is a similarity search of man-mouse footprints versus GENENEST [ 26 ], a database of assembled EST clusters. Database searches are carried out for human and mouse footprints with the BLASTN program [ 27 ]. An E-value cut-off of 10 -4 is applied. Annotation with predicted binding sites The TRANSFAC database [ 28 ] is a repository of experimentally verified binding site sequences and representations thereof. These representations are used for querying the collection of man-mouse CNBs for known binding site patterns. Potential binding sites are detected with TRANSFAC weight matrices by the method of [ 29 ]. Here, the intuition is that there are two random models for a given sequence S : one is given by the signal profile F and the other one by the background model B . Under both models the distribution of weight matrix scores can conveniently be calculated by convolution, since the score is a sum of independent random variables. Probability mass distributions of P F (Score( S )) as well as P B (Score( S )) can be computed by dynamic programming if column scores are reasonably discretized. This allows a fine tuning of the proportions of false positives and negatives for each TRANSFAC weight matrix. Both error levels were set to be equal. All details are given in [ 29 ]. Utility and discussion We now present an overview of the web interface of the database and several example applications. Interface The CORG database is accessible via its home page and offers a redesigned web interface. From the search page one can quickly jump to gene loci via EnsEMBL or other standard identifers (e.g. HUGO symbol, LocusLink identifier, ...). The search query is processed according to the chosen reference source and a list of all matching database entries is returned to the user. This list serves as a springboard to a summary page where the genomic context of the selected gene and its similarities to other upstream regions is visualized as in Figure 1 . Pairwise as well as multiple comparisons are displayed on demand at this stage with a JAVA applet that complies with the JDK 1.1 standard. Alternatively, upstream region sequence and corresponding annotation can be exported in EMBL format (sequence data also in FASTA format). The JAVA applet should run on all JAVA-compatible web browsers. Detailed information about the conserved non-coding block structure are simultaneously shown for multiple upstream regions of different species. If available, annotation information on putative binding sites of transcription factors and EST matches are displayed for the query sequence. The applet facilitates zooming into sequence and annotation. In addition, web links are assigned to sequence features that relate external data sources to the corresponding annotation. CORG data may be also embedded into other viewers or programs via the distributed annotation system ( DAS , [ 30 ]). DAS facilitates the display of distributed data sources in a common framework with respect to a reference sequence. Our DAS server constitutes such an external data source. Position information on all conserved non-coding blocks and mapped promoters is accessible from this DAS server. Each DAS sequence feature provides a link to the corresponding CORG database entry. New DAS sources can be easily added to the ENSEMBL display. A small tutorial on installing external DAS data sources is available on our web page . Additionally, tools for on-site batch retrieval of CORG data will be added to the web portal in the near future. Phylogenetic profiling of binding sites One potential application of CORG is phylogenetic profiling of promoter regions. We define phylogenetic profiling in the context of gene regulation as comparative analysis of presence/absence patterns of binding sites in promoter regions. Here, we consider conserved predicted binding sites and contrast them with validated ones. Serum Response Factor (SRF) promoter SRF, a MADS-box transcription factor, regulates the expression of immediate-early genes, genes encoding several components of the actin cytoskeleton, and cell-type specific genes, e.g. smooth, cardiac and skeletal muscle or neuronal-specific genes [ 31 , 32 ]. Mouse embryos lacking SRF die before gastrulation and do not form any detectable mesoderm [ 33 , 34 ]. SRF mediates transcriptional activation by binding to CArG box sequences (Consensus pattern: CC(AT) 6 GG) in target gene promoters and by recruiting different co-factors. SRF regulates transcription downstream of MAPK signaling in association with ternary complex factors (TCFs) (for a review see [ 35 ]). TCFs bind to ets binding sites present adjacent to CArG boxes in many SRF target gene promoters. Figure 1 gives an overview of the genomic context of human SRF. As expected, the upstream region of SRF shows substantial conservation to its rodent orthologs. Additionally, significant alignments were found in comparisons with fish homologs (one from zebrafish and two from fugu). The same data is presented in the multiple alignment view of the JAVA applet in Figure 2 . This view gives a better idea on the location of alignments in the corresponding source sequences. Note, that the spacing between translation start and alignment is greater in fish than in mammals, which hints at different extension of the promoter region in the two subgroups. Figure 2 Graphical multiple alignment view (JAVA applet). Multiple alignment view of 6 homologous sequences from 5 species . All consistent local similarities in the upstream region of SRF homologs are placed relative to the species-specific translation start sites. The distance of the aligned segment to the translation start site is almost equal for all mammals and larger for the fish. The extent of each upstream region is shown as orange bar. Regions covered by flanking genes would be shown in red. We get a better idea on the cause of sequence conservation by browsing the multiple alignment. Textual information can be obtained by clicking on the alignment boxes. Then, the alignment appears in a pop-up window and may be copied to another destination. In Figure 3 , we used CLUSTAL X ([ 36 ]) to render the conservation structure on to the nucleotide level. Here, a striking observation is the conservation of the regulatory feedback loop of SRF to its own promoter in all species under consideration. So far, this feedback loop was experimentally verified in the mouse system [ 37 ] but could exist in all other species under comparison. Figure 3 Textual multiple alignment view . Multiple alignment as rendered by CLUSTAL X . The largest multiple alignment was retrieved from the JAVA applet by a cut and paste operation and rendered in CLUSTAL X [36]. Conserved binding sites are highlighted by red or blue boxes. Known sites as given in TRANSFAC are marked with a dollar sign [42]. Note that the validated Egr-1 site is only conserved in mammals. This site is bound by the serum-inducible Krox-24 zinc finger protein. Non-coding RNAs Non-coding RNA can be classified as transcribed regulatory elements. Non-coding RNAs are also accessible to the user via the CORG database. Since we were primarily interested in non-coding RNAs rather than small mRNA motifs we restricted our search here to long CNBs. A blast search of our multiple alignments with length L ≥ 50 against the Rfam database [ 38 ] and the microRNA Registry [ 39 ] identifies 21 alignments as 7 distinct microRNAs and a single snoRNA, Table 2 . Table 2 Rfam non-coding RNAs in CORG A + sign indicates that a sequence fragment from the corresponding species (hsa Homo sapiens , mmu Mus musculus , rno Rattus norvegicus , dre Danio rerio , tru Takifugu rubripes ) is contained in the CORG CNB; ∅ indicates that a blast search for an orhologous sequence in the Ensemble database was unsuccessful; n.d. mean no descriptive Ensemble gene annotation. The CNBs containing mir-196a-2 are shifted compared to the known microRNA sequences, preventing the detection of the correct stem-loop structure. The B columns marks whether a candidate was identified by a blast search against the Rfam or microRNA Registry, the A column shows whether a hairpin structure was identified by RNAalifold. p RNAz is the p -value for being an evolutionary conserved RNA secondary structure element returned by RNAz. CNB B A p RNAz ncRNA hsa mmu rno dre tru gene 119596 + + 0.995 mir-34c + + + + ∅ n.d. (BCT-4) 119607 + + 0.938 mir-34b in hsa 119658 + + 0.985 159914 + + 0.998 mir-138-2 + + + + ∅ SLC12A3, n.d. in teleosts 159932 + + 0.999 159939 + + 0.998 194777 + + 0.998 mir-196b + - + + + HOXA9, dre: HOXA9a and HOXA9b 194820 + + 0.999 194839 + + 0.999 194941 + + 0.999 226470 + + 0.999 mir-10a + + + + + HOXB4, dre: HOXB4a and HOXB4b 226514 + + 0.999 226555 + + 0.999 226677 + - 0.004 238163 + + 0.992 mir-10b + + + + + HOXD4, dre: HOXD4a, n.d in tru 238188 + + 0.984 238265 + + 0.994 391314 + - 0.125 mir-196a-2 + + - + + HOXC9, dre: HOXC9a 391315 + - 0.999 391318 + - 0.511 470004 + - 0.218 U93 + + + 0 + n.d. 110374 - + 0.995 IRES ? + + + + + DGCR8 146100 - + 0.891 + + + + 0 Ptf1a 393794 - + 0.999 IRE + + + + + SLCA1 The snoRNA U93 is an unusual mammalian pseudouridinylation guide RNA which accumulates in Cajal (coiled) bodies and it is predicted to function in pseudouridylation of the U2 spliceosomal snRNA [ 40 ]. It appears to be specific for mammals. The genomic copy of the human U93 RNA is located in an intron of a series of reported spliced expressed sequence tags (ESTs); furthermore, it has been verified experimentally that U93 is indeed spliced from an intron [ 40 ]. It was detectable in the CORG footprint dataset because of its location upstream of a conserved putative gene C14orf87 with unknown function. The known microRNAs belong to four different groups. The mir10 and the mir196 precursors are located at specific positions in the Hox gene clusters [ 4 - 7 ]. The mir-196 family regulates Hox8 and Hox7 genes, the function of mir10 is unknown. Substitution pattern of non-coding RNAs For a microRNA we expect a subsequence of about 20 nt that is almost absolutely conserved among vertebrates (the mature miRNA) and a well-conserved complementary sequence forming the other side of the stem from which the mature microRNA is excised. In contrast, the substitution rate should be much larger in the loop region of the hairpin [ 41 ]. mir10 is a good example of this typical substitution pattern, which gives rise to a hairpin structure. The pairwise correlation structure of nucleotides is depicted on top of the multiple alignment in Figure 4 . A different pattern is observed for the Iron Responsive element in the 5'UTR of SLCA1 , a member of the sodium transporter family. This time the substitution pattern does not meet the minimal length of the microRNA definition above. Nevertheless, it is conserved across all vertebrate species as shown in Figure 5 . Figure 4 Alignment and predicted RNA structure of mir-10b . The mir-10b CNB shows the typical pattern of substitutions in a microRNA precursor hairpin: There are two well-conserved arms, of which the mature microRNA is almost absolutely conserved, and a much more variable loop region. [43]. Figure 5 Alignment and predicted RNA structure of the Iron Response Element . The Iron Responsive Element (UTRdb [8] identifier: BB277285) shows a substitution pattern that is different from the hairpin structure in Figure 4. Additional orthologous sequences from the frog Xenopus tropicalis (xtr), the chicken Gallus gallus (gga) and the pufferfish Tetraodon nigroviridis are included. Conclusion We have improved and extended our framework of comparative analysis and annotation of vertebrate promoter regions over previous releases (see [ 20 ]). The following features have been added to the CORG framework: • Mapping of validated promoter regions and proper adjustment of the extent of upstream regions. • Multiple alignments from significant local pair wise alignments. • Novel approach to predict transcription factor binding sites. • Web site offers now a genomic context view (as in Figure 1 ) and an option to export sequence and annotation data. The CORG database is accessible via our web site. The user is guided step-by-step through the process of selecting and analyzing her promoter region of choice. CORG features an interactive viewer based on JAVA technology, which is tailored to detailed promoter analysis. Large-scale studies make direct use of our DAS service or the MySQL implementation of CORG in conjunction with an application interface (contact authors for details). We presented selected application examples from the realm of vertebrate gene regulation. Conserved regulatory elements of different kinds (binding sites, microRNAs and UTR elements) are readily accessible to CORG users. New genomes and annotation will be continuously added to CORG. Availability and requirements The database is freely accessible through the website . Programs, scripts and MySQL database dumps are available from the authors upon request. Authors' contributions Christoph Dieterich built the entire pipeline and some parts of the web interface. Steffen Grossmann annotated transcription factor binding sites and provided parts of the web interface. Andrea Tanzer analyzed known and novel RNA elements in the multiple alignments of the CORG database. Stefan Röpcke set up our database of binding site descriptions. Peter F. Arndt worked on an appropriate alignment scoring scheme. Peter F. Stadler and Martin Vingron initiated this work and provided all necessary infrastructure. Supplementary Material Additional File 1 Distribution of distance between start of transcription and translation . Histogram of observed genomic distances between start sites of transcription and translation in man for 1,700 entries from the EPD. The red and blue line indicates the 90% and 95% quantiles, respectively. Distances greater than 10 6 bp were exluded from the analysis as they mostly occur due to mismappings in the ENSEMBL database. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555765.xml

539046

Human Carrying Capacity and Human Health

The issue of overpopulation has fallen out of favor among most contemporary demographers, economists, and epidemiologists. Discussing population control has become taboo. This taboo could be hazardous to public health

The issue of human overpopulation has fallen out of favor among most contemporary demographers, economists, and epidemiologists. Discussing population control has become a taboo topic. Yet, this taboo has major implications for public health. The silence around overpopulation prevents the global health community from making the necessary link between the planet's limited ability to support its people (its carrying capacity—see sidebar on following page) and health and development crises. In this article, I describe how popular thinking on population control has been shaped over the last 200 years, and how our failure to address the population explosion may be one cause of recent epidemics and social unrest. Overpopulation Concerns Peak, Then Decline The question of human overpopulation and its relationship to human carrying capacity has been controversial for over two centuries. In 1798 the Reverend Thomas Malthus put forward the hypothesis that population growth would exceed the growth of resources, leading to the periodic reduction of human numbers by either “positive checks”, such as disease, famine, and war, or “preventive checks”, by which (in the absence of contraception) Malthus meant restrictions on marriage. This “Malthusian view” was rapidly accepted by most politicians, demographers, and the general public, and remained popular until fairly recently. Malthus's worst fears were not borne out through the century following his death in 1834—food production largely kept pace with the slowly growing global population. However, soon after 1934, the global population began to rise steeply as antibiotics, vaccines, and technology increased life expectancy. By the 1960s, concerns of a mismatch between global population and global food supply peaked—expressed in books such as Paul Ehrlich's 1968 The Population Bomb [1] . This book predicted a future scarred by increasing famine, epidemic, and war—the three main Malthusian positive checks. In 1966, United States President Lyndon Johnson shipped wheat to India to avert a famine on the condition that the country accelerate its already vigorous family planning campaign [2] . Johnson was part of an unbroken series of US presidents concerned with the harmful effects of rapid population growth in developing countries. This line extended (at least) from John F. Kennedy to Jimmy Carter. George H. W. Bush was also sympathetic to this view, prior to becoming vice president in 1981. But the 1970s surprised population watchers. Instead of being a period shadowed by calamitous famine, the new crop strains introduced by the “Green Revolution” (especially grains such as rice, wheat, and maize) caused a dramatic increase in the global production of cereals, the main source of energy in the global diet. Among the development community, despair turned into cautious optimism. By the end of the decade, the public health community felt sufficiently empowered to proclaim “Health for All by the Year 2000”. Average life expectancy continued to zoom upwards almost everywhere—even in sub-Saharan Africa. The introduction of safe contraception contributed to a rapid fertility decline in many countries. But while the rate of global population growth declined from its peak in the late 1960s, the absolute increment of increase in annual global population continued to grow. Most population-related scientists, including food scientists and demographers, as well as US President Jimmy Carter, continued to be very concerned about global overpopulation. In 1970, the father of the Green Revolution, the agricultural scientist Norman Borlaug, was awarded the Nobel Peace Prize. In his Nobel lecture, Borlaug warned that the success of the Green Revolution would buy a breathing space for humankind of three decades, unless equivalent action was taken to reduce fertility rates [3] . China tightened its fertility policy in this decade, introducing its one-child policy in 1979. We are failing to confront the population explosion (Illustration: Sapna Khandwala) Concern for the Third World Fades With hindsight, the 1970s can be seen as the decade when widespread concern about overpopulation started to fade. The social and economic milieu of many developed countries, especially in the US, started to change. US foreign aid, as a percentage of the gross national product, declined from the late 1960s, perhaps in part because of the competing needs of the Vietnam War but also perhaps because of the apparent success of development in the Third World. The economic policies known as Keynesianism, which had been dominant since the end of World War II in many developed nations, came under sustained attack. These policies had placed a high value on full employment and social security. Keynesian policies restrained domestic inequality through high taxation and the promotion of social norms that censured conspicuous consumption (such as company executives exercising restraint in their personal salaries and people buying small houses). Shortly before his death, J. M. Keynes had also been crucially involved in the establishment of the World Bank. Keynes appears to have been personally committed to the advance of global justice, and to the reduction of inequality both within and between nations [4] . The world oil shock in 1973 contributed both to “stagflation”—a combination of rising unemployment with higher prices—and to increased economic power for the oil-producing countries of the Third World. Indeed, the term “Third World” came to be considered pejorative and was replaced by the “South”. Stagflation was interpreted as a failure of Keynesian policy. The demise of Keynesianism was accompanied by a further decline in concern for Third World development among elite economists and the general public. It is unlikely that the issue of global population policy figured into the election that put US President Ronald Reagan into office in 1980. Nevertheless, Reagan's policies were to cement a new orthodoxy about global overpopulation and development strategies. Unlike his republican predecessor, Richard Nixon, Reagan considered concerns about global population size to be “vastly exaggerated” [5] . In the same year, the US surprised the family planning world by abdicating its previous leadership in the effort to promote global family planning, at the International Conference on Population, held in Mexico City in 1984. The US took this position against the strenuous opposition of the Population Association of America, which represented many US demographers [5] . As foreign aid budgets fell, the “Health for All” targets began to slip from reach. Instead, international agencies promoted structural adjustment programs, health charges for patients (“user fees”), and the “trickle down” effect as the best ways to promote development. It is plausible that a fraction of the public who remained concerned about Third World development thought that these new economic policies deserved a chance. Less charitably, the new economic policies also appeared to allow people already financially comfortable to abdicate concern for Third World development because the new orthodoxy asserted that market deregulation, rather than aid, was the royal road to development. The increased domestic inequality of recent decades in developed countries [6] probably also contributed to a reduction in concern for the Third World, as working people have had to struggle harder to keep their position in their own society. It is now clear that market deregulation and generally high birth rates have proven disastrous in many Third World countries. “Health for All”, if recalled at all, is now seen as absurdly optimistic. The failure of development is most obvious in many sub-Saharan countries, where life expectancy has fallen substantially. But life expectancy has also fallen in Haiti, Russia, North Korea, and a handful of other nations [7] . The causes for this decline in life expectancy are multiple and complex. Causes that are usually listed include HIV/AIDS (Zimbabwe and Haiti) [ 8 , 9 ], ethnic hatred (Rwanda) [10] , crop failure (North Korea) [11] , poor governance and poverty (several parts of Africa) [12] , and alcoholism (Russia) [13] . Causal theory is complex. Every cause has a cause, and, increasingly, causes are being considered as a part of causal chains, causal webs, and causal snowballs. Some theorists distinguish between identifiable “proximal” causes and deeper, underlying, or “distal”, causes [14] . Yet, among the multitude of causes that can be identified for declines in either total population or life expectancy, overpopulation is hardly considered, except by dissident public health workers such as Maurice King [15] . Demography, the discipline that would appear to be the most likely holder of the Malthusian baton, is now almost entirely silent about overpopulation in developing countries [16] . Instead, most mainstream demographers appear to consider population ageing and European underpopulation as the most important demographic issues for this century. On the other hand, the role of the rapid demographic transition in China (from large to small families, with an average of two or fewer children) is rarely credited as central to the Chinese economic miracle. Overpopulation: A Cause of Crises in Africa? Often, the carrying capacity of one region at one point in time is boosted by the appropriation of the carrying capacity from other people and even other generations. Such resources include oil, deep sea fish, and the stability of the global climate and ecological systems. But in Rwanda, the most densely populated country in Africa, the importation of such resources has long been limited. Unlike other densely populated countries such as Hong Kong and Holland, Rwanda's economy at the time of its most infamous genocide, in 1994, depended almost exclusively on its primary production [17] . The country had little industry, few exports, and little tourism. The price of its most important export, coffee, had declined steeply just before the genocide [18] . Unlike many Asian countries, Rwanda also received few remittances from Rwandans working as guest workers abroad [17] . Among the many different explanations for the horrific 1994 Rwandan genocide, the possibility of a Malthusian check (also called “demographic entrapment”) is scarcely mentioned [ 17 , 19 ]. A Malthusian check in Rwanda was plausible not only because the total population was too large, but perhaps more importantly because the rate of population growth in Rwanda was faster than the capacity of Rwandan society to process the additional people. As a result, many indicators of development went backwards. The limited agricultural capacity forced many young men into Kigali, causing a concentration of young men with few prospects other than what they might gain from violence. There is even less scientific discussion that entertains the possibility that the sub-Saharan epidemic of HIV/ AIDS may also be a Malthusian check [19] . This is plausible if one applies a conceptual framework that combines the erosion of human carrying capacity through the same rapid population growth seen in Rwanda, with a consequent decline in per capita income and food supply. Furthermore, slowly operating feedbacks occurring as a result of the epidemic further undermined development, including the loss of human capital as teachers died [20] , the loss of agricultural expertise as farmers died [21] , and a deepening debt and loss of productivity from the countless funerals. And leaders in the developed world and many within Africa itself failed to devote the resources and provide the leadership required to quell the epidemic. Conclusion Maurice King refers to the silence on overpopulation as the “Hardinian Taboo”, named after the American ecologist Garett Hardin, who described the taboos that humans use to avoid confronting the need for population control [22] . Daniel Orenstein, at the Center for Environmental Studies at Brown University, has argued that powerful social norms inhibit debate about overpopulation in one of the world's most intractable trouble spots, Israel and Palestine [23] . Whatever the cause of the scarcity of modern academic analysis, the related issues of human carrying capacity and overpopulation deserve fresh consideration. The entrapment model has an explanatory power that is lacking in more superficial causal explanations. Of course, solving entrapment is very difficult, but as with most medical problems, a proper diagnosis will help identify the proper treatment. Human Carrying Capacity Human carrying capacity is the maximum population that can be supported at a given living standard by the interaction of any given human-ecological system. This apparently simple concept has many nuances and is rarely used by population scientists. However, in rejecting this term, purists risk making a terrible conceptual flaw, that of thinking that environmental and human resources are largely irrelevant to human population size. It is irrefutable that human ingenuity and cooperation can increase human carrying capacity [24] . But even so, human welfare will continue to depend on the external world, including for resources such as food and water. Humans are neither computer ciphers nor caged mice. That is to say, while a given area might tolerate a theoretically higher density of human population than it does, the reality of human evolution in distinct groups, separated by culture, religion, and language, means that this theoretical maximum will rarely be attained. A degree of underused carrying capacity can be viewed as a desirable buffer around disparate groups, vital for reducing tension and preventing conflict. Even culturally homogenous groups can outgrow their carrying capacity, as in the case of the Great Hunger in Ireland in the 1840s, when the population crashed because of famine, disease, and emigration. Indeed, Malthusian theory was used, in part, to justify the scanty aid provided to the Irish from Britain, a country that did not identify closely with the Irish.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539046.xml

555942

Gene fusions and gene duplications: relevance to genomic annotation and functional analysis

Background Escherichia coli a model organism provides information for annotation of other genomes. Our analysis of its genome has shown that proteins encoded by fused genes need special attention. Such composite (multimodular) proteins consist of two or more components (modules) encoding distinct functions. Multimodular proteins have been found to complicate both annotation and generation of sequence similar groups. Previous work overstated the number of multimodular proteins in E. coli . This work corrects the identification of modules by including sequence information from proteins in 50 sequenced microbial genomes. Results Multimodular E. coli K-12 proteins were identified from sequence similarities between their component modules and non-fused proteins in 50 genomes and from the literature. We found 109 multimodular proteins in E. coli containing either two or three modules. Most modules had standalone sequence relatives in other genomes. The separated modules together with all the single (un-fused) proteins constitute the sum of all unimodular proteins of E. coli . Pairwise sequence relationships among all E. coli unimodular proteins generated 490 sequence similar, paralogous groups. Groups ranged in size from 92 to 2 members and had varying degrees of relatedness among their members. Some E. coli enzyme groups were compared to homologs in other bacterial genomes. Conclusion The deleterious effects of multimodular proteins on annotation and on the formation of groups of paralogs are emphasized. To improve annotation results, all multimodular proteins in an organism should be detected and when known each function should be connected with its location in the sequence of the protein. When transferring functions by sequence similarity, alignment locations must be noted, particularly when alignments cover only part of the sequences, in order to enable transfer of the correct function. Separating multimodular proteins into module units makes it possible to generate protein groups related by both sequence and function, avoiding mixing of unrelated sequences. Organisms differ in sizes of groups of sequence-related proteins. A sample comparison of orthologs to selected E. coli paralogous groups correlates with known physiological and taxonomic relationships between the organisms.

Background Eschericia coli remains a useful resource to the genomic community as it provides important knowledge which can be applied to the analysis of most microbial genomes. Its central role devolves from two facts; first, the accumulated results of seven decades of laboratory experimentation have identified the function(s) of over half of its gene products; second being a metabolic generalist, E. coli 's metabolic functions are widely shared among other organisms. Common practices of annotation rely, more than one might realize, on the accuracy of the annotation of E. coli 's genes. While searches for sequence matches to unknown genes usually yield a large number of matches, chances are high that firm functional information comes only from experimental studies on E. coli . Because annotations of genes do not always indicate that the assignments are derived, and because derived annotations are used serially for further annotation without experimental confirmation, many genes carry original E. coli annotations. It is therefore important to the entire genome-analyzing community that the data on E. coli gene products be as accurate as possible. Since the original GenBank deposit of E. coli K-12 (U00096), new and updated annotations are available at NCBI (U00096.2) and at more specialized databases including, ASAP [ 1 ], coliBASE [ 2 ], CyberCell [ 3 ], EchoBASE [ 4 ], EcoCyc [ 5 ], GenProtEC [ 6 ], and RegulonDB [ 7 ]. An effort is under way to coordinate the current E. coli annotations [ 8 ]. Over recent years, our work on the E. coli genome has led us to an appreciation of the pernicious role that gene fusions often play as troublemakers in function assignments and in relating groups of sequence similar proteins [ 9 ]. The fusion of two independently functioning genes results in the formation of a composite (multimodular) protein encoding for two independent functions located at separate parts of the protein. This type of fusion is not equivalent to the joining of protein domains, i.e. domains encoding binding sites for a cofactor or a substrate, which is seen in multidomain proteins. An example being the enzyme glyceraldehyde-3-phosphate dehydrogenase which according to the domain databases Pfam [ 10 ] and Superfamiliy [ 11 ] contains two domains, an NAD binding site and a dehydrogenase catalytic site. In our studies the entire protein including both domains represents one independent functional unit with one activity. Multidomain proteins are more prevalent and most often encode one overall function for the gene product [ 12 ]. Annotation involving transfer of function from composite proteins to sequence similar matches requires that the alignment regions be evaluated in order to determine whether all activities or only one of them should be assigned to the matching sequence. Currently fused proteins are themselves not always annotated to reflect that they encode more than one function, and rarely is the location of the separate functions indicated. Different combinations of fused genes are seen in the sequenced genomes, adding potential sources for annotation errors. Errors in functional assignments including those caused by fused genes have been noted years ago [ 13 ] and that such proteins may contribute to propagation of annotation errors in databases [ 14 ]. The fused proteins also interfere with the generation of sequence related protein groups as they link proteins based on their coexistence in a fused protein and not purely based on sequence similarity. Components of fused genes are often not sequence related, so generating protein groups without taking gene fusions into account may result in "mixed" groups of proteins with different sequence relatedness, functions and evolutionary histories. Previous work has been published where we identified fused E. coli proteins from partial alignments between proteins encoded in the E. coli genome [ 15 ]. This work resulted in the identification of 287 multimodular proteins. As our analysis continued and more genome sequences were incorporated in our studies we realized that most of these identified multimodular proteins actually contained multiple domains and had one overall function. We have therefore revised our method of detecting fused proteins. We are making use of sequence information from 50 genomes including E. coli to detect proteins which are fused in the E. coli genome and are present as individual components in one of the other genomes. We have also made use of published experimental data on E. coli gene fusions. As a result the number of fused proteins in E. coli has been reduced to 109. The number of groups of sequence related proteins was also reduced from 609 to 490 since some of the previously identified groups are made up of protein domains catalyzing only part of an overall reaction. This work represents a revision of the state of fused proteins in the E. coli genome their affect on genome analysis both within E. coli and across genomes. Results Multimodular vs. multifunctional proteins To prevent confusion, we define multimodular proteins as those seeming to result from gene fusion in which two independent proteins are connected. Multimodular proteins encode separate functions in different parts of the molecule. These functions might be the same if two like elements have fused, or as we see more often in E. coli , they differ in sequence and activity. Distinctly different, multifunctional proteins are defined as those that carry out more than one reaction or activity in the same part of the protein. Examples of such multifunctional proteins are encoded by the genes cob U, bir A, ubi G, fol D, cys G, tes A, and ndk (for details see gene products at GenProtEC [ 16 ]). A protein that illustrates both properties is the FadB protein of E. coli [ 17 , 18 ]. FadB is a multimodular protein with N-terminal and C-terminal modules. Its N-terminal module is multifunctional with three activities that are catalyzed at the same active site and cannot be spatially separated along the length of the protein. The three activities are 3-hydroxybutyryl-CoA epimerase, delta(3)-cis-delta(2)-trans-enoyl-CoA isomerase, and enoyl-CoA hydratase. The C-terminal module of FadB encodes a single function, 3-hydroxyacyl-CoA dehydrogenase. Adding the N-terminal and C-terminal modules, there are 4 activities for the FadB protein. Identifying multimodular proteins in E. coli In earlier work, before the genomic sequence of E. coli was completed, we saw that sequence similarity among its proteins was widespread [ 9 , 19 ]. After the entire sequence was available, we identified 287 E. coli proteins as being multimodular and encoded by fused genes [ 15 , 20 ]. The modularity of the proteins was inferred from the alignments among E. coli proteins. However, we have since found that many of these so-called multimodular proteins were proteins containing more than one domain and not more than one protein. Such multidomain proteins may appear to encode two functions but in reality encode two or more conserved motifs (i.e. DNA-binding and effector-binding domains of LysR type transcriptional regulators). By including sequence information from other genomes besides E. coli we were able to better distinguish fusions of complete proteins versus the more common fusions of protein domains. Of the 287 proteins previously identified as multimodulars only 70 remained as fused proteins in this study with the remaining representing domain fusions. In the present work, some of the fused proteins were identified by searching the literature for experimental data. Examples of E. coli proteins long known to contain multiple functions encoded at separate parts of the proteins include GlnE [ 21 ], MetL [ 22 ], ThrA [ 23 ], and TyrA [ 24 ]. We have collected such experimentally verified information over time [ 9 ], labeled as multimodular proteins with literature citations in our database GenProtEC [ 16 ]. Other multimodular proteins were identified by selected types of alignments between E. coli proteins and proteins encoded in 50 sequenced genomes. The component proteins of a multimodular protein may be unimodular and unfused in another genome. We looked for alignments between the larger potentially multimodular proteins in E. coli and smaller orthologous proteins that are homologous to only one of the modules (Figure 1a ). Not all gene fusions of E. coli will be detected by this method. For instance elements of a fused gene may have diverged to the point where the component modules no longer have detectable similarity to their homologous counterparts, or the independently existing modules may have been lost from the gene pool of the 50 genomes analyzed, or the 50 organisms may contain only the multimodular form. Figure 1 Identification and sequence similarity of multimodular E. coli proteins. (a) An E. coli protein (gi1787250) aligns with two smaller proteins from C. acetobutylicum , histidinol phosphatase (gi15026114) and imidazoleglycerol-phosphate dehydratatase (gi15023840). The E. coli protein represents a fused or multimodular protein encoding the two functions in separate parts of the protein as indicated by the two non-overlapping alignment regions. Based on the alignment regions, the E. coli protein is separated into two separate components, modules. The modules are identified with the extensions "_1" or "_2" to indicate their location in the gene product as N-terminal or C-terminal, respectively. (b) Sequence similarity between modules of the multimodular proteins is shown. No detectable similarity between the joined modules is indicated by a difference in the module patterns in the cartoon. Similarity is measured by Darwin and indicates that the proteins align at a distance of ≤ 200 PAM units over at least 83 amino acid residues or >45% of the length of the proteins. This level of similarity also reflects whether the modules belong to the same paralogous group. In total we identified 109 E. coli proteins to be multimodular, with 101 containing two modules and 8 containing three modules. The largest number of multimodular proteins joined modules of dissimilar sequence (illustrated in Figure 1b ). An abbreviated list of the modules and their functions is shown in Table 1 . A complete list of the multimodular E. coli proteins is made available: ' [see Additional file 1 ]'. The remaining proteins, 97.5 % of the total, were considered to be unimodular. The average length of the multimodular proteins was 637 residues compared to 309 for the remaining proteins in the chromosome (Figure 2 ). Individual modules from the multimodular proteins were on average 300 residues long, similar to the length of the unimodular proteins. However, the size alone of a protein does not reflect multimodularity as we found many large proteins to be unimodular. Table 1 Examples of multimodular E. coli proteins. Gene Module Start End Gty 1 Module Function thr A b0002_1 1 461 e aspartokinase I, threonine sensitive thr A b0002_2 464 820 e homoserine dehydrogenase I, threonine sensitive rib D b0414_1 1 143 e diaminohydroxyphosphoribosylaminopyrimidine deaminase rib D b0414_2 147 366 e 5-amino-6-(5-phosphoribosylamino) uracil reductase put A b1014_1 1 569 e bifunctional: transcriptional repressor (N-terminal); proline dehydrogenase, FAD-binding (C-terminal) put A b1014_2 618 1320 e pyrroline-5-carboxylate dehydrogenase adh E b1241_1 1 400 e acetaldehyde-CoA dehydrogenase adh E b1241_2 449 891 e iron-dependent alcohol dehydrogenase thi P b0067_1 1 274 t thiamin transport protein (ABC superfamily, membrane) thi P b0067_2 285 536 t thiamin transport protein (ABC superfamily, membrane) mdl A b0448_1 1 310 pt putative transport protein, multidrug resistance-like (ABC superfamily, membrane) mdl A b0448_2 314 590 pt putative transport protein, multidrug resistance-like (ABC superfamily, ATP_bind) mod F b0760_1 1 260 t molybdenum transport protein (ABC superfamily, ATP_bind) mod F b0760_2 261 490 t molybdenum transport protein (ABC superfamily, ATP_bind) hrs A b0731_1 1 178 t PTS family enzyme IIA, induction of ompC hrs A b0731_2 186 454 t PTS family enzyme IIB, induction of ompC hrs A b0731_3 456 628 t PTS family enzyme IIC, induction of ompC ato C b2220_1 1 125 r response regulator ato C b2220_2 145 461 r sigma54 interaction module of response regulator (EBP family) evgS b2370_1 1 935 e histidine kinase of hybrid sensory kinase evgS b2370_2 953 1197 r response regulator of hybrid sensory histidine kinase gln G b3868_1 1 120 r response regulator, two-component regulator with GlnL, nitrogen regulation gln G b3868_2 139 469 r sigma54 interaction module of response regulator (EBP family) kef A b0465_1 1 779 o unknown function module of mechanosensitive channel kef A b0465_2 780 1120 t mechanosensitive channel (MscS family) arg A b2818_1 1 293 o acetylglutamate kinase homolog (inactive) arg A b2818_2 298 442 e N-alpha-acetylglutamate synthase (amino acid acetyltransferase) ydc R b1439_1 1 117 pr putative transcriptional regulator (GntR family) ydc R b1439_2 118 468 pe putative amino transferase rnf C b1629_1 1 448 pc Fe-S binding module of electron transport protein rnf C b1629_2 450 740 o unknown function module of electron transport protein 1 Gene product type: e, enzyme; pe, putative enzyme; r, regulatory protein; pr, putative regulatory protein; t, transport protein; pt, putative transport protein; pc, putative carrier protein; o, unknown function. Figure 2 Size distribution for multimodular and single module proteins. The protein lengths in amino acid residues are shown for single module proteins (□) and for multimodular proteins (■). On average the multimodular proteins are longer than the unimodular proteins, 637 amino acids versus 314 amino acids. The length of a protein alone does not infer multimodularity and long single module proteins are seen. Characteristics of multimodular proteins of E. coli Table 2 shows some characteristics of the modules in the multimodular proteins. The majority of the E. coli modules, 90%, were found to have homologs existing as independent proteins in one of the 50 genomes analyzed. Independent unimodular homologs within E. coli were detected for only 57% of the modules (data not shown). A list of the major types of multimodular proteins is shown in Table 3 . Table 2 Features of multimodular E. coli proteins: No. Modules 109 multimodular proteins 226 101 bimodular proteins 202 8 trimodular proteins 24 with identity to unfused orthologs 203 without identity to unfused orthologs 23 known function 151 putative function 66 unknown function 9 type of protein 1 : enzyme 97 transport protein 85 regulatory protein 26 other 18 1 includes putative assignments Table 3 Types of multimodular proteins. Protein type 1 Protein names 2 Enzyme Aas, AdhE, AegA, ArgA, ArnA, CysG, Dfp, DgoA, DsbD, FadB, FadJ, FtsY, GlcE, GlmU, GlnE, Gsp, HisB, HisI, HldE, HmpA, MaeB, MetL, MrcA, MrcB 3 , NifJ 3 , PaaZ, PbpC, PheA, PolA, PurH, PutA, RbbA 3 , RibD, Rne 3 , ThrA, TrpC, TrpD, TyrA, YdiF, YfiQ, YgfN, YgfT, YjiR Transport protein AlsA, AraG, CydC, CydD, DhaH, Ego, FeoB, FhuB, FruA, FruB, FrvB, HrsA 3 , KefA, MacB, MalK, MalX, ManX, MdlA, MdlB, MglA, ModF, MsbA, MtlA, NagE 3 , PtsA, PtsG, PtsP, RbsA, ThiP, Uup, XylG, YbhF, YbiT, YddA, YejF, YheS, YjjK, YliA, YnjC, YojI, YpdD 3 , YphE Regulatory protein Ada, Aer, ArcB, AtoC, BarA, BglF, CheA, CheB, EvgS, GlnG, KdpD, MalT, RcsC, TorS, YdcR, YfhA, YieN, ZraR Other InfB, MukB 3 , RnfC, YegH, YfcK, YoaE 1 Gene type includes known and putative functions. 2 Protein names derived from gene names. 3 Genes encoding three modules. • Many of the multimodular enzymes function in the biosynthesis or degradation of compounds (amino acids, cofactors, peptidoglycan and fatty acids). • The majority of the multimodular transport proteins encode fusions of components of the ABC superfamily transporters (ATP-binding and membrane component). Also, fusions of the PTS proteins were detected in different combinations. Thirteen proteins contained two or more PTS components, including Hpr, enzymes I, IIA, IIB, or IIC. • Among the multimodular regulatory proteins, two-thirds were part of two-component regulatory systems and contained histidine kinases fused to response regulators. Seldom were known domain subdivisions within these modules detected by the rules we applied. While the fraction of enzymes (39%) is similar to the fraction of enzymes encoded in the genome as a whole (36%), the proportion of multimodular transport proteins (38%) and regulatory proteins (17%) were higher than their proportion genome wide (14% and 8% respectively). The over-representation in transporters and regulators is a reflection of the level of gene duplication seen for these proteins. Large paralogous groups are detected for some of the ABC transporter protein subunits and for components of the two-component regulators. Pairwise similarity of E. coli single modules All unimodular proteins, including the modules obtained from multimodular proteins, were tested pairwise for sequence similarity. Matching all single module E. coli proteins to each other using the AllAllDb algorithm of the Darwin package, we collected all aligned pairs with a similarity score of less than or equal to 200 PAM units, with an alignment of at least 83 residues. Altogether 9,626 unique pairs met these criteria (data available at GenProtEC [ 16 ]). Paralogous groups of E. coli protein modules We used the data on pairwise similarity to assemble groups of proteins of similar sequence that were unlike other proteins in the cell. Besides the PAM less than 200 and alignment length of at least 83 residues, two additional requirements were imposed; that more than 45% of each protein in each pair be aligned, and that a module could not belong to more than one group. A transitive clustering process was used to form the sequence-similar groups [ 9 ]. This grouping method requires only that each member of the group have sequence similarity to at least one other member of the group and does not require a detectable similarity among all the members of a group. Both closely related groups and groups with more divergent proteins were found. We identified 490 sequence-similar or paralogous groups in E. coli ' [see Additional file 2 for a complete list of the sequence-similar E. coli groups and their members]'. Altogether 1946 unimodular proteins belonged to one of the groups. Modules from 94 of the multimodular proteins were present in 61 of the groups. Table 4 shows the power law type of distribution of the number of members in the groups, smaller groups being more abundant than large ones. There were 279 groups of two proteins, and only 10 % of the groups had 7 or more members. As shown in Table 5 , the smaller groups tended to be tight groups in which the majority of sequences were related by our criteria to all or most others in the group. Larger groups were more divergent with a minority of members related to all others. At group size 8 and above, no members have the property of relating to all others. Table 4 Size distribution of paralogous groups. Group size No. Groups 2 279 3 91 4 32 5 31 6 6 7 18 8 7 9 2 10 2 11 3 12 1 13 2 14 2 18 2 20 1 21 1 22 2 24 1 30 2 40 1 43 1 46 1 51 1 92 1 Table 5 Sequence relationships within paralogous groups. Group size No. Groups All See All All See Some 3 92 56 36 4 32 21 11 5 31 7 24 6 6 0 6 7 18 2 16 The largest groups of paralogous enzymes, transport proteins and regulatory proteins are shown in Table 6 , 7 and 8 , respectively. While enzymes represent the largest gene product type in E. coli with known or predicted function, they tend to be present in smaller paralogous groups as compared to the transporters and regulators. Among the larger groups the oxidoreductases and the subunits of oxidoreductases are most common, making up 8 of the top 20 enzyme groups (Table 6 ). Table 6 Paralogous enzyme groups in E. coli . No. Members Group function 20 oxidoreductase, Fe-S-binding 18 oxidoreductase, NAD(P)-binding 18 oxidoreductase 1 , NAD(P)-binding 13 aldehyde oxidoreductase, NAD(P)-binding 13 oxidoreductase, FAD/NAD(P)-binding 11 sugar kinase 10 terminal oxidoreductase, subunit 9 aldo-keto oxidoreductase, NAD(P)-binding 8 phosphatase 8 nucleoside diphosphate (Nudix) hydrolase 8 acyl-CoA ligase 7 glutathione S-transferase 7 RNA helicase, ATP-binding 7 sugar epimerase/dehydratase, NAD(P)-binding 7 alcohol oxidoreductase 7 acyltransferase 7 aminotransferase, PLP-binding 7 decarboxylase, TPP-binding 7 crotonase 7 acyltransferase 1 Contains GroES-like structural domain (SCOP sf50129). Table 7 Paralogous transport protein groups in E. coli No. Members Group function 92 ABC superfamily transport protein, ATP-binding component 51 ABC superfamily transport protein, membrane component 40 MFS family transport protein 24 ABC superfamily transport protein, periplasmic binding component/ transcriptional regulator (GalI/LacR family)/ 22 APC family transport protein 12 ABC superfamily transport protein, membrane component 11 PTS family transport protein, enzyme IIA 9 ABC superfamily transport protein, periplasmic binding component 8 ABC superfamily transport protein, periplasmic binding component 7 GntP family transport protein 7 RND family transport protein 7 ABC superfamily transport protein, membrane component 5 HAAP family transport protein 5 PTS family transport protein, enzyme IIB 5 PTS family transport protein, enzyme I 5 GPH family transport protein 5 NCS2 family transport protein 5 HAAP family transport protein 5 transport protein 5 PTS family enzyme IIC 5 RhtB family transport protein 5 outer membrane porin Table 8 Paralogous regulatory protein groups in E. coli. No. Members Group function 46 LuxR/UhpA or OmpR family transcriptional response regulator of two-component regulatory system 43 LysR family transcriptional regulator 30 GntR or DeoR family transcriptional regulator 22 sensory histidine kinase in two-component regulatory system 14 sigma54 activator protein, enhancer binding protein 14 AraC/XylS family transcriptional regulator 7 ROK family transcriptional regulator/sugar kinase 7 IclR family transcriptional regulator 5 methyl-accepting chemotaxis protein 5 MerR family transcriptional regulator 4 DNA-binding regulatory protein 3 AraC/XylS family transcriptional regulator 3 MarR family transcriptional reguator 3 AsnC family transcriptional regulator ATP-binding components of the ABC superfamily of transport proteins are highly conserved and make up the overall largest paralogous group in E. coli (Table 7 ). The other two components of the ABC superfamily transporters are less conserved with membrane components in groups of 52 or less and periplasmic binding components in groups of 9 or less. Components of the PTS system; enzyme IIA, IIB, IIC and I also formed sequence similar groups. One of the groups classified as a group of transporter proteins actually contains both transport proteins (periplasmic binding components of the ABC superfamily) and regulatory proteins (transcriptional regulators of the GalR/LacI family). These two functional types are sequence related, and all of the proteins contain a common structural domain (SCOP sf53822) for the binding of small molecules [ 25 , 26 ]. The difference lies in the presence or absence of a DNA-binding domain. Response regulators of two-component regulatory systems make up the largest group of regulatory proteins in E. coli (Table 8 ). Sensory histidine kinases of two-component regulatory systems and the sigma54 activating proteins also constitute paralogous groups. A group almost equal in size to the response regulators is the LysR-family of transcriptional regulators. Other large groups of transcriptional regulators are also present. Cross genome comparisons of paralogous groups In addition to using paralogous groups for intra-genomic analyses, the groups were also used in cross genome comparisons (see Table 9 ). The sizes of selected sequence related groups are shown for three bacteria, the closely related enterics E. coli and Salmonella enterica serovar Typhimurium and the more distantly related organism Bacillus subtilis . The sizes of the groups in the closely related bacteria are similar, whereas there are differences in relation to B. subtilis , a gram positive soil organism. For instance, the largest E. coli enzyme group containing Fe-S-binding oxidoreductases was represented by only one homolog in the B. subtilis genome. However, B. subtilis encodes for 31 oxidoreductases homologous to the group of 18 NAD(P)-binding oxidoreductases of E. coli . The number of homologous sugar kinases, respiratory reductase subunits, and nucleoside diphosphate (Nudix) hydrolases appeared overall to be lower in B. subtilis . Table 9 Cross genome comparisons of enzyme groups. Ec 1 So 2 Bs 3 Group function 20 18 1 oxidoreductase, Fe-S-binding 18 14 31 oxidoreductase, NAD(P)-binding 18 13 10 oxidoreductase 4 , NAD(P)-binding 13 13 11 aldehyde dehydrogenase, NAD(P)-binding 13 11 13 oxidoreductase, FAD/NAD(P)-binding 11 16 6 sugar kinase 10 13 5 respiratory reductase, alpha subunit 9 8 8 aldo-keto reductase, NAD(P)-binding 8 7 5 phosphatase 8 8 2 nucleoside diphosphate (Nudix) hydrolase 1 No. proteins in Escherichia coli paralogous group 2 No. sequence matches for E. coli paralogous group in Salmonella typhimurium LT2 3 No. sequence matches for E. coli paralogous group in Bacillus subtilis 4 Contains GroES-like structural domain (SCOP sf50129). Discussion Protein modules vs. protein domains We have attempted to enumerate fused genes in E. coli in earlier work. Although we recognized the difference between independent proteins with complete function, called modules [ 9 ], as opposed to parts of proteins such as motifs and domains, we were not successful in our most recent effort in collecting only complete proteins to the exclusion of domains [ 15 , 27 ]. In earlier work we depended on size as a criterion to eliminate domains, but we know now some domains are large and overlap the lower range of sizes of independent proteins [ 28 ]. We also limited our previous studies to alignments between E. coli proteins. In this report we make use of information from 50 genomes to detect complete and independent protein homologs for the components of the fused E. coli proteins. The need to make use of additional genome sequences is supported by the fact that only 57% of the modules in fused E. coli proteins had unfused homologs within the E. coli genome while 90% had homologs among the 50 genomes. This result suggests that additional fused E. coli proteins might be detected in the future with more available genome sequences. The overall effect of changing the methodology has been to reduce the numbers of multimodular proteins identified in E. coli K-12. As a result of reducing the number of fused proteins, the number of paralogous protein groups was also reduced. The grouping process is based on similarity between the sequences hence many parts of the same proteins remained together in the new groups. The effects of multimodular proteins on annotation of genes For many years we have known that the E. coli contained fused genes and groups of sequence-similar proteins [ 19 ]. Today with the sequence of the entire genome and that of many other microbial genomes, we can quantify the gene fusions in E. coli and apply this information to generate paralogous groups. Even though we find that multimodular proteins are a minor fraction, 2.5%, of the proteins in E. coli K-12 MG1655, they significantly affect the annotation of related genes and the ability to define paralogous genes within a genome. Examples of the types of errors arising in the annotation of fused proteins are shown in Figure 3a . The multimodular protein ThrA (gi1786183) encodes an aspartokinase in the N-terminal module (aa 1–461) and a homoserine dehydrogenase in the C-terminal module (aa 464–820). A sequence similar protein from Lactococcus lactis , gi12723655, aligning only to the N-terminal module is erroneously annotated as having both aspartokinase and homoserine dehydrogenase activities. The correct annotation should be aspartokinase. In a second example, a protein from Bacillus halodurans , gi10174117, aligns to the aspartokinase module of ThrA but is described as homoserine dehydrogenase. The correct assignment should be aspartokinase. Figure 3 Annotation and composition of multimodular proteins. (a) Annotation is complicated by multimodular proteins. An E. coli protein (gi1786183) contains two modules, an N-terminal aspartokinase and a C-terminal homoserine dehydrogenase. Two single module proteins from L. lactis and B. halodurans (gi12723655 and gi10174117) align to the N-terminal aspartokinase module of the E. coli protein. Based on the sequence alignments, both of these proteins should be annotated as aspartokinases. However, errors are seen in the annotation of the L. lactis and B. halodurans proteins stemming from transfer of functions between multimodular proteins and partially aligned sequences without taking into account the alignment regions. (b) Different combinations of modules are seen in multimodular proteins of different organisms. While aspartokinase is fused to homoserine dehydrogenase in E. coli it is fused to DAP decarboxylase in X. fastidiosa . In both organisms the fusions are between enzymes of metabolic pathways, threonine biosynthesis for E. coli and lysine biosynthesis in X. fastidiosa . As shown in Figure 3b , different genes are sometimes fused to the same gene in different organisms. In E. coli an aspartokinase is fused to a homoserine dehydrogenase (gi1766183), while in Xylella fastidiosa , an aspartokinase is fused to a diaminopimelate decarboxylase (gi9106073). One needs to be alert to partial alignments. In this case, the annotation is correct for both activities of the Xylella protein, although the description does not follow the convention of stating the N-terminal activity first, raising the potential for misidentification of the activity of a partial homolog. Generality of gene fusions and remedies The details of gene duplication and divergence and of gene fusions have followed different courses in separate lines of descent of bacteria. The fusions of different gene partners to aspartokinase in E. coli and X. fastidiosa connected proteins acting in the same pathway. However, the pathways are different for the two organisms, threonine biosynthesis for E. coli and lysine biosynthesis in X. fastidiosa . Fusions of genes in a pathway have long been known and also the fusions of different genes in different organisms. In the tryptophan biosynthesis pathway of E. coli both the trp C gene (formerly trpC(F)) and the trp D gene (formerly trp G(D)) encode two enzymes as indicated in their former names. In contrast Rhizobium meliloti has a fusion between the trp E and trp G genes, trp E(G) [ 29 ]. Such differences not taken into account in annotation have generated errors in assignment of activities in some of the tryptophan synthesis proteins in a number of organisms. The variability in gene fusions among bacteria means that definition of multimodular proteins cannot be transferred from one organism to another, but must be worked out by analyzing the partial homology patterns with smaller independent proteins found in other organisms. To promote awareness of fused proteins, databases should list such proteins with their separate component activities and the approximate locations of these; either by start and end residues, or by module location (N-terminal, C-terminal, or Middle for proteins with >2 modules). Such a format has been implemented in GenProtEC [ 16 ]. When analyzing protein sequence alignments, one should make use of information on the alignment lengths and on the percent of each sequence that is involved in the alignment. Such information may hold clues to detecting fused proteins. Properties of paralogous groups of E. coli Groups of unimodular E. coli proteins similar in sequence vary in size from two (simple pairs) up to 92 members (Table 4 ). From pairs to groups of 8, the number of paralogous groups follows a power law. Above size 8, most sizes are represented by just one or two groups. For the smallest groups, two to four members, the degree of sequence similarity (PAM scores) tend to range widely (Figure 4 ). As the groups are larger, a clear distribution around PAM 150 emerges. Perhaps the larger groups are ones whose success is reflected in many duplication events over time with a retained function if the sequence drift is held to the range 100 to 200 PAM units. It appears that choosing 200 PAM as the upper ceiling has not eliminated an important number of groups with highly diverged members. Also, the broad range of degree of relatedness among members of paralogous groups (Table 5 , Figure 4 ) suggests that some types of proteins diverge further than others. The cluser around PAM 150 is populated by large successful paralogous groups, some of which are closely related in catalytic function while others have diverged to more distantly related activities. Figure 4 Sequence similarity of E. coli paralogous protein groups versus the group size. Protein sequences were aligned by the AllAllDb program of Darwin. Multimodular proteins were separated into modules (independent functional units) prior to the Darwin analysis. Alignments with similarities of ≤ 200 PAM units over 83 amino acids and where >45% of the length of both proteins in the pair were aligned were used to generate protein groups. The average PAM distances for the protein pairs in the smaller groups having 2–4 members (▲) and in the larger groups of ≥ 5 members (△) are shown. The smaller groups are more abundant and show a wide range of similarities. The larger groups appear to be more divergent with higher average PAM values clustering around PAM 150. The largest paralogous groups are transporters and regulators (Tables 7 & 8 ). Paralogous groups of enzymes tend to be smaller (Table 6 ). The largest enzyme classes tend to be oxidoreductases or subunits of oxidoreductases, and the relationships among members of these groups point in the direction of shared binding capacities accounting for the sequence relatedness, e.g. Fe-S clusters. In earlier work we found that some sequence related enzymes are alike in their ligand-binding characteristics, others are alike in mechanism of the catalytic action [ 30 ]. Both types of shared properties are seen in Table 6 . The ABC transporters have been a successful formula in bacterial evolution. The ATP-binding subunits maintain detectable sequence similarity. More divergent are the membrane subunits, and least similar are the periplasmic ligand-binding subunits, perhaps understandably divergent as their binding specificities for each transported compound will differ with the properties of the compounds [ 31 ]. One of the groups of periplasmic binding components also contains sequence related transcriptional regulators of the GalR/LacI family, agreeing with previous reports [ 25 , 26 ]. The major difference between these two functions is the presence or absence of a DNA-binding domain. According to Fukami-Kobayashi et al. [ 26 ], the regulators in this group are believed to have arisen by the fusion of a DNA binding domain to an ancestral periplasmic binding protein. The substrate specificity is thought to have evolved subsequently. Only a few of the transporters and regulators in this group bind the same substrates; galactose (MglB and GalR), ribose (RbsR and RbsB) and xylose (XylF and XylR). Among the regulator groups (Table 8 ), the class of two-component regulators is large. The two major activities of sensory histidine kinase and response regulators separate by the rules for grouping modules, but their known internal structures do not emerge. Many other groups are different kinds of transcriptional regulators. Another example of different functions related by sequence has been reported for a class of repressors and kinases, the ROK family [ 32 ]. In this case the two different functions are sequence related via their sugar-binding domains and differ in their DNA-binding or kinase activity. Cross genome comparisons Examining comparable paralogous groups among organisms may provide insight into functional and physiological differences among organisms. Illustration of the possibilities is shown in Table 9 where the sizes of comparable paralogous groups are shown for the closely related enteric bacteria E. coli and S. enterica serovar Typhimurium and the distant gram positive soil organism B. subtilis . Major difference is seen for one category of oxidoreductases. The largest enzyme group in E. coli contains 20 FeS-binding proteins whereas the B. subtilis genome has only one protein of this type. Members of the E. coli group include subunits of formate dehydrogenases, hydrogenases 3 and 4, DMSO reductase, and a NADH dehydrogenase. The presence of elements of the formate hydrogen lyase system and of the DMSO reductase in E. coli but not B. subtilis illustrates information on metabolic differences that emerges from such cross-genome comparisons. B. subtilis does not have the diverse anaerobic respiratory capability of E. coli and S. enterica . Duplication and divergence of this common ancestral gene seems to have taken a different course in the two bacterial lineages. In another example, B. subtilis has made use of one enzyme type to a greater extent than the two enteric organisms. The number of one of the types of NAD(P)-binding oxidoreductases is much larger in B. subtilis (31 proteins) than in the enterics (18 proteins). The B. subtilis enzymes in this group are fatty acid biosynthesis enzymes, agreeing with the known fact that this organism synthesizes a greater variety of fatty acids and has dedicated more of its proteome towards diversifying its fatty acid biosynthetic capabilities [ 33 , 34 ]. Thus sequence similar groups may be used in comparative analysis between genomes, highlighting areas where genetic resources have been expanded, pointing up metabolic differences between organisms. Conclusion • Proteins encoded by fused genes, multimodular proteins, require special attention in genome analysis. Such multimodular proteins contain two or more functional components that are located at separate parts of the protein and that may exist as independent proteins in other genomes. Annotation of the multimodular proteins should include the separate functions and their corresponding locations in the gene product. This will improve transfer of function between the fused proteins and sequences matching their entire length or only the length of one of their module components. Current annotation errors involving fused genes can be remedied by introducing this approach. • The identification of multimodular proteins in E. coli was improved by making use of sequence information from 50 genomes to detect alignments between the fused proteins and smaller, un-fused homologs corresponding to the component modules. The more common multidomain proteins, proteins containing fused sequence domains or motifs that together make up one overall function, were not detected as multimodular proteins by this approach. As a result the current number of fused E. coli proteins was reduced to 109 proteins with 8 containing three modules and 101 containing two modules. The multimodular E. coli proteins consist mainly of enzymes, regulators and transport proteins. Their component modules are often not related by sequence but many are related in that they function in a common pathway or cell role. Components of fused genes appear to vary from genome to genome hence complicating their detection and function assignment. • Multimodular proteins are different from multifunctional proteins in that the latter catalyze more than one reaction in the same region of the protein. • The generation of paralogous or sequence related groups is improved when the modules of multimodular proteins are separated and treated as independent proteins for the grouping process. 490 groups of sequence related E. coli proteins ranging in size from 2 to 92 were generated from the new module data. The smaller groups range widely in degree of relatedness while the larger groups have diverged from one another to about the same extent. Transport proteins and regulatory proteins were found in the larger groups while enzyme groups tended to have fewer members. • Over half of the E. coli proteins belong to paralogous groups, reflecting the prominent role of duplication and divergence in the evolution of the genome. The number and sizes of paralogous groups reflect the distinctiveness of the organisms and they can be used in cross genome comparisons. Methods Sequence sources Protein coding sequences were obtained from GenBank and included the following genomes: Aquifex aeolicus , (AE000657); Archaeoglobus fulgidus , (AE000782); Aeropyrum pernix , (BA000002); Agrobacterium tumefaciens , (AE007869/AE007870); Borrelia burgdorferi , (AE000783); Bacillus halodurans , (BA000004); Bacillus subtilis , (AL009126); Buchnera sp. APS, (BA000003); Campylobacter jejuni , (AL111168); Clostridium acetobutylicum , (AE001437); Chlamydia muridarum , (AE002160); Chlamydophila pneumoniae CWL029, (AE001363); Deinococcus radiodurans , (AE000513/AE001823); Escherichia coli K-12, (U00096); Escherichia coli O157:H7 EDL933, (AE005174); Escherichia coli O157:H7, (BA000007); Haemophilus influenzae , (L42023); Helicobacter pylori 26695, (AE000511); Halobacterium sp. NRC-1, (AE004437); Lactococcus lactis subsp. lactis , (AE005176); Mycobacterium leprae , (AL450380); Mycoplasma genitalium , (L43967); Mycobacterium tuberculosis H37Rv, (AL123456); Methanococcus jannaschii , (LL77117); Mesorhizobium loti , (BA000012); Mycoplasma pneumoniae , (U00089); Mycoplasma pulmonis , (AL445566); Methanobacterium thermoautotrophicum , (AE000666); Neisseria meningitidis MC58, (AE002098); Pseudomonas aeruginosa , (AE004091); Pyrococcus horikoshii , (BA000001); Pasteurella multocida , (AE004439); Pyrococcus abyssi , (AL096836); Rickettsia prowazekii , (AJ235269); Salmonella enterica subsp. enterica serovar Typhi, (NC_003198); Salmonella typhimurium LT2, (AE006468); Shewanella oneidensis MR-1, (NC004347); Sinorhizobium meliloti , (AL591688); Staphylococcus aureus subsp. aureus Mu50, (BA000017); Streptococcus pneumoniae TIGR4, (AE005672); Streptococcus pyogenes M1 GAS, (AE004092); Sulfolobus solfataricus , (AE006641); Synechocystis PCC6803, (AB001339); Thermoplasma volcanium , (BA000011); Thermotoga maritima , (AE000512); Treponema pallidum , (AE000520); Ureaplasma urealyticum , (AF222894); Vibrio cholerae , (AE003852/EC003853); Xylella fastidiosa 9a5c, (AE003849); Yersinia pestis , (AL590842). Analysis of protein sequence similarities Pairwise sequence alignments and scores were generated using the AllAllDb program of Darwin (Data Analysis and Retrieval With Indexed Nucleotide/peptide sequence package), version 2.0, developed at the ETHZ in Zurich [ 35 ]. Maximum likelihood alignments are generated with an initial global alignment by dynamic programming [ 36 - 38 ] followed by dynamic local alignments [ 39 ]. A single scoring matrix is used for these steps. After the initial alignment, the scoring matrix is adjusted to fit the approximate distance between each protein pair to produce the minimum PAM value. PAM units are defined as the numbers of point mutations per 100 residues [ 37 ]. The final report includes PAM distances and variances. For the work reported here, sequence pairs were collected that had alignment lengths of at least 83 amino acids and distances of 200 PAM units or less. We chose the length requirement of 83 residues as it improves the significance of the sequence alignments for the more distantly related protein pairs [ 40 ]. The requirement for at least 83 residues also avoids a class of commonly occurring protein domains smaller than 83 residues that appear widely in many otherwise unrelated proteins (such as small binding sites for a type of substrate, cofactor, or regulator). In addition for this study we removed proteins directly involved in horizontal gene transfer (IS proteins, transposases, and known prophage components) from the dataset. Identification of multimodular proteins Proteins encoded by fused genes were identified from the E. coli literature and from unequal sequence alignments. The literature was searched for E. coli proteins with more than one function encoded at separate parts of the protein. The locations of the alignment regions in the proteins were analyzed for orthologous and paralogous protein pairs. We identified proteins with two or more non-overlapping alignment regions where each region aligned separately to smaller homologs. Figure 1a illustrates the alignment of two unfused proteins with parts of a fused protein. Multimodular proteins so identified were separated into independent modules. Using the pairwise data, start and end positions of the modules were estimated from the many alignment regions and were set to cover as much of the sequence as possible, not only the most conserved regions of all the alignments. No overlap was allowed between any adjacent modules. Generation of internal sequence similar groups (paralogs) The sum of the separated modules from the multimodular proteins and the naturally occurring unimodular proteins of E. coli were aligned against themselves. Protein pairs aligning with >45% of the length of the peptides were used in a transitive grouping process as previously described [ 15 ]. The transitive nature of the process ensures sequence similarity to at least one member of the group and does not require all members of the group to have detectable similarity to one another. This type of clustering allows for more divergent sequences to be grouped. The restriction of PAM value to no more than 200 prevents groups from expanding beyond significant similarity. Authors' contributions MS designed the study, performed the sequence analysis, and participated in the data analysis and in writing the manuscript. MR participated in the data analysis and in writing the manuscript. Supplementary Material Additional File 1 Multimodular E. coli proteins. The table contains a complete list of the multimodular proteins in E. coli . Each module is described by its Gene name, Module Id, Module Start and End positions, Gene type, and Module Product. Click here for file Additional File 2 E. coli paralogous groups and their members. The table contains a complete list of the paralogous protein groups in E. coli . The members of the 409 paralogous groups are indicated by their Group Membership, Module Id, Module Start and End Position, Module Product. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC555942.xml

535898

Verbal autopsy of 80,000 adult deaths in Tamilnadu, South India

Background Registration of the fact of death is almost complete in the city of Chennai and not so in the rural Villupuram district in Tamilnadu, India. The cause of death is often inadequately recorded on the death certificate in developing countries like India. A special verbal autopsy (VA) study of 48 000 adult (aged ≥ 25 yrs) deaths in the city of Chennai (urban) during 1995–97 and 32 000 in rural Villupuram during 1997–98 was conducted to arrive at the probable underlying cause of death to estimate cause specific mortality. Methods A ten day training on writing verbal autopsy (VA) report for adult deaths was given to non-medical graduates with at least 15 years of formal education. They interviewed surviving spouse/close associates of the deceased to write a verbal autopsy report in local language (Tamil) on the complaints, symptoms, signs, duration and treatment details of illness prior to death. Each report was reviewed centrally by two physicians independently. Random re-interviewing of 5% of the VA reports was done to check the reliability and reproducibility of the VA report. The validity of VA diagnosis was assessed only for cancer deaths. Results Verbal autopsy reduced the proportion of deaths attributed to unspecified and unknown causes from 54% to 23% (p < 0.0001) in urban and from 41% to 26% (p < 0.0001) in rural areas in Tamilnadu for adult deaths (≥ 25). The sensitivity of VA to identify cancer was 95% in the age group 25–69. Conclusion A ten day training programme to write verbal autopsy report with adequate feed back sessions and random sampling of 5% of the verbal autopsy reports for re-interview worked very well in Tamilnadu, to arrive at the probable underlying cause of death reliably for deaths in early adult life or middle age (25–69 years) and less reliably for older ages (70+). Thus VA is practicable for deaths in early adult life or middle age and is of more limited value in old age.

Background In developed countries, data on disease-specific mortality by age are readily available from national vital registration. In developing countries, where 80% of the world's deaths occur, estimation of cause of death is more difficult because the levels of coverage of vital registration and reliability of cause of death stated on the death certificate are generally low (especially in rural areas). A reliable assessment of disease-specific mortality rates is not yet possible in many parts of India, either because the underlying cause of the terminal illness was never known or because the relevant information was not recorded. For legal purposes death records do usually subdivide the causes of death into medical and non-medical (external) causes. But once non-medical causes have been excluded, specification of the underlying cause of a death from disease may be inaccurate, misclassified or missing for about 50% of adult deaths. For example, in Chennai, Tamilnadu, (south India) about half of those who died at home soon after the diagnosis of cancer (and whose deaths were therefore, in almost all cases, likely to have been caused by their cancer) do not have cancer mentioned on their death certificate [ 1 ], and for other diseases the problems might well be even worse. Elsewhere, 'Verbal Autopsy' i.e., 'systematic retrospective inquiry of family members about the symptoms and signs of illness prior to death' has been used to help determine the underlying medical cause of death, particularly in childhood [ 2 , 3 ]. For childhood deaths in populations that are not covered by adequate medical services such "verbal autopsies" are now of established value in helping to classify the broad patterns of mortality. Verbal autopsies have also been used to assess the medical causes of maternal deaths [ 4 - 7 ]. Although in India there are about as many deaths in middle age as in childhood, there is less experience with verbal autopsies of adult deaths. A special study on 'verbal autopsy' of adult deaths was conducted in urban and rural areas in the state of Tamil Nadu, south India during 1998–2000. The aims of the study were (a) to develop a verbal autopsy instrument and test its utility to determine the underlying cause of death and (b) to estimate cause specific mortality using underlying cause of death arrived based on verbal autopsy reports. Now we report the type of training given to the field interviewers to interview and write the verbal autopsy report for adult deaths, the procedure followed to arrive at the probable underlying cause of death, the accuracy of the instrument developed and change in the proportions of cause of death based on the underlying cause of death arrived by reviewing 80 000 verbal autopsy reports. Methods A: Training of field interviewers to write a 'Verbal autopsy' report Male non-medical graduates with at least 15 years of formal education were selected. A ten days training was given on verbal autopsy interview techniques and writing verbal autopsy reports. There are four steps in training; 1). introduction to anatomy, signs and symptoms of various diseases, 2). mock interviews, 3). hands-on-training on writing verbal autopsy reports and 4). feed back session. Step 1 This consisted of a basic three day introduction to anatomy, collecting data on history of past illness (refer Appendix I in Additional file 1 ), using symptoms/signs checklist of various diseases (refer Appendix II in Additional file 1 ) and to interview the surviving spouse/close associates or relatives of the deceased, the other members of the community such as neighbours to get data on train of events or circumstances preceding the death. Reports are to include complaints, symptoms, signs, duration of illness and treatment details of the illness prior to death. The following data are to be ascertained (for all deaths due to medical causes) from the respondent to write the verbal autopsy narrative report: • onset of illness prior to death: sudden or gradual • major symptom(s) and associated symptom(s) – in chronological order • If a symptom was present it was used as a filter to define what questions to be asked. For example, the filter symptom for heart attack was chest pain and the associated symptoms were breathlessness, sweating, vomiting and pain in the retrosternal area radiating to hand, shoulder, back etc. Cough for more than 4 weeks was a filter for lung cancer and tuberculosis. For each symptom, the duration should be recorded. Details of additional symptoms are built into the narrative in chronological order, by prompting, if necessary. • progress of the illness • any treatment received : Yes/No • If yes, type of treatment received • details of hospitalization prior to death: ○ name of the hospital (e.g. tuberculosis hospital, cancer hospital, coronary care unit etc), ○ duration of hospitalization, ○ whether discharged from the hospital against medical advice or not. ○ Status at the time of discharge from the hospital: alive/dead • history of similar episodes and treatment(s) given • abstract information related to the illness prior to death from the investigation reports done for any illness close to the time of death (within 6 months prior to the death) / hospital discharge summary etc, if available • If a death certificate is available, copy the cause of death given on the death certificate (In the Tamilnadu study death certificates were available for only 20% of total deaths). • While recording history of adults with long standing illness, the description should include details that occurred in the month preceding the death, with other information recorded in the past history section (Appendix I in Additional file 1 ). • For deaths that occur during pregnancy, delivery, or within six weeks of delivery: use Appendix II (A and B) in Additional file 1 If the respondent is able to give the major symptoms and circumstances leading to death, then additional probing questions are asked about the associated symptoms using the symptoms/signs checklist (Appendix II(A & B) in Additional file 1 ) If the respondent is not able to give sufficient information on the symptoms of the illness prior to death or have difficulty in remembering any major symptom, then get necessary information to rule out non-medical causes of death. When the interviewer is sure that the death was not due to unnatural cause, the following procedure is used to collect necessary data on the symptom. ○ read out the filter symptom/sign of each module in the symptom/sign checklist ○ check responses to each, and note down positive responses ○ Where there is a positive response, additional details on that symptom and associated symptoms, if any, should be obtained. Thus, the methodology of collecting data in the open format using 'symptoms/signs checklist' is an interactive process, with the respondent taking the lead in providing the information, and the interviewer prompting where necessary for more details. The Field Interviewer gathers as much information as possible on the underlying cause of death from the respondent. It is imperative to get a logical and complete history of symptoms, signs, events, investigations and treatment, so that the medical reviewer gets sufficient information to assign a probable specific underlying cause of death. Step 2 In the following two days, mock interviews were organized to illustrate techniques of probing a respondent to get the required information on cause of death as well as how to write the verbal autopsy report in local language (Tamil) in Appendix I in Additional file 1 as stated by the respondent. Step 3 The third component of training included three days of hands on verbal autopsy training in the field. To limit distress over the terminal event, the field visit was carried out at least six months after death. Name of the deceased, father's name (if the deceased was a male) or spouse name (if the deceased was a female), age, gender, informant's name and address of the deceased at the time of death were given to field interviewers to locate the house of the deceased. The Field Interviewers carry Appendix I and II (symptoms/signs checklist) in Additional file 1 to the field. They were blind to the cause of death stated on the death certificate. The Field Interviewer located the house of the deceased based on the data given to him. He introduced himself to the respondent and began the interview. Each one completed twenty reports which were reviewed and feedback was provided two days after completion of field work to maximize quality of writing the verbal autopsy report. Step 4 The final component of training was feed back session for 2 days. This session involved teaching them how to include essential information in report writing. The feedback session mainly focused discussion on reports which did not have a specified underlying cause of death and reports with minimal information to arrive at the probable underlying cause of death; for example, a report may say that a person had a stroke ten days ago but did not specify the type of onset (sudden or gradual, whether the person was conscious or unconscious, had difficulty in speaking or not, which parts of the body may have been affected etc.) or a report may say that the deceased had fever for ten days and died. It did not give details about the fever and other associated symptoms if any. B: Verbal autopsy of 80 000 adult deaths (≥ 25 years at the time of death) in Tamilnadu, South India This special verbal autopsy study was carried out in two areas in Tamilnadu. The Chennai city (urban) with a population of 4.2 million, and the Villupuram district (rural), with a population of 2.5 million were chosen for this study. We have successfully traced 48 000 adult deaths (≥ 25 years at the time of death) in urban area and 32 000 adult deaths in rural Villupuram district and reviewed 80 000 verbal autopsy reports to arrive at the probable underlying cause of death. Mortality data in urban area (Chennai) Information on deaths that occur in Chennai has been maintained manually by trained staff in Chennai Vital Statistics Department (VSD). The following data on deaths that occurred in Chennai during 1995 to 1997 were collected from the death registers in the Vital Statistics Department: deceased name, age, gender, marital status, father/spouse name, informant's name, occupation, place of death, address at the time of death, date of death and recorded cause of death (immediate, underlying and/or contributory). 72,000 deaths occurred during the study period of 1995–97. Of 72 000 found, 5000 deaths were attributed to external causes (unintentional injuries, suicide or homicide) in the death certificate, and were excluded. Of the remaining 67 000 deaths attributed in the VSD to medical causes, 48 000 of the households were successfully visited during 1998–99 to try to assign cause of death by verbal autopsy. Mortality data in rural area (Villupuram district) All formal and informal village records were to be sought to identify all deaths at any age during 1997–98. 41,000 such records were identified and 39 000 of the households were successfully visited during 1999–2000 to try to assign cause of death by verbal autopsy. Of 39 000 deaths, 7000 were before age 25. Feed back sessions and re-interview Feed back sessions were organized regularly throughout the study period to improve the quality of the verbal autopsy reports and 5% of the field visit reports were validated by re-interview one week after completion of the main interview, and blind to its results. This re-interviewing was done by one or other of two special interviewers because knowledge that a resurvey might well take place would ensure reliably motivated fieldwork at the initial survey, and also to check whether there were any systematic defects in the technique of any of the field workers: none were found. The underlying cause of death arrived based on re-interview data was not substantially different from the one arrived based on main interview data. Arriving at underlying cause of death All verbal autopsy reports were centrally reviewed by two physicians independently in order to arrive at "probable underlying cause of death". Each made a diagnosis based on signs, symptoms and sequence of events prior to death given in the verbal autopsy report, which were then coded according to the 9 th International Classification of Diseases, Injuries and Causes of Death [ 8 ]. The same 2 physicians reviewed all the 80,000 verbal autopsy reports. The discrepancies in the underlying causes of death were noted in 5% of verbal autopsy reports. These were discussed and resolved. The disagreement between 2 physicians in arriving at underlying cause of death was noted before classifying causes of death into broad groups. For example, 'Pneumonia' and 'Lower respiratory infection' were grouped under 'Infection'. According to one physician the underlying cause of death was pneumonia and for another physician it was lower respiratory infection. Results Urban study In Chennai city, the study was done in 1998–99 and the verbal autopsy reports were reviewed for 27 726 male deaths and 20 631 female deaths. Table 1 shows about 1100 (M:683, F:456) deaths due to medical causes were reassigned to external causes based on verbal autopsy reports. Deaths from unspecified medical causes and unknown causes decreased from 54% to 23% (p < 0.0001). Table 1 Cause of death from Vital Statistics Department* and based on Verbal Autopsy of 48 000 adult deaths (aged ≥ 25) in Chennai (urban), south India: 1995–97 Causes of death (ICD9 codes) Cause of death in VSD Cause of death based on Verbal Autopsy M (%) F (%) M (%) F (%) Vascular disease (390–415, 418–459) 8319 (30) 5168 (25) 11056 (41) 7435 (37) Respiratory tuberculosis (TB) (011, 012, 018) 1399 (5) 372 (2) 2231 (8) 575 (3) Other respiratory diseases (416, 417, 460–519) 1088 (4) 596 (3) 1597 (6) 855 (4) Neoplasm (140–239) 1163 (4) 1002 (5) 2344 (9) 1999 (10) Infection except respiratory & TB (rest of 1–139, 279.8 [HIV], 320-6, 590, 680-6) 584 (2) 303 (2) 1034 (4) 618 (3) Unspecified medical causes (780-9, 797-9) 12291 (44) 115 11 (56) 4367 (16) 5889 (29) Other specified medical causes 1899 (7) 1045 (5) 4414 (16) 2804 (14) No cause given in VSD (hence probably medical) 983 (4) 634 (3) Nil Nil Total deaths – medical 27 726 20 631 27 043 20 175 Re-assigned by VA to external causes *Excluded from the study 683 456 Total deaths (medical causes+external causes) 27 726 20 631 27 726 20 631 *Deaths(M: 3644; F:1644) that were assigned by the Vital Statistics Department(VSD) to non-medical causes were excluded from the study Rural study In Villupuram district, verbal autopsy report was written for all deaths i.e., deaths due to medical and external causes. So verbal autopsy reports of 19 294 male deaths and 12 494 female deaths were reviewed. Deaths from unspecified medical causes and unknown causes decreased from 41% to 26% (p < 0.0001) (Table 2 ). Table 2 Cause of death from various local records in Villupuram district and based on Verbal Autopsy of 32 000 adult deaths (aged ≥ 25) in Villupuram (rural), south India: 1997–98 Causes of death (ICD9 codes) Cause of death in local records Cause of death based on Verbal Autopsy M (%) F (%) M (%) F (%) Vascular disease (390–415, 418–459) 3351 (20.3) 1614 (14.4) 3928 (24.6) 2404 (22.0) Respiratory tuberculosis (TB) (011, 012, 018) 1659 (10.1) 686 (6.1) 1841 (11.5) 671 (6.1) Other respiratory diseases (416, 417, 460–519) 717 (4.4) 471 (4.2) 1044 (6.5) 728 (6.6) Neoplasm (140–239) 415 (2.5) 594 (5.3) 488 (3.1) 664 (6.1) Infection except respiratory & TB (rest of 1–139, 279.8 [HIV], 320-6, 590, 680-6) 1818 (11.0) 1584 (14.1) 1954 (12.2) 1411 (12.9) Unspecified medical causes (780-9, 797-9) 5829 (35.4) 4565 (40.7) 4173 (26.1) 2737 (25.0) Other specified medical causes 2237 (13.6) 1346 (12.0) 2570 (16.1) 2334 (21.3) No cause given (hence probably medical) 451 (2.7) 343 (3.1) Nil Nil Total deaths – medical 16 477 11 203 15 998 10 949 External causes 2817 1291 3296 1545 Total deaths (medical causes+external causes) 19 294 12 494 19 294 12 494 Validity of verbal autopsy tool The cause of death stated on the death certificate is often inaccurate. Studies, which have been undertaken around the world, show substantial difference (10–40%) between the clinical diagnosis or clinical cause of death and postmortem findings [ 9 - 12 ] and many of the completed death certificates failed to provide relevant information to allow adequate ICD-10 coding [ 13 ]. In India, individuals whose deaths might have been due to external causes are often subjected to postmortem examination, but others are not. So it is not possible to compare (clinical diagnosis of) medical causes of death against postmortem findings in India. Gajalakshmi et al [ 1 ] had done a study in Chennai to determine the sensitivity of the death certificate to identify cancer by comparing the cause of death stated on the death certificate with the morbidity date base of Chennai population-based cancer registry. It was found that the sensitivity of the death certificate to identify cancer as the underlying cause of death was 57%. In Chennai, about 75–80% of cancer patients attend health care facilities at late stage of the disease; about half of those who died at home soon after the diagnosis of cancer (and whose deaths were therefore, in almost all cases, likely to have been caused by their cancer) did not have cancer mentioned on their death certificates. Hence verbal autopsy tool was developed to determine specific cause of death, to compute cause specific death rates. Where a cause recorded on the death certificate in the VSD differed from the underlying cause assigned by the VA, there was often no absolute way of knowing which was correct (where the assigned cause by a medical doctor on the death certificate lacked detail, the VA may well be more reliable, and vice-versa) except for cancer deaths which could be verified with the Chennai cancer registry records. Hence, the validity of VA diagnosis was assessed only for cancer deaths(ICD 9:140–208) by comparing with the stated cause of death in the VSD records and verifying with the Chennai cancer registry records and hospital medical records (only for cancer diagnosis). Chennai Population-Based Cancer Registry is a demographic registry in the network of the Indian Council of Medical Research, Govt. of India and has been functioning since 1982 at the Cancer Institute(WIA), Chennai. Cancer is not a notifiable disease in India. Hence registration has been done by active method. Cancer patients attending the Govt. hospitals are interviewed to collect data on age, sex, address, duration of stay in Chennai city, marital status, mother tongue and educational level. Interviews are done at the houses for those who have been missed by the registry staff during their (Govt.) hospital visit. The clinical data, such as date of cancer diagnosis, method of diagnosis, site of cancer, any spread of the disease, histology, treatment details and status (alive or dead, if dead at the hospital, then, date and cause of death) for all registered patients are abstracted from the hospital medical records. All data on cancer patients attending the private hospitals are abstracted from the hospital records. The mortality data available at the Vital Statistics Department are linked with the Chennai Cancer Registry data base. Therefore the cause of death arrived based on verbal autopsy report was compared with the hospital data on cancer patients available in the Chennai Cancer Registry data base. Table 3 shows that 3053 deaths were identified as being due to cancer by VA. Review of VSD records revealed that 1435 of 3053 deaths as being due to non-cancer causes (majority of deaths were attributed to ill-defined/unknown followed by vascular causes) and 1618 (of 3053 deaths) as being due to cancer. Since the sensitivity of death certificate to identify cancer as underlying cause of death is only 57% in Chennai [ 1 ], all deaths at ages 25–69 included in the present study were verified with Chennai population-based cancer registry records. Table 4 shows that out of 3053 deaths identified by VA as cancer underlying cause of death, 2765 deaths matched with Chennai population-based cancer registry data base and 288 deaths did not. The cancer deaths identified by VSD records and not by VA (n = 107) (Table 3 ) matched with Chennai population-based cancer registry data base. Thus 288 cancer deaths, identified by VA, were not registered in the Chennai population-based cancer registry [ 14 , 15 ]. These were missed by the Chennai population-based cancer registry, both in the routine morbidity and mortality data registration process. We were successful in identifying all 288 cancer deaths, not available in the Chennai population-based cancer registry, in the medical records of the hospitals located in Chennai city. Thus all 3053 cancer deaths identified by VA were confirmed by linking with Chennai Cancer Registry records and hospital medical records. So there were no false positive cancer deaths recorded by VA. The sensitivity of VA to identify cancer was 94% (1618/1725) compared to VSD records and 96% (2765/2872) compared to Chennai population-based cancer registry in the age group 25–69 [ 15 , 16 ] and the Chennai population-based cancer registry missed 9% (288/3160) of total cancer deaths in the early adult life and middle age during the study period of three years. Table 3 Cancer (ICD 9: 140–208) deaths at ages 25–69 by Verbal autopsy (VA) and in Vital Statistics Department records (VSD) in Chennai (urban), South India VSD VA Cancer Noncancer Total Cancer 1618 1435 3053 Noncancer 107 21941 22048 Total 1725 23376 25101 Table 4 Cancer (ICD 9: 140–208) deaths at ages 25–69 by Verbal autopsy (VA) and in Chennai population-based cancer registry in Chennai (urban), South India Cancer Registry VA Registered Not registered Total Cancer 2765 288 3053 Noncancer 107 21941 22048 Total 2872 22229 25101 Discussion The Tamilnadu study on verbal autopsy [ 15 , 16 ] used university graduates since it is very expensive to send professionally trained individuals to field visits, to write verbal autopsy reports. We have found it very difficult to get female graduates willing to do field work. Hence only males were recruited for the field work. Responders of female deaths were usually males who did not hesitate to reveal the circumstances/ symptoms etc prior to death. The participation rate was 100%. The informants were given full information about the objectives of the study and the participation in this study was entirely voluntary basis. The verbal autopsy tool for adult deaths is an open narrative format uses the check list of symptoms and signs with filters to get more information on train of events or circumstances preceding the death. The sensitivity of this tool is 95% (94% compared to VSD records and 96% compared to Chennai population-based cancer registry) in the age group 25–69 during the study period. The validity of this verbal autopsy tool is influenced by the training given to the interviewers, on the immediate random checking of the 5% of interview data and reviewing of the field reports centrally by 2 physicians to arrive at the probable underlying cause of death which is better than that arrived by opinion-based algorithm [ 17 ]. There is little information or literature about validity of cause of death for adults by verbal autopsy in India. As a result of using verbal autopsy method, adult deaths (≥ 25) from unspecified and unknown causes decreased from 54% to 23% (p < 0.0001) in urban and from 41% to 26% (p < 0.0001) in rural areas in Tamilnadu. Ten day training to write verbal autopsy reports followed by constant monitoring of the submitted reports resulted in arriving at the probable underlying cause of death for most of the deaths and to compute broad classification of the underlying causes of about 90% of deaths in early adult life or middle age: in old age, however, the proportion classifiable is substantially lower. The specific causes of death arrived based on verbal autopsy reports were used to estimate death rates for Chennai city [ 16 ] and to estimate the risk of death associated with smoking for broad groups of causes of death[ 18 ]. This methodology of arriving cause of death for adult deaths by verbal autopsy is now being adopted by the Registrar General of India, Govt. of India, for nationwide use in the Sample Registration System(SRS) that consists of 6671 units (4436 rural and 2235 urban) spread across the country covering 1.1 million households and about 6 million population. The SRS is a large demographic survey of vital events occurring in a national random sample of urban and rural areas in India by the Registrar General of India (RGI) to provide annual estimates of age-specific birth and death rates at the national and state levels. This exercise on verbal autopsy for adult deaths along with the verbal autopsy questionnaire developed by the SRS collaborators for deaths that occur at ages less than 15 years and for maternal deaths is expected to yield reliable cause-specific death rates for India. Conclusion A ten day training programme to write verbal autopsy report with adequate feed back sessions and random sampling of 5% of the verbal autopsy reports for re-interview worked very well in Tamilnadu, to arrive at the probable underlying cause of death reliably for deaths in early adult life or middle age (25–69 years) and less reliably for older ages (70+). Our experience shows that the open narrative, if well written, provides adequate information for assigning probable underlying cause of death for adult deaths. Competing interests The authors declare that they have no competing interests. Authors' contributions VG and RP participated in designing the study, analysis and preparation of the report. Development of verbal autopsy tool, training on verbal autopsy method, co-ordination of field work and data management were done by VG. Verbal autopsy review by VG and TS Kanaka, physician. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 Appendix I – used to write the history of past illness and verbal autopsy report for adult deaths (25 years or older) and in Appendix II – Symptoms/signs checklist for adult deaths (≥ 25 years) is given Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC535898.xml

212701

Supersensitive Worms Reveal New Gene Functions

The past ten years saw great progress in the field of molecular genetics, as new tools gave scientists the ability to investigate entire genomes instead of just one or two genes at a time. In this paper, Ronald Plasterk and colleagues developed a systemic approach using Caenorhabditis elegans , a tiny nematode and the first animal to have its genome sequenced, to gather functional information on nearly 400 genes. Many of the systemic approaches to discovering gene function involve either measuring or deleting messenger RNA (mRNA), the molecule that helps translate genes into proteins. The method used here, called RNA interference, or RNAi, follows the deletion approach by taking advantage of a cellular process bearing the same name. In nature, RNAi is thought to be an important part of the innate defense machinery in plants and animals, protecting them from invaders like viruses by interrupting the manufacture of viral proteins. To do this, short double-stranded RNA molecules with complementary sequences to the target gene inhibit the gene's function by disabling mRNA, which effectively shuts down the gene. By mimicking this natural process to turn off selected genes, scientists can find clues to how those genes might normally function by watching what happens when they are taken out of the picture. With the fully sequenced worm genome, it is possible to create interfering RNAs for all of its 20,000 or so genes. And because worms eat bacteria—which can themselves be used to deliver interfering RNAs—worms are the perfect RNAi model organism. The researchers fed the worms RNAi-producing bacteria, then observed the effects on the worm or its offspring to infer the function of the targeted gene. As previously reported, repeating this experiment for every gene in the worm genome, yields about 10% of the worms displaying abnormalities ranging from embryonic death to uncoordinated movement, suggesting defects in genes controlling development or muscle control, respectively. Having previously identified an RNAi-hypersensitive mutant worm strain, Plasterk and his colleagues repeated the experiment in the mutants and report proposed functions for 393 previously unknown genes. The types of abnormalities observed in the short-lived mutations induced by RNAi, they say, resemble the more stable mutations seen in the collection of worm mutations cataloged by worm researchers over the years. Though the DNA alterations for many of these mutations are not yet known, researchers know roughly where they occur in the genome. And the researchers show here that they can use their RNAi experimental results along with what is known about the mutants to identify several of the sequence alterations. They also performed what is believed to be the first analysis in which independently generated large-scale RNAi results were systematically compared to see how variable such RNAi results are, and the results have implications for similar approaches not just in worms but in plants and other animals. Caenorhabditis elegans worms

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC212701.xml

544898

Case-based medical informatics

Background The "applied" nature distinguishes applied sciences from theoretical sciences. To emphasize this distinction, we begin with a general, meta-level overview of the scientific endeavor. We introduce the knowledge spectrum and four interconnected modalities of knowledge. In addition to the traditional differentiation between implicit and explicit knowledge we outline the concepts of general and individual knowledge . We connect general knowledge with the "frame problem," a fundamental issue of artificial intelligence, and individual knowledge with another important paradigm of artificial intelligence, case-based reasoning , a method of individual knowledge processing that aims at solving new problems based on the solutions to similar past problems. We outline the fundamental differences between Medical Informatics and theoretical sciences and propose that Medical Informatics research should advance individual knowledge processing (case-based reasoning) and that natural language processing research is an important step towards this goal that may have ethical implications for patient-centered health medicine. Discussion We focus on fundamental aspects of decision-making, which connect human expertise with individual knowledge processing. We continue with a knowledge spectrum perspective on biomedical knowledge and conclude that case-based reasoning is the paradigm that can advance towards personalized healthcare and that can enable the education of patients and providers. We center the discussion on formal methods of knowledge representation around the frame problem. We propose a context-dependent view on the notion of "meaning" and advocate the need for case-based reasoning research and natural language processing. In the context of memory based knowledge processing, pattern recognition, comparison and analogy-making, we conclude that while humans seem to naturally support the case-based reasoning paradigm (memory of past experiences of problem-solving and powerful case matching mechanisms), technical solutions are challenging. Finally, we discuss the major challenges for a technical solution: case record comprehensiveness, organization of information on similarity principles, development of pattern recognition and solving ethical issues. Summary Medical Informatics is an applied science that should be committed to advancing patient-centered medicine through individual knowledge processing. Case-based reasoning is the technical solution that enables a continuous individual knowledge processing and could be applied providing that challenges and ethical issues arising are addressed appropriately.

Background A meta-level view of science Our aim is to place Medical Informatics in the context of other sciences and to bring coherence in its formal education [ 1 ]. This will necessarily place the discussion at a meta-level view of science, which traditionally was the concern of philosophers. From such a general perspective, science could be defined as "the business of eliciting theories from observations in a certain context , with the hope that those theories will help to understand, predict and solve problems." Also revolving around the "business of creating theories," R. Solomonoff's ideas [ 2 ], summarized in [ 3 ], contribute to the basis of Algorithmic Information Theory (AIT) [ 4 ], a relatively new area of research initiated by A. Kolmogorov, R. Solomonoff and G. Chaitin, and regarded as the unification of Computer Science and Information Theory. According to Solomonoff's view, a scientist's theories are compressions of her observations (i.e., her experimental data). These compressions are used to explain, communicate and manage observations efficiently and, if valid, to help solving problems, understanding and predicting. Intuitively, the higher the compression achieved by the theory, the more "elegant" that theory and the higher its chances of acceptance. This very general perspective of the scientific endeavor also makes science to appear twofold: it comprises the creation of theories (i.e., theory elicitation) as well as their subsequent use in understanding, predicting and solving problems (i.e., theory application). Therefore, science seems to be driven by two opposite forces: that of creating theories, and that of applying those theories to practical applications. The four-dimensional space-time continuum we live in (i.e., our universe) forms the reality (i.e., the context) of all scientific observations. The compression of the immense complexity and dynamicity of this reality in concise "theories of everything" was already demonstrated by Zuse [ 5 , 6 ] and recently Schmidhuber [ 7 ]. These results of theoretical computer science demonstrate the power of human theory elicitation and provide important answers to old questions of science and philosophy. However, their unfeasibility when applied to practical problems, which would be equal to building computing devices capable of running precise simulations of our reality, also widens the gap between theoretical research and practical sciences. For the time being, humanity still needs to divide science and define human knowledge as a collection of individual theories elicited from scientific observations. The immense number of theories that comprise the collective human knowledge about every possible subject, as well as its extraordinary dynamics, have forced us to divide it into what we commonly refer to as knowledge domains , thereby reducing the contexts of our observations to smaller space-time continuums. The attempts to process with computers the knowledge in a domain have taught us that we need to recognize the reality of the "knowledge acquisition bottleneck" [ 8 ] and to not underestimate the importance of common-sense knowledge (see [ 9 ] and [ 10 - 13 ]). The particularities with regard to the context retention, acquisition, representation, transferability and applicability of domain knowledge, causes us to distinguish between different modalities of domain knowledge, and place them on what we refer to as the knowledge spectrum . The knowledge spectrum The knowledge spectrum (Figure 1 ) spans from a complex reality (the source of experimental data and information gathered from observations and measurements) to high-level abstractions (e.g., theories, hypotheses, beliefs, concepts, formulae etc). Therefore, it comprises increasingly lean modalities of knowledge and knowledge representations media and the relative boundaries and relationships between them. Two forces manifest on the knowledge spectrum: that of creating abstractions and that of instantiating abstractions for practical applications. The former is the theory elicitation and is synonymous to processes of context reduction and knowledge decomposition. The latter, theory application, equates to context increase and knowledge composition processes. The engines behind the two knowledge spectrum forces are the knowledge processors, natural or artificial entities able to create abstractions from data and to instantiate abstractions in order to fit reality. Knowledge is traditionally categorized into implicit and explicit (Table 1 ) and ranges from rich representations grounded in a reality, to highly abstracted, symbolic representations of that reality. The classical distinction between data, meta-data, information, knowledge and meta-knowledge is simplified by our subscription to the unified view of Algorithmic Information Theory (AIT) [ 4 ] which recasts all knowledge modalities and their processing into a general framework requiring a Universal Turing Machine, its programs and data represented as finite binary sequences. From this perspective a precise distinction between these modalities becomes unimportant. Implicit knowledge (U, from unobvious, unapparent) is the rich, experiential, sensorial kind of knowledge that a knowledge processor acquires when immersed into an environment (i.e., grounded in an environment), or presented with detailed representations of that environment (e.g., images, models, recordings, simulations). It is very well applicable to specific instances of problems and relies on processing mechanisms such as feature selection, pattern recognition and associative memory. Explicit knowledge (E) is the abstract, symbolic type of knowledge present explicitly in documentations of knowledge such as textbooks or guidelines. It requires a representation language and the capability of a knowledge processor to construe the meaning of concepts of that language. It is applicable to both specific and generic problems and relies on explicit reasoning mechanisms. The distinction between implicit and explicit knowledge are useful to characterize the nature of human expertise, but become problematic when one wants to describe fundamental differences between theoretical and applied sciences: many applied sciences, especially knowledge intensive ones, in addition to general theories of problem solving, also make use of explicit knowledge in order to describe, with various degrees of precision, particular instances of problem solving and theory application. This represents the rationale for further dividing the knowledge spectrum into general and individual knowledge (Table 2 ). General knowledge General knowledge (G) is the explicit, abstract, propositional type of knowledge (e.g., guidelines), well applicable to context-independent, generic problems. However, it is more difficult to use in specific contexts because of the gap between the general knowledge itself and a particular application context. This knowledge gap translates into uncertainty when a general knowledge fact is instantiated to a specific situation. For example, knowing generally that a certain drug may give allergic reactions but being uncertain whether a particular patient may or may not develop any, is an example of what we consider the uncertainty associated with general knowledge. The creation of general knowledge (i.e., abstraction, generalization, context reduction, theory elicitation) is a relevance-driven process done by "stripping away irrelevancies" [ 9 ]. This causes general knowledge to have a lower complexity and be more manageable: "generalization is saying less and less about more and more" [ 9 ]. Formal representations of explicit knowledge have been common in early artificial intelligence (AI) applications in the context of expert system development. They operated under the "closed world assumption" and were meant to make the representation of knowledge manageable, reproducible and clear. However this assumption also rendered the expert systems "brittle" or completely unusable when applied to real world problems [ 14 ]. The completeness necessary for automatic reasoning using explicit reasoning mechanisms can be illustrated with the following formal definition of the concept of "a brick" in a limited, hypothetical world, containing only simple geometric objects such as bricks and pyramids (Figure 2 ) (adapted from [ 15 ]): "being a brick implies three things: 1. first, that the brick is on something that is not a pyramid; 2. second, that there is nothing that the brick is on and that is on the brick as well; and 3. third, that there is nothing that is not a brick and the same thing as the brick." This definition could have the predicate calculus representation: This representation shows that an intelligent agent who has no implicit knowledge of the hypothetical physical world and no capacity of generalization or analogy making, must be explicitly provided with all knowledge necessary to reason about "bricks" in that limited reality. Such approaches are known to suffer from a fundamental shortcoming, the "frame problem." The frame problem Daniel Dennett was the first philosopher of science who clearly articulated the "frame problem" and promoted it as one of the central problems of artificial intelligence [ 16 ] (also see [ 17 ]). Janlert [ 18 ] identifies the frame problem with "the problem of representing change." In [ 14 ] the frame problem is defined as "the problem of representing and reasoning about the side effects and implicit changes in a world description." In order to articulate and circumvent the abstract nature of its definition, Dennett has invented a little story involving three generations of increasingly sophisticated robots. These fictitious robots are products of early artificial intelligence (AI) technology that use automated reasoning based on formal representations similar to the brick example. These particular robots are specifically designed to solve a problem consisting of the retrieval of their life-essential batteries from a room, under the threat of a ticking bomb set to go off soon. Although increasingly sophisticated in their reasoning, all three successive versions of the robot fail: • The first robot fails by missing a highly relevant side effect of pulling the wagon with the batteries out of the room: the ticking bomb sitting on the same wagon was also retrieved, together with the batteries. • The second robot did not finish its extensive, irrelevant side-effect reasoning procedures before the bomb goes off. As Dennett ironically puts it, the robot "had just finished deducing that pulling the wagon out of the room would not change the color of the room's walls and was embarking on a proof of the further implication that pulling the wagon out would cause its wheels to turn more revolutions than there were wheels on the wagon – when the bomb exploded." • The third robot failed because it was " busily (i.e., explicitly) ignoring some thousands of implications it has determined to be irrelevant " and its batteries were therefore lost in the inevitable explosion. The frame problem can therefore be recast as a problem of relevance [ 17 ] (see preface), which is compounded by time constraints. It demonstrates that relevance judgment mechanisms based on general knowledge are time consuming and cause the failure to solve time-constrained decision problems. It is a problem only because in the real world we do have time constraints . Individual knowledge Individual knowledge (I) or instance specific knowledge, on the other hand, is a knowledge modality very well applicable to real problems, because it identifies uniquely and matches precisely an application context. The knowledge gap and uncertainty are reduced but still exist because of our changing reality (time dimension) which may render individual knowledge about a patient collected in the past (e.g., value of blood pressure from a month ago), less applicable in the present or future. Because it preserves context (i.e., it is more grounded), individual knowledge has a higher complexity than general knowledge and hence is more difficult to manage (i.e., has high memory requirements). For example, knowing the drugs and the precise description (e.g., numeric, textual, visual) of the allergic reactions that they caused in a certain person, as well as many other particular knowledge facts about individual, is what we call individual knowledge. The uncertainty and knowledge gap related to the application of such knowledge to future instances of decision making involving that individual are reduced: individual knowledge is supposed to fit very well the application context where it was originally captured. Case-based reasoning Individual knowledge captured from a very specific context (e.g., diagnosing a particular patient with a particular disease) can be extrapolated to similar contexts. The higher the similarity between contexts, the smaller the knowledge gap and instantiation uncertainty and the higher the chances for a successful solution to a new problem. For this reason, individual knowledge processing has become increasingly important for artificial intelligence applications and is defined as the approach to solving new problems based on the solutions of similar past problems [ 14 , 19 - 21 ]. It has several flavors (e.g., exemplar-based, instance-based, memory-based, analogy-based) [ 21 ] which we will refer to in this paper interchangeably, through the generic term of "case-based reasoning" (CBR). There are four steps (the four "RE") that a case-based reasoner must perform [ 14 , 20 , 21 ]: 1. RETRIEVE: the retrieval from memory of the cases which are appropriate for the problem at hand; this task involves processes of analogy-making or case pattern matching; 2. REUSE: the decomposition of the retrieved cases in order to make them applicable to the problem at hand; 3. REVISE: the compositional adaptation and application of the knowledge encoded in the retrieved cases to the new problem; and 4. RETAIN: the addition of the current problem together with its resolution to the case base, for future use. CBR entails that an expert system has a rich collection of past problem-solving cases stored together with their resolutions. CBR also hinges on a proper management of the case base and on appropriate mechanisms for the matching, retrieval and adaptation of the knowledge stored in the cases relevant to a new problem. Ideally, the individual knowledge in a case-base will progress asymptotically towards an exhaustive knowledge base, which represents the "holy grail" of knowledge engineers. From a learning systems point of view, similarly to artificial neural networks [ 22 , 23 ] and inductive inference systems [ 24 ] that learn from training examples, a CBR system acquires new knowledge, stores it in a case base and makes use of it in new problem solving situations. The absolute positions and shapes of boundaries between the four knowledge modalities, although admittedly not as precise as drawn on the knowledge spectrum in Figure 1 , are not of importance for this discussion. However, the relative relationships between knowledge modalities are, and can be represented formally as a Venn diagram (Figure 3 ), which implies that: • Individual knowledge has a higher complexity than the explicit knowledge elicited from the same context. This is equivalent to stating that, for example, the picture of a person encodes more knowledge than the textual description of that person's appearance. • Implicit knowledge is a subset of the individual knowledge. • General knowledge is a subset of the explicit knowledge. • The set of individual knowledge represented explicitly formed by the intersection of individual knowledge with explicit knowledge is a nonempty set. This is equivalent to stating that it is possible, for example, for an explicit textual description to identify a context uniquely (e.g., the complete name and address of a person at a specified moment in time). A meta-level view of Medical Informatics The meta-level overview of sciences and the definitions and properties of the knowledge spectrum and knowledge modalities enable us to draw some fundamental differences between theoretical sciences and applied sciences such as Medicine [ 25 ] and Medical Informatics. From this perspective, theoretical sciences (e.g., theoretical computer science): • Make use of observations which are highly abstract symbolisms and create far more limited contexts of application of their theories, when compared to the complexity of the human body or of any social or biological system, • Have as a primary purpose the creation of general knowledge comprising valid, powerful theories which explain precisely and completely the observations, and therefore, • Include a relatively limited number of precise theories which are evaluated primarily by their power of explaining experimental observations, elegance, generality, and • Are less concerned with the acquisition of the individual knowledge required by the practical implementation and by the application of results to real world problems. Applied sciences such as Medicine and Medical Informatics, on the other hand: • Gather extensively data and observations ( individual knowledge ) from very complex systems [ 9 , 26 ] (e.g., human body), which are characterized by high individual variation and randomness; • Have as a primary purpose not only distilling data and observations into general knowledge , but are also concerned with the implementation details and with the application of theories to individual problem solving (e.g., diagnosis and treatment of real patients), • May lack the incentive to refine existing theories which are objectively wrong as long as practical success is achieved [ 25 ], • Contain very few simple, "elegant" theories ( general knowledge ) that can solve individual problems completely or explain and predict accurately [ 27 ] because of the complexity of the human body and its individual variation and, therefore, • May pursue the application of a multitude of mutually contradictory, poorly grounded, general theories (e.g., the general theory of medical reasoning and the concepts of "diagnosis" and "symptom") [ 1 , 25 ], • Abound in general theories (e.g., guidelines) which are "lossy" (i.e., ignore individual context variation) and which are evaluated statistically by their practical success relative to existing ones (e.g., cancer therapy), • Attempt to make up for the knowledge gap between general knowledge and the reality where knowledge is applied, by employing experienced clinicians who require extensive training and information technology (e.g., decision support), and, in addition, • Are compounded by time-constrained circumstances and largely unsolved ethical issues (e.g., privacy and confidentiality, genomics research). Given the special circumstances of our applied science in the context of other sciences and the increasing recognition of the importance of knowledge processing to Medical Informatics [ 28 ], we propose, as part of the thesis of this paper, that Medical Informatics should complement the traditional quest for general biomedical knowledge with the advance of acquisition, storage, communication and use of individual knowledge . By doing so, Medical Informatics will provide a solution to the problems that arise during the use of general knowledge and, in the same time, will enable clinical research as well as advanced decision support and education of both healthcare providers and patients . Individual knowledge processing equates to a case-based reasoning (CBR) approach that employs collections of patient cases. Currently, such collections are the focus of research on Electronic Health Records (EHR). Envisioned as "womb to tomb" collections of patient-specific data, EHR contain a wealth of data that could be used to support case-based decisions. If EHR are to be used in a CBR context, the issues pertinent to the design of case-bases automatically become pertinent to the EHR design, and the CBR paradigm becomes important to Medical Informatics . The overall knowledge processing capacity of healthcare systems can be thought to be distributed between two sources: human resources (i.e., healthcare professionals) and information technology (Medical Informatics). An ideal CBR approach would increase this knowledge processing capacity by allowing for the automatic processing (acquisition, representation, storage, retrieval and use) of individual knowledge present in increasingly rich knowledge media such as natural language artifacts, images, videos and computer simulations of reality (Figure 4 ). The storage and communication of knowledge are well advanced by current information technology. However, most of the acquisition, retrieval and knowledge use are, and will continue to be the task of professionals until advanced processing (e.g., real-time computer vision, scene understanding and synthesis, image understanding, robotics, natural language understanding) are applicable. Given the widespread use of natural languages as knowledge representation and communication media, it follows that natural language processing (NLP) research is a very important component of Medical Informatics, required to advance the organization and processing of individual knowledge in reusable case-bases. Further, the goal to advance processing of increasingly complex knowledge representations (e.g., natural language, sounds, images, simulations) and create intelligent machines that can hear, see, think, adapt and make decisions, brings Informatics even closer to what traditionally was the concern of Artificial Intelligence (AI) . Finally, because the knowledge processing capacity of human resources tends to remain relatively constant, moving towards the ideal of individual knowledge processing, no matter how slowly, may also have ethical implications because it proves that medical informaticians are trying to do everything they can in order to serve the interest of the individual . Discussion In order to support our thesis, the following discussion will focus initially on fundamental aspects of medical decision-making and biomedical knowledge creation from the standpoint of the knowledge spectrum. This will lead to a discussion of fundamental knowledge representation and processing principles and the proposal of a CBR perspective on EHR, including challenges and potential solutions. Human and computer knowledge processing Decision making in medicine Medicine is a knowledge intensive domain where time-constrained decisions based on uncertain observations are commonplace. In order to successfully cope with such situations, health professionals go through a tedious learning process in which they gain the necessary domain knowledge to evolve from novices to experts. As experts, health professionals have attained, among other things, two important, highly interrelated abilities: • To be able to reduce knowledge complexity by determining efficiently what is relevant for solving a problem in a particular situation, and, • To be able to reduce the knowledge gap between knowledge facts and reality which translates into being able to reduce the uncertainty of knowledge instantiation to a particular context. For example, both the presence and the absence of a past appendectomy are relevant and contribute (potentially unequally) to reducing the uncertainty of instantiation of the biomedical knowledge of an expert to a particular context of a patient with right lower abdominal pain. Fundamental to decision making, relevance judgments and uncertainty reduction seem both closely connected with the quality and quantity of knowledge available for solving a problem as well as with the nature of knowledge processing mechanisms. Studies of expert-novice differences in medicine [ 29 ] have shown that the key difference between novices and experts is the highly organized knowledge structures of the latter, and not the explicit strategies or algorithms they use to solve a problem. This is supported by expert system development experiences which showed that a system's power lies in the domain knowledge rather than in the sophistication of the reasoning strategies [ 14 ]. Studies of predictive measures of students' performance indicate that tests which measure the acquisition of domain knowledge are the best predictors [ 30 ]. The work on naturalistic decision-making (NDM) and the development of psychological models of "recognitional decision-making" such as the Recognition-Primed Decision (RPD) [ 31 - 33 ], suggest the heavy dependence of decision makers on their previous experience of problem-solving and also on their ability to perform mental simulations. The discussion around the amount of problem solving experience of a decision maker becomes critical in time-constrained decision circumstances. The exhaustiveness of the knowledge base and the efficiency of retrieval mechanisms now become paramount to the decision speed. Empirical evidence that shows the existence of "systematic changes of cognitive processes" related to time stress, comes from the studies on the psychology of decision-making under time constraints [ 34 ]. Although most of these studies attest the overall negative effect of time stress on the "effectiveness of decision-making processes" [ 35 ], others [ 31 , 33 ] argue that even extremely time-constrained situations could be handled successfully by human subjects, given enough expertise (i.e., enough problem solving experience). Since humans are able to make sound relevance judgments and reduce instantiation uncertainty of knowledge most of the times, the following questions arise: What is their strategy for increasing the exhaustiveness of their knowledge base while managing its exponential complexity? How do they represent and organize their knowledge and how do they manage time-constrained situations? At least some of these questions have been under intense scrutiny that has resulted in important empirical work on naturalistic decision-making [ 32 , 33 , 36 , 37 ]. Important insights have been gained at the individual but also at the organizational and social levels. Coherent with the importance of the social aspects of decision-making, Armstrong [ 38 ] builds an interesting argument about the Darwinian evolution, social networking and the drive for knowledge discovery of the humanity as being some of the reasons that contribute to the human decision making potential. From the perspective of the knowledge spectrum, it seems reasonable to associate expert decision makers with individual knowledge and novices with the more abstract general knowledge about a subject, available in explicit knowledge artifacts (e.g., textbooks, guidelines). It is also conceivable that mental models of experts span a great length of the knowledge spectrum, causing them to efficiently perform implicit processing (feature selection, pattern recognition, associative recall) and also just-in-time explicit reasoning (Figure 5 ). The ability to move freely across the knowledge spectrum causes experts to efficiently reduce data to abstractions and to create hypotheses and micro-theories through sound relevance judgments. The powerful mental simulations that experts can perform allow them to construe appropriate meanings of concepts and to verify their hypotheses against contexts of reality. Novices, on the other hand, have limited mental models of reality situated towards the abstract region of the spectrum. This causes them to have difficulties with construing appropriate meanings of concepts due to the increased knowledge gaps between their mental models and reality. Novices are therefore unable to make sound relevance judgments and limited in their ability of interpreting data and creating abstractions. They are also usually overwhelmed by the explicit, general knowledge present in textbooks and guidelines and unable to fully construe the meanings of concepts present in such knowledge artifacts. In conclusion, in information and knowledge intensive domains such as medicine, explicit reasoning is important but individual knowledge acquisition (i.e., experience) and processing (i.e., CBR) are crucial for decision-making. Because the nature of expertise seems largely connected with individual knowledge processing, it follows that the evolution of novices into experts is unattainable only by the provision of extensive general knowledge. In addition, not only the individual learning but also the collective sharing of experiences (e.g., case records, personal stories, etc.) between individuals and between generations, contribute to the way humans deal with decision problems . Patient-centered vs. population-centered healthcare The major driving force of science is universally applicable knowledge (i.e., general knowledge). While creating and communicating new knowledge, scientists move across the knowledge spectrum from the data that captures the reality of their experiments and observations towards abstract representations that allow them to communicate their theories. In biomedical research, such an example is the randomized controlled trial (RCT), currently regarded as the gold standard for knowledge creation. The correct design of an RCT is crucial for the validity of the medical evidence obtained. A correct randomization process in RCTs will limit the bias and increase the chance for applicability of the evidence obtained, to a specifically selected group of patients (e.g., "women aged 40–49 without family history of breast cancer"). However, at the same time, the randomization process removes the circumstances of individual cases and creates a knowledge gap between the RCT evidence and future application instances. As with any statistical approach, the RCT-based evidence is best applicable at the population level rather than at the individual level. This depersonalization of medical knowledge and evidence was also noted by others [ 39 , 40 ] and could also be illustrated by the observation that most patients feel relieved when told that the chances of being successfully treated for a certain condition are 99%, for example. Although this is psychologically very positive, the patients should not necessarily be relieved, as they could very well happen to fall among the 1%, for whom things could go wrong and for whom, usually, the RCT-based evidence does not provide additional information. An experienced physician and, from a CBR perspective, a highly efficient case-based reasoner, is most of the times able to individualize the medical decision for a particular patient for whom things are likely to go wrong and fill in the knowledge gap between the RCT evidence and the medical problem at hand. This could lead to avoiding a therapeutic procedure recommended by the medical evidence. The individual knowledge that this decision is based on is usually not provided by the RCT, but is acquired through a tedious process of training. This decision is often so complex that it cannot be easily explained as it becomes heuristic in nature and is motivated by the individual knowledge that a decision maker possesses. Others [ 41 ] have also pointed out that when physicians manage their cases (e.g., diagnosis and treatment), their previous experience allows them to make informed decisions based on heuristics rather than on a sound, complete and reproducible reasoning, such as logical inference based on a predicate calculus representation of a problem. In addition, human experts often disregard probabilistic, RCT-type of evidence and consistently detach themselves from the normative models of classical decision theory (e.g. probability theory, Bayes theory) in favor of heuristics-based approaches. Although prone to occasional failures, heuristics-based decisions are much more efficient in time-constrained and uncertain situations [ 33 ]. From the perspective of the knowledge spectrum, the driving forces of Health Informatics and RCT methodology seem to have opposite directions: while Informatics aims towards individual knowledge and personalized health care, the general knowledge gained through populational studies (e.g., RCTs) targets the ideal of universal applicability (Figure 6 ). The value of a single bit of data (e.g., a Yes/No answer to a specific question such as a past appendectomy) can be very relevant in a decision-making context if it reduces the overall uncertainty of knowledge. However, such individual bits of data are inevitably lost during the creation of general knowledge. Rigorously and expensively collected, general, populational level knowledge is useful only in situations where individual knowledge lacks (e.g., new drugs), providing the decision makers have access to it and are able to apply it to specific situations. However, general knowledge is unlikely to be used as such in many naturalistic decision-making processes, because it does not support the way expert decision makers think. The knowledge gap and inherent instantiation uncertainty manifested in the application of general knowledge does not fully enable the education of providers and patients which would require additional knowledge about individual contexts of successful or unsuccessful application instances. Informatics, on the other hand, by advancing individual knowledge processing, provides an alternative solution to the problems that arise from the use of general knowledge that targets universal applicability. An integral part of individual knowledge, genomic data is already recognized [ 39 , 40 ] as being of extreme importance for a solution to the problems of general knowledge . Knowledge representation by formal methods The application of formal knowledge representations to real problems suffers from a fundamental shortcoming: the frame problem. As explained above (see "The frame problem"), the frame problem can be recast as a problem of relevance. Given the capability of relatively effortless human relevance judgments, the frame problem seems a rather "artificial" creation, difficult to grasp and which usually goes unnoticed. In order to circumvent its abstract nature, Dennett uses a story-telling approach. However, the frame problem also applies to and could be illustrated from the perspective of humans, who in their first years of life, learn and can easily and efficiently reason about the side effects and the implicit changes of the complex four-dimensional spatio-temporal physical world in which they live. As this learning gradually becomes common sense knowledge, it causes us to efficiently determine the relevant implicit changes while ignoring the non-relevant ones for a given situation. For example, such facts as that the clothes we are wearing are moving with us while walking or traveling are most of the times irrelevant given the context of a planned trip. However, if the trip involves some rapid movement through the air such as riding a motorbike, suddenly wearing a sombrero becomes a relevant fact. As experts at managing our physical world, we are able, through an effortless but powerful mental simulation, to determine the relevance of such a particular fact. The recall of our personal experiences of moving fast through the air and of the dragging force of the air becomes paramount. Therefore, intelligent agent must be endowed with efficient mechanisms for determining the relevance of particular facts for a decision . We suggest that what made the robots vulnerable was their creators' choice for knowledge representation and reasoning: the robots did not have quick access to implicit knowledge about the relevance of particular facts (i.e., records of problem solving instances) but only to explicit facts in frames which had to be employed in time-consuming, immense number of explicit relevance judgments about the effects of particular actions. Although they were supposed to be experts at their task, the robots were behaving like novices. The frame problem is not a problem of the knowledge representation per se, but a problem of the choice for representation of knowledge needed to solve time-constrained decisions. In other words, formal representations and logic reasoning work, but not in time constrained, complex situations. From the perspective of our knowledge spectrum, explicit, formal representations sit on the abstract side of the spectrum (Figure 7 ). The retrieval of explicit knowledge representation is currently the subject of the increasingly important field of research of information retrieval (IR). It is commonly accepted that IR is strongly coupled with the notion of intended meaning of concepts: a retrieved document is considered to be relevant to a query if the intended meanings of the authors of a document are relevant to the intended meaning of that query. We propose that "meaning," a property that characterizes all concepts present in explicit knowledge, is intimately connected (if not identical) with the notion of context . According to this rather paradoxical view, meaning, a property which characterizes the abstract side of the knowledge spectrum, is strongly coupled with context which, by definition, is a feature of the reality side of the knowledge spectrum. Therefore, in order to construe meaning appropriately one needs to be able to efficiently move from abstractions towards richer representations of reality. This movement on the knowledge spectrum is necessary in order to fill the knowledge gap between abstract concepts and the richer mental representations required for construing their meaning. Explicit, formal representations attempt to capture general truth and generally applicable problem solving strategies, but become too abstract in nature. Through the abstraction process, which is essentially a reduction driven by the relevance judgments of knowledge creators, the context of a problem is lost. Losing context creates difficulties with construing meaning (which is context-dependent by definition) and widens the knowledge gap between the representation itself and the reality of a future problem-solving instance. The knowledge gap translates into the instantiation uncertainty that characterizes the application of general knowledge to specific problems. Making up for the knowledge gap through explicit relevance reasoning becomes time consuming and consequently takes its toll on the applicability of the representation. In sensitive applications such as medical decision-making and health research, general knowledge may potentially be harmful (e.g., prescribing an highly recommended drug to which a patient has a undocumented allergy). In addition, abstractions and general methods and theories of problem solving and decision making (e.g., guidelines) do not fully enable the education of individuals and the learning from successes and mistakes. Knowledge representation approaches must therefore preserve to the extent possible, the context of a problem-solving instance. By efficiently recalling similar past instances of problem solving and their contexts, intelligent agents are immediately provided with implicit knowledge about relevance, encoded in the retrieved contexts and, in the same time, with more possibilities to reduce the instantiation uncertainty of general knowledge when applied to specific problems. To enable this, informatics research must advance the processing of rich representations of the knowledge encoded in past problem solving cases: this is the definition of CBR research . Knowledge representation by natural language Similar to formal specifications (e.g., predicate calculus) natural language uses abstractions, i.e., concepts. Its richness and power of expression place it in the knowledge spectrum to the left side of formal specifications but to the right side of rich descriptions consisting of images, sounds, video-clips and simulations of reality (Figure 4 ). Natural language has power of expression but loose semantics and inherent ambiguity. However, despite its abstract nature, it remains the indispensable, main knowledge representation and transfer medium between humans. In order to illustrate our point about ambiguity we direct the reader to the previous, natural language definition of the concept of "a brick." Although the definition may look unequivocal, there are subtle ambiguities that make a difference in the predicate calculus representation. The first condition of an object to be "a brick" (i.e., "the brick is on something that is not a pyramid," highlighted in the equations 2 and 3) is an ambiguous natural language construction and could have slightly different formal representations: In (2) this condition has been interpreted as: "the brick being on something IMPLIES that that something is not a pyramid" and was therefore represented as "for all Y, if X is on Y, this implies that Y is not a pyramid." In (3), which is identical to (1) but is repeated to the benefit of the reader, this condition was interpreted as "the brick MUST BE (or is always) on something that is not a pyramid" and that was represented as "there exists Y such as X is on Y and Y is not a pyramid." The first definition is therefore more "relaxed" as it allows the possibility that a brick sits on nothing. The second definition is more restrictive, because it requires the brick to be on something that is "not a pyramid" or otherwise X is not a brick anymore. Therefore, the first definition is more general and defines the concept of "a brick" in such a way that the definition would be true even in a world with no gravity (i.e., the brick is on nothing). In addition, definition (3) does not reject the possibility that an object sits on both another brick and a pyramid, at the same time (Figure 8 ). The point is that, most often, humans receive and transmit knowledge without the deep understanding and completeness required by an exact mathematical representation of the knowledge to be transmitted. This shallowness has also been recognized by others [ 42 ] who are trying to draw natural language processing researchers' attention to the fact that humans are rather superficial in their knowledge acquisition and processing and often make use of "underspecified" representations. Although, since the early days of science, scientists have fallen in love with the pure reasoning approaches, as they were reproducible, unambiguous means to express new knowledge, the problems with the use of classical predicate calculus as a knowledge representation method and of the classical logic inference as a reasoning strategy are discouraging. This is due to the requirements of complete, unequivocal representations, which prevents them from dealing with the messiness of the real world problems. If possessing the necessary knowledge, humans are able to effortlessly fill the knowledge gaps between natural language representations and their richer representations of reality (i.e., mental models), and to easily construe the appropriate meaning of potentially ambiguous concepts. Although current technology allows for its storage, knowledge present in richer media (e.g., images, videos, simulations) is currently very difficult to process (e.g., real-time computer vision, scene understanding and synthesis, image understanding) using today's technology. Because natural languages are used by people universally and allow rich representations that no other language specification can attain, natural language processing (NLP) research is a first step that Informatics should take in order to advance the organization and processing of individual knowledge in case-bases that can be reused. The insights gained will advance knowledge processing towards richer knowledge representation media, will reduce the knowledge processing gap and consequently increase the knowledge processing capacity currently supported largely by human knowledge processors . Memory-based knowledge processing One of the main features of information processing systems is their memory. It is accepted that storage and manipulation of information are necessary for complex cognitive activities in humans [ 43 ]. Memory is also considered crucial for both the "situation recognition" and mental modeling processes which are part of naturalistic decision models [ 33 ]. From a computational point of view, one could easily argue that without a random access memory structure there can be no effective processing. In the context of "the computational architecture of creativity," this argument is clearly outlined in [ 44 ]. It is based on the examination of the classes of computational devices, in the ascending order of their computational power, ranging from finite-state machines to pushdown automata and linear automata. These are paralleled by their corresponding grammars, arranged similarly in the Chomsky hierarchy, consisting of regular grammars, context-free grammars, context-sensitive grammars and of the unrestricted transformational grammars for machines with random access memory [ 44 ]. Recent natural language processing (NLP) research stresses the importance of memorization of individual natural language examples [ 45 ]. The importance of memory is also emphasized in earlier [ 46 ] and more recent models of language processing in humans [ 47 - 49 ]. These converge on the idea that natural language processing, regardless of the processor, is memory-based (i.e., case-based). Additional evidence comes from the fact that most language constructs (e.g. words, phrases) have very low frequencies. In fact, the very low frequency of most words in the English language (i.e., Zipf's law) is known from the 1940s since Zipf's famous book "Human Behavior and the Principle of Least Effort" [ 50 ] which is discussed in [ 51 ]. The main implication of "Zipf's law" is that purely statistical approaches or language processing algorithms that do not memorize training examples will either lose important information or may need extensive data (potentially impossible to collect) in order to be able to retain important features which have extremely low frequencies [ 52 ] and which may be crucial for construing the appropriate meanings of a language's concepts. The tradeoff between learning effort and communication efficiency seems to be biased naturally towards memorization rather than towards logical reasoning. The processing complexity of natural language might therefore not be an intrinsic quality of the algorithms, but rather a function of the memorization capabilities of the language processor, given the sparseness of natural language pattern space. By analogy, the advanced knowledge processing in humans might not be the result of very sophisticated reasoning strategies, but rather the utilization of a limited reasoning apparatus on a huge knowledge base, consisting of rich representations of one's experience. The limitations in reasoning are balanced by complex spatio-temporal pattern recognition capabilities operating on a case base built from years of experience. This case base includes common-sense knowledge. Furthermore, people and computers memorize information differently. Both have a short term, working memory and long-term memory for storing data and information. However, the memory access is carried out in different ways. Computers can reliably store large streams of data, which most of the times have a very well defined spatial and temporal structure (e.g., a movie clip). In contrast, people can only store information and knowledge rather than data and their storage is unreliable, temporally fragmented and spatially incomplete. Computers have very reliable memories capable of error checking at the bit level while the human memory supports only a high-level semantic consistency check. Finally, computers access their memory in a random seek fashion, being able to position their "reading heads" at any position in the data streams in order to extract a certain block of data. People can access their memory by content, by being provided with an incomplete description of a potentially complex, spatio-temporal pattern serving as a retrieval key. Therefore, one of the main differences between computers and humans is that computers have address-based random access memories, while humans possess content-addressable memories. In conclusion, from a case-based reasoning perspective, humans seem to be naturally endowed with the necessary structures for efficient case base acquisition, organization and retrieval while computers do not directly support this way of processing information and knowledge . Pattern recognition, comparison and analogy-making Pattern recognition is an undisputed feature of human cognitive abilities and a research area in its own right. However, it does not seem to be as pervasive as it should, in the information processing systems in current use. Natural language, as a product of human cognition, offers compelling evidence that people are naturally inclined toward processing information using pattern recognition and similarity principles. This evidence is supported by the widespread use of language devices such as the simile and the metaphor. These are examples of comparison and analogy making that humans perform without effort, in contrast to the difficulty of implementing them in the artificial information processing systems [ 53 ]. Analogy making is essential to generating new knowledge and new artifact designs [ 54 - 56 ], as well as to problem solving and inductive reasoning [ 57 , 58 ]. In a case-based reasoning context, the essential tasks of case matching and retrieval rely on pattern recognition, comparison and analogy making. In a decision making process, these mechanisms provide the immediate, implicit access to information about relevance stored in the contexts of similar instances of problems solving. The patterns and analogies that humans are able to handle are often represented by complex spatio-temporal events with a potentially multi-sensorial impact. For example, while humans have no difficulty in understanding a metaphor like "the computer swallowed the disk," an artificial information processing system that has no visual input sensors and which lacks the capability of image understanding, would probably never be able to perceive this particular analogy with the same speed, because of the extensive reasoning and amount of explicit knowledge needed to bring the swallowing process, as it occurs in living things, close to the action of inserting a disk into a computer's disk drive. In addition to operating on high dimensional, spatio-temporal complex patterns, analogy making in humans may also possess a dynamic component that could yield different relevance judgment outcomes, depending of context. A very illustrative example is given by French and Labiouse in [ 59 ], using the concept of a "claw hammer." According to its designed purpose, the "claw hammer" is semantically close to other concepts like "nail," "hit" and "pound." However, it may be dynamically "relocated" or reassigned in the semantic space, through a complex spatio-temporal mental simulation and analogy-making process, to the dynamically created class of "back-scratching devices," in the semantic neighborhood of the "itch," "scratch" and "claw" concepts. Similarly, one could think about the concept of a wooden decoy duck, which inherits properties from at least the "wooden object", "animal duck", "toy" and "hunting gear" classes. This concept may also be dynamically relocated into the semantic neighborhood of any of the classes, depending on the context of use that may be focused on themes such as "combustibles" or "hunting" for example. In the medical domain, the contextual dependence of relevance judgments, classifications and analogies is even more important, as these are often based on uncertain information and may be dynamically reevaluated in the light of new information about the patients or about their diseases. Polyhierarchy and multiple inheritance are indisputable desiderata of terminology systems [ 60 ]. However, building multiple inheritance mechanism using current technology seems very difficult, simply because the number of possible alternative classifications increases exponentially with the number of concepts. It is also very unlikely that this kind of taxonomic dynamicity (e.g., the claw hammer circumstantially classified as a back-scratching-device) of the human semantic space could work on such fixed conceptual structures which are constructed beforehand through learning, in human semantic memory. A more plausible hypothesis is that such ad-hoc classifications are circumstantially created using mechanisms that are closer to a distance calculation between high dimensional, distributed, vector representation of concepts. This is in agreement with neurolinguistic evidence from functional brain imaging studies of the human semantic memory. These studies suggest the existence of distributed feature networks for the representation of object concepts [ 61 ] and help the case for less structured approaches to capturing and representing semantics such as compositional terminology schemes (e.g. as in GALEN-GRAIL [ 62 ] and SNOMED-RT [ 63 ]), latent semantic indexing (LSA) [ 64 - 70 ] and connectionist models [ 49 , 71 , 72 ]. These approaches allow for a multidimensional semantic space where concept features can vary in importance, evolve or change dynamically, accounting for many possible classifications and subtle variations of concept meaning, including the new and the less plausible ones. This contrasts with the fixed or highly structured semantic representation schemas (e.g. fixed knowledge frames, semantic networks, ontologies), which fail to capture concept semantics in a way that provides richness, dynamicity and reusability. The dynamicity of concept meanings and relevance judgments may offer at least one of the reasons why fixed classification schemes, controlled terminology systems or open domain ontologies have not turned out satisfactory. It may also explain why existing lexical databases based on carefully handcrafted knowledge such as WordNet [ 73 ] often contain either too fine-grained or too coarse-grained, "static" semantic information [ 64 ]. In information intensive domains like medicine, concept dynamicity may account for why the development of a universal (i.e., one size fits all) clinical terminology system is so difficult [ 74 ]. From a case-based reasoning perspective, humans are naturally equipped with powerful pattern matching and classification capabilities which allow them to cope with complex, time-constrained relevance judgments, to easily construe meaning of concepts and to tolerate the ambiguity of natural language . Only relatively recently have computers come close to this functionality with the introduction of data mining and machine learning techniques such as self organizing maps and clustering algorithms based on similarity metrics [ 75 ]. In such machine learning approaches, the important problem of feature selection equates to a problems of relevance. CBR enabled EHR – Proposals, Challenges and Solutions Iatrogenic causes are said to be important causes of death in the US [ 76 ]. The reported incidence of adverse effects among patients in Canadian acute care hospitals is 7.5% [ 77 ]. A proposed means to counteract such medical errors is information technology, through the education and decision support offered to health care professionals. One very effective form of medical education is the retrospective analysis of case records where health professionals, both experienced or novices, learn from their own and from others' successes and failures [ 78 ]. Providing that legal and ethical implications such as provider and patient protection are dealt with appropriately, the efficacy of this teaching method can be improved if case records are continuously created, enriched, accumulated and organized on similarity principles. This is possible through a CBR approach of the EHR which, from this perspective, could serve as a comprehensive case base of managed patients that will evolve asymptotically towards an exhaustive knowledge base. Medical errors are also connected with the complex human cognitive task of planning [ 79 ]. CBR approaches, devised originally as a solution to automated planning tasks [ 80 ], have been since used in various applications including healthcare, legal and military (e.g., battle planning) [ 21 ]. This demonstrates a particularly good fit of a medical decision support based on CBR with its human users, the healthcare professionals. Providing that the privacy and confidentiality issues, which are even more stringent in this case, are dealt with appropriately, opening EHR to patients could benefit them [ 81 ]. It is perfectly conceivable that patients could learn from the history of other cases similar to theirs, which could be presented in an anonymized, story-telling format and organized on case similarity principles. It is also possible that patients may be willing to directly provide some of their own case information in order to be matched with previously managed cases, for example in the context of online chronic disease support groups. These principles are already realized in form of bulletin boards, mailing lists and forums, where actual patients interact with each other and occasionally with health professionals and exchange information regarding health related problems ([ 82 ] and [ 83 ]). The unstructured, textual exchange of information in such resources would ideally be moderated by knowledgeable individuals (e.g., providers). Although the automatic processing of text still is not readily available, case matching is possible so far and is performed by the very individuals who are able to offer useful information and knowledge to others, based on the similarity of their own experiences (i.e., their own story). Medicine has always and will always be a case oriented profession. Medical Informatics has recognized this early through the works of various researchers who pioneered the area of decision support systems [ 84 ]. Relevant to CBR work are also the attempts to enhance early decision support systems with domain knowledge from simulated patient cases [ 85 ]. Currently, the exploration of CBR in medical contexts is increasing [ 86 - 94 ]. Regardless of the problem nature, the most important components of a CBR expert system are • The case base, the memory of past problem-solving instances • The case matching or pattern matching procedure which retrieves the relevant cases for a certain problem While humans seem to possess a natural support for these two components, there is still work to be done in order to make the computer support this kind of knowledge acquisition and processing. We envision four important challenges in advancing towards CBR enabled EHRs: 1. Case record comprehensiveness 2. Organization on similarity and associative principles (associative memory) and development of advance data visualization techniques 3. Development of pattern recognition and similarity measures between heterogeneous records 4. Solving ethical issues and provision of privacy and confidentiality measures 1) Case record comprehensiveness EHR comprehensiveness is required because the exhaustiveness of a case base is not only a function of the number of records but also of the richness of each case record. Current knowledge processing technology limits the acquisition and especially the processing of comprehensive EHR records which incorporate structured data, images, video-clips, bio-signals, genomic data, unstructured textual data covering clinical findings, detailed patient history, etc. However, as knowledge processing technology advances and knowledge acquisition bottlenecks are overcome, it might be possible to overcome the heterogeneity and sparseness of EHR and allow the creation of representative case-bases and the organization of knowledge on principles that facilitate similarity based retrieval. Temporal knowledge is also a good example of a heterogeneously represented type of knowledge in the form of potentially non-interoperable standards for dates and times and temporal knowledge of various degrees of precision, embedded in knowledge facts such as "soon after receiving the drug, the patient developed a rash." Currently, for many people, the problem may seems to boil down to devising yet another standard which encompasses all the different temporal representations of dates, times and temporal concepts into a unified, common representation. From a knowledge engineering standpoint, and again currently for many researchers, this may equate to the creation of a comprehensive ontology of temporal knowledge. However, the problem of representing time starts to look like a somewhat limited version of another burning problem of Medical Informatics: that of medical terminologies. The fact that all these issues remain largely unsolved, can only help the case for CBR and for adaptive, empirical methods and approaches to knowledge processing. We believe that such approaches have the potential to cope and overcome the problems with redundant, possibly ambiguous representations, which have arbitrary degrees of precision. Thereby we are specifying a goal towards which the development of EHR should proceed. 2) Organization on similarity and associative principles (associative memory) and development of advanced data visualization techniques Similarity based retrieval is difficult with current database technology. For example, queries to retrieve cases which are similar to a textual description of a given case are difficult to answer. The comprehensiveness of EHR must be complemented with the possibility of indexing its records on similarity principles. Conceptually, the functionality of EHR will be that of an associative memory of cases that will enable the CBR paradigm. The organization of a case-base must be complemented by the development of advanced data visualization techniques that comply with the principles of organization of information by similarity. One example of such data visualization techniques are self-organizing maps [ 75 ]. These models are able to perform cluster analyses on high dimensional data sets and provide a visual display which can help with the navigation through and retrieval of similar cases. For instance, the self-organizing map obtained from the analysis of the Wisconsin Breast Cancer Dataset [ 95 ] used to cluster and classify cases based on their similarity in [ 96 ], could also be used for data visualization and navigation purposes, in a CBR context (Figure 9 ). It also demonstrates how high level abstractions (i.e., benign tumors forming the green cluster on the map) can be derived through an entirely automatic, data driven approach. 3) Development of pattern recognition and similarity measures between heterogeneous records CBR relies on the proper management of the case base and on appropriate mechanisms for matching and retrieval of these case records. All similarity retrieval mechanisms are based on some sort of distance calculation between the problem at hand and the records in the case base, followed by the retrieval of the most relevant ones. Clinical narratives and other EHR components containing unrestricted text represent a particularly difficult challenge for semantic similarity measures. The development of terminology systems based on less structured (e.g., latent semantic indexing, connectionist models) and data-driven approaches will provide the semantic richness, dynamicity and reusability needed for such complex tasks. A concrete example for the potential feasibility of such approaches, is the automated knowledge induction based on contextual similarity modeling ranging from morphological to sentential context [ 97 ] (Figure 10 ). An experimental knowledge processing system can induce automatically the new knowledge fact that Ayercillin , an item unknown to the system and hence not appearing in Figure 10 , is most likely to be a drug, precisely a penicillin. The decision is based on morphological (e.g., "-cillin"), semantic (e.g., six of the similar items are known to be drugs, precisely, penicillins) and pragmatic (e.g., the six, semantically similar items are consistent with the use in a medical context) similarities that help in filtering out the non-relevant information (e.g., book of common prayer). On the same basis, the system can also induce that surgical procedures ending in "-tomy" (e.g., perineotomy, valvulotomy, myringotomy, strabotomy) are usually incisions while those ending in "-ectomy" (e.g., myringectomy, tonsillectomy, splenectomy, nephrectomy) are usually removals, that concepts containing the morpheme "leuco" (e.g., leucocyte, leucothoe, platalea leucorodia) are usually associated with color white while those containing "eryth" (e.g., erythroblast, erythema, erythrina) with color red. However, despite such proof-of-concept applications and other progress in data mining and knowledge extraction from heterogeneous databases, case matching remains largely an open research question. 4) Solving ethical issues, provision of privacy and confidentiality measures We discuss this challenge last, not because it is less important but, on the contrary, because of its potential to become the most important obstacle to individual knowledge processing. The very fact that individual knowledge has the potential to contribute to solving future problems instances, raises the important ethical issue whether such knowledge should be made available to decision makers and researchers. Because the definition of individual knowledge implies the possibility to match it in time and space with an application context, i.e., with a patient, sharing individual knowledge is counterbalanced by the need for privacy and confidentiality. In addition, to further complicate matters, it may turn out that some of the most useful records for future instances of decision making are instances of medical errors or other unexpected events that are unique in their course of events and therefore easily identifiable together with their contexts of development (i.e., patients, providers, family members). The high complexity of individual knowledge renders explicit, manually controlled access to individual knowledge cases and their components unfeasible. The only solution to this problem seems to be of technological nature. Current privacy and confidentiality measures which include de-identification, de-nominalization and scrambling of the unique personal identifiers automated or semi-automated seem insufficient to counteract the potential to identify patients from unique, individual knowledge patterns. As a general approach, we propose that the accurate measuring of similarity of individual knowledge could form the basis of a confidentiality risk assessment. This could be intuitively understood by considering that: • very similar individual knowledge patterns which are in great numbers are a very low threat to the privacy and confidentiality infringement, and, at the other extreme, • stand-alone patterns which possess unique features or combinations of features, are at high risk of privacy and confidentiality breaches. In addition, the provision of privacy and confidentiality could be regarded as a special case of knowledge processing, which involves knowledge about the proper use (e.g., access, modification, transfer) of individual knowledge . This potentially complex, particular case of meta-knowledge processing could be implemented and managed using the principles of CBR paradigm itself, by building case-bases with examples of both proper and improper (simulated, not necessarily real) individual knowledge accesses and that can be compared with future access instances. Overcoming this very important challenge hinges on the possibility to effectively measure the similarity between heterogeneous records and on the advancement of knowledge processing on CBR principles. If successful, CBR research might therefore fulfill a longstanding need for intelligent information processing and advance informatics towards the ideal of individual knowledge processing. This calls for further investigation of information processing models that are, similarly to human experts, capable to efficiently move across the knowledge spectrum. One class of such models is represented by artificial neural networks [ 14 , 23 ], which are highly adaptive information processing models able to create high-level abstractions from raw data, completely automatically [ 75 ] and "learn by themselves" new information processing functions from data. From this perspective, Informatics aligns closely to the goals of AI to create intelligent machines that can hear, see, speak, think, adapt and make decisions. Conclusions CBR provides potential solutions to important problems that, among other, stymy the usefulness of EHR. The natural integration of learning with reasoning and the CBR resemblance to the cognitive models of human decision-making hold the promise to overcome the "brittleness" and "knowledge acquisition bottleneck" of classical expert systems. The CBR applications to the medical field have the potential to offering the training and decision support needed by health professionals and the means towards a true patient-centered healthcare. With a CBR theoretical foundation still in its infancy and with limited medical applications in existence, more research is needed for providing proofs of the feasibility of practical CBR-EHR applications. Challenges in the way to accomplish this include the increasing complexity, ethical issues as well as the paradigm shift that our current computing devices must undergo in order to support the CBR principles of knowledge processing. Summary 1. Science is twofold and is driven by two opposite forces: that of creating theories (theoretical sciences), and that of applying theories to practical applications (applied sciences). Medical Informatics is fundamentally an applied science that should be committed to advancing patient-centred medicine through individual knowledge processing. 2. Case-based reasoning is the technical solution that enables a continuous individual knowledge processing that could be integrated with the Electronic Health Records. 3. Medicine is an information and knowledge intensive domain where time-constrained decision problems can only be solved effectively based on the recollection of similar problems and their solutions (i.e., a case-based reasoning strategy). The collective sharing of experiences is important for making future decisions as well as for learning how to make decisions. 4. Unlike computers, human decision makers possess the components necessary to perform case-based reasoning naturally (i.e., a content addressable memory to organize a case base efficiently by similarity principles, as well as the capability to perform pattern recognition, comparison, and analogy-making). 5. Applying the CBR approach to EHR might be a way to overcome the important obstacles of EHR acceptance and use, providing that technical challenges and ethical issues arising are addressed appropriately. List of abbreviations AI Artificial intelligence AIT Algorithmic Information Theory CBR Case Based Reasoning E Explicit knowledge EHR Electronic Health Records G General knowledge LSA Latent Semantic Indexing I Individual knowledge NDM Naturalistic Decision Making RCT Randomized Controlled Trial RPD Recognition-Primed Decision U Implicit (Unobvious) knowledge Competing interests The author(s) declare that they have no competing interests. Authors' contributions Before the reviews SP researched the paper and provided a first draft. JA, JM critically revised the manuscript three times each and provided their own additions to the initial draft. JA provided more feedback on the cognitive aspects and decision-making as well as writing style and missing references. JM additions were with regard to the writing style, clarity, missing references and the overall organization of the paper. After the reviews SP and JM worked on the responses to reviewers' comments. SP wrote a first revision of the paper. JM provided extensive feedback as well as new references and suggested a major revision that includes recent ideas. JA also commented and made suggestions on the knowledge spectrum model and on the meta-level view on Medical Informatics. SP overhauled the entire paper. JM revised the new version and provided feedback. SP operated the changes and proposed new modifications. JM revised the second draft. JA provided feedback on the second draft of the paper with regard to fundamental aspects of knowledge modalities. SP and JM incorporated the minor changes suggested by the last review. All authors read and approved the final version of the paper. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC544898.xml

521184

Endangered Frogs Coexist with Fungus Once Thought Fatal

Amphibian declines have reached crisis proportions in various parts of the world. In many areas, habitat loss is the likely culprit. But when mass die-offs suddenly occurred in relatively undisturbed habitats, the cause was far less obvious. Fourteen species suffered either extinctions or major declines in the pristine rainforests of Queensland, Australia, between 1979 and 1993. It was suggested in 1996 that some unknown disease had spread through the populations, but no pathogen was discovered until 1998, when the fungus Batrachochytrium dendrobatidis was identified from sick and dead frogs. Since then, several lines of evidence suggest that B. dendrobatidis may be involved in frog declines: the fungus has been found on frogs in afflicted areas; lab studies show that it's highly pathogenic to some frog species; and pathological evidence links it to host mortality. But with little information about the prevalence of this fungal infection in wild frogs, or how the disease impacts frogs in the wild, the causal role of this chytrid fungus remains unclear. To evaluate the effects of B. dendrobatidis on frogs in their natural habitat, Richard Retallick et al. focused on six species living in the high-elevation rainforest streams of Eungella National Park in Queensland, Australia, where frog losses were “particularly catastrophic.” Two species vanished between 1985 and 1986: the Eungella Gastric-Brooding Frog ( Rheobatrachus vitellinus ), which is now thought extinct, and the Eungella Torrent Frog ( Taudactylus eungellensis ), which later reappeared in a few small populations. Other local frog species escaped this period relatively unscathed. Taudactylus eungellensis (Photo: Richard Retallick) Retallick captured frogs from six sites from 1994 to 1998, clipped one or two toe tips from each frog to age and identify them, and then released the frogs back into the wild. At the time, B. dendrobatidis had yet to be identified, but Retallick retained the toe tips, and the authors tested the toes for disease in 2002–2003. Fungal infections were found in two species— T. eungellensis and Litoria wilcoxii / jungguy (the latter consists of two species that are indistinguishable without genetic analysis); the other four species were infection-free. L. wilcoxii / jungguy did not decline to any great extent during the 1985–1986 die-off. The proportion of infected T. eungellensis frogs was greatest at three particular sites, which showed peak infections during cooler months. Prevalence of infection was highest during winter and spring, but did not vary from year to year, suggesting that the infection is now endemic. Fungal infections were found in 27.7% of L. wilcoxii / jungguy frogs, with no evidence that prevalence differed among sites, seasons, or individuals (males, females, or subadults). The probability of recapture was significantly lower for frogs that were already infected when first captured. While this might suggest a correlation between infection and death, it's impossible to distinguish death from simple failure to recapture the animal. On further analysis, McCallum and colleagues found no evidence that survival differed between infected and uninfected frogs, suggesting that this potentially devastating amphibian disease now coexists with the frogs, with little effect on their populations. These results, the authors conclude, “show unequivocally” that remaining populations of T. eungellensis , a rainforest frog listed as endangered, “now persist with stable infections of B. dendrobatidis .” While these findings do not exonerate the fungus as the agent of mass declines, they do rule out the possibility that the fungus caused the decline, then vanished from the area, allowing frog populations to recover. The authors allow that it's possible that B. dendrobatidis did not cause the initial T. eungellensis declines. Or alternately, the fungus could have emerged as a novel pathogen in the ecosystem, causing massive casualties before some form of evolutionary response took hold. Surviving frog populations may have evolved resistance to the pathogen, for example, or less virulent strains of the fungus may have evolved. If it turns out that frog populations can develop resistance to the chytrid fungus, the researchers point out, then a conservation program of captive breeding and selecting for resistance could potentially thwart the extinction of these, and other, critically endangered frogs. A critical next step, then, is to determine whether frogs and fungus do coevolve.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521184.xml

529255

The diagnostic value of endoscopy and Helicobacter pylori tests for peptic ulcer patients in late post-treatment setting

Background Guidelines for management of peptic ulcer patients after the treatment are largely directed to detection of H. pylori infection using only non-invasive tests. We compared the diagnostic value of non-invasive and endoscopy based H. pylori tests in a late post-treatment setting. Methods Altogether 34 patients with dyspeptic complaints were referred for gastroscopy 5 years after the treatment of peptic ulcer using a one-week triple therapy scheme. The endoscopic and histologic findings were evaluated according to the Sydney classification. Bacteriological, PCR and cytological investigations and 13 C-UBT tests were performed. Results Seventeen patients were defined H. pylori positive by 13 C-UBT test, PCR and histological examination. On endoscopy, peptic ulcer persisted in 4 H. pylori positive cases. Among the 6 cases with erosions of the gastric mucosa, only two patients were H. pylori positive. Mucosal atrophy and intestinal metaplasia were revealed both in the H. pylori positive and H. pylori negative cases. Bacteriological examination revealed three clarithromycin resistant H. pylori strains. Cytology failed to prove validity for diagnosing H. pylori in a post-treatment setting. Conclusions In a late post-treatment setting, patients with dyspepsia should not be monitored only by non-invasive investigation methods; it is also justified to use the classical histological evaluation of H. pylori colonisation, PCR and bacteriology as they have shown good concordance with 13 C-UBT . Moreover, endoscopy and histological investigation of a gastric biopsy have proved to be the methods with an additional diagnostic value, providing the physician with information about inflammatory, atrophic and metaplastic lesions of the stomach in dyspeptic H. pylori positive and negative patients. Bacteriological methods are suggested for detecting the putative antimicrobial resistance of H. pylori , aimed at successful eradication of infection in persistent peptic ulcer cases.

Background Treatment of peptic ulcer in accordance with relevant guidelines is becoming a common task for general practitioners [ 1 - 6 ]. In a post-treatment setting, in accordance with guidelines, prompt check-up of treatment results is recommended only in gastric ulcer cases with the use of 13 C-urea breath test ( 13 C-UBT ) [ 2 - 5 ]. In a situation where patients have clinical symptoms after H. pylori eradication therapy, endoscopy is favoured in all peptic ulcer cases [ 6 ]. The aim of endoscopy is to establish the reason for clinical symptoms and to prove presence of peptic ulcer or malignancies, but also to support physicians and patients in the understanding of complaints [ 7 ]. Moreover, endoscopy allows determination of persistent H. pylori infection using endoscopy-based tests. Endoscopic biopsies alone are not considered adequate for confirming eradication of bacteria, although they might provide additional information about gastritis and dysplasia [ 8 ]. Use of more than one method in testing gastric specimens definitely enhances the diagnostic value when assessing the post-treatment H. pylori status [ 9 ]. Our aim was to assess the diagnostic value of different non-invasive ( 13 C-UBT ) and endoscopy-based diagnostic methods (visual endoscopy, classical cytological and histological examination of mucosal specimens, PCR and bacteriological methods) for monitoring patients after eradication therapy in a late post-treatment setting. Methods Patients The study group was formed of 134 consecutive peptic ulcer outpatients who had been treated by 7-day triple therapy with metronidazole, amoxicillin and omeprazole in 1996. The group was observed at the outpatient department of Tartu University Hospital at 4 weeks, at 1 year (1997) and at 5 (2001) years after treatment [ 10 ]. Five years after treatment, 108 patients (81% of the initial group) were available for the follow-up of the clinical course of peptic ulcer. During the 5-year follow-up period only 11 (10 %) patients had relapses of peptic ulcer. For comparison of the diagnostic value of different diagnostic methods in a post-treatment setting, 34 patients were recruited. The inclusion criteria for this study group were resistant upper abdominal pain as the predominant complaint and compliance with all investigations (clinical symptoms, 13 C-UBT , endoscopy, biopsy, bacteriology, PCR and cytology). The studied patients were not NSAID users. Methods The patients passed the Gastrointestinal Symptoms Rating Scale (GSRS) test [ 11 ] in a validated Estonian translation. Dyspeptic syndrome (abdominal pain, heartburn, acid regurgitation, sucking sensation, nausea and vomiting) was registered on the 7-grade Likert scale for assessing severity of symptoms. The mean score of dyspeptic syndrome was calculated for each patient. 13 C-UBT The subjects passed 13 C-UBT drinking 100 mg 13 C -urea; the test meal was citric acid and the time of specimen collection was 30 min. The test was provided, according to a standard protocol, from the Helsinki Keskuskatu Laboratory, Finland. The ratio of 13 CO 2 to 12 CO 2 in expired breath was measured by mass spectrometry and expressed in ml/mmol/kg (δ). An automated breath 13 C analyser (ABCA) with chromatographic purification and a single inlet isotope ratio mass spectrometer (IRMS) were used. A difference of 5‰ in the content (δ 13 C ) was considered positive for H. pylori infection. Endoscopy of the upper gastrointestinal tract The procedure was performed with the gastroscope Olympus-GIF 21. All mucosal defects were registered according to the Sydney classification for endoscopic evaluation [ 12 ]. Gastric ulcer was diagnosed if the ulcer was located at the angulus or above it. Duodenal ulcer was diagnosed if the ulcer was found in the duodenal bulb area. Gastrobiopsy and histological examination Five specimens from the antrum mucosa and five from the corpus mucosa were taken with medium-sized forceps. Two specimens were embedded in paraffin and the paraffin sections were stained using haematoxylin-eosin and Giemsa methods. The mucosal specimens were evaluated histologically according to the Sydney classification: presence of neutrophil infiltration, chronic lymphocytic inflammation, surface epithelial damage, atrophy, intestinal metaplasia, lymphoid follicles and H. pylori colonisation were evaluated on a three-grade scale both for the antrum and the corpus [ 12 - 14 ]. Bacteriological examination One specimen from the antrum and one from the corpus were placed in the Stuart Transport Medium (Oxoid) and taken to the laboratory within two hours for bacteriological examination. The biopsy samples were homogenised with sterile glass powder and under a stream of CO 2 and diluted in the Brucella broth (Oxoid). H. pylori was isolated on the Columbia Agar Base supplemented with 7% horse blood and 1% Vitox (Oxoid) or Isovitalex (BBL). The plates were incubated for 3–7 days at 37°C under microaerobic conditions (CampyBak, BBL or CampyGen, Oxoid). H. pylori was identified by Gram staining and by oxidase, catalase and urease reactions [ 15 ]. The sensitivity of the isolated H. pylori strains to clarithromycin was estimated by E-test. The antibiotic cut-off points employed for the E-test were 1.0 mg/l (NCCLS, 2002). Cytological examination One specimen was used for imprinting the cytology slides from the antrum and corpus mucosa, fixed with 96% ethanol and stained by Acridine Orange (Difco, BBL) [ 16 ]. The cytological specimens were studied under a fluorescence microscope (AXI Phot 2) where the morphotypes and the density of bacterial colonisation were evaluated [ 17 ]. A positive cytological diagnosis was based on the presence of typical helical H. pylori cells on the gastric mucosa and in the mucus layer. PCR For DNA extraction of H. pylori from a frozen gastric biopsy specimen, a previously described procedure was used [ 18 ]. The presence of the glmM gene in each strain was established by PCR using primers, the reaction mixture, and thermal cycling [ 19 , 20 ]. DNA from H. pylori NCTC 11637 (National Collection of Type Cultures, Central Public Health Laboratory, Colindale Ave., London NW9 5HT, England, United Kingdom) and the DNA-free reaction mixture were assayed in separate tubes in each PCR and were run as the positive and negative controls of the reaction, respectively. The PCR products were identified by electrophoresis on 2% agarose gels. Criteria for evaluation H. pylori was assessed positive if at least two tests were positive according to golden standard [ 21 ]. Statistical analysis The data were analysed by Fisher's exact tests using the Jandel SigmaStat 2.0 program. Measurements from the GSRS were expressed as the mean values for dyspeptic syndrome. Ethics The study was carried out in accordance with the Helsinki Declaration and was approved by the Ethics Committee of the University of Tartu. Results Dyspeptic syndrome was found in all 34 cases. The mean GSRS score for the patients varied from 1.2 to 4.3. The applied non-invasive test revealed H. pylori infection in half of the investigated patients: positive 13 C-UBT was found in 17 out of 34 cases. There was no difference between the mean GSRS score values for the H. pylori positive and negative cases (2.8 ± 1.8 vs. 2.9 ± 1.7, p > 0.05). On endoscopy, among the 34 patients, no ulcer or other mucosal defects were observed in 24 cases; erosions in the duodenal bulb were revealed in 6 cases and peptic ulcer was found in 4 cases (2 duodenal ulcers and 2 gastric ulcers). The data of H. pylori status and of the endoscopic finding are presented in Table 1 . Table 1 Comparison of the findings in H. pylori positive and negative cases in a late post-treatment setting Patients (n = 34) Non-invasive method 13 C-UBT (+) n = 17 13 C-UBT (-) n = 17 Invasive methods Endoscopy: Normal 11 13 Duodenal ulcer 2 0 Gastric ulcer 2 0 Erosions 2 4 Cytology: H. pylori (+) 4* Diverse forms of bacteria Histology: H. pylori (+) 17 0 Bacteriology: H. pylori (+) 16 1 PCR: H. pylori (+) 17 0 * typical morphology of H. pylori (the other cases showing diverse forms of bacteria) A poor concordance was found between the visual examination of the gastric and duodenal mucosa on endoscopy and the applied non-invasive and invasive tests of H. pylori (accepting 13 C-UBT , histological examination and PCR as the reference tests). The gastric and duodenal mucosa was visually normal in 11 H. pylori positive cases out of 17. On the contrary, only in 4 H. pylori positive cases did the endoscopic examination reveal the above mentioned peptic ulcers. Among the 6 cases with erosions of the duodenal mucosa, only two patients were H. pylori positive. Comparison of the different diagnostic methods used for the detection of H. pylori is shown in Table 1 . The results of 13 C-UBT and PCR were consistent with the data of histological examination both in 17 H. pylori positive and 17 negative cases. On bacteriological examination, only one case, which was H. pylori positive both by PCR and the histological tests, was H. pylori negative. In contrast, cytological examination assessed typical H. pylori bacterial cells in only 4 of the 17 H. pylori positive cases (24%), while all other cases (both positive and negative for H. pylori by the other methods) displayed abundant bacteria of different morphotypes. The data of the histological findings are presented in Table 2 . Colonisation of the gastric mucosa by H. pylori was detected in 17 patients out of 34. Neutrophil infiltration, chronic inflammation, and surface epithelial damage both in the antrum and corpus mucosa were significantly expressed in the H. pylori positive cases (p < 0.001). Glandular atrophy and intestinal metaplasia were rarely observed both in the antrum and corpus mucosa of the H. pylori negative cases in comparison with the H. pylori positive cases, but the difference was not statistically significant (p > 0.05). Lymphoid follicles were more frequent in the antrum colonised with H. pylori (p < 0.05). Table 2 Gastric mucosal findings (by the Sydney system) in H. pylori positive and negative cases Gastric mucosal findings (Sydney system) H. pylori (+) n = 17 H. pylori (-) n = 17 p values Activity of neutrophil polymorphs Antrum 11/17 0/17 <0.001 Corpus 7/16 0/17 <0.05 Chronic inflammation Antrum 16/17 1/17 <0.001 Corpus 13/16 0/17 <0.001 Surface epithelial damage Antrum 13/17 0/17 <0.001 Corpus 8/16 0/17 <0.001 Glandular atrophy Antrum 7/17 2/17 NS* Corpus 4/16 3/17 NS Intestinal metaplasia Antrum 1/17 2/17 NS Corpus 0/16 2/17 NS Lymphoid follicles Antrum 6/17 0/17 <0.05 Corpus 5/16 2/17 NS * NS, not significant (p > 0.05). Bacteriological investigation revealed H. pylori in 16 biopsy samples of the antral mucosa, while highly (> 256 mg/l) clarithromycin resistant H. pylori strains were found in 3 cases. Discussion Proper diagnostic and therapeutic management of patients with dyspeptic syndrome after H. pylori eradication therapy is of utmost importance for physicians as well for patients [ 7 ]. Several studies [ 22 , 23 ] have demonstrated the reliability of H. pylori tests used before treatment, while post-treatment testing is not yet adequately studied. However, in the case of long-lasting recurrent dyspepsia after H. pylori eradication therapy, endoscopy has been strongly recommended [ 4 ]. Our study shows that endoscopy gives useful information for the general practitioner both in the cases where peptic ulcer is found and in the cases where it is not found. In the case of a normal endoscopic finding, further management depends on the histological finding and on H. pylori status. Since persistent H. pylori positivity is always associated with possible peptic ulcer recurrence, the second line treatment according to bacterial susceptibility should be recommended. In the remaining cases where H. pylori is absent, the gastric mucosa is normal and no ulcer is detected, management of such patients should be aimed at establishment of other possible reasons for their complaints. Usually, a normal endoscopic finding reassures both the doctor and the patient [ 7 ]. A recent study of Ohkusa et al. [ 24 ] showed that even simple careful visual evaluation of the mucosa and the diagnoses of erythema and oedema correlated well with H. pylori infection. On the contrary, our results demonstrate that although all patients with recurrent peptic ulcer were H. pylori positive, the minor visual findings in the other cases were not in concordance with H. pylori colonisation. Usually, the mucosa was visually normal even when H. pylori was found, and, on the contrary, most duodenal erosions occurred in H. pylori negative patients. The clinical data of our patients did not suggest earlier use of NSAID, which would have been one of the main reasons for H. pylori negative erosions. Therefore, after treatment, in presence of complaints, it is important to obtain samples for the investigation of gastric mucosa specimens to enhance the value of endoscopic examination. We completely agree with the opinion that the value of using mucosal specimens for histological evaluation of late post-treatment H. pylori eradication is sometimes underestimated [ 9 ]. The non-invasive H. pylori test alone cannot solve the clinical problem of these patients. In our study, H. pylori negative patients had dyspeptic syndrome as well as gastric mucosal erosions, glandular atrophy and intestinal metaplasia. The last two lesions can presumably be associated with previous H. pylori infection and the follow-up of severe mucosal changes is recommended [ 25 ]. Hence it is evident that follow-up strategy should be considered also in H. pylori negative cases in accordance with endoscopic and histological findings. Our study demonstrates that evaluation of the gastric mucosa with a focus on neutrophil and lymphocyte infiltration and epithelial damage is specific and sensitive for diagnosing H. pylori infection even after treatment, and that the diagnostic value of a histology-based decision is high. Today, the value of mucosal specimens for the post-treatment histological diagnosis of H. pylori is considered low assuming that H. pylori colonisation may be patchy, or coccoid forms are difficult to detect [ 25 ]. We have excluded patchy damage by using 13 C-UBT in parallel with histological investigation. Next, for detecting the coccoid forms of the bacteria, we used additionally PCR method. Our results show that the histological finding of H. pylori completely correlates with the results of 13 C-UBT and PCR both in H. pylori positive and negative cases. This confirms the validity of the histological evaluation of mucosal specimens in the case of recurrent peptic ulcer or erosions. Moreover, in countries with a high rate of H. pylori infection and gastric cancer, it is especially important to follow up patients for detecting dysplasia and malignancies [ 26 - 29 ]. Surprisingly, brush cytology from the mucosa failed to detect H. pylori in cases where it was found by other methods. Cytology is highly evaluated for detection of H. pylori infection, as its agreement with histology is considered to be 100% [ 30 ]. Our results show that when patients had been treated with antibacterial drugs and still had dyspeptic complaints, cytological examination was not suitable for H. pylori detection, as different forms of the bacteria were found. The morphology of the helicobacters could have been modified for coccoid or otherwise non-typical forms. It is possible that some other bacteria might have colonised the mucosa due to reduced colonisation resistance after antibacterial treatment, failure of some intestinal functions or usage of medicines administered to relieve the feeling of discomfort [ 31 - 33 ]. Bacteriological investigation enabled to find a few clarithromycin resistant H. pylori strains, which may result in the failure of repeat triple therapy. As the macrolide clarithromycin is chemically stable and well tolerated [ 34 ], physicians often choose it for treatment of different infections. Therefore, if the physician plans to use macrolides, endoscopy and histological testing should be accompanied by bacteriological investigation. Regarding PCR, its main value, obtaining of fast results, is evidently not so important in post-treatment settings. Conclusions In a late post-treatment setting, patients with dyspepsia should not be monitored only by non-invasive investigation methods; it is also justified to use the classical histological evaluation of H. pylori colonisation, PCR and bacteriology as they have shown good concordance with 13 C-UBT . Moreover, endoscopy and histological investigation of a gastric biopsy have proved to be the methods with an additional diagnostic value, providing the physician with information about inflammatory, atrophic and metaplastic lesions of the stomach in dyspeptic H. pylori positive and negative patients. Bacteriological methods are suggested for detecting the putative antimicrobial resistance of H. pylori , aimed at successful eradication of infection in persistent peptic ulcer cases. Competing interests The author(s) declare that they have no competing interests. Authors' contribution HIM, IK and KLa carried out GSRS, endoscopy and gastrobiopsy. HK recruited patients, collected 13 C-urea breath tests, and performed GSRS. KLõ carried out bacteriological examination. PH performed cytological examination. HA carried out molecular analysis and participated in the writing of the manuscript. HIM performed histological examination and statistical analysis, and participated in the design of the study and in the writing of the manuscript. MM coordinated the study and participated in the completion of the manuscript. All authors have read and approved the final version of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529255.xml

548132

Expression of Bcl-2 and p53 at the fetal-maternal interface of rhesus monkey

To study the apoptosis and its mechanism at the fetal-maternal interface of early gestation, localization of apoptotic cells in the implantation sites of the rhesus monkey on day 17, 19, 28 and 34 of pregnancy were first examine by using the TUNEL technique. The expression of Ki67, a molecular marker of proliferating cells, and two apoptotic proteins, B cell lymphoma/leukaemia-2 (Bcl-2) and P53, were then studied by immunohistochemistry. Apoptotic nuclei were observed mainly in the syncytiotrophoblast. Ki67 was confined almost exclusively to cytotrophoblasts. The localization of Bcl-2 protein follows that of the apoptotic nuclei and its expression level increased as the development of the placenta progressed on. P53 was detected to some extent in cytotrophoblasts and syncytiotrophoblast covering the basal feet of the anchoring villi during the late stage of placentation. Based on these observations, it might be suggested that Bcl-2 could be possible to play an interesting role in limiting degree of nuclear degradation and sustaining cell suvival in the multi-nucleated syncytiotrophoblast cells during early pregnancy, and P53 could also be essential in regulating the trophoblastic homeostasis by controlling its proliferation or apoptosis.

Introduction Apoptosis plays important roles in placentation and embryonic development [ 1 ]. The cells undergoing apoptosis have characteristic structural changes in the nucleus and cytoplasm. The nuclear disintegration involves DNA cleavage into oligonucleosomal length DNA fragments [ 2 - 4 ], and the DNA fragments can be detected by terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP) nick end-labelling (TUNEL) technique. Expressions of apoptotic regulatory molecules, such as Fas, Fas ligand, P53, and the proteins of Bcl-2 family, have been reported in human placenta [ 5 - 8 ]. Bcl-2 and P53 are two of the key players in the apoptotic signaling cascades. Bcl-2, a proto-oncogene first discovered in human follicular lymphoma [ 9 ], is involved in the inhibition of apoptosis and the survival of a variety of cell types [ 10 ]. Bcl-2 protein is located in the membranes of endoplasmic reticulum, nuclear envelope, and mitochondria. Over-expression of Bcl-2 suppresses apoptosis by preventing the activation of caspases that carry out the process. P53 is well known as a tumor suppressor. It is a transcription factor that induces apoptosis mainly through inducing the expression of a batch of redox-related genes [ 11 ] and the down-regulating Bcl-2 [ 12 ]. The expression of Bcl-2 and P53 human placenta has been studied [ 1 , 13 ]. However, their cellular distribution in the implantation site at early stage of pregnancy has not been reported. Because the monkey and the human share a very similar implantation process in terms of timing, morphological changes, and cell types involved [ 14 ], we aimed, in the present study, to investigate the expression, localization of Bcl-2 and P53 in the implantation site of the rhesus monkey, in order to gain some insights to the mechanism of time-dependent apoptosis occurring at the fetal-maternal interface. Materials and methods Animals Healthy adult male and female rhesus monkeys ( Macaca mulatta ) were purchased from the monkey colony of the Primate Research Center (PRC), Kunming Institute of Zoology (KIZ), Chinese Academy of Sciences (CAS). All experimental procedures were approved by the Animal Ethics Committees of both the Institute of Zoology and PRC. The animals were caged individually and were evaluated daily by visual examination of the perineum for menses, with the onset of menses defined as Day 1 of the menstrual cycle. Adult female monkeys with regular menstrual cycles of approximately 28 days were chosen for this study. Female monkeys on Day 11 of their menstrual cycle were caged with a male monkey of proven fertility from previous mating for 3 days. Vaginal smears were examined the next morning for the presence of sperm. The day when the smear was detected as positive for sperms was designated as Day 1 of pregnancy (D1). The presence of a conceptus was confirmed by ultrasound examination. The monkeys were anesthetized by pentobarbital sodium (3 animals each group), and the uteri were removed surgically from early villous to villous placenta stages: on D17, D19, D28 and D34 of pregnancy respectively and cut into pieces, the specimens were quickly washed in cold phosphate-buffered saline (PBS) to remove adherent blood, then placed in cold 4% paraformaldehyde fixative for 16 h at 4°C and further processed through graded dehydration, clearing and embedding in paraffin for immunohistochemistry and TUNEL assay. Part of the specimen was cryopreserved at -70°C for Western blot analysis. Reagents Primary antibodies including rabbit anti-human P53 (SC-6243) and mouse anti-human Bcl-2 (SC-7382) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Their quality and specificity were confirmed by the result of the Western blot analysis of D34 placental proteins (Figure 1 ). Rabbit anti-human cytokeratin (ZA0070), mouse anti-human actin (ZA0001), and mouse anti-Ki67 (ZM0166) were purchased from Zymed laboratories (San Francisco, CA, USA). Biotin labeled secondary antibodies, alkaline phosphatase (AP) conjugated avidin, horseradish peroxidase (HRP) conjugated goat anti-rabbit IgG, HRP-conjugated horse anti-mouse IgG, and AP substrates "Vector-red" were from Vector Laboratories (Burlingame, CA, USA). Digoxigenindideoxy (DIG)-11-dUTP, TdT, blocking reagent, AP-conjugated anti-DIG antibody and 5-bromo-4-chloro-3-indoxyl phosphate/nitro-blue tetrazolium chloride (BCIP/NBT) were purchased from Roche-Boehringer-Mannheim (Mannheim, Germany). Proteinase K was purchased from Merck-Schuchardt (Darmstadt, Germany). Levamisole were purchased from Sigma-Aldrich (St Louis, MO, USA). SuperSignal ® West Pico substrate was from PIERCE Biotechnology (Rockford, IL, USA). Figure 1 Western blot analysis of Bcl-2 and P53 for monkey tissue obtained on D34 of pregnancy . Specific signals of Bcl-2 and P53 proteins were detected. No band was found in the control when antibodies were replaced with normal IgG of the same concentration and origin. Western blot Western blot was done as previously described [ 15 ] with slight modifications to verify the cross-reactive specificity of the antibodies with the monkey tissue. The tissue of implantation sites on D34 of pregnancy was homogenized and the supernatant (50 μg) from centrifugation was run on a 10% SDS-PAGE gel under reduced conditions. After being transferred to the polyvinylidene difluoride membrane, individual lanes were cut and blocked with 5% nonfat milk/PBS for 1 h, followed by incubation at 20°C for 1 h with the primary antibodies (IgG, 0.2 μg/ml) in 5% milk/PBS. The membranes were washed three times, 5 min for each, in 5% milk/PBS and incubated with HRP-conjugated horse anti-mouse IgG (0.2 μg/ml, for Bcl-2) or HRP-conjugated goat anti-rabbit IgG (0.04 μg/ml, for P53) in 5% milk/PBS for 1 h respectively. The membranes were washed in PBS three times 5 min for each, followed by 10 min of incubation with SuperSignal ® West Pico substrate, then exposed on x-ray film. For negative controls, primary antibodies were replaced with normal IgG of the same concentration and origin. TUNEL Apoptotic cells were identified by using the TUNEL technique [ 1 , 16 ]. The procedure was slightly modified based on Gao et al. [ 17 ] as the following. Deparaffinized and hydrated 4 μm sections were first treated with 10 μg/ml proteinase K at 37°C for 20 min, and then subjected to 3'-end-labelling of the DNA with 1 μM DIG-11-dUTP and 1 U/μl TdT at 37°C for 1 h. The sections were washes three times in Tris buffer, and incubated with blocking buffer (100 mM Tris, 150 mM NaCl, pH 7.5, and 1% blocking reagent) for 30 min at room temperature. Next, sections were incubated with the primary AP-conjugated anti-DIG antibody (1:500 in 1% blocking reagent, 100 mM Tris, and 150 mM NaCl, pH 7.5) at room temperature for 2 h, and then washed with Tris buffer. Staining was developed using the standard substrates NBT (337.5 μg/ml) and BCIP (175 μg/ml). Negative controls were similarly processed with the omission of TdT. Immunohistochemistry Serial 4 μm sections of tissue were deparaffinized and rehydrated through degraded ethanol. Antigen retrieval was performed by incubating the sections in 0.01 M citrate buffer (pH 6.0) at 98°C for 20 min followed by cooling at room temperature for 20 min. Non-specific binding was blocked with 5% (v/v) normal goat serum in PBS for 1 h. The sections were incubated with primary antibodies specific for P53 (1 μg/ml), Bcl-2 (2 μg/ml) or Ki67 (2 μg/ml) respectively in 2% goat serum overnight at 4°C. Sections were then washed three times with PBS (10 min each) and incubated with biotinylated secondary antibody (2 μg/ml) at RT for 30 min. 3 × 10 min successive washes were followed by incubation with avidin-AP complex (1:200, RT, 20 min). Sections were developed with standard substrates (337.5 μg/ml NBT and 175 μg/ml BCIP) or Vector Red AP substrates according to the manufacturer's protocol after another three washes. Endogenous AP activity was inhibited by supplement of 1 mM levamisole into substrate. The sections stained with Vector Red substrates were counter-stained using haematoxylin. Sections incubated with normal IgG instead of primary antibody served as negative controls. A double immunostaining technique using the antibodies to cytokeratin and actin was performed to localize the extravillous endovascular trophoblast cells. De-paraffinized sections were incubated with 3% H 2 O 2 in methanol for 10 min at room temperature to quench endogenous peroxidase after antigen retrieval treatment as described above. To detect the cytokeratin signal, the sections were washed (3 × 10 min in PBS), blocked for nonspecific signals, incubated sequentially with primary anti-human cytokeratin anbibody (1 μg/ml, RT, 1 h), secondary biotinylated goat anti-rabbit IgG (2 μg/ml, RT, 30 min), and avidin-peroxidase complex (1:200, RT, 20 min), and developed with DAB substrate solution in a similar way as described above. To detect actin signal, the procedure was repeated one more time with anti-human actin antibody (1 μg/ml, RT, 2 h) as primary antibody, AP conjugated horse anti-mouse IgG (1 μg/ml, RT, 40 min) as secondary antibody, and Vector Red developing AP substrate. As a result, the trophoblast cells were labeled brown and the blood vessel wall red. Microscopic assessment Placental samples from three individual monkeys for each group were analyzed. Experiments were repeated at least three times, and one representative from at least three similar results was presented. The mounted sections were examined using a Nikon microscope. For Ki67, the percentages of immunoreactive cells were assessed on at least 2000 cells in each tissue section; For TUNEL, the percentages of positive nuclei were assessed out of at least 2000 nuclei in each tissue section; For assessment of Bcl-2 staining intensity in cells of different compartments, semi-quantitative subjective scoring was performed by three blinded investigators using a 4-scale system with "-"= nil; "+/-"= weak; "+" = moderate; and "++" = strong as described by Yue et al. [ 18 ]. Results Apoptosis in implantation site of early pregnancy The TUNEL technique was used to identify cell types that underwent apoptosis in the implantation site of rhesus monkey on D17, D19, D28 and D34 of pregnancy. On D17 and D19, apoptotic nuclei were observed in the syncytiotrophoblast layer covering the basal feet of the anchoring villi (Figure 2 A, B , arrowhead) and in the villous stromal cells (Figure 2 A , arrow), but not in the cytotrophoblasts. The positive nuclei in the syncytiotrophoblast was only about 0.06%. On D28 and D34, the apoptotic nuclei were present in the syncytiotrophoblast covering the villi (Figure 2 C, D ), in the villous stromal cells (Figure 2 C, D , arrow), in the syncytiotrophoblast layer covering the basal feet of the anchoring villi (Figure 2 E ), and in the cytotrophoblasts within the cell columns (Figure 2 F ). On D28, the percentage of TUNEL-positive nuclei in the syncytiotrophoblast was 0.21%. As pregnancy progresses, the percentage increased to 0.34% on D34. In maternal compartment, a lot of apoptotic nuclei were detected in the stromal cells (Figure 2 G ) and glandular epithelium (Figure 2 H ). Figure 2 Apoptosis detected by TUNEL at the implantation sites of the rhesus monkey on D17 (A), D19 (B), D28 (C, G) and D34 (D, E, F, H) of gestation. Apoptotic nuclei were stained dark. Arrowhead and arrow in panel A – D indicated the nuclei of syncytiotrophoblast and villous stromal cells respectively. The insets in C and D showed the positive nuclei under a higher magnification. Note the syncytiotrophoblast layer covering the basal feet of the anchoring villi in E and the cell columns in F. G and H represent the stromal cells and glandular epithelial cells respectively in the endometrium. I was the negative control. St, syncytiotrophoblast; CT, cytotrophoblast; Vi, placental villi. Scale bars represent 50 μm. Proliferative activity in implantation site at early pregnancy Ki67 is a protein expressed in cycling cells from G1 to M phases and is widely used as a roliferative marker (19, 20). As shown in Figure 3 , Ki67 was expressed in the cytotrophoblasts and the villous stromal cells, but not in the syncytiotrophoblasts. As pregnancy progresses, the percentage of Ki67-positive cytotrophoblast cells lining the villi decreased from more than 85% on D17 to less than 25% on D34 (Panel A-D, and E, F for a higher magnification). However, the cytotrophoblasts at the proximal tip of cell columns remained highly proliferative (more than 70%) at all stages (Panel G and H). Figure 3 The proliferating activity revealed by Ki-67 immunostaining at implantation sites of the rhesus monkey on D17 (A, E, G), D19 (B), D28 (C) and D34 (D, F, H) of gestation. Panels A-D were under a lower magnification. Ki-67 protein was stained red, and nuclei blue. E and F were the placental villi under a higher magnification. G and H were the anchoring villi under a higher magnification. Vi, placental villi. ST, syncytiotrophoblast. CT, cytotrophoblast. Sc, stromal cell. En, endometrium. Scale bars represent 100 μm. Bcl-2 expression in implantation site at early pregnancy In order to study the mechanisms of the apoptosis observed at the fetal-maternal interface, the expression of Bcl-2 was investigated by using immunohistochemistry. At the early stages of placentation (D17, D19), Bcl-2 was only detected in the syncytiotrophoblast covering the cell columns (Figure 4, A and 4B ) and the extravillous cytotrophoblast (Figure 4C , arrow). At the later stages (D28, D34) it was detected in all the syncytiotrophoblast (Figure 4,D and 4E ), the villous stromal cells (Figure 4F , arrow), and the extravillous endovascular trophoblast cells (Figure 4G ), the fetal origin of these cells were indicated by the anti-cytokeratin antibody staining (Figure 4 G inset, brown), and the vascular wall was stained by anti-actin antibody (red). The pattern of Bcl-2 expression in the syncytiotrophoblast was similar to that of the apoptotic nuclei distribution (Figure 2 ). In the maternal compartment, Bcl-2 could be detected in some stromal cells (Figure 4H ). Notably, the cytotrophoblasts lining the villi, within the cell columns, and the glandular epithelia were negative for Bcl-2 staining. The semi-quantitative expression level of Bcl-2 in different cell types at the various stages was summarized in Table 1 . A gradual increase of Bcl-2 staining was observed in the syncytiotrophoblast as gestation advances. Figure 4 Immunohistochemical staining for Bcl-2 at implantation sites of the rhesus monkey. Bcl-2 staining is red, and nuclear counterstain blue. A, villous plancenta on D17. B, villous plancenta on D19. C, extravillous trophoblast cells in the basal plate of D17. D, villous plancenta on D28. E, villous plancenta on D34. F, villous plancenta on D34 under a higher magnification. G, the extravillous endovascular trophoblast cells; in the inset, the fetal origin of these cells was confirmed by anti-cytokeratin antibody (brown), and their position within the vascular wall was confirmed by anti-actin antibody staining (red). H, decidua. I, negative control. Vi, placental villi. ST, syncytiotrophoblast. CT, cytotrophoblast. Sc, stromal cells. Ge, glandular epithelium. Evc, extravillous cytotrophoblast. Scale bars represent 50 μm. Table 1 Semi-quantitative assessment of the immunohistochemical staining of Bcl-2 in the placenta of rhesus monkey. D17 D19 D28 D34 Syncytiotrophoblast lining the villi +/- +/- ++ ++ Syncytiotrophoblast covering the cell column + + ++ ++ cytotrophoblast lining the villi - - - - extravillous cytotrophoblast + + + + P53 expression in implantation site at early pregnancy The expression profile of P53 was also acquired by using the immunohistochemistry. On D17 and D19, the expression of P53 was only confined to a small number of nuclei in the syncytiotrophoblast (Figure 5,A,B ). On D28 and D34, its expression was observed not only in the syncytiotrophoblast (Figure 5C,D ) but also in the nuclei of cytotrophoblasts lining the villi (Figure 5E ) and within proximal tip of cell columns (Figure 5F ) where a proliferative activity was high as indicated by Ki67 staining (Figure 3 ). Clustered P53-positive nuclei were seen in the syncytiotrophoblast covering the basal feet of the anchoring villi (Figure 5G ), coincident well with the strong apoptosis detected by TUNEL (Figure 2E ). P53 was also expressed in some stromal cells (Figure 5H ) of the uterine endometrium. Figure 5 Immunohistochemical staining for P53 at implantation sites of the rhesus monkey on D17 (A), D19 (B), D28 (C, H), and D34 (D, E, F, G) of gestation. P53 was stained dark in nuclei. A-D were villous placenta under a lower magnification. The inset of panel A shows the staining in the syncytiotrophoblast covering the basal feet of the anchoring villi under a higher magnification. E, staining in villous placenta under a higher magnification. F, staining in cell columns. G, syncytiotrophoblast covering the basal feet of the anchoring villi under a higher magnification. H, the endometrium with arrows indicating stromal cells. ST, syncytiotrophoblast. CT, cytotrophoblast. Scale bars represent 50 μm. Discussion For the first time in present study, we investigated the expression of Bcl-2 and P53 in relation to apoptosis at the fetal-maternal interface of rhesus monkey at the very early stages (D17-D34) of gestation. Villous trophoblasts consist of cytotrophoblasts and syncytiotrophoblast. While cytotrophoblasts possess a brisk mitotic activity during the first trimester of gestation in human, the syncytiotrophoblast is incapable of cell division despite of a metabolic activity [ 1 ]. This fact implies that cell proliferation is differently regulated in these two cell types. The reports on the type of trophoblast cells undergoing apoptosis in the first trimester are controversial [ 1 , 21 , 22 ]. Our results further cleared that the apoptotic nuclei were distributed mainly in the syncytiotrophoblast at the early stages and in the cytotrophoblasts within the cell columns at later stages of pregnancy. In our previous study, Bax expression was found at the Fetal-Maternal Interface of Rhesus Monkey [ 17 ]. Bax is a Bcl-2 family member that promotes cell death susceptibility, possibly by countering the effect of Bcl-2 on cell survival through heterodimer interaction. Bax to Bcl-2 "rheostat" may be a critical factor in regulating apoptosis in multiplicate tissues. As shown in Figure 6 , Bax was found expressed in the placenta and glandular epithelium of endometrium and all kinds of cells in placental villi, and no obvious change was observed between different time points from D17 to D34 in placental villi. Therefore, we speculated that Bcl-2 may play a more important role on controlling the apoptosis in placental villi. The diffusive expression of Bcl-2 in syncytiotrophoblast obtained from the first trimester human placenta has been reported recently [ 7 , 23 - 25 ]. Our observation on the Bcl-2 expression in syncytiotrophoblast at later stages (D28-D34) agreed well with these data. As shown in this study, although Bcl-2 was expressed, apoptotic nuclei still exsisted in the same region. This phenomenon implies that the expression of Bcl-2 is not sufficient to completely inhibit the apoptosis in the syncytiotrophoblast. Therefore, the role of Bcl-2 here becomes an interesting question. Multiple nuclei sharing the same cytoplasm is a morphological characteristic of syncytiotrophoblast. In such cells, the apoptotic signal may be transmitted from one nuclear to another, and cause a spontaneous abortion. Therefore, the number of nuclei undergoing apoptosis in the syncytiotrophoblast should be limited by some mechanism in order to ensure normal embryo development in normal pregnancy [ 1 ]. We speculate that Bcl-2 may be included in this mechanism. The major role of apoptosis-associated Bcl-2 expression in the syncytiotrophoblast might be to limit the nuclear degradation to a special area and inhibit the spread of cell apoptosis signals to the other nuclei sharing the same cytoplasm, thus sustain cell survival in these multi-nucleated cells. Toki et al has also suggested that Bcl-2 might play a major role in avoiding the possible excessive nuclear degradation in syncytiotrophoblast [ 26 ]. Further studies, however, are needed to prove this speculation. The immunostaining for Bcl-2 was also detected in part of the extravillous and endovascular cytotrophoblast in our study. These subtypes of cytotrophoblast lost the capacity of proliferation (Ki-67-negative), but they did not undergo apoptosis (negative in TUNEL assay). Therefore, we hypothesize that Bcl-2 may also participate in regulation of the extravillous trophoblast apoptosis by stimulating the cellular survival. Figure 6 Immunohistochemical staining for Bax at implantation sites of the rhesus monkey on D28. Bax staining is brown, and nuclear counterstain blue. A, villous plancenta, positive staining was found in all the cells. B, endometrium, glandular epithelium was positive for Bax staining. Vi, placental villi. Ge, glandular epithelium. P53 was partly identified in some nuclei of the syncytiotrophoblast with the same position of apoptotic nuclei, in the basal feet of the anchoring villi in particular, but it is not clear whether the P53 was co-localized with the apoptotic signals. Activation of P53 in some cell types leads to either the cessation of cell growth or apoptosis [ 27 ]. Therefore, P53 protein might be related to cell cycle arrest or apoptosis in syncytiotrophoblast during early stage of placentation. Low level of P53 staining was detected in the cytotrophoblasts during the earlier stages of gestation (D17 and D19). However, at the later stages (D28 and D34), the expression was observed predominantly in the nuclei of cytotrophoblasts. The presence of P53 in cytotrophoblast in the primate was consistent with that observed in the human first trimester placenta [ 8 ]. Indeed, the TUNEL staining showed that the apoptosis seldom happened in the cytotrophoblast, with the exception of cytotrophoblast at proximal tip of cell columns during later stages of placentation (D28, D34) where a high proliferative activity and P53 expression were detected. This finding supports the hypothesis that a physiological upregulation of the P53 tumour suppressor gene might be a mechanism for controlling excessive trophoblastic proliferation in normal placentation [ 26 , 28 ]. It is known that early pregnancy is unique in its methods of cell proliferation control, the existing data suggest that some growth factors and transcription factors from the embryo and endometrium, such as CSF-1, VEGF, and transcription factors of the helix-loop-helix family, provide at least part of this control [ 29 ]. In addition, other studies found maternal age and some diseases, such as diabetes can also influence the apoptotic and proliferative activities in trophoblast cells [ 30 , 31 ]. Further investigations are required to uncover which endocrine event regulates the expression of Bcl-2 and P53.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548132.xml

529269

Detecting imbalanced expression of SNP alleles by minisequencing on microarrays

Background Each of the human genes or transcriptional units is likely to contain single nucleotide polymorphisms that may give rise to sequence variation between individuals and tissues on the level of RNA. Based on recent studies, differential expression of the two alleles of heterozygous coding single nucleotide polymorphisms (SNPs) may be frequent for human genes. Methods with high accuracy to be used in a high throughput setting are needed for systematic surveys of expressed sequence variation. In this study we evaluated two formats of multiplexed, microarray based minisequencing for quantitative detection of imbalanced expression of SNP alleles. We used a panel of ten SNPs located in five genes known to be expressed in two endothelial cell lines as our model system. Results The accuracy and sensitivity of quantitative detection of allelic imbalance was assessed for each SNP by constructing regression lines using a dilution series of mixed samples from individuals of different genotype. Accurate quantification of SNP alleles by both assay formats was evidenced for by R 2 values > 0.95 for the majority of the regression lines. According to a two sample t-test, we were able to distinguish 1–9% of a minority SNP allele from a homozygous genotype, with larger variation between SNPs than between assay formats. Six of the SNPs, heterozygous in either of the two cell lines, were genotyped in RNA extracted from the endothelial cells. The coefficient of variation between the fluorescent signals from five parallel reactions was similar for cDNA and genomic DNA. The fluorescence signal intensity ratios measured in the cDNA samples were compared to those in genomic DNA to determine the relative expression levels of the two alleles of each SNP. Four of the six SNPs tested displayed a higher than 1.4-fold difference in allelic ratios between cDNA and genomic DNA. The results were verified by allele-specific oligonucleotide hybridisation and minisequencing in a microtiter plate format. Conclusions We conclude that microarray based minisequencing is an accurate and accessible tool for multiplexed screening for imbalanced allelic expression in multiple samples and tissues in parallel.

Background Single nucleotide polymorphisms (SNPs) are highly abundant in the human genome, appearing on average at 0.1% of the nucleotide positions [ 1 ]. Thus, each gene or transcriptional unit will contain multiple SNPs that potentially give rise to sequence variation between individuals and tissues on the level of RNA. Recent studies indicate that differences in the expression levels of the alleles of heterozygous SNPs may occur frequently for human genes [ 2 - 6 ]. Imbalanced allelic expression was detected in foetal liver or kidney tissues for more than half of 602 genes analysed, and one third of the genes displayed more than four-fold differences in allelic expression [ 3 ]. Another study detected lower levels of allelic imbalance for one fifth of 129 genes analysed in lymphoblastoid cell lines [ 4 ]. Non-synonymous SNPs in coding regions of genes may be functional by altering an amino acid, which in turn may affect the structure and function of the encoded protein, while synonymous SNPs may have functional consequences by affecting the stability or folding of mRNA transcripts. Intronic SNPs may give rise to alternatively spliced mRNAs, while SNPs in 5'- or 3'-untranslated mRNA regions may affect the stability or processing of the RNA. Moreover, SNPs in non-protein coding regions of genes that affect binding of regulatory factors may cause imbalanced expression of SNP alleles. This form of genetic variation has been suggested as a common cause of both normal and disease-related inter-individual variation in complex phenotypes [ 7 ]. Clearly, methods with high accuracy that can be used in a high throughput setting are needed for systematic surveys of expressed sequence variation and its molecular causes. Owing to the high sequence specificity of nucleotide incorporation by DNA-polymerases, single nucleotide primer extension has proven to allow quantitative determination of SNPs in genomic DNA in several studies and assay formats (for a review, see Syvänen 2001 [ 8 ]). A frequently used quantitative application of the method is to determine SNP allele frequencies in pooled DNA samples [ 9 - 13 ]. The rationale for detecting imbalanced expression of the two alleles of a heterozygous SNP by minisequencing is to measure the ratio between the amounts of labelled nucleotides incorporated in the minisequencing reactions for the two SNP alleles in RNA (cDNA) samples from the tissue of interest. These ratios are then compared to the corresponding ratio measured in genomic DNA, where the two alleles are present in an equimolar ratio [ 2 , 4 , 14 - 16 ]. Imbalanced expression of the alleles of a SNP is revealed by a difference in the ratios measured in the RNA and DNA samples. We are currently using microarray based minisequencing for multiplex genotyping of SNPs. Our custom-made microarrays permit the genotyping of up to 100 SNPs in 80 samples per standard microscope slide, either using immobilised minisequencing primers [ 17 , 18 ] or using a "tag-array" format [ 19 , 20 ] of the method [ 13 ]. The purpose of this study was to evaluate the performance of these two microarray formats in quantitative determination of SNP alleles on the RNA level as alternatives with higher multiplexing capacity than previously used primer extension methods in which the SNPs are analysed in individual reactions. Using these systems, we were able to detect significant differences in the amounts of the two alleles of heterozygous SNPs on the RNA level. Results We used a panel of ten coding SNPs in five genes to choose the optimal microarray based minisequencing strategy for multiplex, quantitative genotyping of SNPs in DNA and RNA samples. The selected SNPs were located in genes shown by reverse transcriptase PCR analysis to be expressed in one or both of two endothelial cells lines, HUVEC (human umbilical vein endothelial cells) and HAEC (human aortic endothelial cells) that served as our model cell lines in this study (data not shown). We evaluated two formats of microarray based minisequencing by performing five parallel assays with each method for each sample in the evaluation. The SNPs were analysed in both DNA polarities and the evaluation of the methods was based on the DNA polarity yielding the highest signal-to-noise ratio. In Method I, immobilised minisequencing primers are extended with fluorescently labelled ddNTPs in reactions performed on the microarray surface after annealing of the multiplex PCR products to the primers [ 18 , 21 ]. In Method II, cyclic primer extension reactions are performed in solution in the presence of 5'-tagged minisequencing primers, PCR products and fluorescent ddNTPs [ 22 , 23 ]. After the cyclic reactions the extended primers are captured on a microarray surface carrying immobilised oligonucleotides complementary to the 5'-tag sequences on the minisequencing primers. Both these systems are performed in an "array-of arrays" format developed previously in our laboratory [ 24 ]. We analysed a dilution series with mixtures of DNA from two individuals with different genotypes for the panel of ten SNPs in both DNA polarities. The genotyping results from these mixtures of known amounts of the two SNP alleles are expressed as the signal ratio between the fluorescence signals corresponding to the two alleles of each SNP. The quantitative analysis of these ten SNPs is illustrated in Figure 1 by regression lines, in which the mean signal intensity ratios are plotted as a function of the known allelic ratios in the mixed samples. The coefficient of determination (R 2 ), which describes how well the regression line fits the data points, was used to assess the accuracy of quantification of the SNP alleles by Methods I and II. As can be seen in Table 2 , the R 2 values are close to one for most of the SNPs analysed, demonstrating little scatter of the data points around the regression line. For Method I, six of the ten SNPs analysed have R 2 values ≥ 0.95, while for Method II the R 2 values are ≥ 0.95 for eight of the SNPs. Thus, accurate quantification of SNP alleles is possible by both methods. The slopes of the regression lines vary between the ten SNPs as well as between the two methods (Figure 1 ). A regression line with a steep slope usually corresponds to a high R 2 value, as observed for the SNP rs5930 LDLR analysed by Method I and SNP rs5331 EDNRB analysed by Method II. A flat slope does not necessarily imply less accurate quantification, as exemplified by the SNP rs4331 ACE, where Method II yielded a flat slope with a higher R 2 value than Method I. We also determined the sensitivity of the methods for detection of a minority allele. The detection limit was defined as the percentage of the minority allele in the mixed sample, for which the signal ratio differed from the signal ratio in the corresponding homozygous sample with a p-value < 0.05 in a two sample t-test. Depending on the genotype of the DNA samples used for the dilution series, determination of the lower limit of detection was possible for seven of the ten SNPs with allele ranges 0–50% or 0–100% in the mixed samples (Table 1 ). For the remaining three SNPs with the allele range 50–100%, the smallest percentage of an allele that could be distinguished from a heterozygous genotype was identified by the same approach. Using Method I, we were able to detect less than 5% of the minority allele for two SNPs (rs1042713 ADRB2 and rs5925 LDLR) and less than 9% for rs4331 ACE, rs1042719 ADRB2, rs5351 EDNRB and rs5930 LDLR (Table 2 ). Method II allowed more sensitive detection of minority alleles than Method I. Less than 2% was detectable for the SNPs rs1042713 ADRB2, rs1042719 ADRB2 and rs5351 EDNRB, and less than 9% was detectable for the SNPs rs4331 ACE, rs5925 LDLR, rs5930 LDLR and rs1433099 LDLR (Table 2 ). For the SNPs rs1042714 ADRB2, rs1042718 ADRB2 and rs1799983 NOS3, we were able to measure 4–14% deviations from the heterozygous genotype (Table 2 ). These results show that the amount of SNP alleles can be accurately determined on the DNA level by Methods I and II using reference samples with the two SNP alleles present in known ratios. Next, the performance of the two methods in quantitative analysis on the RNA level was assessed. The ten SNPs were first genotyped in genomic DNA (gDNA) from the HUVEC and HAEC cells to identify those SNPs that were heterozygous in either or both cell lines. Three SNPs in the low density lipoprotein receptor gene (LDLR; rs5925, rs5930 and rs1433099) were heterozygous in the HAEC cell line, and one SNP in each of the genes encoding angiotensin I converting enzyme (ACE rs4331), β 2 -adrenergic receptor (ADRB2 rs1042719) and endothelin receptor type B (EDNRB rs5351) were heterozygous in the HUVEC cell line. These SNPs were genotyped in cDNA produced from total RNA extracted from the cells with the corresponding gDNA as reference samples using both methods. Table 3 presents the mean fluorescence signals with coefficients of variation (CV) obtained in five parallel reactions for the six SNPs in cDNA and gDNA from the HUVEC and HAEC cells. For the heterozygous SNPs the largest difference in the variability between parallel reactions was observed between SNPs, with the lowest CV values (3.6 – 8.6 %) for the rs1042719 ADRB2 SNP, and the highest CV values (13 – 41%) for the rs1433099 LDLR SNP. No systematic differences in the variability of parallel reactions were observed between Method I and Method II, or between cDNA and gDNA. Table 4 shows the differences in mean signal intensity ratios between the cDNA and gDNA assays for the six SNPs that were heterozygous in HUVEC or HAEC cells, respectively, together with the corresponding normalized cDNA/gDNA ratios. The SNPs in the ACE, ADRB2 and EDNRB genes displayed significant imbalanced expression in the HUVEC cells using both methods. For the SNP rs4331 ACE, the signal intensity ratio based on the raw data obtained by Methods I and II differed from each other, but despite this large difference, both methods yielded similar levels of allelic imbalance for this SNP after normalisation against the signal ratio in gDNA (Table 4 ). Only for one of the three LDLR SNPs (rs5930), the difference in fluorescence intensity ratios between cDNA and gDNA from HAEC cells reached statistical significance by both methods. Allelic imbalance of the LDLR gene was detected for the LDLR SNP rs5925 using Method II only. To test that the results on imbalanced allelic expression detected by the multiplexed microarray based methods represents the true biological situation in the cells, we analysed the heterozygous SNPs in five replicate RNA samples prepared from HUVEC or HAEC harvested at different time points from different cell culture flasks. We also analysed the three LDLR SNPs in five replicate reverse transcription reactions from the same RNA sample prepared from HAEC cells. For this analysis we used our first generation solid-phase minisequencing assay for individual SNPs in a microtiter plate format. The concordant cDNA/gDNA ratios from these control experiments from independent cell and RNA samples presented in Table 5 show that the detected allelic imbalance was not caused by the procedures for RNA extraction or cDNA synthesis. Finally, we verified the results obtained by microarray-based minisequencing for three of the SNPs by real-time PCR with allele specific hybridization probes (TaqMan). Table 4 shows these results together with the corresponding results by solid-phase minisequencing in a microtiter plate format. Allelic imbalance was detected with statistical significance for the SNP rs1042719 ADRB2 and the SNP rs1433099 LDLR by both methods. Particularly for the SNP rs1042719 ADRB2, the cDNA/gDNA ratios obtained by the two reference methods were highly similar to the results from the microarray-based methods presented in Table 4 , as well as with each other. As for the microarray-based Method II, the difference in signal ratios between cDNA and gDNA measured by the TaqMan assay for the SNP rs5925 LDLR did not reach statistical significance due to large variation between parallel assays. Analysis of the SNP rs1433099 LDLR by the reference methods confirms the imbalanced expression of the LDLR receptor alleles. Discussion The purpose of our study was to evaluate microarray based minisequencing for multiplexed detection and quantification of imbalanced expression of SNP alleles, as a prelude to further large scale screening for allelic imbalance. We found no significant differences in the performance of our two "in house" methods, minisequencing with primers directly immobilised on the microarrays (Method I)[ 18 ] and the "tag-array" format, based on cyclic minisequencing followed by capture on microarrays using immobilised complementary "tag" probes (Method II) [ 23 ]. Both methods showed a linear relationship between SNP allele ratios and the signal intensity measured in the four-colour fluorescence minisequencing assay for all SNPs. With respect to accuracy assessed by coefficients of variation (CV) between five parallel assays both methods performed equally well, and the CV values between parallel assays were indistinguishable between genomic DNA and reverse transcribed cDNA samples. The sensitivity of detecting a SNP allele present as a minority in a sample was defined as the percentage for which the signal ratio differed from the signal ratio in the corresponding homozygous sample with a p-value < 0.05 in a two sample t-test. The sensitivity differed between SNPs, and range from 1% to 9%, with a trend to be slightly better using the "tag-array" system (Method II). In several cases the p-values were lower than 0.05 (Table 2 ), which indicates that in practice the sensitivity of detection would be lower than the stringent limit set here. The sensitivity of our multiplex microarray based minisequencing methods compares well with the sensitivity of other single nucleotide primer extension assays performed for individual SNPs in recent studies [ 4 , 25 - 27 ]. It is notable that the largest differences in accuracy and sensitivity were observed between SNPs. Some of the SNP-to-SNP differences are likely due to differences is the accuracy and efficiency of incorporation of the four different fluorescently labelled nucleotide analogues by the DNA polymerase [ 13 , 26 ] as well as to other sequence context dependent factors. The large variation between parallel assays for the SNP rs1433099 LDLR prevented detection of the allelic imbalance for the LDLR gene, while imbalance was detected by the SNP rs5930 LDLR using both methods. This result demonstrates that it is preferable to analyse more than a single SNP in each gene in systematic screening for allelic imbalance in gene expression. As more data from primer extension assays accumulate, it may be possible to improve the accuracy of the system by improving the SNP selection and assay design further with the aid of algorithms developed based on this data [ 28 , 29 ]. Comparison of the relative amounts of the alleles of six SNPs on the RNA (cDNA) level to heterozygote SNPs in genomic DNA revealed four SNPs with imbalanced expression of the two alleles. A three-fold increase in the expression of the T-allele for the SNP rs4331 ACE was the most pronounced difference observed. In our study, 1.4–1.5-fold differences in allelic expression levels were detectable. The sensitivity of detecting a minority allele in our system would allow the distinction between 10-fold reduction in the expression of an allele and monoallelic expression, for example as a result of imprinting. Owing to its potential for high throughput screening of large numbers of samples, we have also performed a preliminary evaluation of the commercial SNPstream genotyping system (GenomeLab, Beckman Coulter) that also utilises the "tag-array" primer extension strategy in a semi-automated 384-well microtiter plate format for detection of imbalanced allelic expression [ 30 ]. The same trend of imbalanced allelic expression was observed for each of the SNPs, which is encouraging for future studies of imbalanced allelic expression in a high throughput semi-automated way. Other studies that have used fluorescent single base primer extension assays report that 1.2 – fold to 1.5 – fold differences in allelic expression are detectable [ 2 , 4 , 5 ]. Primer extension methods based on direct measurement of fluorescent signals, including the microarray-based methods evaluated here, are likely to provide better accuracy and sensitivity for allele quantification than homogeneous primer extension based on fluorescence polarisation [ 31 , 32 ], in which the allele quantification relies on measurement of small differences between large polarization signals. It is also reassuring for future large scale detection of imbalanced allelic expression that the accuracy of our methods seemed to be similar for cDNA and genomic DNA. Analysis of replicate RNA samples from different batches of both cell lines using a microtiter plate format of the minisequencing method evidenced for the biological authenticity of the allelic imbalance detected using minisequencing in the microarray format. The data obtained from independent cell samples also indicate an acceptable reproducibility of RNA extraction, RNA storage and cDNA synthesis. Another important factor besides sample to sample variation that may affect the accuracy of the relative allele quantification is the amount of mRNA subjected to the analysis. At a low copy number of mRNA, the stochastic distribution of the RNA templates may be a major source of variation [ 33 ]. The reason for the large variation between parallel assays for the LDLR receptor gene observed with all four methods used in our study may reflect a low expression level of the LDLR gene in the HAEC cells. Moreover, the amount of gene specific transcript in each RNA sample may vary which makes it difficult to perform balanced multiplex RT-PCRs to screen for allelic imbalances in several genes in one reaction. A similar minisequencing strategy as the one used for determination of imbalanced expression between SNP alleles can also be used for determination of the relative expression levels of highly homologous genes [ 15 ] and for determination of alternatively spliced transcripts [ 34 ], a resolution that is beyond the capacity of traditional microarray based RNA expression profiling. Conclusions Here we demonstrated the applicability of two formats of microarray based minisequencing for detecting imbalanced expression of SNP alleles. The accuracy and sensitivity of both systems allow detection of 1.4- to 10-fold differences in the expression levels of the two alleles of heterozygous SNPs. The microarray-based minisequencing systems utilise widely available reagents and equipment, and can thus easily be established "in-house". Moreover, the system is flexible with respect to number of SNPs and samples to be analyzed. Systematic quantitative screening of genetic diversity on the RNA level in multiple individuals and tissues will be a future approach in the elucidation of the molecular mechanisms that regulate gene expression. Methods DNA and RNA samples DNA samples from 30 volunteer donors were genotyped by Methods I and II to identify individuals of different genotypes for the panel of ten SNPs analysed. The SNPs are described in the section "SNPs and primers" below. DNA (10 ng/μl) from one individual was serially diluted 2:1 into DNA (10 ng/μl) from a second individual, to yield a series of DNA samples with different ratios between the SNP alleles. These mixed DNA samples were used for construction of quantification standard curves. Depending on the genotype of each SNP in the two individuals whose DNA was mixed, dilution series of samples with different allelic ranges were obtained for the ten SNPs, as specified in Table 1 . Human Umbilical Vein Endothelial Cells (HUVEC) and Human Aortic Endothelial Cells (HAEC) (Cascade Biologics, Inc., Portland, OR, USA) were grown in Medium 200 with Low Serum Growth Supplement (LSGS Kit, Cascade Biologics, Inc., Portland, OR, USA) at 37°C in a humidified atmosphere of 5% CO 2 . Cells from the cultures were harvested at 80% confluence according to the manufacturer's instructions. Total RNA was isolated from the cells using the TRIZOL ® Reagent (GIBCO BRL, Paisley, Scotland) and the RNA samples were stored at -70°C until use. High quality RNA with A 260 /A 280 ratio over 1.9 and intact ribosomal 28S and 18S RNA were used for cDNA synthesis. The RNA samples were treated with 1 U RQ1 RNase-free DNase (Promega, Madison, WI, USA) per μg RNA. Two to 2.5 μg total RNA was subjected to first strand cDNA synthesis using SuperScript™ II (RNase H - Reverse Transcriptase, Invitrogen, Carlsbad, CA, USA) reagents in a 20 μl volume. DNA was extracted from the cells using GenElute™ Mammalian Genomic DNA Kit (Sigma, St Louis, MO, USA) and stored at -20°C until use. PCR The fragments comprising the SNPs were PCR-amplified in individual reactions using 10–15 ng genomic DNA or one tenth of the cDNA products, 0.2 mM dNTPs, 1U AmpliTaq ® Gold DNA polymerase (Applied Biosystems, Foster City, CA, USA), 1.5 mM MgCl 2 , and 0.2–0.3 μM of primers in 50 μl of 10 mM Tris-HCl pH 8.3 and 50 mM KCl. The PCR conditions were initial activation of the enzyme at 95°C for 10 min followed by 35 cycles of 95°C for 1 min, 56°C for 1 min and 72°C for 1 min and a final extension at 72°C for 7 min in a Thermal Cycler PTC225 (MJ Research, Watertown, MA, USA). The amplified fragments were combined and concentrated to 60 μl using Microcon ® YM-30 Centrifugal Filter Devices (Millipore Corporation, Bedford, MA, USA). SNPs and primers Ten SNPs located in coding regions of genes known to be expressed in HUVEC and HAEC cells were analysed. Information on the SNPs, including dbSNP [ 35 ] ID number and nucleotide variation is given [see Additional file 1 ] together with the sequences of the minisequencing primers. The primers for PCR and minisequencing were designed using the Oligo Primer Analysis software v6.65 (Molecular Biology Insights Inc., Cascade, CO, USA). Preparation of microarrays The minisequencing primers or the complementary tag-oligonucleotides were covalently immobilised on CodeLink™ Activated Slides (Amersham Biosciences, Uppsala, Sweden) by the mediation of a NH 2 -group in their 5'- or 3'-end, respectively. The oligonucleotides were applied in duplicates to the slides at a concentration of 25 μM in 150 mM sodium phosphate pH 8.5 using a ProSys 5510A instrument (Cartesian Technologies Inc, Irvine. CA, USA) equipped with one Stealth Micro Spotting pin (SMP3B, TeleChem International Inc., Sunnyvale, CA, USA) to minimise the variation between spots in different "subarrays". The oligonucleotides were spotted in an "array-of-arrays" configuration that facilitates analysis of 80 individual samples in parallel on each microscope slide [ 24 ]. In each "subarray" a fluorophore-labelled oligonucleotide was included as a control for the immobilisation process. A reference oligonucleotide, complementary to a synthetic template included in the minisequencing reaction mixtures to monitor the difference in incorporation efficiency of the four nucleotides by the DNA polymerase, was also included in each "subarray". Finally, an oligonucleotide designed not to hybridise to any of the oligonucleotides present in the reaction mixture was included in each "sub-array" to be used for background corrections. After printing, the slides were incubated in a humid chamber for at least 24 hours, followed by treatment with ethanolamine according to the manufacturer's instruction. The slides were then stored desiccated in the dark until use. Minisequencing using immobilised primers (Method I) Aliquots of 7.5 μl of the concentrated PCR products were analysed in five parallel "subarrays" for each sample, essentially as described previously [ 18 ]. The PCR products were allowed to anneal to the immobilised oligonucleotides. After washing, the extension reactions were performed with 0.75 U of Thermo Sequenase™ DNA polymerase (Amersham Biosciences, Uppsala, Sweden) and 0.35 μM Texas Red-ddATP, Tamra-ddCTP, R110-ddGTP and Cy5-ddUTP (Perkin Elmer Life Sciences, Boston, MA, USA) in Thermo Sequenase™ reaction buffer in a total volume of 15 μl, followed by washing of the slide. Minisequencing using "tag-arrays" (Method II) Five parallel reactions with a 4.5 μl aliquot of the concentrated PCR products were analysed for each sample, as described in detail in [ 23 ]. Excess of PCR primers and dNTPs were removed by treatment with 5 U of exonuclease I and 1 U of shrimp alkaline phosphatase (USB Corporation, Cleveland, OH, USA). The cyclic minisequencing reactions were performed in the presence of the 20 tagged primers at 10 nM concentration, 0.1 μM Texas Red-ddATP, Tamra-ddCTP and R110-ddGTP, 0.2 μM Cy5-ddUTP (Perkin Elmer Life Sciences, Boston, MA, USA) and 1 U of Thermo Sequenase™ DNA polymerase (Amersham Biosciences, Uppsala, Sweden) for 55 cycles of 95°C and 55°C for 20 s each. The extension products were allowed to anneal to the immobilised complementary tag oligonucleotides at 42°C for 2.5 hours followed by washing of the slide. Solid-phase minisequencing in a microtiter plate format PCR was run with one of the primers biotinylated. The biotinylated PCR products were immobilised in a microtiter plate coated with streptavidin (Combiplate 8, Labsystems, Helsinki, Finland) and the unbiotinylated strand was removed with alkali treatment [ 9 , 15 ]. The minisequencing mixture, containing the appropriate tritium labelled dNTP (Amersham Biosciences, Uppsala, Sweden), AmpliTaq ® DNA polymerase (Applied Biosystems, Foster City, CA, USA) and the minisequencing primer was added. The extension reaction was allowed to proceed for 10 min at 50°C. The extended primers were released with alkali and the amount of incorporated tritium labelled nucleotide was measured. Hybridisation with allele-specific TaqMan probes Primers and probes for the TaqMan assays were designed by Applied Biosystems as Assay-by-Design (rs1042719 ADRB2 and rs5925 LDLR) or Assay-on-Demand (rs1433099 LDLR) service. The probes for the two alleles were labelled with the reporter dyes FAM and VIC respectively. The sequences of the primers and probes for the SNPs rs5925 LDLR and rs 1042719 ADRB2 are found in [ Additional file 1 ]. The primer and probe sequences for the SNP rs1433099 LDLR were not made available to us by ABI since this SNP is included in their Assay-on-Demand program. Real time quantitative PCR was run in 25 μl TaqMan Universal PCR Master Mix (Applied Biosystems) with 200 nM of both labelled TaqMan probes, 900 nM PCR-primers and 10 ng genomic DNA or one tenth of the cDNA products. The PCR conditions were initial activation of the enzyme at 95°C for 10 min followed by 60 cycles of 95°C for 15 sec and 60°C for 1 min in a ABI7000 instrument (Applied Biosystems, Foster City, CA, USA). The signal intensity ratios were calculated based on normalised ΔRn fluorescence values obtained from the assay during the exponential phase of PCR. The ΔRn values were retrieved from cycle 38 for the SNP rs1042719 ADRB2, cycle 42 for the SNP rs5925 LDLR and cycle 43 for the SNP rs1433099 LDLR. Imbalanced expression of the SNP alleles was determined by a t-test as described below. Signal detection and data analysis In Methods I and II fluorescence was measured using a ScanArray ® Express instrument (Perkin Elmer Life Sciences, Boston, MA, USA) with the excitation lasers Blue Argon 488 nm, Green HeNe 543.8 nm, Yellow HeNe 594 nm and Red HeNe 632.8 nm with the laser power set to 80% and the photomultiplier tube gain adjusted to obtain equal signal intensities from reaction control spots for all four spectra. The fluorescence signals were extracted using the QuantArray ® analysis 3.1 software (Perkin Elmer Life Sciences, Boston, MA, USA). The mean of the fluorescence signals for the duplicate spots was corrected for the average background in each "sub-array" separately. The data was handled and interpreted using the Microsoft ® Excel program. The genotype for each individual SNP was assigned by calculating a ratio between the fluorescence signals for the two alleles. Coefficients of determination (R 2 ) were assigned by linear regression analysis of the relationship between the signal intensity ratios determined from the minisequencing assay and the known allelic ratios in the mixed samples for the quantification standard curves. Two-sample t-tests with two-tailed significance levels assuming unequal variance were performed to determine the lowest level of detection of a specific allele for the quantification standard curves and to evaluate the imbalanced expression of the two alleles of the SNPs in the cell lines. Authors' contributions UL participated in the design of the study and in RNA and DNA extraction, and performed all the laboratory work involving "in-house" minisequencing methods, performed the statistical calculations and drafted the manuscript. MF cultured the cells, performed RNA and DNA extraction, performed the assays with the reference method, and provided input to the manuscript. AD performed the assays using the SNPstream system. A-CS conceived the study, participated in its design, coordination and in preparation of the manuscript. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Additional file1 is a pdf-file with information on the SNPs, including dbSNP ID number, nucleotide variation and the sequences of the primers and probes used in the microarray based minisequencing and TaqMan assays respectively. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529269.xml

539052

Sensitive, Noninvasive Detection of Lymph Node Metastases

Background Many primary malignancies spread via lymphatic dissemination, and accurate staging therefore still relies on surgical exploration. The purpose of this study was to explore the possibility of semiautomated noninvasive nodal cancer staging using a nanoparticle-enhanced lymphotropic magnetic resonance imaging (LMRI) technique. Methods and Findings We measured magnetic tissue parameters of cancer metastases and normal unmatched lymph nodes by noninvasive LMRI using a learning dataset consisting of 97 histologically proven nodes. We then prospectively tested the accuracy of these parameters against 216 histologically validated lymph nodes from 34 patients with primary cancers, in semiautomated fashion. We found unique magnetic tissue parameters that accurately distinguished metastatic from normal nodes with an overall sensitivity of 98% and specificity of 92%. The parameters could be applied to datasets in a semiautomated fashion and be used for three-dimensional reconstruction of complete nodal anatomy for different primary cancers. Conclusion These results suggest for the first time the feasibility of semiautomated nodal cancer staging by noninvasive imaging.

Introduction Most primary malignancies spread systemically via lymphatic dissemination [ 1 ]. For example, the finding of axillary nodal metastases predicts a much shorter disease-free survival in breast cancer [ 2 ]. The total nodal tumor burden (number of affected nodes and metastatic tumor volume) affects prognosis even more severely [ 3 ]. Accurate lymph node staging also remains a cornerstone in choosing the most appropriate therapy for a given stage. Therapeutic intervention of metastatic lymph nodes [ 4 ], prophylactic radiation of frequently affected drainage routes [ 5 ], and systemic therapies [ 6 ] all have been shown to improve survival. Genetic profiles identifying metastatic tumors [ 7 ], serum biomarkers, and proteomic profiles are currently being developed to identify patients at risk [ 8 , 9 ]. No direct genetic profile, however, has been demonstrated to date to accurately predict the presence of human nodal metastases in a given patient. Rather, surgical approaches, such as sentinel lymph node biopsy or lymph node dissection, are still commonly used. Careful histological analysis includes mapping, bisectioning, and rapid staining in the frozen tissue laboratory. Higher diagnostic accuracies can be achieved by serial sectioning (50 μm) and by immunohistochemical staining [ 10 , 11 ]. Noninvasive imaging studies are commonly used during the workup of primary malignancies. Typically, lymph nodes are diagnosed by tomographic techniques (computed tomography [CT], magnetic resonance imaging [MRI]) as malignant when their short axis is >10 mm in size [ 12 ]. Such size criteria, however, have been shown to be unreliable [ 13 ]. Similarly, the detection of cancer in nonenlarged (occult) nodes is often quite low by positron-emission tomography (PET) and single photon emission computed tomography imaging. For example, small nodal metastases (< 5 mm) are often missed by PET imaging in patients with breast cancer [ 14 ]. More recently, it has become possible to image anatomic regions at submillimeter resolutions by MRI, with excellent spatial coverage and reduced motion artifacts. The development [ 15 , 16 ] and clinical introduction of lymphotropic magnetic nanoparticles has been shown to significantly improve diagnostic accuracies of MRI for nodal staging (LMRI) in prostate cancer [ 17 ]. These nanoparticles serve as probes for lymphatic anatomy and function and enhance tumor detection through abnormal distribution patterns in malignant nodes [ 17 , 18 ]. Despite the advances of LMRI for cancer staging, image analysis has been challenging and occasionally controversial. Traditional analysis has been based on a reader's identification of certain structural abnormalities that can be variable, given differences in acquisition parameters and interpretation criteria [ 19 , 20 , 21 ]. Furthermore, it has been challenging to quickly and accurately analyze large datasets generated by LMRI. The goal of the current study was to develop and test technologies that would vastly improve the accuracy of current LMRI nodal staging. Specifically we set out to (a) determine whether unique magnetic parameters existed and could be used for semiautomated image analysis and (b) whether the technique could be applied to different primary cancers. Here we provide the first comprehensive analysis of tissue parameters validated against histopathology as an end point. Methods Study Design The Institutional Review Board approved the current study and all patients signed informed consent. The study was divided into a learning ( n = 97 lymph nodes with known histopathology) and a test dataset ( n = 216 lymph nodes with known histopathology; Table 1 ). Assignment into datasets was done in temporal fashion. The learning dataset represented retrospective cases at outset of the study, and the test dataset represented prospective cases collected during a 1-y interval. In the learning set, 55% of the nodes were benign, and 45% of the nodes were malignant. The learning dataset was obtained from 36 patients (24 male, 12 female, age 28–85 y, mean 59.7 y) with histologically proven primary genitourinary malignancies (prostate, 21; bladder, 9; testes, 5; ureter, 1). All patients completed the MRI study and then underwent surgical resection ( n = 26) and/or nodal biopsy ( n = 10). The investigated nodes had a mean short axis diameter of 10.5 mm (range 3–39 mm). Table 1 Overview of Patient Datasets The test dataset was obtained from 34 patients (25 male, nine female, age 30–82 y, mean 58.9 y) with histologically proven malignancies from different primaries ( Table 1 ), including prostate ( n = 18), breast ( n = 7), penile ( n = 4), bladder ( n = 2), testes ( n = 2), and colon ( n = 1). Seventy-nine percent of the nodes were benign and 21% of the nodes were malignant. The nodes in the test dataset had a mean short axis diameter of 10.0 mm (range 3–39 mm). Both datasets included the full spectrum of normal nodes to completely replaced nodes. MRI MRI was performed at 1.5 T (System 9X, General Electric Medical Systems, Milwaukee, Wisconsin, United States) using phased-array coils. All images were archived on DICOM PACS servers (MIPortal, CMIR and Siemens Medical Systems, Erlangen, Germany; and Impax RS 3000, AGFA Technical Imaging Systems, Richfield Park, New Jersey, United States) for subsequent analysis. Images of the pelvis ( n = 56) extended from the pubic symphysis to just above the level of aortic bifurcation. In patients with primary testicular cancers ( n = 7) imaging was extended superiorly to include the renal hilum and retroperitoneum. In patients with breast cancer ( n = 7) we obtained MR images of the bilateral axillae, including the internal mammary and supraclavicular regions. All patients were imaged with identical pulse sequences and timing parameters. Imaging was performed before and 24 h after intravenous ferumoxtran-10 administration (Combidex, Advanced Magnetics, Cambridge, Massachusetts, United States; 2.6 mg Fe/kg diluted in normal saline and infused over a 20-min period using a 5-μm filter). The acquired pulse sequences included (a) axial T2-weighted fast spin-echo (TR/TE, 4500/80; flip angle, 90°; field of view, 24–28 cm; slice thickness, 3 mm; matrix, 256 × 256; number of excitations, 2; in-plane resolution, 1.2 mm); (b) a T1-weighted two-dimensional gradient-echo sequence obtained in different anatomical planes (TR/TE 175/1.8; flip angle, 80°; field of view, 22–30 cm; slice thickness, 4 mm; matrix, 128 × 256; in-plane resolution, 2.0 mm); (c) an axial T2-weighted dual TE gradient-echo (TR/TE 2100/14–24; flip angle, 70°; field of view, 26–28 cm; slice thickness, 3 mm; matrix, 160 × 256; in-plane resolution, 1.7 mm); and (d) a three-dimensional (3D) T1-weighted gradient echo sequence; TR/TE 4.5–5.5/1.4; flip angle, 15°; field of view, 24–28 cm; slice thickness, 1.4 mm; matrix, 256 × 256; in-plane resolution, 1.0 mm). The above listed imaging sequences and parameters had previously been optimized to reduce motion artifacts, maximize signal-to-noise ratio (SNR), and provide diagnostically useful images of the pelvis, abdomen, and chest within clinically acceptable time limits. The T2-weighted fast spin-echo sequence, in (a) above, was primarily used for qualitative nodal detection, and hence a square pixel with more than one acquisition was obtained. The two-dimensional axial T1-weighted gradient-echo sequence, in (b) above, was chosen to achieve adequate anatomical coverage within a short imaging time. The axial dual-echo gradient-echo sequence, in (c) above, was developed specifically for this project to provide artifact-free datasets for quantitative image analysis. A matrix size of 160 × 256 was chosen for this sequence to achieve a balance between the upper limits for imaging time while reducing image noise. Finally, a 3D T1-weighted sequence was obtained, in (d) above to provide a dataset for vascular maximum intensity projection (MIP) reconstructions. Quantitative Image Analysis All image analysis was performed on archived DICOM images using different software packages (e.g., custom-built software such as CMIR-Image, MGH, Boston, Massachusetts, United States; Syngo, Siemens Medical Systems; Advantage Windows, General Electric Medical Systems). Lymph nodes were identified by readers who manually placed kernels onto each node for automated boundary detection and calculation of nodal dimensions and volumes. The thus identified regions of interest (ROIs) encompassed the entire lymph node (not only portions of it) and were used for quantitative signal-intensity (SI) measurements (see Table 2 ). Serial measurements of nodal dimensions on different pulse sequences or time points varied less than 2%. Table 2 Frequency of Imaging Parameters in Learning Dataset A number of quantitative tissue parameters were calculated either as differences between pre- and postcontrast scans (δ) or as single-value analysis on postcontrast scans (see Table 2 ). The lymph node/muscle (LNM) ratio was calculated by dividing signal intensities of an entire lymph node by that of adjacent muscle using a similar-sized ROI, drawn manually. The nodal SI change was calculated by obtaining SI before and after contrast administration. The nodal SNR was calculated by obtaining SD/SD noise . The T2* was calculated in nodal ROIs on dual TE images using CMIR-Image. T2* maps were constructed by performing fits of a standard exponential relaxation model (S = Ke –TE/T2* ) to the data on a pixel-by-pixel basis. Only pixels with intensity greater than a threshold level (2X of noise) were considered during the fitting process. Pixel variance was obtained from post-MR images. Comparative visual analysis included short axis measurements, and identification of heterogeneity, large focal defects, and central hyperintensity, according to criteria previously established [ 12 , 17 ]. To determine the diagnostic accuracy of the different tissue parameters in the learning dataset, we determined sensitivity, specificity, and predictive values for each parameter alone and in combination ( Table 3 ). The most discriminatory parameters were then applied to the test dataset ( Table 4 ). Table 3 Discriminatory Power of Imaging Parameters in Learning Dataset PPV, positive predictive value; NPV, negative predictive value Table 4 Application of Quantitative Parameters to Test Dataset ( n = 216) a Includes short axis > 10 mm or round > 8 mm PPV, positive predictive value; NPV, negative predictive value In the final set of semiautomated image analysis, 3D reconstructions were obtained for nodal mapping onto vascular anatomy using MIP projections. While the MIP projections do not aid in the differentiation between malignant and benign lymph nodes, they are invaluable in providing anatomic content to the dozens of lymph nodes identified. In particular, MIP images were generated interactively from postcontrast, fat-saturated, volumetric interpolated breath-hold images to outline vascular anatomy. The evaluated lymph nodes characterized as benign or malignant (by T2*/variance analysis) were then superimposed on the volumetric 3D images, using customized software (Advantage Windows, General Electric Medical Systems). Statistical Analysis Data were expressed as mean ± standard deviations (SD) and medians. All statistical testing was performed using GraphPad Prism (GraphPad Software, San Diego, California, United States). The significance between two individual groups was determined using the nonpaired Student's t -test (e.g., benign and malignant datasets in Figure 1 ). For the more discriminatory datasets alternative-free-response receiver operating characteristic curves were plotted. Ratios for cut-off single-value parameters were defined to yield highest sensitivity and specificity. Accuracy for a given parameter was expressed as the area under the curve (A z ), and values are summarized in Table 4 . Figure 1 Tissue Parameters in Learning Dataset Nodal tissue parameters for benign and malignant nodes are shown before (A and B) and after (C–E) intravenous administration of magnetic nanoparticles. Note the insensitivity of conventional MRI (A and B), better separation using single-value analysis (C and D) and excellent separation using two-value analysis (E). Histology All lymph nodes were sampled histologically within 2 wk of the MRI (mean: 6 d; range: 2–14 d). The analysis was done in surgically resected lymph nodes ( n = 55; both benign and malignant nodes) or in fine needle aspirates and core biopsies ( n = 15; malignant nodes only), implementing careful mapping procedures to correlate nodes. Surgically excised nodes were sectioned at 10–20 μm intervals after bihalving and were stained with hematoxylin-eosin. Results Learning Dataset The learning dataset consisted of 97 histologically validated lymph nodes from 36 patients with different primary malignancies (see Table 1 ). The mean short axis diameter was 10.5 mm (range 3–39 mm) with 56 of the 97 nodes (58.3%) measuring less than 10 mm, that is, below the traditional imaging cutoff for malignancy (“occult nodes”). Table 2 summarizes the incidence of different visual, comparative (before and after contrast administration), and semiautomated (postcontrast administration only) parameters in the two different groups. Figure 1 is a graphical representation of overlaps between malignant and benign groups for different parameters listed in Table 2 . Table 3 summarizes sensitivities, specificities, and predictive values for the different quantitative imaging parameters. Sensitivities of metastasis detection by visual image analysis ranged from 50%–94%, however, often with lower specificities. Volumetric measurements, in particular, were insensitive markers of malignancy in nonenlarged nodes (see Table 3 ). In contradistinction, image analysis of pre- and postcontrast image sequences resulted in higher specificities and sensitivities (see Table 3 ). Comparative differences between benign and malignant nodal groups were highest for T2* and pixel variance measurements (see Table 3 ). Of all the semiautomated parameters tested alone, T2* measurements showed the highest sensitivity (93%; 95% confidence interval: 82%–98%) and specificity (94%; 95% confidence interval: 84%–99%) in the learning dataset (see Figure 1 and Table 3 ). Of all the semiautomated parameters tested in combination, T2* measurements combined with pixel variance analyses postcontrast showed the highest sensitivity (98%; 95% confidence interval: 88%–99%) and specificity (94%; 95% confidence interval: 82%–98%) in the learning dataset ( Figure 1 E). Using the dual-value analysis, there was one malignant outlier in the benign dataset (the lymph node was 3 mm in overall size, with few malignant cells seen on histology, and probably too small for analysis) and two benign outliers in the malignant dataset (both these nodes showed hyalinosis replacing more than 50% of the nodal architecture). Test Dataset To determine whether feature extraction would be accurate for prospective nodal staging, we utilized the above criteria against a larger test dataset encompassing 216 validated lymph nodes from 34 patients, including different primaries (see Table 1 ). The sensitivity, specificity, and predictive values of the most discriminatory parameters of this prospective analysis are summarized in Table 4 . We primarily focused on semiautomated image analysis of postcontrast scans because of the high sensitivity and specificity determined in the learning dataset. T2* measurements showed a sensitivity of (93%; 95% confidence interval: 82%–99%) and a specificity of (91%; 95% confidence interval: 85%–96%). Combined T2* and pixel variance analysis achieved a sensitivity of 98% (95% confidence interval: 88%–99%) and a specificity of 92% (95% confidence interval: 87%–96%) comparable to that of the learning set and much superior to currently used size criteria. Using the dual-value analysis, there were two malignant outliers in the benign dataset (both of these nodes were less than 3 mm in overall size and probably too small for analysis—similar to the learning dataset) and three benign outliers in the malignant dataset (two of these nodes had hyalinosis replacing more than 50% of the nodal architecture and one had macrocalcifications). More important, all the misclassified nodes occurred in individual patients rather than in the same patient and, hence, did not affect the overall nodal staging on a patient-by-patient basis in this dataset. Image Reconstruction Video 1 Automated 3D Reconstruction of Pelvic Nodal Anatomy Utilizing semiautomated feature extraction to identify lymph nodes and image analysis (based on T2* and pixel variance), we subsequently proceeded to map individual lymph nodes onto vascular anatomy in the different anatomic drainage patterns. Figure 2 summarizes the different steps in image analysis. Figure 3 and Video 1 shows an example of a 45-y-old patient with colorectal cancer undergoing semiautomated nodal staging. In this particular patient, MRI identified six positive lymph nodes (< 10 mm each), reconstructed as a 3D dataset, whereas all positive lymph nodes were missed by PET scans. Figure 4 and Video 2 show reconstructions and analyses from a patient with a breast cancer primary with bilateral nodal metastases. Note the high spatial resolution allowing the detection of a 3-mm nodal metastasis. Figure 2 Steps in Semiautomated Image Analysis Semiautomated image analysis involves recognition and automated segmentation of each lymph node (A), quantitation of magnetic tissue parameters (T2*, variance of pixel values; [B]), comparison of extracted tissue parameter to a database (C), and 3D reconstruction of nodal anatomy onto vascular anatomy (D). Figure 3 Pelvic Nodal Staging Nodal staging in patient with colorectal cancer. A PET scan using 18 FDG as a tracer (A) and a CT scan (B) were interpreted as negative for nodal metastases. LMRI identified six small pelvic lymph nodes ([C] and [D]; red arrowheads), which had magnetic parameters of malignancy. Semiautomated reconstruction (E) identifies multisegmental metastases, subsequently proven histologically (F). For 3D reconstruction of pelvic nodal anatomy see Video 1 . Figure 4 Breast Cancer Mapping Patient with breast cancer prior to sentinel lymph node biopsy. (A) Conventional axillary MRI shows nonenlarged lymph nodes that do not meet the size criteria of malignancy (white bar = 5 mm). (B) Following intravenous administration of nanoparticles, a single 3-mm intranodal metastasis was correctly identified. (C) Ex vivo MRI of sentinel node specimen confirms metastasis. (D) Semiautomated nodal analysis and reconstruction correctly juxtaposed solitary lymph node metastases adjacent to two normal lymph nodes. (E) Correlative histopathology confirms the diagnosis. For 3D reconstruction of axillary nodal anatomy see Video 2 . Video 2 Automated 3D Reconstruction of Axillary Nodal Anatomy Discussion We show that it is feasible to extract various quantitative tissue parameters to predict the likelihood of nodal metastases in vivo. These results are highly relevant in cancer staging because they provide evidence that (a) quantitative tissue parameters enable diagnosis of lymph node metastases while reducing interobserver variability and (b) that semiautomated reconstructions allow spatially more extensive mapping than is currently possible. Metastases to lymph nodes occur during growth of most primary malignancies, and their presence mandates the need for more extensive and systemic therapy. Nodal cancer staging currently relies on invasive procedures (surgical lymph node dissection, sentinel lymph node resection, biopsy) with significant morbidity and cost [ 22 , 23 ], or insensitive tomographic imaging methods [ 24 ]. For example, detection sensitivities using size criteria with state-of-the-art multislice CT are as low as 50%, whereas PET imaging of nonenlarged nodes has equally low sensitivities [ 14 ]. Based on the observation that nanoparticulate solutions accumulate in nodal macrophages upon systemic injections [ 25 , 26 ], lymphotropic superparamagnetic preparations have been developed [ 16 ]. In earlier clinical trials (using lower spatial resolution sequences), metastases of 1–2 mm have been detected [ 17 ], whereas as few as 1,000 tumor cells have been detected in nodes in experimental mouse models [ 18 ]. Despite these advances, it has been difficult to acquire images of sufficiently high resolution and to derive parameters to automate diagnosis. The data presented here indicate that unique magnetic parameters allow identification of nodal metastases and accurate 3D reconstructions, including surgically inaccessible lymph nodes. The significance of the above findings is 3-fold. First, the ability to directly and noninvasively monitor nodal tumor involvement represents a powerful diagnostic tool for cancer. Accurate staging represents the cornerstone for triaging patients to either localized or to more aggressive and systemic therapies. Second, the method described here was sensitive for the limited subsets of primary cancers tested. It is reasonable to hypothesize that such analysis could be applied to staging of other common primaries. In particular, lung, colorectal, genitourinary, and head and neck cancers could benefit from this staging procedure. In addition to nodal staging, the nanoparticle-enhanced MRI can also be used to measure microvascularity in primary tumors [ 27 ] and to improve the detection of liver metastases [ 28 ]. Third, our results are significant because the semiautomated staging method is highly accurate and reduces variability in visual image analyses between different observers. The LMRI staging technique is believed to be clinically relevant in several key areas. First, LMRI may play a significant role in avoiding unnecessary surgeries, that is, those in node-positive patients. Second, since LMRI can detect lymph nodes outside traditional surgical fields, this information may influence surgical approaches. In colorectal cancer, LMRI may provide a “sentinel-node-like” guide to staging. Third, it is likely that LMRI would be useful to identify appropriate patients to receive neoadjuvant chemotherapy prior to surgery. Currently, neoadjuvant therapy is often reserved for postoperative patients, once the nodal status has been determined. Fourth, LMRI may be particularly useful to guide radiation therapy by mapping the complete nodal status onto bony and vascular landmarks. Finally, LMRI could be used to avoid invasive diagnostic procedures, which are not part of therapy. For example, LMRI could replace lymphangiography, mediastinoscopy, or endoscopic ultrasound for nodal staging. Our findings have a number of direct implications for technology development and in clinical care. Accurate measurements of T2* relies on motion artifact-free multiecho pulse sequences that are not routinely available on clinical scanners at spatial resolutions required for nodal staging. Such sequences will have to be implemented and combined with postprocessing tools to simplify and semiautomate analysis. Similar software approaches are already used routinely in lung nodule characterization [ 29 ] or screening for breast cancers [ 30 ]. We predict that in the case of LMRI, such automation routines will be highly specific, given the unique mechanism of image contrast. As a proof-of-principle, we implemented approaches to identify, segment, analyze, and display nodal information. While the current technology is already highly accurate, we anticipate further improvements with hardware and software advances. We hope that this will ultimately translate into clinical practice and replace unnecessary intervention. Patient Summary Background When deciding on treatment for patients with cancer, it is very important to assess whether the cancer has spread to lymph nodes—both to help decide what treatment a patient should have and what the eventual outcome might be. Previous ways of finding involved lymph nodes included taking out the nodes by surgery, ultrasound, and CT and MRI scans. What Does This Study Show? A solution of magnetic nanoparticles that tend to go to lymphoid organs was injected and then tracked by MRI. The pattern of the particles was abnormal when there was metastasis in the nodes, and it was possible to train a computer to recognize this abnormality. The authors developed the program in one group of patients and then tested it in another group, in which they were able to correctly predict whether the nodes were involved in about nine of ten nodes. In addition, they could use the information to display a virtual picture of the involved nodes. What Does This Study Mean for Patients? The technique will need to be validated in a larger group of patients, and by other investigators. However, it means that it is potentially possible to work out much more precisely, and with less chance of error, whether lymph nodes are involved in cancer. Hence, treatment can be better planned, and if surgery is needed to remove nodes for analysis, then this technique could ensure that the surgery is as minimal as possible. Where Can I Get More Information? RadiologyInfo, a public information site developed by the American College of Radiology and the Radiological Society of North America: http://www.radiologyinfo.org/ Medline Plus, which has health information from the National Library of Medicine: http://www.nlm.nih.gov/medlineplus/cancer.html

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539052.xml

523234

Controlling the Timing of Gene Expression during Organ Development

For more than 2,000 years, from the time of Aristotle onwards, it was thought that the complete body plan of human beings (and that of other animals) was present in the fertilized egg. During pregnancy, a preformed miniature human being, or homunculus, grew bigger and bigger; development was simply a process of growth. Then, in the mid-18th century, Carl Friedrich Wolff described how the chick gut, basically a tube, forms from an initially flat sheet of cells, overthrowing at a stroke the preformation theory of embryology. We now know that development is a complex series of coordinated processes that transforms the amorphous ball of cells produced from the fertilized egg by cell division into an intricate body containing numerous specialized tissues and organs. And we are beginning to understand how a wide array of transcription factors—proteins that bind to regulatory sequences within genes to control their expression—guide the sequential stages involved in development. It seems that these factors form regulatory networks that control the temporal and spatial waves of gene expression that underlie and are required for organized body building. Susan Mango and her colleagues are studying the role of transcription factors in controlling organ development. The organ they are studying—the pharynx of the nematode worm—is relatively simple. This muscular tube, which passes bacteria (the food of this small soil-dwelling organism) from the mouth to the midgut, contains fewer than 100 cells of only seven different types. To get an overall picture of the regulatory sequences within genes that are involved in the temporal control of pharyngeal development, the researchers identified 339 candidate pharyngeal genes by comparing gene expression profiles in mutant worm embryos that had excess pharyngeal cells with those in mutant embryos lacking pharyngeal cells. Then, by referring to a database that details gene expression patterns in nematode worms and embryos, the researchers classified 37 of their candidate genes as having early-onset expression and 34 as having late-onset expression. Next, the scientists carefully examined the DNA sequence of each gene for candidate regulatory regions that might contribute to its temporal regulation. Of nine candidate motifs revealed by this search, six functioned as regulatory sites in in vivo assays. The researchers estimated that these six elements, together with sites that bind PHA-4—a member of a family of transcription factors that are important in digestive tract development in many animals—account for the timing of onset of expression of about half of the nematode's pharyngeal genes. Finally, the researchers used combinations of the newly discovered temporal regulation sites and PHA-4 sites in a genome-wide search that predicted pharyngeal genes and their time of onset of expression with greater than 85% accuracy. Fluorescent reporter genes expressed in the developing C. elegans foregut From these results and those of previous studies, Mango and her colleagues propose a model to explain how the temporal control of pharyngeal gene expression needed for pharynx development is achieved. The earliest time for pharyngeal gene expression, they suggest, is determined by how well PHA-4 sticks to a particular gene's binding site. However, gene expression only occurs if other factors that bind to the regulatory sites are also present, and the exact combination of these factors determines which gene is active at any given time. The identity of these factors remains to be discovered. Nevertheless, at least for this simple organ, we now have a much better idea of how the complex process of organ formation is controlled at a molecular level, and it is likely that similar regulatory networks will underlie the formation of other organs as well.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC523234.xml

493285

Transient spleen enlargement in peripheral blood progenitor cell donors given G-CSF

The administration of granulocyte colony-stimulating factor (G-CSF) to peripheral blood progenitor cell (PBPC) donors causes spleen length to increase, but the duration of enlargement is not known. Eighteen healthy subjects were given 10 μg/kg of G-CSF for 5 days and a PBSC concentrate was collected by apheresis. Ultrasound scans were used to assess craniocaudal spleen length before and after G-CSF administration. Mean spleen length increased from a baseline length of 10.7 ± 1.3 cm to 12.1 ± 1.2 cm on the apheresis day (p < 0.001). Ten days after apheresis, spleen length fell to 10.5 ± 1.2 cm and did not differ from baseline levels (p = 0.57), but in 3 subjects remained 0.5 cm greater than baseline length. Increases in spleen length in PBPC donors are transient and reversible.

Background Peripheral blood progenitor cell (PBPC) concentrates donors are routinely given granulocyte colony-stimulating factor (G-CSF) to increase the concentration of circulating PBPCs and hence the number of progenitors that can be collected by apheresis. Typically 10 to 16 μg/kg of G-CSF are given subcutaneously daily for 4 to 6 days prior to the collection [ 1 - 3 ]. The administration of G-CSF to healthy PBPC concentrates donors is very safe, but there have been four reports of spontaneous rupture of the spleen and splenectomy in healthy allogeneic PBPC donors given G-CSF [ 4 - 7 ]. While spontaneous rupture of the spleen in PBSC donors given G-CSF is rare, the administration of G-CSF for five days causes spleen length to increase in almost all healthy donors [ 8 , 9 ]. The increase in length is highly variable, but the mean increase is approximately 13%. Spleen length begins to return to baseline levels quickly, but it is not known how long it takes to return to baseline. In a previous study of 20 PBPC donors given 10 μg/kg of G-CSF for 5 days, we found that spleen length measured four days after the last dose of G-CSF was less than the length on the day of apheresis but greater than baseline values [ 8 ]. Since allogeneic PBPC donors may be at risk for splenic rupture while the spleen is enlarged, it is important to determine when spleen size returns to baseline levels. The purpose of this study was to determine if spleen length returns to baseline 10 days after G-CSF-mobilized PBPC concentrates are collected by apheresis from healthy subjects. Methods Study design All of the subjects were in good health and were donating G-CSF-mobilized PBPC concentrates for laboratory investigations. The donors were given 10 μg/kg of G-CSF (Filgrastim, Amgen, Thousand Oaks, CA) daily for 5 days, and a PBSC concentrate was collected approximately 2 hours after the last G-CSF dose was given. PBPC concentrates were collected with a CS3000 blood cell separator (Baxter Health Care Corporation, Round Lake, IL). Spleen length was evaluated by ultrasound examination three times: prior to the administration of the first dose of G-CSF, on the day of apheresis, and 10 or 11 days after apheresis. This study was approved by the Institutional Review Board of the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland. Spleen length assessment Craniocaudal spleen length was assessed using ultrasound (Acuson Aspen Advanced, Siemens Medical Solutions Ultrasound Division, Mountain View, CA) with a sector transducer (Acuson, 4V1, 4.0 mHz frequency, Siemens Medical Solutions). The intra-observer error for measuring spleen length using ultrasound is 4.9 mm when healthy subjects are evaluated at separate settings [ 10 ]. Blood counts and chemistries Complete blood counts were performed with an automated cell counter (Cell Dyne 4000, Abbott Diagnostics, Santa Clara, CA). CD34+ cell counts were performed using a flow cytometer (Beckman Coulter, Miami, FL). Statistical analysis Spleen lengths measured before and after the G-CSF course were compared using 2-tailed paired t-tests. Spleen length changes were also compared among males and females and Caucasians and non-Caucasians using 2-tailed t-tests. The percent change in spleen length was compared with blood counts, CD34+ cell counts, and donor age using linear regression. Results and Discussion The median age of the 18 healthy subjects was 34 years old and ranged from 22 to 55 years of age. Eight of the subjects were male, 13 of the donors were Caucasian, 3 were African American, and 2 Asian. Apheresis day spleen length increased above baseline length in 17 of 18 donors (Figure 1 ). The mean spleen length increased from a baseline level of 10.7 ± 1.3 cm to 12.1 ± 1.2 cm on the day of apheresis (p < 0.001). The mean increase in length was 1.4 ± 0.8 cm or 13.1 ± 8.9%. Figure 1 Spleen length changes in healthy subjects donating G-CSF-mobilized PBPC concentrates. Eighteen subjects were given 10 μg/kg of G-CSF for 5 days and a PBPC concentrate was collected approximately 2 hours after the last dose of G-CSF. Spleen length was measured by ultrasound before G-CSF was given (day 1), immediately after the PBPC concentrate collection (day 5), and approximately 10 or 11 days after the collection (day 15 or 16). Spleen length was measured 10 or 11 days after apheresis in all 18 subjects. The mean spleen length 10 days after apheresis fell to 10.5 ± 1.2 cm and was less than the apheresis day spleen length (p < 0.001). There was no difference between the 10-day post-apheresis and pre-G-CSF spleen length (p = 0.57). The spleen length 10 days after apheresis was less than the apheresis day length in all 17 donors whose spleen length increased. However, the spleen length 10 days after apheresis remained more than 0.5 cm greater than baseline spleen length in 3 subjects (Table 1 ). All 3 were Caucasian and 2 were female. Their spleen length remained 6.7%, 11.0%, and 10.5% greater than baseline levels. Two of these subjects had relatively large increases in spleen length of 18.3% and 26.3%, but their spleen length fell considerably 10 days after apheresis. It is likely that the spleen returned to baseline length in these 2 subjects shortly after the third ultrasound was preformed. The other subject's spleen increased only 12.0% in length and 10 days after apheresis had changed little. It is not certain when this subject's spleen returned to baseline length. Table 1 Peripheral Blood Stem Cell Donors Whose Spleen Length 10 Days After Apheresis Remained More than 0.5 cm Greater than Baseline Length Spleen Length (cm) Donor Age (Yrs) Gender Race Baseline Apheresis Day 10 Days Post-Apheresis Enlargement 10 Days Post-Apheresis 7 23 Female Cauc 10.4 12.3 11.1 0.7 13 22 Female Cauc 10.0 11.2 11.1 1.1 14 54 Male Cauc 9.5 12.0 10.5 1.0 Cauc = Caucasian PBPC donors with the largest increase in spleen size may be at the greatest risk for spontaneous splenic rupture. In order to determine if subject age, gender, race, or post-G-CSF blood counts affected the magnitude of spleen enlargement, we assessed the relationship between these factors and percent change in spleen length in the 18 subjects in this study and 20 subjects studied previously [14]. There was no difference in the spleen length increase between males and females (12.3 ± 9.7% versus 14.8 ± 8.0%, p = 0.40) or between Caucasians and non-Caucasians (13.5 ± 8.4% versus 13.6 ± 10.4%; p = 0.98). Spleen length increase was not related to donor age (r = 0.13). In addition, spleen length increase was not related to preapheresis CD34+ (r = 0.04), WBC (r = 0.05), neutrophil (r = 0.07), lymphocyte (r = -0.14), monocyte (r = -0.04), and platelet counts (r = 0.19) or hemoglobin level (r = -0.04). Conclusions Healthy PBPC concentrate donors given G-CSF should be warned that their spleens will be enlarged for a brief time and that they may be at risk of splenic rupture. Most donors are likely at risk for splenic rupture only during the time of G-CSF administration and for about 10 days after the completion of the G-CSF course. Since splenic enlargement may persist for longer periods in some donor, until more data are available it may be worthwhile to counsel PBPC donors to avoid activities that could lead to abdominal and splenic trauma for 2 to 3 weeks after the last dose of G-CSF.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC493285.xml

548668

Maps of the Sri Lanka malaria situation preceding the tsunami and key aspects to be considered in the emergency phase and beyond

Background Following the tsunami, a detailed overview of the area specific transmission levels is essential in assessing the risk of malaria in Sri Lanka. Recent information on vector insecticide resistance, parasite drug resistance, and insights into the national policy for malaria diagnosis and treatment are important in assisting national and international agencies in their control efforts. Methods Monthly records over the period January 1995 – October 2004 of confirmed malaria cases were used to perform an analysis of malaria distribution at district spatial resolution. Also, a focused review of published reports and routinely collected information was performed. Results The incidence of malaria was only 1 case per thousand population in the 10 months leading up to the disaster, in the districts with the highest transmission. Conclusion Although relocated people may be more exposed to mosquito bites, and their capacity to handle diseases affected, the environmental changes caused by the tsunami are unlikely to enhance breeding of the principal vector, and, given the present low parasite reservoir, the likelihood of a malaria outbreak is low. However, close monitoring of the situation is necessary, especially as December – February is normally the peak transmission season. Despite some losses, the Sri Lanka public health system is capable of dealing with the possible threat of a malaria outbreak after the tsunami. The influx of foreign medical assistance, drugs, and insecticides may interfere with malaria surveillance, and the long term malaria control strategy of Sri Lanka, if not in accordance with government policy.

Background After the tsunami hit Sri Lanka on 26 December 2004, news reports and public health agencies warned against the possibilities of an increase of vector borne diseases, in particular malaria and dengue. Immediately after the disaster, an estimated 860,000 people were displaced and more than 820 emergency camps established throughout the affected areas [ 1 ]. By 14 January, approximately 440,000 people were still sheltered in approximately 460 emergency camps [ 2 ]. Maps of the tsunami affected area, are presented elsewhere [ 3 ]. Malaria in Sri Lanka is of a highly unstable nature and has historically fluctuated greatly over the years and with significant seasonal differences. Sixty-five to eighty percent of the malaria cases are caused by Plasmodium vivax and the remainder by Plasmodium falciparum [ 4 ]. Recently, an overview of the spatial and temporal distribution of malaria in Sri Lanka over the period 1995 – 2002 was published in this journal [ 5 ]. The present publication aims at providing an update on the recent malaria situation, to October 2004 inclusive, and to discuss factors of relevance which may help in assessing the potential of the tsunami and ensuing events for exacerbating the malaria situation. Methods Malaria maps were based on monthly records over the period January 2004 – October 2004 (the most recent month for which data recording was complete at the time of writing) of microscopically confirmed malaria parasite positive blood smear readings, at district spatial resolution. These were collected by the Anti Malaria Campaign (AMC) Directorate of the Ministry of Health from aggregated disease records reported by governmental hospitals and mobile clinics. Additionally, in the temporal analysis, monthly data by district for the period 2001 – 2002, and data by sub district for 1995 – 2000 as described by Briët et al . [ 5 ] were used. The quality of routinely collected information on malaria is described elsewhere [ 5 ]. As denominator for the incidence calculations, population estimates for 2001 and beyond were made by exponential interpolation (and extrapolation to December 2004) (Figure 1 ) as follows. For the districts Mannar, Vavuniya, Trincomalee and Batticaloa, that were not or incompletely enumerated in the 2001 census because of limited access of the government to these conflict affected areas, the 2001 mid-year population was taken from data posted by the North East Provincial Council [ 6 ]. For all other districts, the 2001 mid-year population was taken from data posted by the Department of Census and Statistics [ 7 ]. The natural annual (mid-2001 to mid-2002 and mid-2002 to mid-2003) population growth rates for Jaffna, Kilinochchi, Mullaitivu, Mannar, Vavuniya, Trincomalee and Batticaloa were taken as the average annual growth rates of all the other districts, calculated from mid year population statistics estimated by the Department of Census and Statistics. For all other districts, these growth rates were calculated for each district separately. For mid 2003 to mid 2004 and beyond, the growth rates for mid-2002 to mid-2003 were used. Further, the number of internally displaced persons (IDPs) was taken into account [ 8 ]. For each month and for each district, the net number of immigrants was calculated as the total number of IDPs moved to or within a district since 2001, minus the number of IDPs moved from or within that district. This net number of immigrants was then distributed over the months proportionately to the monthly statistics of IDPs moved to or within a district. Additionally, the number of monthly immigrants from India was taken into account. Figure 1 Population. Map of population by divisional secretariat division in Sri Lanka estimated for mid December 2004. One dot represents 1,000 people. Sources: Department of Census and Statistics , North East Provincial Council and UNCHR . A focused review of literature has been performed, identifying crucial information for the outbreak preparedness and control during the emergency phase. The intent was not to present a complete review of malaria in Sri Lanka but to provide information useful for an assessment of the current situation. A general review of malaria in Sri Lanka can be found in Konradsen, Amerasinghe et al . [ 4 ]. Results and discussion Present malaria situation and parasite reservoir The country-wide malaria incidence increased from January 1996 to January 2000, with the typical seasonality of high peaks around January and lower peaks around June – July, but it has decreased dramatically since January 2000 (Figure 2 ). Figure 3 shows that the recent decrease in the overall malaria incidence in the country is predominantly due to a decrease in incidence in the districts of Vavuniya and Kilinochchi in the north. The decrease was least in the district of Ampara, making it the most malarious district during January to October 2004 (Figures 4 and 5 ). Although districts on the east coast which were badly affected by the tsunami had been relatively malarious in 2004 as compared to the rest of the country, the maximum of around 1 case per 1000 people over a 10 month period in these districts is remarkably low. The total number of malaria cases in 2003 was 10,510, the lowest since the resurgence of malaria in 1968 when the eradication campaign failed [ 9 ]. The year 2004 promises to be three times lower with only 3,037 cases recorded up to October, as opposed to 9,682 cases recorded during January – October 2003. The low incidence is not related to a decline in collection effort, which has decreased only marginally (Figure 2 ). At the time of writing, malaria incidence information for the months of November and December was still incomplete. In November 2004, without the figures for the non endemic districts Gampaha and Kalutara, and data from a few medical institutions in Mannar and Mullaitivu missing, thus far only 230 cases were recorded. In the malaria endemic districts, December, January and February are normally the months with the highest malaria incidence [ 5 ], so a rise in case numbers may normally be expected. However, neither the district authorities nor the Epidemiology Unit of the Ministry of Health have reported any malaria cases from the affected areas for 30 December 2004 – 13 January 2005, based on the spot checks performed and the review of available health information [ 10 ]. Asymptomatic infections of P. falciparum and P. vivax and dormant stages of P. vivax normally provide the parasite reservoir for bridging periods of low seasonal transmission (with unsuitable conditions for mosquito vectors). Under the present policy of administering primaquine in addition to chloroquine (see section on diagnosis and treatment), the reservoir of dormant stages of P. vivax will be low and this will delay a possible outbreak. It must be emphasized that the low level of malaria transmission in the recent past does not guarantee that localized or even island wide epidemics will not occur. In the past, even after periods of very low levels of malaria transmission, outbreaks have occurred, often due to constraints placed on the public health system, by unusual rainfall patterns or by yet unexplained factors. Figure 2 Monthly parasite and blood smear examination incidence patterns. Monthly parasite incidence patterns of P. falciparum and P. vivax malaria combined per 1000 population (red line on logarithmic scale), blood smears examined per 1000 population (black line on logarithmic scale), and percentage of blood smears positive for malaria (blue line) from January 1995 to October 2004 in Sri Lanka. Figure 3 Trends of parasite incidence. Trends of parasite incidence of P. falciparum (red bars) and P. vivax (blue bars) malaria over the years November 1995 – October 1996 (bar on far left) to November 2003 – October 2004 (bar on far right), at district resolution. The height of the bars in the legend represents an annual parasite incidence of 10 cases per 1000 persons. Figure 4 Parasite incidence of Plasmodium vivax. Map of the districts of Sri Lanka with P. vivax malaria cases per 1000 population over the period January – October 2004. Figure 5 Parasite incidence of Plasmodium falciparum. Map of the districts of Sri Lanka with P. falciparum malaria cases and mixed infections of both P. vivax and P. falciparum per 1000 population over the period January – October 2004. Capacity of health care services and disease surveillance An important factor to consider in the current situation is the capacity of the existing health care service. Following the tsunami the Sri Lanka Ministry of Health reported 22 hospitals and nine administrative buildings damaged or completely destroyed, mostly in Ampara and Trincomalee districts [ 11 ]. It has been reported that at least 40 doctors and hundreds of other medical staff have died as a consequence of the tsunami and a much higher number injured or in other ways affected by the disaster [ 12 ]. However, both the central government departments and organizations in the field report sufficient medical staff. Even in the conflict affected areas in the north and east, the AMC has been able to monitor malaria and react timely with control measures to outbreaks since the peace process started in 2002. Also, the AMC has long standing experience with mobile clinics for malaria detection and treatment in remote areas. Lack of co-ordination among the many government departments, international aid agencies, non-governmental organizations and private individuals involved in the first phase of the emergency continues to be an important issue weeks into the disaster. According to the Ministry of Health media reports, more than 600 foreign doctors are now working in the affected areas, but few, if any, are registered with the Sri Lanka Medical Council or other relevant authorities [ 13 ]. With doctors from many countries, language barriers are also a perceived problem. In some places, central stocks of medical supplies were destroyed, including the Regional Medical Supply Division in the Ampara District. However, sufficient drugs have been imported during the days and weeks following the disaster. The World Health Organization has drawn up plans for antimalarials, insecticides and spray equipment to be made available on request. Although the increased capacity at the district and provincial levels has improved the co-ordination, a risk remains that local needs for health care are not adequately covered in spite of the availability of significant resources. In some parts of the island, especially areas in the east, affected both by the destruction caused by the tsunami and by exceptionally heavy rainfall in the weeks following, distribution of drugs has been problematic and this has left certain communities vulnerable. Whereas the overall capacity to provide treatment and routine malaria control activities, in general, has not been severely hampered, the routine health information system will have been constrained by the large number of autonomous health camps set up, and their lack of integration with the established surveillance system. It is essential to establish a system for monitoring malaria in the affected areas. Many people are moving back to their old place of residence trying to rebuild livelihoods and it will be essential for the public health authorities to keep contact with these communities to prevent an increase in malaria going unnoticed. Diagnosis, treatment and drug resistance In Sri Lanka, microscopy on blood smears or use of rapid diagnostic test kits have been the standard diagnostic procedure, and precedes the prescription of drugs to the patient. In the current situation, with the many small health clinics established within emergency camps, it is likely that the use of rapid diagnostic kits would be the more feasible means of confirmation. The first line drugs recommended for malaria treatment in Sri Lanka is still a chloroquine and primaquine (PQ) combination for cases of P. vivax as well as P. falciparum infection. Primaquine is not administered to children below one year, and those with known G-6PD enzyme deficiency, and for pregnant mothers. So far, there have been no reports of chloroquine-resistant P. vivax infections in Sri Lanka. The first chloroquine-resistant P. falciparum case was reported in 1984 [ 14 ]. Up to 62% in vivo chloroquine resistance has been recorded in malarious areas [ 5 , 15 - 17 ]. For chloroquine resistant cases of P. falciparum the government recommended drug is sulphadoxine-pyrimethamine (SP). However, SP is not recommended for the last trimester of pregnancy, first six weeks of lactation and for children below two months of age. The first SP-resistant case of P. falciparum was reported in 1992 in Polonnaruwa district. Up to 1999, five to six cases have been reported by the AMC. More recently (January – June 2002), SP resistant P. falciparum has been documented in the Northern Province [ 17 ]. For SP resistant cases quinine is recommended, but only as an in-patient treatment. In the current emergency situation, with many (foreign) doctors working autonomously, the diagnosis and treatment practices may depart from the established government guidelines and new antimalarials are also likely to be brought in. Moreover, the current practice of restricting SP to government hospitals will be difficult to enforce. Similarly, introduction of low quality and obsolete drugs will be difficult to counter at community level at the current stage of supervisory capacity and co-ordination level. Drugs have been reported stolen from warehouses, allegedly finding their way to private trade establishments [ 18 ]. Overall, it is crucial that the development of drug resistance is monitored closely and inappropriate drugs are actively phased out of the market to avoid later complications in case management. Environmental changes and vector breeding The seawater brought inland by the tsunami has mixed with monsoon rainwater to form puddles of varying salinity. Also, thousands of muddy surface water puddles have been created as a result of destruction and rehabilitation activities that are already underway. The brackish puddles are expected to favour the breeding of Anopheles subpictus sibling species B, which is a well-known coastal breeding species in Sri Lanka. However, it has not been directly incriminated as a field vector in Sri Lanka, despite its susceptibility to P. falciparum [ 19 ]. Nevertheless, Abhayawardana et al. [ 20 ] found peak malaria transmission in coastal areas of Puttalam in the presence of An. subpictus sibling species B and the complete absence of Anopheles culicifacies (the main malaria vector in Sri Lanka), and suggested that this An. subpictus sibling may have a role in transmission. It is noteworthy, that freshwater An. subpictus (which is now known to consist of a mixture of species A, C and D), which breeds in muddy rain fed puddles, has been consistently incriminated in malaria transmission in many inland areas of Sri Lanka [ 4 ]. Another species that is likely to breed prolifically in muddy rain-fed pools is Anopheles vagus . This species has been linked as a vector responsible for a malaria outbreak in southern Sri Lanka [ 4 , 21 ]. On present evidence, neither An. subpictus nor An. vagus , are likely to cause major malaria epidemics but could, at high density, be responsible for focal outbreaks that need quick action. Thus, it is important that an entomological monitoring programme be set up in the period leading up to and during the south west monsoon that is expected during May to June 2005 in the tsunami affected western and southern Sri Lanka. It should be noted that the infamous Asian brackish water breeding malaria vector Anopheles sundaicus , which is a threat in the tsunami-affected areas in Indonesia, Myanmar, and the Andaman and Nicobar islands [ 22 ], does not occur in Sri Lanka. The main vector in Sri Lanka is An. culicifacies type E [ 23 , 24 ], which breeds mainly in pools formed in river and stream beds, and therefore, its density is mostly dependent on temporal and spatial variations in rainfall and river flow. Anopheles culicifacies also breeds in abandoned gem mining pits, agricultural wells and to a lesser extent in pools in agricultural water reservoirs [ 4 ]. It is unlikely that the rubble constituting a major part of the landscape in the affected areas creates breeding opportunities for An. culicifacies , unless it blocks waterways and creates pooling. Post-tsunami development activities may revive banned sand mining practices in rivers. If this happens, clear water pools created by these sand mining activities may serve as breeding sites for An. culicifacies [ 4 ]. Overall, it is very unlikely that the principal vector of malaria in Sri Lanka will breed prolifically in brackish water habitats or other habitats that may be created during the post tsunami reconstruction phase. Similarly, the principal dengue vector in Sri Lanka, Aedes aegypti , does not breed in saline water [ 25 ]. However, it may find plenty of rainwater-filled containers amidst the rubble created by the disaster for it to breed. Vector control strategies and insecticide resistance The Colombo based Head Office of the AMC gives the overall guidelines for island wide vector control, while each province works out a plan for control activities based on the distribution and level of malaria transmission. Several malaria vector control interventions are currently employed within the country. In all districts, residual insecticide spray activities are focused on areas where malaria transmission has been established by confirmed malaria cases. The control of anopheline larvae using mostly chemicals focuses on sites close to human habitation. Small-scale application of larvivorous fish and environmental modifications are also carried out. Since 1997, mosquito nets, which are biannually treated with insecticide, are distributed free of charge in malarious areas. During the last two years, the main control effort has been through these nets. Since January 2004, 80,000 nets with long lasting insecticide have been distributed. Also, nets are available for purchase from outlets in most parts of the country. Studies in Sri Lanka over the 1990s on An. culicifacies and a range of potential secondary vectors such as An. subpictus and An. vagus have shown high level of resistance to either organochlorines, organophosphates or to both groups of insecticides [ 4 , 26 - 28 ]. DDT and Malathion are no longer recommended since An. culicifacies and An. subpictus has been found resistant. Currently, synthetic pyrethroids such as Cyfluthrin, Deltamethrin, Etofenprox, and Lambda-cyhalothrin are being used in the country. At present, Fenitrothion is the only organophosphate used for vector control. A study conducted by Abhayawardana from 1990 to 1992 on An. subpictus found 68% and 54% susceptibility to Malathion and Fenitrothion, respectively, for inland species (sibling species A), whereas for coastal species (primarily sibling species B) it was 100%. However, the latter was found resistant to permethrin [ 20 ]. From several districts it was reported that, as a result of the tsunami, organisations have brought in insecticides not normally used or no longer recommended for vector control in Sri Lanka (P. Amerasinghe, personal communication). Vector resistance, in the light of the introduction of new insecticides, needs to be monitored and if necessary action should be taken. Exposure of the affected community The majority of the people initially affected by the disaster are still living in emergency camps or in other places close to the coast. At the time of writing, to the best of our knowledge, relatively few people have moved from areas of low or no malaria transmission to areas of high transmission. However, during the next phases, when people may be resettled in semi-permanent and later in permanent housing, communities may be relocated from areas where they have had no malaria experience to malarious areas. In these situations, the communities' capacity to cope with malaria infection will be low. Despite distribution of nets to many camps, and intensified vector control in some areas, people in the emergency camps (schools, temples, mosques, etc.) and those returning to damaged houses are more exposed to mosquito bites than in pre-disaster housing, due to the open nature of the shelter. Additionally, most families will have lost mosquito nets or other means to protect against mosquito bites. It is more difficult to assess the protective effect of tents that have been set up in most of the semi-permanent camps established. The location of semi permanent and permanent settlements may have a significant effect on the risk of infection. Epidemiological studies from other parts of Sri Lanka have shown that people living within 750 m of a stream with An. culicifacies breeding, were at significantly higher risk for malaria than people living further away [ 29 ]. Conclusions This paper provides maps of both P. vivax and P. falciparum malaria incidence distribution on the island of Sri Lanka at district resolution in the 10 months preceding the tsunami, and an analysis of monthly malaria incidence in the country since January 1995. The malaria incidence was historically low, which implies a limited parasite reservoir in the human population. In spite of the fact that the months of December - February are normally the peak period for transmission, given the transmission level in the months leading up to the disaster, the risk of a large-scale outbreak seems to be limited. However, the low transmission levels over the past years may also have made people less alert to possible outbreaks, and the population would have less protective immunity towards the disease. The environmental changes resulting from the tsunami do not create particular opportunities for breeding of the principal malaria vector An. culicifacies but potential does exist for less important species such as An. subpictus and An. vagus . People living in emergency camps or returning to pre-disaster areas of residence are at higher risk of mosquito bites than normal. In spite of the emergency, the capacity of the public health authorities to perform malaria preventive and curative interventions remains high and essential supplies and staff capacity is not a problem. However, co-ordination of assistance and maintaining a strong surveillance system remain significant areas of concern. Increased attention to the establishment of a monitoring system including both parasitological and entomological parameters is recommended. Likewise, the large inflow of donated drugs and insecticides outside government control will potentially have long term implications on malaria control and case management, and especially the quality of administered drugs and the development of drug resistance requires careful monitoring. Authors' contributions GNLG collected the malaria data. OJTB checked the data, calculated incidence, made the maps. FK, FPA and PHA performed the focused literature review. All authors helped write, read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548668.xml

539250

BAG-1 haplo-insufficiency impairs lung tumorigenesis

Background BAG-1 is a multifunctional co-chaperone of heat shock proteins (Hsc70/Hsp70) that is expressed in most cells. It interacts with Bcl-2 and Raf indicating that it might connect protein folding with other signaling pathways. Evidence that BAG-1 expression is frequently altered in human cancers, in particular in breast cancer, relative to normal cells has been put forward but the notion that overexpression of BAG-1 contributes to poor prognosis in tumorigenesis remains controversial. Methods We have evaluated the effect of BAG-1 heterozygosity in mice in a model of non-small-cell lung tumorigenesis with histological and molecular methods. We have generated mice heterozygous for BAG-1, carrying a BAG-1 null allele, that in addition express oncogenic, constitutively active C-Raf kinase (SP-C C-Raf BxB) in type II pneumocytes. SP-C C-Raf BxB mice develop multifocal adenomas early in adulthood. Results We show that BAG-1 heterozygosity in mice impairs C-Raf oncogene-induced lung adenoma growth. Lung tumor initiation was reduced by half in BAG-1 heterozygous SP-C C-Raf BxB mice compared to their littermates. Tumor area was reduced by 75% in 4 month lungs of BAG-1 haploinsufficient mice compared to mice with two BAG-1 copies. Whereas BAG-1 heterozygosity did not affect the rate of cell proliferation or signaling through the mitogenic cascade in adenoma cells, it increased the rate of apoptosis. Conclusion Reduced BAG-1 expression specifically targets tumor cells to apoptosis and impairs tumorigenesis. Our data implicate BAG-1 as a key player in oncogenic transformation by Raf and identify it as a potential molecular target for cancer treatment.

Background BAG-1 is a multifunctional protein that is expressed in most cells. Originally identified as a Bcl-2 binding protein [ 1 ], other interaction partners of BAG-1 were described, including the serine threonine kinase C-Raf [ 2 ]. The C-terminal "BAG domain" of BAG-1 mediates the interaction with the Hsc70 and Hsp70 heat shock proteins [ 3 ], molecular chaperones that bind proteins in non-native states assisting them to reach a functional active conformation [ 4 ]. BAG-1 acts as a nucleotide exchange factor in this activation cycle [ 3 ]. The above findings indicated that BAG-1 might connect protein folding with other signaling pathways. Signaling networks promoting cell growth and proliferation are frequently deregulated in cancer [ 5 ]. The classical mitogenic cascade transmits stimuli from growth factor receptors via Ras, Raf, MEK and ERK to the cell nucleus [ 6 ]. C-Raf, like A- and B-Raf kinases also act at the outer membrane of mitochondria to augment cell survival [ 7 , 8 ]. Previously we had observed the stimulation of C-Raf kinase activity by BAG-1 in vitro [ 2 ]. Ras and B-Raf mutations have been found in various human cancers [ 9 , 10 ]. Evidence that BAG-1 expression is frequently altered in human cancers, in particular in breast cancer, relative to normal cells has been put forward but the notion that overexpression of BAG-1 contributes to poor prognosis in tumorigenesis remains controversial [ 11 ]. Methods Animals Mice used in these studies were generated and maintained according to protocols approved by the animal care and use committee at University of Würzburg. To inactivate the BAG-1 gene, we constructed a vector where exons 1 and 2 are replaced with a neomycin resistance gene. A phage clone with a 15-kb genomic insert from mouse strain 129/Sv spanning all seven exons of BAG-1 was identified and characterised using standard methods. The targeting construct contained 1,1-kb from the BAG-1 locus upstream of the neomycin resistance gene of plasmid pPNT [ 12 ] and 6-kb downstream. The upstream arm of 1,1 kb is located 5' to the start codon in the first exon of BAG-1 and the 3' arm of 6 kb is located downstream of exon 2. The mutation was introduced into embryonic stem cells by homologous recombination. Positive clones were identified by Southern blot analysis. Germline transmitting chimeras were obtained and bred to C57BL/6 mice. Further details will be described elsewhere. Heterozygous BAG-1 mice were genotyped by a PCR assay. The targeted BAG-1 allele was detected with primers P1 (5'-GAG TCT CCC GAT CCC TTT TCC), located upstream of exon 1, and P2 (5'-GAT TCG CAG CGC ATC GCC TT), located in the neomycin resistance gene, yielding a product of 600 base pairs. BAG-1 heterozygous mice were backcrossed at least three times onto C57BL/6 background before crossing with SP-C C-Raf BxB mice. Lung tumour mice expressing oncogenic C-Raf BxB were backcrossed at least six times onto C57BL/6 background. Western blot For the analysis of BAG-1 expression, lung lysates of the indicated genotypes were separated on 12,5% polyacrylamide-SDS (sodium dodecyl sulphate) gel, transferred to nitrocellulose Protran BA83 membrane (Schleicher&Schüll) and probed with rabbit anti-BAG-1 (FL-274) antibody (1:250, Santa Cruz Biotechnology). Amounts of protein were determined by Bradford protein assay to ensure equal protein loading for the analysis. Blots were developed using the appropriate horseradish peroxidase coupled secondary antibody and the ECL system (Amersham Pharmacia Biotech). Subsequently, the membrane was stripped and reprobed with rabbit antibody to glyceraldehyde 3 phosphate dehydrogense (1:2000, ab9485, Abcam Ltd.). Histopathology and immunohistochemistry Animals were sacrificed and lungs were fixed under 25 cm water pressure with 4% paraformaldehyde and embedded in paraffin. 5 μm sections were stained with hematoxylin and eosin and analysed. Pictures were taken using a Leica DMLA microscope and a Hitachi HV-C20A colour camera. Immunohistochemical staining to detect activated caspase-3, phospho-ERK (extracellular signal-regulated kinase), PCNA (proliferating cell nuclear antigen) have been described elsewhere [ 13 ]. Apoptotic, PCNA and p-ERK indices were determined by evaluating randomly chosen adenomas or fields of normal lung in 3–4 sections and determining the percentage of positive cells per 2000 cells at ×400. Results and discussion BAG-1 heterozygosity impairs C-Raf driven tumorigenesis In order to assess the functional role of BAG-1 on tumorigenesis, we have generated a null allele of BAG-1. To inactivate the BAG-1 gene, exons 1 and 2 were replaced with a neomycin resistance gene. This strategy was chosen to disrupt the expression of all known isoforms of BAG-1 which are generated by alternate translation initiation of a single mRNA; the start codons are present in exons 1 and 2. Western blot analysis of liver protein extracts of BAG-1 deficient embryos showed the complete loss of all BAG-1 protein isoforms. Embryos homozygous for this allele died at midgestation at around E13,5, but the heterozygous animals (BAG-1 +/- ) are normal. A comprehensive description of the BAG-1 -/- phenotype is subject of another manuscript. Previously, we had generated a lung cancer mouse model by targeting constitutively active C-Raf kinase (SP-C C-Raf BxB) to the lung [ 14 ]. These mice develop multifocal adenomas early in adulthood. Based on the observation, that BAG-1 can activate C-Raf [ 2 ], we asked whether heterozygosity for BAG-1 would affect C-Raf BxB driven adenoma growth. We observed that lung tumour initiation was reduced by half in 1, 2 and 4 months old BAG-1 +/- mice transgenic for SP-C C-Raf BxB compared to their BAG-1 +/+ littermates. Tumour area was reduced by 75% in 4 month lungs of BAG-1 haploinsufficient mice compared to mice with two BAG-1 copies, see Figure 1 . The histological picture emphasises the difference in adenoma formation between a representative SP-C C-Raf BxB/BAG-1 +/+ and SP-C C-Raf BxB/BAG-1 +/- lung. The difference in the staining intensity of the two lung sections derives mainly from the observation that the adenoma cells have a tendency to bind more intensively hematoxylin and eosin compared to normal lung cells. Thus, reduction of the BAG-1 gene dosage impairs the oncogenic activity of C-Raf in vivo. Reduced BAG-1 expression in BAG-1 heterozygous lungs Quantitative immunoblots demonstrated that the specific BAG-1 protein concentration in the lungs of BAG-1 +/- mice was half the amount of BAG-1 +/+ littermates, see Figure 2a . Moreover, immunohistochemical staining showed that BAG-1 was expressed in adenoma cells, see Figure 2b . There was no obvious difference in the BAG-1 immunohistochemistry of SP-C C-RafBxB/BAG-1 +/+ and SP-C C-RafBxB/BAG-1 +/- lungs. Tumour cells of BAG-1 heterozygous mice show increased apoptosis Concerning the molecular mechanism how a reduction of the BAG-1 protein expression in the heterozygous mice would impair tumorigenesis, we determined the fraction of apoptotic cells. Staining for activated caspase-3 revealed indistinguishable apoptosis in healthy regions of the lung of 1 month old SP-C C-Raf BxB mice with either one or two BAG-1 alleles, in line with the unaltered, normal lung structure of BAG-1 +/- mice. In the adenomas, however, we observed a significant increase of apoptotic cells in BAG-1 +/- SP-C C-Raf BxB mice compared with their SP-C C-Raf BxB/BAG-1 +/+ littermates, see Figure 3a . This mechanism of action of BAG-1 on the regulation of cell survival is compatible with the phenotype of embryonic day 12,5 BAG-1 null embryos. Immunohistochemical staining for activated caspase-3 and trypan blue staining of dissociated cells showed hypocellularity and elevated levels of apoptosis in the livers of BAG-1 -/- embryos (unpublished observations). Proliferation and p-ERK signalling are unaffected in BAG-1 heterozygous mice To exclude the alternative mechanism that the decreased level of BAG-1 expression in heterozygous animals would cause reduced cell proliferation in the adenomas, we performed proliferating cellular antigen (PCNA) staining. No significant differences were observed in the fraction of proliferating adenoma cells between SP-C C-Raf BxB animals heterozygous or wild type for BAG-1, see Figure 3b . Also, the percentages of adenoma cells positive for Ki-67, another proliferation marker and Bmi-1, a chromatin-associated protein expressed in stem cells, were not affected by the BAG-1 heterozygosity (not shown). Furthermore, staining of lung sections for phosphorylated ERK revealed no quantitative differences in the adenomas of SP-C C-Raf BxB animals heterozygous or wild type for BAG-1, see Figure 3c . Thus, signalling through the mitogenic cascade was not affected by the BAG-1 heterozygosity in the adenoma cells. Conclusions Tumours often are highly dependent on signalling pathways promoting cell growth or survival and may become hypersensitive to downregulation of key components within these signalling cascades. This study identifies BAG-1 as a protein specifically required at wild type expression levels for the survival of tumour cells and reveals it as potential anticancer target. Since many key components of survival pathways are regulated by interaction with (co-)chaperones [ 15 ], our finding is not without precedent but novel insofar as we have uncovered that reduced BAG-1 expression specifically targets tumour cells to apoptosis and impairs tumorigenesis. Whether this effect on adenoma cell survival requires that BAG-1 interacts with C-Raf or Hsc70/Hsp70 or with both partners requires additional studies. Questions concerning specific roles of the different BAG-1 isoforms were not addressed with this BAG-1 deficient mouse as both isoforms of BAG-1, p50 and p32 are absent in protein extracts of knock-out embryos. Another setting where BAG-1 has a physiological role is the heart, where up-regulation of BAG-1 after ischemia rescues cells from apoptosis [ 16 ]. A possible model combining the findings of this report and other data indicates that BAG-1 functions as an activator of C-Raf at the outer mitochondrial membrane where enzymatically activated C-Raf finds apoptosis-related targets such as BAD [ 17 ], see Figure 4 . We can purify overexpressed C-Raf either in an enzymatically inactive form in a complex with Hsp70 or in an enzymatically active form in a complex with Hsp90/50 (unpublished observations), and BAG-1 is proposed to regulate this activation with ATP generated in the mitochondria. Experiments dealing with this questions are currently ongoing. Therefore, the therapeutic efficacy of a standard chemotherapeutic agent [ 13 ] should be increased dramatically by co-application with a BAG-1 inhibitor, since it would target the adaptability of cancer cells to environmental stress and overcome their genetic plasticity. One way to reduce BAG-1 expression is through use of RNA interference-based gene silencing, in particular as BAG-1 overexpression has been observed in human tumours [ 11 ]. Drugs that bind to the ATP binding site of Hsc70/Hsp70 might also be expected to be effective as they would inhibit the interaction of BAG-1 with the ATPase domain of heat shock proteins. Such new specific BAG-1 inhibitors may be identified, aided by the known three-dimensional structure of the BAG domain [ 18 , 19 ]. Competing interests The author(s) declare that they have no competing interests. Authors' contributions RG caried out the molecular and histological studies and participated in the design and co-ordination of the study. BWK carried out the histological and immunohisto-chemical studies. GC participated in the histological and immunohistochemical experiments. URR participated in the design and co-ordination of the study. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539250.xml

539278

Effects of the group I metabotropic glutamate receptor agonist, DHPG, and injection stress on striatal cell signaling in food-restricted and ad libitum fed rats

Background Chronic food restriction augments the rewarding effect of centrally administered psychostimulant drugs and this effect may involve a previously documented upregulation of D-1 dopamine receptor-mediated MAP kinase signaling in nucleus accumbens (NAc) and caudate-putamen (CPu). Psychostimulants are known to induce striatal glutamate release, and group I metabotropic glutamate receptors (mGluR) have been implicated in the cellular and behavioral responses to amphetamine. The purpose of the present study was to evaluate whether chronic food restriction increases striatal MAP kinase signaling in response to the group I mGluR agonist, DHPG. Results Western immunoblotting was used to demonstrate that intracerebroventricular (i.c.v.) injection of DHPG (500 nmol) produces greater activation of ERK1/2 and CREB in CPu and NAc of food-restricted as compared to ad libitum fed rats. Fos-immunostaining induced by DHPG was also stronger in CPu and NAc core of food-restricted relative to ad libitum fed rats. However, i.c.v. injection of saline-vehicle produced greater activation of ERK1/2 and CREB in CPu and NAc of food-restricted relative to ad libitum fed rats, and this difference was not seen when subjects received no i.c.v. injection prior to sacrifice. In addition, although DHPG activated Akt, there was no difference in Akt activation between feeding groups. To probe whether the augmented ERK1/2 and CREB activation in vehicle-injected food-restricted rats are mediated by one or more GluR types, effects of an NMDA antagonist (MK-801, 100 nmol), AMPA antagonist (DNQX, 10 nmol), and group I mGluR antagonist (AIDA, 100 nmol) were compared to saline-vehicle. Antagonist injections did not diminish activation of ERK1/2 or CREB. Conclusions These results indicate that a group I mGluR agonist induces phosphorylation of Akt, ERK1/2 and CREB in both CPu and NAc. However, group I mGluR-mediated signaling may not be upregulated in food-restricted rats. Rather, a physiological response to "i.c.v. injection stress" is augmented by food restriction and appears to summate with effects of the group I mGluR agonist in activating ERK1/2 and CREB. While the augmented cellular response of food-restricted rats to i.c.v. injection treatment represents additional evidence of enhanced CNS responsiveness in these subjects, the functional significance and underlying mechanism(s) of this effect remain to be elucidated.

Background Chronic food restriction increases central sensitivity to rewarding and motor-activating effects of psychostimulants and direct dopamine receptor agonists [ 1 ]. Corresponding adaptive changes at the cellular level include increased psychostimulant-induced DA release in nucleus accumbens core [ 2 ], and upregulation of D-1 DA receptor-mediated MAP kinase signaling in dorsal and ventral striatum, with consequent increased activation of CREB and the immediate early gene, c-fos [ 3 , 4 ]. Changes in striatal glutamate receptor function have not been investigated but are of interest in light of findings that (i) psychostimulants induce striatal glutamate release [ 5 - 7 ], (ii) amphetamine-induced activation of striatal MAP kinase, CREB, and c-fos is attenuated by a group I metabotropic glutamate receptor antagonist [ 8 ], and (iii) the augmented effects of striatal D-1 receptor stimulation in food-restricted rats include hyperphosphorylation of the NMDA receptor NR1 subunit [ 9 ]. The group I mGluRs (mGluR1/5) are of particular interest. Group I mGluRs are densely expressed in striatal medium spiny neurons [ 10 ] and activate phospholipase C, resulting in hydrolysis of phosphoinositides and activation of Ca 2+ -dependent signaling cascades [ 11 ]. The group I mGluR agonist, DHPG, increases the phosphorylation of extracellular signal-regulated kinase (ERK) and the transcription factor CREB when infused into dorsal striatum [ 12 ], as do cocaine injected systemically and the D-1 agonist SKF-82958 injected into the lateral ventricle [ 4 , 13 ]. DHPG also elicits hyperlocomotion resembling that induced by DA receptor agonists, and the effect is not attenuated by the D-1 DA receptor antagonist, SCH23390 [ 10 ]. It is therefore possible that glutamate and group I mGluR function contribute to the changes in psychostimulant-induced behavioral responses and striatal cell signaling in food-restricted subjects. The purpose of the first experiment of this study was to compare dorsal and ventral striatal cell signaling in response to intracerebroventricular (i.c.v.) administration of the group I mGluR agonist, DHPG, in ad libitum fed and food-restricted rats. A dose of 500 nmol was used based on pilot work and published reports indicating that this dose is sufficient to exert behavioral effects [ 14 ] and stimulate PI hydrolysis [ 11 ] while being below threshold for producing convulsive behavior [ 15 ]. A second experiment was based on results of the first experiment and sought to elucidate the observed increase in striatal ERK1/2 and CREB phosphorylation in food-restricted rats injected i.c.v. with saline-vehicle. Results Experiment 1 In both the caudate-putamen and nucleus accumbens ERK1/2 phosphorylation was increased by food restriction (Cpu: F 1,18 = 10.8, p < .005; NAc: F 1,18 = 8.2, p < .01) and by DHPG injection (Cpu: F 1,18 = 6.6, p < .02; NAc: F 1,18 = 10.6, p < .005). However, there was no interaction between feeding condition and drug treatment in either brain region (Cpu: F 1,18 = 0.6; NAc: F 1,18 = 0.3). This analysis supports the impression (see Figures 1 & 2 , top panel) that food restriction did not actually increase the group I mGluR-mediated response to DHPG; rather, increased signaling in the food-restricted group, irrespective of i.c.v. injection treatment, appears to have summated with DHPG to produce a greater net effect in the food-restricted than ad libitum fed group. An identical pattern of results was obtained for CREB phosphorylation (Figures 1 & 2 center panel; Cpu: F diet; 1,18 = 6.2, p < .025; F drug; 1,18 = 3.9, p = .06; F diet × drug; 1,18 = 1.2; NAc: F diet; 1,18 = 12.9, p < .0025; F drug; 1,18 = 15.3, p < .001; F diet × drug; 1,18 = 2.7, p > .10). Interestingly, Akt phosphorylation was induced by DHPG but did not differ between food-restricted and ad libitum fed groups in either the Cpu (F drug; 1,18 = 4.5, p < .05; F diet; 1,18 = 0.4; F diet × drug; 1,18 = 0.4) or NAc (F drug; 1,18 = 12.7, p < .0025; F diet; 1,18 = 0.4; F diet × drug; 1,18 = 2.4, p > .10; Figures 1 & 2 bottom panel). In previous studies, striatal Fos-immunostaining was greater in food-restricted relative to ad libitum fed rats injected i.c.v. with d-amphetamine or the D-1 DA receptor agonist SKF-82958, but not saline-vehicle [ 3 , 16 ]. Therefore, in a small sample of ad libitum fed and food-restricted subjects, brains were processed for Fos-immunostaining and revealed stronger DHPG-induced staining in CPu (t(7) = 3.6, p < .01) and NAc core (t(7) = 2.8 p < .025) but not NAc shell (t(7) = 1.1) of food-restricted, relative to ad libitum fed, subjects (see Figure 3 ). During the period between DHPG injection and sacrifice, rats, which remained in their home cages, did not display any obvious behavioral responses to the drug. Experiment 2 In rats without lateral ventricular cannulas that received no injection treatment prior to sacrifice, ERK1/2 phosphorylation did not differ between feeding groups (Figure 4 ; Cpu: t(8) = 1.2; NAc: t(8) = 1.0), nor did CREB phosphorylation (Cpu: t(8) = 1.2; NAc t(8) = 0.1). In food-restricted rats injected with GluR antagonists prior to sacrifice, neither ERK1/2 nor CREB phosphorylation differed from that observed in food-restricted rats injected with saline-vehicle (Figure 5 ). Discussion Adding to the results of Choe and Wang who demonstrated DHPG-induced activation of ERK1/2 and CREB in dorsal striatum [ 12 ], the present study has demonstrated DHPG-induced activation of ERK/12 and CREB in both CPu and NAc following lateral ventricular infusion. While i.c.v. injection allows for the possibility that these effects were not directly mediated by striatal group I mGluRs, the proximity of striatum to the lateral ventrical and the high density of mGluRs in these structures support the likelihood of direct effects. Direct effects are also supported by the observed activation of Akt which mediates group I mGluR signal transduction [ 17 ]. As previously seen in rats challenged with the D-1 dopamine receptor agonist, SKF-82958, DHPG-induced activation of ERK1/2, CREB and c-Fos were greater in food-restricted than ad libitum fed rats. However, the pattern of results in this study – i.e. augmented ERK1/2 and CREB activation in food-restricted rats injected with vehicle, and no difference in Akt activation between feeding groups – suggests that group I mGluR signaling was not enhanced. Rather, a cellular-activating effect of the injection procedure was enhanced by food restriction and summated with the effect of DHPG. This response of vehicle-injected subjects was clearly a response to some aspect of the i.c.v. injection procedure because similarly food-restricted rats that were not challenged in any way prior to sacrifice displayed levels of pERK1/2 and pCREB that were essentially identical to those of ad libitum fed rats. Furthermore, the augmented activation of ERK1/2 and CREB in vehicle-injected food-restricted rats was not mediated by Akt because Akt, though activated by DHPG, did not differ between food-restricted and ad libitum fed rats injected with vehicle or DHPG. Attribution of the augmented DHPG-induced ERK1/2 and CREB phosphorylation to a "nonspecific" response to i.c.v. injection distinguishes the group I mGluR from the D-1 DA receptor. Although a modestly enhanced ERK response to i.c.v. vehicle infusion was seen in the NAc of food-restricted rats in the prior study, the dramatically increased activation of ERK1/2 and CREB by SKF-82958 in CPu and NAc was dissociable from any response to the injection procedure [ 4 ]. Even mild stressors are known to stimulate DA [ 18 ] and Glu [ 19 , 20 ] release in striatum, and it is possible that food-restriction augments this physiological response or the cell signaling induced by it. Because the ionotropic (AMPA and NMDA) as well as group I mGluRs are abundant in striatum and all three receptor types mediate MAP kinase signaling [ 12 , 21 , 22 ], corresponding antagonists were injected i.c.v. and compared to vehicle. It was reasoned that if the physiological response to the injection procedure involves one of these GluRs, ERK1/2 and CREB phosphorylation would be decreased, relative to vehicle, in the group(s) receiving the corresponding antagonist. The MK-801 treatment was also a potential probe for mediation by D-1 DA receptors because D-1 DA agonist-induced activation of ERK1/2 and CREB are dependent on the NMDA receptor in food-restricted subjects [ 9 ]. Results of this test did not provide support for mediation by one of the GluR types. Because only one dose of each antagonist was used, these results must be considered preliminary. However, considering the proximity of striatum to the lateral ventricle, and the fact that the doses used have exerted measurable cellular, physiological or behavioral effects in other studies [ 23 - 26 ] there is some doubt about involvement of GluRs in the effect of "injection stress". An interesting possibility to consider is mediation of the response by brain-derived neurotrophic factor (BDNF). Food-restricted rats have elevated striatal BDNF levels which appear to be involved in an enhanced neuroprotective response to diverse insults [ 27 ]. BDNF activates striatal ERK1/2 which, unlike glutamate-induced activation of Akt, ERK1/2 and CREB is not blocked by a PI 3-kinase inhibitor [ 21 ]. It is therefore possible that an enhanced BDNF-mediated activation of ERK1/2 and CREB in food-restricted rats represents a neuroprotective response to intraventricular infusion. Conclusions The present results indicate that a group I mGluR agonist activates Akt, ERK1/2 and CREB in both the CPu and NAc. Further, activation of ERK1/2, CREB, and c-Fos are stronger in food-restricted than in ad libitum fed rats. The augmented response is not attributed to increased Group I mGluR function but, instead, to an augmented response of food-restricted rats to the i.c.v. injection procedure. Some evidence casting doubt on attribution of the latter response to GluRs was obtained. It will be of interest to evaluate whether striatal MAPK signaling in food-restricted rats is generally augmented in response to stressors or whether this response is peculiar to i.c.v. infusion. In addition, it will be of interest to evaluate whether this physiological response is mediated by BDNF and its TrkB receptor. Methods Subjects and surgery All experimental procedures were approved by the New York University School of Medicine Institutional Animal Care and Use Committee and were performed in accordance with the "Principles of Laboratory Animal Care" (NIH publication number 85-23, revised 1996). Subjects were male Sprague-Dawley rats (375–425 g) housed individually in plastic cages with free access to food and water except when food restriction conditions applied. Animals were maintained on a 12-h light/dark cycle, with lights on at 07:00 h. Rats were anesthetized with ketamine (100 mg/kg; i.p.) and xylazine (10 mg/kg; i.p.) and stereotaxically implanted with a 26-gauge guide cannula (Plastics One, Roanoke, VA USA) in the right lateral ventricle. The cannula was permanently affixed to the skull by flowing dental acrylic around it and four surrounding mounting screws. Patency of the guide cannula was maintained with an occlusion stylet. Several days after surgery, cannula placements were confirmed by demonstration of a vigorous and short latency (i.e. <60 s) drinking response to 50 ng of angiotensin II. Food restriction and habituation Seven days following surgery, half the subjects were put on a food restriction regimen whereby a single 10 g meal of Purina (Gray Summit, MO USA) rat chow was delivered at approximately 17:00 h each day. Rats continued to have ad libitum access to water. Once body weight had declined by 20–25% (approximately 15 days) daily food allotments were titrated for an additional week to maintain stable body weight. During the 3 weeks required for food-restricted rats to attain and stabilize at target body weights, all rats were habituated, on five occasions, to the handling and injection procedures to be employed on the terminal day of the experiment. Drug treatment In Experiment 1, six food-restricted and six ad libitum fed rats received i.c.v injections of sterile 0.9% saline (5 μl). Six food-restricted and six ad libitum fed rats received i.c.v injections of the group I metabotropic glutamate receptor agonist, DHPG (500 nmol in 5 μl; Tocris Cookson, Ellisville, MO, USA). An additional five food-restricted and four ad libitum fed rats received i.c.v. injections of DHPG and brains were processed for Fos-immunostaining. In Experiment 2, six unoperated food-restricted rats and six unoperated ad libitum fed rats were prepared and habituated as above and received no unusual handling or drug injection on the terminal day of the experiment. In the second part of Experiment 2, twenty four food-restricted rats with lateral ventricular cannulas were prepared and habituated as in Experiment 1. On the terminal day of the experiment, six were injected i.c.v. with saline vehicle (5.0 μl), six were injected with the AMPA glutamate receptor antagonist, DNQX (10 nmol; Sigma-Aldrich, St. Louis, MO), six were injected with the NMDA glutamate receptor antagonist, MK-801 (100 nmol; Sigma-Aldrich), and six were injected with the group I mGluR antagonist, AIDA (100 nmol; Tocris Cookson). Antagonist doses were chosen on the basis of prior findings [e.g. [ 23 , 26 , 28 ]] and/or being just below threshold for producing noticeable motoric abnormalities. For i.c.v injection, solutions were loaded into a 30 cm length of PE-50 tubing attached at one end to a 250-μl Hamilton syringe filled with distilled water and at the other end to a 33-gauge injector cannula which extended 1.0 mm beyond the implanted guide. The 5.0 μl injection volume was delivered over a period of 95 s. One minute following injection, internal cannulas were removed, stylets replaced, and animals were returned to home cages. Lysate preparation Prior studies have indicated that MAP kinase activation is transient and the optimal time-point to study phosphorylation of ERK 1/2 after physiological or pharmacological treatment is 15–20 min [ 29 ]. Therefore, all rats, except the twelve unoperated/uninjected rats, were killed 20 min after injection by brief exposure to CO 2 followed by decapitation. The twelve unoperated/uninjected rats were simply removed from home cages and exposed to CO 2 followed by decapitation. Brains were rapidly removed and immediately frozen in powdered dry ice. Five hundred-micrometer sections were cut using an IEC Minotome cryostat, and CPu and NAc were micropunched, under an Olympus dissecting microscope, from a series of 8 consecutive frozen sections. The tissue was then homogenized in 10 volumes of 50 mM Tris-HCl, pH 7.5 containing 50 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1 mM Na 3 VO 4 , 40 mM β-glycerophosphate, 50 mM NaF and 5 mM Na 4 P 2 O 7 , 1% Tx-100, 0.5 μM okadaic acid, 0.5% sodium deoxycholate and 0.1% SDS, followed by centrifugation and protein determination using BCA reagent kit as described by the manufacturer (Pierce) Supernatants were mixed with 5 × SDS-PAGE sample buffer, boiled for 5 min, cooled on ice and kept at -80°C until future use. Western blotting Protein (10–30 μg per lane) was separated by electrophoresis on precast 10% polyacrylamide gels (Cambrex, East Rutherford, NJ, USA). Precision Plus protein standard molecular weight markers (Bio-Rad, Hercules, CA, USA) were also loaded to assure complete electrophoretic transfer and to estimate the size of bands of interest. The gels were transferred to nitrocellulose membrane (Osmonics) for 2 h, with a constant voltage of 100 volts. Membranes were blocked for 1 hr at room temperature with blocking buffer, 5% non fat dry milk in 50 mM Tris-HCl, pH 7.5 containing 150 mM NaCl and 0.1% Tween 20 (TBS-T), then probed overnight at 4°C using primary monoclonal antibodies for phospho-(Thr202/Tyr204)-p44/42 ERK1/2 (mouse monoclonal, 1:2000; Cell Signaling, Beverly, MA, USA), or polyclonal antibodies for phospho-Akt (Ser 473) (rabbit polyclonal, 1:1000 dilutuion, Cell Signaling), and phospho (Ser 133) CREB (rabbit polyclonal, 1:2000; Upstate Biotechnology, Lake Placid, NY, USA). Total levels of ERK1/2, Akt and CREB were detected on the same blots using anti-rabbit p42/44 ERK1/2 antibody 1:2000, (Cell Signaling), anti-rabbit total Akt (1:2000 dilution, Cell Signaling), or anti-rabbit CREB antibody (1:2000 dilution, Calbiochem). After detection of phosphorylated ERK1/2, phospho-Akt and CREB blots were stripped with 25 mM Glycine, pH 2.0 containing 1% SDS for 30 min at room temperature, washed six times in TBS-T buffer, blocked in blocking buffer for 1 h and then incubated overnight at 4°C in total ERK1/2, total Akt or CREB antibody. After probing with primary antibodies and washing with TBS-T buffer (3 × 5 min), membranes were incubated with 1:2000 dilution horseradish peroxidase conjugated anti-mouse or 1: 2000 dilution anti-rabbit IgG (Cell Signaling). Proteins were visualized using a chemiluminescense ECL kit (Pierce). Densitometric analysis of the bands was performed using NIH image software. Phospho-p42/44 MAPK, phospho Akt and phospho CREB values were normalized to total p42/44 MAPK, Akt and CREB values respectively. Immunohistochemistry Ninety min after DHPG injection rats were anesthetized with sodium pentobarbital (50.0 mg/kg,i.p.) and transcardially perfused with isotonic phosphate buffered saline (PBS) followed by 4% paraformaldehyde in PBS. Brains were then removed and maintained in 20% sucrose at 4°C for 48 h. Forty μm sections were cut on an IEC Minotome cryostat and collected in a cryoprotective solution. Fos immunostaining was carried out using a rabbit polyclonal c-Fos antiserum (Oncogene Science-Calbiochem, La Jolla, CA) and the avidin-biotin peroxidase complex (ABC; Vector laboratories). Sections were washed in 1% sodium borohydride followed by PBS and incubated for 2 hrs in 4% normal goat serum plus 1% BSA in PBS containing 0.2% Triton X-100 (Sigma-Aldrich) to block nonspecific binding. This was followed by incubation, overnight, with rabbit polyclonal c-fos antiserum (1:5000 dilution). Following several PBS washes, sections were incubated with a secondary antiserum (Vector, Burlingame, CA) for 60 min and subsequently reacted with avidin-biotin complex (ABC) (Vector). The peroxidase reaction was visualized with a chromogen solution containing 100 mM nickel sulfate, 125 mM sodium acetate, 10 mM imidazole, 0.03% diaminobenzidine (DAB), and 0.01% hydrogen peroxide at pH 6.5. Sections were then mounted on chrome-alum coated slides, dehydrated, and coverslipped. Objective counting of c-Fos positive cells in CPu at coronal levels +1.7 and -0.3 mm, and NAc core and shell at coronal level +1.5 mm in relation to bregma [ 30 ] was accomplished with a light microscope (Olympus, CK2) equipped with a Sony XC-77 video camera module coupled to an MCID image analysis system (Imaging Research Inc., St. Catherines, Canada). For each region, bilateral grain counts from three to five consecutive sections were measured to arrive at an average bilateral value per rat. Because results obtained from anterior and posterior levels of CPu were essentially the same, results from the two levels were combined and averaged for each rat. Data analysis For each Western blot, film exposure time was set as needed to visualize distinct bands in the control samples of each experiment. Immunoblots were analyzed using NIH imaging software. For each blot, relative phospho-protein levels were calculated from the ratio of optical density of the phosphorylated protein/total protein to correct for small differences in protein loading. In addition, tubulin levels were analyzed in several representative gels and no differences were observed between treatment groups. Results were expressed by comparison to the normalized control, which in Experiment 1 was defined as the ad libitum fed group injected with vehicle. In the first part of Experiment 2, unoperated/uninjected ad libitum fed rats served as control. In the second part of Experiment 2, food-restricted rats injected with i.c.v. saline vehicle served as control. Differences between treatment conditions in Experiment 1 were analyzed by two-way analysis of variance (ANOVA; feeding condition × injection treatment). Differences between treatment conditions in the first and second parts of Experiment 2 were analyzed by student's t-test and one-way ANOVA, respectively. Authors' contributions YP conducted the majority of immunoblotting experiments plus the immunohistochemistry experiment, contributed to experimental design and assisted in manuscript preparation. YB provided technical supervision of immunoblotting experiments and assisted in manuscript preparation. KC contributed to design of the study, assisted in all experiments, and wrote the final draft of the manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539278.xml

514498

Mapping the distribution of Loa loa in Cameroon in support of the African Programme for Onchocerciasis Control

Background Loa loa has recently emerged as a filarial worm of significant public health importance as a consequence of its impact on the African Programme for Onchocerciasis Control (APOC). Severe, sometimes fatal, encephalopathic reactions to ivermectin (the drug of choice for onchocerciasis control) have occurred in some individuals with high Loa loa microfilarial counts. Since high density of Loa loa microfilariae is known to be associated with high prevalence rates, a distribution map of the latter may determine areas where severe reactions might occur. The aim of the study was to identify variables which were significantly associated with the presence of a Loa microfilaraemia in the subjects examined, and to develop a spatial model predicting the prevalence of the Loa microfilaraemia. Methods Epidemiological data were collected from 14,225 individuals living in 94 villages in Cameroon, and analysed in conjunction with environmental data. A series of logistic regression models (multivariate analysis) was developed to describe variation in the prevalence of Loa loa microfilaraemia using individual level co-variates (age, sex, μl of blood taken for examination) and village level environmental co-variates (including altitude and satellite-derived vegetation indices). Results A spatial model of Loa loa prevalence was created within a geographical information system. The model was then validated using an independent data set on Loa loa distribution. When considering both data sets as a whole, and a prevalence threshold of 20%, the sensitivity and the specificity of the model were 81.7 and 69.4%, respectively. Conclusions The model developed has proven very useful in defining the areas at risk of post-ivermectin Loa -related severe adverse events. It is now routinely used by APOC when projects of community-directed treatment with ivermectin are examined.

Background Knowledge of the spatial distribution of Loa loa is important in countries involved in the African Programme for Onchocerciasis Control (APOC) and is also now a significant issue for the Global Programme to Eliminate Lymphatic Filariasis (GPELF). Both programmes rely on the wide-scale distribution of anti-helminthic drugs to poor communities using community-directed drug distribution schemes. A problem was first observed in Cameroon where a series of reports of severe and sometimes fatal encephalopathic reactions to ivermectin (Mectizan ® ) in individuals with high Loa loa microfilarial counts was made [ 1 - 3 ]. Similar problems may occur when albendazole (used to control lymphatic filariasis) is distributed in Loa loa endemic areas although evidence for this is contradictory [ 4 - 7 ]. Loa loa is associated with tropical "eye worm" (migration of adult worms across the sub-conjunctiva), Calabar swelling, oedemas and prurities but is not considered as pathogenic as other filarial worms, and is consequently less well studied. It is transmitted by species of horse-flies ( Chrysops spp.), most commonly Chrysops dimidiata and C. silacea which inhabit the forest areas of West and Central Africa, extending to the Ethiopian border. Severe and fatal reactions to ivermectin have been associated with individuals with high Loa loa microfilarial loads (more than 30,000 microfilariae (mfs) per ml of blood) and those with more than 50,000 mfs/ml are considered at very high risk [ 1 , 8 - 10 ]. Changes in the protocols for drug administration and post-treatment surveillance in areas considered to be at-risk of severe reactions to ivermectin have been implemented [ 11 , 12 ], but the detailed geographic distribution of Loa loa remains unclear [ 13 ]. Such information is essential in terms of developing safe treatment, and above all surveillance strategies across the region and, given the vast area which may be affected, it is recognized that rapid assessment methods must be developed to evaluate the risk of severe reactions in communities co-endemic for loiasis and onchocerciasis. Recent studies have shown that there is a close relationship between intensity of microfilariae infection and prevalence rates of Loa loa [ 14 , 15 ] suggesting that a distribution map based on prevalence of infection alone (and not intensity, which would require time-consuming counting of mfs) would provide sufficient information to delineate areas of high risk of severe reactions. The aim of our study was to develop a map indicating the areas where Loa loa infection may be high enough (i.e. with a prevalence of Loa microfilaraemia in adults exceeding 20%) that poses an operational problem for drug distribution by the Community-Directed Treatment with Ivermectin (CDTI) strategy [ 16 ]. As a preliminary step we created a risk model for Loa loa in West and Central Africa based on the relationship of crude Loa loa prevalence data (obtained from a literature search) to a wide range of environmental variables. Initial results suggested that land/forest cover derived from NOAA-AVHRR satellite data (i.e. collected by the Advanced Very High Resolution Radiometer on board the satellite series operated by the National Ocean and Atmospheric Administration of the USA), and soil type (from the FAO digitised soil map of Africa), are significant predictors and a preliminary risk map was produced [ 17 ]. However, when this first model was tested against field data from Cameroon it was found to poorly represent areas of high risk of infection in certain districts [ 18 ]. Possible reasons for this include: (a) the low spatial resolution of the satellite and environmental data used (1 km) which was unable to identify narrow gallery forests in savannah areas and/or (b) the use of prevalence data from various sources which had not been standardised by age or sex. In order to improve the quality of epidemiological data being used to develop the model, we created a new prevalence database from a series of surveys conducted by the team of the Institut de Recherche pour le Développement (IRD) at the Centre Pasteur du Cameroun (CPC). A second independent data set, collected as part of a project supported by the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR) to calibrate a rapid clinical diagnosis of Loa loa (RAPLOA) [ 15 ], was used to validate the results of the model. Methods Epidemiological data Epidemiological data used to create the model was obtained from a series of field studies undertaken by the IRD-CPC laboratory in Cameroon during the period 1991–2001 in which the presence of Loa loa parasites in the blood of individuals was assessed. The data come from five regions: (a) the forest-savannah mosaic area of the Mbam and Kim division (Central province); (b) the degraded forest area of the Lekie division (Central province); (c) the dense forest area of the southern part of the Central province; (d) the highland savannah area of the Western Province; and (e) the savannah areas of the districts of Banyo and Bankim, in the Adamaoua province. The Ministry of Public Health of Cameroon provided ethical approval for these surveys. The total number of subjects included in the analysis was 14,225 individuals for 94 villages. Study participants consisted of individuals over the age of 5 years who gave their consent or for whom consent was obtained from the parent or guardian. Individuals who had had filaricidal drugs, such as ivermectin or diethylcarbamazine, in the previous five years were not included in the study. The arrival of the research team in the villages was announced to the population one week before the event through the local authorities and teachers. The objective of the examinations were clearly explained, and it was stated that each individuals results would be returned. On the given day, the team settled at a central place of the village (usually at the chief's home, or in a school), and all the volunteers were examined between 10.00 and 16.00 hours. A standardized quantity of capillary blood was obtained with a non-heparinized micro-capillary tube. Prior to 1994, the amount taken was 30 μl, and thereafter, 50 μl. After Giemsa staining, the slides were examined under a microscope and the presence of Loa mfs was recorded. The latitude and longitude of all study villages were obtained from either the ordinance survey map or a global positioning system. Thus the data set included village name, longitude and latitude, alongside data on individuals examined (age, sex, standard size of blood sample taken and presence/absence of Loa loa infection). A summary of the epidemiological data used is presented in Table 1 . Table 1 Epidemiological data sets used in development of the spatial model for predicting the prevalence of Loa microfilaraemia Region Sex* Blood sample (μl) Prevalence of Loa loa Age range No. subjects No. villages Banyo-Bankim M 50.00 22.21 5–98 1783 16 Banyo-Bankim F 50.00 20.00 5–90 1815 South of Central Province M 30.00 32.81 5–80 381 5 South of Central Province F 30.00 27.37 5–80 453 South of Central Province M 50.00 24.76 5–86 941 17 South of Central Province F 50.00 18.49 5–90 1206 Lekie Department M 30.00 26.33 5–91 1052 14 Lekie Department F 30.00 19.10 5–92 1340 Lekie Department M 50.00 23.96 5–80 359 3 Lekie Department F 50.00 21.41 5–90 369 Mbam et Kim M 30.00 12.37 5–99 1293 24 Mbam et Kim F 30.00 4.53 5–95 1281 Western Province M 50.00 6.22 15–99 916 15 Western Province F 50.00 3.76 15–99 1036 * M = male and F = female General principles of the analysis The aim of the study was to identify variables which were significantly associated with the presence of a Loa microfilaraemia in the subjects examined. Besides data on individuals (age, sex, size of blood sample), we investigated whether some variables describing the environment of their place of residence (see list of variables below) were significantly related to the microfilaraemia. For these environmental variables, distance operatives indicating whether or not villages were within 5 km of potential at-risk variables were created. Distance operatives were based on the normal dispersal range of Chrysops which has been shown to be within 5 km of their breeding sites [ 19 ]. In order to design a valid modelling structure, univariate analysis was first undertaken of the relationship between, on the one side, individual and environmental variables, and on the other side the variable of interest, i.e. the presence/absence of Loa loa mfs in the individual. Then we developed a series of logistic regression models (multivariate analysis) to describe variation in the prevalence of Loa loa microfilaraemia using individual level co-variates (age, sex, μl of blood taken for examination) and village level environmental co-variates. Logistic regression is useful to predict the presence or absence of a characteristic or outcome based on values of a set of predictor variables. It is similar to a linear regression model but is suited to models where the dependent variable is dichotomous. Logistic regression coefficients can be used to estimate odds ratios for each of the independent variables in the model. In our example the dichotomous variable is the presence or absence of Loa loa infection in each individual in the sample. Those variables which did not have a significant relationship with age and sex adjusted prevalence were excluded from further analysis. All variables were analysed at individual level. Environmental variables and data sets included in the analysis The distribution of the human population in sub-Saharan Africa is relatively well described from population census data obtained at the national and sub-national level. The population data for 1990 used in this analysis was obtained from the Global Resource Information Database of the United Nations Environment Programme (UNEP-GRID), Eros Data Centre [ 20 ]. These population density surfaces (with 1 km resolution) have been derived from a model including population estimation for countries, population of major urban centres and transportation network and accessibility. They have been used in numerous studies involving population distribution in Africa, including studies on estimating the burden of malaria [ 21 ]. The USGS (United States Geological Survey) hydrologic digital dataset provides detailed descriptions of the topography of the area, including elevation (Digital Elevation Model, DEM), slope, aspect (direction of maximum rate of change in elevation between each cell and its eight neighbours and representing direction of slope), flow accumulation (defining amount of upstream area draining into each cell) [ 22 ]. Satellite data indicating the 'greenness' of the environment (Normalised Difference Vegetation Index, NDVI) was obtained from the VEGETATION sensor launched in 1998 onboard the SPOT 4 satellite system. NDVI images have been specifically tailored to monitor global land surfaces' parameters on a daily basis with a medium spatial resolution of one km [ 23 ]. They permit monitoring of seasonal and inter-annual variation in vegetation status and have been shown elsewhere to be important correlates of the spatial and temporal changes in the distribution of insect vectors of disease [ 24 ]. Free access is given to 10 daily synthesized maximum value composite products 3 months after insertion in the VEGETATION archive [ 25 ]. Thirty-six decadal NDVI images were obtained from SPOT VEGETATION satellite sensor data archive for 1999. The VEGETATION data used consists of 10 daily NDVI products, compiled from daily synthesis over the previous ten days, for the entire African continent, to which both radiometric corrections and geometric corrections have been applied to enhance product quality. The true value of NDVI was calculated from the 8 bit decadal images (1–255) using ((digital number*0.004)-0.1). From the true values images, mean, minimum, maximum, median, standard deviation were calculated for the year 1999. In order to take into account the flight distance of the vector, 5 km buffers were created around each study village and the mean value of the NDVI variable was obtained and used in the development of the logistic regression model. Validation of the model Epidemiological data used to validate the model was obtained from a multicentric study supported by TDR which was designed to assess the relationship between clinical and parasitological indicators of Loa loa endemicity [ 15 , 26 ]. The surveys were conducted by three research teams (comprising epidemiologists, parasitologists, social scientists and clinicians) based in Buea and Yaoundé, both in Cameroon, and in Calabar, Nigeria. Only data from Cameroon are used in our validation process. The Buea study sites were in South-West and North-West Provinces of Cameroon; in these areas, a total of 4532 individuals over the age of 15 years, living in 42 villages, participated in the survey. The study sites for the Yaoundé team were located in the East Province of Cameroon, where 3181 persons of the same age, living in 32 localities, were examined. A standardized questionnaire was administered to participants from whom finger-prick blood samples were collected and examined for Loa loa mfs. Model validation was undertaken through correlating model outputs with the independent data set. Results Univariate analysis While Loa loa prevalence when regressed against population density was not found to be significantly correlated, a strong and significant linear relationship ( r 2 = 0.7699; P < 0.001) was found with age below 40 years (Figure 1 ). In our analysis we used this linear relationship for all individuals below 40 but then treated all those above 40 as though they remained at this age since any subsequent change with age was deemed non-significant. This broken stick approach was considered to be a reasonable strategy and much simpler than fitting a non-linear model. Figure 1 Relationship between prevalence of Loa loa microfilaraemia and age. Prevalence rates were significantly ( P < 0.001) higher in males than in females (19.8% n = 6725 and 15.2% n = 7500 respectively). Sex was therefore entered (as a categorical variable) in the logistic regression model. As one would expect, prevalence rates were significantly higher ( P < 0.001) in individuals from whom the blood smear was prepared using a volume of 50 μl as compared with those from whom 30 μl of blood were taken. Blood sample size was therefore entered (as a categorical variable) in the logistic regression model. After investigating the value of the DEM and its associated files (slope, aspect, flow accumulation) in predicting microfilaraemia prevalence using univariate and multivariate regression statistics, the DEM alone was chosen as a predictive variable for the development of a logistic regression model. Because the relationship between Loa loa prevalence and the DEM was found to be non-linear (Figure 2 ) the DEM data was divided up into 250 m interval classes and treated as a categorical variable. Figure 2 Relationship between prevalence of Loa loa microfilaraemia and elevation. Since univariate analysis of the SPOT VEGETATION NDVI data against prevalence indicated a non-linear relationship [Figure 3a and 3b ], the satellite data were entered into the model as a series of numeric variables (the original, the square and the cube) for the mean, the minimum, the maximum and the standard deviation. Standard deviation and maximum NDVI were most strongly correlated with prevalence using univariate analysis. Figure 3 Relationship between prevalence of Loa loa microfilaraemia and standard deviation of NDVI (a); and between prevalence of Loa loa microfilaraemia and maximum annual NDVI (b). Multivariate analysis Separate and combined models were created which included all individual level co-variates and altitude for both buffered and non-buffered SPOT VEGETATION data variables. Final model choice was based on (a) the simplicity and biological validity of the model (b) the predictive capacity of the model when assessed using bootstrap methodology (where half of the data was extracted randomly from the data set and tested against model results developed from the remaining half) and (c) the ability of the model to be extended over the large geographic area involved in the APOC programme. Thus, besides the three individual-level covariates (sex, age and blood sample size), only three village level environmental co-variates were kept in our model: maximum annual NDVI, standard deviation of the NDVI, and elevation. In the final model, the probability of a female individual being infected with Loa loa is represented by: 1/(1+e - z ), where (for our example a female aged ≥ 40 years, and a blood sample of 50 μl): [Z] = ([Sex]*0.221) + ([Age] * 0.049) - ([Blood sample size]*0.542) + ([Maximum NDVI 2 ] * 74.38) - ([Maximum NDVI 3 ] * 58.816) + ([Standard deviation NDVI] * 11.788) + (([Elevation] < 250) * 1.133) + (([Elevation] = 250-500) * 0.865) + (([Elevation] = 500-750) * 0.981) + (([Elevation] = 750-1000) * 1.171) - (([Elevation] = 1000-1250) * 0.623) - (([Elevation] = 1250-1500) * 1.516) - (([Elevation] = 1500-1750) * 1.342) - (([Elevation] = 1750-2000) * 0.669) - 22.883 (constant), where [Sex] = 0 for females, and 1 for males; and [Blood sample size] = 0 for 30 μl, and 1 for 50 μl. Thus for our example, age, sex and blood sample size variables are presented for a fixed value because of their non-spatial nature. The probability of infection increases (+) with increasing value for sex (0 versus 1), with increasing age, and with increasing standard deviation of NDVI and Maximum NDVI 2 ; it decreases (-) with Maximum NDVI 3 (indicating that the relationship to maximum NDVI is non linear), and increases for categories of elevation between 0 and 1000 metres after which it decreases. The probability of infection decreases with increasing value of blood sample size (0 versus 1) even if larger blood sample should normally allow better detection of the infection. This could be due to the fact that larger samples have been taken often in places where infection rates are low, in the highlands, and this is presumably a result of interacting with the other variables. A similar model can be made for males and individuals at any other age. The model was developed using 50% of the data (randomly selected) and then tested on the remaining 50% of the data. When the model results were plotted against the observed data an r 2 of 0.6033 was obtained (Figure 4 ). In addition, when one takes the threshold of 20% as the Loa microfilaraemia above which there is a risk of post-ivermectin encephalopathy, the sensitivity and specificity of the model (respectively: the proportion of villages with a measured prevalence ≥ 20%, and which were correctly predicted as such by the model; and the proportion of villages with a measured prevalence <20%, and correctly identified as such by the model) were found to be 77.1 and 80.0%, respectively (Table 2 ). Figure 4 Relationship between observed prevalence of Loa loa microfilaraemia and predicted prevalence. Table 2 Distribution of the villages examined, according to their observed and predicted prevalence of Loa microfilaraemia Observed prevalence <20% ≥ 20% Predicted prevalence First dataset <20% 48 8 (collected by IRD-CPC) ≥ 20% 12 27 Dataset used for validation <20% 20 5 (collected by TDR) ≥ 20% 18 31 Mapping the model results: an Environmental Risk Map for Loa loa In order to standardize the variables relating to individuals (which cannot be mapped) we chose values that best represent the average adult population (above the age of 15) i.e. 50:50 male – female ratio, age 40 and a blood sample of 50 μl. Age 32 was the approximate average age of individuals in the data sets used to develop the model. Using age 40 (the average age of individuals >15 years of age) in the model was therefore likely to slightly over-represent infection in these data sets but the final model would be directly comparable with current policy to exclude individuals below the age of 15 from rapid epidemiological surveys for loiasis [ 14 , 15 ]. In order to create a model, which included both males and females, two separate models were created, one for each sex and the mean of the two model results taken. Validation of the model Since the chosen model was based on environmental variables, which could be mapped across the entire country, it was possible to extrapolate the model results to the whole of Cameroon (Figure 5 ) and compare them with the results of the independent TDR survey. Correspondence between observed prevalence from the TDR verification data set and the model results was found to be very close: the sensitivity and the specificity of the model, when using the prevalence threshold of 20%, were 86.1 and 52.6%, respectively. Five villages were observed to have high prevalence rates (>20%) despite being model predictions for prevalence below 20% (Table 2 ). Two villages, Nguri and Ngu, located in the North-West province, were classified as extremes (Figure 6 ). Inaccuracies in the georeferencing of the village location or the spatial data used to create the model may account for this, as could the possibility that the local population regularly visit the Chrysops infested area nearby. The three other villages (Ntem, Boum and Baktala) are all villages from the Eastern Province of Cameroon which were sited on very localised areas where model predictions for <20% prevalence occurred. These villages were all within 1 km of areas where model results predicted >20% prevalence. Figure 5 Predictive model of Loa loa prevalence for Cameroon overlaid with the observed prevalence data. Figure 6 Inset of verification villages in North-West province of Cameroon (note the position of Nguri and Ngu). When one considers both datasets as a whole, and the prevalence threshold of 20%, the sensitivity and the specificity of the model were 81.7 and 69.4%, respectively. Discussion Detailed information on distribution of disease and levels of endemicity is a prerequisite for effective planning of control programmes. This has been an important element in the planning of the African Programme for Onchocerciasis Control (APOC) where community-directed treatment with ivermectin (Mectizan ® ) has been targeted at areas of hyper- and meso-endemicity, identified by Rapid Epidemiological Mapping [ 27 ]. The expansion of the APOC programme in Cameroon has been delayed due to severe (sometimes fatal) adverse reactions in patients receiving ivermectin who were co-infected with Loa loa. This has led to caution in defining new areas for the expansion of the programme and a requirement for identifying areas of high risk of Loa loa using a potentially rapid and extensive methodology. Thomson et al . [ 17 ] developed a preliminary model using satellite mapping based on existing knowledge of distribution and prevalence, and recently a method of rapid assessment using a simple questionnaire, called RAPLOA, has been developed [ 15 ]. This paper further develops the satellite-derived risk map of Loa loa by using improved satellite data sets and detailed information on infection rates in villages in various ecological zones in Cameroon. The resulting model is able to predict prevalence risk throughout southern Cameroon with a greater accuracy than hitherto available. The final model has been chosen because of the good fit between observed and predicted prevalence and is based on elevation and 1 km SPOT VEGETATION data. In order to understand the value of the risk map it is important to note that (a) while the 1 km resolution of the environmental data cannot reveal the small and localised muddy breeding sites of the Chrysops vectors, their general habitat is well captured by the model and it is recognised that the 1 km resolution is much less than the dispersal capacities of the adult flies; and (b) the model is not dependant on the seasonal or inter-annual variability of the vector density, since this is not reflected in fluctuations in Loa microfilaraemias, because the adult worm life-span is much longer (4–17 years). Among the areas in which the prevalences of Loa loa are particularly high, according to the Environmental Risk Map (ERM), two should be pointed out, because they are also those where most of the cases of serious adverse events (SAEs) have been recorded so far. The first one is the Lekie Department (Central Province), in the Sanaga valley, where 53 of the 63 probable or possible Loa loa encephalopathy cases recorded between 1989 and 2001 in Cameroon have occurred [ 3 ]. Besides the high level of endemicity, this cluster of cases may be related to the high population density there, as well as the proximity of the area with the capital, Yaoundé, which probably led to a high reporting rate; however, other hypothesis should be considered, such as specific susceptibility of the local human populations, or special pathogenicity of the local Loa "strain" [ 28 , 29 ]. The latter possibility is currently investigated in Cameroon, and in the Mayumbe forest (Democratic Republic of Congo), another area where some 15 cases of fatal SAEs were reported in December 2003. The second area of interest shown by the ERM in Cameroon is the Tikar plain, near Bankim, between the Western, North-West and Adamaoua provinces, where several cases have been recently reported, though the area's vegetation is of shrub savanna type. The ERM also shows those areas where CDTI projects against onchocerciasis may be implemented in a near future, and where the risk of SAEs is probably high: the Eastern and Southern Provinces, and the north-western part of the Littoral Province. While the use of satellite-derived environmental data for evaluation of disease risk has been developed for several vector borne diseases (malaria, Rift Valley fever, visceral leishmaniasis, tick-borne encephalitis) [ 24 ], limited use of such data in public health programmes for the control of infectious diseases has been achieved. This paper defines a model which identifies areas of potential high risk of severe adverse reactions to ivermectin, and will contribute to APOC's programme development by enabling resources to be effectively targeted to areas deemed at risk. The Technical Consultative Committee of APOC, and the Mectizan Expert Committee have developed a series of recommendations aimed at "facilitating effective detection and management of SAEs following treatment with Mectizan in known and suspected Loa loa endemic areas". Three types of mass treatment strategies have been defined, according to the levels of endemicity of onchocerciasis and of loiasis, the latter being defined by the prevalence of Loa microfilariae (<20 % versus ≥ 20%) or of history of eye worm passage (<40 % versus ≥ 40%). In those areas where onchocerciasis is meso- or hyperendemic, and the prevalence of Loa mfs exceeds 20%, a number of detailed measures should be taken, regarding training of community distributors and medical personnel, availability of medical supplies, duration of distribution, surveillance of the treated populations, and management of the patients who develop a SAE. Though lower, the risk that SAE occur in areas where the prevalence of Loa microfilaraemia is lower than 20% is not nil, and thus guidelines have also been developed regarding the strategy to apply in such situations [ 30 ]. The model developed is by no means a definitive product but provides a basis for decision making in terms of where rapid epidemiological surveys for loiasis should now be targeted. Modelling in this way permits an iterative process between field epidemiologist and modeller which not only means that decisions are made on the best available data but that such data is updated rapidly as new survey results are entered in the database and the model refined. Models are valuable in so far as they reflect reality. Given the complexity of disease transmission processes where human, parasite, vector and environment interact, it is impossible to think that all relevant factors can be incorporated into a general model which can be applied on a regional scale. What is significant here is the fact that important decisions need to be made now with regard to the likely spatial extent of the distribution of Loa loa . As it stands, the model presented here uses environmental features known to be associated with the biology of the vector, is robust for the area of data collection and predicts areas within Cameroon where Loa loa has been found in the past. It provides a rapid assessment methodology of areas where adverse reactions to ivermectin may occur and will be further developed in conjunction with the RAPLOA procedure which APOC intends to apply to CDTI areas potentially endemic for Loa loa [ 15 ]. The next step in the process will be to update the model with new data (e.g. the verification data set) from all the countries where loiasis is endemic, and to explicitly represent uncertainty in the model outputs so that decision makers will be able to assess for themselves the quality of the model results for their area of interest. List of abbreviations APOC African Programme for Onchocerciasis Control CDTI Community-Directed Treatment with Ivermectin CPC Centre Pasteur du Cameroun DEM Digital Elevation Model ERM Environmental Risk Map for Loa loa GIS Geographic Information System GPELF Global Programme to Eliminate Lymphatic Filariasis IRD Institut de Recherche pour le Développement Mfs Microfilariae NDVI Normalised Difference Vegetation Index NOAA-AVHRR National Ocean and Atmospheric Administration – Advanced Very High Resolution Radiometer RAPLOA Rapid Assessment of the Prevalence and Intensity of Loa infection SAE Serious Adverse Event SPOT Satellite Pour l'Observation de la Terre TDR UNICEF/UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases UNEP-GRID United Nations Environment Programme-Global Resource Information Database USGS United States Geological Survey Competing interests None declared. Authors' contribution MCT designed the study, planned the analysis, supervised the modelling, interpreted the results, and wrote the paper. VO prepared the data, developed the model, prepared the maps, and wrote the paper. JK and JG collected the field data used for the development of the model. SW, IT and PE collected the field data used for the validation of the model. JHR supervised the study thanks to which the validation data were collected. DHM proposed the study, contributed to its design and interpretation of the results. MB designed the study, planned the analysis, collected the field data, interpreted the results, and wrote the paper.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514498.xml

549204

Structural comparison of metabolic networks in selected single cell organisms

Background There has been tremendous interest in the study of biological network structure. An array of measurements has been conceived to assess the topological properties of these networks. In this study, we compared the metabolic network structures of eleven single cell organisms representing the three domains of life using these measurements, hoping to find out whether the intrinsic network design principle(s), reflected by these measurements, are different among species in the three domains of life. Results Three groups of topological properties were used in this study: network indices, degree distribution measures and motif profile measure. All of which are higher-level topological properties except for the marginal degree distribution. Metabolic networks in Archaeal species are found to be different from those in S. cerevisiae and the six Bacterial species in almost all measured higher-level topological properties. Our findings also indicate that the metabolic network in Archaeal species is similar to the exponential random network. Conclusion If these metabolic network properties of the organisms studied can be extended to other species in their respective domains (which is likely), then the design principle(s) of Archaea are fundamentally different from those of Bacteria and Eukaryote. Furthermore, the functional mechanisms of Archaeal metabolic networks revealed in this study differentiate significantly from those of Bacterial and Eukaryotic organisms, which warrant further investigation.

Background Classification of biological organisms is of fundamental importance to evolutionary studies. It is commonly believed that there are three domains of life: Archaea, Bacteria and Eukaryote. Currently, the most popular classification method is the so called "molecular approach", in which polymorphism information in DNA or protein sequence is exploited to assess the phylogenetic relationships among species [ 1 , 2 ]. To a large extent, this is a "local" approach since the choice of sequence for comparison greatly affects the final result, "lateral gene transfer" (LGT) and thus the resulting "genome chimerism" further complicates the situation [ 3 ]. A new "system" approach that takes "global" properties of each organism into consideration serves as a potential alternative to overcome this shortcoming. Indeed, recent advances in system biology and increasingly available genomic databases have made it possible to rebuild biological networks from genomic data and have offered opportunity for such a "system" approach [ 4 ]. Podani and co-workers [ 5 ] proposed classifying organisms based on two kinds of network indices: the Jaccard index, which measures proportions of common sets of nodes in two networks, and Goodman-Kruskal γ function, which measures the similarity between rankings of nodes in two networks. They studied metabolic and information network structures of 43 organisms using these two measures under the hypothesis that network structure and the network design principle(s) behind them contain phylogenetic information. Ma and Zeng [ 6 ] conducted a more extensive phylogenetic classification study on 82 fully sequenced organisms based on different cellular function systems (enzyme, reaction, and genes) at the genomic level. They constructed phylogenetic tree based on Jaccard index and Korbel's definition, and concluded that in general, the classification based on network indices are in good agreement with the one obtained by analyzing the 16S rRNA using molecular approach. These studies seem to support the notion that significant differences in the network design principle(s) exist among the three domains of life [ 7 ]. These differences may reflect on the different approaches that organisms take to organize their entire systems to serve their special needs in the environment they live during the evolutionary history. Motivated by these encouraging results, in this manuscript, we went on to conduct a thorough comparison of network structural properties which provide further and more compelling evidences that significant differences exist among the network design principle(s) in organisms from the three domains of life. Restricted by the theoretical network structural studies, there are not many deterministic and informative topological measurements available [ 8 - 11 ]. The established measurements can be roughly divided into two categories: higher-level (global) properties and low-level (local) properties. The difference between the two is that one needs to know the whole network in order to calculate the higher-level property measures (e.g. average path length) while the low-level properties can be worked out locally (e.g. marginal degree of individual node) [ 9 ]. We use three groups of topological measurements (both low and higher-level) that address different aspects of the network structure. The first group contains network indices such as average clustering coefficient, average path length [ 12 ]. The second group is composed of degree distributions (both marginal and bivariate joint degree distributions) [ 8 - 11 , 13 ]. The third group is composed of network motif profiles that are recently shown to represent the network design principle(s) and global statistical properties of the network when aggregating together [ 14 - 16 ]. These measurements have been well studied in the network literatures, and are able to capture most aspects of network degree information. Single cell model organisms such as E. coli and S. cerevisiae have been studied intensively in biochemistry, cell biology and genetics; hence the rebuilt networks in those organisms present the best chance to approximate the true underlying network. Moreover, single cell organisms are less likely to have experienced the Whole Genome Duplication (WGD), which might drastically change the network structure [ 17 , 18 ]. As a result, we selected eleven single cell organisms to study their network structural properties: one Eukaryote: S. cerevisiae ; six Bacteria: E. coli , V. cholerae , R. solanacearum , B. subtilis , L. lactis , S. coelicolor ; and four Archaea: S. solfataricus , S. tokodaii , M. acetivorans , T. acidophilum . There are three main types of intracellular networks: the protein-protein interaction network, the transcriptional regulation network and the metabolic network. The first two are rebuilt by using high throughput techniques such as yeast two-hybrid system, in vivo pull down assay or DNA microarray, which are subject to high uncertainties, and the resulting networks may not be good approximation to biological complexity [ 19 - 22 ]. On the other hand, the metabolic network is derived from metabolic pathways, many of which are inferred from biochemical experiment-defined stoichiometries of many reactions [ 23 ]. It is well known that central pathways contain "hub nodes" of the whole metabolic network [ 24 , 25 ] and are also main building blocks of the so-called Giant Strongly Connected Component (GSCC) and Giant Weakly Connected Components (GWCC) [ 26 ]. The former is defined as the largest cluster of nodes within which any pair of nodes is mutually reachable from each other, and the latter is defined as the largest cluster of nodes within which each pair of nodes is connected in the underlying undirected graph [ 10 ]. Therefore, our high confidence in the structure of GSCC and GWCC, based on experimentally verified pathways, guarantees high confidence in whole network structure. The long history of biochemical studies of enzymes ensures relatively low false positive and low false negative rates of connections. Therefore, we decided to use metabolic networks in single cell organisms to compare network topological properties in the three domains of life. Results In constructing metabolic networks, Ma and Zeng [ 28 ] argued that connections through "current metabolites", which is referred to as cofactors in biochemistry such as ATP, ADP, H 2 O, should be removed from metabolic networks. We followed their suggestions by removing such "current metabolites" before conducting the following analysis. Group I measures: network indices Before checking different types of network topological measurements, we visually compared different metabolic networks (Fig. 1 ). Metabolic networks in S. cerevisiae and the six Bacterial species appear much more heterogeneous than Archaeal metabolic networks. It is well known that the so-called exponential random network (marginal degree distribution follows a Poisson distribution, see Methods for details) appears homogeneous while scale-free network (marginal degree distribution follows a power-law distribution, see Methods for details) appears more heterogeneous and modular [ 9 ]. Calculations of the two classic network indices, average clustering coefficient and average betweenness (see Methods for definition) also indicate that the metabolic networks in S. cerevisiae andthe six Bacterial species are more clustered and modular than those in the four Archaeal species (Table 1 , Fig. 2 ). From Table 1 and Fig. 2 , it is evident that the Clustering Coefficient (C) and Betweenness (B) did a better job in separating Archaeal species from non-Archaeal species than Average Path Length (L) and Diameter (D). Note that since we removed connections through "current metabolites" when constructing metabolic networks, our average path lengths are much longer than those reported in Jeong et al. [ 25 ] but similar to those reported in Ma and Zeng [ 28 ]. To avoid the confounding effects stemming from different network sizes, we calculated the so-called concentrations (number of appearances of subgraphs divided by the number of nodes with edges or arcs (directed edges), see Methods for details) of three-node subgraphs and four-node subgraphs. The concentration of subgraphs is an objective measure of the extent of clustering and modularity of the network [ 8 , 9 ]. It is observed that the concentrations of subgraphs in S. cerevisiae and the six Bacterial metabolic networks are much higher than those in Archaeal metabolic networks (Fig. 3 ). Group II measures: degree distributions Marginal degree distributions Recently, a variety of real-life networks are found to share the "scale-free" property, i.e. the marginal degree distribution follows a power-law distribution [ 25 , 29 - 31 ]. Our analysis demonstrates that the outgoing and incoming marginal degree distributions in metabolic networks also follow the power-law distribution. A simple linear model fits the log-transformed data well (except for the incoming degree distributions for most of the Archaea) which indicates that in general, the power-law model is appropriate to capture the structure of degree data (Fig. 4 ). Parameters were estimated using the Least Square method. The results together with goodness of fit measure R 2 and 95% individual confidence intervals are summarized in Table 2 and Table 3 . The estimated power-law index γ is around -0.3 in all cases and the estimated log-transformed scaling parameter α ranges within 2.0 to 2.5. These indicate that marginal degree distribution, which is a low-level (local) topological property measure, although showed some distinction, is not enough to effectively differentiate networks from different domains. Overall, metabolic networks in most of the species we studied seem to follow the power-law distributions and thus are "scale-free". The fact that the incoming degree distributions of most Archaeal species we studied do not follow power-law well (Fig. 4B ) suggests that networks in Archaeal species tend to be less "scale-free" and more "random-like" compared to those of the non-Archaeal species. As we have shown, marginal degree distribution alone does not reveal the fundamental network structural differences between the Archaeal species and the non-Archaeal species. Simulation studies have shown that randomized networks preserving marginal degree distribution can be quite different in terms of global (higher level) topological properties such as average clustering coefficient [ 9 ]. In metabolic networks, we are unable to determine the preferred types of reactions based on just marginal substrate or product degree distributions. Since the metabolic network is rebuilt from chemical reactions, joint behavior of substrate and product in reactions should be more informative than disjoint behavior of metabolites. Therefore, we calculate the joint degree distributions hoping to gain more insight into the network organization. Joint degree distributions Joint degree distribution measures and describes correlation between connectivities of neighboring nodes. N ( K 0 , K 1 ) is defined as the number of edges connecting nodes of connectivity K 0 to those of connectivity K 1 . For metabolic networks, which are directed, N ( K out , K in ) is used to measure the number of arches where substrate (node) with out-connectivity K out transforms to product with in-connectivity K in . This quantity reflects intrinsic properties of the network and can be used to distinguish different types of networks. For instance, we can test whether N ( K out , K in ) of a particular network differs significantly from that of the random network. To be specific, we calculate , where ( K out , K in ) represents the mean of random variable N ( K out , K in ) in a large number (say, 1000) of random networks simulated by an edge-rewiring algorithm proposed by Maslov and Sneppen [ 13 ], ( K out , K in ) denotes the estimated standard deviation of N ( K out , K in ). The p -value can then be obtained by compare Z to a standard normal distribution. Comparing with "properly" randomized network ensembles allows us to concentrate on those statistically significant patterns of the complex network that are likely to reflect the design principle(s) [ 13 ]. We calculated statistically significant correlation profiles (Z-score profiles, see Methods for details) for the metabolic network in each organism (Fig. 5 ). The Z-score profiles of the four Archaeal species are similar to each other but quite different from those in S. cerevisiae and the six Bacterial species. Although the dark red regions of the Z-score profiles in Archaeal species are quite different in scale, they all seem to differ significantly from the random network preserving the corresponding marginal degree distribution in a similar way ( p -value < 0.1). Looking into the correlation profiles more carefully, we found that the number of statistically significant positive ( K out , K in ) increases in the order of S. cerevisiae , the six Bacterial species and the four Archaeal species. The significant Z-score of certain observation N ( K out , K in ) implies that the chemical reaction between substrates with out-degree K out and products with in-degree K in are statistically significant. We define substrates whose K out >= 2 or products whose K in >= 2 as versatile metabolites. Thus, the above trend implies that the preference to employ reactions involving versatile metabolites increases in the order of S. cerevisiae , the six Bacterial species and the four Archaeal species. Correspondingly, the variety of metabolites decreases in the above order and so does the number of distinct enzymes or variety of enzymes because of the high specific binding of metabolites and enzyme. This is consistent with the biological facts that S. cerevisiae (Eukaryote) encodes a greater variety of enzymes than Bacterial and Archaeal species. Group III measure: Network Motif The network motif is defined to be recurring and non-random building blocks of the network [ 14 , 15 ]. Just like sequence motif, which is an over-represented and biologically meaningful DNA or protein sub-sequence, network motif is an over-represented and biologically meaningful subgraph. Network motif has been shown to be informative of network design principle(s) and network structure. It was found that over 80% of the nodes in the E. coli transcription regulation network are covered by network motifs [ 14 ]. Dobrin et al. [ 16 ] recently discovered that in the E. coli transcriptional regulatory network, "individual motifs aggregate into homologous motif clusters and a supercluster forming the backbone of the network and play a central role in defining its global topological organization." More importantly, network motifs capture the information that is likely to be missed by the correlation profiles because motif actually describes the number of appearances of certain configurations of multiple nodes, and therefore nicely complement with the correlation profiles [ 9 ]. One might argue that there are certain amount of overlaps between the information they capture but the motif profile does not capture the degree information of the connecting nodes, which may be the most powerful feature of the correlation profiles. We searched for all of the 13 three-node subgraphs and all of the 199 four-node subgraphs in the metabolic networks of eleven species. The results showed that the three-node motif profiles found in S. cerevisiae and the six Bacterial species are identical while there is no three-node motif found in any of the four Archaeal networks (Fig. 6 ). Also there is no common four-node motif shared by Archaeal species and S. cerevisiae /Bacterial species while two four-node motifs (id4702, id4950) are shared by the latter ( Additional file 1 ). Among all the 13 possible three-node subgraphs, six of them have one pair of nodes not directly connected. Abundance of such subgraphs will lower the extent of clustering and modularity of the network. As expected, we found that all three-node motifs identified in S. cerevisiae and the six Bacterial species form triangles (Fig. 6 ). It may explain our main finding that metabolic networks in non-Archaeal species are more clustered and modular than those in Archaeal species. Discussion Based on our comparison of network structural properties beyond network indices, we were able to gain more insight into the structural differences across the three domains of life. Having shown that the metabolic network is "scale-free", we further showed that metabolic networks in the four Archaeal species are closer to "exponential random network" [9:Ch2, [ 11 ]] than those in S. cerevisiae and the six Bacterial species. The reasons are the following: First, the Archaeal metabolic networks are visually more homogeneous among themselves compared to their counterparts in the non-Archaeal species. In random networks, any pair of nodes is equally likely to be connected. The network topology should look homogeneous given that the size of network is large enough. The "scale-free" network, on the other hand, features a highly modular and heterogeneous topology since the marginal degree is power-law distributed [ 8 , 9 ]. Moreover, the marginal degree distributions of the metabolic networks in non-Archaeal species fit the power-law model better than Archaeal species (Table 2 and Table 3 ). Second, the average clustering coefficient and average betweenness of Archaeal metabolic networks are much smaller than those in S. cerevisiae and the six Bacterial species. The same is true for the concentrations of three-node and four-node subgraphs. As pointed out by Watts and Strogatz, real-life networks show strong clustering or network transitivity while exponential random network does not [ 12 ]. Third, there is no three-node motif and fewer four-node motifs found in Archaeal metabolic networks compared to non-Archaeal metabolic networks. In particular, the ubiquitous feed-forward loop (FFL) motif found in networks from biology (including metabolic networks in S. cerevisiae and the six Bacteria species in this study) to neurology and engineering fields was not found in any of the four Archaeal metabolic networks (Fig. 6 ). Since motifs are statistically significant subgraphs compared to "properly" randomized network ensembles, no motif or fewer than usual motifs found in a real-life network indicates that the network structure is closer to that of a random network. It has been shown by Milo et al. [ 15 ] that concentration of FFL motif is insensitive to the network size within E. coli transcription regulation network, but diminishes to zero in increasingly larger random networks. This also supports that Archaeal metabolic networks are closer to randomized network ensembles than other real-life networks. The metabolic networks in Archaea are both "random-like" and "scale-free", which might exert profound influences on their adaptability to the hostile environment. Archaeal species are typically restricted to marginal habitats such as hot springs or areas of low oxygen concentration and can assimilate different kinds of inorganic carbon and nitrogen sources. Indeed, the chemical structure and component of the macromolecules such as protein and lipid make significant contributions to the organism's adaptability to the environment. The seemingly ad hoc network organization (both "random-like" and "scale-free") in Archaeal species might also enabled them to survive in those extreme physiological conditions. Archaeal species might employ some biologically significant subgraphs (rather than statistically significant motifs) which can not be detected by current motif searching algorithm [ 15 ]. This makes the Archaeal metabolic networks appear random in statistical sense (not statistically significantly different from random networks) but not in biological sense. Our comparison results showed that many network structural properties measured in Archaeal species are different from those of non-Archaeal species. However, the hidden anthropomorphic factors might account for some of the differences observed. Specifically, the drastic differences of topological profiles between the metabolic networks of Archaeal species and non-Archaeal species may be partially explained by the fact that significantly less extensive metabolic pathway studies have been conducted in Archaeal species [ 32 ]. Robustness of topological profiles against random perturbations can alleviate the impact to a certain extent but is unable to eradicate it [ 9 ]. Conclusions Our network analysis results showed that in most of higher-level (global) topological properties measured, metabolic networks in the four Archaeal species are similar to each other but significantly different from those in S. cerevisiae and the six Bacterial species. This provides further evidence that the metabolic network structures and consequently the design principle(s) in the four Archaeal species are very different from those in S. cerevisiae (Eukaryote) and the six Bacterial species. Our finding that the metabolic networks in Archaeal species possess many properties of the exponential random network begs for better understanding of the design principle(s) in biological networks, which may be revealed by further systematic analyses. For example, locate and align conservative pathways such as glycosis between E. coli or S. cerevisiae and Archaeal species to understand the functional mechanisms of Archaeal metabolic networks. Methods Data source Chemical reaction data was obtained from metabolic database in Ma and Zeng [ 28 ], which consists of five related tables: reaction , enzyme , react , connect and organism . We compiled a new table from this database excluding any inconsistent or redundant connections between metabolites (details below). SQL was used to query the database. Identify and remove inconsistency Inconsistent connections refer to pairs of metabolites that have conflicting reversibility annotation. It is caused by the fact that a pair of metabolites can be in more than one reaction and the reversibility of these reactions can be different. For example, NAD + and Nicotinamide is a pair of metabolites in two reactions: 1) NAD + + L-Arginine = Nicotinamide + N 2 (ADP-D-ribosyl)-L-arginine 2) NAD + + H 2 O -> Nicotinamide + ADPribose. (Note that here the role of NAD + is NOT "current" metabolite, and hence connections established through it should NOT be removed). Reaction 1 is a reversible reaction while reaction 2 is not. We annotated an edge between the two metabolites as long as there was at least one reversible reaction that both of them were involved. For example, the type of connection between NAD + and Nicotinamide is edge (undirected connection). This step could be summarized as "edge ← edge + arc". Identify and remove redundancy There are also numerous redundant connections where the same pair of metabolites switch their roles between substrate and product in two or more different irreversible reactions. For example: 1) UDPglucose + N-Acylsphingosine = UDP + Glucosylceramide 2) Glucosylceramide + H 2 O = D-Glucose + N-Acylsphingosine . ( N-Acylsphingosine and Glucosylceramide is a pair of metabolites that switch their roles in two irreversible reactions). In case of redundancy, we annotated an edge between the pair of metabolites rather than the two arcs because they could be converted to each other through two reactions. This step could be summarized as "edge ← arc + arc". Definitions of some network topological measurements Clustering coefficient (C) We define two kinds of clustering coefficients for each node in the directed metabolic networks, i.e. C in and C out . C in measures the average clustering coefficient of the node representing the product that can be generated from its first-order "nearest neighbors" through chemical reactions. C out measures the average clustering coefficient of the node that generate its first-order "nearest neighbors" through chemical reactions. The larger the coefficients, the more clustered and modular the network appears to be. Betweenness (B) The betweenness for any node n i in the network is defined as , where g jk is the number of shortest paths between node j and node k . g jk ( n i ) is the number of shortest path between node j and node k containing node n i , g is the total number of nodes with edges/arcs. C B ( n i ) needs to be multiplied by two in the case of directed network [ 27 ]. The average betweenness is defined as: . Higher value of betweenness indicates the network is more clustered and modular. Average path length (L) Watts and Strogatz [ 12 ] defined the average path length as , where d ( j , k ) is the shortest path length between node j and node k (distance), V represents the set of all nodes with edges/arcs of the graph, and g is the number of nodes with edges/arcs. Diameter (D) The diameter of the directed graph G is the longest geodesic between any pairs of nodes. The geodesic is the shortest path between a pair of nodes. Pajek [ 33 ] was used to calculate the average betweenness, average path length and diameter. Concentration of subgraphs (S) Wasserman and Katherine [ 27 ] defined the subgraph as follows: A graph G s is a subgraph of G if the set of nodes of G s is a subset of the set of nodes of G , and the set of lines in G s is a subset of the lines in the graph G . Let M be the number of subgraphs, and N be the number of nodes with edges or arcs. Then the "concentration of subgraph" is defined as C = M/N . A high value of C indicates the network is more clustered and modular. Mfinder1.1 [ 15 ] was used to calculate both M and N . Marginal degree distribution calculations The marginal degree distribution of each network is calculated from the Boolean adjacency matrix A , a matrix of 0 or 1. Zero means there is no connection between nodes, and 1 the opposite. The outgoing degree of the node i , k out ( i ) is defined as , where . The incoming degree of the node i , k in ( i ) is defined as . Simple regression analyses of marginal degree distributions The power-law degree model was first log transformed into linear model, i.e. log P ( K i ) = γ log( K i ) + log( α ) + ε i ( i = 1,2,...,n ), γ and α are parameters, ε i is the residual. K i is the degree and P ( K i ) is the corresponding probability. Based on the fitted linear model, we made statistical inference including parameter estimation and individual confidence intervals on the estimates using the Least Square method. Correlation profile calculations Statistically significant correlation profiles were calculated using Matlab code downloaded from Dr. Maslov's website [ 34 ]. The adjacency matrix of the network is the input. Motif profiles calculations According to Milo et al.[ 15 ] , a subgraph is referred to as a motif if the following criteria are met: 1) Its empirical p -value is smaller than a pre-specified threshold, e.g. 0.01. 2) The number of appearances in real networks with distinct sets of nodes is larger than another pre-specified cut-off value, e.g. 4. 3) The number of appearances in real networks is significantly larger than that in randomized networks, i.e. . N real and N rand represent the number of certain subgraphs detected in real-life network and randomized networks, respectively. This is to avoid the situation where some common subgraphs are detected as motifs that have only slight differences in N real and N rand but have a narrow spread of distribution in randomized networks [ 14 , 15 ]. Motif profiles are generated using the Mfinder program. This program and the motif dictionary were downloaded from Dr. Uri Alon group's website [ 35 ]. Authors' contributions DZ and ZSQ conceived and designed the study; DZ wrote the computer code, analyzed the data and draft the manuscript. Both authors read and approved the final manuscript. Supplementary Material Additional File 1 Four-node motifs found in the metabolic networks in different species. The number of connecting nodes for each network is shown. For each motif, the numbers of appearances in real networks ( N real ) and in randomized networks ( N rand ± SD , all values rounded) are shown. The p -values of all motifs are less than 0.01, as determined by comparing to 1000 randomized networks. Each motif occurs at least four times in one network. Other restrictions apply. Motifs were detected and generated using program found in Milo et al. [ 15 ] and the motif dictionary therein. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549204.xml

549562

Surface expression, single-channel analysis and membrane topology of recombinant Chlamydia trachomatis Major Outer Membrane Protein

Background Chlamydial bacteria are obligate intracellular pathogens containing a cysteine-rich porin (Major Outer Membrane Protein, MOMP) with important structural and, in many species, immunity-related roles. MOMP forms extensive disulphide bonds with other chlamydial proteins, and is difficult to purify. Leaderless, recombinant MOMPs expressed in E. coli have yet to be refolded from inclusion bodies, and although leadered MOMP can be expressed in E. coli cells, it often misfolds and aggregates. We aimed to improve the surface expression of correctly folded MOMP to investigate the membrane topology of the protein, and provide a system to display native and modified MOMP epitopes. Results C. trachomatis MOMP was expressed on the surface of E. coli cells (including "porin knockout" cells) after optimizing leader sequence, temperature and medium composition, and the protein was functionally reconstituted at the single-channel level to confirm it was folded correctly. Recombinant MOMP formed oligomers even in the absence of its 9 cysteine residues, and the unmodified protein also formed inter- and intra-subunit disulphide bonds. Its topology was modeled as a (16-stranded) β-barrel, and specific structural predictions were tested by removing each of the four putative surface-exposed loops corresponding to highly immunogenic variable sequence (VS) domains, and one or two of the putative transmembrane strands. The deletion of predicted external loops did not prevent folding and incorporation of MOMP into the E. coli outer membrane, in contrast to the removal of predicted transmembrane strands. Conclusions C. trachomatis MOMP was functionally expressed on the surface of E. coli cells under newly optimized conditions. Tests of its predicted membrane topology were consistent with β-barrel oligomers in which major immunogenic regions are displayed on surface-exposed loops. Functional surface expression, coupled with improved understanding of MOMP's topology, could provide modified antigens for immunological studies and vaccination, including live subunit vaccines, and might be useful to co-express MOMP with other chlamydial membrane proteins.

Background Every Gram-negative bacterium in the order Chlamydiales is an obligate intracellular pathogen [ 1 ]. The organisms are dimorphic, and alternate between free-living, infectious "elementary bodies" (EBs) endocytosed by mucosal cells into vesicular inclusions, and metabolically active, intracellular "reticulate bodies" (RBs). RBs replicate and redifferentiate into EBs before being released to infect neighboring cells, and infections (including Chlamydia muridarum pneumonitis, an important animal model) are often complicated by a damaging immune response and chronic inflammation. Human genital C. trachomatis infections are associated with ectopic pregnancy and infertility, and serovars that target ocular membranes can lead to trachoma and blindness. Chlamydophila pneumoniae ( Ch. pneumoniae ) causes pneumonia in the elderly, and colonization of the placenta by Ch. abortus causes abortion in ewes (and, occasionally, in women). Uniquely among bacteria, the chlamydial outer membrane (OM) is reinforced by a network of disulphide bonds [ 2 ]. Treatment of EBs with Sarkosyl produces "chlamydial OM complexes" (COMCs) [ 3 ] containing three relatively detergent-resistant, cysteine-rich proteins: the Major Outer Membrane Protein (MOMP), encoded by ompA , and OmcB and OmcA, encoded by omp2 and omp3 , respectively. MOMP (~40 kDa) is expressed in both EBs and RBs [ 4 ]. It contains extensive β-sheet secondary structure and forms large pores [ 5 , 6 ], similar to β-barrel porins found in other outer bacterial membranes (e.g. E. coli OmpF). The MOMPs encoded by different C. trachomatis serovars share five well-conserved regions and four "variable sequence" (VS) domains [ 7 , 8 ]. C. trachomatis VS domains, and homologous regions in MOMPs from other species, could correspond to cysteine-rich surface-exposed loops in a porin β-barrel, and EB MOMP is oxidised and highly cross-linked, making the OM very stable. RBs in contrast are osmotically active with reduced, mainly monomeric, MOMP [ 9 ]. MOMP's pore-forming ability is enhanced by reduction [ 5 ], compatible with a link between reversible disulphide bond formation and the developmental stage of the bacteria. Supporting this idea, DTT-reduced EBs tend to resemble RBs [ 5 ], and native MOMP is monomeric when solubilised in SDS under reducing conditions, but forms monomers, dimers, trimers, tetramers and even larger complexes [e.g. [ 6 , 10 , 11 ]] under oxidising conditions. C. trachomatis MOMP is highly immunogenic. Antibodies to the protein neutralised EB infectivity [ 12 ], and triggered approaches to generate MOMP-based vaccines [e.g. [ 13 , 14 ]]. However, as implied earlier, the immunopathology of chlamydial infections is complicated [ 15 ], with T H1 type immune responses as well as specific antibodies (T H2 responses). MOMP is not equally immunogenic in all spp ., and it also stimulates T-cell division, including CD4+ and CD8+ T-cells, enhancing IFN-γ secretion [ 16 ]. C. trachomatis MOMP will probably need to be modified to form a safe and effective subunit vaccine, emphasizing the importance of understanding its structure in more detail. OmcA and OmcB (the other main components of the COMC) are present as approximately 1 OmcB:2 OmcA:5 MOMP [ 17 ]. Ch. psittaci (formerly known as C. psittaci ) OmcA is a 9 kDa lipid-anchored protein with 14 cysteine residues [ 18 ], while OmcB (60 kDa) contains 37 cysteines [ 19 ]. The Omc proteins may not be integral membrane proteins. Reduced OmcB is water-soluble, and although OmcA remains membrane-associated, it can be readily solubilised when reduced [ 20 ], and neither protein was detected on the surface of intact EBs by immunogold labeling [ 21 ]. Regardless of their membrane association, OmcB appears to be extensively cross-linked in the periplasm of EBs, forming disulphide bonds with both MOMP and OmcA. Appropriately, both Omc proteins are expressed late in the developmental cycle (from a bicistronic operon), as RBs are reorganized into EBs [ 22 ], consistent with the idea that RB MOMP is functional and exchanges nutrients and other factors (possibly including signaling molecules) with the host cell. Extensive disulphide cross-linking in EBs may inactivate the porin, and prevent expansion of the growing bacterial cell wall. Although MOMP is of major biological and clinical interest, chlamydia only grow in eukaryotic cells, and MOMP is difficult to isolate and purify because it can aggregate when oxidized, or interact with other cysteine-rich chlamydial proteins. As a result, many groups have expressed recombinant MOMP in E. coli using full-length ompA genes that include the signal sequence to target the translated protein to the OM. Although leadered MOMP can be expressed in a heterologous system [ 23 - 25 ], this approach has proved to be highly problematic, because the protein tends to misfold and aggregate. Koehler et al . [ 26 ] demonstrated surface-exposure, but with a dramatic reduction in cell viability, including OM disruption and substantial cell lysis (i.e. unincorporated, periplasmic MOMP may have been exposed). Jones et al . [ 27 ] co-reconstituted recombinant MOMP with endogenous E. coli porins, and showed altered solute permeabilities in liposome-swelling assays. Although attributed to novel porin activity, this could have reflected modification of endogenous porins. Wyllie et al . [ 28 ] pursued an alternative approach with truncated versions of Ch. abortus and Ch. pneumoniae MOMP, and obtained small amounts of folded proteins without prior denaturation and refolding, sufficient for incorporation into planar bilayers and single-channel recording. Other expression systems, pioneered because of their potential for vaccine delivery, include mammalian COS cells [ 29 ] and Vibrio cholerae [ 30 ]. PorB (37 K), a second putative porin, is also surface-exposed in chlamydia [ 31 ]. Recombinant PorB specifically transported dicarboxylates in liposome-swelling assays [ 32 ], although. it was used with a C-terminal His tag. The terminal residues of porins normally meet to complete a transmembrane β-strand, and may even be linked by a salt bridge. Being integral to the protein fold, additional terminal residues might affect the conformation and, therefore, the specific function of a porin. We expressed PorB as well as MOMP to help determine the factors affecting chlamydial porin expression, but because of these theoretical concerns concerning porin folding and function, we avoided tagged proteins in the present study, and built on previous work with leadered constructs. We developed improved conditions for the surface expression of MOMP in E. coli cells, and demonstrated unambiguously by single-channel recording that recombinant C. trachomatis MOMP folded and formed a functional protein in the absence of many endogenous porins. We showed that MOMP can insert into the outer membrane of E. coli cells and form SDS-sensitive oligomers in the absence of cysteine residues, and generated a "working model" of the topology of MOMP to provide structural hypotheses that could be tested by engineering the recombinant protein. Results Optimised MOMP expression in E. coli cells Our first objective was to obtain properly folded recombinant chlamydial porins in the outer membranes of E. coli cells. Building on previous work (e.g. [ 26 ]), BL21(DE3) cells were transformed with pET- ompA or pET- porB constructs, and expression was induced by 1 mM IPTG at 37°C after growth to an OD 600 of 0.6. Compared to the expression of non-leadered proteins (which accumulate in cytoplasmic inclusion bodies), cells expressing leadered porins must transport the immature full-length porin across the inner membrane, cleave the leader sequence in the periplasmic space, and fold and insert the mature protein into the OM. Expression of mature, leaderless C. trachomatis MOMP did not inhibit growth compared to non-transformed cells, in contrast to substantial inhibition with full length MOMP (Fig. 1A ). To investigate whether different leader sequences could improve processing, C. trachomatis MOMP was expressed with the OmpT leader rather than its native leader. Initial growth rates were comparable to those shown by non-transformed cells, and similar to cells expressing mature MOMP (i.e. MOMP without a leader sequence), although the cultures again showed a reduced final cell density. We next investigated the expression of MOMPs from other chlamydial spp . to determine whether the observed effects were specific to C. trachomatis MOMP, and we also expressed C. muridarum PorB to exclude a universal problem with the expression of all putative chlamydial porins in E. coli . The constructs had different effects on cell viability (Fig. 1B ). Bacteria expressing C. muridarum MOMP grew more slowly than bacteria expressing C. trachomatis MOMP, although the bacteria continued to grow slowly throughout the entire period of induction. The growth of bacteria expressing Ch. abortus MOMP or C. muridarum PorB was markedly reduced, and the density decreased after 30 min. The "recovery" at later stages reflected multiplication of non-expressing cells in the presence of β-lactamase released from dead or dying cells (growth ceased on fresh Ampicillin plates, data not shown). We then changed the leader sequences. The growth of cells expressing C. trachomatis MOMP and C. muridarum PorB was improved by replacing the native chlamydial leader with the E. coli OmpT leader, and the decrease in optical density occurred later in the induction and continued more slowly. In contrast, no significant improvement was seen when Ch. abortus MOMP was expressed with the OmpT leader (data not shown). We also expressed full-length constructs in E. coli BL21(DE3)omp8 cells lacking expression of the endogenous porins LamB, OmpA, OmpC and OmpF [ 34 ]. Toxicity was more pronounced than in unmodified BL21 cells, and after establishing conditions for detergent extraction of recombinant MOMP ( Additional Data File #1 ), expression conditions were further optimised to improve the yield of processed, recombinant protein. Native and OmpT-leadered C. trachomatis MOMP constructs were induced rapidly at 37°C with 1 mM IPTG or slowly at 16°C with 0.1 mM IPTG (Fig. 2 ). At 37°C both versions of MOMP were expressed, and by 4 hours about half the protein was processed, as shown by the doublet band of OM-associated MOMP with and without its signal sequence (Fig. 2 ). The ~2 kDa difference between the cleaved and non-cleaved protein bands (38 kDa and 40 kDa, respectively), is similar to the difference seen when leadered versions of E. coli OmpF are expressed). There was a slight decrease in total protein when MOMP was expressed with its native leader at 16°C, but the proportion of processed protein was unchanged. Although protein decreased following slow induction of MOMP containing the OmpT leader, most of the protein was processed. Based on these observations, slow induction of native-leadered MOMP was carried out in different growth media for prolonged periods. After growing for 6 hours, cultures in LB medium plateaued at an OD 600 ~0.85, after which the cells began to lyse. In contrast, cells cultured in more supportive SOC medium continued to grow steadily, and began to plateau about 12 hours after induction ( Additional Data File #2 ). Processing and surface expression of mutagenised and engineered MOMPs Given the known difficulties associated with protein misfolding and aggregation (e.g. [ 23 - 26 ]), a particular problem for chlamydial MOMPs compared to other bacterial porins, our next objective was to determine whether MOMP was actually inserted into the E. coli outer membrane. Although recombinant MOMP was associated with the OM fraction following subcellular fractionation, the observation that its leader sequence was not always cleaved (Fig. 2 ) suggested that some leadered protein co-fractionated with OMs, possibly as a peripheral membrane protein. This raised the possibility that even cleaved recombinant proteins might not be fully integrated into the OM. To determine whether processed MOMP was actually inserted into (and across) the OM, we carried out whole cell immunoblots to probe for the presence of MOMP epitopes on the surface of intact E. coli BL21 cells. Because of the importance of reduced temperature (Fig. 2 ), we carried out inductions for whole cell immunoblotting at 37°C, 16°C and an intermediate temperature of 25°C. MOMP was incorporated into the OM at both 25°C and 16°C, when induced in the presence of either 1 mM or 0.1 mM IPTG, respectively. Expression and processing were more rapid at 25°C, and because the presence of some unprocessed protein was irrelevant in this experiment, we induced the cells at 25°C for 2 hrs. Non-transformed BL21 cells, or cells transformed with an empty plasmid, and BL21 cells transformed with constructs encoding mature, leaderless C. trachomatis MOMP, or with OmpT-leadered MOMP and native leadered-MOMP, were applied to a nitrocellulose membrane (avoiding methanol-activated PVDF, and the risk of OM permeabilisation and exposure of periplasmic MOMP), and probed with anti-MOMP pAb (Fig. 3A ). The absence of a signal from control cells and cells expressing MOMP in its non-leadered, mature form confirmed the incubation and blotting conditions did not cause cell lysis and expose unincorporated protein. Both OmpT- and native-leadered MOMP were detected on the cell surface (Fig. 3A , whole cell blots), confirming they were inserted into the OM. Unfortunately, BL21omp8 cells were too fragile to survive the same blotting procedure. SDS-PAGE analysis of OG-solubilised OM fractions (Fig. 3A , middle panel) confirmed MOMP expression and processing, although parallel immunoblots (Fig. 3A , lower panel) showed faint additional bands of ~40 kDa for the leadered proteins, indicating that processing was incomplete, as expected. Parallel immunofluorescence data (Fig. 3B ) showed MOMP was confined to cytoplasmic inclusion bodies containing the mature protein when the appropriate cells were fixed and permeabilised before staining (Fig. 3B , panel b). As expected, staining was absent when the antibody was applied before permeabilisation (data not shown). However, OM staining was seen for MOMP expressed with both the OmpT leader and the native leader (panels c and e, respectively). When these cells were permeabilised before staining (panels d and f, respectively), immunoreactive protein was also noted internally, as expected (e.g. Fig. 3A , lower panel), although reduced or absent in BL21omp8 cells induced for 12 hrs at 16°C in more supportive SOC medium (Fig. 3B , inset in panel d). We concluded that MOMP constructs encoding appropriate leaders could be expressed in E. coli , cross the inner membrane, and be processed in the periplasm. Furthermore, under modified incubation and induction conditions (especially at reduced temperatures, and in the relatively supportive medium SOC), MOMP could be folded and incorporated into the outer membrane. Membrane topology of MOMP Having confirmed that C. trachomatis MOMP was inserted into the OM of E. coli cells, we set out to investigate how the protein was organized in the membrane. While noting that predictive algorithms must always be deployed with care, and with reference to established findings for a given protein, we first analyzed MOMP's primary sequence for membrane crossings using a neural network trained with OM proteins of known structure [ 36 ]. The analysis (Fig. 4A ) showed 16 membrane crossings. As expected, the VS domains of C. trachomatis MOMP generally corresponded to regions of the protein predicted to be extracellular. We then reanalyzed the sequence using two β-strand prediction programs (Fig. 4B ). The combined analysis revealed a total of 16 strands, corresponding numerically to the initial "membrane crossing" prediction (which does not on its own appear to be sufficient to identify the specific extramembrane domains). We discarded the strand coinciding with VS1 in B2TMPRED (see Methods) because VS domains are likely to be extracellular loops, and inserted an extra strand between G210 and S218 to bring the chain back across the membrane, so that all 4 VS domains remained external. Minor adjustments were made to accommodate known constraints on β-strand organization and porin structures [ 39 , 40 ]. The final working model (Fig. 5 ) provided testable hypotheses concerning the pattern of transmembrane folding. All the cysteine residues were predicted to be accessible for inter- or intrasubunit disulphide bond formation or cross-linking with other proteins. Most were predicted to be external, but two were periplasmic. Although one thiol group was in a predicted transmembrane domain, it faced the central water-filled pore rather than the lipid bilayer, where it could potentially interact with a cysteine thiol on a pore-confined loop. We designed four C. trachomatis MOMP constructs (with intact cysteines and native leaders, to correspond exactly in these respects to the "wild-type" protein) in which substantial regions of VS domains 1, 2, 3 or 4 (shown in Fig. 6A ) were deleted, to test the prediction that these domains are surface-exposed loops that can be shortened without compromising the main β-barrel fold and membrane insertion. The region removed from VS1 was G63 to Y87; from VS2, E141 to F156; from VS3, Y220 to G238; and from VS4, D278 to T318. Our strategy (see Methods) resulted in some mutations. Most were conservative changes (M62T in VS1, T239V in VS3 and A277V in VS4), apart from G219D in VS3. However, our topology prediction placed this residue in an external loop, where the additional charge was unlikely to be significant. We also generated another pair of constructs with deletions of either one or two of the predicted β-strands between VS domain 1 and VS domain 2 (summarized in Fig. 6B–C ), in an attempt to disrupt the formation of OM-inserting β-barrels. These constructs were designated: Δβ5, with removal of E95 to F111 (with no residue changes) and Δβ5,6, with removal of F97 to A129 (with 2 changes, E95D and M96V). Before expressing the cDNAs encoding putative loop or strand deletions, we re-examined the expression and OM insertion of full-length C. trachomatis MOMP using a construct in which all 9 cysteine residues (Fig. 4A , circles; Fig. 5 , shaded residues) were replaced by alanine. The results (Fig. 7A ) were similar to those for the non-mutagenised protein, showing that folding and membrane insertion could proceed without cysteine residues and without the controlled formation of disulphide bonds (as it may do in RBs). We then expressed each of the "loop-deleted" MOMP proteins in BL21 cells. All four were detected on the cell surface (Fig. 7B ), demonstrating incorporation into the OM. In contrast, recombinant proteins expressed from constructs with putative β-stand deletions were not detectable on the surface of E. coli cells (Fig. 7 ). We considered the unusual possibility that all the epitopes in the "strand-deleted" proteins might have been unreactive in the E. coli membrane, due to masking or oligomerisation, but suspension of the cells in Tris (rather than phosphate) buffer (100 mM NaCl, 50 mM Tris-HCl, pH 7.4), or the addition of 2 mM EDTA, failed to "unmask" any immunoreactivity ( Additional Data File #3 ). MOMP forms oligomers in the E. coli outer membrane Native MOMPs are difficult to purify free from other chlamydial proteins [ 6 ], precluding firm conclusions about native subunit structure, especially in the absence of protein (cysteine) oxidation. In preliminary investigations of the subunit organisation of recombinant MOMP, we noted that the recombinant protein did not form SDS-resistant oligomers ( Additional Data File #4 ). However, unlike trimeric E. coli porins [ 40 ], oligomers of isolated MOMP, away from their normal membrane environment [ 6 ], may be unstable in the presence of SDS, so we subjected detergent-solubilised OM extracts to large-scale non-denaturing GE chromatography in milder detergents. For these and all subsequent experiments, MOMP was expressed in BL21omp8 cells with the OmpT leader (in SOC medium, at 16°C), to exclude heterooligomers containing endogenous E. coli porins, and minimize uninserted periplasmic protein, respectively. We carried out GE chromatography in LDAO or Zwittergent 3–14 (having previously noted these to be cheaper but equally effective detergents to replace OG, Additional Data File #1 ), with excess (5 mM) DTT in the presence of MOMP cysteine residues (calibrating the column in the presence of detergent). Under these conditions, MOMP appeared to form oligomers containing 2–4 subunits, although some recombinant MOMP always formed higher-order oligomers (Fig. 8 ). Similar results were obtained after repeating each experiment at least twice. The apparent subunit stoichiometry of recombinant MOMP depended on the detergent, with putative dimers in LDAO, and trimers or tetramers in Zwittergent 3–14, depending on the presence or absence of cysteine residues, respectively. However, it should be emphasized that only the major quaternary species was identified in each case. The presence of oligomers in LDAO or Zwittergent 3–14 contrasted with the absence of SDS-resistant oligomers during SDS-PAGE, and oligomer formation even in the absence of cysteine residues argued against an essential role for disulphide bonds. We also investigated the subunit organization of MOMP by covalent cross-linking following expression and insertion into BL21omp8 OMs, by removing DTT to allow in situ cysteine oxidation by dissolved oxygen. OM proteins were then incubated in SDS sample buffer with or without reducing agent at room temperature for 10 mins, separated by SDS-PAGE, and detected by Western blotting (Fig. 9 ). Reduced MOMP appeared as a single band of ~38 kDa, but non-reduced MOMP occupied several distinct bands. SDS-denatured, monomeric MOMP appeared as a band of ~38 kDa (labeled "denatured monomer"), corresponding to the reduced sample. However, monomeric MOMP also formed a band of ~35 kDa, running "ahead" of its normal apparent molecular mass, as previously seen with "folded" porin monomers [ 41 , 42 ]. Additional, fainter bands at higher molecular masses corresponded to dimers, tetramers and possible trimers (~80 kDa, ~160 kDa and ~120 kDa, respectively), similar to the findings following GE chromatography, with an upper band of aggregated protein that failed to enter the gel. Surface-expressed MOMP is functional Fully processed and correctly folded MOMP should function as a porin-like ion channel [ 6 ]. We tested this crucial prediction by expressing "wild-type" full-length recombinant C. trachomatis MOMP in BL21omp8 cells which express only a small subset of native E. coli porins, and not OmpF or OmpC [ 34 ]. We then functionally reconstituted solubilised BL21omp8 OM protein GE fractions in voltage-clamped planar lipid bilayers. Fractions containing "oligomeric" MOMP complexes gave rise to large-conductance, porin-like ion channels (Fig. 10 ). Similar channels were recorded irrespective of whether the detergent was LDAO or Zwittergent 3–14 (using fractions corresponding to 195 ml or 180 ml, respectively). The channels were voltage-dependent, closing at relatively high holding potentials (e.g. + or - 100 mV), but remaining open around 0 mV. The single-channel conductance in symmetric 500 mM KCl was 480 ± 19 pS (mean ± SEM, n = 6 independent experiments), and the reversal potential in 500 mM vs 50 mM KCl ( cis vs trans ) was -31 ± 1.5 mV (mean ± SEM, n = 9 independent experiments). This corresponded to a relative cation vs anion selectivity of 3.8 under these specific ionic conditions. Control preparations (detailed under Methods), including membrane proteins from control BL21omp8 cells subjected to the same experimental conditions, where OM proteins were solubilised and subjected to GE chromatography in exactly the same way, did not give rise to similar channel activity (6 experiments). Discussion Functional reconstitution of recombinant C. trachomatis MOMP at the single-channel (single molecule) level from cells lacking many endogenous porins provides very strong evidence that MOMP adopted its native fold when expressed in E. coli under suitable conditions. Although a leadered version of recombinant chlamydial MOMP was expressed and functionally analysed previously [ 27 ], membranes containing the protein were co-reconstituted with endogenous E. coli porins for liposome-swelling studies. Although MOMP may have contributed additional porin-like activity, functional modification of endogenous porins could not be ruled out. Interestingly, the successful expression and processing of recombinant chlamydial porins in E. coli cells depends on the precise leader sequence, as well as on the specific protein. PorB is less "toxic" with its native leader, in contrast to MOMP, which is less "toxic" with the E. coli OmpT leader, and native-leadered C. muridarum MOMP is less deleterious to E. coli than Ch. abortus MOMP. Although a full investigation of the role of leader sequences could not be undertaken here, it is known that successful OM insertion, as well as prior transport across the inner membrane and processing, is also signal sequence-dependent. For example, a large proportion of E. coli LamB porins with signal sequence mutations remained "tethered" to the inner membrane (probably by their unprocessed signal sequence), even though the protein was also closely associated with the OM [ 43 ]. For C. trachomatis MOMP, use of the Omp-T leader and induction at 16°C (not induction at 37°C, as previously employed), in either "wild-type" cells or "porin knockout" cells in a supportive medium (SOC), provides improved processing and OM insertion, and there is also significant insertion at 25°C in "wild-type" E. coli . The single-channel properties of C. trachomatis MOMP are consistent with previous data on bacterial [ 40 ] and putative chlamydial [ 6 ] porins. In particular, the channels show "bell-shaped" voltage-dependent gating and are mainly open around ~0 mV, with very high conductances (close to the saturating conductances predicted for a large water-filled pore) and poor ionic selectivity, showing only a slight preference (~4:1) for cations over anions (using a Nernst-Plank analysis because relatively wide, water-filled porin channels are probably electroneutral [ 35 ], and poorly-described by electrodiffusion theory). The channels often appeared in groups of three, as might be expected for a trimeric "triple-barrelled" porin (e.g. Fig. 10 ). However, unless the channels were randomly incorporated into the bilayer (which is difficult to demonstrate), these complexes may represent a selected sub-population. Despite the lack of sequence similarity to known bacterial porins, a combination of different predictive approaches (none of which was entirely satisfactory in isolation), set in the context of elegant and pioneering work from many laboratories on the properties of VS domains, predicted that C. trachomatis MOMP, like putative porins in the intracellular pathogens Burkholderia thailandensis and B. pseudomallei [ 44 ], could be a 16-stranded β-barrel. Our working model pays due attention to the construction principles for β-barrels [ 39 , 40 ]. The N and C termini complete final strand 16, the periplasmic turns are short, and most external loops are long and include the immunogenic VS domains. The barrel surface in contact with the bilayer consists largely (though not exclusively) of hydrophobic side chains, and all 18 strand residues with charged side chains project into the pore to line the central water-filled central channel. 6 cysteines lie in extracellular loops, and 2 periplasmic cysteines lie on opposite sides of the barrel where they are unlikely to form an intrasubunit disulphide bond, although they could form intersubunit bonds, or bonds with other proteins. A single membrane thiol projects into the barrel pore, where it could be involved in disulphide bond formation if a loop (e.g. L1) were to fold into the barrel. Our working model for the membrane topology of C. trachomatis MOMP differs in some significant respects from the recent prediction for C. muridarum MOMP [ 45 ] (which was based partly on hydrophobicity plots). Although both studies predict that MOMPs are 16-stranded β-barrels with an average strand length of ~8 residues, periplasmic thiols are absent from the C. muridarum prediction. This would preclude the significant interactions with OmcB and OmcA, described in the Background. We also assigned L2, 4, 6 and 7 as C. trachomatis VS domains, not L2, 3, 5 and 6, the homologous regions in C. muridarum MOMP. Experimental tests of the predicted membrane topology of C. trachomatis MOMP are consistent with our model, because individual VS domains can be substantially truncated without preventing incorporation of the protein into the bacterial OM. If MOMP is a β-barrel porin, as suggested, and VS domains are confined to specific extracellular loops, it is conceivable that MOMP can continue to fold into a β-barrel in the absence of one of these domains. On the other hand, the removal of β-strands would disrupt folding. Removal of a single strand, bringing periplasmic residues into direct contact with external residues, is predicted to be particularly destructive to the global fold. Removal of more than one strand might be better tolerated, provided the β-barrel can form with a significantly reduced diameter. In practice, it appears that C. trachomatis MOMP cannot accommodate either type of strand modification. GE chromatography suggested that MOMP forms oligomers in the presence of Zwittergent or LDAO, and in line with these findings, in situ cysteine cross-linking of recombinant MOMP in E. coli OMs revealed oligomeric MOMP complexes, together with a species of folded or partially-folded MOMP monomers containing at least one intramolecular disulphide bond. This species contrasts with reduced, denatured MOMP monomers seen when chlamydial MOMP is solubilised directly from OMs (or native EBs [ 6 ]). However, the exact stoichiometry of MOMP oligomers in the E. coli OM remains uncertain because our size estimates for the oligomers, and thus their stoichiometries, may be too high because of uncorrected bound detergent. Also, it is clear that the stability of MOMP oligomers is detergent-dependent. Native Ch. abortus MOMP forms SDS-resistant oligomers of ~100 K [ 6 ], unlike the SDS-unstable MOMP oligomers isolated from E. coli OMs. We speculate that this may be because native MOMP oligomers are stabilised by interactions with other chlamydial components (e.g. co-purified Omp90 [ 6 ]), and possibly also by disulphide bonds. Disulphide bond formation (whether transient or permanent) does not appear to be essential during protein folding and OM insertion, because a cysteine-free mutant can be fully processed (Fig. 7A ) and can also form oligomers. Overall, our results show that the subunit stoichiometry of detergent-solubilised MOMPs expressed and processed in E. coli is detergent-dependent, that MOMP subunits can be cross-linked by disulphide bridges, and that folded monomers contain at least one intrasubunit disulphide bond (Fig. 8 ). Conclusions C. trachomatis MOMP, an immunodominant, cysteine-rich, chlamydial surface protein of crucial importance in the immune response to infection, is a major subunit vaccine target. However, unlike many other bacterial porins, it has been difficult to refold from inclusion bodies or to achieve and demonstrate functional surface expression. This study is the first to report unambiguous functional analysis, by single-channel recording, of recombinant chlamydial MOMP recovered from bacterial outer membranes. The modified expression system described in the present study provided a means to test specific hypotheses provided by a working model for the C. trachomatis protein. However, although our results are consistent with a working model of MOMP as a 16-stranded β-barrel, more mutations or other approaches are needed before a specific model can be accepted. The protein also formed oligomers, even in the complete absence of cysteine residues. The surface display of modified, functional MOMP in E. coli cells (potential vehicles for a live, subunit vaccine), together with a working topological model, could guide the removal of unwanted or harmful epitopes from engineered proteins, and it might also be possible to display external loops containing specific MOMP epitopes on other porin "scaffolds" in living cells. However, it is important to note that such approaches will be limited if essential disulphide bonds in the native chlamydial envelope, including bonds involving non-MOMP cysteines, stabilise the conformation of key immunogenic VS domains. Methods DNA manipulations C. trachomatis ompA (corresponding to X62918, from the Da serovar) and Ch. abortus ompA were cloned without their leader sequences into the Nde -I/ Nco -I sites of pET22b(+) (Novagen) after destroying an internal Nde -I site in C. trachomatis ompA by Quik-Change PCR mutagenesis (Stratagene). This did not alter the encoded protein. C. muridarum ompA and porB were amplified with and without their leaders from genomic DNA and cloned into the Nde -I/ Bam- HI sites and Nde -I/ Nco -I sites, respectively, of the same vector ( C. muridarum ompA also required null mutation removal of an internal Nde -I site). The E. coli OmpT protease leader sequence or the native C. trachomatis MOMP leader sequence was added to the 5' end of the leaderless C. trachomatis and Ch. abortus inserts by sequential gene extension PCR using three overlapping primers. A 5' Nde -I site was again used to provide the starting methionine codon in the final full-length construct. Quik-Change PCR was also used to create pairs of unique internal restriction sites in native-leadered C. trachomatis ompA to permit the deletion of specific domains by plasmid restriction and religation [ 33 ]. These sites were: for VS1, Age -I; for VS2, Bcl -I; and for VS3, VS4 and the predicted β-strands, Aat- II. Successful deletions were confirmed by hemi-nested single-colony PCR (using Taq polymerase) to identify clones that could be amplified by gene-spanning primers but not by primers complementary to regions that had been removed. We also generated C. trachomatis MOMP expression constructs containing inserts in which all 9 cysteine residues (C26, C29, C33, C102, C115, C182, C184, C207 and C335) were replaced by alanine using Quik-Change PCR. Most of the modifications were carried out in a pSTBlue-I/NovaBlue system, and the fidelity of each insert was confirmed by automated DNA sequencing (MWG Biotech). Protein expression and recovery E. coli BL21(DE3) or BL21(DE3)omp8 [ 34 ] cells were harvested from cultures of LB (Luria-Bertani) medium (10 g/l Bacto tryptone, 5 g/l yeast extract, 10 g/l NaCl, pH 7.0) or SOC medium (20 g/l Bacto tryptone, 5 g/l yeast extract, 0.5 g/l NaCl, 20 mM glucose, pH 7.0) by centrifugation at 6,000 × g for 5 mins after inductions as described in the Results section, and washed in 50 ml phosphate buffered saline (PBS). The cell pellet was resuspended in 5 ml TEN buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 100 mM NaCl) containing 1 mg lysozyme and incubated for 30 min. at room temperature. Following sonication (6 × 15 s, 6 μm amplitude, Sanyo Soniprep 150 sonicator) the cell lysate was incubated with 20 U/ml Benzonase (Novagen) for 15 min at room temperature. OM fragments were pelleted by centrifugation at 15,000 × g for 10 min, and washed twice in 20 ml TEN buffer. Membrane proteins were solubilised by resuspending the pellet in 6 ml solubilisation buffer containing 50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 50 mM NaCl and 10 mM DTT with either 1% (w/v) octyl glucoside (OG, Anatrace), 1% (w/v) lauryl (dodecyl) dimethylamine oxide (L(D)DAO, Anatrace) or 1% (w/v) Zwittergent 3–14 (Anzergent 3–14, Anatrace), and incubating at 37°C for 1 hour. The solution was clarified by ultracentrifugation (Beckman TLA-100) for 20 mins at 100,000 rpm. Protein concentrations were determined after TCA precipitation. SDS-PAGE and Western blotting Unless otherwise indicated, SDS-PAGE was carried out under reducing conditions using 10–12% (w/v) gels. Molecular masses were estimated from plots of relative mobility vs the logarithm of the molecular mass of Precision Plus unstained protein markers (BioRad). For Western blotting, proteins were electrophoretically transferred to PVDF membranes under conditions compatible with the transfer of high-MW proteins including native MOMP oligomers [ 6 ]. The membranes were blocked in 5% (w/v) non-fat milk in PBS-T (0.005% (v/v) Tween-20 in PBS) then incubated in 1:5000 goat anti- C. trachomatis MOMP antibody (Fitzgerald International) for 1 hour at room temperature. Following 2 × 30 sec and 3 × 5 min washes in PBS-T, membranes were incubated in 1:10,000 HRP-conjugated anti-goat/sheep antibody (Sigma) for 1 hr at room temperature. After washing, immunoreactive proteins were detected by ECL. Whole cell immunoblotting and immunofluorescence 10 ml of LB medium was seeded 1:100 with cultures grown to saturation overnight, and incubated until the OD reached 0.6. The cells were pelleted by centrifugation (6,000 g × 10 mins) and resuspended in fresh medium. Following incubation at the selected temperature for 10 mins, 0.1–1 mM IPTG was added and incubation was continued for another 2–16 hrs. Intact cells were harvested by gentle centrifugation (4,500 g × 5 mins) and washed in 1 ml PBS. The pellets were resuspended in 200 μl PBS, and 10 μl was applied to a nitrocellulose membrane and allowed to dry. The membrane was blocked and probed with anti- C. trachomatis MOMP polyclonal antibody as described above. Immunofluorescence was carried out as described previously [ 26 ], with fixation and permeabilisation either before or after immunostaining, using 1:200 dilutions of the above primary antibody and fluorescein-conjugated anti-goat secondary antibody (Sigma). The cells were then observed by bright field, phase contrast and fluorescence microscopy using a Leica TCS-NT confocal microscope. Gel-exclusion chromatography Solublised OM proteins were separated by GE chromatography using a high resolution 26/60 HiLoad Superdex 200 prep grade column (Amersham Pharmacia Biotech) freshly equilibrated in 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 50 mM NaCl and 5 mM DTT (omitting the latter for the cysteine-less MOMP mutant). The buffer also contained either 0.05% (w/v) LDAO or 0.05% (w/v) Zwittergent 3–14. 2 ml aliquots of solubilised OM proteins (containing up to 10 mg protein, solubilised as described earlier under protein expression) were loaded, and the column was eluted with the same buffer for 800 min. at a flow rate of 0.5 ml/min. 5 ml fractions were collected and 10 μl of each protein-containing fraction was deposited onto a pre-prepared PVDF membrane and probed for MOMP as described earlier under Western blotting. The column ( V t 320 ml) was calibrated in the presence of detergent using standard proteins. V 0 (the void volume) was 115 ml, and K av was calculated as ( V e - V 0 )/( V t - V 0 ), where V e is the elution volume. Bilayer reconstitution and single-channel analysis Planar bilayers were cast from diphytanoyl phosphatidylcholine (Avanti) between two 0.5 ml chambers containing 50 mM KCl, 20 mM Tris-HCl (pH 8.0) and 1 mM DTT, designated cis and trans [ 6 ]. The cis chamber was voltage clamped with respect to the trans chamber using an Axon 200B amplifier or a Biologic RK300 amplifier. 1–5 μl aliquots of pre-diluted solubilised proteins (containing up to 10 ng protein and no more than 5 ng detergent) were added to the cis chamber, followed by aliquots of 5 M KCl to raise the salt concentration to 500 mM. Channel incorporation usually occurred within 30 min, accelerated by switching the holding potential between +/- 60 mV. Experimental protocols were programmed and the digitised data were low-pass filtered (1 kHz, 8-pole Bessel-type response) and recorded using pClamp8 software (Axon Instruments), and analysed offline. The bilayer potential was slowly and repeatedly ramped between -100 mV and +100 mV (each sweep taking 32 s) in the presence of an asymmetric (500 mM vs 50 mM, cis vs trans ) gradient of KCl, or with equimolar 500 mM or 1 M KCl. At least 3 voltage ramps were recorded and analysed for each experiment, and equilibrium recordings were obtained at defined holding potentials. Holding potentials refer to the cis chamber, and upgoing deflections represent net movement of cations from cis to trans or of anions from trans to cis . Relative ionic permeabilities were determined from the equilibrium solution of the Nernst-Planck flux equations [ 35 ]. When the cation and anion fluxes are equal: E r is the equilibrium (zero current, or reversal) potential in asymmetric KCl, and R, T and F have their usual significance. (the permeability ratio of K + to Cl - ) was calculated using appropriate activity coefficients ( a ) from standard tables. Control experiments were carried out using equivalent amounts of detergent and with equivalent amounts of solubilised OM proteins purified from non-transformed bacteria, selecting identical GE column fractions. Membrane topology prediction The number of membrane crossings was predicted using a neural network based-outer membrane protein topology prediction program trained with known porins [ 36 ]. We discounted a predicted membrane crossing very near the N-terminus that was only apparent after numerical rounding. β-strands were predicted independently by similar computational approaches using B2TMPRED [ 37 ] and TMBETA [ 38 ], respectively. The three predictions were combined and adjusted manually, taking account of the accessibility of the VS domains of C. trachomatis MOMP and the known characteristics of antiparallel, amphipathic β-barrel strands in porins. Authors' contributions HEF and HM carried out most of the experiments. HEF analysed and organised the data, and drafted the first version of the manuscript. RHA conceived the overall project, provided experimental guidance, carried out some of the experiments, and redrafted the manuscript. All the authors read and approved the final manuscript. Supplementary Material Additional File 1 Detergent extraction of recombinant MOMP. Immunoblot with ECL detection following SDS-PAGE of OM proteins (10 μg per lane) from BL21 cells expressing OmpT-leadered C. trachomatis MOMP, induced for 2 hrs at 37°C. OM proteins (see Methods) were solubilised in 1% (w/v or v/v) octylglucoside, Triton X-100, Zwittergent 3–14 or LDAO, as indicated. NB & B are non-boiled and boiled samples, respectively. NE = non-expressing (control) cells. Click here for file Additional File 2 Optimisation of C. trachomatis MOMP expression and processing in BL21omp8 cells. Growth curves of BL21omp8 cells expressing C. trachomatis MOMP with its native leader, in LB or SOC medium (means ± SEM, n = 4). Click here for file Additional File 3 MOMP epitopes are not unmasked by Tris buffer or EDTA in whole cell immunoblots. Control BL21 cells and cells expressing "strand-deleted" constructs were suspended in 100 mM NaCl containing 50 mM Tris-HCl (pH 7.4) with or without 2 mM EDTA, applied to nitrocellulose membranes, and probed with anti- C. trachomatis MOMP polyclonal antibody. Click here for file Additional File 4 Recombinant MOMP does not form SDS-resistant oligomers. SDS-PAGE and immunoblot analysis of C. trachomatis MOMP expressed with its native leader in BL21omp8 cells at 16°C (induced for 12 hrs in the presence of 0.1 mM IPTG). Lanes 1 & 2 contain 10 μg non-boiled and boiled OM proteins, respectively, solubilised in 1% (w/v) OG. Note successful transfer of high-MW proteins. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549562.xml

526370

Inhibitory effects of proanthocyanidins from Ribes nigrum leaves on carrageenin acute inflammatory reactions induced in rats

Background The anti-inflammatory effects of proanthocyanidins (PACs), isolated from blackcurrant ( Ribes nigrum L.) leaves, were analysed using carrageenin-induced paw oedema and carrageenin-induced pleurisy in rats. Results Pretreatment of the animals with PACs (10, 30, 60 and 100 mg/kg, i.p.) reduced paw oedema induced by carrageenin in a dose and time-dependent manner. PACs also inhibited dose-dependently carrageenin-induced pleurisy in rats. They reduced (A) lung injury, (B) pleural exudate formation, (C) polymorphonuclear cell infiltration, (D) pleural exudate levels of TNF-α, IL-1β and CINC-1 but did not affect IL-6 and IL-10 levels. They reduced (E) pleural exudate levels of nitrite/nitrate (NOx). In indomethacin treated rats, the volume of pleural exudate was low, its content in leukocytes and its contents in TNF-α, IL-1β, IL-6 and IL-10 but not in NOx were reduced. These data suggest that the anti-inflammatory properties of PACs are achieved through a different pattern from those of indomethacin. Conclusion These results suggest that the main mechanism of the anti-inflammatory effect of PACs mainly lies in an interference with the migration of the leukocytes. Moreover, PACs inhibited in vivo nitric oxide release.

Background Proanthocyanidins are compounds, naturally occurring in various plants, with anti-inflammatory [ 1 , 2 ] and anti-arthritic activities [ 3 ]. They are reported to prevent skin aging and heart diseases, they scavenge oxygen free radicals and inhibit UV radiation-induced peroxidation [ 4 - 10 ]. We have isolated prodelphinidins and procyanidins, proanthocyanidins (PACs) from blackcurrant ( Ribes nigrum L., Grossulariaceae) leaves which are used in European traditional medicine for the treatment of inflammatory disorders such as rheumatic diseases [ 11 ]. Majority of these compounds are water soluble monomers and oligomers (2 to 3 units) consisting of flavan 3-ol monomer units linked together by mostly C-4 to C-8 (Figure 1 ) and to a lesser extent C-4 to C-6 bindings. Few tetramers are also found. Figure 1 Chemical structure of proanthocyanidins. Where R = H, it is a procyanidin: catechin (R 1 = H and R 2 = OH) and epicatechin (R 1 = OH and R 2 = H); Where R = OH, it is a prodelphinidin: gallocatechin (R 1 = H and R 2 = OH) and epigallocatechin (R 1 = OH and R 2 = H). Previously, we have observed that, in vitro, these compounds profoundly affect the metabolism of chondrocytes : they increase the secretion from these cells of type II collagen and proteoglycans while they decrease the secretion of prostaglandin E2 (PGE2) [ 12 ]. On the other hand, while these compounds inhibited purified cyclo-oxygenase-1 and cyclo-oxygenase-2, they did not reduce the release of thromboxane B2 and PGE2 from human in vitro stimulated platelets and neutrophils respectively [ 12 ]. Moreover, PACs might influence the contractile status of smooth muscles of blood vessels : intravenous and intraperitoneal injection of PACs induced a drop of the blood pressure without a significant bradycardia [ 13 ]. This effect counteracts the hypertensive activity of norepinephrine. The present studies were designed to evaluate the potential anti-inflammatory activities of these compounds, in vivo , on carrageenin-induced paw oedema and pleurisy in rats. This latter inflammatory reaction allowed us to examine the influence of PACs not only on the exudate volume and polymorphonuclear cell accumulation but also on the release of several cytokines, IL-1β, TNF-α, IL-6, IL-10, CINC-1 and of nitric oxide (NO). These cytokines and NO are among the more important mediators involved in inflammatory processes [ 14 - 16 ]. Results Influence of PACs on rat paw oedema Carrageenin-induced oedema was significantly inhibited by PACs dose-dependently (Figure 2 ). This inhibitory effect was efficient from 2 h after the carrageenin injection for the two upper doses of PACs and was significative 4 h after the carrageenin administration for all doses of PACs. The maximum inhibitory effect of PACs reached 63% at 4 h after carrageenin, time of the maximum development of the oedema. Figure 2 Time course of inflammatory reaction induced by injection of carrageenin 1% in rat hind paw and its antagonism by PACs (10, 30, 60 and 100 mg/kg -1 ). Inflammation is expressed as the increase of the rat paw volume (ml) from 0 to 4 h following injection of carrageenin. The volume of the paw was reduced by PACs at the four doses tested and the inhibition is time and dose-dependant. Each value is the mean ± s.e. mean of n = 6 experiments. *P < 0.05 versus carrageenin. Influence of PACs on the carrageenin-induced pleurisy In control rats, the volume of the exudate collected 4 h after carrageenin injection reached 0.87 ± 0.18 ml per rat (n = 12). This exudate contained a large number of cells, mostly (> 95%) polymorphonuclear leukocytes (PMNs). The total leukocytes number in the exudate was 119.71 ± 29.29 × 10 6 per rat (Figure 3A ). PACs significantly reduced the volume of the exudate in a dose-dependent relationship, showing a maximum inhibitory effect (48%) from the dose of 30 mg/kg which was not increased by the upper doses of PACs. As expected, the volume of the exudate was reduced in indomethacin-treated rats. On the other hand, PMNs infiltration (Figure 3B ) was significantly inhibited by PACs in a dose-dependent way and by indomethacin. Figure 3 Effect of indomethacin and PACs on carrageenin-induced pleurisy. At 4 h after carrageenin injection, the volume of the exudate (A) was reduced by PACs (10, 30, 60 and 100 mg/kg) and indomethacin (10 mg/kg) administration. The accumulation of polymorphonuclear cells (PMNs, B) in the pleural cavity was inhibited by all tested drugs. Each value is the mean ± s.e. mean of n = 6 experiments. °P < 0.05 versus sham. *P < 0.05 versus carrageenin. Effects of PACs on the release of cytokines High levels of TNF-α, IL-1β, IL-6, IL-10 and CINC-1 were found in pleural exudates induced by carrageenin (Figure 4 ). Indomethacin reduced the level of the five cytokines studied while PACs lowered significatively the levels of TNF-α (Figure 4A ), inhibited the release of IL-1β (Figure 4B ) but did not affect IL-6 levels (Figure 4C ) and IL-10 production (Figure 4D ). PACs also lowered significantly CINC-1 levels (Figure 4E ). Figure 4 Effect of indomethacin and PACs on cytokines release in pleural exudate. Pleural injection of carrageenin caused by 4 h an increase in the release of the cytokines, tumor necrosis factor alpha (TNF-α, A), interleukin-1β (IL-1β, B), interleukin-6 (IL-6, C), interleukin-10 (IL-10, D) and cytokine-induced neutrophil chemoattractant-1 (CINC-1, E). TNF-α, IL-1β and CINC-1 levels were reduced by PACs, but IL-6 and IL-10 levels were not modified. Indomethacin lowered the level of all these cytokines. Each value is the mean ± s.e. mean of n = 6 experiments. °P < 0.05 versus sham. *P < 0.05 versus carrageenin. Effect of PACs on nitrite/nitrate (NOx) levels in pleural exudate The pleural exudate of carrageenin-treated rats contained a large amount of NOx (716 ± 32 μM; n = 6) (Figure 5 ). The amount of NOx in pleural exudate of rats treated with 10 mg/kg indomethacin was similar to the content found in the control group. On the other hand, PACs, at 30 mg/kg, significantly decreased the amounts of NOx in pleural exudate from 51%. Figure 5 Effect of PACs and indomethacin on NOx formation in pleural exudate. Production of NOx release was not significantly affected by pretreatment of rats with indomethacin (10 mg/kg, intraperitoneally) while PACs caused an inhibition in NOx production. Each value is the mean ± s.e. mean of n = 6 experiments. °P < 0.05 versus sham. *P < 0.05 versus carrageenin. Histological examination of lung sections Histological examination of lung sections revealed significant tissue injury (Figure 6 ) when compared with lung sections taken from saline-treated rats (Figure 6A ). Lung withdrawn from rats treated with carrageenin showed oedema, tissue injury and an extensive infiltration of the tissue by PMNs (Figure 6B ). Pretreatment of rats with indomethacin (10 mg/kg, i.p.) or PACs (30 mg/kg, i.p.) showed a reduced lung injury as well as a decrease in the infiltration of PMNs (Figures 6C,6D ). Figure 6 Effect of PACs on lung injury. When compared to lung sections taken from control animals (A), lung sections from carrageenin-treated rats (B) demonstrated interstitial haemorrhage and polymorphonuclear leukocyte accumulation. Lung sections from a carrageenin-treated rat that had received PACs (30 mg/kg) (C) or indomethacin (10 mg/kg) (D) exhibited reduced interstitial haemorraghe and a lesser cellular infiltration. Original magnification: × 125. Discussion Proanthocyanidins (PACs) from Ribes nigrum leaves reduced the inflammatory reactions induced by carrageenin in rats : the extent of the paw oedema was halved, the volume of the pleural exudates and its content in TNF-α, IL-1β, CINC-1 and NOx were reduced, the infiltration of leukocytes into the lungs and the accumulation of leukocytes into the pleural cavity were largely diminished. PACs have been reported to be able to scavenge free radicals and NO [ 17 ]. This property could be an explanation of the reduction of NOx level in the pleural fluid after PACs treatment. According to Ialenti et al [ 18 ], during the development of carrageenin-induced pleurisy, the main role of NO is the inhibition of leukocytes migration to the inflammatory site. However, in rats pretreated with PACs, the level of NOx and of leukocytes are simultaneously reduced. This result suggests that PACs could more or less directly affect the transmigration of leukocytes. The development of carrageenin-induced inflammatory reactions in rats results from the activation of the kinin system, the accumulation of leukocytes and the release of several mediators such as prostanoids and cytokines [ 19 , 20 ]. Indeed, these inflammatory reactions are greatly reduced in kininogen-deficient rats, in animals pretreated with kinin-antagonists and in leucopenic rats [ 19 , 21 ]. Previous studies [ 22 ] have demonstrated that PACs can reduce other inflammatory reactions such as the oedemas induced in rats by nystatin and concanavalin-A in which the kinin system is not involved [ 19 ] but in which leukocytes play a major role [ 23 ]. The comparison of the major determinants of these three kinds of reactions, all inhibited by PACs, is another argument suggesting that the main target explaining the anti-inflammatory activity of PACs would be the involvement of leukocytes. Pro-inflammatory cytokines TNF-α, IL-1β and IL-6 are sequentially released in the pleural exudates induced by carrageenin in rat [ 14 ]. These cytokines cause chemotaxis to attract granulocytes and monocytes and then, migrating leukocytes produce, in turn, further cytokines, such as TNF-α and IL-1β, and other pro-inflammatory mediators [ 15 ]. IL-6 has been proposed as a crucial mediator for the development of carrageenin-induced pleurisy and for the accumulation of leukocytes in the inflammatory site. Indeed, in carrageenin-induced pleurisy in IL-6 knock-out mice, the degree of plasma exudation, leukocyte migration and the release of TNF-α and IL-1β were greatly reduced. Moreover, a positive feedback plays an important part in the development of the oedema as levels of TNF-α and IL-1β are attenuated in IL-6 knock-out mice [ 24 ]. PACs did not affect the level of IL-6 and of IL-10, an anti-inflammatory cytokine, but reduced the pleural content of TNF-α, IL-1β and leukocytes. This result indicates that the release of IL-6 does not depend on the presence of leukocytes, of TNF-α and IL-1β on one hand, and, on the other hand, suggest that the main target of PACs would be the accumulation of leukocytes and the associated release of inflammatory mediators. TNF-α plays an important role in promoting and amplifying lung inflammation through the release of chemotactic factors such as CINC-1 (rat IL-8), an important mediator that promotes the migration of neutrophils [ 25 ] and oesinophils [ 26 ]. CINC-1 can increase the expression of LFA-1 integrin on rat neutrophils [ 27 ] and because expression of leukocyte adhesion molecules such as E-selectin is dependent on CINC [ 28 ], the inhibition of CINC-1 levels in pleural exudates by PACs may exert both direct and indirect effects on neutrophil vascular adhesion and extravascular migration. PACs probably acts by disrupting TNF-α, IL-1β, CINC-1 and PMNs accumulation pathways. One of the mechanism for the anti-inflammatory effect of PACs may be attenuation of the migration of PMNs in the exudate, because CINC-1, a representative cytokine for PMNs migration in rats, is suppressed by PACs in parallel with PMNs number dose-related fashion. Although, clarification for the precise mechanism would remain in future study. Recently, grape seed proanthocyanidins have been demonstrated to reduce the expression of soluble adhesion molecules, ICAM-1, VCAM-1 and E-selectin in the plasma of systemic sclerosis patients [ 29 ]. The same compounds have been shown to inhibit TNF-α-induced V-CAM-1 expression in human umbilical vein endothelial cells cultures [ 30 ]. A possible mechanism of the anti-inflammatory effect of PACs would be an interference with the expression or the effect of adhesion molecules. This interference would result in a reduction of polymorphonuclear cell migration and subsequently in a reduction of the release of pro-inflammatory factors such as TNF-α and IL-1β. Injection of carrageenin into the pleural cavity induces the accumulation of leukocytes, a release of cytokines, the expression of inducible NO synthase and of cyclo-oxygenase-2, and thus the release of large amounts of nitric oxide and of prostanoids [ 16 ]. The inhibitory effect of PACs on the accumulation of leukocytes and on the release of TNF-α and IL-1β could have resulted in a decrease in the induction of inducible NO-synthase and of cyclo-oxygenase-2 and finally of plasma exudation. Comparatively, some animals have been treated with indomethacin. The inhibitory effect of this well-known non-steroidal anti-inflammatory drug is larger than that obtained with PACs. Indomethacin greatly reduced plasma exudation, nearly suppressed the accumulation of leukocytes and decreased the levels of the cytokines while, it did not modify the pleural content of NOx. Indomethacin is known to inhibit the cyclooxygenase-1 and -2 responsible of the release of PGE 2 production. The peak of cyclooxygenase-2 activity measured by prostanoid levels in carrageenin-induced pleural exudates spreads from 2 to 6 h after irritant injection [ 31 , 32 ]. Both IL-6 and IL-10 release are, in part, stimulated by PGE 2 [ 33 , 34 ]. An inhibition of PGE 2 production by high doses of indomethacin could result in a downregulation of IL-6 and IL-10 production [ 35 , 36 ]. Moreover, Cuzzocrea et al [ 24 ], using carrageenin-induced pleurisy in IL-6 knock out mice, showed that IL-1β and TNF-α production in the pleural exudates is, at least, partly IL-6 dependent. Our results showing a reduction in the levels of IL-1β, TNF-α, IL-6, IL-10 and CINC-1 by indomethacin four hours after the induction of the pleurisy, could be mainly explained through the inhibition of PGE 2 and IL-6 pathways. Conclusions In conclusion, we have shown that proanthocyanidins isolated from Ribes nigrum leaves interfere with the accumulation of circulating leukocytes, associated with a reduction of pro-inflammatory factors such as TNF-α, IL-1β and CINC-1, a decrease of NOx level and a decrease in plasma exudation. Methods Animals We used male Wistar rats, weighing 250 – 300 gm. The animals were maintained on a standard laboratory diet with free access to water. The experiments were conducted as approved by the Animal Ethics Committee of the University of Liège, Belgium. Paw oedema Rats were pretreated with an intraperitoneal administration of saline or PACs (10, 30, 60 and 100 mg/kg). Thirty minutes later, lambda carrageenin, (0.1 ml, 10 mg/ml) was injected into the plantar region of the right hind paw. Each experimental group contained six animals. Paw volume was measured using a water plethysmometer (Ugo Basile) before and 1 h, 2 h and 4 h after the injection of carrageenin. After 4 h, the animals were anaesthetized with a large dose of sodium pentobarbital (80 mg/kg). Carrageenin-induced pleurisy Rats were pretreated with an intraperitoneal injection of saline, PACs (10, 30, 60 or 100 mg/kg) or indomethacin (10 mg/kg) 30 min before the intrapleural injection of the irritant. They were then anaesthetized with ketamine HCl (75 mg/kg) and carrageenin (0.2 ml, 10 mg/ml) or saline (0.2 ml) was administered into the right pleural cavity. Each experimental group contained 6 animals. Four hours later, the animals were anaesthetized with sodium pentobarbital (80 mg/kg). The chest was carefully opened and the pleural cavity rinsed with 2.0 ml saline solution containing heparin (5 U/ml). Exudates and washing solutions were removed by aspiration and the total volume measured. Exudates with blood were rejected. Exudates were aliquoted and kept frozen at -32°C. After removal of the exudates, lungs were withdrawn and fixed for one week under 30 cm pressure with 10% formaldehyde aqueous solution containing 0.480 M Na 2 HPO 4 and 0.187 M KH 2 PO 4 (pH 7.2) at room temperature. They were then dehydrated by graded ethanol and embedded in Paraplast. Tissue sections (thickness 7 μm) were deparaffinized with UltraClear, stained with hematoxylin-eosine and examined using light microscopy. The volume of the exudates was calculated by subtracting the volume of the washing solution (2.0 ml) from the total volume recovered. A sample of each exudate was diluted in phosphate buffer and total leukocyte count was performed using a hemocytometer. The levels of IL-1β, TNF-α, IL-6 and IL-10 in the exudates were measured using a colorimetric commercial ELISA kit (Biosource, Nivelles, Belgium) with a lower detection limit of 4, 3, 8 and 5 pg/ml, respectively. The levels of CINC-1 in the exudates were measured using a colorimetric commercial ELISA kit (Amersham Biosciences, Freiburg, Germany) with a lower detection limit of 0.49 pg/ml. The amount of NOx (nitrite/nitrate) present in the exudates was determined using a microplate assay method (Calbiochem, Leuven, Belgium) based on Griess reaction after reduction of NO 3 - to NO 2 - with a lower detection limit of 1 μM. Extraction and purification of proanthocyanidins Proanthocyanidins from Ribes nigrum leaves were extracted and isolated according to a previously described method [ 37 ]. A voucher sample (RN 210590) has been deposited in the Pharmaceutical Institute of Liège, Belgium. Briefly, leaves were powdered separately and then extracted at room temperature with acetone (70 % v/v in water). The acetone was removed under vacuum at 40°C. The resulting aqueous solution was freeze-dried. Isolation was carried out by MPLC on reversed-phase RP8 with water-acetone (9:1) to obtain a total proanthocyanidin-enriched fraction (PACs). Materials We used ketamine-HCl from Pfizer (Bruxelles, Belgium), sodium pentobarbital from Ceva (Bruxelles, Belgium) and heparin from B. Braun Medicals (Diegem, Belgium). PACs and lambda carrageenin (Sigma, Bornem, Belgium) were dissolved in saline. Indomethacin (Merck, Sharp and Dohme, Leuven, Belgium) was dissolved in Tris-HCl (0.15 M, pH 7.4). Statistical evaluation Results are given as mean ± standard error of the mean (s.e. mean) of N observations. For the oedema paw studies, a Mixed Procedure SAS (normal distribution) was used to compare difference of least square means. For the pleurisy studies, data sets were examined by one-way analysis of variance (ANOVA) followed by a Scheffe post-hoc test. A P -value of less than 0.05 was considered significant. Authors' contributions NG carried out PACs isolation, animal experimentation, immunoassays, lung sections and statistical analysis. MT coordinated and participated to the PACs isolation. LA coordinated the PACs isolation. JD participated in animal experimentation, conceived of the study and participated in its design and coordination.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526370.xml

514895

Co-transcriptional folding is encoded within RNA genes

Background Most of the existing RNA structure prediction programs fold a completely synthesized RNA molecule. However, within the cell, RNA molecules emerge sequentially during the directed process of transcription. Dedicated experiments with individual RNA molecules have shown that RNA folds while it is being transcribed and that its correct folding can also depend on the proper speed of transcription. Methods The main aim of this work is to study if and how co-transcriptional folding is encoded within the primary and secondary structure of RNA genes. In order to achieve this, we study the known primary and secondary structures of a comprehensive data set of 361 RNA genes as well as a set of 48 RNA sequences that are known to differ from the originally transcribed sequence units. We detect co-transcriptional folding by defining two measures of directedness which quantify the extend of asymmetry between alternative helices that lie 5' and those that lie 3' of the known helices with which they compete. Results We show with statistical significance that co-transcriptional folding strongly influences RNA sequences in two ways: (1) alternative helices that would compete with the formation of the functional structure during co-transcriptional folding are suppressed and (2) the formation of transient structures which may serve as guidelines for the co-transcriptional folding pathway is encouraged. Conclusions These findings have a number of implications for RNA secondary structure prediction methods and the detection of RNA genes.

Background Most of the existing computational methods for RNA secondary structure prediction fold an already completely synthesized RNA molecule. This is done either by minimizing its free energy (e.g. done by MFOLD [ 1 - 3 ] and by the programs of the VIENNA package [ 4 - 8 ]) or by maximizing the probability under a model whose parameters can incorporate a variety of different sources of information, e.g. comparative information, free energy and evolutionary information (e.g. [ 9 ], TRNASCAN-SE [ 10 ], PFOLD [ 11 , 12 ] and QRNA [ 13 ]). All of these programs, including those that predict folding pathways by folding an already synthesized RNA sequence [ 14 , 15 ], therefore disregard the effects that co-transcriptional folding may have on the RNA's functional secondary structure. They essentially aim to predict the thermodynamic RNA structure , i.e. the secondary structure that minimizes the free energy of the molecule. However, theoretical studies of RNA molecules [ 16 ] indicate that the thermodynamic structure of even moderately long RNA molecules need not necessarily correspond to the functional structure which confers the desired functionality within the organism to the RNA molecule. RNA molecules are known to fold as they emerge during transcription [ 17 , 18 ]. Transcription is a directed process of variable speed, during which the 5' end of the RNA molecule is synthesized before its 3' end. Hydrogen-bonds at the 5' end of the RNA molecule can thus form earlier in time than hydrogen-bonds involving the 3' end of the molecule. The thus emerging secondary structure elements can be transient or not, depending on their stability, their formation times and the availability and stability of competing alternative pairing partners. The directedness and also the speed of transcription can influence both the folding pathway and the functional secondary structure of the RNA molecule. We call this phenomenon sequential or co-transcriptional folding and call the resulting secondary structure the kinetic structure of the RNA molecule. Co-transcriptional folding leads to the formation of temporary secondary structure elements [ 18 , 19 ]. The time that it takes to form and replace these transitory structure elements may successively narrow down the set of accessible folding pathways and may thereby guide the folding towards an ensemble of secondary structures which contains the desired functional secondary structure. However, these temporary secondary structure elements can also have distinct biological functions, e.g. in viroids [ 19 ] and as initial sites for protein anchoring during pre-mRNA transcription [ 20 ]. Based on experimental and theoretical investigations, Harlepp et. al. [ 21 ] and Isambert et. al. [ 22 ] found that temporary structures may form during transcription. All these results suggest that temporary secondary structure elements may play an important role in the correct folding of RNA sequences. The speed of transcription also has an effect on folding which can be investigated by varying the nucleoside triphosphate concentration [ 19 ] or by transcribing RNA genes with viral polymerase T7 which has faster elongation during transcription than bacterial polymerases [ 23 , 24 ]. Both decreasing and increasing the natural speed of transcription can yield inactive transcripts [ 23 , 24 ]. Recent in vitro investigations of the Tetrahymena ribozyme [ 25 ] show that its co-transcriptional folding in vitro is twice as fast as the refolding of the entire RNA molecule under the same conditions and that both lead to the same functional folding. Moreover, they find that the co-transcriptional folding in vitro is still much slower than in vivo. Among the multitude of biochemical processes which are known to occur transcriptionally [ 26 , 27 ], some processes act in order to prevent the mis-folding of RNA molecules. RNA chaperones are proteins which are believed to help refold mis-folded RNA structures by promoting intermolecular RNA-RNA annealing through non-specific interaction [ 28 ]. Without RNA chaperones, moderately long GC-rich helices have dissociation half-times of up to 100 years [ 29 ]. This time can be significantly reduced by RNA chaperones, which preferentially bind stretches of unfolded RNA and thereby decrease the kinetic barrier between the correct and incorrect secondary structure elements [ 28 ]. Specific RNA-binding proteins are also known to promote RNA folding by either guiding its folding or stabilizing its correct structure [ 30 , 31 ]. The hnRNP proteins non-specifically bind pre-messenger RNA and help in the splicing process [ 32 ]. RNA sequences can also promote the proper folding of other RNA sequences. It is known, for example, that the temporary interaction with highly conserved leader sequences of bacterial rRNA-operons is needed for the proper formation of 30S ribosomal subunits and the maturation of 16S rRNA [ 33 , 34 ]. All these experimental and the few theoretical findings suggest that co-transcriptional folding may play an important role in the correct folding of RNA molecules. They also show that the functional structure may only be a transient one which is available during a certain time span and that the functional structure need not correspond to the structure which would dominate the ensemble of structures after an infinite time span. Little is known whether co-transcriptional folding is mainly governed by the specific or non-specific binding of proteins (or other molecules) which target the emerging RNA or whether the primary structure of the RNA molecule itself conveys the desired properties to guide its own correct co-transcriptional folding. In this paper, we propose several statistics in order to detect, if and how co-transcriptional folding influences RNA sequences. Using these statistics, we show that the effects of co-transcriptional folding are widespread in RNA genes. Methods Theory We want to show that an RNA sequence is organized in such a way to help the formation of the functional secondary structure during transcription. We aim to support this hypothesis by detecting two different features: • Possible competitors of helices in the functional structure are suppressed. When the 3' end of a helix that is part of the final secondary structure emerges during transcription, the number of possible competitors for the 5' part of the helix should be as low as possible in order to promote the formation of the correct helix. • The folding pathway is engineered. During transcription, several temporary helices are formed which may guide the folding process. We investigate these features using several statistics which are based on the known primary and secondary structures of our RNA sequences. A crucial point in investigating these features is to define a set of statistics that have expectation of zero in the H 0 case, when we suppose no co-transcriptional folding. However, verifying that these statistics have an expectation value of zero in the H 0 case cannot simply be achieved by analyzing random sequences. Indeed, even generating random sequences is not trivial. First, it is hard to reliably predict the minimum free energy structure for the randomized sequences as most secondary structure prediction algorithms discard pseudo-knots and, even without pseudo-knots, predict only on average about 70 % of the base-pairs correctly. In addition, there is no guarantee that the secondary structure with the lowest free energy would correspond to the functional one. Second, even if the random sequences are generated by a shuffling algorithm which keeps the given secondary structure fixed, it cannot be guaranteed that the fixed structure remains the correct one for the new primary sequence. Generating random sequences therefore provides no straightforward solution for obtaining a H 0 statistics with expectation value zero. We circumvent this problem by studying pairs of statistics, where both statistics have the same, unknown expectation value in the H 0 case and where one statistics has a bias away from the H 0 expectation value in case of co-transcriptional folding, while the other statistics is not affected by co-transcriptional folding. By studying the difference of these two statistics, we thus gain a new statistics with expectation value zero in the case of no co-transcriptional folding and an expectation value larger or smaller than zero in the case of co-transcriptional folding. The statistics (which we will define in detail below) measure the presence of alternative helices which compete for at least one base-pair with the helices of the known secondary structure. These competing alternative helices are required to consists of at least min stem = 9 consecutive base-pairs of type {G - C, C - G, A - U, U - A, G - U, U - G} and are calculated by a dynamic programming procedure in which the known primary and secondary structure of the RNA is fixed, see Figure 1 for the definition of a competing, alternative helix. We checked that we obtain qualitatively similar results for smaller and larger min stem values (data not shown). While calculating all helices of at least min stem length, we test which of these helices constitute competing alternatives to helices of the known secondary structure and record each such competing case in one of our statistics. These alternative helices may be part of a pseudo-knotted structure and we do not discard them. As each of the two bases i and of a base-pair in a known helix can have a competing alternative base-pairing partner within an alternative helix and as this alternative partner can either be found 5' (before), 3' (behind) or between the two strands of the known helix, all cases can be classified into six different classes. Of these six, we discard the two classes where the alternative helix falls between the two strands of the known helix as this un-paired loop region is typically too short to accommodate an alternative helix of at least min stem length. The remaining four classes, see Figure 2 , can be sub-divided into two cis- and two trans- alternative classes, depending on whether the known base-pairing partners lie between the alternative base-pairing partners (trans) or not (cis). The four statistics 3' cis, 3'trans, 5'cis and 5'trans that we use correspond to these four classes. Figure 1 Definition of a competing, alternative helix. Pictorial definition of a competing, alternative helix. The known base-pair between sequence positions i and has to have at least two other directly adjacent base-pairs within the known secondary structure (right) and the competing, alternative helix has to contain an alternative base-pair between sequence positions i and c ( c is the competitor of ) which has to be contained within a helix of minimum stem length (left). Figure 2 Definition of the statistics. Pictorial definitions of the four configurations 3' cis , 3' trans , 5' cis and 5' trans which correspond to the four statistics used to measure the directedness of RNA folding. Sequence positions i and form a base-pair within the known secondary structure. Sequence position c is an alternative base-pairing partner for i (but according to the base-pairing rules therefore not for ) within a competing, alternative helix of a minimum length min stem . See the text for more explanation. It is important to note that even without co-transcriptional folding, the destabilizing effects of competing cis- and trans -alternative helices are not necessarily the same as the stacking energies are not symmetric with respect to the 5' → 3' direction of the RNA sequence [ 3 ]. In addition, alternative cis -pairing partners are closer to the known pairing partners than trans -pairing partners and may thus lead more easily to incorrect helices. We may therefore compare only cis -competitors with other cis -competitors and trans -competitors with other trans -competitors. This yields two possible comparisons: 3' trans versus 5' trans and 5' cis versus 3' cis, see Figure 2 , with which we can measure the effects of co-transcriptional folding. We proceed as follows to detect if co-transcriptional folding takes place: For every RNA sequence of the data set, we detect events of type 3' cis, 3'trans, 5'cis and 5'trans , where an alternative helix competes with a known helix. Each such event is given two different weights, see Table 1 for an overview of definitions: (1) a weight of 1/ ( d·log ( l )), where d is the distance between the two competing helices and l is the length of the sub-sequence 5' or 3' of the known helix on which the competing helix falls, or (2) a weight of | G | / ( d ·log( l )), where the former weight is multiplied by the absolute value of the free energy G of the competing, alternative helix. The factor 1/ d gives alternative helices that are far away from the known helix a smaller weight than closer ones. The factor 1 /log (l) accounts for the fact that log (l) is proportional to the expected sum of 1 /d statistics for a sub-sequence of length l (i.e. the integral ). The free energy factor G in the second type of weights gives stable alternative helices which have a larger impact on the folding pathway a greater weight than helices which are easily unfolded. Statistics derived from weights of type 1 /(d log( l )) are denoted by an index p (for plain) and those of type | G | / ( d ·log( l )) by an index g (for free energy). By summing the weighted counts for each of the four classes of events, we thus arrive at eight different scalar values which characterize each RNA sequence: 3'Trans x , 3'Cis x , 5'Trans x and 5'Cis x for x ∈ { p,g }. Table 1 Definitions of the different statistics. Definitions of the different statistics used. i and denote the sequence positions of a base-pair in the known structure, c is an alternative pairing partner for i (but according to the base-pairing rules therefore not for ), L is the length of the RNA sequence, N is the number of sequences in the data set and the index x indicates the type of weight used. Please refer to the text for a description of how alternative pairing partners are calculated. x p plain weights g free energy weights 3'cis x 1/(( c - i ) log( L - i )) | G ci |/(( c - i ) log( L - i )) 3'trans x 1/(( c - ) log( L - )) | G ci |/((c - ) log( L - )) 5' cis x 1/(( i - c ) log( i )) |G ic |/(( i - c ) log( i )) 5' trans x 1/(( - c )log( )) |G ic |/(( - c ) log( )) cis x 5' cis x - 3' cis x trans x 3' trans x - 5' trans x 3' Cis x Σ #3' cis 3' cis x 3' Trans x Σ #3' trans 3' trans x 5' Cis x Σ #5' cis 5' cis x 5' Trans x Σ #5' trans 5' trans x Cis x 5' Cis x - 3' Cis x Trans x 3' Trans x - 5' Trans x where x ∈ { p,g }, y ∈ {3' Cis , 3' Trans , 5' Cis , 5' Trans , Cis , Trans } We can now define the two statistics which are capable of measuring the two main types of asymmetry within each RNA sequence: Cis := 5' Cis - 3' Cis Trans := 3' Trans - 5' Trans which can calculate for both types of weights. Without co-transcriptional folding, the expectation value of these two statistics is zero. Co-transcriptional folding induces two types of asymmetries by suppressing the number of alternative helices which compete with the final helices (indicated by an increased number of configurations, see Figure 2 ) and by promoting the formation of transient helices which guide the correct folding (indicated by an increased number of configurations). Both types of effects are indicated by an expectation value larger than zero for the respective statistics. Without co-transcriptional folding, the introduced statistics have an expectation of zero, moreover, the distributions should be symmetric. The number of positive cases (pos) thus follows a binomial distribution with parameter p = 0.5 and the statistic where n is the number of all cases, approximately follows a standard normal distribution. If this value is sufficiently positive, we have to reject the hypothesis that co-transcriptional folding is not encoded within RNA genes. Data All 16S rRNA, 23S rRNA as well as Group I and Group II type intron sequences with completely known secondary structures were downloaded from the Comparative RNA Web (CRW) Site [ 35 , 36 ], resulting in 304 16S rRNA, 84 23S rRNA, 15 Group I intron and 6 Group II intron sequences from three main taxonomical units (Archea, Bacteria, Eukaryotes) and two organelles, see Table 2 . Table 2 Composition of the two data sets. Taxonomic unit all 16S rRNA 23S rRNA Group I Group II Data set A Archea 28 22 6 0 0 Bacteria 277 232 45 0 0 Eukaryotes 41 35 6 0 0 Chloroplasts 6 6 0 0 0 Mitochondria 9 9 0 0 0 Sum 361 304 57 0 0 Data set B Eukaryotes 15 0 0 15 0 Bacteria 5 0 5 0 0 Chloroplasts 5 0 5 0 0 Mitochondria 23 0 17 0 6 Sum 48 0 27 15 6 Organellar 23S rRNA sequences frequently contain Group I introns and recent research revealed that the 23S rRNA of several hyperthermophilic bacteria also have Group I intron [ 37 ]. Other species only rarely have introns in rRNA genes, however, some 16S rRNA introns are known [ 38 ]. rRNA genes in bacteria are encoded in the so-called rrn-operon (see for example [ 39 ]). The canonical order of rRNA genes in the rrn-operon is 16S-23S-5S, but some exceptions to this rule are known. In Vibrio harvey, the order is 23S-16S-5S [ 40 ], but not in Vibrio cholerae [ 41 ] and Vibrio parahaemolyticus [ 42 ], whose 16S rRNA sequences were downloaded from the Comparative RNA Web Site. We divided the gathered sequences into two sets: data set A which consists of all RNA sequences that are thought to correspond to the originally transcribed sequence units and data set B which contains all those RNA sequences that are known to differ from the originally transcribed sequence units. Data set B thus contains the Group I and II intron sequences, organellar and hyperthermophilic bacteria 23S RNA sequences. As we neither know the sequence nor the secondary structure of the original transcript units from which the sequences of data set B were derived, we are limited to detecting the effects of co-transcriptional folding within these shorter sequences. We expect this to be much more difficult than in sequences that correspond to the originally transcribed sequence units as co-transcriptional folding introduces long range effects which are harder to detect the shorter the investigated sub-sequence gets. See Table 2 for a detailed overview of the composition of each data set. Results We calculated the 3'Cis x , 3'Trans x , 5'Cis x and 5'Trans x values for both types of weights, i.e. x ∈ { p,g }, for each sequence in the two data sets. From these values we then derived each sequence's Cis x and Trans x values, again for both x types of weights. Their distributions are shown in Figure 3 . Averaging over the values of all sequences in each of the two data sets resulted in the final values shown in Table 3 . Figure 3 Distribution of Cis and Trans values. Distribution of Cis and Trans values for the sequences of data sets A and B and both types of weights (plain (p) or free energy based (g)). The area under each curve has been normalized to one to allow a direct comparison between the two data sets. Table 3 Average values for different statistics. Final values of the different statistics which were obtained by averaging the values of each sequence in the data set. The error shown is the standard deviation. dataset A 0.215 ± 0.009 0.461 ± 0.032 0.285 ± 0.009 0.382 ± 0.032 0.070 ± 0.004 0.079 ± 0.026 B 0.298 ± 0.040 0.562 ± 0.086 0.296 ± 0.043 0.521 ± 0.075 -0.003 ± 0.015 0.041 ± 0.082 dataset A 2.916 ± 0.106 6.236 ± 0.431 3.710 ± 0.111 5.134 ± 0.354 0.794 ± 0.061 1.102 ± 0.384 B 3.392 ± 0.406 7.033 ± 1.050 3.362 ± 0.456 6.380 ± 0.954 -0.030 ± 0.184 0.653 ± 1.253 The first thing to note in Figure 3 is that all distributions follow approximately a symmetric distribution, thus confirming our theoretical considerations, and that the distributions of data set B are always shifted towards lower values with respect to the corresponding distributions for data set A which are always centered around average values larger than zero. The mean values of Cis and Trans in Table 3 are positive for data set A for both types of weights, indicating the influence of co-transcriptional folding, whereas they are closer to zero or even negative in the case of data set B. A Cis value larger zero means that configurations of type outnumber those of type , see Figure 2 . The formation of potential transient helices involving base-pairs between c and i that can later yield to the final secondary structure element containing the base-pair between i and thus seems to be encouraged. However, these transient structure elements may not be too stable if they are to guide rather than impede the proper folding. The presence of transient helices could thus be further substantiated by showing that these transient helices are less stable than the final helix. In contrast to the configuration, the competing ic helices in the case are suppressed as they lie 3' of the final helix and thus emerge later in time during co-transcriptional folding. A Cis value larger than zero can therefore be explained by the presence of temporary helices which may guide the formation of the final, functional secondary structure during co-transcriptional folding. A Trans value larger than zero means that configurations are less frequent than configurations, see Figure 2 . In the configuration, both c and are competing pairing partners for i as they both emerge before i during transcription. This may lead to the formation of wrong ci helices, whereas the order of pairing partners in the configuration has a lower risk of mis-folding due the c emerging only after the and thus only after the helix could have already formed. In addition, 3'Trans > 3'Cis in Table 3 can be interpreted as a stabilization of the final, functional secondary structure. Imagine that the hydrogen bounds of the or helix temporarily break up. In the case of the 3'Trans configuration, the pairing partners come in the order along the RNA sequence, whereas they come in the order in the 3'Cis configuration. In the order, the c part is in vicinity to the i part, so the possibility of ending up with a wrong refolding due to a ic helix is larger than in the case. Overall, we can thus conclude from the average values in Table 3 , that the sequences of data set A are tailored towards co-transcriptional folding, whereas we cannot reliably detect the effects of co-transcriptional folding within data set B. We detected co-transcriptional folding in data set A by showing that the final secondary structure is actively stabilized (3'Trans > 3'Cis), that the formation of temporary helices may guide the structure formation and that these helices may thus be used to actively engineer a folding pathway (Cis > 0) and that secondary structure elements which may interfere with the formation of the final, functional secondary structure during co-transcriptional folding are suppressed (Trans > 0). In order to quantify the influence of co-transcriptional folding further, we calculated two statistics, a t-test for the hypothesis that the given statistics have an expectation value of zero as well as the p-value of the number of positive cases for our two co-transcriptional folding indicators, see Table 4 . The high p-values for data set B imply that the presence of co-transcriptional folding is not well supported in this data set. However, the corresponding indicators strongly support co-transcriptional folding within data set A. Table 4 Statistical significance of results. p-values of t-test for the hypothesis that the final values in Table 3 have an expectation value of zero as well as the p-values for the hypothesis that the number of positive cases follows a binomial distribution with parameter 0.5. dataset A B p-value for t-test p-value for pos p-value for t-test p-value for pos < 0.0001 < 0.0001 0.5733 0.6137 < 0.0001 < 0.0001 0.5650 0.6137 0.0012 < 0.0001 0.3093 0.8068 0.0021 < 0.0001 0.3011 0.5000 Discussion Recent experimental studies [ 23 , 24 , 19 ] have shown that the proper speed of transcription helps the correct folding of RNA molecules. In addition, theoretical studies [ 16 ] indicate that the functional structure of an RNA need not correspond to the minimum free energy structure, even for moderately long RNA molecules. These findings suggest that co-transcriptional folding may play a decisive role in the formation of functional RNA structures. Although our statistics are able to reveal two general effects of co-transcriptional folding within data set A, we cannot conclude that they would be powerful enough to serve as a reliable indicator of co-transcriptional folding for single RNA sequences, as some of the sequences in data set A may not correspond to the originally transcribed sequence units. In addition, all of our statistics consider only a first order effect of co-transcriptional folding by studying alternative helices for the known helices, but do not take higher order effects into account as e.g. alternative helices of alternative helices etc. Based on computer simulations, H. Isambert et. al. [ 43 ] conjecture that pseudo-knotted motifs are common in co-transcriptional folding. Pseudo-knotted structures are explicitly included in our statistics, as the corresponding calculations naturally allow for alternative helices which are part of a pseudo-knot and as we do not reject them. Conclusions To summarize, our findings show that co-transcriptional folding is a guiding principle in the formation of functional RNA structure and that it can influence both the primary and potential secondary structures of an RNA molecule. This has several implications. Current algorithms for RNA secondary structure prediction can probably be improved by adopting co-transcriptional folding as a guiding principle rather than only free energy minimization. This may hopefully provide the extra information needed to be able to reliably detect RNA genes [ 44 ]. Several groups have already come up with computer algorithms which attempt to fold an RNA sequence co-transcriptionally [ 45 - 48 , 22 ]. These findings also have implications for computational methods which infer the phylogeny of RNA sequences, as these consider only co-evolution within the base-pairs of the functional helices, but discard any information due to the conservation of folding pathways and may hence mis-estimate evolutionary times. Similar arguments hold for all comparative studies that aim to detect functional secondary structure elements, since co-evolution of nucleic acids does not necessarily imply that these nucleic acids are base-paired in the final functional secondary structure. As evolution probably not only selects for the correct functional secondary structure, but also for a suitable folding pathway, it should be possible to detect the effects of co-transcriptional folding also in a comparative way. Most importantly, co-transcriptional folding should lead to a better understanding of how RNA sequences fold. This should in turn enable us to also understand why some RNA sequences mis-fold and fail to function properly in the organism. Even though protein folding is known to differ in many respects from RNA folding, they also have some features in common [ 49 ]. One of the obvious similarities is that both proteins and RNA sequences are synthesized in a directional process. It would thus be interesting to investigate if protein folding is also influenced by co-translational folding. In this study, we neither attempted to study the effects that co-transcriptional folding may have on sequences that are transcribed together (e.g. genes in an operon) nor to study the influence that the binding by proteins or RNA sequences or RNA editing may have on the co-transcriptional folding pathway and the final, functional RNA structure. This will almost certainly require more refined investigation methods, but we hope that this study provides enough insight and motivation to start to tackle these exciting questions. Authors' contributions I.M.M. proposed this work and contributed the main idea for the statistics. I.M. selected the data and evaluated the statistical significance of the results. Both authors shared the programming tasks and the writing of the manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514895.xml

300675

Structural Mechanism Shows How Transferrin Receptor Binds Multiple Ligands and Sheds Light on a Hereditary Iron Disease

Iron is an essential nutrient for sustaining life-forms as diverse as plankton and humans. But too much iron, or too little, can spell trouble. Mammalian cells maintain the proper balance partly with the help of a specialized cell surface protein called the transferrin receptor (TfR). TfRs bind to the iron-carrying transferrin protein (Fe-Tf) and escort their cargo to the cell's interior. (To learn more about iron metabolism, see the primer by Tracey A. Rouault in this issue [DOI: 10.3171/journal.pbio.0000079 ].) This receptor also binds the hereditary hemochromatosis protein (HFE), which is mutated in individuals who have the common iron-overload disorder hereditary hemochromatosis. While the molecular pathway that mediates cellular intake of iron through the TfR is known, it was not clear just how TfR assists in iron release to the cell and how it binds HFE and transferrin. By introducing multiple mutations in human TfRs, Pamela Bjorkman and colleagues identified functional binding sites for transferrin in both its iron-loaded and iron-free (apo-Tf) forms and for HFE. From these data, the researchers mapped out a scenario of the dynamic interactions between receptor and ligands (the bound molecule) and worked out a structure-based model for the mechanism of TfR-assisted iron release from Fe-Tf. Bjorkman's lab, which had previously solved the structures of both HFE and HFE bound to the TfR, used their structural information to investigate how the proteins interact, which amino acid residues are required for binding, whether the two ligands bind differently to the receptor, and how HFE binding affects transferrin binding. They found that Fe-Tf and HFE occupy the same or an overlapping site on the receptor, but since transferrin is much larger than the HFE protein, it appeared that transferrin could also interact with other parts of TfR. And it remained to be seen whether TfR discriminated between the iron-loaded and iron-free states of transferrin. In this study, Bjorkman and colleagues expanded their library of TfR mutants to clarify the transferrin binding signature on TfR and to see how the TfR mutations affect the way apo-Tf and Fe-Tf interact with the receptor. They characterized the binding affinities of 30 TfR mutants to HFE and Fe-Tf and to apo-Tf, and they report that mutations in 11 of the TfR residues interfere with either one or both forms of transferrin. Four of these residues are essential for transferrin binding and are conserved in all known TfR DNA sequences. Since residues that didn't have much impact are not conserved, the scientists say the results are likely to describe transferrin–TfR interactions for other species as well. As expected, the most critical residues required for transferrin binding fall within the receptor's helical domain and have significant physical overlap with residues required for HFE binding; though some residues that are required for apo-Tf binding do not affect Fe-Tf binding. Bjorkman et al. also identify additional residues in another domain on TfR (called the protease-like domain) that support Fe-Tf but not apo-Tf binding, confirming that the receptor binding footprints of the two metal-binding states of transferrin are indeed different. With a structural model showing where Fe-Tf and apo-Tf bind to the receptor, they could evaluate how they bind and thus explain how the receptor mediates iron release. By suggesting a mechanism through which TfR binding regulates iron release, this structural model of the transferrin–TfR complex will bolster efforts to elucidate the molecular details of this process. Confirmation that transferrin and HFE do indeed compete for docking privileges reveals a possible role for HFE in maintaining iron homeostasis and will provide valuable insights into the dysregulation that leads to the warehousing of iron and resulting tissue and organ damage associated with hemochromatosis. Ribbon diagram of transferrin receptor homodimer

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC300675.xml

549210

An evaluation of the metabolic syndrome in the HyperGEN study

Background In 2001 the National Cholesterol Education Program (NCEP) provided a categorical definition for metabolic syndrome (c-MetS). We studied the extent to which two ethnic groups, Blacks and Whites were affected by c-MetS. The groups were members of the Hypertension Genetic Epidemiology Network (HyperGEN), a part of the Family Blood Pressure Program, supported by the NHLBI. Although the c-MetS definition is of special interest in particular to the clinicians, the quantitative latent traits of the metabolic syndrome (MetS) are also important in order to gain further understanding of its etiology. In this study, quantitative evaluation of the MetS latent traits (q-MetS) was based on the statistical multivariate method factor analysis (FA). Results The prevalence of the c-MetS was 34% in Blacks and 39% in Whites. c-MetS showed predominance of obesity, hypertension, and dyslipidemia. Three and four factor domains were identified through FA, classified as "Obesity," "Blood pressure," "Lipids," and "Central obesity." They explained approximately 60% of the variance in the 11 original variables. Two factors classified as "Obesity" and "Central Obesity" overlapped when FA was performed without rotation. All four factors in FA with Varimax rotation were consistent between Blacks and Whites, between genders and also after excluding type 2 diabetes (T2D) participants. Fasting insulin (INS) associated mainly with obesity and lipids factors. Conclusions MetS in the HyperGEN study has a compound phenotype with separate domains for obesity, blood pressure, and lipids. Obesity and its relationship to lipids and insulin is clearly the dominant factor in MetS. Linkage analysis on factor scores for components of MetS, in familial studies such as HyperGEN, can assist in understanding the genetic pathways for MetS and their interactions with the environment, as a first step in identifying the underlying pathophysiological causes of this syndrome.

Background Metabolic and physiologic disorders for cardiovascular disease (CVD) and type 2 diabetes (T2D), including abdominal obesity, insulin resistance, hyperglycemia, dyslipidemia, and hypertension often cluster. This cluster is frequently identified as the "metabolic syndrome" (MetS). Reaven [ 1 ] related MetS to the presence of resistance to insulin-mediated glucose disposal, glucose intolerance, hyperinsulinemia, increased triglycerides, decreased high-density lipoprotein cholesterol, and hypertension. Later, the definition of MetS was extended to include obesity, inflammation, microalbuminuria, and abnormalities of fibrinolysis and of coagulation [ 2 - 4 ]. Clearly, insulin resistance is not considered equivalent to MetS [ 5 , 6 ]. Grundy et al. [ 7 ], at a recent National Heart, Lung, and Blood Institute (NHLBI) /American Heart Association (AHA) National Conference, concluded that abdominal obesity is strongly associated with MetS. Sonnenberg et al. [ 8 ] have hypothesized that increased adipose tissue mass contributes to the development of MetS by triggering an increase in proinflammatory adipokines, especially the tumor necrosis factor-α, which may play a role in the pathogenesis of dyslipidemia, insulin resistance, hypertension, endothelial dysfunction, and atherogenesis. Although several studies have targeted MetS, its genetic determination and its pathophysiology remain unclear [ 9 ]. Different definitions and multivariate statistical approaches have been applied to characterize the increasing high-risk MetS premorbid state. Recently, special attention has received the categorical definition of metabolic syndrome (c-MetS) of the National Cholesterol Education Program Adult Treatment Panel III (NCEP) [ 2 ]. The NCEP definition (see Material and Methods) has especially two components, its usefulness in the clinical diagnosis of MetS and its association with recommendations for its therapeutic treatment. Based on the NCEP c-MetS definition, it is reported that 20 to 25 percent of the U.S. adult population has MetS. This represents a high prevalence of the syndrome in the general population [ 10 , 11 ]. In addition, employing the multivariate statistical method factor analysis (FA) different studies in different sampled populations have documented the underlying latent traits of MetS [ 4 , 12 - 17 ]. Meigs [ 3 ] has reported that FA in different studies has yielded on average 2 to 4 latent traits (factors) of MetS. Different studies have found different numbers of latent factors, depending on the type and number of the original risk factors analyzed, sampled population(s), methods utilized, including the statistical rotation method, and decisions about how many factors appeared statistically meaningful. The objective of this study was to exemplify important facets of the MetS in the HyperGEN study. Two MetS aspects were assessed: a. The trait characterized as the categorical MetS (c-MetS) was studied according to the NCEP definition; b. The underlying (latent) traits or clusters of MetS (q-MetS) were evaluated by performing FA with and without Varimax rotation on 11 risk factors. All data were grouped by ethnicity and gender. Subgroups were created by excluding T2D participants, under the assumptions that T2D individuals may have a different pattern of glucose and insulin levels. Finally, our goal was to compare the expression of c-MetS and q-MetS in the Hypertension Genetic Epidemiology Network (HyperGEN) study. Results Sample size and relationships among original risk factors For c-MetS the sample sizes varied from 2,025 observations for fasting triglycerides (TG) to 2,300 for high density lipoprotein (HDL) cholesterol in Blacks, and from 2,171 observations for TG to 2,471 for HDL in Whites. In the HyperGEN study, a high percentage of individuals have body waist (WAIST) and systolic blood pressure (SBP) / diastolic blood pressure (DBP) above the NCEP thresholds (Figure 1 ). Whites tended to have greater percentages of participants with TG and HDL beyond the NCEP thresholds than Blacks. The prevalence of c-MetS was 34 and 39 percent in Blacks and Whites, respectively. Figure 1 Categorical MetS (c-MetS) in the HyperGEN Study For q-MetS, the sample sizes and variables studied are summarized in Table 1 (the statistics were similar when T2D subjects were excluded, results not shown). After participants with missing data for any of the 11 original variables were excluded, this resulted in a sample of 1,422 Blacks and 1,470 Whites with complete data. The samples reduced to 1,173 Blacks and 1,322 Whites when T2D participants were excluded. Table 1 Original Data Included in Factor Analysis Blacks (N = 1422) Whites (N = 1470) Variable Units Kurtosis Skewness Mean Std Dev Kurtosis Skewness Mean Std Dev BMI kg/m 2 0.51† 0.72 32.04‡ 7.54 0.67 0.77 28.86 5.57 GLUC mg/dl 0.68 0.70 107.37 44.06 1.43 -0.55 100.80 26.38 INS μU/ml 0.34 0.06 10.55 9.24 0.42 0.23 7.45 5.90 LDL mg/dl 0.31 0.33 118.89 36.63 0.36 0.26 116.41 31.56 HDL mg/dl 0.46 0.22 53.58 15.17 0.17 0.18 48.70 14.09 TG mg/dl 0.41 0.35 101.32 56.03 0.17 0.20 144.77 75.24 SBP mm Hg 0.46 0.68 128.45 21.78 0.44 0.62 120.50 18.33 DBP mm Hg 0.85 0.60 74.08 11.59 0.42 0.44 68.96 9.94 WAIST cm 0.51 0.64 102.05 17.78 0.74 0.71 99.22 15.48 WHR ratio 0.11 0.10 0.90 0.08 0.17 0.05 0.92 0.09 %BF % 0.44 0.17 40.00 12.06 0.33 0.57 33.41 9.29 †Kurtosis and skewness are reported after the data were transformed (where necessary) and adjusted (see Material and Methods); ‡number of observations, mean, and standard deviations represent measures from the final sample in factor analysis In terms of the participants with complete data, the age at clinic visit had a mean of 46 and a standard deviation of 13 years in Blacks, and a mean of 51 and a standard deviation of 14 years in Whites. Overall, when compared to Whites, Blacks tended to have a higher body mass index (BMI), fasting plasma glucose (GLUC), fasting insulin (INS), HDL, SBP, DBP, WAIST, and percent body fat (%BF), similar low density lipoprotein (LDL) cholesterol and waist to hip ratio (WHR), and lower TG. Kurtosis after adjustments varied from 0.11 for WHR to 0.85 for DBP in Blacks, and 0.17 for WHR to 1.43 for GLUC in Whites, which demonstrates normal distributions for the traits in study and also for the factors created by performing FA (see Material and Methods). Pearson correlations among the 11 adjusted and normally distributed variables are presented in Table 2 . The lower triangle correlations correspond to all data, whereas the upper triangle refers to the data excluding T2D participants. In both ethnicities, strong correlations were observed among BMI, WAIST, %BF, INS, and WHR. GLUC correlated significantly with INS (inversely, because GLUC was inversely squared transformed). HDL cholesterol had a significantly negative correlation with TG and INS. SBP was significantly correlated with DBP. Interestingly, WAIST correlated higher with BMI (about 0.9) than with WHR (about 0.7) in all groups. These inter-correlations determine the structure of the factors. Table 2 Correlation Matrix of the Variables Included in Factor Analysis Variables ** Blacks: Upper Triangle: All Data Excluding T2D (N = 1422 / 1173) BMI INS GLUC LDL HDL TG SBP DBP WAIST WHR %BF Lower Triangle:All Data BMI 0.56‡ - 0.33‡ 0.15‡ - 0.22‡ 0.15‡ 0.18‡ - 0.04 0.91‡ 0.45‡ 0.78‡ INS 0.51‡ - 0.50‡ 0.16‡ - 0.37‡ 0.32‡ 0.04 - 0.08† 0.57‡ 0.44‡ 0.49‡ GLUC - 0.29‡ - 0.40‡ - 0.16‡ 0.25‡ - 0.26‡ - 0.06 0.03 - 0.33‡ - 0.26‡ - 0.28‡ LDL 0.13‡ 0.11‡ - 0.11‡ - 0.18‡ 0.23‡ - 0.00 - 0.05 0.14‡ 0.11† 0.17‡ HDL - 0.20‡ - 0.35‡ 0.23‡ - 0.16‡ - 0.41‡ 0.02 0.06 - 0.25‡ - 0.25‡ - 0.20‡ TG 0.17‡ 0.32‡ - 0.31‡ 0.20‡ - 0.41‡ 0.05 - 0.01 0.20‡ 0.28‡ 0.16‡ SBP 0.17‡ 0.02 - 0.07* 0.02 0.05 0.06 0.76‡ 0.16‡ 0.12† 0.06 DBP - 0.04 - 0.08† 0.04 - 0.02 0.08† - 0.00 0.74‡ - 0.03 0.03 - 0.06 WAIST 0.90‡ 0.52‡ - 0.32‡ 0.14‡ - 0.23‡ 0.22‡ 0.15‡ - 0.03 0.70‡ 0.75‡ WHR 0.44‡ 0.42‡ - 0.32‡ 0.10† - 0.25‡ 0.30‡ 0.11‡ 0.02 0.69‡ 0.407‡ %BF 0.76‡ 0.44‡ - 0.24‡ 0.16‡ - 0.18‡ 0.16‡ 0.03 - 0.06* 0.74‡ 0.37‡ Variables Whites: Upper Triangle: All Data Excluding T2D (N = 1470 / 1322) BMI INS GLUC LDL HDL TG SBP DBP WAIST WHR %BF Lower Triangle:All Data BMI 0.55‡ - 0.30‡ 0.01 - 0.21‡ 0.20‡ 0.19‡ 0.07* 0.89‡ 0.46‡ 0.77‡ INS 0.52‡ - 0.30‡ - 0.01 - 0.35‡ 0.34‡ 0.22‡ 0.15‡ 0.52‡ 0.36‡ 0.45‡ GLUC - 0.32‡ - 0.31‡ - 0.08* 0.13‡ - 0.16‡ - 0.12† - 0.09* - 0.30‡ - 0.25‡ - 0.26‡ LDL 0.02 - 0.02 - 0.05 - 0.04 0.08* 0.05 0.07 0.04 0.08* 0.07* HDL - 0.22‡ - 0.37‡ 0.21‡ - 0.02 - 0.43‡ 0 0.01 - 0.20‡ - 0.21‡ - 0.12† TG 0.20‡ 0.33‡ - 0.21‡ 0.06* - 0.45‡ 0.13† 0.09† 0.22‡ 0.23‡ 0.20‡ SBP 0.17‡ 0.19‡ - 0.13‡ 0.04 0.00 0.10† 0.70‡ 0.15‡ 0.13‡ 0.09† DBP 0.03 0.09† 0.00 0.06 0.01 0.06 0.67‡ 0.06 0.09† 0.03 WAIST 0.89‡ 0.49‡ - 0.33‡ 0.04 - 0.21‡ 0.23‡ 0.14‡ 0.02* 0.70‡ 0.77‡ WHR 0.46‡ 0.35‡ - 0.25‡ 0.06 - 0.23‡ 0.25‡ 0.11† 0.07* 0.69‡ 0.44‡ %BF 0.76‡ 0.43‡ - 0.25‡ 0.08† - 0.12‡ 0.18‡ 0.11† 0.01 0.76‡ 0.43‡ *p < 0.05 † < 0.01 ‡ < 0.0001 **Variables were adjusted for age and center within ethnicity and gender (see Material and Methods) A negative correlation between GLUC and INS is result of the inverse squared transformation of the original GLUC FA with no rotation identified a factor (Factor 1) that was loaded mostly by central obesity, obesity risk factors and INS (Table 4, see Additional file 1 ). It explained 21 percent to 32 percent of the variance of the original risk factors. Three other factors, "Obesity," "Blood pressure," and "Lipids," were identified. In Blacks, the "Obesity" factor was primarily loaded by BMI, INS, WAIST, and %BF. In Whites, exclusion of T2D participants led to a similar pattern. However, in all data in Whites, the first factor represented a stronger mixture of central obesity, obesity and INS, leaving the fourth factor with mainly WHR loading. In order to distinguish the second factor from the first one, we labeled the first as "Central obesity" factor and the second as "Obesity" factor. In both ethnicities, blood pressure (BP) gave rise to a separate factor. Also the "Lipids" remained as a separate factor and was dominated mainly by HDL and TG. INS was associated mostly with the "Obesity" and "Lipids" factors. In the case of FA with Varimax rotation, again, 4 distinct factors explained about 60 percent of the variance in the original variables (Table 4, see Additional file 1 ). We are labeling them as "Obesity," "BP," "Lipids," and "Central obesity". Factor loadings less than 0.1 were not listed in Tables 3 and 4 (see Additional file 1 ). The first factor alone explained 23 percent to 25 percent of the variance in the original risk factors, while the fourth factor explained 7 percent to 9 percent of the variance. The "Obesity" factor (Factor 1) loaded mainly BMI, WAIST, WHR, %BF, and INS. SBP and DBP loaded separately. A distinct factor ("Lipids") loaded mainly HDL cholesterol, TG, INS and GLUC in Blacks and HDL cholesterol, TG and INS in Whites. The fourth factor contained a higher loading for WHR than for WAIST. Similar factor loadings were present in the samples when T2D participants were excluded (Table 4, see Additional file 1 ). Table 3 Factors, Loadings, and Sums of Squared Loadings in All Data (Males (M) and Females (F)), and by Gender (M, F) (Varimax Rotation) Factor 1 (Obesity) Sample BMI* %BF WHR WAIST INS GLUC SBP DBP LDL HDL TG SS Loadings Blacks M+F (1422) 0.95† 0.71 0.32 0.86 0.42 -0.20 0.13 2.49 M (483) 0.92 0.71 0.65 0.95 0.46 -0.15 0.11 0.11 -0.21 0.14 2.91 F (939) 0.96 0.71 0.23 0.84 0.40 -0.22 0.14 2.43 Whites M+F (1470) 0.94 0.77 0.40 0.86 0.46 -0.26 0.14 -0.11 0.11 2.69 M (721) 0.94 0.76 0.55 0.93 0.48 -0.26 -0.13 0.16 2.99 F (749) 0.94 0.81 0.31 0.84 0.46 -0.28 0.18 -0.11 0.11 2.69 Factor 2 (BP) Sample BMI %BF WHR WAIST INS GLUC SBP DBP LDL HDL TG SS Loadings Blacks M+F (1422) 0.99 0.76 1.58 M (483) 0.13 0.83 0.83 1.42 F (939) 0.99 0.77 1.59 Whites M+F (1470) 0.13 0.88 0.76 1.39 M (721) 0.14 0.87 0.80 1.41 F (749) 0.12 -0.14 0.98 0.65 0.11 1.45 Factor 3 (Lipids) Sample BMI %BF WHR WAIST INS GLUC SBP DBP LDL HDL TG SS Loadings Blacks M+F (1422) 0.22 0.19 0.31 0.24 0.52 -0.43 0.27 -0.57 0.64 1.49 M (483) 0.17 0.12 0.29 0.18 0.48 -0.19 -0.72 0.65 1.37 F (939) 0.18 0.11 0.29 0.21 0.49 -0.41 0.32 -0.55 0.59 1.32 Whites M+F (1470) 0.23 0.17 0.27 0.21 0.51 -0.27 -0.68 0.64 1.41 M (721) 0.14 0.41 -0.18 -0.18 -0.75 0.61 1.20 F (749) 0.25 0.19 0.24 0.23 0.5 -0.34 0.11 0.22 -0.70 0.60 1.44 Factor 4 (Central Obesity) Sample BMI %BF WHR WAIST INS GLUC SBP DBP LDL HDL TG SS Loadings Blacks M+F (1422) 0.90 0.40 0.15 -0.13 0.11 1.03 M (483) -0.19 0.37 0.11 0.2 F (939) 0.93 0.44 0.15 -0.16 0.11 1.13 Whites M+F (1470) 0.16 0.69 0.47 -0.10 0.76 M (721) 0.11 0.57 0.27 -0.11 0.13 0.45 F (749) 0.92 0.44 0.14 1.10 * Variables were adjusted for age and center within ethnicity and gender (see Material and Methods) GLUC negative loadings are result of inverse squared power transformation of the original GLUC;† Loadings ≥ 0.4 are in bold FA with Varimax rotation was performed also by gender (Table 3 ). Between genders, the factors loaded in a similar fashion in Blacks and Whites. WHR and WAIST reflected gender differences in their loadings in factors 1 and 4. Discussion The fact that MetS is more prevalent in the HyperGEN study as compared to the average of the US adult population [ 10 , 11 ], is probably due to selection bias arising from the hypertension selection criterion applied in HyperGEN (see Material and Methods). Although BP was an important contributor to the categorical definition of the MetS (c-MetS), other risk factors such as WAIST, HDL and TG were also important. The HyperGEN Whites had a higher prevalence of c-MetS than the Blacks. Blacks had a higher percentage of participants with WAIST, BP, and GLUC beyond the NCEP thresholds. In this study, FA of 11 potential risk factors for CVD and T2D yielded 4 latent factors, explaining about 60 percent of the variance in the original risk factors. Using the maximum likelihood estimation (MLE) method in S-PLUS (Insightful Corporation software), we found that the model p-values were significant, suggesting that additional factors may exist. However, although additional factors must exist to explain approximately 40 percent of the variance in the original variables, none of the remaining factors individually can explain more than about 5 percent of the variance. Therefore, we concluded that the quantitative structure of MetS can be described in terms of three to four factors when no rotation was performed, and four factors with Varimax rotation. One may even argue whether the fourth factor in the Varimax rotation is very meaningful. We chose to retain the "Central Obesity" factor particularly because it tends to reflect the well-known gender asymmetry (Table 3 ). FA without and with Varimax rotation can be useful in different settings. We believe that FA without any rotation makes more sense when investigating the pattern of risk factor clustering in the MetS. On the other hand, gene finding studies can be enhanced with Varimax rotation because, it is easier to find genes each of which influences a different (uncorrelated) factor. Therefore, depending on the goals of a study, rotation may or may not be used. We regard this flexibility as strength of the FA method. We tested the pattern of the factors between genders only for FA with Varimax rotation. This pattern was stable among ethnicities between genders (Table 3 ). WHR on Factors 1 and 4 and WAIST on Factor 4 had statistically different loadings in males and females. Another characteristic of the HyperGEN study was that at least two participants in each sibship had hypertension. A large proportion of the hypertensive participants have used anti-hypertensive and anti-cholesterol medications. Maison et al. [ 13 ] have compared medicated and unmedicated groups to see any implications of the medication use in FA. They applied FA to 9 risk factor changes over time, and separated data into groups of treated and untreated for hypertension and dyslipidemia. They found 3 and 4 factors respectively in males and females, the "BP", "Glucose," "Lipid," and "BMI, WHR, INS and TG" factors, which were similar between treated and untreated groups. It is a common belief that T2D participants may have a different expression of INS and GLUC, therefore it may influence also the factors pattern in the MetS analysis. In the present study, we found a consistency of the factors before and after removing type 2 diabetics. This finding is supported also by Hanson et al. [ 17 ] who studied two samples of Pima Indians classified as T2D and non-T2D. They identified consistently 4 latent factors out of 10 risk factors in the two samples, with a relative variation only on insulin loadings. They found that the "Insulinemia" factor explained 30 percent of the original variance, "Body size" 20 percent, "BP" 15 percent, and "Lipids" 14 percent. In our study, the INS variable loaded with BMI, WAIST, %BF, and also with lipids. INS was present mainly in 2 and sometimes in 3 factors with loadings mostly in the "Obesity" and "Lipids" factors. Other studies have provided similar results about the latent traits of MetS [ 4 , 16 , 17 ]. FA studies cited here, and other studies described by Meigs [ 3 ], have elements in common with our study: 2 to 4 factors were identified; BMI, INS, WHR and WAIST are major contributing risk factors; SBP and DBP load in a separate factor; INS was associated with more than one factor and mainly with obesity and lipids. Our study and several others have shown that FA is a useful method for studying the underlying traits of MetS. Nevertheless this methodology has not passed without been criticized. Lawlor et al. [ 18 ] reviewed 22 published studies of the MetS, all based on FA. None of the studies had clearly identified whether they used FA for exploratory analysis or for hypothesis testing purposes. Such ambiguous use was regarded by Lawlor et al [ 18 ] as a major misuse of FA in the study of the metabolic syndrome. In fact exploratory FA and the hypothesis-testing (confirmatory) FA are two distinct methods. Basically, the two analyses use different constraints and different approaches in the analytical software. The exploratory FA is driven by the data (example is our study). On the other hand, the hypothesis-testing FA is only performed with some prior knowledge of possible loadings for different risk factors. One applies confirmatory FA on the data to test if the factor structure of the hypothesized model specifying the interrelationships among the original variables and the latent factors included in the model is true or not. A detailed example of the confirmatory FA of the metabolic syndrome is provided by Shen et al. [ 16 ]. If c-MetS and q-MetS are explaining the same disorder in two different aspects in the HyperGEN study, can FA contribute to MetS gene finding? MetS is recognized as a precursor for cardiovascular disease and type 2 diabetes [ 19 ]. There are several studies that have used FA for understanding the complexity of the MetS. Our study brings in more evidence that FA provides not only insights about the latent factor traits for MetS, but it produces factor scores for each of the MetS domains at the same time. Can factor scores be used in genetic analysis, such as in linkage analysis? The concept of a latent factor has much (intellectual) parallel with the concept of a latent gene. Much like a latent gene might have pleiotropic effects on several correlated phenotypes (original risk factors), several correlated risk factors load onto a latent factor. This makes FA very attractive from a genetic analysis point of view since, unlike individual risk factors each of which may entail several genes, each factor is likely to involve only a few genes which simplify their discovery. We believe that FA is useful for complex disease gene finding. In a near future motivated from this analysis we plan in the HyperGEN study to perform linkage analysis on the trait established by c-MetS and also on factor scores created by q-MetS, for identifying essential MetS putative genes / QTLs. Conclusions These analyses demonstrated that obesity and hypertension were the most important factors contributing to the MetS in the HyperGEN Study. Three to four distinct factor domains were identified depending on the FA rotation applied and decisions made. Results support the hypothesis that MetS is a compound phenotype, where obesity and its relationship to lipids and insulin are clearly the driving force of MetS. Insulin may play a connecting role between obesity and lipid domains. In genetic analysis, it is well known that categorical data, especially a complex trait such as MetS, encounter reduced power as compared to quantitative variables. Therefore, we suggest that genetic analysis should be performed on specific combinations of traits that belong to a factor. It is possible that some common genes may exist in the pathways for the factors identified. Linkage analysis investigating putative quantitative trait loci for MetS factor domains can be a first step which may help discover the underlying mechanisms, or generate new hypotheses, in finding the causes of MetS. Material And Methods Data collection and MetS definition The sample represents data from the HyperGEN network, part of the Family Blood Pressure Program, supported by the NHLBI as described by Williams et al. [ 20 ] and Province et al. [ 21 ]. The ethnicity was recorded as a self-reported demographic category. In the HyperGEN study sibships were recruited, each with at least 2 members who were hypertensive before age 60. Also, parents and offspring of some of the hypertensive sibs, as well as random samples of unrelated Blacks and Whites, were recruited, totaling 4,781 participants. Insulin measurements are not available in a part of the sample and therefore the sample size was smaller for FA. Also, participants with missing values for any of the quantitative risk factors used in the definition of MetS were excluded. A detailed account is provided in the Results section. A participant was classified as having T2D if (s)he had a fasting plasma glucose value ≥ 126 mg/dl, or is a current user of hypoglycemic medication or insulin that was documented at examination in the clinic, or if diabetes was reported in the HyperGEN questionnaire. Also, an age at diagnosis ≥ 40 years was required for T2D individuals [ 22 ]. c-MetS according to the NCEP definition, was identified in participants by the simultaneous presence of 3 or more of the following conditions: WAIST > 102 cm in men, and > 88 cm in women; TG ≥ 150 mg/dl; high density lipoprotein (HDL) < 40 mg/dl in men, and < 50 mg/dl in women; systolic blood pressure (SBP) ≥ 130 mm Hg and/or diastolic blood pressure (DBP) ≥ 85 mm Hg or using antihypertensive medications; GLUC ≥ 110 mg/dl or on treatment for diabetes [ 2 ]. Factor analysis was founded on 11 variables: BMI expressed as the ratio of body weight divided by body height squared (in kg/m 2 ); WAIST measured at the level of the umbilicus in cm; WHR defined as waist circumference divided by hip circumference; GLUC in mg/dl; INS in μU/ml (where fasting time was defined as ≥ 12 hours before blood draw); LDL in mg/dl; HDL in mg/dl; TG in mg/dl; Sitting SBP in mm Hg; DBP in mm Hg (SBP and DBP were measured three times after the subject was asked to sit for five minutes, with the mean of the second and third measurements of each variable being used in the analysis); %BF derived from the bioelectric impedance measurements based on the Lukaski formula [ 23 ]. Statistical Analysis TG and INS had skewed distributions. A relatively skewed distribution was also present for HDL. Log transformation brought these variables distributions to approximately normal. GLUC and %BF were highly kurtotic. Using Box-Cox transformation, it was found that the inverse of the squared transformation of GLUC (1/GLUC 2 ) and the squared transformation of %BF (%BF 2 ) reduced the excess kurtosis considerably. The procedure transreg in SAS (version 9 for PC) was employed for finding power transformations. There were two field centers recruiting Blacks and four field centers recruiting Whites. Accordingly, dummy (0,1) field center variables, one for Blacks and three for Whites, were created. All 11 risk factors were adjusted within ethnicity and gender for age, age 2 , age 3 , and field center effects using stepwise regression analysis within ethnicity and gender by employing SAS (SAS version 8.2 for Linux). Any variables with outliers beyond ± 4 standard deviations (SD) were also adjusted for heteroscedasticity of the variance. After the adjustments for each variable, outliers beyond ± 4 SD were eliminated. Each final adjusted variable was standardized to a mean 0 and variance 1. Prevalence of c-MetS was estimated with the FREQ procedure of SAS. The multivariate method of factor analysis was employed for reducing a group of risk factors to a subset of latent factors. The primary goal of FA is to identify the interrelationships among a set of variables. In this study FA was used for exploratory analysis, because there was no a priori information about the structure underlying the variables. FA can be used also for a confirmatory analysis, when validation (or refutation) of a postulated structure is sought. In either case, FA seeks parsimony by summarizing a large group of interrelated variables (risk factors for a complex disease such as MetS) in terms of a small number of latent factors, thereby reducing the dimensionality. Theoretical statistical descriptions of FA can be found in the literature [ 24 - 26 ]. FA was performed with S-PLUS 6.0.1 software by using the factanal function, in which the MLE was employed. FA evaluated latent factors underlying the 11 original variables. FA was performed with and without the Varimax rotation. "No rotation" achieves the simplest latent factor structure, in the extreme case loading any variable in one of the factors and almost negligible loadings in the rest of the factors. That is the reason why some studies (extracting factors with no rotation) find a concentration of the major variables' loading on the first factor. This is also the reason why some investigators named the first factor in their studies as the "Metabolic syndrome" factor [ 3 , 27 ]. Conversely, when Varimax rotation is applied, the objective is to maximize the independence of the clusters for variables that load onto factors. This is achieved by loading in separate factors distinct combinations of the interrelated risk factors. A loading of 0.4 or larger was considered as a significant contribution of an original variable to a factor. List of abbreviations used MetS, metabolic syndrome; c-MetS, categorical MetS; q-MetS, latent traits of MetS; FA, Factor analysis; MLE, maximum likelihood estimate; T2D, type 2 diabetes; CVD, cardiovascular disease; BMI, body mass index; INS, fasting insulin; GLUC, fasting glucose; WHR, waist to hip ratio; SBP, systolic blood pressure; DBP, diastolic blood pressure; BP, blood pressure; TG, fasting triglycerides; LDL, low density lipoprotein cholesterol; HDL, high density lipoprotein cholesterol; %BF, percent body fat. Competing interests The author(s) declare that they have no competing interests. Authors' contributions All authors were equally involved in designing the MetS study, evaluating statistics, interpreting the data, writing the manuscript, and organizing the figure and tables. Supplementary Material Additional File 1 Table 4. This table contains information on factor loadings result of FA with and without rotation performed on 11 risk factors Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549210.xml

546335

Opposing Fat Metabolism Pathways Triggered by a Single Gene

Regulating metabolism of fat is an important challenge for any animal, from nematodes to humans. Central players in the regulatory network are the nuclear hormone receptors (NHRs), which are transcription factors that turn on or off a set of target genes when bound by specific lipid molecules. NHR genes number 48 in mammals, and a surprising 248 in nematodes. Despite the difference in quantity, there are some structural similarities between NHRs in these two groups, in particular, between the nematode gene nhr-49 and the mammalian HNF4 . In this issue, Keith Yamamoto and colleagues show that nhr-49 controls two different aspects of fat metabolism, which interact to form a feedback system controlling the consumption and composition of fats in the nematode. Regulation of fat content and lifespan in C. elegans Using RNAi to suppress gene expression, the researchers discovered that when nhr-49 was absent, the lifespan of the nematode was reduced by more than 50%, and the animal displayed numerous gross abnormalities in the gut and gonad. This was accompanied by unusually high fat content in the larvae. By using quantitative PCR to measure output of fat and glucose metabolism genes, the researchers showed that deletion of nhr-49 changed expression of 13 of these genes, with the most dramatic effects occurring within two metabolic pathways: mitochondrial lipid oxidation and fatty acid desaturation. Oxidation degrades lipids to release energy, explaining the build-up of fat in nhr-49 -suppressed larvae. One of nhr-49 's normal functions is to increase expression of the mitochondrial acyl–Coenzyme A (CoA) synthetase gene acs-2 . A principal role for mitochondrial acyl-CoA synthetases is to “activate” free fatty acids for transport into mitochondria, where they are oxidized. This process involves attaching a CoA group to a free fatty acid, and often serves as a rate-limiting step in lipid oxidation. Indeed, the authors found that suppression of acs-2 alone was sufficient to reproduce the highfat phenotype, while overexpression of acs-2 rescued the phenotype even in the absence of nhr-49 . Fatty acid desaturation is the process of converting saturated fats into unsaturated ones, by forming one or more double bonds between adjacent carbons in the tail. This process is catalyzed by fatty acid desaturase enzymes. nhr-49 increases expression of several desaturases, most importantly fat-7 , which converts stearic acid to oleic acid; deletion of nhr-49 more than doubled the proportion of stearic acid compared to oleic acid. RNAi interference of fat-7 alone produced two interesting results. First, it shortened the nematode life span, suggesting this was the primary pathway through which nhr-49 suppression exerted that same effect. Second, it produced some effects that were opposite those of nhr-49 suppression: specifically, it reduced rather than increased fat content, and it increased rather than reduced expression of acs-2 . These results show that in its normal actions, nhr-49 sets in motion two opposing pathways: it increases acs-2 , which leads to reduction of fat content, and it increases fat-7 , which, by reducing acs-2 , increases fat content. Surprisingly, this behavior links nhr-49 most closely not to HNF4 , with which it shares the most structural similarity, but to another type of mammalian NHR, called peroxisome proliferator-activated receptors (PPARs). Further investigation of this link may lead to better understanding of the functions of PPARs, and provide opportunities for altering their function for treatment of fat metabolism disorders such as diabetes and obesity.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546335.xml

548683

FAM20: an evolutionarily conserved family of secreted proteins expressed in hematopoietic cells

Background Hematopoiesis is a complex developmental process controlled by a large number of factors that regulate stem cell renewal, lineage commitment and differentiation. Secreted proteins, including the hematopoietic growth factors, play critical roles in these processes and have important biological and clinical significance. We have employed representational difference analysis to identify genes that are differentially expressed during experimentally induced myeloid differentiation in the murine EML hematopoietic stem cell line. Results One identified clone encoded a previously unidentified protein of 541 amino acids that contains an amino terminal signal sequence but no other characterized domains. This protein is a member of family of related proteins that has been named family with sequence similarity 20 (FAM20) with three members (FAM20A, FAM20B and FAM20C) in mammals. Evolutionary comparisons revealed the existence of a single FAM20 gene in the simple vertebrate Ciona intestinalis and the invertebrate worm Caenorhabditis elegans and two genes in two insect species, Drosophila melanogaster and Anopheles gambiae . Six FAM20 family members were identified in the genome of the pufferfish, Fugu rubripes and five members in the zebrafish, Danio rerio . The mouse Fam20a protein was ectopically expressed in a mammalian cell line and found to be a bona fide secreted protein and efficient secretion was dependent on the integrity of the signal sequence. Expression analysis revealed that the Fam20a gene was indeed differentially expressed during hematopoietic differentiation and that the other two family members (Fam20b and Fam20c) were also expressed during hematcpoiesis but that their mRNA levels did not vary significantly. Likewise FAM20A was expressed in more limited set of human tissues than the other two family members. Conclusions The FAM20 family represents a new family of secreted proteins with potential functions in regulating differentiation and function of hematopoietic and other tissues. The Fam20a mRNA was only expressed during early stages of hematopoietic development and may play a role in lineage commitment or proliferation. The expansion in gene number in different species suggests that the family has evolved as a result of several gene duplication events that have occurred in both vertebrates and invertebrates.

Background Hematopoietic differentiation is a complex process whereby multiple functionally and morphologically distinct cell types arise from a population of pluripotent hematopoietic stem cells (PHSCs) [ 1 ]. The accurate and efficient regulation of hematopoietic development is controlled by a large number of regulatory proteins that have been identified over the past few decades. These regulatory molecules include the hematopoietic growth factors (HGFs), soluble proteins that recognize specific receptors on the surface of sub-populations of hematopoietic cells, thereby initiating signal transduction pathways that modulate the differentiation, proliferation, and/or survival of target cells [ 2 ]. The identification of regulators of hematopoiesis has been an ongoing effort for many years and has benefited from the existence of accessible cell line models as well as the characterization of genes affected by somatic mutations associated with specific human leukemias [ 3 ]. We have used a pair of factor-dependent murine cell lines to identify novel genes expressed within distinct hematopoietic lineages as an approach to the identification of novel candidate genes for development of diagnostic and therapeutic approaches to leukemia. The EML and MPRO cell lines were both established by infecting murine bone marrow cells with a retrovirus expressing a dominant negative retinoic acid receptor α (RARα) protein [ 4 , 5 ]. The infected cells were selected in the presence of either stem cell factor (SCF) or granulocyte/macrophage colony stimulating factor (GM-CSF). EML are SCF-dependent and resemble uncommitted hematopoietic progenitor cells. They can be induced to differentiate to the promyelocyte stage of granulopoiesis in the presence of interleukin-3 (IL-3) and high doses of all trans retinoic acid (atRA) [ 4 , 6 ]. Terminal neutrophil differentiation of EML cells can be induced by replacement of IL-3 and SCF with GM-CSF. MPRO cells are GM-CSF-dependent and can be induced to differentiate to neutrophils by adding high doses of atRA to the culture medium. The expression patterns of a number of genes expressed during hematopoiesis have been examined in EML and MPRO cells and generally agree with the patterns observed in other cell systems and in primary hematopoietic cells. Thus, EML and MPRO provide a powerful system for the identification and characterization of novel genes expressed within the hematopoietic lineage. We have employed the representation difference analysis technique [ 7 ] to identify cDNAs representing genes expressed at higher levels in EML cells 72 hours after induction of differentiation than in uninduced cells. We describe the identification of a clone derived from an uncharacterized putative secreted protein. We have performed a comparative genomics analysis and determined that this protein is the founding member of an extended family of highly related proteins. This family contains three members in mammalian species, one or two members in invertebrate or simple vertebrate species and five or six members in fish. We have determined that one family member is a secreted glycoprotein and describe the expression pattern of the human and mouse genes in tissues and during hematopoietic differentiation. Results Identification of differentially expressed genes by representational difference analysis (RDA) Total RNA was prepared from EML cells grown in the presence of SCF alone (0 hour) or in medium supplemented with IL-3 and atRA for 72 hours. The RNA was converted to cDNA and subjected to three rounds of RDA as previously described [ 6 ]. Six differentially-expressed clones were identified [ 6 , 8 ]. Clone number 1623 was chosen for further analysis and the differential expression of this gene was confirmed by Northern blot analysis. 1623 mRNA was essentially undetectable in the 0 hour sample but readily detectable in the 72 hour sample (data not shown and see figure 10 ). Sequence analysis of clone 1623 The initial cDNA isolated by RDA was a 273 bp fragment that appeared to contain the coding sequence of the C-terminus of a protein that was not present at that time in public databases. To identify the full open reading frame of this cDNA, we first performed rapid amplification of cDNA ends (RACE) in both the 5' and 3' directions. Extension in the 3' direction revealed the presence of a consensus polyadenylation signal located 154 nucleotides downstream of the putative translation stop codon. Extension in the 5' direction yielded an additional 443 bp of sequence containing a contiguous ORF. Comparison of this extended sequence to public databases identified a cDNA (NM_017565) that was identical to clone 1623 in the region of overlap. This cDNA was isolated from a mouse mammary tumor but no functional analysis had been performed. We designed PCR primers based on the published sequence and confirmed that the cDNA isolated from EML cells was identical to the published sequence. The full length ORF was 1623 bp in length and encoded a protein of 541 amino acids (Figure 1A ). The protein did not contain any recognizable motifs when examined using domain mapping software such as SMART or Profilescan (see Methods). However, a putative amino terminal signal sequence was identified using the SignalP analysis program. The cDNA mapped to an 11 exon gene located on mouse chromosome 11E1 (Figure 1B ). The gene spanned approximately 60,000 bp of genomic sequence and the relatively large size of the gene is primarily due to the fact that the first intron is greater than 44,000 bp in length (Figure 1C ). Figure 1 Characterization of the full length mouse 1623 (Fam20a) cDNA and genomic sequence. A. The full length cDNA derived from the original RDA clone was isolated using a combination of 5' and 3' rapid amplification of cDNA ends (RACE) procedures, comparisons to public databases, and amplification of putative full length clones by PCR. The full open reading frame was 1623 bp in length and encoded a 541 amino acid protein. The locations of regions conserved within the subsequently identified FAM20 family are indicated using underlines. Eight cysteine residues that are also conserved within the family are indicated in bold and four putative N-glycosylation sites are indicated in red type. B. The distribution of the 11 exons of the mouse Fam20a gene is shown with the exons indicated using numbers. A consensus polyadenylation signal is located downstream of the terminal exon. C. The sizes of the 11 exons and 10 introns of the Fam20a gene are shown. Identification of a family of related genes The full length cDNA and the encoded protein were compared to sequences in public databases. We reported previously a weak similarity to a protein named Fjx1, which is the mouse orthologue of a Drosophila protein named four-jointed [ 8 ]. However, the degree of sequence identity between these proteins was low (16%) and thus the search was extended to include uncharacterized proteins. We first identified two other mouse proteins that displayed significant similarity to, but were distinct from, the query sequence. One protein (Accession number NP_663388 aka Riken C530043G21) was 409 amino acids in length and displayed 27% identity to the query sequence while the second (NP_085042) was a truncated version of a 579 amino acid protein that displayed 40% identity to the query sequence. We have subsequently discovered that these proteins are members of a highly related family and this family has received the official name "family with sequence similarity 20" (FAM20) from the Human Genome Organization Gene Nomenclature Committee. The protein derived from our original 1623 cDNA is named Fam20a in mouse and the other two family members are named Fam20b and Fam20c, respectively. Continued searches of public databases revealed the existence of related proteins in several other species. Each mammalian genome contains genes encoding three members that are orthologous to the three mouse proteins mentioned above. The accession numbers for the relevant cDNAs in human and rat are listed in Table 1 and we have also identified the same number of related sequences in other mammalian genomes, including the pig, cow and dog (data not shown). However, most of these sequences are incomplete and will not be described in detail here. Table 1 Accession numbers for vertebrate FAM20 family members 1 Family Member Human Mouse Rat Fugu Zebrafish Ciona FAM20A NM_017565 NM_153782 XM_221067 (BK001521) SINFRUP00000146987 (BK001515) ND ND FAM20B NM_014864 NM_145413 XM_222770 SINFRUP00000138548 (BK001520) CAI11712 AK115425 FAM20C NM_020223 NM_030565 XM_221975 (BK001522) SINFRUP00000140879 (BK001516) SINFRUP00000141732 (BK001518) SINFRUP00000163431 (BK001517) BK001519 ENSDARP00000009272 ENSDARP00000005688 ENSDARP00000028589 ND 1. Accession numbers refer to nucleotide sequences in GenBank except in the Fugu and Zebrafish listing where the accession numbers refer either to predicted peptides in the Ensembl database (entries beginning SINFRUP or ENSDARP) or Genbank (entry beginning CAI). Entries beginning with BK refer to predicted sequences in the Third Party Annotation database arising from this study. ND: None detected. In order to gain further information concerning the origin of the FAM20 family, we searched for related sequences in several other invertebrate and vertebrate organisms. The ascidian Ciona intestinalis is a model for a basal chordate organism and has emerged as a powerful model for evolutionary and developmental studies [ 9 , 10 ]. In particular, many gene families or subfamilies are represented by single members in C. intestinalis and thus the identification of an orthologue in this organism can provide useful information for evaluating the evolutionary origin of the members of a gene family. Consistent with this concept, we identified a single cDNA and the corresponding genomic locus in C. intestinalis that displayed significant sequence similarity to the mammalian FAM20 genes and proteins (Table 1 ). Complete genome sequences are also available for several invertebrate species and two related sequences were identified in Drosophila melanogaster and Anopheles gambiae with one family member in Caenorhabditis elegans (Table 2 ). Finally, analysis of genomic cDNA and protein databases for the pufferfish ( Fugu rubripes ) and zebrafish ( Danio rerio ) revealed the presence of six and five family representatives, respectively (Table 1 ). The gene numbers in these various species are listed on an idealized evolutionary tree in Figure 2 and suggest that the FAM20 gene family has undergone a complex set of gene duplications in both the invertebrate and chordate lineages. Table 2 Accession numbers for invertebrate FAM20 family members 1 Family Member Drosophila Mosquito C. elegans FAM20A ND ND ND FAM20B NM_170079 NM_206490 EAA08010 EAA13434 NM_078126 Fam20C ND ND ND 1. Accession numbers refer to nucleotide sequences in GenBank. ND: None detected. Figure 2 Evolutionary distribution of FAM20 gene number. An idealized evolutionary tree (modified from [10]) is shown with the number of FAM20 genes identified in several genomes as described in the text. The gene numbers are supportive of a single gene duplication event occurring in invertebrates (at least in insects) and multiple gene duplication events occurring in higher vertebrates. Assignment of subfamily relationships To elucidate the nature of these putative gene duplications, we sought to assign the individual sequences from the various species into subfamilies based on protein sequence and gene structure, specifically using the number and size of exons in the latter case. Initially, the exon distribution of each of the Fam20 members in three mammalian species (human, mouse and rat) was compared and revealed obvious inter and intra orthologue similarities (figure 3A ). Each FAM20A gene contained 11 exons and exon sizes were identical in these three species. Likewise, each FAM20B gene contained 7 exons that were identical in size in human, mouse and rat. The FAM20C genes each contained 10 exons and only exon 1 displayed any variation in size amongst these three species. In intra-orthologue comparisons, the exons in FAM20B and FAM20C genes clearly aligned with exons in the FAM20A genes, with small variations (in multiples of three bases) in the size of the internal exons in FAM20B. FAM20B lacks exons corresponding to exons 2–4 of FAM20A while FAM20C lacks exon 3. In addition, exons 8 and 9 in FAM20A and FAM20C are represented by a single exon in FAM20B that is identical in size to the combined exons in the other two genes. Thus, the three mammalian genes are highly evolutionarily related and presumably are derived from a common ancestral gene. Figure 3 Assignment of FAM20 family members to subfamilies. A. Exon size and distribution of mammalian FAM20 members. The exons within each FAM20 gene in human, mouse and rat are indicated with the number of base pairs indicated within each exon. The sizes of exons that differ in size from the FAM20A genes are indicated. B. A dendrogram showing the relationships between FAM20 proteins from human (Hs), mouse (Mm), rat (Rn), Fugu rubripes (Fr), Danio rerio (Dr), D. melanogaster (Dm), A. gambiae (Ag), C. intestinalis (Ci) and C. elegans . The accession numbers of the cDNA sequences from which each protein sequence was derived are shown in parentheses except in the case of the mosquito family members where the accession number is used as the gene/protein name. Accession numbers for zebrafish peptide sequences are listed in Table 1. The FAM20 nomenclature has not been extended to the invertebrate sequences and the previous gene names have been used for Drosophila and C. elegans family members. The subfamily assignment of each family member is shown on the right. C. Exon number and size distribution within Fugu Fam20 members. The accession number of each sequence within the Third Party Annotation database is shown at left and family assignment based on dendrogram position and exon distribution is shown on the right. To assign the genes identified in other species to these three subfamilies, we performed a global comparison of the peptide sequences derived from 25 of the identified family members listed in Tables 1 and 2 . One zebrafish protein (from FAM20A) was omitted as its sequence is incomplete. A dendrogram showing the results of this comparison is presented in Figure 3B . As expected, the mammalian orthologues clustered together and thus defined the subfamilies. All of the invertebrate proteins and the single protein identified in C. intestinalis clustered with FAM20B proteins, suggesting that this represents the ancestral branch of the FAM20 family. A single protein from Fugu and zebrafish clustered with the FAM20A and FAM20B family members while two Fugu and two zebrafish proteins clustered with FAM20C members. However, two Fugu proteins and one zebrafish protein clustered on a separate branch between FAM20A and FAM20B. In order to determine the subfamily to which these proteins belonged, we made use of the high degree of conservation of exon size and number noted in the mammalian genes (Figure 3C ). The exon number and size of the Fugu and zebrafish genes encoding the two proteins assigned to FAM20A and FAM20B were consistent with their membership in these families. The only variations noted were a slightly larger exon 2 in the Fugu Fam20a gene and the division of exon 1 into two exons in the Fugu Fam20b gene. As in the mammalian family members, the sizes of the terminal exons varied more than the internal exons. The other four Fugu genes displayed exon distributions consistent with membership in FAM20C, despite the clustering of two of the encoded proteins between FAM20A and FAM20B. We have assigned each of these proteins to FAM20C with number suffixes (c1, c2, etc.) to designate individual genes and proteins. Each of these genes maps to distinct genomic loci and thus represents independent genes and not splicing variants of a smaller number of genes (data not shown). The gene structures of the three zebrafish family members were also consistent with this family assignment (data not shown). Comparisons of the derived protein sequences within each subfamily are shown in figures 4 , 5 , 6 . Figure 4 Sequence alignment of FAM20A protein sequences. The complete protein sequences of FAM20A members were compared using the AlignX component of the VectorNTI sequence analysis suite of programs. Identical amino acids are outlined in yellow, and similar residues are indicates in light blue. Conserved regions 1, 2 and 3 are underlined (see below). Gaps are indicated with dashes and the sequences are from human (H), mouse (M), rat (R) and puff erfish (F). Figure 5 Sequence alignment of FAM20B protein sequences. The complete protein sequences of FAM20B members are presented as described in figure 4. The sequences are from human (H), mouse (M), rat (R), pufferfish (F), zebrafish (D) and C. intestinalis (Ci). Figure 6 Sequence alignment of FAM20C protein sequences. The complete protein sequences of FAM20C members are presented as described in figure 4. The sequences are from human (H), mouse (M), rat (R), pufferfish (Fcl-4) and zebrafish (Dcl-3). Features of FAM20 proteins All of the identified FAM20 protein sequences contain putative signal sequences at their amino termini but no other functional domains were unambiguously detected using several different annotation search software programs. In order to search for potential functional domains, we compared the sequences of all family members. These comparisons revealed that the greatest similarity was located within the carboxy-terminal two thirds of each protein (Figure 7A ). We have named this region the conserved C-terminal domain (CCD) and it overlaps with a domain listed in the CDD database at NCBI as DUF1193. The CCD contains three distinct regions that are more highly conserved within all members of the family than the surrounding sequences (named conserved regions 1, 2 and 3 in figure 7A ) and the consensus sequences for each conserved region were derived (figure 7B ). Amino acids that are essentially invariant in all family members have been indicated in bold type and the heptapeptide DRHHYE in CR2 is the longest contiguous sequence that is conserved in all members of the family. A set of eight cysteine residues is also perfectly conserved within the CCD of each family member that may participate in inter-or intramolecular disulphide bond formation. Figure 7 Schematic representation of the structural features of FAM20 family members. A. Structural features of FAM20A showing domains and residues conserved within the entire family. Key: SS: signal sequence; CCD: conserved C-terminal domain; CR: conserved region; Cys: cysteine residues conserved within CCD (indicated with asterisk). B. Consensus sequences were derived for CR1, CR2 and CR3 using a global comparison of all the family members listed in Tables 1 and 2. Residues that are invariant or only differ in one sequence are indicated in bold. Non-conserved residues are indicated with an x and positions with more than one common residue are shown below the main sequence. Fam20a is a secreted protein As the putative signal sequence was the only known domain identified in all family members, we next tested whether this sequence is functional. Signal sequences are commonly found on proteins that are directed to the endoplasmic reticulum (ER) and either retained there or processed and transported into the Golgi apparatus and secreted from the cell. Many proteins are glycosylated during their transit through the ER and Golgi apparatus and the mouse Fam20a protein contains four potential sites for N-glycosylation (indicated in red type in figure 1 ). As Fam20a does not contain an ER retention signal, we predicted that it should be detected in the medium of expressing cells. A mammalian expression vector was constructed that contained the full length mouse Fam20a coding sequence fused to a C-terminal Myc epitope tag and a hexahistidine sequence to permit purification. The plasmid was transfected into monkey kidney COS-1 cells and total protein was isolated from both the cells and the cell medium. Proteins in the cell medium were first processed on a Nickel column to isolate and concentrate the recombinant Fam20a protein and both protein samples were analyzed by immunoblotting using an antiserum specific for the Myc epitope. The predicted molecular weights of the full length and processed forms of Fam20a are 61,500 and 57,500, respectively, and a recombinant form of the protein synthesized in rabbit reticulocyte lysates was run alongside as a molecular size marker. The recombinant protein migrated just below the 62,000 mol.wt. size marker (Figure 8A and 8B , lane 5); however, the proteins detected in both the cell medium and cell extract migrated slower (lane 3). To test whether this slower migrating band represented a glycosylated form of Fam20a, the protein samples were treated with the enzyme N-glycosidase F (PNGaseF). The protein detected after enzyme treatment migrated more rapidly than the untreated protein and comigrated with the recombinant form of the protein (compare lanes 3, 4 and 5). We noted a second band that migrated slightly more slowly than the recombinant protein in the PNGase F treated cell extracts that may represent an alternatively modified form of Fam20a (Figure 8B , lane 4). To confirm that Fam20a is a secreted protein, we also exposed Fam20a-expressing cells to Brefeldin A, a fungal metabolite that specifically blocks transport from the ER to the Golgi apparatus, and examined the effects on Fam20a secretion. Brefeldin A treatment resulted in a consistent decrease in the amount of Fam20a detected in the cell medium (Figure 8C , compare lanes 5 and 6). Thus, Fam20a is a secreted glycoprotein. Figure 8 Fam20a is a secreted protein. COS-1 cells were transfected with either an empty expression vector (-) or one encoding mouse Fam20a with a C-terminal myc epitope tag and proteins were isolated from either the medium (panel A) or the cells (panel B). The proteins were analyzed by immunoblotting using a Myc tag-specific antiserum. Samples in lanes 2 and 4 of each blot were pre-treated with protein N-glycosidase prior to analysis to remove glycosyl groups. A recombinant form of Fam20a synthesized in rabbit reticulocyte lysates (TnT) was included on each gel as a size marker. The position of glycosylated and deglycosylated Fam20a is indicated using arrowheads and cross reacting material detected in the medium is indicated using asterisks. The location of molecular size markers is shown on the left of each gel. C. Protein samples from the medium of transfected cells that were untreated or treated with Brefeldin A were analyzed by immunoblotting using the Myc tag-specific antiserum. As Brefeldin A was resuspended in DMSO, the untreated cells were exposed to DMSO alone as a vehicle control. The amount of Fam20a detected in the medium of Brefeldin A treated cells was consistently lower than that observed in untreated cells (indicated using an arrowhead). Fam20a secretion requires a functional signal sequence We next tested whether the integrity of the signal sequence was required for Fam20a secretion. Signal sequences typically contain a high proportion of hydrophobic amino acids and 19 of the first 34 amino acids of Fam20a are hydrophobic (Figure 9A ). Therefore, we expressed a Fam20a protein lacking the first 23 amino acids (FAM20a(Δ23)) in COS-1 cells and examined secreted and intracellular proteins by immunoblotting (Figure 9B ). Glycosylated FAM20a(Δ23) protein was not detected in the medium (compare lanes 2 and 3) and immunoreactivity that comigrated with the unglycosylated recombinant protein was detected in the cell extract (lane 6). We also compared the subcellular location of the FAM20a(Δ23) protein to the wild type protein using GFP fusion proteins (figure 9C ). The wild type Fam20a-GFP proteins displayed perinuclear and cytoplasmic staining consistent with ER localization. In contrast, the Fam20a (Δ23)-GFP protein was absent from the cytoplasm and appeared to be exclusively localized within the nucleus. To ensure that this effect was not a consequence of a gross change in protein structure due to the deletion of 23 amino acids, we also constructed an expression vector encoding a Fam20a protein with a two amino acid substitution within the putative signal sequence (Figure 9A ). These changes (Leu 14 –Leu 15 to Asp-Glu) were predicted to disrupt the signal sequence without grossly altering the protein structure. Again the mutant protein displayed nuclear staining and was absent from the ER (Figure 9D ). These results confirm that an intact signal sequence was necessary for secretion of Fam20a and that secretion was accompanied by prominent localization of the protein to the ER. Figure 9 Secretion of Fam20a requires an intact signal sequence. A. Schematic representation of the putative signal sequence of Fam20a. The predicted cleavage site is indicated with a red arrowhead. The two amino acid substitutions introduced in the SSmut construct and the sequence remaining in the Δ23 mutant construct are shown. B. Immunoblot analysis of Fam20a and Fam20a(Δ23) protein levels in transfected COS-1 cells. The position of the glycosylated form of Fam20a (which is absent in Fam20a(Δ23) transfected cells is indicated with an arrowhead. C. Fluorescence images of COS-1 cells expressing either Fam20a-GFP or Fam20a(Δ23)-GFP. The wild type protein was observed within the cytoplasm, predominantly in a structure that is likely to be the ER. The mutant protein was primarily localized to the nucleus. D. Immunofluorescence images of Fam20a and Fam20a (SSmut) proteins as detected by antiserum directed against the C-terminal Myc epitope. The wild type protein was again detected in the ER and the mutant protein primarily in the nucleus. The cells have been counterstained with DAPI to delineate the nucleus. Figure 10 RT-PCR analysis of mouse Fam20 mRNA levels during differentiation of EML and MPRO cells. Total RNAs were prepared from EML (panel A) or MPRO (panel B) cells at the indicated timepoints during myeloid and granulocytic differentiation. cDNAs prepared from each sample were amplified using primer pairs specific to each mouse family member. The PCR products were analyzed by agarose gel electrophoresis and stained using Gelstar SYBR Green DNA stain. GAPDH was used as a loading control. Expression patterns of FAM20 genes during myeloid differentiation We originally identified Fam20a as a differentially expressed mRNA in EML cells induced to differentiate along the myeloid lineage. To determine whether Fam20b and Fam20c are also expressed during hematopoiesis, we performed RT-PCR analysis of cDNAs prepared at various times during experimentally-induced differentiation of EML and MPRO cells using primers specific to each family member. Fam20a mRNA levels were low in uninduced EML cells maintained in the presence of SCF and increased during the subsequent 72 hours of incubation in atRA and IL-3 (Figure 10A ). EML cells mature to the promyelocyte stage of neutrophil differentiation under these conditions and can subsequently be differentiated into neutrophils by adding GM-CSF in place of SCF and IL-3. Fam20a mRNA levels decreased during terminal neutrophil differentiation in EML cells and also in MPRO cells induced to undergo the same differentiation process in the presence of atRA (Figure 10A and 10B ). Fam20b and Fam20c mRNAs were readily detected in both cell lines and their levels did not vary dramatically during the differentiation process in either cell line (Figure 10A and 10B ). Expression patterns Of FAM20 genes in human tissues Although we originally isolated Fam20a from a hematopoietic cell line, cDNAs and ESTs derived from each of the FAM20 family members have been isolated from non-hematopoietic tissues (data not shown). Therefore, we examined the expression patterns of the three genes in a panel of cDNAs derived from various human tissues. FAM20A displayed the most restricted expression pattern with high levels in lung and liver and intermediate levels in thymus and ovary (Figure 11 ). Low levels of FAM20A mRNA were detected in several other tissues. FAM20B and FAM20C were expressed in a wider variety of tissues and their expression patterns were very similar. Figure 11 RT-PCR analysis of human FAM20 mRNA levels in human tissues. A panel of commercially available human cDNAs prepared from the indicated tissues was analyzed by PCR using primer pairs specific for each of the human FAM20 family members. GAPDH was again used as a loading control although large variations were observed in the GAPDH signal in the different tissues. Discussion Several classes of secreted proteins, including the colony stimulating factors or hematopoietins, are important regulators of hematopoietic differentiation and function [ 11 ]. These molecules are of clinical significance due to their use in stimulating hematopoiesis in patients with neutropenias and other hematological disorders [ 11 ]. Consequently, the identification of novel secreted proteins that display specific spatiotemporal expression patterns in hematopoietic cells is of great interest. In this report, we describe the identification and initial characterization of a new family of secreted glycoproteins expressed within the hematopoietic lineage as well as several other cell types and tissues. The family has been named FAM20 to indicate the fact that the members are related by sequence similarity rather than a specific shared function and contains three members in mammals. We anticipate that the members will acquire new names as their specific functions are determined. The family contains three separate subfamilies which are referred to as FAM20A, FAM20B and FAM20C in humans. FAM20 proteins: features and potential functions At the present time, the specific function(s) of the FAM20 proteins is unknown. Analysis of the sequences of the proteins from various species failed to reveal obvious similarities to known functional domains except for an N-terminal signal sequence. Expression studies clearly demonstrated that the mouse Fam20a protein is a secreted protein and that disruption of the signal sequence prevented the detection of the glycosylated form of the protein in cell media. Surprisingly, disruption of the signal sequence resulted in redistribution of intracellular Fam20a from a cytoplasmic compartment that is likely to be the ER to the nucleus. It is unclear whether this redistribution is functionally significant and is presumably due to the presence of a cryptic nuclear localization signal (NLS) in the protein. NLSs are generally comprised of stretches of basic residues [ 12 ] and several candidate regions rich in basic amino acids are present in Fam20a that could direct the mislocalized protein to the nucleus. Overall, the FAM20 proteins vary in length between 400–670 amino acids with the FAM20B family members being the shortest and the FAM20C members generally being the longest. The variation in length is due to differences in the length of the less conserved N-terminal region. The FAM20 proteins are further characterized by a highly conserved C-terminal region of approximately 350 amino acids that we named the conserved C-terminal domain (CCD). We noted the presence of three regions that are more highly conserved amongst all family members, with the most invariant extended sequence being the DRHHYE heptapeptide within conserved region 2. The function of the peptide is as yet unknown but three possible functions can be proposed. First, the histidines within this sequence could be involved in the coordination of metal ions that may be required for FAM20 protein function. Second, FAM20 proteins may be enzymes and this sequence may be a component of the active site of the enzyme. Interestingly, a weak match was detected between this region of certain FAM20 family members and a conserved domain within the phosphatidylinositol 3- and 4-kinases [ 13 ]; however additional experiments must be performed to address the relevance of this similarity. Third, the highly charged nature of this peptide suggests that it may be located on the surface of FAM20 proteins, where it may participate in protein:protein interactions. Although we have been unable to identify any other proteins in public databases that contain the exact sequence, the sequence RHHYE is found between amino acids 41–45 in the N-terminal region of the viral infectivity protein (Vif) from human immunodeficiency virus 1 (Accession number AAQ09611) [ 14 ]. Vif enhances HIV-1 infectivity by blocking the antiviral activity of the nucleotide editing enzyme APOBEC3G [ 15 ]. Vif exerts its inhibitory effect by binding to and inducing the degradation of APOGEC3G and also by blocking its translation [ 16 , 17 ]. The APOBEC3G interacting domain is located within the N-terminal region that contains the FAM20-related pentapeptide although the specific involvement of this sequence has not yet been investigated [ 16 ]. Thus, this sequence within conserved region 2 may be a site for protein:protein interaction. However, APOBEC3G or related proteins are unlikely candidate binding partners as they are intracellular proteins. We also noted that one of the Drosophila FAM20 proteins (NM_170079; CG31145) was identified as a protein that interacted with the Dynein light chain protein Dlc90f [ 18 ]. Dyneins are motor proteins involved in intracellular transport [ 19 ]. It appears unusual that a secreted protein would directly interact with a motor protein; therefore, this may represent a specific interaction of this particular family member in fruitfly. Evolution of the FAM20 gene family Tremendous progress has been made over the past decade in the sequencing of genomes from species at different positions on the evolutionary tree [ 20 - 25 ]. This vast amount of information can be used for comparative studies to elucidate the evolutionary origin of members of gene families. We have reported here the identification of orthologues of the FAM20 members in two insect species ( D. melanogaster and A. gambiae ), a simple chordate ( C. intestinalis ), three mammals ( H. sapiens, M. musculus and R. norvegicus ) and two fish ( F. rubripes and D. rerio ). In each case, the identification of these genes as bona fide transcription units is supported either by direct experimental evidence for the human and mouse genes, or by the existence of multiple EST sequences in public databases. A single FAM20 gene was identified in C. intestinalis , which is considered to be a representative of a basal chordate related to the common ancestor of humans and other higher chordates [ 10 ]. The C. intestinalis genome encodes approximately 16,000 genes and generally contains single representatives of superfamilies in higher vertebrates, thereby permitting the elucidation of likely evolutionary origins of genes within these families [ 9 ]. The C. intestinalis FAM20 protein clustered with the FAM20B subfamily members, as did the family members identified in invertebrate species. Therefore, we propose that the FAM20B subfamily contains the direct descendents of the ancestral FAM20 gene and that the FAM20A and FAM20C subfamilies result from duplication and subsequent evolution of this ancestral gene [ 26 ]. This pattern is consistent with the 2R hypothesis of genome evolution proposed by Ohno in 1970 [ 27 ]. In the FAM20 case, the loss of one gene at some stage of higher vertebrate evolution would need to be hypothesized to explain the final paralogue number of three in mammals. A single round of gene duplication appears to have occurred subsequent to the divergence of the nematode and insect lineages, giving rise to the two paralogues in fruit fly and mosquito that both cluster within the FAM20B subfamily. A further round of gene duplication has been proposed to have occurred in fish [ 28 ] and the existence of five to six FAM20 genes in pufferfish and zebrafish is consistent with this hypothesis. However, it is interesting that the expansion appears to have occurred exclusively in the FAM20C subfamily. This pattern could be explained either by two successive rounds of gene duplication of a small genomic region that included the original FAM20C representative, or by two rounds of duplication of larger genomic segments followed by gene conversion of FAM20A and FAM20B descendents (or parents) to yield four FAM20C members. Analysis of the genomic regions surrounding each of the FAM20C genes in fish will be necessary to distinguish between these two possibilities. Nevertheless, the expansion of the FAM20C subfamily in fish suggests that these proteins have acquired specific functions required within these species. FAM20 expression patterns in mammalian cells and tissues Expression analysis of the three FAM20 family members in mammalian tissues and hematopoietic cells showed that FAM20A is expressed in a much more restricted pattern that the other two members. Importantly, Fam20a was also the only member to display obvious differential expression in hematopoietic cells undergoing myeloid differentiation. Fam20a mRNA levels were highest during intermediate stages of differentiation of EML cells, at a time when many cells are becoming committed to the myeloid lineages and undergoing extensive proliferation [ 4 , 8 , 29 ]. EML cell cultures also give rise to a small number of cells from B cell and erythroid lineages under these conditions but these cells do not survive after SCF and IL-3 is removed from the medium and replaced with GM-CSF. Thus, the decrease in Fam20a mRNA levels after the 72 hour timepoint could be interpreted in two ways. First, Fam20a may be expressed in one of the lineages that cannot survive in GM-CSF. Second, Fam20a may be expressed specifically in cells committed to the myeloid lineage and its expression may decrease during terminal granulocytic differentiation. The similarities in Fam20a expression patterns in EML cells after the 72 hour timepoint and in MPRO cells suggests that the second explanation is most likely. Therefore, we propose that Fam20a is primarily expressed in cells committed to the granulocytic lineage and presumably plays a role in either lineage commitment of cell proliferation. The hypothesis is currently under investigation using gene disruption techniques and through further characterization of the Fam20a protein. Methods Cell culture and Representational Difference Analysis (RDA) EML and MPRO cells were cultured as described previously [ 6 , 30 ]. Briefly, EML cells were cultured in Iscove's modified Dulbecco's medium containing 20% horse serum (Atlanta Biochemical, Norcross, GA) supplemented with 10% BHK-MKL conditioned medium (CM) as a source of stem cell factor (SCF). The cells were differentiated by adding 10% of WEHI-3 CM as a source of IL-3 and all trans retinoic acid (atRA, 1 × 10 -5 M, Sigma) for 72 hours. Terminal granulocytic differentiation was induced by removing IL-3 and SCF and adding 10% BHK-HM5 CM as a source of granulocyte/macrophage-colony stimulating factor (GM-CSF). MPRO cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS, Hyclone, Logan, UT) and 10% BHK-HM5 CM. Terminal differentiation was induced by adding 1 – 10 -5 M atRA to the culture medium. For collection of RNA samples at different timepoints, 10 cm dishes were seeded with 150,000 cells and cultured under identical conditions. Cells were harvested from individual wells at 24 hour intervals for RNA preparation. Total RNA was prepared from isolated cells using a modified guanidium isothiocyanite/phenol extraction procedure as described previously [ 31 ]. Poly A + RNA prepared from EML cells at zero hours (stem cell stage) and 72 hours (promyelocyte stage) was converted to double stranded cDNA and three rounds of RDA was performed as described previously [ 6 ]. cDNA clones representing six putative differentially expressed genes were identified and named according to clone number. Initial characterization of these clones was described earlier [ 6 , 8 ] and clone number 1623 was examined further in this study. Cos-1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM, BioWhittaker, Walkersville, MD) containing 10% FBS and Penicillin/Streptomycin (BioWhittaker) under standard cell culture conditions. For transfections, cells were plated at a density of 2.5 × 10 4 per well in six well plates and transfected using the Effectene Transfection Reagent (Qiagen, Valencia, CA) under conditions recommended by the manufacturer. Isolation of a full length 1623 cDNA Analysis of the sequence of the original 1623 cDNA isolated from the RDA experiments revealed that it contained the coding sequence from the C-terminus of an uncharacterized protein. 5' and 3' rapid amplification of cDNA ends (RACE) was performed using Marathon-Ready cDNA from mouse spleen using the Advantage cDNA PCR kit (both from Clontech) as described previously [ 30 ]. The RACE reactions were performed using gene specific primers designed based on the original 1623 sequence or on sequences identified in early RACE reactions and adaptor primers provided with the cDNA. The 5' RACE products were compared to sequences in public databases and identified a mouse cDNA (NM_017565; aka DKFZp434F2322) that was identical to the extended 1623 sequence. PCR primers were designed that overlapped the putative initiation and stop codons of the ORF encoded by NM_017565 and used to PCR amplify products from cDNAs prepared from undifferentiated MPRO cells. The product of this PCR reaction was subcloned and sequenced in its entirety. The sequence was identical to the original clone (which was isolated from mammary tissue) indicating that the same mRNA is expressed in both tissues. The full length clone (now named Fam20a) encoded a 541 amino acid protein. Sequence analysis All basic sequence manipulations were performed using the VectorNTI suite of sequence analysis programs (Informax/Invitrogen Corp., Carlsbad, CA). Identification of FAM20 members from different species was initially achieved using standard BLAST searches at NCBI . Genome specific searches were performed either through NCBI or Ensembl or through genome specific websites ( Ciona intestinalis : ; Fugu rubripes : ). Searches were performed either as nucleotide-nucleotide (blastn) searches or as protein-translated nucleotide (tblastn) searches. In general, these searches were sufficient to identify gene regions encoding the most highly conserved regions of the proteins, particularly in the fish and Ciona genomes. To assemble complete genes, genomic regions containing the identified regions of similarity were downloaded into VectorNTI and searches were performed for individual exons, initially derived from mammalian genes but subsequently, from potential orthologues in fish or lower vertebrates, depending on the sequence being examined. Putative exons were examined for the presence of consensus splice donor and acceptor sites and complete sets of exons were assembled into cDNAs for translation and further comparisons. These results were compared against genes assembled by two gene prediction programs, FGENESH: and GENSCAN: , however, these programs often made incorrect assignments that required manual assessment of the results. The protein sequences derived from each of the examined species were primarily used in the comparisons to define subfamily assignments and alignments were performed in the AlignX program in VectorNTI using the blosum62 scoring matrix. Protein domain searches were performed using the Simple Modular Architecture Research Tool (SMART) and/or Profilescan . Signal sequence searches were performed using the SignalP analysis program [ 32 ]. Sequences derived by annotation of genome sequences in this study have been submitted to the Third Party Annotation database at NCBI under accession numbers BK001515 to BK001522. The FAM20 family name has been approved by the Human Genome Organization nomenclature committee. Plasmid construction The Fam20a and Fam20aΔ23 were constructed in the pEGFP-N1 vector by PCR cloning. Xho I and Hind III restriction sites were engineered into the PCR primers (Fam20a: 5'GGCCTCGAGGCCATGCCCGGGCTGCGCAGG3', 5'GGCAAGCTTGCTCGTCAGATTAGCCTG3'; Fam20aΔ23: 5'GGCCTCGAGGCCATGTACTTCCACCTCTGGCCG3' and reverse primer was as above). Similar strategy was used to clone Fam20a and Fam20aΔ23 in the pcDNA3.1 vector (Forward primers were same as above; Fam20a and Fam20aΔ23 reverse primers: 5'GGCAAGCTTGGGCTCGTCAGATTAGCCTG3'). The Fam20a(SSmut) was generated using a PCR-based site-directed mutagenesis kit (Quickchange, Stratagene). All of the sequences were verified by sequencing using an ABI automated sequencer. Western blotting Whole cell extracts were prepared from transfected COS-1 cells by lysing directly in 2X Laemmli Sample Buffer as described previously [ 31 ]. Purified His-tagged proteins were prepared as described below and mixed with 2X Laemmli Sample Buffer (120 mM Tris-HCl (pH 6.8), 10% Glycerol, 3.3% SDS, 0.2 M dithiothreitol, 0.004% Bromophenol Blue). Equal amounts of total cellular and purified secreted proteins were separated by 12% SDS-PAGE and transferred to nitrocellulose membranes (Micron Separations, Westborough, MA). Membranes were blocked in 5% non-fat milk in TBST (100 mM Tris-HCl (pH 7.5), 0.9% NaCl, 0.1% Tween-20) for 1 hour. Myc-tagged Fam20a proteins were detected using an anti-myc mouse monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Horseradish peroxidase-conjugated donkey anti-mouse antibody (Promega, Madison, WI) was used for the secondary antibody. The immune complexes were detected using SuperSignal chemiluminescence detection kit (Pierce, Rockford, IL). In vitro transcribed and translated Fam20a was generated using the TnT T7 coupled reticulocyte lysate system (Promega) as described previously [ 33 ]. Purification of His-tagged proteins from cell media COS-1 cells were transfected as described above and media were collected after 48 hours and mixed with freshly made 10X binding buffer (500 mM NaH 2 PO 4 -H 2 O (pH 8.0), 150 mM NaCl, 10 mM Imidazole). His-tagged proteins were purified using Ni-NTA Magnetic Agarose Beads (Qiagen, Valencia, CA) according to manufacturer's recommendations. Following the purification step they were either used directly in western blot analysis or for deglycosylation experiments. Purified proteins were subjected to enzymatic deglycosylation using the GlycoPro deglycosylation kit (ProZyme, San Leandro, CA) following the manufacturer's protocol. Immunofluorescence COS-1 cells were plated on glass coverslips and transfected with the indicated constructs the following day. After 48 hours of incubation, the cells were fixed in 1% formaldehyde for 30 minutes at room temperature. Coverslips were incubated in PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 (pH 7.3)) containing 1% Triton X-100 for 10 minutes to permeabilize the cell membranes. They were then transferred to PBST (PBS containing 0.1% Tween-20) and incubated for 30 minutes. Mouse monoclonal anti-Myc antiserum (Santa Cruz) was used at 1:2,000 dilution in PBST containing 1% bovine serum albumin (1 hour at room temperature). Coverslips were then washed with PBST three times and incubated with the secondary antibody (Alexa Fluor 488 [Molecular Probes, Eugene, OR]) for 1 hour. Coverslips were washed three times in PBST and once in water for 15 minutes each, and were mounted on glass slides. The cells were observed by epifluorescence microscopy using an Axiovert 135 TV microscope (Zeiss, Gottingen, Germany). Images were captured with a Kodak DC290 zoom camera and analyzed with MetaMorph 6.0 (Universal Imaging Corporation, Downingtown, PA). Reverse transcription-polymerase chain reaction Total RNA was purified as described above from EML and MPRO cells at 24 hour timepoints during the differentiation process of each cell line The primers were designed using the Vector NTI sequence analysis suite of programs (InforMax, Frederick, MD) and were follows: MmFam20a forward 5'-catagaggcccacggcgagcg-3' and reverse 5'-atggaatggggcaacag gggc-3'; Mmfam20b forward 5'-tggacaggattctgggtttc-3' and reverse 5'-ccagggatgtcgatgtttct-3'; MmFam20c forward 5'-agcagacgagagagcaggag-3' and reverse 5'-cggatctccttggtcatgtt-3'; HsFAM20a forward 5'-ctggcaggaaaagagtg-3' and 5'-cccgaacttggtgaacatct-3', HsFAM20b forward 5'-ccctgaagagacaccagaagagc-3' and reverse 5'-gaaacccagaatcctgtcca-3', HsFAM20c forward 5'-ggctcacgttccacattggt-3' and reverse 5'-aaagtcagggggtgtctcct-3'.; mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward 5'-aatggtgaaggtcggtg tgaac-3' and reverse 5'-gaagatggtgatgggcttcc-3'; human (GAPDH) forward 5'-tgaaggtcggagtcaacggatttggt-3' and reverse 5'-catgtgggccatgaggtccaccac-3'. GAPDH was used as control for cDNA integrity. Single stranded cDNA was reverse transcribed from 2 μg of total RNA using 400 Units of Superscript RNase H - Reverse Transcriptase (Invitrogen, Carlsbad, CA), 0.125 mM dNTPs, 10 mM DTT, and 1 μM oligo (dT) 15 in a total volume of 50 μl for 1 hour at 37°C. 2 ul of the RT reaction was then mixed with 1 ul of 10X PCR Buffer (500 mM Tris-HCl (pH8.3), 2.5 mg/ml crystalline BSA and MgCl 2 at 10, 20 or 30 mM (Idaho Technology Inc., Idaho Falls, ID), 0.2 mM each dNTP, 0.05 μM of each primer and 1.25 Units Taq DNA polymerase (Fisher Scientific, Pittsburgh, PA) in a total volume of 10 μl. Reactions were loaded into a capillary tube and PCR cycles were carried out using the Rapidcycler Thermal cycler (Idaho Technology). Annealing temperatures and Mg 2+ concentrations were initially optimized for each primer. A commercial set of tissue cDNAs, (Human Multiple Tissue cDNA (MTC) Panels, Clontech, Palo Alto, CA) was used for the RT-PCR analysis of human FAM20 family members. List of Abbreviations FAM20: family with sequence similarity 20; EML: Erythroid, myeloid, lymphoid cell line; MPRO: mouse promyelocyte cell line; GM-CSF: granulocyte/macrophage-colony stimulating factor; PHSC: pluripotent hematopoietic stem cell; HGF: hematopoietic growth factor; RAR: retinoic acid receptor; atRA: all trans retinoic acid; SCF: stem cell factor; IL-3: interleukin-3; RDA: representational difference analysis; RACE: rapid amplification of cDNA ends; CCD: conserved C-terminal domain; GFP: green fluorescent protein; CM: conditioned medium. Author Contributions DN isolated full length Fam20a cDNA, performed transfection experiments and drafted the manuscript. HY performed RT-PCR assays. IN assisted in construction of expression vectors. SS performed genomics analyses. EC and EGB provided reagents and assistance in performing RT-PCR assays. YD performed the original representational difference analysis and SCW directed the project and performed genomics analyses. SCW also wrote the final draft of the manuscript and all authors read and approved this version.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548683.xml

300885

A Truly Broad View of Gene Expression Spotlights Evolution and Diversity

Bioinformatics and microarrays have given scientists powerful new tools to investigate the structure and activity of genes on a global scale. Rather than studying just a few genes, scientists can analyze tens of thousands within and across species. Microarrays flag which genes are expressed under particular cellular conditions in an organism, while genome sequencing offers clues to gene function and regulation. By comparing the genomic properties of different species, scientists can spot patterns that help them identify functional and regulatory elements, learn about genome structure and organization, and gain a better understanding of the evolutionary forces that shape life on Earth. The potential of these technologies to reveal insights into the fundamental structure and function of biological systems continues to grow along with the wealth of gene sequence and expression data—but the ability to interpret and merge these datasets lags behind the ability to collect them. In an effort to overcome these limitations, Sven Bergmann, Jan Ihmels, and Naama Barkai developed a comparative model that integrates gene expression data with genomic sequence information. Because functionally related genes are expected to be coexpressed in different organisms and because the sequence of some of these functionally related genes may also be conserved between organisms, Bergmann and colleagues hypothesized that “conserved coexpression” could serve as an indicator of gene function on a genomic level. (Conserved genes are those that have changed little since they first evolved. Conserved coexpression describes functionally related genes that are activated together in different species.) But first they had to determine whether coexpression was conserved among species. Analyzing the gene expression profiles of six distantly related organisms—bacteria, yeast, plant, worm, fruitfly, and human—the researchers found that functionally related genes were indeed coexpressed in each species. The most strongly conserved sets of coexpressed genes are associated with core cellular processes or organelles. These results indicate that conserved coexpression can improve the interpretation of genome sequence data by providing another functional indicator for homologous sequences. Since functionally related genes are expressed together in different organisms, it would be reasonable to think their regulatory networks are also conserved. To explore this idea, the researchers grouped coexpressed genes and their regulatory elements into “transcription modules” for each organism. They found significant variation in the number, organization, and relative importance of these modular components. Which components contributed most to an organism's global transcription program, for example, depended on the organism. But they also found that the transcription networks are highly clustered—meaning that genes connected to a specific gene are also connected to each other. This finding indicates that gene expression programs, regardless of their size or individual components, are highly modular. Each transcriptome contains modules that have been conserved over time along with “add-on” modules that reflect the needs of a particular species. This modularity supports the notion that variation between and among species arises from the diversity of gene expression programs. Although the regulatory details of individual gene groups varied, the researchers found common ground in the overall landscape of the expression data. The transcription programs exhibit properties typical of dynamically evolving “real-world” networks that are designed to perform in uncertain environments and to maintain connections between elements independent of scale. These properties were originally identified in studies of social networks and the World Wide Web, but they aptly describe the real-world challenges of the cell. Studies of dynamically evolving networks show that nodes (i.e., genes and proteins) added at an early stage (much like highly conserved genes) are more likely to develop many connections, acting as a hub. Following these organizational principles, transcription networks would have a relatively small number of highly connected “hub genes”—though a much higher number than one would expect in a random network. And that is what the authors observed: the networks they constructed from the expression data had the expected number of highly connected hub genes, which tend to be essential and conserved among organisms. Since these highly connected genes are likely to have homologues in other organisms, they can serve as powerful and efficient tools for assigning function to the thousands of uncharacterized sequences found in sequence databases. This model presents a framework to explore the underlying properties that govern the design and function of the cell and provides important clues—in the form of conserved transcription modules—to the evolutionary building blocks that generate diversity. Regulatory relations among transcription modules

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC300885.xml

546334

Novel Enzyme Shows Potential as an Anti-HIV Target

At just 9.8 kilobases, the HIV genome pales in comparison to the 3.2 gigabases of its human and nonhuman primate targets. The compact retrovirus encodes just 14 proteins, which play different roles in promoting viral infection and virulence. As a retrovirus, HIV uses the host's cellular machinery—including RNA polymerases, which carry out transcription—to copy its RNA genome into DNA and infiltrate human chromosomal DNA. Once the virus is integrated (now called a provirus), its genes can be transcribed. Adept as HIV is in exploiting its host's molecular resources, the virus can't establish a foothold without the services of its skeleton crew. The HIV transcription factor Tat (“transactivator of transcription”), for example, is an essential regulator of HIV gene expression. Without Tat, HIV transcripts don't reach full length and can't effect viral replication. In a new study, Melanie Ott and colleagues identify an enzyme that regulates viral transcription by modifying Tat. The regulation of HIV genes depends on a complex interplay between proviral DNA, cellular proteins and transcription factors, and Tat. Unlike most transcription factors, Tat activates transcription by binding to RNA, specifically to a bulging “stem-loop” structure that forms at one end of all viral transcripts called the trans -acting responsive element (TAR). Tat binding to TAR requires recruiting the enzyme cyclin-dependent kinase 9 (CDK-9) to the HIV promoter (where transcription begins). CDK-9 chemically modifies the RNA polymerase and enhances its transcribing efficiency. The transcription process—including the labyrinthine protein–protein and protein–DNA (and in the case of Tat, protein–RNA) interactions—is highly regulated. One process that figures prominently in this regulation is acetylation, which adds an acetyl group (a molecule made of oxygen, hydrogen, and carbon) to a molecule or protein. Histone acetylation was long thought to influence transcription by regulating the structure and function of chromatin, which is an assembly of proteins (mostly histones) and DNA. Another, more widely accepted model proposes that histone acetylation controls transcription by recruiting cofactors required for transcription. Acetylation and deacetylation enzymes can also target other proteins. Of the three classes of deacetylases known to modify human histones, the sirtuins (SIRT1–7) appear to preferentially target a number of nonhistone proteins. Ott and colleagues first tested the ability of all seven SIRT proteins to deacetylate Tat by placing them in a test tube with Tat proteins. Though three SIRT enzymes caused Tat acetylation, only one, SIRT1, is a nuclear enzyme, like Tat, suggesting that SIRT1 might work similarly in living cells. Recycling of Tat through deacetylation by SIRT1 Ott and colleagues went on to show that transcription via Tat occurs in the presence of SIRT1, but not when SIRT1's catalytic center is removed. Experiments using cells taken from transgenic mice lacking SIRT1 demonstrated that introducing human SIRT1 enzymes increased Tat's transcriptional effects in a dose-dependent manner, while treating cells with the small molecule HR73, a derivative of a molecule that inhibits the yeast version of the SIRT1 protein, caused a 5-fold reduction in HIV transcription. The authors propose a cycle of transcriptional transactivation in which SIRT1 deacetylates Tat at the HIV promoter. Deacetylated Tat associates with CyclinT1 and TAR, and leads to transcription. Tat acetylation dissociates Tat from CyclinT1 and TAR, and transfers Tat to the elongating polymerase complex. Since acetylated Tat can't recruit CyclinT1 and CDK-9, the authors explain, a new round of transcription requires that new, unacetylated Tats are produced or existing Tats are deacetylated. Thus, efficient viral replication depends on adequate Tat supplies. And since HIV gene expression relies on SIRT1's enzymatic activity, inhibiting SIRT1 could prove to be a promising anti-HIV therapy. Future study will have to verify whether inhibiting SIRT1 can successfully put the brakes on HIV transcription and control the virus (See also “A New Paradigm in Eukaryotic Biology: HIV Tat and the Control of Transcriptional Elongation” [DOI: 10.1371/journal.pbio.0030076 ]).

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546334.xml

545199

Climate Drives the Meningitis Epidemics Onset in West Africa

Background Every year West African countries within the Sahelo-Sudanian band are afflicted with major meningococcal meningitis (MCM) disease outbreaks, which affect up to 200,000 people, mainly young children, in one of the world's poorest regions. The timing of the epidemic year, which starts in February and ends in late May, and the spatial distribution of disease cases throughout the “Meningitis Belt” strongly indicate a close linkage between the life cycle of the causative agent of MCM and climate variability. However, mechanisms responsible for the observed patterns are still not clearly identified. Methods and Findings By comparing the information on cases and deaths of MCM from World Health Organization weekly reports with atmospheric datasets, we quantified the relationship between the seasonal occurrence of MCM in Mali, a West African country, and large-scale atmospheric circulation. Regional atmospheric indexes based on surface wind speed show a clear link between population dynamics of the disease and climate: the onset of epidemics and the winter maximum defined by the atmospheric index share the same mean week (sixth week of the year; standard deviation, 2 wk) and are highly correlated. Conclusions This study is the first that provides a clear, quantitative demonstration of the connections that exist between MCM epidemics and regional climate variability in Africa. Moreover, this statistically robust explanation of the MCM dynamics enables the development of an Early Warning Index for meningitis epidemic onset in West Africa. The development of such an index will undoubtedly help nationwide and international public health institutions and policy makers to better control MCM disease within the so-called westward–eastward pan-African Meningitis Belt.

Introduction Meningococcal meningitis (MCM) has affected Sahelian Africa for centuries and became endemic over the past 25 y. During the 1980s, the World Health Organization (WHO) registered between 25,000 and 200,000 disease cases per year, with about 10% of them resulting in death, and with the highest infection rates observed in younger children [ 1 ]. MCM became, therefore, a public health concern in the poorest regions in the world following the severe drought at the end of the 1970s. MCM is an infection of the meninges, caused by the bacteria Neisseria meningitidis , that causes high death rates in African communities. The agent is highly contagious, and person-to-person aerial transmission occurs through respiratory and throat secretions [ 2 ]. Interaction between different environmental parameters (e.g., immune receptivity of individuals, a poor socioeconomic level, the transmission of a more virulent serotype [such as the recent emergence of the serogroup W135 in West Africa], social interactions [such as pilgrimages, tribe migrations, and meetings], and some specific climatic conditions) may play a part in MCM disease outbreaks and spread within local populations [ 2 ]. Among favorable conditions for the resurgence and then dispersion of the disease, climatic conditions may be important as environmental forces inducing periodic fluctuations of disease incidence. Recent findings concerning the population dynamics of some infectious diseases have clearly identified the importance of climate as a major driver [ 3 , 4 ]. MCM outbreaks in West Africa usually start at the beginning of February, and then disappear in late May. The geographical distribution of disease cases is called the “Meningitis Belt” and is roughly circumscribed to the biogeographical Sahelo-Sudanian band [ 5 , 6 ]. This Sahelo-Sudanian region has a dry winter, dominated by northern winds, called the Harmattan, followed by a wet season starting in spring with the monsoon. The co-occurrence in both space and time of MCM disease cases and climate variability within the Sahelo-Sudanian area suggests that the occurrence of MCM might be directly related to climate. So far, very few studies have tried to quantify the potential linkages that could exist between climate and MCM outbreaks ( Figure 1 ). Figure 1 The “Meningitis Belt” in West Africa Modified from the WHO [ 9 ]. The winter climate causes damage to the mucous membranes of the oral cavity through dry air and strong dust winds, and creates propitious conditions for the transmission of the bacteria responsible for MCM; low absolute humidity and dust may enhance meningococcal invasion by damaging the mucosal barrier directly or by inhibiting mucosal immune defenses. In contrast, higher humidity during both the spring and summer seasons strongly reduces disease risk by decreasing the transmission capacity of the bacteria [ 7 , 8 ], and MCM epidemics generally stop with the onset of rainfall [ 9 ]. In addition to the seasonal cycle, the link between climate and meningitis has also been documented at the interannual scale in northern Benin, where Besancenot et al. [ 10 ] have suggested a positive relationship between low absolute humidity and interannual variability in meningitis. Meanwhile, although the global influence of climate is quite clear, the effects of climatic variability on MCM population dynamics are still only partially known because of the mixing of different processes acting at different spatial hierarchical scales, and the interactions between disease outbreaks and medical, demographical, and socioeconomic conditions. Most studies thus far have focused on very small spatial scales, and the need remains to discriminate between local properties and potential large-scale effects in disease patterns, to go beyond data heterogeneities and idiosyncratic details in order to identify important disease patterns influenced by large-scale forces such as climate variability (see Methods). The aim of the present work is thus to document the climatic context of MCM disease outbreaks and population dynamics in a highly affected Sahelian country, that is, Mali, and to show, if it exists, the presence of a correlation between climate and seasonal resurgence of disease. The idea behind the present study was to explain MCM disease dynamics in Mali in a statistically robust way, which will permit us to propose some tools for predicting disease risks for the benefit of public health. Methods The Scaling-Up Approach: From Local to Global Scales Recent findings concerning the population dynamics of some infectious diseases have clearly identified the importance of climate as a major driver [ 3 , 4 ]. With evidence of the impact of large-scale meterological phenomena such as El Niño on infectious disease patterns, modern epidemiology is now confronted with a scale problem in the identification of the spatiotemporal scales that might be relevant to explain patterns and processes [ 11 ]. Since most previous studies have focused on very small spatial scales, there is a need for “bottom-up” approaches to discriminate between local properties and potential large-scale effects on disease patterns. One of the simplest assumptions of these “bottom-up” approaches is the assumption that local scales are random processes overlaying a driving large-scale phenomenon such as climate variability. As such, this scaling-up approach seeks to point out the emerging patterns conditioned by the large-scale processes, with a random or deterministic function f such that: Local data = f (large-scale forces, idiosyncratic details). The aggregation of local data in the scaling-up approach is a simple way to go beyond data heterogeneities and idiosyncratic details so that only the important disease effects of large-scale forces remain. That is the rationale for our study: to show that large-scale phenomena such as the seasonal Harmattan winds over the whole of Mali can contribute to the periodic resurgence of MCM and its variation in time on a national scale, even if this aggregate analysis for the entire country is not able to capture variations at smaller space scales. Epidemiological Data: The WHO Weekly Reports The diagnosis of MCM is based on both physical examination and on evaluation of the cerebrospinal fluid (CSF) from a lumbar puncture. As the disease is characterized by a sudden onset of intense headache, fever, nausea, vomiting, photophobia, and stiff neck, in association with neurological symptoms (lethargy, delirium, coma, and convulsions), the WHO [ 9 ] recommends that the clinical diagnosis include an examination for meningeal rigidity, neurological signs, purpura, blood pressure, and focal infection. A lumbar puncture and CSF examination are then used to confirm the diagnosis based on physical examination and to identify the meningococcus [ 9 ]. This diagnosis is the basis for disease surveillance and case reporting using a standard case definition for MCM ( Box 1 ) that can be implemented in any health-care setting. Meningitis reports are included in the weekly reports of notifiable diseases and are aggregated at different spatial scales from the health site to the country level. The present study is based on these weekly reports by the WHO's Department of Communicable Disease Surveillance and Response of cases and deaths due to MCM over the whole of Mali from 1994 to 2002. Box 1. Standard Case Definition of MCM Modified from the WHO [ 9 ] This case definition allows the detection of cases of meningococcal septicemia. Suspected case of acute meningitis a . Sudden onset of fever (>38.5 °C rectal or 38.0 °C axillary) with stiff neck. In patients under 1 y of age, a suspected case of meningitis occurs when fever is accompanied by a bulging fontanelle. Probable case of bacterial meningitis b . Suspected case of acute meningitis as defined above with turbid CSF. Probable case of MCM b . Suspected case of either acute or bacterial meningitis as defined above with Gram stain showing Gram-negative diplococcus or ongoing epidemic or petechial or purpural rash. Confirmed case c . Suspected or probable case as defined above with either positive CSF antigen detection for N. meningitides or positive culture of CSF or blood with identification of N. meningitides. a Often the only diagnosis that can be made in dispensaries (peripheral level of health care). b Diagnosed in health centers where lumbar punctures and CSF examination are feasible (intermediate level). c Diagnosed in well-equipped hospitals (provincial or central level). The Typical Seasonal Pattern of the Disease In this study, an epidemic is defined in terms of population dynamics following Anderson and May [ 12 ] and Grenfell and Dobson [ 13 ]. This definition considers disease resurgence and its variation in time and allows us to focus on the cyclic character of the disease resurgence each year, even if the number of annual cases is low, and it makes it easier to find temporal correlations between climate and disease. To represent a typical seasonal cycle of a meningitis epidemic, we computed the weekly mean of standardized anomalies M ( w ) of the number of cases as where X̄ y and σ y represent, respectively, the mean and the standard deviation of the 54 weekly values of cases X y ( w ) for the year y, and N = 9 represents the number of years for the 1994–2002 period. The Onset of the Epidemic We determined the week of the onset of the epidemic for each year by characterizing a breaking slope in the annual cycle of the number of cases. The dates of the breaking slope have been determined objectively by using the Mann–Whitney–Pettitt test [ 14 ], which is a nonparametric test used here to detect a “change point” in a time series. A change point is defined as a point on either side of which values are on average higher or lower than the whole of the other data points. Considering the studied time series X y ( w ), we computed U y ( w ) as where 2 ≤ w ≤ 54 and U y (1) = V y (1), where V y ( w ) is defined by Then the the most significant change point of the year y is the point for which the value |U y ( w )| is maximized. The probability P y ( w ) of a given week being a change point is defined by where T = 54 (the length of the time series in weeks). Atmospheric Data: The NCEP/NCAR Reanalysis The National Centers for Environmental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR) have completed a reanalysis project with a current version of the Medium Range Forecast model [ 15 ]. This dataset consists of a reanalysis of the global observational network of meteorological variables (wind, temperature, geopotential height [i.e., the height of a pressure surface above mean sea level], humidity on pressure levels, surface variables, and flux variables such as precipitation rate), with a “frozen” state-of-the-art analysis and forecast system at a triangular spectral truncation of T62, performing data assimilation throughout the period from 1948 to the present. This analysis enables circumvention of problems involving previous numerical weather prediction analyses due to changes in techniques, models, and data assimilation. Data are reported on a 2.5° × 2.5° grid every 6 h (00.00, 06.00, 12.00, and 18.00 UTC) on 17 pressure levels from 1,000 hPa to 10 hPa, a good resolution for studying synoptic weather systems. For this study we used the wind speed fields at 1,000 hPa (near the surface) for the 9 y of the 1994–2002 period: first we averaged the four outputs of each day, and then we averaged these daily means for each week to obtain a weekly value. Computation of the Harmattan Wind Index The principal component analysis (PCA) [ 16 ] is a multivariate procedure that extracts the common variance that exists in a set of variables. The main use of PCA is to reduce the size of a dataset while retaining as much information as possible in principal components (PCs), which are linear combinations of the initial variables. In this study, this technique has been used in order to summarize the spatiotemporal variability of wind fields at low levels. As the input data are spatial objects (grid points), the PCA gives for each mode a spatial pattern associated with a time series (the PC). We performed the PCA on the weekly values from 1994 to 2002 of wind speed at 1,000 hPa over the Mali window (in red in Figure 1 ) by taking into account all the grid points from 12.5° N to 25° N and from 12.5° W to 2.5° E. The input matrix was thus composed of 6 × 7 = 42 loadings (the number of grid points over the spatial window) and 486 scores (the number of weeks in the 1994–2002 period). Data was first standardized in order to extract the correlation matrix C = X′X, where X represents the input matrix and X′ its transpose. The α th PC time series ψ α can thus be obtained by a linear combination of the initial variables through where u α is the α th eigenvector of the correlation matrix C associated with the eigenvalue λ α . The α th spatial pattern is the correlation map between the initial wind fields and the α th PC time series. The examination of the different spatial patterns and PC time series (not shown here but previously discussed in [ 17 ]) reveals a close relationship between Harmattan wind dynamics and the third PC with negative values, which represents a strong wind in southern Mali. The Harmattan wind index of the study thus represents the third PC, with a temporal pattern very similar to the seasonal cycle of wind speed associated with the Harmattan winds over Mali. Results The “Epidemic Seasonality” in Mali The weekly records of the WHO's Department of Communicable Disease Surveillance and Response of cases and deaths due to MCM for the 1994–2002 period allowed us to describe the seasonal evolution of MCM epidemics in Mali. Two important parameters were used: the date of the onset of the epidemic and the date of the seasonal maximum number of cases. We determined for each year the week of the onset of the epidemic (here called “ W o ”) as determined by a breaking slope in the annual cycle of the number of cases. The dates of breaking slope have been determined objectively by using the Mann–Whitney–Pettitt test [ 14 ], which is a nonparametric test used here to detect a change point in a time series. This test has the advantage of being adapted to small samples, giving the point of the most significant change and the probability of it being a significant change point (see Materials and Methods). The mean date of epidemic onset fell between the fifth and sixth week of the year (7–15 February), with a standard deviation of about 2 wk (5.2 ± 1.7 wk). For the 1994–2002 period, the maximum number of cases occurred between week 13 and week 14, that is, between 1 April and 15 April, with a standard deviation of about 2 wk (13.7 ± 1.6 wk) ( Figure 2 ). Figure 2 The Seasonal Periodicity of Meningitis Cases Mean seasonal pattern of the number of cases of MCM over the 1994–2002 period in standardized anomalies (bars). The red curve represents the same evolution, but in composite mean, using the week of epidemic onset as the reference date, W o , each year. Time series in red is shown from “ W o − 3 wk” to “ W o + 30 wk.” In order to mitigate the effect of strong variability from one year to another during the 1994–2002 period, we computed the average of standardized anomalies of the number of cases (bars in Figure 2 ) to represent a typical seasonal pattern of a meningitis epidemic. The first 5 wk are characterized by negative anomalies. The average length of the “epidemic year,” as defined by the number of consecutive weeks with positive anomalies, is 4 mo (16 wk). To improve the description of the seasonal pattern of the epidemic and to reduce noise due to the variability of the onset date year to year, we determined the composite mean of the number of cases over the 1994–2002 period by using the week of epidemic onset for each year as the respective reference date, W o . The red curve of Figure 2 shows the mean seasonal course before and after the onset of the epidemic, showing an abrupt increase of the number of cases—the “upward phase”—until the sixth week after the onset, a highly active period of the disease—the “active phase”—from “ W o + 6” to “ W o + 10,” followed by a decrease of the number of cases—the “downward phase”—until the end of the epidemic around 16 wk after the onset. Both upward and downward phases lasted on average 1.5 mo. The Atmospheric Circulation during an Epidemic in Mali Rainfall distribution over West Africa is controlled by the meridional migration of the intertropical convergence zone following the seasonal excursion of the sun [ 18 ]. The latitudinal shift of this zone of high humidity and instability leads to an opposition of two main annual regimes: the bimodal regime of the Guinean latitudes (from the equator to 7° N) with two rainy seasons during spring and autumn, and the unimodal regime of monsoon over Sudano-Sahelian Africa and succession of a dry winter and a wet summer [ 19 ]. North of the intertropical convergence zone, the intertropical front is defined as the confluence line between moist southwesterly monsoon winds and dry northeasterly Harmattan [ 20 , 21 ]. The seasonal progression of this system, involving a migration toward the summer pole of moisture and winds converging in the low layers, can be documented by using weekly fields surface wind speed from NCEP and NCAR data, which provide gridded atmospheric parameters with a 2.5° resolution [ 15 ]. The relationship between atmospheric circulation and the seasonal course of the MCM epidemic in Mali was studied using a regional index summarizing the spatiotemporal evolution of the low-layer circulation. This index was obtained from a dominant mode of a PCA (see Materials and Methods) applied to weekly fields of surface wind speed over the 1994–2002 period in Mali. The seasonal pattern shows the Harmattan wind dynamics ( Figure 3 ) with negative values representing strong winds, and positive values representing weak winds, in the southern part of Mali, the area under study in the present work. Figure 3 Temporal Patterns of Epidemics and Climate Weekly means of the Harmattan wind index over the 1994–2002 period and mean seasonal pattern of the number of cases of MCM (in standardized anomalies). Using this atmospheric index, we defined the date of “winter maximum” as the first minimum of the wind index for each year of the period 1994–2002. The winter maximum thus corresponds to the week where wind speed is the strongest. The mean date of winter maximum is around the sixth week, with a standard deviation of 2 wk, corresponding to the week when the Intertropical Front is located at its southern latitude. The Harmattan wind index shows a temporal pattern very similar to the that of the number of cases of MCM, with a clear breaking slope at the sixth week, on 15 February, corresponding to the onset of the epidemic and to the winter maximum, and with a recession of the disease at the 16th week concomitant with the onset of the wet season in the south part of Mali in early May. It is interesting to note that although they are determined from two different datasets, the mean weeks of winter maximum and of the onset of the epidemic are identical, 7–15 February. This coherence is reinforced by a very strong correlation between the two dates (0.92) for the years 1994 to 2002. Figure 4 shows the linear regression analysis between week of winter maximum and week of epidemic onset. Although the number of years under consideration is low, the scatter plot points out the close statistical linkage between these two events, suggesting that the winter maximum explains more than 85% of the variance in the week of epidemic onset in Mali: An earlier winter is associated with an earlier onset of the epidemic, and a later winter with a later onset. However, even though the results of the correlation analysis are strongly significant, the high R 2 is partially due to the low number of considered years (only nine); this low number is the main limitation of the present analysis. Figure 4 The Onset of Epidemics and the Winter Maximum Scatter plot of the week of epidemic onset and the week of winter maximum over the 1994–2002 period. Discussion In this paper and a previous publication of ours [17], by using the weekly number of cases of MCM disease in Mali and large-scale fields of surface wind speed, we clearly identify a strong relation between climate and the seasonal pattern of MCM cases in Mali. It is shown that the onset of disease outbreak is characterized by a clear breaking slope in the seasonal cycle of the number of cases at the sixth week of the year, that is, 15 February. The computation of an atmospheric index based on surface wind speed over Mali points out that this abrupt shift is also present in the atmospheric signal, corresponding to the winter wind maximum, when Harmattan winds are the strongest in Sahelo-Sudanian Africa. The similarity in the seasonal patterns of both Harmattan winds and MCM disease cases is obvious, with a strong correlation between the week of winter maximum and that of the onset of epidemic. Similar results, not illustrated here, have been obtained by using surface temperatures and specific humidity for the computation of atmospheric indexes, attesting to the robustness of the analysis. However, whatever the climatic index is used, this analysis does not allow us to link the intensity of the “epidemic” (the annual number of cases) to the intensity of winter in terms of absolute humidity and surface wind speed. This lack of a relation may be due to the time series length, with an insufficient number of years to study interannual variations, or it may imply that the climatic influence is limited to explaining the occurrence of the seasonal cycle of the epidemic and its geographical range distribution, but not its intensity. Although they fail to forecast epidemic intensity, such climatic indexes, with their correlation with the onset and the seasonal course of the epidemic in Sahel, provide an important means of disease monitoring and prediction in Africa. Indeed, the seasonal pattern of humidity and Harmattan winds can be easily tracked, thus promoting the emergence of an Early Warning Index (EWI) for the onset of MCM epidemics. The seasonal forecast of this EWI based on Harmattan winds could thus be implemented routinely by using comprehensive coupled models of the atmosphere, oceans, and land surface that provide a degree of predictability of climate fluctuations with a seasonal lead time in many parts of the world [ 22 ]. The ability of the climate models to predict the winter maximum could be tested by using the outputs of the Development of a European Multi-Model Ensemble System for Seasonal to Interannual Prediction (DEMETER) project, which was conceived and funded under the European Union Fifth Framework Environment Programme. The principal aim of DEMETER was to advance the concept of multimodel ensemble prediction by using a number of state-of-the-art global-coupled ocean–atmosphere models and to produce a series of 6-mo multimodel ensemble hindcasts. The DEMETER project already has application partners in agronomy and in tropical disease prediction [ 22 ]. This EWI parameter, in association with other environmental parameters implicated in disease resurgence [ 23 ], could help to more precisely characterize disease risk maps at regional scales. The natural extension of this work is to relate this information on the timing of disease outbreaks with specific spatial environmental characteristics at finer scales, in an Early Warning System based on the monitoring of the impact of climate variability and environmental change on epidemic occurrence in West Africa. Recent findings by Molesworth et al. [ 23 ] have already quantified the relationship between the environment and the location of the epidemics to propose a model based on environmental variables and to identify regions at risk for meningitis epidemics. The combination of the EWI for MCM epidemic onset and risk maps at regional scales could be a starting point to more optimally direct national and international health policy strategies and to optimize mass vaccination campaigns. In addition, more general measures can be taken by national authorities to improve the control of MCM disease, such as closing markets and schools and discouraging social gatherings when an outbreak is likely to occur [ 9 ]. Patient Summary Background Climate is known to be one of the factors that can affect when and how epidemics occur; for example, floods often increase the risk of waterborne disease. However, there are many more subtle climatic changes that might also be important in affecting when and how diseases occur. What Did the Researchers Do? They looked at the relationship between a recurring epidemic of a disease called meningococcal meningitis in Mali in West Africa and the local climatic conditions, especially the winds. Meningococcal meningitis is a serious infection of the lining of the brain and spinal cord by a bacterium (called Neisseria meningitides ). These researchers had previously published some detailed work on the local climate in a French journal. In this paper they have focussed more on the aspects that deal with disease. They found out that over several years the onset of the epidemic coincided with the peak of the winds. Who Will Use These Results? People who would find these results useful are those who plan for epidemics. Such information will allow them to plan in advance, and even predict whether an epidemic will occur at all. However, these results were based on only the years between 1994 and 2002, and so will need to be confirmed in more years.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545199.xml

509422

Use and improvement of microbial redox enzymes for environmental purposes

Industrial development may result in the increase of environmental risks. The enzymatic transformation of polluting compounds to less toxic or even innocuous products is an alternative to their complete removal. In this regard, a number of different redox enzymes are able to transform a wide variety of toxic pollutants, such as polynuclear aromatic hydrocarbons, phenols, azo dyes, heavy metals, etc. Here, novel information on chromate reductases, enzymes that carry out the reduction of highly toxic Cr(VI) to the less toxic insoluble Cr(III), is discussed. In addition, the properties and application of bacterial and eukaryotic proteins (lignin-modifying enzymes, peroxidases and cytochromes) useful in environmental enzymology is also discussed.

Introduction Chromate reductases are a group of enzymes that catalyze the reduction of toxic and carcinogenic Cr(VI) to the less soluble and less toxic Cr(III). These proteins have recently raised enormous interest because of their central role in mediating chromium toxicity and their potential use in bioremediation and biocatalysis. Chromate (Cr(VI)) is generated as by-product of various industrial processes such as leather tanning, chrome-plating, pigment production and thermonuclear weapon manufacture [ 1 ]. Its high water solubility facilitates a rapid leaching, provoking a wide dispersion capable to contaminate drinking water supplies. Therefore, the characterization of enzymes that reduce chromate, as well as the study of their induction patterns and gene expression are relevant to complete our understanding of chromium metabolism in order to minimize the toxicity of this compound in the environment. The chromate-reducing activities have been located in the cell membrane or in the cytoplasm of many bacteria [ 2 ]. Their ubiquities in many different organisms suggest that they might share a common role in, for example, physiological redox sensing or detoxification. Recently, two novel dimeric flavoproteins with chromate reductase activity, ChrR (from Pseudomonas putida ) and YieF (from E. coli ) have been purified and characterized [ 1 ]. These enzymes were able to transform chromate to the less toxic Cr(III). However, while ChrR was not a pure two-electron reducer of chromate, YieF was able to catalyze a three-electron reduction. The role of ChrR and YieF in protection against chromate toxicity was also investigated and the results suggested that both enzymes may have an important role in protection against chromate toxicity [ 1 ]. The ability of some microorganism and their enzymes to remove toxic pollutants has been recently reviewed [ 3 - 6 ]. The identification and characterization of the degradative pathways functioning in microorganism have been the starting point for biotechnological and environmental applications [ 3 ]. Discussion The intensive industrial and agricultural development has been considered as responsible for a widespread contamination of soil, air and groundwater with toxic pollutants, which are harmful for human health and the environment [ 6 ]. These contaminants enter the environment through different paths, which may include direct application, combustion processes and natural emissions. Major contaminants are polycyclic aromatic hydrocarbons (PAHs), petroleum hydrocarbons, phenols, polychlorinated biphenyls, azo dyes, organophosphorus pesticides and heavy metals [ 3 ]. In particular, Cr(VI) is a common pollutant due to the use of chromium compounds in tanning and other industries. Chromate shares structural similarities with sulphate ion (SO 4 -2 ) and may be introduced in eukaryotic and bacterial cells by the sulphate transport system [ 1 ]. In bacteria, flavoenzymes such as glutathione reductase reduce Cr(VI) by a one electron transfer leading to the formation of the highly unstable radical Cr(V) and the flavin semiquinone form of the enzyme. Both species undergo a further redox cycle in which Cr(VI) is re-generated by one-electron transfer to oxygen, producing and accumulating reactive oxygen species (ROS). The appearance of relatively large quantities of ROS, and the consequent oxidative stress are responsible for the toxic effects and cellular damage attributable to the presence of Cr(VI). On the other hand, trivalent chromium Cr(III) is water insoluble, less bio-available and less toxic [ 1 ]. Thus, the strategies employed to eliminate chromate toxicity would involve its reduction to Cr(III) by chemical or biological means. While chemical methods are expensive at the large scale required to decontaminate waste sites, microorganisms are commonly used for environmental purposes through the exploitation of their natural catalytic activities. Enzymatic treatments have a minimal impact on ecosystems, as they present no risk of biological contamination. Furthermore, enzymes can act over a wide range of pH, temperature and ionic strength and also may be active in the presence of high concentrations of organic solvents in which major pollutant molecules are soluble [ 6 ]. Several bacterial enzymes that can be used in bioremediation have been described; they include mainly oxidative enzymes such as mono-and dioxygenases but their use is restricted by the need of cofactors, which can only be efficiently regenerated inside the microorganism [ 6 ]. In the last two decades bioremediation has explored the use of the catalytic machinery of white rot fungi to remove toxic pollutants. White rot fungi comprise all those fungi capable to degrade lignin, a polyphenolic polymer highly resistant to bacterial biodegradation. Many strains from the genera Pleurotus, Bjerkandera, Phanerochaete , and Trametes produce extracellular enzymes with ligninolytic activity. These enzymes are often referred to as lignin-modifying enzymes and include mainly Lignin peroxidase, Manganese dependent peroxidase and laccase [ 4 ], though some authors have reported other related enzymes such as a Mn-independent MnP activity [ 7 ]. Besides lignin-modifying enzymes, several other enzymes such as the heme-containing peroxidases, chloroperoxidase and horseradish peroxidase, and the non-enzymatic hemeproteins, hemoglobin and cytochrome c , are able to oxidize organic compounds in the presence of hydrogen peroxide. An interesting feature of these enzymes is their remarkable low specificity towards substrates that arises from their own catalytic mechanism. In vivo , peroxidases use endogenous low-molecular weight compounds, called mediators, to generate free radicals capable to carry out a wide variety of reactions such as oxidations, bond cleavage, hydroxylations, polymerization and demethylation [ 4 ]. Several research efforts have been focused on the ability of peroxidases to degrade pollutants such as PAH's, azo dyes and organophosphorus pesticides [ 8 - 10 ]. Strong regulations have been established to push the industrial sector to develop new programs destined to a greater environmental care. Nowadays, industry is strongly dependent on petroleum and its derivatives as a source for raw materials and energy. There are still large reserves of crude oil, which are heavy oils with a high content of sulphur and heavy metals. The use of these fuels generate a great pollution, being one of their most important environmental impacts the formation of the acid rain which takes place by the sulphur oxide production during combustion. Redox enzymes may encounter fields of application not only in the bioremediation of polluted environments, but also in the development of novel clean technologies to avoid or diminish the environmental contamination. Biocatalytic methods for sulphur removal from straight-run diesel fuel have been developed [ 11 ]. The removal of heavy metals from the petrophorphyrin-rich fraction of asphaltenes has also been reported [ 12 , 13 ]. Thus, enzymes can play an important role in the development of alternative or complementary biotechnological processes with potential application in polluting industries. Despite their potential application in bioremediation and clean processes, the activity of oxidative enzymes may be limited, among other factors, by the low bioavailability of the pollutants and by the relatively low operational stability of the enzyme under the environmental conditions required to carry out the bioremediation. Several strategies to increase the catalytic activity of peroxidases have been proposed, including chemical modification of the enzyme [ 14 , 15 ] and genetic tools [ 16 ]. Recently, with the cloning and expression in suitable hosts, larger amounts of the desired enzymes may be produced, facilitating their characterization and their direct use in environmental applications. Further, through the use of novel techniques such as directed molecular evolution [ 17 ], proteins designed specifically for bioremediation could be made available in a not distant future. In the last few years there has been an extensive research in the application of laboratory evolution for tailoring redox enzymatic systems (laccases, peroxidases, cytochrome P450 monooxygenases) to improve their activities and stabilities against temperature or organic solvents [ 18 - 20 ]. The application of this powerful approach for bioremediation issues is coming up, but first a big effort in the high-throughput (HTP) screening methodology must be done. So far, little has been reported on the optimisation of suitable HTP for the detection of xenobiotics [ 21 ]. Therefore, the success in the enzyme evolution for environmental issues will be highly dependent on the automation of HTP. Microbial genomics is a new emerging field that enables us to look at parts of the environment that were, until recently, masked to us. Present estimations suggest that more than 99% of the microorganisms in most environments (also those subjected to chronic contamination) are not amenable to grow in pure culture, and thus very little is known about their enzymatic activities. We can now access the genomes of non-culturable microorganisms through creating the "so-called" metagenomic-libraries and identify protein-coding genes and biochemical pathways that will shed some light on their properties and function [ 22 ]. New enzymatic systems found in contaminated areas can be used as parental types for some rounds of directed evolution with the main aim of improving the catalytic performance for their use towards solving a broad range of environmental problems. Conclusion The even more strict regulations on hazard wastes has forced to the development of new environmentally compatible strategies to substitute or complement the conventional ones. Chemical technologies are expensive when applied to large scale and in many cases are technically not feasible. On the other side, by exploitation of the huge diversity of natural activities and metabolic pathways presented by microorganisms, new strategies can be envisaged. The use of oxidative enzymes as biocatalysts for environmental purposes presents a promising potential due to their low specificity and low energetic requirements. However, further characterization of new biocatalysts is needed. The use of novel technologies such as molecular directed evolution may have a large impact in the tailoring and further application of enzymes not only for bioremediation but also for the development of friendly environmental technologies.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509422.xml

300884

Structure and Implications of JAMM, a Novel Metalloprotease

Proteins may be the workhorse of the cell, but when a cell can synthesize one protein in a matter of minutes, chances are some will become obsolete. Though many proteins put in years of productive service, others quickly outlive their usefulness and can even damage the cell. Proteins that help form bone and muscle, for example, function for years while regulators of mitosis and cell proliferation might finish their jobs in seconds. Such short-timers are soon tagged as superfluous by a chain of small proteins called ubiquitin, which marks the proteins for degradation in an enzyme called the proteasome. Once in the proteasome, these proteins are broken down and can then be recycled for more productive ventures. A massive structure by cellular standards, the proteasome consists of multiple subunits, including a cylindrical core particle called 20S, which catalyzes degradation, and regulatory complexes called 19S caps, which form lid and base structures at both ends of the core. While the structure and biomechanics of the 20S core have been well characterized, much less is known about the functional mechanics of the regulatory complexes. The lid--base complex recognizes only ubiquitin-tagged proteins, which are then unfolded so they can enter the proteasome. But first ubiquitin chains must be detached from the protein, a task performed by an enzyme in the proteasome called Rpn11 isopeptidase. How the lid–base complex removes the ubiquitin tag, unfolds the protein, and shuttles it into the proteasome's core is not clear. Now Raymond Deshaies and colleagues present the structure of a homolog of the 19S lid's isopeptidase enzymatic center and provide new insights into these questions. The proteasome Rpn11 subunit contains a key region called the JAMM motif, which Deshaies' lab has shown previously is required for the proteasome to remove ubiquitin tags. For the work discussed in this paper, the researchers set out to understand how the proteasome strips off ubiquitin tags from proteins about to be destroyed by determining the three-dimensional structure of the JAMM motif. The researchers tested many genes to look for a JAMM-containing protein that would crystallize properly and found one in the heat-loving prokaryote Archaeoglobus fulgidus . After determining the structure of the JAMM protein (called AfJAMM), the researchers discovered that AfJAMM looks nothing like the well-known deubiquitinating enzymes. But the arrangement of a set of amino acids that binds a zinc ion and forms the proposed active site of AfJAMM does resemble that found in a well-known protein-degrading metalloprotease called thermolysin, even though in other respects AfJAMM and thermolysin have very different features. The researchers mutated amino acid residues in another JAMM protein called Csn5 (they expected these residues to be critical for isopeptidase activity as well, based on comparisons of the AfJAMM and thermolysin structures) and found that the residues are indeed important for Csn5 function. These results suggest that JAMM does indeed represent a novel family of metalloproteases. As for the wider function of JAMM proteins, the researchers speculate that these proteins are likely to be involved in a variety of important regulatory systems since they appear in life forms that lack ubiquitin and ubiquitin-like proteins. The crystal structure reported in this paper will provide a valuable tool for investigations into the underlying structural and functional mechanisms of these enzymes. And it may have important therapeutic implications. Proteasome inhibitors are promising anticancer therapies—fighting cancer by blocking machinery required by rapidly dividing cells. In the hopes of developing more targeted therapies, scientists are trying to fine-tune their control of the ubiquitin system and the proteasome. Inhibiting the JAMM domain of enzymes like Csn5, which remove ubiquitin-like tags from proteins upstream of the proteasome, for example, might just do the trick. The active site of JAMM

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC300884.xml

554094

Ultrasound imaging versus morphopathology in cardiovascular diseases. Coronary collateral circulation and atherosclerotic plaque

This review article is aimed at comparing the results of histopathological and clinical imaging studies to assess coronary collateral circulation in humans. The role of collaterals, as emerging from morphological studies in both normal and atherosclerotic coronary vessels, is described; in addition, present role and future perpectives of echocardiographic techniques in assessing collateral circulation are briefly summarized.

In the past 25 years, the concept of a compensatory function of the coronary collaterals (or anastomoses) – i.e. vessels that join different coronary arteries or branches – has been practically cancelled from the mind of cardiologists since cineangiography shows that the onset of coronary heart disease (CHD) occurs independently of their presence. The assumption, therefore, was and is that they have no compensatory meaning [ 1 ] and coronary obstruction causes ischemia. A crucial and questionable assumption which disregards solid and recognized pathological data and supports invasive therapies, the diagnostic gold standard being the coronary cineangiography. In many cardiological centers, at the first chest discomfort, the latter is the guide for emergency angioplasty + stent or surgical bypass when a coronary ostruction is found; with the belief that a severe coronary stenosis causes angina pectoris, its occlusion an acute myocardial infarct (AMI) or sudden death (SD) and chronic ischemia explains hibernating myocardium. By injection under controlled pressure of plastic materials through the aorta, casts of coronary arteries, including coronary ostia, in normal and pathological hearts were obtained. They gave an objective tridimensional view of anatomy, different patterns of coronary distribution and overall collaterals in relation to coronary lumen reduction. The method allowed a histologic control of the myocardium [ 2 - 4 ]. The casts of normal coronary arteries showed a smooth surface without identations easily identified when even a minor lumen reduction was present. In hearts of normal people dead by accident without pathological findings at autopsy, homocoronary (between branches of the same coronary artery) and intercoronary (between different coronary arteries) anastomoses were present everywhere joining at any level the intramural branches. Only in two of more than 600 hearts, superficial collaterals between extramural coronary arteries were seen and sampled for histology. The diameter of the innumerable normal collaterals ranged from 20 (maximal penetration of plastic injection) to 350 microns, frequently assuming a corkscrew aspect, possible adaptation to the contraction cycle of the myocardium (Figure 1 ). The first conclusion was that arterial intramural system, including the terminal bed, is an anastomotic network, at least from the anatomical viewpoint. Figure 1 Coronary anastomoses or collaterals. A) intercoronary ventricular and (B), atrial. C) homocoronary anastomoses. Note the innumerous collaterals joining different intramural branches at any level of their course. They have frequently a corkscrew aspect (D) visible also histologically (E), as adaptation to cardiac contraction-relaxation cycle. In hypertrophic hearts with normal coronary arteries and in normal hearts of patients with chronic hypoxia, e.g. anemia, collateral diameter and length were increased in the whole intramural system (500 microns; Figure 2 ). The more impressive change was seen in presence of coronary stenosis greater than 70% with a diameter and length exceeding 1000 microns and several centimeters respectively (Figure 3 ). The other peculiarity was that collateral enlargement was strictly related to a stenosis filling distal tract of the obstructed vessel ( satellite anastomoses ); when more than one severe stenoses exist each one had its own satellite collaterals. However, an identical obstruction located at the same level of an artery might show relatively few highly enlarged collaterals (the only ones visible by cineangiography), or numerous relatively small collaterals (Figure 4 ). A finding possibly due to a redistribution of blood flow consequent to newly formed severe stenoses or an infarct. In the latter condition, all vessels within the necrotic tissue disappear ( avascular area seen in plastic casts; Figure 5 ) and the surviving collaterals at periphery will further enlarge since the pressure gradient distal to the coronary obstruction persists. Figure 2 Vessel changes in relation to modification of the cardiac mass. A) atrophic heart with acquired serpentoid form of extramural vessels due to cardiac mass reduction, and minor intramural vascularity. The contrary is seen in cardiac hypertrophy (B) in which the extramural arteries increase in length and diameter (but not in number) to adapt themselves to the greater myocardial mass. Similarly, the same enlargement is seen in the intramural branches. Cor pulmonale, in which condition the right ventricle may become greater than the left one, is an extreme example of adaptation of extramural (C) and intramural, including collaterals (D). No histologic evidence exists of new vessel formation. The cardiac vein show a similar behaviour. Figure 3 Collateral enlargement in topographical relation (satellite) with severe stenosis or occlusion. A double occlusion of LAD (anterior view) and occlusion of RCA (posterior view) apparently compensated by enlarged collaterals in a non cardiac patient dead from brain hemorrhage. B, similar condition in cases with RCA occlusion (arrow) without corresponding myocardial infarct with numerous homo and intercoronary collaterals of the anterior wall (C), and (D) septum. Occlusion of LAD without evidence of other stenotic changes of the coronary arteries in a 39-year-old woman with rheumatic heart disease and mitral insufficiency. In this case, arteritis was documented histologically by sampling before corrosion. An acute infarct (avascular area at the apex, arrow) was present. F, a single, high enlarged collateral from LCX, supplying the distal tract of an occluded LAD. Note, numerous normal anastomoses. This indicates that ischemia is not the cause (no diffuse enlargement of all collaterals in the whole ischemic area) but rather pressure gradient induces selective compensatory routes. Figure 4 Different aspects of collateral compensation in presence of the same occlusive pattern of LAD. A, relatively few very enlarged collaterals and (B) numerous relatively small collaterals. This divergency may be due to progressive atherosclerotic obstruction of other main vessels or lost of the intramural vasculature, including collaterals, following an infarct. Chart C shows all the possibilities of flow redistribution. The histology of the enlarged anastomoses corresponds to a capillar-like wall, even in the rare extramural collaterals with rudimentary focal tunica media (C). D), enlarged collaterals in a case of anomalous origin of LAD from the pulmonary artery and (E,G) different aspects of giant capillaries (or plexus) in various stages of an acute/old infarction. The absence of new vessel formation is well documented in recent infarcts associated with endocardial thrombus (G). In the latter numerous new vessels form in the granulation tissue repair of the thrombus in contrast to their absence in infarct (arrow; postmortem coronary injection for vessels identification). Figure 5 Avascular area of an infarct. By plastic cast (A anterior, B posterior view) or postmortem angiogram (C) the infarcted zone (arrow) lacks of intramural vessel injection ("avascular area"). Stretching of the necrotic myocardium and secondary vascular damage with wall degeneration and thrombosis (D), explain this vascular "sequestration" which occurs in early phase. This may indicate a blockage without possibility of therapeutical intervention via blood flow within the infarcted myocardium. Note that the avascular area in this AMI case documented histologically, depended from LAD without evidence of occlusion or severe stenosis. The occluded vessel (arrow) was (B) the RCA, the distal part of which was filled by numerous anastomoses. No myocardial damage was seen in its territory. By dissection even an expert pathologist, the diagnosis could be of myocardial infarction following occlusion of the RCA. E) obliterative intimal hyperplasia in arterioles around a seven days old infarct with early repair process. Another satellite collateral system is annexed around and within the atheroclerotic plaque. Plastic casts and histological serial sections showed an extensive vascularization limited only at plaque level and formed by giant adventitial capillary-like vessels filled by intracoronary radiopaque injected material, connecting secondary branches proximal and distal to the stenosis as well as new vessels formed within the atherosclerotic intima i.e. arterioles, with a well developed tunica media, related to angiomatous plexuses which open in the residual lumen (Figure 6 ). This plaque satellite system may explain why by cineangiography the coronary tract distal to stenosis is immediately filled while in its absence a delay or flow reduction should be expected. Figure 6 Vascularization of a coronary atherosclerotic plaque showing different aspects of neovascularization. By serial sections of postmortem injected plaques, giant advential capillary-like vessels (A) are connected with secondary branches proximal and distal to the plaque and with new arterioles (B) with a well developed tunica media (indication of functioning blood flow), within the thickened, atherosclerotic intima in turn joined through angiomatous plexuses (C) to the residual lumen (D) E) plastic casts of plaques with different aspects of vascularization. Both homo-intercoronary and plaque collateral systems are anatomical structures capable to adapt in particular pathological conditions. The question is whether or not they are able to prevent ischemia and compensate an occlusion which by cineangiography appears as a "cut off" of a vessel without imaging of its distal tract. It must be stressed that in postmortem casts with coronary occlusion the latter was always injected through collaterals. In 87% of AMI patients, within four hours from clinical onset, a cineangiographic occlusion was observed and in 88% of cases undergone emergency bypass surgery, a "layered thrombus" was recovered "proximal to stenosis" [ 5 ]; a thrombus due to plaque rupture [ 6 - 8 ] causing the infarct or sudden death. In discussing this dogma the first need is to review the function of the collaterals. Collateral function Capillary function in presence of normal coronary arteries In normal hearts and in pathologic hearts with normal coronary arteries, the collaterals, due to their capillary structure, participate to the metabolic exchange as terminal capillary bed. This means a much greater extent of the exchange surface which invalidates any "one myocardial / one capillary" model to study the delivery of any substance from capillary to myocardial cell. The myocardial interstitium is crossed by a myriad of "endothelial" vessels in any direction. Compensatory function in presence of coronary obstruction The demonstration of tridimensional collateral enlargement by casts indicates, per se, that there was an increased blood flow. Their adequacy to compensate one or more severe coronary obstructions is documented by the following main facts: 1. At the first episode of coronary heart disease (CHD) in apparently healthy people acting their normal life, 89% with a fatal AMI had one or more (47%) severe atherosclerotic stenosis greater than 70% ;65% of sudden and unexpected death (SUD) showed the same finding in one or more (35%) vessels; 66% of non cardiac patients dead from other diseases and 39% of normal subjects dying from accident had the same severe atherosclerotic stenosis in one or more (40% and 16% respectively) coronary arteries (Table 1 ). At histology, all plaques were old lesions preexisting months or years without any evidence of CHD despite a stressful life and in absence of a myocardial infarct. The only explanation is that the collateral system was able to fully compensate the blood flow reduction consequent to the stenoses. Table 1 Maximal atherosclerotic lumen diameter reduction and number of main arteries with severe (≥ 70%) stenosis Source Acute myocardial infarct Sudden death unexpected Non cardiac atherosclerotic Patients Accidental death in normal people 1st chronic 1st chronic Cases 145 55 133 75 100 97 % Lumen reduction 0 3 - 10 - 7 8 <50 3 1 18 - 10 20 50–69 10 - 18 5 17 31 70–79 30 8 21 8 11 19 80–89 45 11 39 14 24 13 ≥ 90 54 35 27 48 31 6 No. arteries ≥ 70% 1 61 16 40 13 26 22 2 49 22 34 26 18 13 > 3 19 16 13 31 22 3 1 st episode, in apparently normal people without extensive monofocal myocardial fibrosis Chronic, in subjects with history of coronary heart disease and/or extensive myofibrosis. 2. Myocardial infarct size measured planimetrically was not related to the number of severe coronary stenoses found in each AMI case (Table 2 ) as should be. More severe coronary stenoses should determine a higher ischemia resulting in larger infarcts. Table 2 Lack of correlation between number of severe (≥ 70%) coronary stenoses and acute infarct size (% left ventricular mass) in 200 consecutive and selected fatal cases. Source Acute myocardial infarct Cases 200 97 103 ≤ 20 size > 20 Lumen reduction < 69 7 10 ≥ 70 90 93 in 1 39 38 2 37 34 ≥ 3 vessels 14 21 p < 0.05 for trend 3. No relation between the total vascular territory of obstructed coronary artery and infarct size which often extended in territories of non stenosed or occluded vessels. In vivo hypokinetic zones expand in well perfused region [ 9 ]. 4. The relatively frequent finding of a coronary occlusion without an infarct. 5. In an experiment done in a leading dog lab, a controlled coronary stenosis, maintained for few days and then occluded, did not determine any dysfunction or infarct because a dramatic increase of collateral flow [ 10 - 12 ]. These are the main facts supporting the concept that collaterals shown postmortem succeed in limiting or abolishing ischemia induced by coronary obstruction and question the existence of chronic ischemia due to coronary atherosclerosis since a plaque takes time to develop while collaterals [ 10 , 11 ] adapt itself quickly as soon a pressure gradient between stenosis and distal vessel is established. On the other hand, there is no demonstration of a possible failure, both acute or chronic, of collaterals; including spasm since they have not tunica media. The inability of cineangio imaging to visualize collateral systems is explained by its very limited power of resolution of all intramural vessels and by the selective injection of radiopaque labelled blood flow in one coronary artery competing with non labelled flow from the other coronary artery. Only very enlarged intercoronary anastomoses can be seen cineangiographically without any value in relation to cardiac dysfunction. Acute ischemia induced by balloon inflation at angioplasty may depend on sudden occlusion by compression of the collateral plaque system. Active coronary atherosclerotic plaque according to cineangio imaging Active plaque means an impending infarct expressed by a variety of angiographic signs as irregular lumen, haziness with ill-defined margins, smudge appearance, inhomogeneity, opacification, luciencies, persistence of radiopaque material, etc. Signs difficult to correlate with postmortem findings since terminal changes can not be excluded. They may represent the irregular vascularization of the atherosclerotic plaque opacified by the injected radiopaque material. Worthy of note is that cineangio defects can persist unchanged per years [ 13 ]. Cineangio coronary occlusion The very high frequency of coronary occlusion seen angiographically in AMI patients (see above) does not correspond to that observed in pathological studies in which the mean figure is 50% for AMI and 29% for SUD patients. Nevertheless, different selection of material, divergent definition and an absence of a correct correlation of all pertinent variables give reason of dissimilar conclusions. In 200 selected AMIs and 208 SUD cases the unique cause of occlusion was a thrombus found in 41% and 29% respectively. In AMI group it correlated significantly with a lumen reduction greater than 70% (93%), length of plaque more than 6 millimeters (95%), its concentric shape (100%), prevailing atheroma (84%), medial neuritis (92%) infarct size greater than 50% (86%). SUD cases showed a similar behavior. In reality, both clinicians and pathologists observe a phenomenon which started before, missing its onset and sequence of events to distinguish whether primary or secondary. In only one case reported in literature [ 14 ], this sequence and histological examination of the whole heart was possible in a 45 year old man suffering a two months unstable angina. At coronary cineangiography there were two critical stenoses of the left anterior descending branch (LAD), one proximal and another distal to the origin of diagonal branch and a critical stenosis in the first tract of the right coronary artery (RCA). An antero-septal-lateral hypokinesis was documented. After the fourth left coronary injection, in absence of any symptom or sign and cineangio imaging changes, a first ECG showed downsloped ST segment. The latter persisted for other four LAD injections when the vessel disappeared, again, without any subjective and objective signal. Intracoronary vasodilator and fibrinolytic agents, successful angioplasty in reopening critical stenoses, surgical bypass in rapid sequence were performed without re-establishing flow. Only for few short periods a reflow occurred with an imaging of occlusion which from the distal tract ascended till the origin of LAD (Fig. 7 ) and not at the site of angioplastically reopened stenoses. An interesting note is that a severe chest pain started after angioplasty, 90 minutes from the first ECG change. The patient survived an extensive myocardial infarction and 12 months later underwent heart transplantation because irreversible congestive heart failure. We had the opportunity to examine the heart removed at surgery confirming a large (40% of the total left ventricular mass) antero-septal-lateral scar, end result of the infarct, scattered foci of fibrosis everywhere, absence of small vessel disease, colliquative myocytolysis expression of congestive failure, severe lumen reduction by sclerosis of LAD (90%) – despite it showed a normal lumen at bypass surgery – and vein graft (80%) without evidence of thrombosis, RCA occlusion by an organized thrombus located in an atherosclerotic plaque with 90% lumen reduction, medial neuritis i.e. lympho-plasmacellular inflammation of nerves closed to the tunica media, in all atherosclerotic plaques, absence of an infarct in RCA territory. Figure 7 Cineangiographic monitoring in a patient with non occlusive LAD stenosis (A) who developed an extensive infarct without angiographic occlusion. The subsequent imaging of occlusion began distally (B) and ascended to the origin (C) of the vessel (arrow) indicating that the angiographic "pseudocclusion" was due to stasis for increased peripheral resistance and not for primitive thrombosis, not shown morphologically (see text). One case is only one case but when for the first time shows how the events developed, it becomes a precious mile stone for our knowledge demonstrating that the cineangio occlusion was a pseudocclusion namely a blood flow stasis in LAD secondary to an increased intramyocardial resistance. The first main question is how many of the 87% cineangio occlusion are pseudocclusion and whether the "layered" thrombus recovered at bypass surgery was a true thrombus or a coagulum which frequently show a layering of blood elements not seen in thrombus formation. "Red" thrombus, namely a coagulum, is frequently and erroneously considered as thrombus. The second question concerns the nature of increased intramyocardial resistance: spasm of intramural arterial vessels or their extravascular compression by an asynergic myocardium? The first sign of CHD is hypokinesis of a myocardial zone which particularly in systole may compress vessels. Any time there is an increase of the intraventricular pressure with bulging of hypokinetic myocardium such a compression may abolish blood flow with subsequent infarction. In the reported patient location and infarct size corresponded to the hypokinetic area observed before the infarct onset. A last comment deserves the supposition that small atherosclerotic plaques undetectable at cineangio, may rupture causing an infarct. A supposition based on the cineangio finding of a non critical stenosis observed in a vessel tributary of a territory in which an infarct will develop. Since, when the latter occurred, stenoses in other non supplying vessels did not show a further lumen reduction, the conclusion was that even the plaque related to infarction had a non critical lumen reduction [ 15 ]. A conclusion that ignores the following two main facts. First that no one pathological study demonstrated the rupture of a small plaque associated with a thrombus occluding a normal or mild stenotic lumen. Second, myocardial asynergy by increasing intramyocardial resistance, promotes plaque progression by an increased dynamic stress on wall of the supplying extramural artery. For instance, in the previous case both LAD and vein graft with a normal lumen at surgery, in 12 months became critically stenotic (90% and 80% respectively). Regional myocardial dysfunction is an important cofactor in accelerating atherosclerosis lesion in related artery. Target of ultrasound diagnosis: the present and the future In the past years, clinical methods available to measure collateral flow have been too crude and showed major limitations, thus contributing to debate and confusion about the functional relevance of collateral circulation in the human myocardium. Coronary angiography allows visualization of collateral vessels having a diameter ≥100 μm, that actually prevents the majority of them from being detectable with this technique [ 16 , 17 ]. On the other hand, scintigraphic perfusion imaging techniques have limited spatial resolution [ 18 ]. Intracoronary wedge pressure and Doppler flow velocity measurements clearly demonstrated the presence of considerable collateral flow even in patients without angiographic evidence of collaterals [ 19 , 20 ], but they are invasive and not suitable for routine clinical use. With the introduction of new generation echo contrast agents and advanced ultrasound techniques, myocardial contrast echocardiography (MCE), an ultrasound imaging technique that utilizes physiologically inert gas-filled microbubbles as red blood cell tracers, has gained importance for the non-invasive assessment of blood flow at the level of myocardial perfusion [ 21 , 22 ]. Although evaluation of viability is the main clinical application of MCE [ 23 ], indirect assessment of collateral derived myocardial perfusion has been described in different clinical and experimental settings. In patients with severe left coronary artery disease, the placement of a graft to the posterior descending coronary artery was found to improve the collateral derived peak contrast effect within the anterior left ventricular wall [ 24 ]. In a series of subjects with healed myocardial infarction and total occlusion of the culprit vessel, a correlation was found between angiographic collateral grade and peak contrast effect after contralateral intracoronary contrast injection [ 25 ]. Collateral perfusion detected by MCE paralleled changes detected by radiolabeled microspheres during thrombosis and vasodilator administration in a canine model [ 26 ]. The usefulness of MCE has been confirmed in subjects without coronary occlusion where it was able to map the myocardial territory perfused by coronary collateral flow and to evidence immediate reduction of perfusion when collateral flow was abolished by angioplasty [ 27 ]. In patients with no prior myocardial infarction undergoing coronary angiography, intracoronary MCE effectively quantified coronary collateral flow, as demonstrated by the linear correlation existing between peak echo contrast effect and collateral flow index determined by intracoronary wedge pressure [ 28 ]. On the other hand, a strong correlation was reported between collateral receiving area at MCE and regional wall motion score index in patients with coronary occlusion, thus providing evidence that collateral derived perfusion is a good indicator of preserved regional function [ 29 ]. Likely, the grade of collateral flow on MCE was inversely correlated to the infarct size and was able to predict functional improvement following coronary revascularization [ 30 ]. Using an experimental model of chronic ischemia, it was found that not only the presence of collaterals can be identified by MCE, but also that temporal and spatial development of collateral circulation can be tracked serially [ 31 ]. Finally, intravenous MCE has been recently reported to provide qualitative and quantitative evaluation of collateral blood flow in the presence of an occluded infarct-related artery, and to emerge as the only predictor of true collateral blood flow among other markers [ 32 ]. All these reports as a whole support the concept that MCE provides important information on collateral flow and represents a promising mean for evaluating the status of coronary collateral circulation in clinical practice. Some important caveat , however, have to be taken into account. First, although the peak contrast pixel intensity has been reported as the most accurate of the variables obtained to measure collateral flow, there is a remarkable scatter in the correlation between peak pixel intensity and true collateral flow [ 33 ]. Second, it is known that regional contrast heterogeneity is common, resulting in frequent false positive perfusion defects [ 34 ]. Finally, coronary collateral vessels may cause additional dilution of contrast affecting the transit rate calculation. Further technical improvements may contribute in the near future to ensure standardization of the acoustic window and provide a quantitative evaluation of collateral flow. These issues appear to be of crucial importance to turn the echocardiographic assessment of coronary collateral flow into a ready-to-go clinical tool. Besides the attempt to obtain direct echocardiographic assessment, coronary collateral circulation can indirectly affect the result of diagnostic stress testing with the use of echocardiographic technique. Increased vulnerability to myocardial ischemia induced by pharmacological coronary vasodilation was reported consistently with the hypothesis of a facilitated steal phenomenon in the presence of good collateral circulation [ 35 ]. On the other hand, the role of collaterals against echocardiographically-assessed stress-induced myocardial ischemia is controversial, some Authors reporting a protective [ 36 ] and others a neutral [ 37 ] effect. However, dobutamine-induced wall motion worsening in myocardial territories supplied by occluded epicardial vessels has been reported in case of evident collateral circulation [ 38 ], thus emphasizing the importance of a preserved, though reduced, blood flow to distinguish jeopardized myocardium from necrotic tissue. Differently, the ability of low-dose dobutamine stimulation to identify myocardial regions with a high probability of functional improvement after revascularization seems to be independent of both severity of underlying coronary stenosis and degree of collateralization of the involved coronary vessel [ 39 ]. The application of low-frequency ultrasound to intravascular microbubble contrast agents has been receiving attention in the last few years due to its potential therapeutic application, primarily as targeted gene delivery systems [ 40 ]. Further evidence from experimental studies has shown small capillary ruptures in exteriorized rat skeletal muscle [ 41 ], intact mouse muscle [ 42 ] and rabbit myocardium [ 43 ] to follow the application of ultrasound power. However, capillary rupturing via microbubble destruction with ultrasound is able to enhance arterioles per muscle fiber, arteriole diameters, and maximum nutrient blood flow in skeletal muscle [ 44 ]; thus, it may be tailored to stimulate an arteriogenesis response that restores hyperemia blood flow following arterial occlusion [ 45 ]. The potential of this method to become a clinical tool for stimulating blood flow to organs affected by occlusive vascular disease and, in particular, to the myocardium represents an interesting track for future research involving the application of ultrasound technology in the ischemic heart disease. Final consideration on coronary atherosclerotic plaque Any hypothesis on the pathogenic role of a plaque and its activity and vulnerability should consider all interrelated variables for a correct interpretation of findings. When only one or few variables are investigated erroneous conclusions can be reached. An atherosclerotic plaque is always an active structure since its progression depends on a sequence of events due to a variety of correlated phenomena; while vulnerability is just an hypothesis which believe that some findings indicate a risk of plaque rupture. The known variables are: degree of lumen reduction, shape, length, satellite collaterals, tunica media changes, inflammatory reaction per se and associated with media nerves (medial neuritis), survival (Table 3 ) macrophagic repair process, inflammation, vascularization hemorrhage, proteoglicans, atheroma, calcification, smooth muscle cell and elastic fiber hyperplasia, rupture, thrombosis, various factors released from all involved cells, hemodynamic pressure stresses, regional myocardial asynergy, spasm plus still unknown variables to be included. Table 3 Occlusive coronary thrombus versus significantly main correlated variables. Percentage distribution Source Acute myocardial infarct Sudden unexpected death Cases Total 200 208 Cases+occlusive thrombus% 41 15 Lumen reduction% ≤ 69 7 - 70–79 33 16 80–89 35 47 > 90 24 38 Length stenosis mm ≤ 5 6 6 5–20 38 19 > 20 56 75 Concentric 100 94 Atheromatous 84 75 Medial neuritis 92 92 Infarct size % ≤ 10 20 - 11–20 32 - 21–30 48 - 31–40 44 - 41–50 78 - > 50 86 - Survival days ≤ 2 29 3–10 51 11–30 45 Survival minutes < 10 - 12 10–60 - 23 61–180 - 30 Most studies analized few variables mainly observed in animals after hypercholesterol diet or in familial hypercholesterolemia. A pattern [ 46 , 47 ] totally different from that seen in general population and CHD. Furthermore, myocardial infarction is not synonymous of sudden/unexpected death, thrombus is a totally divergent structure from coagulum, collaterals can not be ignored and meaning of the coronary atherosclerotic plaque can be interpreted in another way. The presence of functioning collaterals induces a particular hemodynamic condition within the residual lumen at the plaque level with proximal flow reduction counterbalanced by distal collateral flow. Any time there is a regional asynergy (Figure 8 ) with increasing intramural resistance, stasis in related artery will result in blockage of flow within the lumen with the most favourable situation for intimal hemorrhage, rupture, and thrombosis as secondary phenomena and not primary cause of an infarct. It is hard to believe that occlusion of a pinpoint lumen already compensated by collaterals is the cause of an infarct and rupture of a cap causes infarct or sudden death; being clear that any acute coronary syndrome is an etiopathogenetic entity which can not be caged in any unifying theory [ 48 ]. In the next review on different types of myocardial damage, this argument will be further reconsidered. Figure 8 The coronary thrombus is a multivariant phenomenon (A), including medial neuritis. Its location in severe (≥70) stenosis associated with other factors (retrograde collateral flow, reduced fibrinolytic activity, etc, see text) justifies the concept that is a secondary phenomenon. Any time there is an increased peripheral resistance (B) (spasm, intramural extravascular compression following infarction, etc), stasis in related main vessel and in collaterals both outside and within the plaque is expected with hemorrhage, plaque rupture and trombosis (C). On the other hand, it is difficult to accept that acute occlusion of a pin-point lumen bypassed by preexisting functioning collaterals (D) may result in infarct necrosis or sudden death. Even experimentally occlusion of a severe "chronic" (7 days) stenosis does not produce any ischemic dysfunction. Authors' contributions Prof. Giorgio Baroldi contributed to the conception and organization of this review and to the final comments. Dr. Riccardo Bigi and Dr. Lauro Cortigiani summarized the use of ultrasound techniques in atherosclerotic plaque imaging.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554094.xml

490028

Genome-Wide Survey of Cohesin: A Molecular Guardian of Genomic Fidelity

At a fundamental level, the continuity of life depends on cell division. Humans generate many millions of cells per second just to stay alive, with most cell types dividing and multiplying repeatedly during a lifetime. Details of cell division vary from cell to cell and organism to organism, but certain features are universal, including what is arguably a cell's most crucial task: the faithful duplication and segregation of its genetic material. During mitosis, a cell copies its nuclear DNA, then splits into two identical daughter cells, a process that involves moving the replicated chromosomes (called sister chromatids) toward opposite ends of the cell. After chromosomes replicate, a protein complex called cohesin binds the sister chromatids together. Cohesion helps the cell distinguish between the copies, which in turn aids proper distribution. Improper sister chromatid segregation can yield an abnormal number of chromosomes (called aneuploidy) in the daughter cells, a condition associated with cancer. During meiosis—the cell division that produces egg and sperm cells—aneuploidy causes a number of congenital disorders, including Down's syndrome. PeakFinder automates identification of peaks in ChIP data To end up in their appropriate positions, sister chromatids must establish attachments to tentacle-like protein polymers called spindle microtubules, which emanate from spindle poles at opposite ends of a cell. Cohesion between the chromatids makes these bipolar attachments possible and keeps sister chromatids from separating after they attach to the spindle. Cohesion occurs along the length of a chromosome and is particularly strong around centromeres, the pinched region of a chromosome. Centromeres, in turn, assemble another protein complex called the kinetochore, which mediates the attachment of chromosomes to spindle microtubules; together, they guide chromosomes to their respective destinations. Cohesin's binding locations were discovered by removing chromatin—the mass of DNA and proteins that forms chromosomes—from cells, and purifying the regions associated with cohesin. These studies looked at cohesin's binding distribution either genome-wide or at select regions of a few chromosomes. Here, two research groups use a similar approach to provide a broader picture in their analysis of cohesin binding in the budding yeast Saccharomyces cerevisiae , a favorite system for cell biologists. In the first paper, Jennifer Gerton and colleagues generated a map for the entire yeast genome of locations where cohesin binds to chromosomes during meiosis and mitosis. In the second paper, Paul Megee and colleagues found that centromeres attract large concentrations of cohesin to their flanks and that the assembly of these cohesin domains is mediated by centromere–kinetochore complexes. Gerton's group reports that large regions surrounding centromeres have “intense” cohesin binding. These binding sites correlate with DNA base composition—DNA is composed of four chemical bases, or nucleotides, that are referred to as A, C, G, and T—showing a strong association with AT-rich regions. In meiotic chromosomes, cohesin binding sites are interspersed between the DNA double-strand breaks that initiate the exchange of genetic information characteristic of meiosis, perhaps keeping the chromatids attached without interfering with genetic recombination. Most striking, the authors note, is the observation that cohesin binding changes according to the cell's gene transcription program. Cohesin prefers DNA that lies between active transcription zones and is unceremoniously displaced from regions where RNA transcripts are being made (a process called elongation). This suggests that elongation through a region and cohesion binding may be incompatible. These observations support previous work indicating that DNA sequences required for the replication and segregation of chromosomes must be protected from transcription to function properly. Whatever the explanation, this finding begs the question of how more complicated genomes can accommodate these two seemingly contradictory processes. Megee's group investigated whether all yeast chromosomes have these large centromere-flanking cohesin regions and whether the centromeres and DNA sequences that surround them somehow facilitate the assembly of cohesin complexes. By removing centromeres and generating cells incapable of assembling kinetochores, the researchers show that the assembly of these cohesin regions is mediated solely by the centromere–kinetochore complex. What's more, inserting centromeric DNA sequences in abnormal chromosomal locations produced new cohesin-assembling regions around these “neo” centromeres. The kinetochores' influence appears to stretch over tens of thousands of DNA bases and serves chromatid segregation in two crucial ways: by recruiting high levels of cohesin to centromeres' sides, which attaches chromatids to their bipolar spindles, and by attaching chromatids to microtubules, which provides their passage to the cell's opposite sides. The maintenance of genomic integrity, the authors conclude, likely relies on the coordination of these essential functions.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC490028.xml

539287

Soy versus whey protein bars: Effects on exercise training impact on lean body mass and antioxidant status

Background Although soy protein may have many health benefits derived from its associated antioxidants, many male exercisers avoid soy protein. This is due partly to a popular, but untested notion that in males, soy is inferior to whey in promoting muscle weight gain. This study provided a direct comparison between a soy product and a whey product. Methods Lean body mass gain was examined in males from a university weight training class given daily servings of micronutrient-fortified protein bars containing soy or whey protein (33 g protein/day, 9 weeks, n = 9 for each protein treatment group). Training used workouts with fairly low repetition numbers per set. A control group from the class (N = 9) did the training, but did not consume either type protein bar. Results Both the soy and whey treatment groups showed a gain in lean body mass, but the training-only group did not. The whey and training only groups, but not the soy group, showed a potentially deleterious post-training effect on two antioxidant-related related parameters. Conclusions Soy and whey protein bar products both promoted exercise training-induced lean body mass gain, but the soy had the added benefit of preserving two aspects of antioxidant function.

Background Many male exercisers avoid soy protein because there is a perception that it is inferior to proteins like whey for supporting lean boss mass gain. This perception persists even though there are no studies comparing whey and soy for effects on lean body mass gain. Soy may actually help promote lean body mass gain by the antioxidants associated with soy protein. Antioxidants are agents, either consumed in the diet or made by the body, which work against molecular damage due to oxidant reactions caused by free radicals, which are reactive molecules with an unpaired electron [ 1 ]. Soy protein isolate contains a mixture of antioxidants including isoflavones, saponins, and copper, a component of a number of antioxidant enzymes [ 2 ]. Body free radical production seems to be particularly high during exercise, and the resulting oxidant stress appears to contribute to muscle damage and fatigue [ 3 ]. This damage and fatigue could conceivably limit progress in exercise training by slowing muscle recovery between exercise workouts. This could limit lean body mass gain during an exercise program. If soy protein can promote lean body mass gain at least as well as whey, there may be one advantage to consuming soy protein. Soy protein contains antioxidants which may not only help with lean body mass gain, but which can also promote other aspects of health. Antioxidant actions are thought to work against the onset and severity of many diseases and health problems [ 1 ]. This may be particularly important during exercise training, which in some cases, depletes antioxidant capacities and/or increases oxidant stress [i.e. [ 4 , 5 ]]. This may explain why high degrees of chronic exercise can be detrimental. For example, some athletes show increases in histochemical muscle lesions as well as high cancer mortality, which have been linked to prolonged periods of exercise [ 6 , 7 ]. However, this area has been controversial since some studies suggest that long term exercise training produce body adaptations which increase antioxidant defenses [i.e. [ 8 , 9 ]]. Either way, soy protein antioxidants could conceivably exert beneficial effects during exercise training, either by restricting antioxidant depletion or by enhancing antioxidant capacity increases. The present study compared a soy protein product to a whey protein product in subjects undergoing a 9 week weight training program. Subjects were evaluated for lean body mass gain and changes in antioxidant status. The latter was done using one measurement of a component of antioxidant capacity and one for a component of oxidant stress. The former was based on an assay called plasma antioxidant status which assesses the ability to scavenge certain chemically generated radicals. The oxidant stress parameter was plasma myeloperoxidase, a measure of neutrophil activation, which is associated with increased secretion of superoxide radical [ 1 ]. Methods Subjects This study was approved by the Human Subjects Review Committee for Biomedical Sciences at The Ohio State University. All subjects signed an informed consent form. Male subjects, aged 19–25, were recruited from the Sport, Fitness and Health Program courses at The Ohio State University to participate in the present 9-week study. All subjects were considered experienced weightlifters with at least 1 year or more experience in strength training, which was confirmed by a questionnaire. Subjects were reported to be non-smokers, non-vegetarians, not currently taking supplements of any kind, and having no major health problems (i.e., diabetes, cardiovascular disease, etc.). All subjects had a body mass index (BMI) of less than 30. Strength Training Program At the start of the study, each subject was put on a common strength training program to strictly follow for the duration of the 9 week study. Subjects were given either workout 1 or workout 2. The two workouts were identical with the exception of exercise order and were designed to prevent subjects in the strength training classes from having to perform the same exercises at the same time. Midway through the program, subjects with workout 1 were given workout 2 and vice versa in order to maintain consistency. The strength training protocol was 3 sets of 4–6 repetitions for 14 exercises so that strength was the variable being maximized. The following exercises were performed to work all major muscle groups: 1) chest press; 2) chest fly; 3) incline press; 4) lat pull-down; 5) seated row; 6) military press; 7) lateral raise; 8) preacher curl; 9) bicep curl; 10) supine tricep extension; 11) seated tricep extension; 12) leg press; 13) calf raise; and 14) abdominal crunches. Protein Treatments Subjects were randomly assigned in a double-blind manner to either a soy, whey, or control group. The controls did the exercise program but did not consume a protein product (n = 9/each group). The soy protein product was DrSoy ® Bars, which contained 11 grams of protein and an assortment of micronutrients. The whey bars were made using the same recipe as the DrSoy ® Bars except that whey protein was substituted for soy protein. The products were supplied to study personnel in plain wrappers with different colors for each product. The color code was unknown to the subjects and study personnel who were in contact with the subjects. Each subject was instructed to consume 3 bars per day for the 9-week training period. This was in addition to the subjects' self-selected diet. Subjects were instructed not to change eating patterns during the course of the study. The time of the day when the bars were consumed was recorded daily in the subject's fitness log so that compliance could be monitored. Measurements Lean body mass was analyzed by hydrostatic weighing. Each subject performed at least 3 efforts and an average reading was taken. Blood was drawn into heparin tubes before and after the 9 week treatment period on a day when the subjects did not exercise. Blood was spun at 3000 × g and the plasma was stored at -70°C until analysis. Unfortunately, a problem during blood processing made some plasma samples unavailable for analysis. Plasma was analyzed for free radical scavenging capacity using the Total Antioxidant Status Assay Kit from Calbiochem-Novachem Corp. (San Diego, CA). Plasma myeloperoxidase was analyzed using an ELISA kit from Calbiochem-Novachem. Statistical analysis Statistical analysis was done by the Jump 3.1 program (SAS Institute, Cary, NC), with significance at p < 0.05. For each parameter and treatment group, values prior to the 9 week treatment were compared to values after treatment by paired, 2-tailed Student's t-test. In addition, for lean body mass, the changes in values for soy treatment were compared to the change in values for the other two groups by Tukey test. Results Baseline subject characteristics are given in Table 1 . Exercise training plus soy or whey treatments each produced a statistically significant increase in lean body mass, but the training alone did not (Figure 1 ). A comparison of the change in lean body mass for the soy group versus the change in the whey group did not show a significant difference (Figure 2 ). Plasma radical scavenging capacities fell in the whey and training alone groups, while the myeloperoxidase values rose in those same two groups (Figures 3 and 4 ). The values were unchanged in the soy group (Figures 3 and 4 ). Table 1 Subject characteristics. WHEY SOY CONTROL (Training Alone) AGE 20.36 ± 0.34 21.67 ± 0.24 20.44 ± 0.63 HEIGHT (cm) 180 ± 1.55 179 ± 1.30 178 ± 1.81 WEIGHT (kg) 81 ± 2.81 79 ± 2.49 79 ± 0.48 LBM (kg) 67 ± 1.96 66 ± 2.30 67 ± 1.65 Values are means ± SEM. Figure 1 Lean body mass pre- and post-treatment . Values are % lean body mass (kg) ± SEM from 9 subjects per group. *Significantly different from pre-treatment values (paired t-test, p < 0.05) Figure 2 Percent change lean body mass . Values are % change in lean body mass ± SEM. *Different letters indicate significantly differences between groups (Tukey test, p < 0.05) Figure 3 Plasma antioxidant status . Values are mM of trolox equivalents ± SEM (N = 5 for control and whey, 8 for soy) *Significantly different from pre-treatment values (paired t-test, p < 0.05) Figure 4 Plasma myeloperoxidase . Values are mg/L ± SEM (N = 5 for control and whey, 8 for soy) *Significantly different from pre-treatment values (paired t-test, p < 0.05) **Significantly different from pre-treatment values (paired t-test, p < 0.01) Discussion In this study, soy and whey were both effective at increasing lean body mass with exercise training, but the soy had the added advantage of inhibiting two negative effects of training on antioxidant status. The percent change in the radical scavenging capacity (total antioxidant status) seen with training alone and training plus whey was substantial compared to the differences typically seen for these types of measurements[ 11 - 13 ]. The lean body mass data seen here contradicts the common, but unconfirmed notion that soy is inferior to whey for promoting lean body mass gain. It should be noted, however, that the general trend for this study may or may not be duplicated for other study designs. For example, the time frame used here, 9 weeks, is not overly long for seeing lean body mass gain, which may explain why the training alone did not produce an effect on lean body mass gain. Thus, the effects of soy or whey on lean body mass gain versus training alone may be more pronounced than in longer studies. It should also be noted that the training program used here emphasized low exercise repetitions in subjects not used to this type of training. In addition, this study included only subjects that were still relatively early in their training experience, and placed no restriction on Calorie intake. These design considerations were geared toward gaining bulk and power. The effects of whey or soy on lean body mass might be different in a design that emphasizes higher repetitions or Calorie restriction in other types of subjects. In addition, it can be noted that the current study diet intervention used bars which included added micronutrients. Thus, this study did not determine if the effects of the soy or whey protein required co-administration of micronutrients. It is not known whether the negative effects of training seen here for antioxidant status in the whey plus training alone groups would continue upon longer training. The current state of knowledge concerning exercise training effects on antioxidant defenses does not present a clear pattern [i.e. [ 4 , 5 , 8 , 9 ]], possibly because of the highly variable circumstances involved in different studies such as training intensity, types of exercise done, types of antioxidant measures used, fitness level of the subjects, length of training, and dietary patterns of the subjects. These variables may help explain why some studies find training-induced declines in antioxidant defense while others find no change or even an increase. Nonetheless, the present study suggests that soy protein intake can promote antioxidant function during training which could be helpful no matter what the effects of training by itself. Another unresolved issue is whether the effects on lean body mass seen here for the two proteins were due to increased total protein intake or other factors. In regard to the former, the data regarding the amount and type of protein intake necessary to produce optimal strength training gains is conflicting. While a diet meeting the current RDA for protein intake (0.8 g/kg body mass) may be sufficient for the sedentary individual, recent studies suggest dietary protein exceeding that of the RDA is needed for muscle hypertrophy [ 14 , 15 ]. One of the difficulties in deriving an exact protein recommendation for exercisers is that total energy intake has not been consistent in the studies. In some studies, total energy intake was low, which can cause an abnormally high percentage of energy output to be derived from protein [ 15 , 16 ]. In the present study, a 3 day diet record gave no indication that Calorie intake was low (data not shown). If soy and whey promotion of lean body mass gain was not due to increased total protein intake, which remains uncertain, then other factors were responsible. In the case of soy protein, there are associated antioxidants [ 2 ]. As presented in the Introduction, this could conceivably help indirectly with lean body mass gain. In the case of whey, the content of essential amino acids, especially those with sulfur, may be conducive to promoting lean body mass gain [i.e. [ 17 , 18 ]]. In summary, soy and whey protein bars both supported lean body mass gain in conjunction with a short term power-based weight training program, but only the soy bar prevented a training-induced drop in antioxidant capacities. Competing interests Author AB owns the company that produces the soy bars used in the study. Authors' contributions ECB planned and carried out specifics of the intervention. RAD conceived the general aims of the study and chose the blood measurements. AB invented the protein bars and planned specifics of the nutrition intervention. STD planned the general aspects of the exercise intervention.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539287.xml

539244

The role of transforming growth factor-beta (TGF-beta) during ovarian follicular development in sheep

Background Recently, several members of the transforming growth factor-beta (TGF-beta) superfamily have been shown to be essential for regulating the growth and differentiation of ovarian follicles and thus fertility. Methods Ovaries of neonatal and adult sheep were examined for expression of the TGF-betas 1–3 and their receptors (RI and RII) by in situ hybridization using ovine cDNAs. The effects of TGF-beta 1 and 2 on proliferation and differentiation of ovine granulosa cells in vitro were also studied. Results The expression patterns of TGF-beta 1 and 2 were similar in that both mRNAs were first observed in thecal cells of type 3 (small pre-antral) follicles. Expression of both mRNAs continued to be observed in the theca of larger follicles and was also present in cells within the stroma and associated with the vascular system of the ovary. There was no evidence for expression in granulosa cells or oocytes. Expression of TGF-beta 3 mRNA was limited to cells associated with the vascular system within the ovary. TGFbetaRI mRNA was observed in oocytes from the type 1 (primordial) to type 5 (antral) stages of follicular growth and granulosa and thecal cells expressed this mRNA at the type 3 (small pre-antral) and subsequent stages of development. The TGFbetaRI signal was also observed in the ovarian stroma and vascular cells. In ovarian follicles, mRNA encoding TGFbetaRII was restricted to thecal cells of type 3 (small pre-antral) and larger follicles. In addition, expression was also observed in some cells of the surface epithelium and in some stromal cells. In granulosa cells cultured for 6 days, both TGF-beta 1 and 2 decreased, in a dose dependent manner, both the amount of DNA and concentration of progesterone. Conclusion In summary, mRNA encoding both TGF-beta 1 and 2 were synthesized by ovarian theca, stroma and cells of the vascular system whereas TGF-beta 3 mRNA was synthesized by vascular cells. Luteinizing granulosa cells also responded to both TGF-beta 1 and beta 2 in vitro. These findings in sheep are consistent with TGF-beta potentially being an important autocrine regulator of thecal cell function and possibly a paracrine regulator of ovarian cell function at various development stages.

Background Members of the transforming growth factor-beta (TGF-β) superfamily are important intraovarian growth factors [ 1 - 6 ]. Three key members of the TGF-β subfamily, namely TGF-β1, TGF-β2 and TGF-β3, have been shown to be produced by ovarian cells [ 7 - 13 ]. However, the cellular distribution of these proteins varies between species. Likewise, the effects of TGF-βs on granulosa cell function also vary between species. In rodents, TGF-βs are potent stimulators of granulosa cell proliferation [ 14 - 16 ] whereas in other species, such as cattle and pigs, these growth factors have only a mild stimulatory or even inhibitory effect [ 17 - 20 ]. Likewise, TGF-βs stimulate progesterone synthesis from rodent granulosa cells [ 21 - 23 ] where inhibitory effects are observed in granulosa cells collected from sheep, cattle and pigs [ 17 , 24 - 26 ]. Exceptional ovulation rates and sterility have been observed in lines of sheep with mutations in two members of the TGF-β superfamily, namely growth differentiation factor 9 or bone morphogenetic protein 15 or one of their receptors, activin like kinase-6 [ 6 , 27 ]. However, little is known about the roles of other members of the TGF-β superfamily in sheep and thus the potential interactions of members of the TGF-β superfamily are unclear. The objectives of this study in sheep were to localize the ovarian cellular types expressing mRNA encoding TGF-β1, TGF-β2, TGF-β3, TGFβRI and TGFβRII and to determine the effects of TGF-β1 and TGF-β2 on granulosa cell proliferation/survival and progesterone production in vitro . Methods Generation of cDNAs encoding a portion of the coding region of genes of interest Except where indicated, laboratory chemicals were obtained from BDH Chemicals New Zealand Ltd (Palmerston North, New Zealand), Invitrogen (Auckland, New Zealand) or Roche Diagnostics N.Z. Ltd. (Auckland, New Zealand). Total cellular RNA was isolated from ovine ovary using TRIzol according to manufacturer's instructions. First strand cDNA was produced from total cellular RNA using a poly t primer. Complementary DNAs encoding a portion of the coding sequence of the genes were isolated using standard PCR techniques. For individual cDNAs generated, primer sequences and annealing temperature are given in table 1 . Resulting PCR products were ligated into appropriate vectors and their nucleotide sequence determined by automated sequence analysis (Waikato DNA Sequencing Facility; The University of Waikato; Hamilton, New Zealand). These sequences were compared with known sequences to confirm identity using the GAP program of GCG (Wisconsin Package Version 10.2, Genetics Computer Group; Madison, Wisconsin). All sequences were >80% identical to those listed as reference sequences in table 1 indicating that the ovine homologue of the respective genes had been obtained. Table 1 GenBank reference numbers used for primer design, primer sequences, annealing temperatures, and GenBank accession numbers for the resulting ovine sequence for the various genes amplified. Gene Reference: (Genbank #) Forward Primer (5' to 3') Reverse Primer (5' to 3') Annealing temperature Genbank # (resulting sequence) TGF-β1 NM_011577 ggaattcatgccgccctcggggctgcgg (EcoR I site and bases 867–888) ggtctagatcagctgcacttgcaggagcg (Xba I site and bases 2040–2020) 63°C ND TGF-β2 M19154 ggaattcatgcactactgtgtgctgagc (EcoR I site and bases 468–488) ggtctagagctgcatttrcaagacttkac (Xba I site and bases 1794–1773) 64°C AY656797 TGF-β3 J03241 ggaattcgcaaagggctctggtggtcctgg (EcoR I site and bases 277–299) ggtctagaccagttctcctccaagttgcgg (Xba I site and bases 1206–1186) 62°C AY656798 TGFβRI U97485 cacagatgggctttgctttg (bases 180–199) ccttgggtaccaactatctc (bases 1007–988) 50°C AY656799 TGFβRII S69114 gtcctgtggacgcgcat (bases 80–97) aggagcacatgaagaaagtc (bases 449–430)* 50°C AY656800 TGFβRII (for PCR) various gccaacaacatcaaccac gggtcrtggtcccagca 53°C AY751461 TGFβRII (internal for PCR) AY751461 tcgccgaggtctacaagg atgccctggtggttgagc 55°C N/A * Sequence is based on the corresponding ovine sequence obtained from an ovine est clone. ND, the complete sequence of the clone was not determined, as the ovine TGF-β1 sequence is known. The clone was sequenced from both ends and resulting sequence compared to the known ovine TGF-β1 sequence to confirm identity of the isolated cDNA. N/A as the sequence overlaps that of AY751461. Collection of tissue samples All experiments were performed in accordance with the 1999 Animal Welfare Act Regulations of New Zealand. All animals had ad lib access to pasture and water and lambs were kept with their mothers until just prior to tissue collection. Romney ewes and lambs were killed by administration of a barbiturate overdose (Pentobarbitone; 200 mg/kg, Southern Veterinary Supplies, Christchurch, New Zealand). Recovered ovaries were fixed in 4% (w/v) phosphate-buffered paraformaldehyde and embedded in paraffin wax. In Situ Hybridization Cellular localization of mRNAs was determined using the in situ hybridization protocol described previously with minor modifications [ 28 ]. Sense and anti-sense RNA probes were generated from cDNA encoding the gene of interest with T7, T3 or SP6 RNA polymerase using the Riboprobe combination system (Promega, Dade Behring Diagnostics Ltd., Auckland, New Zealand). For all in situ hybridizations, 4–6 μm tissue sections were incubated overnight at 55°C with 45,000 cpm/μl (approximately 48,000 dpm/μl) of 33 P-labelled antisense RNA. Non-specific hybridization of RNA was removed by RNase A digestion followed by stringent washes (2 × SSC, 50% formamide, 65°C and 0.2 × SSC at 37°C). Following washing, sections were dehydrated, air dried and coated with autoradiographic emulsion (LM-1 emulsion; Amersham Pharmacia Biotech, New Zealand). Emulsion-coated slides were exposed at 4°C for 3 weeks, developed for 3 1/2 minutes in D19 developer (Eastman Kodak, Rochester, NY), development was stopped using a 1 minute incubation in 1% acetic acid and slides were fixed with a 10 minute incubation in Ilfofix II (Ilford Limited, Cheshire, England). Sections were stained with hematoxylin and then viewed and photographed using both light and dark field illumination on an Olympus BX-50 microscope (Olympus New Zealand Ltd., Lower Hutt, New Zealand). Non-specific hybridization was monitored by hybridizing at least two tissue sections from each age group (lamb and adult) with approximately equal concentrations of the sense RNA for each gene. Hybridization was considered to be specific when the intensity of silver grains, as measured by visual assessment, over a cellular type was greater than that observed in the area of the slide not containing tissue. For all genes, hybridization of the sense RNA over the tissue section was similar or lower in intensity to that observed on the areas of the slide not containing tissue of both the sense and antisense hybridized slides and thus was considered non-specific. Follicular classification Classification of follicles was based on the system outlined by Lundy et al. [ 29 ]. Briefly, type 1/1a follicles consist of an oocyte surrounded by a single layer of flattened or mixed flattened and cuboidal cells. Type 2 follicles contain 1 < 2 layers of cuboidal granulosa cells whereas type 3 follicles contain 2 < 4 layers of cuboidal granulosa cells. Type 4 follicles have >4 layers of granulosa cells and a well defined theca but have not yet formed an antrum. Type 5 follicles have multiple layers of granulosa cells, a well defined theca and a defined antrum. All follicles with signs of degeneration (i.e. pyknotic granulosa cells, lack of a distinct basement membrane or degenerate oocytes) were excluded from the study. Ovarian sections from a minimum of eight animals, including at least three lambs and three adults, were examined for each gene studied. In addition, each follicle class was observed in a minimum of three animals. No differences in expression patterns between lamb and adults ovaries were noted in this study. Granulosa cell culture Ovaries were collected from ewes following slaughter at the local abattoir, transported back to the laboratory at room temperature, washed in 3% bleach solution in PBS for 5 minutes, rinsed twice in PBS and stored in Leibovitz media containing 0.1% BSA, 100 U/ml penicillin and 100 μg/ml streptomycin. Follicles approximately 1–2 mm in diameter were dissected away from the ovaries and stored in Leibovitz media until collection of granulosa cells. The granulosa cells were collected by cutting follicles in half followed by manual scraping of cells from the follicular wall using a wire loop. Oocytes and follicular debris were removed from the cells using a micro-glass pipette. Remaining cells were collected by centrifugation at 300 g for 5 min at room temperature, washed once in 5 mls Leibovitz media, twice in 5 mls McCoys media (Sigma, Auckland, New Zealand) with 100 U/ml penicillin, 100 μg/ml streptomycin and 2 mM GlutaMAX-1 and resuspended using a syringe and needle. Cell viability was determined using trypan blue exclusion and 100,000 viable cells per well (250 μl total volume) were added in McCoys media containing 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM GlutaMAX-1, 5 ng/ml selenium (Sigma), 10 ng/ml insulin (Sigma), 5 μg/ml apo-transferrin (Sigma), 30 ng/ml androstenedione (Sigma), 3 ng/ml ovine FSH (purified in our laboratory; 1.4 X USDA-oFSH-19-SIAFP RP2), 1 ng/ml IGF-1 (Long-R3, GroPep, Adelaide, SA 5000, Australia) with varying doses (0–10 ng/ml) of purified human TGF-β1 and recombinant human TGF-β2 (R & D Systems, Minneapolis, MN). Cells were cultured at 37°C in a 5% CO2 incubator. Every 48 hours, 200 μl of media was removed from each well and replaced with 200 μl of warmed media that had been prepared at the start of the culture and stored at 4°C. Media samples from the last 48 hours of treatment were collected on day 6 of treatment and frozen at -20°C for later determination of progesterone concentrations by RIA. Unattached cells were removed by 2 washes with McCoys media at 37°C. Attached cells were lysed by incubating cells at 37°C in 100 μl distilled water for 1–2 hours followed by freezing at -70°C. All treatments were performed at least in triplicate with three independent pools of granulosa cells. Within an assay, individual values outside of 20% of the mean value for the treatment were discarded. Points in which at least 2 of the replicates were not within 20% of each other were regarded as missing data. This occurred for the 10 ng TGF-β1 measure of DNA in a single pool of granulosa cells. Measurement of DNA The amount of DNA present in each well was determined by comparing binding of Hoechst 33258 dye (Sigma, final concentration in well of 10 μg/ml) in samples to calf thymus DNA standard measured with a Wallac 1420 plate reader at 350 nm for excitation and 460 nm for emission. Sensitivity of the assay (+ two SD of control buffer value) was 33 ng per well and the intra- and inter-assay co-efficients of variation (CV), based on variability of the 100, 250, 1000 and 2500 ng standard curve points were 3.9% and 8.8%, respectively. No samples were below the sensitivity of the assay. Measurement of Progesterone Concentrations of progesterone in media were determined by RIA as described [ 30 ]. The sensitivity of the assay (90% maximum binding) was 13 pg/ml and the intra- and inter-assay CV, averaged for a standard pool sample at approximately 20%, 50% and 80% binding, was 8.3% and 19.7%, respectively. No samples were below the sensitivity of the assay. Determination of expression of TGFβRII mRNA in cultured granulosa cells Granulosa cells were collected as described above and either frozen immediately after collection or plated in 6 well culture dishes at a density of 1.0–1.5 × 10 6 viable cells per well in 2 mls of control (i.e. no TGF-β) culture media described above for 48 hours. At this time, unattached cells were removed by washing the wells twice with PBS. RNA was collected using TRIzol according to the manufacturer's instructions. First strand cDNA was produced from total cellular RNA using the SuperScript™ preamplification system for first strand cDNA synthesis. An initial PCR was performed with 4 week old ovary RNA to obtain the ovine sequence of a region of the TGFβRII gene which spans introns 4 and 5 in the human sequence (AY675319) and a second set of primers was designed based on the ovine sequence (see table 1 ). Expression of TGFβRII was determined by PCR using the Qiagen Taq PCR core Kit (Biolab Scientific Limited) and the internal ovine primers listed in table 1 with the following conditions: initial denaturing cycle of 3 minutes at 94°C followed by 40 cycles of denaturing at 94°C for 1 minute, annealing at 55°C for 1 minute and extension at 72°C for 2 minutes and a final extension at 72°C for 10 minutes. cDNA generated from a 4 week old lamb ovary was run as a positive control whereas replacement of cDNA with water was used as a negative control. Expression of TGFβRII was assessed by visualization of DNA bands of the correct size following gel electrophoresis. Identity of product was confirmed by sequencing. Statistical analysis Concentration of progesterone per μg DNA was calculated for individual wells before averaging for each treatment within each assay. Points in which at least 2 of the replicates were not within 30% of each other were regarded as missing data. Changes in the concentrations of progesterone in media and DNA content after culture were analysed with the general linear model procedure of SAS. Replicate was included in the model as baseline progesterone and DNA values varied among the granulosa cell pools. Differences between least square means were evaluated using least significant differences and were considered significant when p < 0.05. Data presented are least square means. The standard errors of least square means were 0.7 ng/well, 0.2 μg/well and 0.5 ng/μg for progesterone, DNA and p4 per DNA, respectively. Results In situ hybridization TGF-β1 The mRNA for TGF-β1 was not observed in granulosa cells or oocytes of any follicles (Figure 1a,1b , table 2 ). However, TGF-β1 mRNA was observed in stromal and/or thecal cells of type 3 follicles, in the theca interna of type 4 and 5 follicles and also in the stroma and cells of the vascular system. Within the theca interna, the cells closest to the basement membrane usually had more intense signal than those further away (Figure 1a,1b ). Figure 1 Localization of expression of mRNA encoding TGF-βs in ovine ovaries. Panels a and b contain corresponding light field and dark field views of a type 5 follicle from an adult ewe hybridized to TGF-β1 antisense RNA. Silver grains indicating hybridization of the TGF-β1 antisense RNA are observed concentrated in thecal (t) cells close to the basement membrane with no specific hybridization observed in either the granulosa cells (gc) or oocyte (o). The inset in panel b contains a dark field view of the same area of the tissue hybridized to the TGF-β1 sense RNA. Note the lack of specific concentration of silver grains over any cellular type. Panels c and d contain several type 5 (5) follicles and a type 4 (4) follicle in a 4 week old lamb hybridized to TGF-β2 antisense RNA. Note the lack of hybridization in the oocytes and granulosa cells of the type 4 and 5 follicles and the concentration of silver grains in thecal cells around the follicles as well as the stromal cells between the follicles and scattered cells of the surface epithelium (se). Observe the equal distribution of silver grains over the thecal cells. The inset in panel d contains a dark field view of the same area of the tissue hybridized to the TGF-β2 sense RNA. Note the lack of specific concentration of silver grains over any cellular type. Panels e and f contain corresponding light field and dark field views of an ovarian section obtained from an adult ewe hybridized to TGF-β3 antisense RNA. There is a lack of hybridization in the section including the granulosa and thecal cells of the types 4 (4) and 5 follicle (5) as well as the oocyte of the type 4 follicle, stroma tissue and the surface epithelium (SE). Panels g and h contain light field and dark field views of a blood vessel (v) from an adult ewe hybridized to TGF-β3 antisense RNA. Observe the specific hybridization in the wall of the vessel (v ). Panel i contains a dark field view of the same area of the tissue hybridized to TGF-β3 sense RNA. Note the lack of specific concentration of silver grains over any cellular type. Scale bar equals approximately 100 μm for all panels. Table 2 Summary of expression of mRNA encoding TGF-βs and receptors in ovine ovary. Gene Follicular type Stroma Vascular System 1/1a 2 3 4 5 TGFβ1 - - t t t + + TGFβ2 - - t t t + + TGFβ3 - - - - - - + TGFβRI o o o, gc, t o, gc, t o, gc, t + + TGFβRII - - t t t + + +, expression observed; -, expression not observed; o, oocyte; gc, granulosa; t, theca TGF-β2 The pattern of expression of mRNA encoding TGF-β2 was similar to that observed for TGF-β1, with hybridization limited to the thecal cells of type 3 and larger follicles (Figure 1c,1d , table 2 ). However, hybridization within the thecal layer appeared evenly distributed in contrast to the signal for TGF-β1 (compare panels a, b and c, d in Figure 1 ). Expression of TGF-β2 mRNA was also observed in some surface epithelium and stromal cells as well as cells associated with the vascular system. TGF-β3 Expression of TGF-β3 mRNA was exclusive to cells associated with the vascular system of the ovary. Expression was not observed in the granulosa, theca, or oocyte of any follicle examined (Figure 1e,1f,1g,1h,1i , table 2 ). TGFβRI Expression of TGFβRI mRNA was observed in oocytes of all types of follicles (Figure 2a,2b,2c,2d , table 2 ). Granulosa and thecal cells of type 3 and larger follicles also expressed TGFβRI mRNA (Figure 2c,2d ). Signal was also observed in the surface epithelium, stromal cells (Figure 2a,2b,2c,2d and luteal tissue (data not shown). Figure 2 Localization of expression of mRNA encoding TGF-β receptors in ovine ovaries. Panels a and b contain corresponding light field and dark field views of several small follicles from a 4 week old lamb following hybridization to the TGFβRI antisense RNA. Note specific hybridization in the oocytes of types 1/1a follicles (1) and type 2 follicles. Observe that some cells of the surface epithelium also express TGFβRI. Panels c and d contain corresponding light field and dark field views of a type 5 follicle from a 4 week old lamb following hybridization to the TGFβRI antisense RNA. Note the hybridization signal in the granulosa (gc), theca (t) and oocyte (o) of the type 5 follicle. Signal was also observed in many stromal cells. The inset in panel d contains a dark field view of the same area of the tissue hybridized to TGFβRI sense RNA. Observe the lack of specific concentration of silver grains over any cellular type. Panels e and f contain corresponding light field and dark field views of several small follicles from a 4 week old lamb following hybridization to the TGFβRII antisense RNA. Note the lack of specific hybridization in the type 1/1a and 2 follicles. Expression was observed in the theca of type 4 and 5 follicles however, note the lack of expression in the granulosa cells and oocytes of these follicles. Note also that some cells of the surface epithelium also express TGFβRII. The insert in panel f contains a dark field view of the same area of the tissue hybridized to TGFβRII sense RNA. Note the lack of specific concentration of silver grains over any cellular type. Panels g and h contain corresponding light field and dark field views of a type 5 follicle as well as several type 1/1a follicles from a 4 week old lamb ovary hybridized to the TGFβRII antisense RNA. Note that hybridization is limited to the theca (t) of the type 5 follicle and several stromal cells and is not observed in the granulosa cells (gc) or oocyte (o) of the type 5 follicle. In addition, signal is observed in the stroma around the type 1/1a follicles (1) but is not observed in the type 1/1a follicles. Scale bar equals approximately 100 μm for all panels. TGFβRII Expression of TGFβRII mRNA was not observed in types 1,1a or 2 follicles (Figure 2e,2f,2g,2h , table 2 ). Also, in larger follicles TGFβRII mRNA was not detected in granulosa cells or oocytes (Figure 2e,2f,2g,2h ). In type 3 and larger follicles, expression of TGFβRII was localized to the theca interna (Figure 2e,2f,2g,2h , table 2 ). As was observed with TGF-β1, expression of TGFβRII within the theca was most intense in the cells adjacent to the basement membrane (Figure 2e,2f,2g,2h ). Signal was also observed in some cells of the surface epithelium (Figure 2e,2f ), and in stroma (Figure 2e,2f,2g,2h ) and luteal tissue (data not shown). Effects of TGF-βs on granulosa cell function in vitro and expression of TGFBRII in cultured cells Both TGF-β1 and TGF-β2 inhibited progesterone synthesis of cultured granulosa cells, whether expressed as a function of number of cells placed in culture (Figure 3 , top panel) or as a function of DNA content at the end of culture (Figure 3 , bottom panel) with significant affects observed with as little as 0.1 ng/ml of either TGF-β. Treatment with either TGF-β also reduced DNA content at the termination of culture (Figure 3 , middle panel). For both variables, no differences were observed between the effect of TGF-β1 and TGF-β2 at any dose of growth factor tested. In contrast to the lack of detectable expression of the TGFβRII mRNA observed in situ , freshly isolated or cultured granulosa cells expressed mRNA for the TGFβRII when assessed by RT-PCR (Figure 4 ). Figure 3 Effects of TGFβs on granulosa cell function. Effects of TGF-β1 and TGF-β2 on secretion of progesterone during the last 48 hours of culture (top), content of DNA at the termination of culture (middle) and progesterone concentration per μg of DNA. Values are expressed as the LS mean from 3 separate experiments. The dose of either TGF-β1 or TGF-β2 is indicated along the bottom of the graphs. For each variable, asterisk(s) indicates values that are different from the control (0) value (* p < 0.05; ** p < 0.01, *** p < 0.001). Comparisons were also made between the values obtained for TGF-β1 and TGF-β2 at each dose; however, no significant differences were observed at any dose tested. Figure 4 Expression of TGFβRII in cultured granulosa cells. Determination of expression of TGFβRII in granulosa cells immediately following collection and following 48 hours of culture. Lanes 1–3 contain PCR products (766 bases) following amplification with ovine TGFβRII primers from 3 separate pools of granulosa cells at the time of collection, lanes 4–6 contain PCR products (766 bases) following amplification with ovine TGFβRII primers from 3 separate pools of granulosa cells collected 48 hours after culture, lane 7 contains the negative control water blank whereas lane 8 contains the PCR product from the positive control 4 week old ovary sample. Migration of DNA molecular weight standards are indicated on the left hand side. Discussion In the ewe, expression of TGF-β1 and TGF-β2 mRNA in the follicle was limited to thecal cells during all stages of follicular growth examined. Furthermore, expression of TGF-β3 mRNA was not observed in any follicular cells. This is in contrast to the observed expression patterns for these proteins in other species where TGF-β1 and TGF-β2 have been localized to granulosa as well as thecal cells and sometimes also to the oocytes of many species [ 8 , 12 , 13 , 31 - 33 ]. Also in contrast to sheep, expression of TGF-β3 in cattle and cats was observed in the oocyte, theca and granulosa of follicles at various stages of development [ 12 , 13 ]. In pigs, the theca interna has been proposed to be the major source of TGF-β since granulosa cells express TGF-β1 mRNA without seeming to make the protein [ 11 ]. Similarly, expression of TGF-β2 mRNA has been observed in bovine oocytes, but no detectable TGF-β activity was observed [ 17 ], although other studies have demonstrated TGF-β protein in the oocytes using immunocytochemistry [ 12 ]. In addition, granulosa cells isolated from pigs and cattle produce little if any TGFβ bioactivity when cultured in vitro [ 9 , 11 ]. Thus, it seems likely that in some species, follicular TGF-β activity originates primarily from the thecal cells, with control of activity possibly occurring at several levels including gene transcription (this study), protein translation [ 11 ] or activation of the protein [ 17 ]. Similar to what has been observed in other species [ 32 , 34 , 35 ], expression of TGFβRI mRNA was observed in several different cell-types of the sheep ovary including the oocyte, granulosa cells, thecal cells, stroma, luteal cells and surface epithelium. While expression of TGFβRII mRNA was also observed in stroma, luteal cells and the surface epithelium, its expression within the follicle was limited to the theca. A similar pattern of expression for the TGFβRII mRNA was observed in mouse follicles, with expression most prominent in the theca and barely detectable in granulosa cells [ 8 ]. However, using immunocytochemistry, strong staining for TGFβRII has been observed in granulosa cells with no to little staining in oocytes and in the theca in other species [ 13 , 32 , 35 - 37 ]. The reasons for these observed differences in localization of the TGFβRII are uncertain but may be due to differences in techniques or species differences. Expression of mRNA encoding all three TGF-β isoforms and the TGF-β type I and II receptors were observed in cells associated with blood vessels and both receptor types and TGF-β1 and 2 mRNAs were observed in the stroma surrounding follicles indicating a potential role for TGF-β in regulating certain functions in the ovarian stroma and vascular network. TGF-β is known to be important in regulating angiogenesis [ 38 , 39 ]. Moreover, in the ovary, both TGF-β1 and TGF-β3 mRNAs are upregulated during revascularization following autotransplantation of rat ovaries [ 40 ] further supporting a role for these factors in regulating vascular function. The much more restricted pattern of expression of TGFβRII mRNA in sheep indicates that the TGFβRI may well be involved with other type II receptors in the signalling of other members of the transforming growth factor family. In agreement with this, TGFβRI has recently been shown to be involved in signalling of the oocyte-derived GDF-9 along with BMPRII [ 41 , 42 ]. In other species, GDF-9 has been shown to regulate granulosa cell mitosis and differentiation [ 6 ] and has been shown to be essential for normal follicular growth and development in both mice [ 43 ] and sheep [ 44 , 45 ]. Thus, expression of TGFβRI mRNA as well as BMPRII [ 46 , 47 ] in granulosa cells is probably mediating the effects of GDF-9. Localization of both of these receptors in granulosa cells from the type 3 (secondary) stage of development onwards is consistent with the presence of normal primary but not secondary follicles in both sheep [ 44 ] and mice [ 43 ] lacking biologically active GDF-9. Interestingly, in sheep, TGFβRI mRNA and BMPRII [ 46 , 47 ] are also both localized in oocytes from the type 1 (primordial) stage onwards suggesting that GDF-9 may also regulate oocyte function in this species. The suppression of progesterone production and DNA content in granulosa cell cultures by TGF-β1 or TGF-β2 is similar to inhibitory to mild stimulatory effects observed in bovine, ovine and porcine granulosa cell cultures and contrary to the strong stimulatory effects observed in rodents [ 11 , 14 - 18 , 20 - 26 ]. The decreased DNA content observed following treatment accounts for some, but not all, of the decrease observed in progesterone concentration in the granulosa cell cultures. The suppression of progesterone synthesis indicates an anti-differentiative role for this growth factor as has been observed for other members of the TGF-β superfamily. The decreased content of DNA observed following culture could be related to a suppression of granulosa cell proliferation or survival. Since TGF-β can stimulate apoptotic pathways in concert with other factors [ 48 , 49 ], a role for TGF-β in regulating apoptosis of ovarian cells has been proposed. No differences in the efficacy of TGF-β1 and TGF-β2 were observed in ovine granulosa cells. Similarly, TGF-β1 and TGF-β2 were equally efficacious in stimulating inhibin production in luteinized human granulosa cells [ 50 ] and in modulating gonadotrophin receptor expression in both rat and porcine granulosa cells [ 51 ]. Interestingly, while both TGF-β1 and TGF-β2 mRNA were synthesized by the theca interna, their spatial patterning within the theca was quite different. TGF-β1 mRNA was concentrated in the thecal cells closest to the basement membrane, similar to the localization observed for the TGFβRII mRNA. In contrast, TGF-β2 mRNA expression was observed throughout the thecal layer. The role, if any, of the apparent differential regulation of these two isoforms in subtypes of thecal cells is currently unknown. Given the potent effects of both TGF-β1 and TGF-β2 on granulosa cell function in vitro , the lack of detectable expression of TGFβRII mRNA in these cells using in situ hybridization was very surprising. There are several potential explanations for these apparent conflicting results. It is possible that TGFβRII is expressed in ovine granulosa cells and the technique utilized simply failed to detect this message. The detection of mRNA encoding TGFβRII in isolated granulosa cells both before and after culture using RT-PCR would seem to support this assumption. However, it is possible that the isolation and culture of the granulosa cells potentially could be inducing expression of TGFβRII as most all cells in culture express TGFβRII [ 52 ]. In addition, strong expression of TGFβRII mRNA in luteal tissue is also consistent with up regulation of the TGFβRII in these cells as induction of progesterone synthesis by the ovine granulosa cells can be considered to indicate at least a partial luteinization of these cells. Finally, it is also possible that TGFβs are using another member of the type II receptor family to mediate their effects. The existence of a second type II receptor capable of mediating TGF-β effects is supported by the inability of cell lines expressing TGFβRII to bind to TGFβ2 but not TGFβ1 [ 53 ] and cell lines responsive to TGF-β without a detectible type II TGFβR [ 52 ]. Conclusions Expression of mRNAs encoding TGF-β1 and TGF-β2 as well as both type I and II TGF-β receptors were observed in the theca of small growing follicles indicating that TGF-βs may be regulating thecal cell function in an autocrine manner. Expression of mRNA encoding TGF-β type I and II receptors is also observed in luteal cells, stroma, the vascular system and surface epithelium suggesting that TGF-βs may also regulate other cell types in the sheep ovary. Since granulosa cells showed no evidence of expressing any of the TGF-β ligands and expression of the TGF-β type II receptor was equivocal, it seems likely that any TGF-β effects in granulosa cells in vivo are due to paracrine or endocrine actions and possibly regulated through an alternative type II receptor. Authors' contributions AHB, LDQ and LJH cloned the ovine TGF-βs and receptors, completed sequencing projects and alignments, and performed the in situ hybridizations and PCRs. SL and KLR performed the granulosa cell bioassays including progesterone and DNA measurements. JLJ and KPM designed and co-ordinated the experiments, performed statistical analysis and drafted the manuscript. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539244.xml

524214

The big problem of the missing cytology slides

Cytology slides are often unique and irreplaceable. Unlike surgical pathology cases, where additional paraffin sections can be cut, cytology slides often cannot be duplicated because there are only a few direct smears or the diagnostic material is present on a single slide. Cytology slides are often "sent out" to other physicians, laboratories or hospitals, typically so that the pathologist at the institution where the patient will receive treatment can review the slides. Less often, a cytology lab sends out the slides for a second opinion or as part of the discovery process in a lawsuit, where they may or may not be defendants. Rarely, unique and irreplaceable cytology slides are lost. This article presents a hypothetical scenario that is based on reported state appellate court decisions. The article discusses some of the legal issues that will affect the defendant cytologist/cytology lab and the "expert cytologist," and suggests some steps a cytologist/cytology lab can take to minimize the risk of repercussions from a lost unique and irreplaceable cytology slide.

1. What is already known on this topic? A WestLaw search (similar to PubMed, but searches state and federal cases and statutes as well as commentary) uncovered only a handful of reported appellate cases that directly applied to the issue of lost cytology slides. There are many cases and statutes dealing with lost and altered evidence (other than cytology slides), but the circumstances surrounding cytology slides are unique. I did not see any previous reviews or commentary specifically addressing the topic. Notwithstanding the relative paucity of on-point legal authority, the issue is one commonly addressed by cytology laboratories and cytologists and many have approached the issue thoughtfully and prudently. 2. What is not highlighted and what this review would answer? This review synthesizes the available legal cases and presents to practicing cytologists a short, relatively concise summary in the format of a hypothetical case. The review seeks to incorporate the legal issues with the practical aspects of running a cytology laboratory. The review answers how some state courts might approach the problem of lost and irreplaceable cytology slides, and offers general ideas to cytologist for minimizing the risk from lost slides. The big problem of the missing cytology slides The issues surrounding the decision whether to send out diagnostic patient slides are more important in cytology than in surgical pathology because the cytology slides are typically unique and irreplaceable. Unlike surgical pathology cases, where additional paraffin sections can be cut, cytology slides often cannot be duplicated because there are only a few direct smears or the diagnostic material is present on a single slide. Slides are routinely "sent out" for a variety of reasons. Most commonly, slides are sent to another institution because the patient's pathology slides will be reviewed before treatment. Some smaller laboratories with only one or two pathologist may send out slides as part of their quality assurance procedures. More rarely, but increasingly, slide's are requested as part of existing or contemplated litigation. Slides might be lost in any of these circumstances. A laboratory or hospital might loose the slides and not discover their absence until the slides are requested. This article will discuss the approach the US legal system takes in addressing what happens when cytology slides are lost, and what steps a prudent laboratory might take to manage the risk. The article also intends to improve cytopathologists' awareness and understanding of some of the legal issues that arise when evidence is missing and perhaps promote an international comparative discussion of alternative legal approaches. In the US, most of medical malpractice law is made and interpreted by state legislatures and state courts. What follows is a hypothetical story based on several legal decisions made by appellate state courts in the United States and reported in the legal literature. Although based on real cases and available to the public, the names have been changed. Importantly, none of this is intended as legal advice and readers should consult their attorney about specific questions. Facts Rugged Labs (RL) is a small, independent laboratory. Part of Rugged Labs' work involves providing Big Giant Lab (BGL) with overflow services for cytopathology. In December of 1995, BGL bought out RL. In 1994 and 1995, a RL cytotechnician interpreted two Pap smears from 30-year-old Ms. Penny as normal. In 1996 a cervical biopsy from Penny showed adenocarcinoma. Ms. Penny underwent a radical hysterectomy, which confirmed the invasive endocervical adenocarcinoma. Ms. Penny is alive today, but endured extended post-operative hospitalization. At the time of the hysterectomy, Penny Plaintiff's oncologist requested that the 1994 and 1995 PAP smear slides be sent to Dr. Experta, who interpreted both Pap smears as containing "abnormal cell groups consistent with adenocarcinoma." Dr. Experta also reviewed the biopsy and in a note concluded that the cells on the Pap smears were consistent with the adenocarcinoma diagnosed on the cervical biopsy. Approximately 6 months later, Plaintiff Penny decided to sue for medical malpractice based on failure to diagnose her endocervical adenocarcinoma on the Pap smears. At some time before the plaintiff filed her lawsuit Dr. Experta's assistant apparently mailed the slides back to BGL. The Pap smear slides are lost, presumably in the mail, by BGL or Dr. Experta's office. RL, the independent laboratory, asked the trial court to grant it summary judgment on the basis that the evidence was lost and no questions of fact remained. Summary judgment means that there are no outstanding questions of fact and the court needs to decide only questions of law. Summary judgment means there is no trial with a jury or judge hearing and weighing evidence. The question of law RL wanted the court to determine on summary judgment was that the lost slides substantially prejudiced RL and summary judgment was, therefore, appropriate. RL included in its motion for summary judgment an affidavit from an expert stating that she could not give an opinion without having the slides to look at. An affidavit by Dr. Experta's secretary stated that the slides were returned to BGL by US mail. BGL submitted an affidavit attesting they did not lose the slides. The parties to the lawsuit stipulated that the slides were lost. The trial court agreed with RL that there should be summary judgment in RL's favor, reasoning that no questions of fact needed to be answered and that a trial would unduly prejudice RL because of the spoliation of evidence. For a summary of the facts, please see Fig. 1. The Appellate Court's Opinion The Plaintiff appealed and the state appellate court reversed the trial court, concluding that the trial court made a mistake in not allowing the case to go to trial. The appeals court reasoned that the case turned on a question of fact; "The slide either showed the presence of cancer cells or it did not." The appellate court envisioned the trial as follows: the cytotechnologist from RL would "testify to her conclusions" and the plaintiff would have Dr. Experta testify to her conclusions. There should be a trial because there were questions of fact including, in the words of the majority, whether the slide had "cancer cells" or not. The majority also discounted the affidavit from BGL, stating that the conclusory statement that BGL had not lost the slides was not valid; just because BGL couldn't find the slides did not mean they never had them. In a footnote, the majority noted that although they aren't accusing anyone, they couldn't help but notice that missing the slides benefited BGL and RL. One judge dissented in the three-judge panel that decided the case. The dissent focused on two points. First, the dissent emphasized that the defendant RL had nothing to do with loosing the slides. Although the plaintiff, Ms. Penny, had no "direct role" in the loss of evidence, the dissent treated Dr. Experta as an agent of the plaintiff and concluded that Dr Experta should have sent the slides back to RL not GBL. The dissent also reasoned that its approach would "encourage experts to treat more carefully evidence delivered into their hands." Secondly, the dissent emphasized that the defendant RL is disadvantaged by the loss of the slide and the plaintiff has gained a significant advantage. The dissent sees a trial where the cytotechnologist's faces a serious credibility problem because her testimony will come across as blatantly self-serving. The cytotech will be limited to "opining that he made no error." Moreover, the fact-finder may conclude that the cytotechnologist is wrong since many may presume that the expert pathologist must be right. Finally, the defendants cannot obtain their own expert to bolster their case, because no cytology slides remain for review. The dissent concludes that summary judgment for RL was proper because the trial court had good reasons to conclude that the lost evidence was going to unduly prejudice the defendant RL. Issues The most obvious issue this case brings up is the difficulty in deciding what to do when it isn't clear who lost the evidence, or when a third party lost the evidence. In contrast, is the situation where one party is responsible for inadvertently losing the evidence or, even worse, where one party deliberately loses or destroys evidence. This is called spoliation of evidence and includes meaningful alteration of the evidence. The court will determine the severity of the sanction for spoliation by the degree of willfulness or bad faith and the extent of the prejudice suffered by the non-responsible party. For example, in one case, surgeons performed a hepatectomy after a small needle core liver biopsy was interpreted as cancer. The hepatectomy specimen showed only cirrhosis and no cancer was found. The small needle liver biopsy was subsequently lost by the hospital. The patient sued for medical malpractice, claiming that the core biopsy was misdiagnosed and lead to an unnecessary hepatectomy. The trial court instructed the jurors that they could "draw the strongest possible inference against [the hospital] as to what the lost cytology slides would have shown." In other words, losing the slide means your opposition can make the slide show whatever they want. Interestingly, in the hepatectomy case, the defendants prevailed by arguing that the diagnosis of cancer and the decision to undergo resection were reasonable, regardless of the cytology results, because the patient had active hepatitis B and a suspicious liver mass on imaging. They argued that even with a negative cytology result the surgeons would have gone ahead with the hepatectomy. But our case is different. Neither party, according to the majority, is responsible for losing the slides, or put another way, either party might be responsible for losing the slides. Several judicial options exist and none are neutral. One is to impose no sanctions and to proceed as usual, only without the slides, as the majority opinion advocated. This likely favors the plaintiff, particularly when there is a subsequent surgical specimen with cancer. A second option is to try to determine which party the court thinks is more prejudiced by the lost slides and then impose a legal remedy, as the trial court did in our hypothetical case by granting summary judgment. The difficulty is that it will not be clear which side is more prejudiced until the slide is recovered. The critical question of whether the negative diagnosis fell below the standard of care can likely be adequately answered only if the parties and their experts can review the slide. A third option might be to not allow Dr. Experta's testimony if the defendant can show that the expert knew or should have known that there was going to be litigation [ 1 ]. Arguably Dr. Experta should have returned the slides with greater care, regardless of whether litigation was contemplated or whether, as it seems in this case, the slides were sent to routinely review pathology slides before treatment. Interestingly, each option leads to a different result based on bias about which party is ultimately more responsible or more likely responsible for loosing the slides and a bias about the standard of care. The majority's opinion included a note that the benefit to BGL from losing the slides can't be ignored, a clear statement that the plaintiff had no responsibility for losing the slides, and a simplistic view about the standard of care reflected by the statement that "the slide either showed the cancer cells or it did not." The majority's subtext is that the plaintiff did nothing wrong and they weren't completely sure about the laboratory. The dissenting opinion, in contrast, treated Dr. Experta as the plaintiff's agent and was dissatisfied that Dr. Experta returned the slide to BGL instead of RL, even though at the time BGL had purchased RL. The dissent also showed a more nuanced understanding of the standard of care, conveying skepticism about Dr. Experta's look back conclusion that the Pap smears were "consistent" with adenocarcinoma. The dissent reasoned that perhaps the cells are consistent with malignancy only in the retroscope and that a defense expert might reasonably conclude that it was "not below the standard of care to determine the biopsy negative" were the slides available for review. Comment The facts often surrounding a cytopathology medical malpractice case are that there is a subsequent biopsy or surgical specimen with a discrepant diagnosis. If, as in the hypothetical Penny vs. GBL , a court does not dismiss the plaintiff's case, the absence of the cytology slide will likely impact the defendant cytologist more adversely because many people will assume that the cancerous cells were on the slide and the cytologist or cytotechnologist missed the cancer, which, after all, was present on the subsequent biopsy. In the hypothetical's facts, this was particularly true since Dr. Experta had already opined that the Paps were "consistent with adenocarcinoma." The defendant cytologist/cytology lab is at a serious disadvantage if the court allows Dr. Experta's opinion as admissible evidence. The majority's comment that GBL benefited from the lost slides and the appellate court's decision to allow the plaintiff to go to trial, suggests the court's bias that they believed Dr Experta's interpretation was the correct one. Admittedly, a cytology lab or cytologist does benefit if slides are lost and the court does not attach any responsibility for losing the slides to the potential laboratory or cytologist defendant, because the plaintiff will not have enough evidence to prevail. Similarly, if the slides are discarded after the legal time periods the likelihood of a successful lawsuit is slim because the plaintiff will not have enough evidence, and the defendant complied with legal requirements regarding slide retention. The hypothetical case of Penny vs GBL differs. Remember that once the court allows the case to go to trial, GBL only benefits from absent slides if the defendant initially misinterpreted the slides. If the slides contained no malignant cells, and Dr. Experta over-interpreted the slides, then GBL is prejudiced by not being able to show the slide. Cytology slides are typically in possession the cytology labs. The slides may be sent out for a variety of reasons, but at least initially the lab has possession of the cytology slides. This is both an advantage and a disadvantage for the lab. The disadvantage is that if slides are lost while in the lab's possession a court will typically see it as the lab's responsibility to safeguard the slides. This allows the plaintiff to have the jury infer whatever is best for the plaintiff's case; that the slide had malignant cells when the cytology diagnosis was benign or that there were only benign cells when the diagnosis was malignant, as in the hepatectomy case. The advantage for the cytology lab is that it is in a position of relative control. The lab can implement a system to help reduce the chances of losing a slide and reduce the risk if a slide is lost. Although the lab may want to employ the help of an attorney experienced in these matters, there are several steps every cytology lab can take. First, the lab should ensure that cytology slides are retained for the time that the current federal CLIA regulations, applicable state regulations and the CAP checklist require. All glass cytology slides must be retained for at least 5 years and fine needle aspiration slides retained for 10 years. (Some state regulations may require longer times.) The CAP checklist also requires policies for "protecting and preserving the integrity and retrieval of original slides in cytopathology" and "to ensure defined handling and documentation of the use, circulation, referral, transfer and receipt of original slides to ensure availability of materials for consultation and legal proceedings." Keeping careful records about when and which slides are released to whom is essential for reducing the risk of losing slides. In the send out cases where slides are sent out for a routine second opinion not sought in contemplation of a law suit or because a patient will be treated elsewhere, the lab might obtain the borrower's explicit written agreement that the borrower has responsibility for the slides with an explicit provision about a duty to indemnify the cytology lab for any losses due to a lost slide or slides. A documented telephone call to request the return of tardy slides may also be worthwhile. In cases where slides are requested in contemplation of a lawsuit, the lab may be able to implement a policy that review of slides is done at the lab, ensuring that the lab retains possession. Alternatively, the lab may pursue or agree to a court order requiring production of the slides that clearly addresses who is responsible for the slides and includes an indemnity clause in the event the slides are lost. It is reasonable to make a distinction between sending out non-reproducible slides to an institution that will treat the patient and sending out non-reproducible slides to the plaintiff's expert witness. It is, therefore, important that the lab understands the purpose for which the slides are requested. A documented telephone conversation may clarify the purpose of the outside review and allow the lab to appropriately "triage" the case. Laboratory administrators should understand that the cytology lab serves a public function in safeguarding cytology slides. Court's have recognized a public policy reason to have the laboratory safeguard slides, in part so that the slides are available in the event of malpractice litigation. At the same time, many states have statutes that give patients the right to examine and copy their medical records (the recent federal HIPAA does the same). These two propositions are not mutually exclusive. One state case considered the question of whether a patient/plaintiff had a right to "immediate possession of pathology slides" and concluded that the plaintiff did not. The court decided that the patient's rights did not exceed the patient's statutory right in the slides. In other words the court was not going to find a common law, or customary, right in the slides that gave the patient a greater right than the applicable statute. The decision noted that the legislative history of the statute included remarks that pathology slides were part of the medical record. The judges then approached the second question about what to do when the "medical record" cannot be duplicated, as with a Pap smear or other cytology slide. In answering this question, the court noted that hospitals and laboratories have public as well as private duties. One of their public duties is to retain slides so that the slides are available in the event of malpractice litigation. This meant that patients do not have a legal right to possess parts of the medical record that cannot be duplicated. The court, however, did not grant the lab complete authority to never release the slides. Since the public policy reason depended fundamentally on preserving slides to help the legal system run smoothly, the decision reminded the laboratory or hospital that, pursuant to the clear terms of a statute, it must send the original slides to a "licensed institution, laboratory or physician" at the patient's written request. The dissent characterized the pivotal issue differently and concluded that the patient had a right to immediate possession because the slides contained the patient's cells and she had the right "to control one's body." The dissent reasoned that recent advancement in genetic science and the accompanying difficult privacy issues raised by genetic information strengthened the patient's right to possession of the cells on the glass slide. The majority addressed this argument and, citing the well known case of Moore vs Regents of the University of California [ 2 ], reminded the reader that no court has recognized that a patient has property rights in cells taken for diagnostic purposes. I mention the dissent to emphasize that the issues surrounding possession and use of glass slides are complex and evolving and it often difficult to predict what a court will say. In summary, lost slides can be a problem for cytology laboratories and cytologists, whether responsible for losing the slides or not. The prudent cytologist will minimize the risk of lost slides, because, as the hypothetical case of Penny v GBL illustrates, lost slides can result in problems for cytologists and the cytopathology laboratory A good place to start is to ensure that existing CLIA regulations, applicable state regulations and other guidelines, such as the CAP checklist, are in place in the laboratory. Thinking about the issue and implementing appropriate risk management strategies with the help of an attorney are also prudent measures. Table 1 The top 10 take home messages 1. Spoliation of evidence includes meaningful alteration of the evidence as well as losing the evidence. 2. A court may look at which party benefits from losing the unique and irreplaceable slides as well as which party was last in possession of the slides and impose appropriate sanctions. 3. One remedy a court may impose on the party responsible for losing the slides is to allow the factfinder, whether jury or judge, to infer that the slides show whatever is best for the opposition's case. 4. Cytologists should adhere to federal, state and respected published checklists regarding how long to keep slides and implement lab policies regarding the circulation and transfer of original slides. 5. A cytology lab should consider calling to find out the reason why a slide is requested so that it can respond appropriately. 6. A cytology laboratory should consider sending unique and irreplaceable cytology slides by registered mail. 7. Some state legislatures and courts recognize a public policy reason for cytology labs to retain and protect cytology slides. 8. Although patients have a right to inspect slides, they typically do not have a legal right to immediate possession. 9. A cytology lab should consider implementing a policy that requires review of the slides at the lab when slide review is in contemplation of a law suit. If this is not possible, then the lab can agree to a court order that clearly spells out who is responsible if slides are lost with indemnification to the lab for lost slides. 10. When in doubt, the prudent cytologist should contact their attorney.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC524214.xml

548319

Iron homeostasis in neuronal cells: a role for IREG1

Background Iron is necessary for neuronal function but in excess generates neurodegeneration. Although most of the components of the iron homeostasis machinery have been described in neurons, little is known about the particulars of their iron homeostasis. In this work we characterized the response of SH-SY5Y neuroblastoma cells and hippocampal neurons to a model of progressive iron accumulation. Results We found that iron accumulation killed a large proportion of cells, but a sub-population became resistant to iron. The surviving cells evoked an adaptative response consisting of increased synthesis of the iron-storage protein ferritin and the iron export transporter IREG1, and decreased synthesis of the iron import transporter DMT1. Increased expression of IREG1 was further substantiated by immunocytochemistry and iron efflux experiments. IREG1 expression directly correlated with iron content in SH-SY5Y and hippocampal cells. Similarly, a high correlation was found between IREG1 expression and the rate of iron efflux from SH-SY5Y cells. Conclusions Neuronal survival of iron accumulation associates with increased expression of the efflux transporter IREG1. Thus, the capacity of neurons to express IREG1 may be one of the clues to iron accumulation survival.

Background Because of its intense oxidative metabolism, the brain consumes a high fraction of total oxygen generating large amounts of reactive oxygen species [ 1 , 2 ]. Although brain antioxidant defenses function properly during most of human life, a number of neurodegenerative processes which involve redox-active iron accumulation become evident with age [ 3 - 5 ]. Iron is a pro-oxidant that in the reductive intracellular environment catalyses hydroxyl radical formation through the Fenton reaction [ 6 ]. At present, the crucial components of the iron homeostasis machinery have been identified. Thus, current efforts should be directed to the understanding of the mechanisms that regulate cellular iron levels and antioxidant defenses. This is of primary importance for the development of strategies to ameliorate iron accumulation and oxidative damage in neurons. In vertebrates, cellular iron levels are post-transcriptionally controlled by the activity of iron regulatory proteins (IRP1 and IRP2), cytosolic proteins that bind to structural elements called iron-responsive elements (IREs). IREs are found in the untranslated region of the mRNAs of the major proteins that regulate cellular iron homeostasis: the transferrin receptor, involved in plasma-to-cell iron transport, and the iron-storage protein ferritin. IRP2-/- mice are born normal but in adulthood develop a movement disorder characterized by ataxia, bradykinesia and tremor [ 7 ]. IRP1-/- mice are normal with slight misregulation of iron metabolism in the kidney and brown fat [ 8 ]. Thus, IRP2 seems to dominate the physiological regulation of iron metabolism whereas IRP1 seems to predominate in pathophysiological conditions. Iron is internalized into cells by the import transporter DMT1. Four DMT1 isoforms have been identified that differ in both the N-and the C-termini [ 9 ]. Two of the isoforms have a 3' iron responsive element (IRE) in their mRNA. Additional variation is given by exons 1A and 1B in the 5' end. Expression of DMT1 in response to iron availability follows a pattern similar to transferrin receptor [ 10 ], but its control by the IRE/IRP system is not clear [for review see [ 11 ]]. A new iron transporter, IREG1, also known as ferroportin or MTP1, was recently described [ 12 , 13 ]. The protein is expressed mainly in enterocytes and macrophages [reviewed in [ 14 ]]. In enterocytes IREG1 is responsible for iron efflux during the process of intestinal iron absorption, while in Kupffer cells IREG1 mediates iron export for reutilization by the bone marrow [ 15 ]. The presence of both DMT1 and IREG1 has been described in neurons, glioma cells and astrocytes [ 16 - 18 ]. The presence of IREG1 in neurons opens the possibility that they may be able to down-regulate intracellular iron concentration through its expression. In this study we examined iron homeostasis in SH-SY5Y neuroblastoma cells and hippocampal neurons. We found that iron accumulation killed a large proportion of cells, but a sub-population became resistant to iron accumulation developing an adaptative mechanism intended to decrease intracellular iron content. Results Iron accumulation and cell death Iron accumulation was determined in SH-SY5Y cells grown to confluence and then cultured for two days in media containing from 1.5 to 80 μM iron (Figure 1A ). Total cell iron increased with increasing extracellular iron, reaching a plateau at 40–80 μM Fe (Figure 1B ). The observed increase in cell iron was accompanied by increases in the labile iron pool (Figure 1C ). Iron accumulation indeed caused loss of cell viability, with hippocampal neurons demonstrating higher sensitivity than SH-SY5Y cells to iron treatment (Figure 2 ). Nevertheless, a sub-population of cells survived to high iron concentrations. It was of interest to inquire into the processes underlying this adaptation, since they could help to understand iron accumulation observed in a number of neurodegenerative diseases. Consequently, we characterized the components of the iron homeostasis machine during the process of iron accumulation. Ferritin and DMT1 regulation Ferritin, the main iron-storage protein in mammalian cells, is considered the first line of defense against iron overload. Increasing iron from 1.5 to 5 μM produced a robust 4-fold increase in cell ferritin content (Figure 3A ). Further increases in iron induced additional increases in ferritin. At 80 μM extracellular iron, ferritin increased 11-fold compared to the basal 1.5 μM iron condition. In molar base, ferritin increased more than iron. The iron to ferritin mol : mol ratio decreased from 1500 at 1.5 μM Fe to 400 at 10 μM Fe to 200 at 80 μM Fe (Figure 3B ). We further characterized iron homeostasis in SH-SY5Y cells by examining the expression of the iron importer DMT1 (Figure 4 ). A 3.5-fold decrease in DMT1 protein expression was observed when iron increased from 1.5 to 80 μM. The presence of DMT1 even at high iron concentration explains the sustained iron uptake observed at 80 μM Fe [ 19 ]. Thus, DMT1 activity persisted even under conditions of iron accumulation that preceded cell death. IREG1 expression and functionality Given that the presence of IREG1 in the central nervous system has been reported [[ 16 ]], it was of interest to examine if it participates in neuronal iron homeostasis. Western blot analysis revealed that SH-SY5Y cells expressed anti-IREG1 reactive bands of 65.3 and 122.1 KDa molecular weight (Figure 5A ). Densitometric analysis revealed a 10-fold increase in the 122.1 KDa band in the 1.5 to 80 mM Fe range while the 65.3 band had a minor increase (Figure 5A ). Both bands were eliminated if the antibody was incubated with the immunogenic peptide before the assay (Figure 5B ). A similar pattern was obtained with an independent anti-IREG1 antibody (the kind gift of Dr. David Haile). Thus, it is most likely that the 65.3 and 122.1 KDa bands correspond to the monomer and dimer of IREG1. The stability of the 122.1 KDa band was dependent of the concentration of b-mercaptoethanol in the sample buffer since increasing b-mercaptoethanol produced a shift in the 122.1 KDa /65.3 KDa band ratios (Figure 5C ). It is possible that in neuronal cells Ireg1 tends to form S-S bridged dimers resistant to the electrophoresis conditions. The presence of IREG1 in SH-SY5Y neuroblastoma cells and hippocampal neurons was further documented by immunocytochemistry. IREG1 was detected in both types of cells, with a predominantly cytosolic distribution pattern (Figure 6 ). The levels of IREG1 expression were directly proportional to the amount of iron in the culture. Thus, it was determined by two independent methods that IREG1 expression in neuronal cells increased with cell iron content. Efflux of iron from neurons has never been reported. In view of the presence of IREG1, we tested whether SH-SY5Y cells actually had an iron efflux function. To that end, iron efflux from cells pre-cultured for 2 days with varied iron concentrations was determined by atomic absorption spectrometry. This method was preferred to the use of radioisotopic iron since the latter could underestimate a putative iron efflux because of isotope dilution with the pre-existing iron pool (see Figure 1B ). SH-SY5Y cells had discrete but measurable iron efflux activity (Figure 7A ). The iron efflux rate increased markedly in the 20–80 μM Fe range (Figure 7B ). Interestingly, the efflux rate correlated closely with the presence of the 122.1 KDa band, while the correlation between efflux activity and the 62.5 KDa band was weaker (Figure 7C ). Discussion The number of neurological diseases associated with iron accumulation in the brain underlines the need for increased knowledge of the mechanisms of brain iron homeostasis. In this study we show that iron accumulation by SH-SY5Y neuroblastoma cells and hippocampal neurons resulted in cell death of part of the population, while another fraction survived by adapting the expression of iron homeostasis proteins. Iron content increased significantly as a function of Fe in the culture up to 20–40 μM Fe, increasing very little thereafter up to 80 μM Fe. Cell iron increase was accompanied by increased ferritin content. The increase in ferritin more than compensated for the increase in iron. Iron to ferritin mol ratios of 1500, 260 and 190 were obtained for 1.5, 20 and 80 μM Fe in the culture media. Thus, the IRE/IRP system of SH-SY5Y cells over-responded to iron accumulation in terms of ferritin expression. Despite the increase in ferritin, the LIP increased between 1.5 and 80 μM Fe. This finding clearly indicates that in SH-SY5Y cells the level of labile iron is a function of total iron, even in the presence of ample ferritin supply. It is possible that ferritin-stored iron contributes to the LIP each time that ferritin undergo lisosomal degradation. Iron accumulation was accompanied by a marked decrease in DMT1 expression. Nevertheless, some DMT1 persisted even at 40–80 μM iron. The persistence of DMT1 at high iron concentrations could underline the continuous iron uptake observed under these conditions [ 19 ]. This is curious because at 40–80 μM Fe cells were dying. Sustained DMT1 expression points to the inability of neuronal cells to shut-off iron uptake and the need for additional defense mechanisms to prevent iron-mediated cell death. The discovery of increased IREG1 expression in response to cell iron accumulation is a major break-through in the understanding of cell survival under conditions of iron accumulation. Total IREG1, and especially a putative IREG1 dimer, increased markedly in the 20–80 μM Fe range. Thus, in SH-SY5Y cells IREG1 is up-regulated by increased cell iron. Expressed IREG1 was functional since it associated with increased iron efflux activity. Iron efflux activity in astrocytes [ 18 ] and neurons (this work) indicate that iron efflux from brain cells is a dynamic process, and highlights the importance of iron transporters as determinants of iron accumulation. The regulation of IREG1 expression is unknown but seems to be cell-specific. In enterocytes, IREG1 expression is induced by iron deficiency [ 13 ] while in macrophages iron increases IREG1 expression [ 20 ]. The findings reported here indicate that in neuronal cells IREG1 has a macrophage-like regulation. This is certainly the case for cells in the 40–80 μM range that survived to iron accumulation. IREG1's predominantly cytosolic distribution pattern is similar to that of Kupffer cells [ 12 ]. Again, this distribution points to macrophage-like behavior of neuronal IREG1. In examining brain biopsies from Alzheimer's patients an intriguing question arises: Why do some neurons die or present evident signs of degeneration while others in the vicinity show a normal phenotype? Extrapolating on the data presented here, it is tempting to hypothesize that surviving neurons induce IREG1 expression while sick neurons do not. Nevertheless, at present we cannot exclude that other regulatory molecules may play a pivotal role under these conditions. Conclusions Hippocampal neurons and SH-SY5Y cells displayed an active system to regulate iron content. Nevertheless, this system was unable to block iron accumulation which resulted in death of part of the cell population. Another fraction of the cell population developed an adaptative mechanism that includes decreased expression of the import transporter DMT1 and increased expression of ferritin and the efflux transporter IREG1. The finding that neurons regulate the expression of functional IREG1 opens new avenues for the understanding and possible treatment of iron-related neurodegenerative processes. Methods Antibodies and immunodetection Antibody D-1, prepared against the C-terminal end of the IRE-containing isoform of DMT1 was used as described previously [ 10 ]. Additionally, a rabbit polyclonal antibody against peptide CGPDEKEVTKENQPNTSVV, corresponding to the consensus sequence of human, rat and mouse carboxyl-terminal sequence of IREG1, was obtained from BioSonda, Chile . Western analysis Cell extracts, cells were prepared treating cells with lysis buffer (50 μl per 1 × 106 cells of 10 mM MOPS, pH 7.5, 3 mM MgCl2, 40 mM KCl, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 0.5 μg/ml aprotinin, 0.7 μg/ml pepstatin A, 5% glycerol, 1 mM dithiothreitol, 0.1% Triton X-100). The mixture was incubated for 15 min on ice and centrifuged for 10 min at 5,000 × g. Protein concentrations were determined using the bicinchoninic acid (BCA) protein assay. The supernatant was stored at -70°C. For Western analysis, 30 micrograms of protein from each sample were boiled in Laemmli sample buffer for 5 min and subjected to SDS-PAGE on a 7.5% acrylamide gel. Proteins were transferred to nitrocellulose membrane and blocked for 1 hr at 25°C with 5% nonfat dry milk in blocking saline (20 mM Tris, 0.5 M NaCl, 0.05% Tween-20). Membranes were incubated with primary antibody overnight at 4°C, rinsed with blocking saline and incubated with horseradish peroxidase-conjugated anti-rabbit IgG antibody for 1 hr at 25°C. Transferred proteins were detected with a peroxidase-based chemiluminiscence assay kit (SuperSignal, Pierce Chem. Co., Rockford, IL). Chemiluminiscence was detected using a Molecular Imager FX device (Bio-Rad, Hercules, CA). The bands were quantified by densitometry using the Quantity One (Bio-Rad) software. Cell culture and iron challenge Human neuroblastoma SH-SY5Y cells (CRL-2266, American Type Culture Collection Rockville, MD), were seeded at 1 × 105 cells in 2-cm 2 plastic wells and cultured in a 5 % CO 2 incubator in MEM/F12 medium supplemented with 10 % fetal bovine serum and 5 mM glutamine. The medium was replaced every two days. Under these conditions, doubling time was about 48 hours. After 8 days in culture, the culture reached a steady-state number of cells. At this time, cells were challenged with iron for the next two days as described [ 19 ]. In brief, low-iron culture media was supplemented with either 1, 5, 10, 20, 40 or 80 μM Fe 3+ as the complex FeCl 3 -sodium nitrilotriacetate. Cell viability was quantified by the MTT assay (Molecular Probes, OR) following the manufacturer's instructions. This model of iron loading attempts to replicate neuronal iron accumulation that occurs during life [ 4 ]. Hippocampal neurons were prepared from E18.5 rat embryos [ 21 ]. Neurons were plated over poly-L-lysine coated cover slips at 100,000 cells/cm 2 . Cultures were maintained in 10% bovine serum until 3 hours after plating, when the culture medium was replaced with medium containing B27 supplement [ 22 ]. After 3 days in culture, the cells were challenged with iron as described above. Labile iron pool The intracellular labile or reactive iron pool of neuroblastoma cells was determined as described [ 23 , 24 ]. The increase in fluorescence after the addition of SIH chelator is directly proportional to the iron labile pool, i.e., iron in complexes with affinity constant < 10 6 . Immunocytochemistry Cells grown in cover slips were sequentially fixed with 2% and 4% parafolmaldehyde (PFA) in Eagles' MEM, and then washed three times with phosphate-buffered saline (PBS). The fixed cells were permeabilized with Triton-X-100 (0.2%) in PBS at room temperature for 3 min and blocked with defatted milk (10%) in PBS for more than 1 h. The cells were incubated with anti-IREG1 antibody (1:500) overnight at 4°C, washed with PBS and then incubated with Alexa-546-conjugated goat anti-rabbit IgG. The labeled cells were observed with a Zeiss LSM 510 Meta confocal laser scanning microscope. Data analysis Variables were tested in triplicates, and experiments were repeated at least twice. Variability among experiments was <20%. One-way ANOVA was used to test differences in mean values, and Turkey's post-hoc test was used for comparisons (In Stat program from GraphPad Prism). Differences were considered significant if P < 0.05. Authors' contributions MTN conceived of the study, participated in its design and coordination and drafted the manuscript. PA performed the experiments with hippocampal neurons, did the ferritin assays and participated in the analysis and interpretation of data. MN optimized the immunocytochemistry detection of IREG1, performed the confocal microscope observations and participated in the analysis and interpretation of data. VT did the Western blot, labile iron pool and viability assays and contributed to the discussion of the results. MA set up the method to determine total Fe concentration, did the sample and control measures of iron and participated in the analysis and interpretation of data. All coauthors participated in refining the text.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548319.xml

545601

Development of a method for screening short-lived proteins using green fluorescent protein

A method for identifying short-live proteins using a GFP-fusion cDNA library for monitoring degradation kinetics is described.

Background Cellular proteins differ widely in their lability, ranging from those that are completely stable to those with half-lives measured in minutes. Proteins with a short half-life are among the most critical to the cell. Regulated degradation of specific proteins contributes to the control of signal transduction pathways, cell-cycle control, transcription, apoptosis, antigen processing, biological clock control, differentiation and surface receptor desensitization [ 1 , 2 ]. Rapid turnover makes it possible for the cellular level of a protein to change promptly when synthesis is increased or reduced [ 3 ]. Furthermore, degradation rate is itself subject to regulation. For instance, inflammatory stimuli cause the rapid degradation of IκBα, the inhibitor of NFκB, resulting in the activation of that transcription factor [ 4 - 6 ]. Analysis of labile proteins has been time-consuming and labor-intensive. The most definitive form of analysis requires pulse-chase labeling cells and immunoprecipitation extracts. In vitro assay of degradation is simpler than in vivo analysis, but an in vitro assay system may not fully mimic the degradation of proteins in the cells. Genome-wide functional screening and systemic characterization of cellular short-lived proteins has received little attention [ 7 ]. GFP, the green fluorescent protein from the jellyfish Aequorea victoria , has been widely used to monitor gene expression and protein localization [ 8 ]. Recently, we demonstrated that fusion of GFP to the degradation domain of ornithine decarboxylase [ 9 ], a labile protein, can destabilize GFP [ 10 ] and that the degradation of an IκB-GFP fusion protein can be monitored by GFP fluorescence [ 11 ]. These studies demonstrate that introducing GFP as a fusion within the context of a rapidly degraded protein does not alter the degradation properties of the parent molecule, and that the GFP moiety of the fusion protein is degraded along with the rest of the protein. GFP fluorescence, which provides a sensitive, rapid, precise and non-destructive assay of protein abundance, can therefore be used to monitor protein degradation [ 12 ]. Furthermore, fluorescence associated with single cells can be analyzed using fluorescence-activated cell sorting (FACS), a technology easily adapted to high-throughput screening [ 13 ]. We developed a GFP-based, genome-wide screening method for short-lived proteins. We made a GFP fusion expression library of human cDNAs and introduced the library into mammalian cells. Transfected cells were FACS-fractionated into subpopulations of uniform fluorescence. Individual subpopulations were treated with cycloheximide (CHX) to inhibit protein synthesis and re-sorted after 2 hours of treatment. Sorting was gated to recover cells with a fluorescent signal that was diminished compared to the population mode. Repeated application of this process resulted in a high yield of clones that encode labile fusion proteins. Results The selection scheme is shown in Figure 1 . GFP-cDNA expression libraries were transfected into mammalian cells and cells fractionated into subpopulations, each with a narrow range of fluorescence intensities. Subpopulations were then twice enriched for cells with the desired characteristics. Plasmid DNAs were recovered from the selected cells, subjected to sequence analysis and functionally verified. We made the expression libraries with modified pEGFP C1/C2/C3 vectors by cloning the cDNAs downstream of EGFP. The titer of the library was found to be high: around 10 6 cell transformants per microgram of DNA. In addition, we confirmed by PCR amplification that 95% of clones contained a cDNA insert larger than 800 base-pairs (bp) (data not shown). The libraries were thus deemed to be useful for screening short-lived proteins in mammalian cells. We used 293T cells as the recipient. These cells offer two advantages. First, they express the SV40 large T antigen. This allows the library plasmids, which contain an SV40 origin of replication, to be highly replicated. Plasmids can therefore be recovered easily. Second, 293T cells have high transfection efficiency. After we introduced the GFP-fusion libraries into the mammalian cells, the transfected cells were easily separated by FACS from non-transfected cells or cells transformed by non-productive constructs. We imposed selection for cells that became less bright within 2 hours of exposure to cycloheximide (CHX), a protein synthesis inhibitor. We chose a short treatment time to avoid selecting cells that became dimmer as a result of secondary responses other than rapid turnover of the GFP tagged proteins. To enrich for cells that are susceptible to CHX treatment, we started with a cell population that has an approximately log-normal fluorescence histogram distribution, with a working range of 1.5 to 4.5 logs. We used FACS fractionation to divide this population into five subpopulations (R2, R3, R4, R5, R6) of ascending brightness, gating each on successive one-half log 10 intervals of fluorescence (Figure 2 ). Each subpopulation (R2-R6) was divided into two; one portion was treated with 100 μg/ml CHX for 2 hours and the other left untreated. Subpopulations were then reanalyzed to determine whether they had retained a distribution consistent with the gating criteria used to obtain this narrow subpopulation and were susceptible to CHX treatment. We found that subpopulations R3 and R4 were susceptible to CHX treatment (Figure 3 ), whereas R5 and R6 did not change their fluorescence properties in response to CHX (data not shown). The fluorescence intensity of R2 was too low to detect after CHX treatment. The lack of susceptibility of the brighter R5-R6 subpopulations was most likely the result of their expressing predominantly stable proteins, which would be expected to provide more intense fluorescence. We selected R4 for further screening in this study. We collected 10 6 cells from the shifted population, the left shoulder of the population observed in the CHX-treated but not in the untreated R4 cells (Figure 3 ). Plasmid DNAs were recovered from the sorted cells and were propagated in Escherichia coli , resulting in a total of 400 clones. The individual clones were stored in 15% glycerol LB medium in a 96-well format. To perform second-round selection, we grouped the 400 clones into 12 pools, each composed of approximately 33 clones. The individual pools of clones were cultured and used for plasmid preparation. We transfected these 12 groups of plasmid DNA into 293T cells and again subjected them to FACS analysis and gating as before. The EGFP-C1 vector was used as a control. Because enhanced green fluorescent protein (EGFP) is a stable protein, its fluorescence intensity would not be changed by treatment with CHX. We found that eight of the 12 groups showed a decrease of the fluorescence intensity peak by 30-50% (compared to untreated cells) after 2 hours of CHX treatment. In four out of 12 groups, no change in fluorescence intensity was detected. To isolate individual clones with the desired property, we randomly chose one of the eight CHX-responsive groups and characterized individual clones. We analyzed 30 clones from this group by individually transfecting them into 293T cells and determining the half-life by FACS-based analysis of CHX chase kinetics. We found out that 22 clones showed a decrease in fluorescence intensity ranging from 30 to 90% after treatment with CHX for 2 hours. Assuming first order kinetics of turnover, this single-time-point experiment implies that the proteins corresponding to these 22 clones have a range of half-lives ranging from about half an hour to 3-4 hours (Table 1 ). The 22 clones were partially sequenced and BLAST used to search for similar protein sequences in the National Center for Biotechnology Information (NCBI) public database. Of these, 19 corresponded to annotated genes in GenBank and the remaining three to unknown genes. Sequencing analysis also indicated that the inserts of these clones corresponded to full-length or near full-length translation reading frames. As no data are available on the intracellular turnover kinetics of the 19 identifiable proteins, we picked three clones - splicing factor SRp30c, a guanine nucleotide-binding regulatory protein (G protein), and cervical cancer 1 proto-oncogene protein - and examined their turnover by CHX chase and western blot analysis. These three clones (Table 1 , numbers 5, 19 and 26) were estimated in the fluorescence-based screen to have diverse turnover kinetics; two of them have a half-life of less than 1 hour while the third turns over somewhat more slowly. To confirm these estimates of turnover by a means independent of GFP fluorescence, 293T cells were transfected with these clones, treated with CHX and periodically sampled over the next 3 hours. Western blot analysis of cell extracts with antibody to GFP showed that the abundance of all three fusion proteins diminished in the presence of CHX (Figure 4a ). The half-life of the proteins determined by western blot analysis was similar to that determined by FACS analysis. Two of the proteins showed a half-life of about 1 hour, while the proto-oncogene protein appears to initiate abrupt degradation within about 2 hours of treatment with CHX. The results for all three proteins are thus consistent with those observed using the fluorescence-based screening method. As positive and negative controls, we similarly analyzed cells expressing a destabilized version of EGFP, d1EGFP, whose short half-life has been previously characterized [ 10 ], and a stable EGFP protein (Figure 4b ). Sequencing analysis indicated that these three GFP fusion cDNAs do not contain a full-length coding sequence. SRp30c cDNA is missing 17 amino acids at its amino terminus, G protein 20 amino acids, and proto-oncogene p40 three amino acids. To exclude the possibility that the missing amino acids or the fused GFP domain contribute artifactually to protein liability, we amplified the full-length coding sequences of these three genes and expressed them as Myc fusion proteins. Their turnover was examined by CHX chase and western blot analysis with antibody to the Myc tag (Figure 5 ). Turnover rates assessed in this way were similar to those of the GFP fusion proteins obtained from library screening, ruling out the presence of these artifacts. This technology is subject to two kinds of false-positive results. First, fusion to a detection tag such as GFP or Myc may affect the folding of tagged proteins, which could accelerate their turnover. Second, expression of the fusion proteins under the control of viral promoter elements could result in overexpression, with concomitant misfolding or failure to associate with endogenous interaction partners. To rule out these artifacts, we measured the degradation of native non-fusion endogenous counterparts of two of the proteins we identified, those for which antibodies were available. Turnover of the proteins associated with clone 19 and clone 25 was measured by CHX chase and western blot analysis. The results (Figure 6 ) demonstrated that the half-life of clone 19, a guanine nucleotide-binding regulatory protein (G protein), was less than 1 hour and the half-life of clone 25, heat-shock 70 kD protein (hsp70), was about 1 hour. The turnover of the native proteins is thus at least as fast as that of the corresponding clones analyzed in the screen, suggesting that the technology can accurately identify short-lived proteins. Discussion The abundance of a given cellular protein is determined by the balance between its rate of synthesis and degradation. The two are of equal importance in their effect on the steady-state level. Furthermore, degradation determines the rate at which a new steady state is reached when protein synthesis changes [ 3 ]. Despite its importance, degradation, the 'missing dimension' in proteomics [ 7 ], has received far less comprehensive attention than synthesis. This deficiency has arisen because developing the tools for a proteome-wide study of protein turnover is technically challenging. Proteins that are labile tend to be present at low abundance, and methods for characterizing turnover time are laborious. We have developed an efficient and rather specific screen by combining GFP fluorescence, as a high-throughput measure of protein abundance, with pharmacologic shutoff of protein synthesis. Of 30 clones that were recovered from the screen (Figure 1 ) and individually examined by CHX treatment and FACS analysis, 22 (73%) are associated with proteins with a half-life of less than 4 hours. Given the relative rarity of rapidly degraded proteins in the proteome [ 14 ], this result demonstrates the specificity of the screening method. We have so far analyzed a restricted subset of the clones that were recovered in our screening procedure - 30 clones present in one of eight positive pools (among 12) from the R4 population. A second population, R3, appears to be equally rich in clones responsive to CHX. Extrapolation from this small sample implies that perhaps 300-400 (that is, 22 × 8 × 2) clones within the GFP-cDNA library may be found to be associated with proteins that are labile according to our secondary screening criterion. In contrast to the results with the less bright R3 and R4 cell populations, the failure to detect a CHX-sensitive subpopulation among the brighter R5-R6 cells is consistent with the expectation that labile proteins tend to be of lower abundance than more stable proteins. For some of the proteins uncovered in this survey, rapid turnover can be rationalized as intrinsic to their cellular function. SRp30c factor (accession number U87279) is responsible for pre-mRNA splicing. Alterative splicing is a commonly used mechanism to create protein isoforms. It has been proposed that organisms regulate alternative splice site selection by changing the concentration and activity of splicing regulatory proteins such as SRp30c in response to external stimuli [ 15 ]. The finding that SRp30c is a short-lived protein is consistent with its postulated regulatory function. The G proteins are a ubiquitous family of proteins that transduce information across the plasma membrane, coupling receptors to various effectors [ 16 , 17 ]. About 80% of all known hormones, neurotransmitters and neuromodulators are estimated to exert their cellular regulation through G proteins. The G protein (accession number M69013) shown here to short-lived is a G protein α subunit that transduces signals via a pertussis toxin-insensitive mechanism [ 18 ]. Like other pertussis toxin-insensitive G proteins such as the Ga12 class, it causes the activation of several cytoplasmic protein tyrosine kinases: Src, Pyk2 (proline-rich tyrosine kinase 2) and Fak (focal adhesion kinase) [ 19 ]. However, it is not known how this G protein is regulated. Its rapid turnover suggests a testable mechanism of its regulatory activation. Cervical cancer 1 proto-oncogene protein p40 (accession number AF195651), is a third protein shown here to turn over rapidly, but its function is unknown. Further studies of its turnover may provide important information on its function and regulation. In mammalian cells, proteasomes have the predominant role in the degradation of short lived proteins, whereas lysosomal degradation appears to be quantitatively less important [ 20 ]. Determining the mechanism that cells use to degrade the proteins uncovered by the method described here will require the use of specific inhibitors [ 21 ]. Before degradation, most short-lived proteins are covalently coupled to multiple copies of the 76-amino-acid protein ubiquitin [ 22 ], a reaction catalyzed by a series of enzymes [ 23 ]. These ubiquitinated proteins are recognized by the 26S proteasome and degraded within its hollow interior [ 24 ]. This system of regulated degradation is central to such processes as cell-cycle progression, gene transcription and antigen processing. A few proteins have been found to be exceptions [ 25 , 26 ]; like ODC, they do not require ubiquitin modification for degradation by the proteasome. In most cases it is not clear how short-lived proteins are selected to be modified and degraded. Some rapidly degraded proteins have been shown to contain an identifiable 'degradation domain'. Removal of this degradation domain makes such proteins stable, and appending this domain to a stable protein reduces its stability. Such a degradation domain has been identified in a number of short-lived proteins, including the carboxy terminus of mouse ODC [ 6 , 27 ] and the destruction box of cyclins [ 28 ]. In some cases, the signal is a primary sequence - like the PEST sequence [ 29 , 30 ]. However, the identifiable structural features of such degradation domains are not sufficiently uniform to provide a reliable guide to identifying labile proteins. The method we have described does not use ubiquitin conjugation as a search criterion. This approach thus has the potential to discover labile proteins regardless of whether ubiquitin modification plays a role in their turnover. Once a large and representative sample of short-lived proteins is identified, a search for structural motifs among these proteins may facilitate the discovery of those motifs which correlate to protein degradation. Conclusions In this study we have developed an innovative technology to identify labile proteins using GFP-fusion expression libraries. Using this technology we have discovered short-lived proteins in a high-throughput format. This technology will greatly facilitate the discovery and study of short-lived proteins and their cellular regulation. Materials and methods Construction of GFP-cDNA expression libraries Messenger RNAs from brain, liver, and the HeLa cell line (Clontech) were used as templates for cDNA synthesis, using a cDNA synthesis kit from Stratagene according to the manufacturer's recommendation, with some modifications. First-strand cDNA was synthesized using an oligo(dT) primer-linker containing an Xho I restriction site and with StrataScript reverse transcriptase. Synthesis was performed in the presence of 5-methyl dCTP, resulting in hemimethylated cDNA, which prevents endogenous cutting within the cDNA during cloning. Second-strand cDNA was synthesized using E. coli DNA polymerase and RNase H. Adaptors containing Eco RI cohesive ends were introduced into the double-stranded cDNA, which were then digested with Xho I. The cDNAs contained two different sticky ends: 5' Eco RI and 3' Xho I. The cDNAs were separated on a 1% SeaPlaque GTG agarose gel in order to collect those larger than 800 bp. After extracting cDNAs from the agarose gel with AgarACE-agarose-digesting enzyme followed by ethanol precipitation, the cDNAs were directionally cloned into EGFP-C1/2/3 expression vectors with three open reading frames (ORFs) (Clontech). The vectors were modified within the multiple cloning sites in order to be compatible with the cDNA orientation. By this means, cDNA ORFs were aligned to the carboxy terminus of EGFP. The host cell used for plasmid transfection and expression, 293T, expresses the SV40 large T antigen. Therefore, the cDNA EGFP-C1/2/3 vector containing the SV40 origin of replication can replicate independently from chromosome DNA in the host cells, which facilitates the recovery of plasmid DNAs from the host cells. Transfection of the libraries into 293T cells 293T cells were cultured at 37°C in DMEM (Invitrogen) supplemented with 10% FBS, 1% nonessential amino acids and 100 U/ml penicillin, 0.1 mg/ml streptomycin. One day before transfection, cells were seeded in 10-cm plate in 10 ml growth medium without antibiotics. Transfection was performed using Lipofectamine 2000 reagent according to the manufacturer's instructions. Samples (25 μg) of a cDNA library were diluted in 1.5 ml Opti-MEM (Invitrogen). Lipofectamine 2000 was diluted in 1.5 ml Opti-MEM and mixed with diluted DNA. After 20 min incubation, the DNA-Lipofectamine 2000 complex was added to the cells. The cells were incubated for 16 h before analysis. FACS analysis of GFP-expressing cells Cells were harvested by trypsinization, washed, and resuspended in DMEM. Cytometric analysis and sorting were performed using a hybrid cell sorter combining a Becton Dickinson FACStarPLUS optical bench with Cytomation Moflo electronics (Stanford Beckman Center shared facility). Green fluorescence was measured using a 525/50 band pass filter. Gates were set to exclude cellular debris and the fluorescence intensity of events within the gated regions was quantified. Fluorescence-activated cell sorting was performed with a lower forward scatter threshold to detect transfected cells while ensuring that debris and electronic noise were not captured as legitimate events. Transfection efficiency was so high that normal voltages for detecting GFP were reduced. For fractionation, the cell population was gated on the basis of the fluorescence intensity. Cells were sorted at a rate of 8,000 events/sec. 10 6 cells were collected in 12 × 75 mm glass tubes containing 200 μl serum to enhance the cell survival rate. For short-lived protein screening, sorted cells were recultured in a 12-well plate and treated with or without 100 μg/ml CHX for 2 h. The cells then were collected and subjected to FACS analysis and sorting. The cells showing a decrease in fluorescence intensity with CHX treatment were collected for further analysis. Plasmid recovery Plasmid DNA was extracted from sorted cells using a Qiagen mini-plasmid preparation kit. Plasmid DNAs were eluted in water and transformed into electro-competent DH10B E. coli (Invitrogen). Bacterial colonies were transferred to 96-well plates containing LB with 50 μg/ml kanamycin and 30% glycerol. After overnight growth at 37°C, the colonies are stored at -80°C. Plasmid DNAs were prepared from individual clones, sequenced and BLAST searches performed against the NCBI database. Construction of Myc-tagged full-length coding sequences of genes To obtain full-length coding sequence of the genes, we amplified them with a human full-length cDNA kit (Panomics) according to the manufacturer's instructions. The full-length coding sequences of cDNAs were then cloned into the pCMV-Myc vector (Clontech) for expression in 293T cells. Western blot analysis of protein degradation The plasmid DNAs of individual clones were prepared and transfected into 293T cells. The transfected cells, with or without CHX treatment, were collected in PBS and cell lysates were prepared by sonication. Proteins were resolved by SDS-polyacrylamide gel electrophoresis and transferred to a membrane. Fusion proteins were detected using a polyclonal antibody against GFP (Clontech), a monoclonal antibody against the Myc epitope (Sigma), a polyclonal antibody against G protein (Santa Cruz) or an antibody against Hsp70 (Santa Cruz). Bands were visualized with SuperSignal West Pico kit (Pierce). Additional data files Additional data file 1 contains the original data used to perform this analysis and is available with the online version of this paper. Supplementary Material Additional data file 1 The original data used to perform this analysis Click here for additional data file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545601.xml

545198

Seasonal Patterns of Infectious Diseases

Why is that many infectious diseases, like cholera, malaria, and meningococcal meningitis, show seasonal patterns? And how can we accurately determine these patterns?

Meningococcal meningitis in western Africa shows recurrent seasonal patterns every year. Epidemics typically start at the beginning of February and last until May. We can try to explain the observed patterns on the basis of some seasonally varying environmental factor that favors disease transmission. Air dryness produced by strong dust winds is the most likely candidate. But while there are qualitative “stories” of this kind in the literature for many seasonal phenomena, convincing quantitative evidence to support them remains largely elusive. Instead, we tend to see weak associations between environmental and transmission variables when measured by simple, linear correlations. The study of meningococcal meningitis in Mali by Sultan and colleagues in this issue of PLoS Medicine is a remarkable exception [1] . The study reports a strong association between the yearly onset of epidemics and a large-scale regional index for atmospheric circulation related to the Harmattan winds in Sahelo-Sudanian Africa. The Importance of Seasonality Why is a focus on the seasonality of infectious diseases and its variation from year to year so important? Isn't it more important for us to instead understand the effects of long-term climate change on human health? At first sight, understanding seasonal patterns seems disconnected from understanding the impact of long-term climate change. However, seasonal patterns are one major pathway for the subtle but potentially drastic effects of climate change on disease dynamics. Long-term climate change affects seasonal patterns through the lengthening of the transmission season and the crossing of environmental and demographic thresholds that underlie seasonal outbreaks [2] . Thus, identifying the specific environmental factors underlying seasonal transmission is a critical step towards predicting and understanding how long-term environmental trends in mean climate and their variability will impact human health. The Problem of Scale One important difficulty in uncovering seasonal drivers of infectious diseases is to identify the appropriate scale of analysis. The relationship between disease and climate described by Sultan and colleagues only becomes apparent at large spatial scales. The authors argue that these large scales are necessary to eliminate “idiosyncratic” variability in the relationship between cases and climate at the local level. In other words, there are only weak correlations between seasonal variations and climate variables at small scales because of the multiple other factors that play a local role and act as noise. But we should be cautious about the suggestion that appropriate larger scales will always resolve the problem of local variability and present strong linear associations between climate and disease. Public health measures might require predictions not only at national and regional scales, but also at a variety of smaller scales. Moreover, one important source of variation in how infectious diseases respond to climate is the fraction of susceptible individuals in the population. This fraction varies over time as the result of immunity acquired by previous infection, and by the input of births and migrants into the pool of susceptible people. The constant waxing and waning of this pool of hosts underlies the intrinsic potential of the population dynamics of infectious diseases to oscillate and create epidemic outbreaks. The tendency of these intrinsic cycles to go up and down in synchrony at different locations in space will determine whether susceptibility levels act as noise at small scales or, alternatively, whether their effect must be considered in conjunction with climate at larger scales. Because the number of susceptible individuals is a hidden variable in most epidemiological analyses, recently proposed methods for its reconstruction from data on cases must be combined with studies on climate variation if we are to understand the interaction between susceptibility levels and climate variation [ 3 , 4 , 5 ]. The problem of scale also arises when we need to identify the appropriate timing (the temporal window) to detect strong associations between disease outbreaks and environmental covariates. This is particularly important when strong couplings between environment and transmission occur only transiently. This seems to be the case for cholera in Bangladesh, where couplings are strong during El Niños, but considerably weaker the rest of the time [6] . Intermittent couplings provide insight into how the system might behave if pushed into specific dynamic regions by a change in climate. Intermittent couplings also suggest the existence of thresholds in the response to climate, an area of research that remains in need of quantitative approaches. Seasonal Drivers May Be Elusive Besides scale, specific seasonal drivers are often elusive because of the simpler reason that in nature seasonality is ubiquitous. Multiple and covarying drivers have been proposed for the seasonal nature of cholera, including temperature, rainfall, and plankton blooms [7] . Yet the specific roles of these drivers in the bimodal seasonal cycle of cholera, and particularly in the second peak in endemic regions in south Asia, have not been convincingly shown ( Figure 1 ) [8] . We still don't have predictive explanations of the geographic variation in seasonal patterns. We won't find such explanations by considering the average seasonal pattern; instead, we must consider the anomalies in amplitude and onset of the peaks that occur in different years. Figure 1 The Role of Rainfall in Driving the Seasonal Nature of Cholera Is Unclear This photograph was taken during a cholera and nutrition survey during flooding in Bangladesh in 1974. In Bangladesh, monsoon rains appear to have a seasonal “dilution” effect on transmission, producing a decrease in cholera cases during that season. We don't know whether extreme rains also produce a lagged increase in cases later on in the cycle. In other parts of the world, cases typically peak during the rainy season. (Photo: Jack Weissman, Centers for Disease Control and Prevention) Ecologists have considered seasonality in mathematical models of the population dynamics of infectious diseases. Models of populations with seasonally forced, dynamic interactions (births, deaths, aggregation, or disease transmission) reveal an array of possible responses, from simple yearly cycles, through cycles that repeat with longer periods, to irregular chaotic fluctuations. Some models also predict intermittent switching between different dynamic infectious disease behaviors. But typical models consider only simple seasonal forcing functions (mathematical functions that are periodic in time and therefore describe in a generic way the seasonal variation in the transmission rate or some other seasonal parameter—a sine wave is an example). There are some important exceptions to this—some models do incorporate more complicated seasonal forcing functions that describe the actual processes underlying the seasonal drivers of transmission. Examples are models of childhood diseases that describe the regular stopping and starting of school terms [ 9 , 10 , 11 ], and recent malaria models that include the seasonal dynamics of mosquito births and pathogen incubation as functions of temperature and rainfall [12] . The explicit way in which models treat seasonal environmental drivers may be critical in addressing the links between within-year seasonal cycles and those of longer period that are observed in many infectious diseases. For meningococcal meningitis, we still need to examine the connection between the seasonal association described by Sultan and colleagues and the previously proposed role of humidity in inter-annual cycles [13] . The Complexity of Infectious Disease Dynamics Sultan and colleagues' study is exceptional in that it illustrates a clear relationship between an external environmental variable and the initiation of disease outbreaks. In contrast, many studies seeking environmental drivers are plagued by the many confounding factors, particularly the impact that other components of global change have on the transmission dynamics of infectious diseases. Thus, when we examine datasets for malaria, we must also consider the evolution of drug resistance and a growing human population that is increasingly forced to live in areas that are marginal for agricultural production but optimal for malaria transmission. Given this complexity, a serious limiting factor to quantitative analyses and predictive models of ecological and disease patterns is the lack of long-term disease records with similar data collected over a network of spatial locations. The handful of extremely valuable records that have allowed progress in understanding long-term patterns in disease dynamics pale in comparison to the spatiotemporal coverage available for climate studies and modeling. The need to resolve these issues of scale and confounding variability only underscores the urgency and importance of maintaining and developing systematic surveillance programs for infectious diseases around the world.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545198.xml

509423

Similar group mean scores, but large individual variations, in patient-relevant outcomes over 2 years in meniscectomized subjects with and without radiographic knee osteoarthritis

Background Epidemiological studies have, so far, identified factors associated with increased risk for incident or progressive OA, such as age, sex, heredity, obesity, and joint injury. There is, however, a paucity of long-term data that provide information on the nature of disease progression on either group or individual levels. Such information is needed for identification of study cohorts and planning of clinical trials. The aim of the study was, thus, to assess the variation in pain and function on group and individual level over 2 years in previously meniscectomized individuals with and without radiographic knee osteoarthritis (OA). Methods 143 individuals (16% women, mean age at first assessment 50 years [range 27–83]) were assessed twice; approximately 14 and 16 years after isolated meniscectomy, with a median interval of 2.3 years (range 2.3–3.0). Radiographic OA (as assessed at the time of second evaluation) was present in the operated knee in 40%, and an additional 19% had a single osteophyte grade 1 in one or both of the tibiofemoral compartments. Subjects completed the self-administered and disease-specific Knee injury and Osteoarthritis Outcome Score (KOOS). Results There were no significant changes in the group mean KOOS subscale scores over the 2-year period. However, a great variability over time was seen within individual subjects. Out of 143 subjects, 16% improved and 12% deteriorated in the subscale Pain, and 13% improved and 14% deteriorated in the subscale ADL ≥ 10 points (the suggested threshold for minimal perceptible clinical change). Similar results were seen for remaining subscales. Conclusion Group mean scores for this study cohort enriched in incipient and early-stage knee OA were similar over 2 years, but pain, function and quality of life changed considerably in individuals. These results may be valid also for other at risk groups with knee OA, and motivate further careful examination of the natural history of OA, as well as properties of the OA outcome instruments used. Longitudinal outcome data in OA studies need to be analyzed both on an individual and a group level.

Background Drugs that may slow or halt the breakdown of cartilage and other joint tissues in osteoarthritis (OA) and possibly improve symptoms and function are now being developed in the pharmaceutical industry. The potential availability of disease modifying OA drugs has focused attention on our relative lack of information on the 'natural disease history' of OA with regard to changes in symptoms, functional limitations, joint structure and other markers of disease change [ 1 ]. Epidemiological studies have identified factors associated with increased risk for incident or progressive OA, such as age, sex, heredity, obesity, and joint injury, pain, alignment, or laxity. There is, however, a paucity of long-term data that document the rate and nature of natural OA disease progression on either group or individual levels. Such information is needed for identification of study cohorts and planning of clinical trials of disease modifying OA drugs. Even more importantly, knowledge of natural disease progression in different patient groups will be needed to select those future groups that may benefit from such drugs. Only a few of the previously published studies have presented information on longitudinal variation in pain and function in the natural history of knee OA. The "Bristol 500 OA study" noted, that although pain changed little on a group level over a 3-year follow-up period, it varied greatly in individuals, with some subjects reporting marked improvements. Similarly, a minority improved functionally [ 2 - 4 ]. Yet another report suggested that most patients with OA attending rheumatology clinics do not deteriorate radiographically or symptomatically over an 11-year period [ 5 ]. A more recent report stated that 42–44% of community-recruited knee OA individuals did not change in physical functioning over a 3-year study [ 6 ]. Most investigations of the natural history of OA have been concerned with radiographic rather than clinical changes. For example, it was reported that the radiographic Kellgren and Lawrence classification score of 1 could represent incipient OA and be predictive of later development of more advanced radiographic features of OA [ 7 ]. MRI may be more responsive to change in early-stage OA than plain radiography [ 8 ]. However, outcome is usually heterogeneous: study subjects may report improvement or deterioration while they do not change radiographically over the time period assessed. It may also be that a few individuals alone generate much of any change detected at group level [ 9 - 11 ]. A further confounding factor in the longitudinal assessment of OA is the potential influence of the population from which the study group was recruited; a study group recruited from e.g. a specialist outpatient clinic is likely to have, on the average, more severe disease and may be at different risk to progress over time than a study group recruited from the community. The objective of this investigation was to assess both group and individual variation in knee pain, function and quality of life over two years in a study group enriched in incipient and early-stage radiographic knee OA. Methods Patients Approval was obtained from the Research Ethics Committee of the Medical Faculty of Lund University, Sweden. All patients who underwent meniscectomy between 1983 and 1985 were identified by searching the surgical records at the Department of Orthopedics, Lund University Hospital. In this period 552 meniscectomies were performed. Inclusion and exclusion criteria (Figure 1 ) were used to identify 264 former patients who, in 1998, were sent a self-administered questionnaire evaluating their knee-specific symptoms and knee function. Figure 1 Flow chart presenting the inclusion and exclusion criteria for patients. ACL = anterior cruciate ligament, PCL = posterior cruciate ligament, OA = osteoarthritis. Out of 211 individuals (80%) who returned the questionnaires, 6 were excluded because they matched one of the exclusion criteria. At 2 years after the first assessment 5 subjects had died, but the remaining 200 individuals were asked to provide a second evaluation using an identical questionnaire. Replies were received from 143 (72%). Of these 143 participants, 102 were meniscectomized by open surgery, and 41 by arthroscopy. Nineteen underwent an additional meniscus operation in the index knee. All re-operations were performed within 3 years after the original meniscectomy. Twenty-three participants were treated with subsequent meniscectomy of the contralateral knee. One of them underwent high tibial osteotomy and 1, because of OA, received a knee prosthesis in the contralateral knee. Data concerning subsequent surgeries were based on the medical records of Lund University Hospital and on self-reported information. Radiographic assessment At the time of the participants' second evaluation with questionnaires, standing anteroposterior (AP) radiographs of both knees were taken in 15 degrees of flexion using a CGR Phasix 60 generator at 70 kV, 16 mA, film-focus distance 1.5 m (CGR, Liège, Belgium). Ten out of the 143 participants (7%) declined the radiographic examination. All AP radiographs of the tibiofemoral joints from the follow-up were assessed for joint space narrowing (JSN) and osteophytes according to the atlas from Osteoarthritis Research Society International (OARSI) [ 12 ]. The presence of these features was graded on a 4-point scale (range 0–3, with 0 = no evidence of bony changes or JSN). We considered radiographic knee OA to be present if any of the following criteria was achieved in any of the 2 tibiofemoral compartments: JSN ≥ grade 2 or the sum of the 2 marginal osteophyte grades from the same compartment ≥ 2, or JSN grade 1 in combination with an osteophyte grade 1 in the same compartment [ 13 , 14 ]. This cut-off approximates grade 2 knee OA or worse based on the Kellgren and Lawrence scale [ 15 ]. Disease-specific questionnaire The Knee injury and Osteoarthritis Outcome Score (KOOS, Swedish version LK 1.0) is a 42-item self-administered knee-specific questionnaire based on the WOMAC Osteoarthritis Index [ 16 , 17 ]. KOOS was developed to be used for short- and long-term follow-up studies of knee injuries, and it comprises 5 subscales: Pain, Symptoms, Activities of Daily Living (ADL), Sports and Recreation Function (Sport/Rec) and knee-related Quality of Life (QOL). A separate score ranging from 0 to100, where 100 represents the best result, is calculated for each subscale. The questionnaire and scoring manual can be downloaded from . The KOOS is valid, reliable and responsive in follow-up of meniscectomy [ 17 ], anterior cruciate ligament reconstruction [ 18 ] and total knee replacement for OA [ 19 ]. The participants completed the KOOS questionnaire answering questions on their operated index knee. Change The minimal perceptible clinical improvement (MPCI) represents the difference on the measurement scale associated with the smallest change in the health status detectable by the individual. Since the KOOS questionnaire contains the full and original version of the WOMAC LK 3.0 index, we used the MPCI as described for WOMAC [ 20 ]. Thus, a level of 10 points or more of improvement or decline was operationally used as a cut-off representing a clinically perceptible difference. The sensitivity of the questionnaire has been established [ 21 ]. Data collection and statistics If questions were left unanswered in any part of the questionnaire, we returned the questionnaire to be completed. The questionnaires were then completed fully. The Mann-Whitney U-test was used to determine differences between the groups. P -values for categoric data were calculated with Fisher's exact test. All tests were 2-tailed and a P -value of ≤ 0.05 was considered statistically significant (SigmaStat, version 2.0, for Windows). Results Group level The study group comprised 143 individuals, of whom 23 (16%) were women. The participants' mean age at the first follow-up was 51 (range 27–83) years. The assessment was carried out twice: at approximately 14 and 16 years after the surgery, with a median interval of 2.3 (range 2.3 to 3.0) years. Fifty-three (40%) of the 133 individuals who had undergone radiographic examination had radiographic tibiofemoral OA in their index (operated) knee (21% women, age range 29–83, mean 53) and 80 were classified as non having OA (11% women, age range 27–82, mean 50). An additional 25 (19%) (not classified as radiographic OA) had a single osteophyte grade 1 in either one or both tibiofemoral compartments. Mean scores for the KOOS subscales at the first assessment did not change significantly over the 2-year study period (Table 1 ). Moreover, there were no significant changes in group mean subscale scores over 2 years when participants were divided into those with or without radiographic OA in the index knee (Table 1 , Figure 2 ). However, individuals with radiographic OA scored worse at both examinations than did those without radiographic OA. The differences between those with and without OA were statistically significant for KOOS Pain Δ = 11 points ( P = 0.004), other Symptoms Δ = 9 points ( P = 0.013), ADL Δ = 10 points ( P = 0.003), Sport/Rec Δ = 17 points ( P = 0.005), and QOL Δ = 16 points ( P = 0.003) assessed in 2000, and in the dimensions Sport/Rec Δ = 14 points ( P = 0.020) and QOL Δ = 12 points ( P = 0.041) evaluated in 1998. Table 1 KOOS scores overall and in patients without and with radiological signs of OA KOOS subscales Patients p-values Total group non-ROA ROA non-ROA vs. ROA n = 143 n = 80 n = 53 1998 2000 1998 2000 1998 2000 2000 pain mean 85 84 88 87 79 76 0.008 median 94 94 94 94 86 83 SD 20 21 16 18 24 25 range 19–100 25–100 39–100 25–100 19–100 25–100 symptoms mean 85 84 87 87 80 78 0.013 median 93 89 93 93 89 82 SD 19 18 17 16 23 21 range 14–100 14–100 25–100 18–100 14–100 14–100 ADL mean 88 88 90 91 83 81 0.004 median 99 97 99 99 94 90 SD 18 18 15 15 23 21 range 18–100 31–100 44–100 34–100 18–100 31–100 sports/rec mean 69 68 74 76 60 57 0.007 median 80 80 80 85 60 60 SD 31 32 28 28 34 34 range 0–100 0–100 0–100 0–100 0–100 0–100 QOL mean 75 73 78 78 67 63 0.005 median 81 81 81 84 69 63 SD 26 27 23 23 30 30 range 0–100 6–100 25–100 6–100 0–100 13–100 Mean, median, standard deviation and range of KOOS scores overall and in patients without and with radiological signs of OA. Note that 10 patients out of 143 did not undergo radiographic examination. P -values for comparison between KOOS subscale results in patients with and without OA in year 2000 are presented. Figure 2 Group mean KOOS scores for patients assessed in 1998 and 2000. Group mean KOOS scores for patients with (n = 53) and without (n = 80) radiographic osteoarthritis (ROA) assessed in 1998 and 2000. Possible score range 0 to 100, with 100 representing the best result. ADL – Activities of Daily Living, QOL – knee-related Quality of Life. Bars present ± 95% confidence intervals. The bars going upwards have wider caps. Note vertical axis break. We analyzed separately those subjects (N = 57) that did not participate in the second assessment. Their mean KOOS scores at the first examination did not differ significantly from the remainder of the study cohort, indicating little or no inclusion bias for the second follow-up (data not shown). The scores in the 5 patients that underwent additional surgery (e.g. osteotomy, knee arthroplasty) did not differ significantly from the rest of the group. Individual study subject changes In spite of the lack of change on a group level, we found substantial intra-individual variability in the questionnaire subscale scores measured 2 years apart. Out of the total 143 study subjects, 40 had either improved or deteriorated (n = 23 (16%) and n = 17 (12%), respectively) 10 points or more for the KOOS subscale Pain. Of the 23 subjects who had improved in their pain score by these criteria, 14 had also improved in the subscale Symptoms, 17 in ADL, 16 in Sports/Rec, and 17 in QOL. Only 1 of these subjects deteriorated in Symptoms, and 2 in Sports/Rec, none in the other subscales. Of the 17 subjects who deteriorated in Pain, 13 similarly deteriorated in Symptoms, 12 in ADL, 10 in Sports/Rec, and 10 in QOL. When evaluating those who had undergone radiographic examination, there were no significant differences in variability detected whether the subject had radiographic tibiofemoral OA or not ( P = 0.24, Table 2 ). Table 2 The percentage of patients improving, not changing, or deteriorating for KOOS subscales over time non-ROA ROA KOOS subscales cut-off n = 80 n = 53 + no change -- + no change -- % % pain 10 13 76 11 21 66 13 20 6 88 6 8 87 6 symptoms 10 16 69 15 26 55 19 20 6 86 8 13 77 9 ADL 10 9 79 13 19 64 17 20 5 86 9 15 79 6 sports/rec 10 19 60 21 28 42 30 20 11 76 13 21 64 15 QOL 10 20 56 24 26 57 17 20 5 88 8 15 75 9 The percentage of patients, with and without radiographic osteoarthritis (ROA), improving, not changing, or deteriorating for KOOS subscales over the 2 year study period. For definition of ROA see methods. Two cut-offs for change (≥ 10 and ≥ 20 points) are presented. We also evaluated a stricter cut-off of 20 points or more as used for the OARSI responder criteria, as opposed to minimal clinically perceptible change [ 22 ]. With this cut-off, in total 19 patients fulfilled the criterion for improvement or deterioration (n = 9 (6%), n = 10 (7%), respectively) in KOOS Pain. Among the subjects with radiographic OA, 3 (6%) improved and 4 (7%) deteriorated by 20 points or more. Corresponding numbers for those without radiographic OA were 5 (6%) for both improvement and deterioration. In order to explore these changes in more detail, the subjects were divided into quartiles, according to KOOS Pain score at the first assessment (Figure 3 ). The most noticeable changes were found in the quartile representing the worst scores: 21 of 36 (58%) subjects showed a change of 10 points or more in either direction. A corresponding change was seen in 11 (31%) individuals from the second worst quartile and in only 9 (25%) from the second best and best quartiles (6 and 3 subjects, respectively). Comparable results were seen for the other subscales of KOOS (data not shown). Figure 3 KOOS Pain subscale. Patients are divided into 4 subgroups (quartiles) according to the score at entry. Each line represents one patient visualizing the score in 1998 (left endpoint of line) and in 2000 (right endpoint of the same line). Discussion We found no significant change over 2 years in the average patient-relevant outcome scores for this study group of individuals who had undergone meniscectomy about 15 years earlier, even though the group was highly enriched in early-stage and incipient radiographic knee OA. However, we found substantial change in the self-report for individual subjects over the same time period. The generally worse KOOS scores for the individuals with radiographic knee OA, compared to those without, are consistent with earlier reports. Thus, the Baltimore Longitudinal Study of Aging reported that patients with a Kellgren-Lawrence score of 1 were almost twice as likely to report ever having knee joint pain compared with those who had a score of zero. The strength of the association increased with increasing Kellgren and Lawrence score [ 23 ]. Similarly, there was in meniscectomized individuals evidence for a graded increase in pain and functional limitations with increasing severity of radiographic signs of OA [ 24 ]. However, a discrepancy between knee pain and radiographic features of knee OA has also been noted, both cross-sectionally and longitudinally [ 3 , 24 , 25 ]. Depression and lack of muscle strength have been shown to better explain pain than radiographic findings [ 26 - 28 ]. Individual vs. group analysis Few reports have explored OA symptom variation on an individual level [ 2 - 4 ]. A detailed comparison of our results with earlier reports is difficult, since they were conducted before validated and patient-relevant OA disease-specific measurement tools had been widely introduced. The "Bristol OA 500" were patients with advanced radiographic knee OA and a mean age of 65 years recruited from a hospital based rheumatology clinic. In contrast, the mean age of the present study cohort was 50 years, with 2/5 having mild-moderate radiographic OA, and another 1/5 incipient radiographic changes. Further, the cohort reported on here was recruited from a group of individuals that had undergone isolated meniscectomy 15 years earlier, but independent of their subsequent symptom level or disease history. The mean scores of our study group were relatively good and not representative of subjects with advanced OA seeking medical care. The rationale for investigating this particular cohort at this time after surgery was its enrichment in early-stage knee OA, and that it consequently may represent a study group suitable for future pharmacological disease-modifying intervention. We assessed our patients at an interval of 2 years; this period of time being suggested as a minimum for clinical trials of disease modification in OA to detect both structural and symptom change [ 29 ]. It could be that the findings reported here are valid only for post-injury, secondary OA, or for this particular cohort. However, the criteria and delimitations for posttraumatic OA compared to primary OA have recently been shown to be much less clear than thought [ 13 , 14 , 30 ], and meniscal pathology is common also in primary, garden-variety, knee OA [ 31 ]. Tibiofemoral OA was observed in 53 out of 133 patients who were underwent radiographic examination. Isolated patellofemoral OA was rare and, since it did not affect the final results, was not taken into account. A further argument favoring the general applicability of the present results is the concordance of our findings with other longitudinal studies on OA [ 2 - 5 , 32 ]. Methodological issues We applied the criteria for minimal perceptible clinical improvement (MPCI) obtained for the WOMAC; since KOOS contains the WOMAC items and is similar in format. The KOOS subscale ADL is equivalent to the WOMAC subscale Function, while new items have been added to the KOOS subscales Pain and Symptoms. The dimensions assessed by the KOOS subscales Sport and Recreation Function and knee-related Quality of Life are not assessed by the WOMAC. The MPCI for the WOMAC is in the range of 8 to 12 points on a 0–100 scale [ 20 ]. This threshold coincides with the change observed in KOOS scores between 3 and 6 months postoperatively when assessing rehabilitation following reconstruction of the anterior cruciate ligament and concurs with the OARSI definition of moderate improvement in the knee pain assessment for clinical trials in OA [ 18 , 22 ]. However, the OARSI responder criteria were designed for the evaluation of the patient's response to oral NSAID and intra-articular treatment and may differ for other interventions. It may be argued that the subject-related changes observed in this study represent inherent instrument instability. However, validation studies of KOOS support the reproducibility and stability of the KOOS instrument [ 17 - 19 ]. Test-retest data on the KOOS subscale pain obtained from 75 patients about to undergo knee arthroscopy [ 17 ] was used to determine the number of subjects improving, deteriorating or not changing over an average period of 5 days. The proportion of subjects changing over 5 days was approximately half of that changing over 2 years in the present study, in further support that the variation in the present study cannot be explained solely by instrument noise (data not shown). A 'frame shift' in the priorities of the individual patient may occur during long term studies. However, we suggest that a significant frame shift is unlikely to have occurred over this 2 year study period of a cohort with a mean age of 50 years. Significant change of KOOS scores over time were noted in 1/3 of the cohort studied. About half of those who changed clinically improved. This was true in particular for patients with lower (worse) baseline scores. It is thus possible that the lower proportion of 'changers' among those with better baseline scores may have been, at least in part, due to a ceiling effect. Conclusions We conclude that despite unchanged group mean scores over 2 years, pain, function and quality of life change considerably over time in individuals, in this study cohort enriched in incipient and early-stage knee OA. These findings may be applicable also to other at risk patient groups in different phases of OA development, and motivate further careful examination of the natural history of OA, as well as properties of the OA outcome instrument used. We suggest that longitudinal OA study data should be analyzed both on the individual and group level. Our findings may have further relevance to clinical trials of OA that seek to document long term benefit in the form of symptom improvement and structural improvement. It is clear that much additional effort will need to be spent on selection of groups at high risk of progression of symptoms and structural joint change, and the identification of predictors for deterioration. Our results also suggest that the use of responder criteria may be an important aspect of analyzing the outcome of such trials [ 22 , 33 ]. Authors' contributions EMR and LSL planned study and collected the data. PTP performed the statistical analysis and drafted the manuscript. ME formed the database of patients and participated in the statistical analysis. ME, EMR, LSL corrected the manuscript. All authors read and approved the manuscript

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC509423.xml

545600

Bioconductor: open software development for computational biology and bioinformatics

A detailed description of the aims and methods of the Bioconductor project, an initiative for the collaborative creation of extensible software for computational biology and bioinformatics.

Background The Bioconductor project [ 1 ] is an initiative for the collaborative creation of extensible software for computational biology and bioinformatics (CBB). Biology, molecular biology in particular, is undergoing two related transformations. First, there is a growing awareness of the computational nature of many biological processes and that computational and statistical models can be used to great benefit. Second, developments in high-throughput data acquisition produce requirements for computational and statistical sophistication at each stage of the biological research pipeline. The main goal of the Bioconductor project is creation of a durable and flexible software development and deployment environment that meets these new conceptual, computational and inferential challenges. We strive to reduce barriers to entry to research in CBB. A key aim is simplification of the processes by which statistical researchers can explore and interact fruitfully with data resources and algorithms of CBB, and by which working biologists obtain access to and use of state-of-the-art statistical methods for accurate inference in CBB. Among the many challenges that arise for both statisticians and biologists are tasks of data acquisition, data management, data transformation, data modeling, combining different data sources, making use of evolving machine learning methods, and developing new modeling strategies suitable to CBB. We have emphasized transparency, reproducibility, and efficiency of development in our response to these challenges. Fundamental to all these tasks is the need for software; ideas alone cannot solve the substantial problems that arise. The primary motivations for an open-source computing environment for statistical genomics are transparency, pursuit of reproducibility and efficiency of development. Transparency High-throughput methodologies in CBB are extremely complex, and many steps are involved in the conversion of information from low-level information structures (for example, microarray scan images) to statistical databases of expression measures coupled with design and covariate data. It is not possible to say a priori how sensitive the ultimate analyses are to variations or errors in the many steps in the pipeline. Credible work in this domain requires exposure of the entire process. Pursuit of reproducibility Experimental protocols in molecular biology are fully published lists of ingredients and algorithms for creating specific substances or processes. Accuracy of an experimental claim can be checked by complete obedience to the protocol. This standard should be adopted for algorithmic work in CBB. Portable source code should accompany each published analysis, coupled with the data on which the analysis is based. Efficiency of development By development, we refer not only to the development of the specific computing resource but to the development of computing methods in CBB as a whole. Software and data resources in an open-source environment can be read by interested investigators, and can be modified and extended to achieve new functionalities. Novices can use the open sources as learning materials. This is particularly effective when good documentation protocols are established. The open-source approach thus aids in recruitment and training of future generations of scientists and software developers. The rest of this article is devoted to describing the computing science methodology underlying Bioconductor. The main sections detail design methods and specific coding and deployment approaches, describe specific unmet challenges and review limitations and future aims. We then consider a number of other open-source projects that provide software solutions for CBB and end with an example of how one might use Bioconductor software to analyze microarray data. Results and discussion Methodology The software development strategy we have adopted has several precedents. In the mid-1980s Richard Stallman started the Free Software Foundation and the GNU project [ 2 ] as an attempt to provide a free and open implementation of the Unix operating system. One of the major motivations for the project was the idea that for researchers in computational sciences "their creations/discoveries (software) should be available for everyone to test, justify, replicate and work on to boost further scientific innovation" [ 3 ]. Together with the Linux kernel, the GNU/Linux combination sparked the huge open-source movement we know today. Open-source software is no longer viewed with prejudice, it has been adopted by major information technology companies and has changed the way we think about computational sciences. A large body of literature exists on how to manage open-source software projects: see Hill [ 4 ] for a good introduction and a comprehensive bibliography. One of the key success factors of the Linux kernel is its modular design, which allows for independent and parallel development of code [ 5 ] in a virtual decentralized network [ 3 ]. Developers are not managed within the hierarchy of a company, but are directly responsible for parts of the project and interact directly (where necessary) to build a complex system [ 6 ]. Our organization and development model has attempted to follow these principles, as well as those that have evolved from the R project [ 7 , 8 ]. In this section, we review seven topics important to establishment of a scientific open source software project and discuss them from a CBB point of view: language selection, infrastructure resources, design strategies and commitments, distributed development and recruitment of developers, reuse of exogenous resources, publication and licensure of code, and documentation. Language selection CBB poses a wide range of challenges, and any software development project will need to consider which specific aspects it will address. For the Bioconductor project we wanted to focus initially on bioinformatics problems. In particular we were interested in data management and analysis problems associated with DNA microarrays. This orientation necessitated a programming environment that had good numerical capabilities, flexible visualization capabilities, access to databases and a wide range of statistical and mathematical algorithms. Our collective experience with R suggested that its range of well-implemented statistical and visualization tools would decrease development and distribution time for robust software for CBB. We also note that R is gaining widespread usage within the CBB community independently of the Bioconductor Project. Many other bioinformatics projects and researchers have found R to be a good language and toolset with which to work. Examples include the Spot system [ 9 ], MAANOVA [ 10 ] and dChip [ 11 ]. We now briefly enumerate features of the R software environment that are important motivations behind its selection. Prototyping capabilities R is a high-level interpreted language in which one can easily and quickly prototype new computational methods. These methods may not run quickly in the interpreted implementation, and those that are successful and that get widely used will often need to be re-implemented to run faster. This is often a good compromise; we can explore lots of concepts easily and put more effort into those that are successful. Packaging protocol The R environment includes a well established system for packaging together related software components and documentation. There is a great deal of support in the language for creating, testing, and distributing software in the form of 'packages'. Using a package system lets us develop different software modules and distribute them with clear notions of protocol compliance, test-based validation, version identification, and package interdependencies. The packaging system has been adopted by hundreds of developers around the world and lies at the heart of the Comprehensive R Archive Network, where several hundred independent but interoperable packages addressing a wide range of statistical analysis and visualization objectives may be downloaded as open source. Object-oriented programming support The complexity of problems in CBB is often translated into a need for many different software tools to attack a single problem. Thus, many software packages are used for a single analysis. To secure reliable package interoperability, we have adopted a formal object-oriented programming discipline, as encoded in the 'S4' system of formal classes and methods [ 12 ]. The Bioconductor project was an early adopter of the S4 discipline and was the motivation for a number of improvements (established by John Chambers) in object-oriented programming for R. WWW connectivity Access to data from on-line sources is an essential part of most CBB projects. R has a well developed and tested set of functions and packages that provide access to different databases and to web resources (via http, for example). There is also a package for dealing with XML [ 13 ], available from the Omegahat project, and an early version of a package for a SOAP client [ 14 ], SSOAP, also available from the Omegahat project. These are much in line with proposals made by Stein [ 15 ] and have aided our work towards creating an environment in which the user perceives tight integration of diverse data, annotation and analysis resources. Statistical simulation and modeling support Among the statistical and numerical algorithms provided by R are its random number generators and machine learning algorithms. These have been well tested and are known to be reliable. The Bioconductor Project has been able to adapt these to the requirements in CBB with minimal effort. It is also worth noting that a number of innovations and extensions based on work of researchers involved in the Bioconductor project have been flowing back to the authors of these packages. Visualization support Among the strengths of R are its data and model visualization capabilities. Like many other areas of R these capabilities are still evolving. We have been able to quickly develop plots to render genes at their chromosomal locations, a heatmap function, along with many other graphical tools. There are clear needs to make many of these plots interactive so that users can query them and navigate through them and our future plans involve such developments. Support for concurrent computation R has also been the basis for pathbreaking research in parallel statistical computing. Packages such as snow and rpvm simplify the development of portable interpreted code for computing on a Beowulf or similar computational cluster of workstations. These tools provide simple interfaces that allow for high-level experimentation in parallel computation by computing on functions and environments in concurrent R sessions on possibly heterogeneous machines. The snow package provides a higher level of abstraction that is independent of the communication technology such as the message-passing interface (MPI) [ 16 ] or the parallel virtual machine (PVM) [ 17 ]. Parallel random number generation [ 18 ], essential when distributing parts of stochastic simulations across a cluster, is managed by rsprng . Practical benefits and problems involved with programming parallel processes in R are described more fully in Rossini et al. [ 19 ] and Li and Rossini [ 20 ]. Community Perhaps the most important aspect of using R is its active user and developer communities. This is not a static language. R is undergoing major changes that focus on the changing technological landscape of scientific computing. Exposing biologists to these innovations and simultaneously exposing those involved in statistical computing to the needs of the CBB community has been very fruitful and we hope beneficial to both communities. Infrastructure base We began with the perspective that significant investment in software infrastructure would be necessary at the early stages. The first two years of the Bioconductor project have included significant effort in developing infrastructure in the form of reusable data structures and software/documentation modules (R packages). The focus on reusable software components is in sharp contrast to the one-off approach that is often adopted. In a one-off solution to a bioinformatics problem, code is written to obtain the answer to a given question. The code is not designed to work for variations on that question or to be adaptable for application to distinct questions, and may indeed only work on the specific dataset to which it was originally applied. A researcher who wishes to perform a kindred analysis must typically construct the tools from scratch. In this situation, the scientific standard of reproducibility of research is not met except via laborious reinvention. It is our hope that reuse, refinement and extension will become the primary software-related activities in bioinformatics. When reusable components are distributed on a sound platform, it becomes feasible to demand that a published novel analysis be accompanied by portable and open software tools that perform all the relevant calculations. This will facilitate direct reproducibility, and will increase the efficiency of research by making transparent the means to vary or extend the new computational method. Two examples of the software infrastructure concepts described here are the exprSet class of the Biobase package, and the various Bioconductor metadata packages, for example hgu95av2 . An exprSet is a data structure that binds together array-based expression measurements with covariate and administrative data for a collection of microarrays. Based on R data.frame and list structures, exprSets offer much convenience to programmers and analysts for gene filtering, constructing annotation-based subsets, and for other manipulations of microarray results. The exprSet design facilitates a three-tier architecture for providing analysis tools for new microarray platforms: low-level data are bridged to high-level analysis manipulations via the exprSet structure. The designer of low-level processing software can focus on the creation of an exprSet instance, and need not cater for any particular analysis data structure representation. The designer of analysis procedures can ignore low-level structures and processes, and operate directly on the exprSet representation. This design is responsible for the ease of interoperation of three key Bioconductor packages: affy , marray , and limma . The hgu95av2 package is one of a large collection of related packages that relate manufactured chip components to biological metadata concerning sequence, gene functionality, gene membership in pathways, and physical and administrative information about genes. The package includes a number of conventionally named hashed environments providing high-performance retrieval of metadata based on probe nomenclature, or retrieval of groups of probe names based on metadata specifications. Both types of information (metadata and probe name sets) can be used very fruitfully with exprSets : for example, a vector of probe names immediately serves to extract the expression values for the named probes, because the exprSet structure inherits the named extraction capacity of R data.frames . Design strategies and commitments Well-designed scientific software should reduce data complexity, ease access to modeling tools and support integrated access to diverse data resources at a variety of levels. Software infrastructure can form a basis for both good scientific practice (others should be able to easily replicate experimental results) and for innovation. The adoption of designing by contract, object-oriented programming, modularization, multiscale executable documentation, and automated resource distribution are some of the basic software engineering strategies employed by the Bioconductor Project. Designing by contract While we do not employ formal contracting methodologies (for example, Eiffel [ 21 ]) in our coding disciplines, the contracting metaphor is still useful in characterizing the approach to the creation of interoperable components in Bioconductor. As an example, consider the problem of facilitating analysis of expression data stored in a relational database, with the constraints that one wants to be able to work with the data as one would with any exprSet and one does not want to copy unneeded records into R at any time. Technically, data access could occur in various ways, using database connections, DCOM [ 22 ], communications or CORBA [ 23 ], to name but a few. In a designing by contract discipline, the provider of exprSet functionality must deliver a specified set of functionalities. Whatever object the provider's code returns, it must satisfy the exprSets contract. Among other things, this means that the object must respond to the application of functions exprs and pData with objects that satisfy the R matrix and data.frame contracts respectively. It follows that exprs ( x ) [ i,j ] , for example, will return the number encoding the expression level for the i th gene for the j th sample in the object x , no matter what the underlying representation of x . Here i and j need not denote numerical indices but can hold any vectors suitable for interrogating matrices via the square-bracket operator. Satisfaction of the contract obligations simplifies specification of analysis procedures, which can be written without any concern for the underlying representations for exprSet information. A basic theme in R development is simplifying the means by which developers can state, follow, and verify satisfaction of design contracts of this sort. Environment features that support convenient inheritance of behaviors between related classes with minimal recoding are at a premium in this discipline. Object-oriented programming There are various approaches to the object-oriented programming methodology. We have encouraged, but do not require, use of the so-called S4 system of formal classes and methods in Bioconductor software. The S4 object paradigm (defined primarily by Chambers [ 12 ] with modifications embodied in R) is similar to that of Common Lisp [ 24 ] and Dylan [ 25 ]. In this system, classes are defined to have specified structures (in terms of a set of typed 'slots') and inheritance relationships, and methods are defined both generically (to specify the basic contract and behavior) and specifically (to cater for objects of particular classes). Constraints can be given for objects intended to instantiate a given class, and objects can be checked for validity of contract satisfaction. The S4 system is a basic tool in carrying out the designing by contract discipline, and has proven quite effective. Modularization The notion that software should be designed as a system of interacting modules is fairly well established. Modularization can occur at various levels of system structure. We strive for modularization at the data structure, R function and R package levels. This means that data structures are designed to possess minimally sufficient content to have a meaningful role in efficient programming. The exprSet structure, for example, contains information on expression levels ( exprs slot), variability ( se.exprs ), covariate data ( phenoData slot), and several types of metadata (slots description , annotation and notes ). The tight binding of covariate data with expression data spares developers the need to track these two types of information separately. The exprSet structure explicitly excludes information on gene-related annotation (such as gene symbol or chromosome location) because these are potentially volatile and are not needed in many activities involving exprSets . Modularization at the R function level entails that functions are written to do one meaningful task and no more, and that documents (help pages) are available at the function level with worked examples. This simplifies debugging and testing. Modularization at the package level entails that all packages include sufficient functionality and documentation to be used and understood in isolation from most other packages. Exceptions are formally encoded in files distributed with the package. Multiscale and executable documentation Accurate and thorough documentation is fundamental to effective software development and use, and must be created and maintained in a uniform fashion to have the greatest impact. We inherit from R a powerful system for small-scale documentation and unit testing in the form of the executable example sections in function-oriented manual pages. We have also introduced a new concept of large-scale documentation with the vignette concept. Vignettes go beyond typical man page documentation, which generally focuses on documenting the behavior of a function or small group of functions. The purpose of a vignette is to describe in detail the processing steps required to perform a specific task, which generally involves multiple functions and may involve multiple packages. Users of a package have interactive access to all vignettes associated with that package. The Sweave system [ 26 ] was adopted for creating and processing vignettes. Once these have been written users can interact with them on different levels. The transformed documents are provided in Adobe's portable document format (PDF) and access to the code chunks from within R is available through various functions in the tools package. However, new users will need a simpler interface. Our first offering in this area is the vignette explorer vExplorer which provides a widget that can be used to navigate the various code chunks. Each chunk is associated with a button and the code is displayed in a window, within the widget. When the user clicks on the button the code is evaluated and the output presented in a second window. Other buttons provide other functionality, such as access to the PDF version of the document. We plan to extend this tool greatly in the coming years and to integrate it closely with research into reproducible research (see [ 27 ] for an illustration). Automated software distribution The modularity commitment imposes a cost on users who are accustomed to integrated 'end-to-end' environments. Users of Bioconductor need to be familiar with the existence and functionality of a large number of packages. To diminish this cost, we have extended the packaging infrastructure of R/CRAN to better support the deployment and management of packages at the user level. Automatic updating of packages when new versions are available and tools that obtain all package dependencies automatically are among the features provided as part of the reposTools package in Bioconductor. Note that new methods in R package design and distribution include the provision of MD5 checksums with all packages, to help with verification that package contents have not been altered in transit. In conclusion, these engineering commitments and developments have led to a reasonably harmonious set of tools for CBB. It is worth considering how the S language notion that 'everything is an object' impacts our approach. We have made use of this notion in our commitment to contracting and object-oriented programming, and in the automated distribution of resources, in which package catalogs and biological metadata are all straightforward R objects. Packages and documents are not yet treatable as R objects, and this leads to complications. We are actively studying methods for simplifying authoring and use of documentation in a multipackage environment with namespaces that allow symbol reuse, and for strengthening the connection between session image and package inventory in use, so that saved R images can be restored exactly to their functional state at session close. Distributed development and recruitment of developers Distributed development is the process by which individuals who are significantly geographically separated produce and extend a software project. This approach has been used by the R project for approximately 10 years. This was necessitated in this case by the fact no institution currently has sufficient numbers of researchers in this area to support a project of this magnitude. Distributed development facilitates the inclusion of a variety of viewpoints and experiences. Contributions from individuals outside the project led to the expansion of the core developer group. Membership in the core depends upon the willingness of the developer to adopt shared objectives and methods and to submerge personal objectives in preference to creation of software for the greater scientific community. Distributed development requires the use of tools and strategies that allow different programmers to work approximately simultaneously on the same components of the project. Among the more important requirements is for a shared code base (or archive) that all members of the project can access and modify together with some form of version management system. We adopted the Concurrent Versions System [ 28 , 29 ] and created a central archive, within this system, that all members of the team have access to. Additional discipline is needed to ensure that changes by one programmer should not result in a failure of other code in the system. Within the R language, software components are naturally broken into packages, with a formal protocol for package structure and content specified in the R Extensions manual [ 30 ]. Each package should represent a single coherent theme. By using well defined applications programming interfaces (APIs) developers of a package are free to modify their internal structures as long as they continue to provide the documented outputs. We rely on the testing mechanisms supported by the R package testing system [ 30 ] to ensure coherent, non-regressive development. Each developer is responsible for documenting all functions and for providing examples and possibly other scripts or sets of commands that test the code. Each developer is responsible for ensuring that all tests run successfully before committing changes back to the central archive. Thus, the person who knows the code best writes the test programs, but all are responsible for running them and ensuring that changes they have made do not affect the code of others. In some cases changes by one author will necessitate change in the code and tests of others. Under the system we are using these situations are detected and dealt with when they occur in development, reducing the frequency with which error reports come from the field. Members of the development team communicate via a private mailing list. In many cases they also use private email, telephone and meetings at conferences in order to engage in joint projects and to keep informed about the ideas of other members. Reuse of exogenous resources We now present three arguments in favor of using and adapting software from other projects rather than re-implementing or reinventing functionality. The first argument that we consider is that writing good software is a challenging problem and any re-implementation of existing algorithms should be avoided if possible. Standard tools and paradigms that have been proven and are well understood should be preferred over new untested approaches. All software contains bugs but well used and maintained software tends to contain fewer. The second argument is that CBB is an enormous field and that progress will require the coordinated efforts of many projects and software developers. Thus, we will require structured paradigms for accessing data and algorithms written in other languages and systems. The more structured and integrated this functionality, the easier it will be to use and hence the more it will be used. As specific examples we consider our recent development of tools for working with graph or network structures. There are three main packages in Bioconductor of interacting with graphs. They are graph , RBGL and Rgraphviz . The first of these provides the class descriptions and basic infrastructure for dealing with graphs in R, the second provides access to algorithms on graphs, and the third to a rich collection of graph layout algorithms. The graph package was written from scratch for this project, but the other two are interfaces to rich libraries of software routines that have been created by other software projects, BOOST [ 31 , 32 ] and Graphviz [ 23 ] respectively, both of which are very substantial projects with large code bases. We have no interest in replicating that work and will, wherever possible, simply access the functions and libraries produced by other projects. There are many benefits from this approach for us and for the other projects. For bioinformatics and computational biology we gain rapid access to a variety of graph algorithms including graph layout and traversal. The developers in those communities gain a new user base and a new set of problems that they can consider. Gaining a new user base is often very useful, as new users with previously unanticipated needs tend to expose weaknesses in design and implementation that more sophisticated or experienced users are often able to avoid. In a similar vein, we plan to develop and encourage collaboration with other projects, including those organized through the Open Bioinformatics Foundation and the International Interoperability Consortium. We have not specifically concentrated on collaboration to this point in part because we have chosen areas for development that do not overlap significantly with the tools provided by those projects. In this case our philosophy remains one of developing interfaces to the software provided by those projects and not re-implementing their work. In some cases, other projects have recognized the potential gains for collaboration and have started developing interfaces for us to their systems, with the intent of making future contributions [ 33 ]. Another argument in favor of standardization and reuse of existing tools is best made with reference to a specific example. Consider the topic of markup and markup languages. For any specific problem one could quickly devise a markup that is sufficient for that problem. So why then should we adopt a standard such as XML? Among the reasons for this choice is the availability of programmers conversant with the paradigm, and hence lower training costs. A second reason is that the XML community is growing and developing and we will get substantial technological improvements without having to initiate them. This is not unusual. Other areas of computational research are as vibrant as CBB and by coordinating and sharing ideas and innovations we simplify our own tasks while providing stimulus to these other areas. Publication and licensing of code Modern standards of scientific publication involve peer review and subsequent publication in a journal. Software publication is a slightly different process with limited involvement to date of formal peer review or official journal publication. We release software under an open-source license as our main method of publication. We do this in the hope that it will encourage reproducibility, extension and general adherence to the scientific method. This decision also ensures that the code is open to public scrutiny and comment. There are many other reasons for deciding to release software under an open-source license, some of which are listed in Table 1 . Another consideration that arose when determining the form of publication was the need to allow an evolutionary aspect to our own software. There are many reasons for adopting a strategy that would permit us to extend and improve our software offerings over time. The field of CBB is relatively volatile and as new technologies are developed new software and inferential methods are needed. Further, software technology itself is evolving. Thus, we wanted to have a publication strategy that could accommodate changes in software at a variety of levels. We hope that that strategy will also encourage our users to think of software technology as a dynamic field rather than a static one and to therefore be on the lookout for innovations in this arena as well as in more traditional biological ones. Our decision to release software in the form of R packages is an important part of this consideration. Packages are easy to distribute, they have version numbers and define an API. A coordinated release of all Bioconductor packages occurs twice every year. At any given time there is a release version of every package and a development version. The only changes allowed to be made on the release version are bug fixes and documentation improvements. This ensures that users will not encounter radical new behaviors in code obtained in the release version. All other changes such as enhancements or design changes are carried out on the development branch [ 34 ]. Approximately six weeks before a release, a major effort is taken to ensure that all packages on the development branch are coordinated and work well together. During that period extensive testing is carried out through peer review amongst the Bioconductor core. At release time all packages on the development branch that are included in the release change modes and are now released packages. Previous versions of these packages are deprecated in favor of the newly released versions. Simultaneously, a new development branch is made and the developers start to work on packages in the new branch. Note that these version-related administrative operations occur with little impact on developers. The release manager is responsible for package snapshot and file version modifications. The developers' source code base is fairly simple, and need not involve retention of multiple copies of any source code files, even though two versions are active at all times. We would also like to point out that there are compelling arguments that can be made in favor of choosing different paradigms for software development and deployment. We are not attempting at this juncture to convince others to distribute software in this way, but rather elucidating our views and the reasons that we made our choice. Under a different set of conditions, or with different goals, it is entirely likely that we would have chosen a different model. Special concerns We now consider four specific challenges that are raised by research in computational biology and bioinformatics: reproducibility, data evolution and complexity, training users, and responding to user needs. Reproducible research We would like to address the reproducibility of published work in CBB. Reproducibility is important in its own right, and is the standard for scientific discovery. Reproducibility is an important step in the process of incremental improvement or refinement. In most areas of science researchers continually improve and extend the results of others but for scientific computation this is generally the exception rather than the rule. Buckheit and Donoho [ 35 ], referring to the work and philosophy of Claerbout, state the following principle: "An article about computational science in a scientific publication is not the scholarship itself, it is merely advertising of the scholarship. The actual scholarship is the complete software development environment and that complete set of instructions that generated the figures." There are substantial benefits that will come from enabling authors to publish not just an advertisement of their work but rather the work itself. A paradigm that fundamentally shifts publication of computational science from an advertisement of scholarship to the scholarship itself will be a welcome addition. Some of the concepts and tools that can be used in this regard are contained in [ 36 , 37 ]. When attempting to re-implement computational methodology from a published description many difficulties are encountered. Schwab et al. [ 38 ] make the following points: "Indeed the problem occurs wherever traditional methods of scientific publication are used to describe computational research. In a traditional article the author merely outlines the relevant computations: the limitations of a paper medium prohibit complete documentation including experimental data, parameter values and the author's programs. Consequently, the reader has painfully to re-implement the author's work before verifying and utilizing it.... The reader must spend valuable time merely rediscovering minutiae, which the author was unable to communicate conveniently." The development of a system capable of supporting the convenient creation and distribution of reproducible research in CBB is a massive undertaking. Nevertheless, the Bioconductor project has adopted practices and standards that assist in partial achievement of reproducible CBB. Publication of the data from which articles are derived is becoming the norm in CBB. This practice provides one of the components needed for reproducible research - access to the data. The other major component that is needed is access to the software and the explicit set of instructions or commands that were used to transform the data to provide the outputs on which the conclusions of the paper rest. In this regard publishing in CBB has been less successful. It is easy to identify major publications in the most prestigious journals that provide sketchy or indecipherable characterizations of computational and inferential processes underlying basic conclusions. This problem could be eliminated if the data housed in public archives were accompanied by portable code and scripts that regenerate the article's figures and tables. The combination of R's well-established platform independence with Bioconductor's packaging and documentation standards leads to a system in which distribution of data with working code and scripts can achieve most of the requirements of reproducible and replayable research in CBB. The steps leading to the creation of a table or figure can be clearly exposed in an Sweave document. An R user can export the code for modification or replay with variations on parameter settings, to check robustness of the reported calculations or to explore alternative analysis concepts. Thus we believe that R and Bioconductor can provide a start along the path towards generally reproducible research in CBB. The infrastructure in R that is used to support replayability and remote robustness analysis could be implemented in other languages such as Perl [ 39 ] and Python [ 40 ]. All that is needed is some platform-independent format for binding together the data, software and scripts defining the analysis, and a document that can be rendered automatically to a conveniently readable account of the analysis steps and their outcomes. If the format is an R package, this package then constitutes a single distributable software element that embodies the computational science being published. This is precisely the compendium concept espoused in [ 36 ]. Dynamics of biological annotation Metadata are data about data and their definition depends on the perspective of the investigator. Metadata for one investigator may well be experimental data for another. There are two major challenges that we will consider. First is the evolutionary nature of the metadata. As new experiments are done and as our understanding of the biological processes involved increases the metadata changes and evolves. The second major problem that concerns metadata data is its complexity. We are trying to develop software tools that make it easier for data analysts and researchers to use the existing metadata appropriately. The constant changing and updating of the metadata suggests that we must have a system or a collection process that ensures that any metadata can be updated and the updates can be distributed. Users of our system will want access to the most recent versions. Our solution has been to place metadata into R packages. These packages are built using a semi-automatic process [ 41 ] and are distributed (and updated) using the package distribution tools developed in the reposTools package. There is a natural way to apply version numbers so users can determine if their data are up to date or if necessary they can obtain older versions to verify particular analyses. Further, users can synchronize a variety of metadata packages according to a common version of the data sources that they were constructed from. There are a number of advantages that come from automating the process of building data packages. First, the modules are uniform to an extent that would not be possible if the packages were human written. This means that users of this technology need only become acquainted with one package to be acquainted with all such packages. Second, we can create many packages very quickly. Hence the labor savings are substantial. For microarray analyses all data packages should have the same information (chromosomal location, gene ontology categories, and so on). The only difference between the packages is that each references only the specific set of genes (probes) that were assayed. This means that data analysts can easily switch from one type of chip to another. It also means that we can develop a single set of tools for manipulating the metadata and improvements in those tools are available to all users immediately. Users are free to extend data packages with data from other, potentially proprietary, sources. Treating the data in the same manner that we treat software has also had many advantages. On the server side we can use the same software distribution tools, indicating updates and improvements with version numbering. On the client side, the user does not need to learn about the storage or internal details of the data packages. They simply install them like other packages and then use them. One issue that often arises is whether one should simply rely on online sources for metadata. That is, given an identifier, the user can potentially obtain more up-to-date information by querying the appropriate databases. The data packages we are proposing cannot be as current. There are, however, some disadvantages to the approach of accessing all resources online. First, users are not always online, they are not always aware of all applicable information sources and the investment in person-time to obtain such information can be high. There are also issues of reproducibility that are intractable as the owners of the web resources are free to update and modify their offerings at will. Some, but not all, of these difficulties can be alleviated if the data are available in a web services format. Another argument that can be made in favor of our approach, in this context, is that it allows the person constructing the data packages to amalgamate disparate information from a number of sources. In building metadata packages for Bioconductor, we find that some data are available from different sources, and under those circumstances we look for consensus, if possible. The process is quite sophisticated and is detailed in the AnnBuilder package and paper [ 41 ]. Training Most of the projects in CBB require a combination of skills from biology, computer science, and statistics. Because the field is new and there has been little specialized training in this area it seems that there is some substantial benefit to be had from paying attention to training. From the perspective of the Bioconductor project, many of our potential users are unfamiliar with the R language and generally are scientifically more aligned with one discipline than all three. It is therefore important that we produce documentation for the software modules that is accessible to all. We have taken a two-pronged approach to this, we have developed substantial amounts of course material aimed at all the constituent disciplines and we have developed a system for interactive use of software and documentation in the form of vignettes and more generally in the form of navigable documents with dynamic content. Course materials have been developed and refined over the past two to three years. Several members of the Bioconductor development team have taught courses and subsequently refined the material, based on success and feedback. The materials developed are modular and are freely distributed, although restrictions on publication are made. The focus of the materials is the introduction and use of software developed as part of the Bioconductor project, but that is not a requirement and merely reflects our own specific purposes and goals. In this area we feel that we would benefit greatly from contributions from those with more experience in technical document authoring. There are likely to be strategies, concepts and methodologies that are standard practice in that domain that we are largely unaware of. However, in the short term, we rely on the students, our colleagues and the users of the Bioconductor system to guide us and we hope that many will contribute. Others can easily make substantial contributions, even those with little or no programming skills. What is required is domain knowledge in one field of interest and the recognition of a problem that requires additional domain knowledge from another of the fields of interest. Our experience has been that many of these new users often transform themselves into developers. Thus, our development of training materials and documentation needs to pay some attention to the needs of this group as well. There are many more software components than we can collectively produce. Attracting others to collaboratively write software is essential to success. Responding to user needs The success of any software project rests on its ability to both provide solutions to the problems it is addressing and to attract a user community. Perhaps the most effective way of addressing user needs is through an e-mail help list and one was set up as soon as the project became active. In addition it is important to keep a searchable archive available so that the system itself has a memory and new users can be referred there for answers to common questions. It is also important that members of the project deal with bug reports and feature requests through this public forum as it both broadcasts their intentions and provides a public record of the discussion. Our mailing list (mailto: bioconductor@stat.math.ethz.ch ) has been successful: there are approximately 800 subscribers and about 3,000 email messages per year. Attracting a user community itself requires a method of distributing the software and providing sufficient training materials to allow potential users to explore the system and determine whether it is sufficient for their purposes. An alternate approach would be to develop a graphical user interface (GUI) that made interactions with the system sufficiently self-explanatory that documentation was not needed. We note that this solution is generally more applicable to cases where the underlying software tasks are well defined and well known. In the present case, the software requirements (as well as the statistical and biological requirements) are constantly evolving. R is primarily command-line oriented and we have chosen to follow that paradigm at least for the first few years of development. We would of course welcome and collaborate with those whose goal was in GUI development but our own forays into this area are limited to the production of a handful of widgets that promote user interaction at specific points. Users have experienced difficulties downloading and installing both R and the Bioconductor modules. Some of these difficulties have been caused by the users' local environments (firewalls and a lack of direct access to the internet), and some by problems with our software (bugs) which arise in part because it is in general very difficult to adequately test software that interacts over the internet. We have, however, managed to help every user, who was willing to persist, get both R and Bioconductor properly installed. Another substantial difficulty that we had to overcome was to develop a system that allowed users to download not just the software package that they knew they wanted, but additionally, and at the same time, all other software packages that it relies on. With Bioconductor software there is a much larger inter-reliance on software packages (including those that provide machine learning, biological metadata and experimental data) than for most other uses of R and the R package system. The package, reposTools contains much of the necessary infrastructure for handling these tasks. It is a set of functions for dealing with R package repositories which are basically internet locations for collections of R packages. Once the basic software is installed, users will need access to documentation such as the training materials described above and other materials such as the vignettes, described in a previous section. Such materials are most valuable if the user can easily obtain and run the examples on their own computer. We note the obvious similarity with this problem and that described in the section on reproducible research. Again, we are in the enjoyable situation of having a paradigm and tools that can serve two purposes. Other open-source bioinformatics software projects The Open Bioinformatics Foundation supports projects similar to Bioconductor that are nominally rooted in specific programming languages. BioPerl [ 42 ], BioPython [ 43 ] and BioJava [ 44 ] are prominent examples of open-source language-based bioinformatics projects. The intentions and design methodologies of the BioPerl project have been lucidly described by Stajich and colleagues [ 45 ]. BioPerl In this section we consider commonalities and differences between BioPerl and Bioconductor. Both projects have commitments to open source distribution and to community-based development, with an identified core of developers performing primary design and maintenance tasks for the project. Both projects use object-oriented programming methodology, with the intention of abstracting key structural and functional features of computational workflows in bioinformatics and defining stable application programming interfaces (API) that hide implementation details from those who do not need to know them. The toolkits are based on highly portable programming languages. These languages have extensive software resources developed for non-bioinformatic purposes. The repositories for R (Comprehensive R Archive Network, CRAN) and Perl (Comprehensive Perl Archive Network, CPAN) provide mirrored WWW access to structured collections of software modules and documents for a wide variety of workflow elements. Development methodologies targeted at software reuse can realize large gains in productivity by establishing interfaces to existing CPAN or CRAN procedures instead of reimplementing such procedures. For reuse to succeed, the maintainer of the external resource must commit to stability of the resource API. Such stability tends to be the norm for widely-used modules. Finally, both languages have considerable interoperability infrastructure. One implication is that each project can use software written in unrelated languages. R has well-established interfaces to Perl, Python, Java and C. R's API allows software in R to be called from other languages, and the RSPerl package [ 46 ] facilitates direct calls to R from Perl. Thus there are many opportunities for symbiotic use of code by Bioconductor and BioPerl developers and users. The following script illustrates the use of BioPerl in R. > library(RSPerl) > .PerlPackage("Bio::Perl") > x <- .Perl("get_sequence", "swiss", "ROA1_HUMAN") > x$division() [1] "HUMAN" > x$accession() [1] "P09651" > unlist(x$get_keywords()) [1] "Nuclear protein" "RNA-binding" [3] "Repeat" "Ribonucleoprotein" [5] "Methylation" "Transport" ... The .PerlPackage command brings the BioPerl modules into scope. .Perl invokes the BioPerl get_sequence subroutine with arguments "swiss" and "ROA1_HUMAN". The resulting R object is a reference to a perl hash. RSPerl infrastructure permits interrogation of the hash via the $ operator. Note that RSPerl is not a Bioconductor-supported utility, and that installation of the BioPerl and RSPerl resources to allow interoperation can be complicated. Key differences between the Bioconductor and BioPerl projects concern scope, approaches to distribution, documentation and testing, and important details of object-oriented design. Scope BioPerl is clearly slanted towards processing of sequence data and interfacing to sequence databases, with support for sequence visualization and queries for external annotation. Bioconductor is slanted towards statistical analysis of microarray experiments, with major concerns for array preprocessing, quality control, within- and between-array normalization, binding of covariate and design data to expression data, and downstream inference on biological and clinical questions. Bioconductor has packages devoted to diverse microarray manufacturing and analysis paradigms and to other high-throughput assays of interest in computational biology, including serial analysis of gene expression (SAGE), array comparative genomic hybridization (arrayCGH), and proteomic time-of-flight (SELDI-TOF) data. We say the projects are 'slanted' towards these concerns because it is clear that both projects ultimately aim to support general research activities in computational biology. Distribution, documentation and testing BioPerl inherits the distribution paradigm supported by CPAN. Software modules can be acquired and installed interactively using, for example perl -MCPAN -e shell . This process supports automated retrieval of requested packages and dependencies, but is not triggered by runtime events. Bioconductor has extended the CRAN distribution functionalities so that packages can be obtained and installed 'just in time', as required by a computational request. For both Perl and R, software modules and packages are structured collections of files, some of which are source code, some of which are documents about the code. The relationship between documentation and testing is somewhat tighter in Bioconductor than in BioPerl. Manual pages and vignettes in Bioconductor include executable code. Failure of the code in a man page or vignette is a quality-control event; experimentation with executable code in manual pages (through the example function of R) is useful for learning about software behavior. In Perl, tests occupy separate programs and are not typically integrated with documentation. Details of object-oriented procedure Both R and Perl are extensible computer languages. Thus it is possible to introduce software infrastructure supporting different approaches to object-oriented programming (OOP) in various ways in both languages. R's core developers have provided two distinct approaches to OOP in R. These approaches are named S3 and S4. In S3, any object can be assigned to a class (or sequence of classes) simply by setting the class name as the value of the object's class attribute. Class hierarchies are defined implicitly at the object level. Generic methods are defined as ordinary functions and class-specific methods are dispatched according to the class of the object being passed as an argument. In S4, formal definition of class structure is supported, and class hierarchy is explicitly defined in class definitions [ 12 ]. Class instances are explicitly constructed and subject to validation at time of construction. Generic methods are non-standard R functions and metadata on generic methods is established at the package level. Specific methods are dispatched according to the class signature of the argument list (multiple dispatch). Overall, the OOP approach embodied in S4 is closer to Dylan or Scheme than to C++ or Java. Bioconductor does not require specific OOP methodology but encourages the use of S4, and core members have contributed special tools for the documentation and testing of S4 OOP methods in R. OOP methodology in Perl has a substantial history and is extensively employed in BioPerl. The basic approach to OOP in Perl seems to resemble S3 more than S4, in that Perl's bless operation can associate any perl data instance with any class. The CPAN Class::Multimethod module can be used to allow multiple dispatch behavior of generic subroutines. The specific classes of objects identified in BioPerl are targeted at sequence data (Seq, LocatableSeq, RelSegment are examples), location data (Simple, Split, Fuzzy), and an important class of objects called interface objects, which are classes whose names end in 'I'. These objects define what methods can be called on objects of specified classes, but do not implement any methods. BioJava, BioPython, GMOD and MOBY Other open bioinformatics projects have intentions and methods that are closely linked with those of Bioconductor. BioJava [ 44 ] provides Dazzle, a servlet framework supporting the Distributed Annotation System specification for sharing sequence data and metadata. Version 1.4 of the BioJava release includes java classes for general alphabets and symbol-list processing, tools for parsing outputs of blast-related analyses, and software for constructing and fitting hidden Markov models. In principle, any of these resources could be used for analysis in Bioconductor/R through the SJava interface [ 46 ]. BioPython [ 43 ] provides software for constructing python objects by parsing output of various alignment or clustering algorithms, and for a variety of downstream tasks including classification. BioPython also provides infrastructure for decomposition of parallelizable tasks into separable processes for computation on a cluster of workstations. The Generic Model Organism Database (GMOD) project targets construction of reusable components that can be used to reproduce successful creation of open and widely accessible databases of model organisms (for example, worm, fruitfly and yeast). The main tasks addressed are genome visualization and annotation, literature curation, biological ontology activities, gene expression analysis and pathway visualization and annotation. BioMOBY [ 47 ] provides a framework for developing and cataloging web services relevant to molecular biology and genomics. A basic aim is to provide a central registry of data, annotation or analysis services that can be used programmatically to publish and make use of data and annotation resources pertinent to a wide variety of biological contexts. As these diverse projects mature, particularly with regard to interoperability, we expect to add infrastructure to Bioconductor to simplify the use of these resources in the context of statistical data analysis. It is our hope that the R and Bioconductor commitments to interoperability make it feasible for developers in other languages to reuse statistical and visualization software already present and tested in R. Using Bioconductor (example) Results of the Bioconductor project include an extensive repository of software tools, documentation, short course materials, and biological annotation data at [ 1 ]. We describe the use of the software and annotation data by description of a concrete analysis of a microarray archive derived from a leukemia study. Acute lymphocytic leukemia (ALL) is a common and difficult-to-treat malignancy with substantial variability in therapeutic outcomes. Some ALL patients have clearly characterized chromosomal aberrations and the functional consequences of these aberrations are not fully understood. Bioconductor tools were used to develop a new characterization of the contrast in gene expression between ALL patients with two specific forms of chromosomal translocation. The most important tasks accomplished with Bioconductor employed simple-to-use tools for state-of-the-art normalization of hundreds of microarrays, clear schematization of normalized expression data bound to detailed covariate data, flexible approaches to gene and sample filtering to support drilling down to manageable and interpretable subsets, flexible visualization technologies for exploration and communication of genomic findings, and programmatic connection between expression platform metadata and biological annotation data supporting convenient functional interpretation. We will illustrate these through a transcript of the actual command/output sequence. More detailed versions of some of the processing and analysis activities sketched here can be found in the vignettes from the GOstats package. The dataset is from the Ritz laboratory at the Dana Farber Cancer Institute [ 48 ]. It contains data from 128 patients with ALL. Two subgroups are to be compared. The first group consists of patients with a translocation between chromosomes 4 and 11 (labeled ALL1/AF4). The second group consists of patients with a translocation between chromosomes 9 and 22 (labeled BCR/ABL). These conditions are mutually exclusive in this dataset. The Affymetrix HGu95Av2 platform was used, and expression measures were normalized using gcrma from the affy package. The output of this is an object of class exprSet which can be used as input for other functions. The package hgu95av2 provides biological metadata including mappings from the Affymetrix identifiers to GO, chromosomal location, and so on. These data can, of course be obtained from many other sources, but there are some advantages to having them as an R package. After loading the appropriate packages we first subset the ALL exprSet to extract those samples with the covariates of interest. The design of the exprSet class includes methods for subsetting both cases and probes. By using the square-bracket notation on ALL, we derive a new exprSet with data on only the desired patients. > data("ALL") > eset <- ALL[, ALL$mol %in% c("BCR/ABL", "ALL1/AF4")] Next we find genes which are differentially expressed between the ALL1/AF4 and BCR/ABL groups. We use the function lmFit from the limma package, which can assess differential expression between many different groups and conditions simultaneously. The function lmFit accepts a model matrix which describes the experimental design and produces an output object of class MArrayLM which stores the fitted model information for each gene. The fitted model object is further processed by the eBayes function to produce empirical Bayes test statistics for each gene, including moderated t -statistics, p -values and log-odds of differential expression. The log 2 -fold changes, average intensites and Holm-adjusted p -values are displayed for the top 10 genes (Figure 1 ). We select those genes that have adjusted p -values below 0.05. The default method of adjusting for multiple comparisons uses Holm's method to control the family-wise error rate. We could use a less conservative method such as the false discovery rate, and the multtest package offers other possibilities, but for this example we will use the very stringent Holm method to select a small number of genes. > selected <- p.adjust(fit$p.value[, 2]) < 0.05 > esetSel <- eset [selected, ] There are 165 genes selected for further analysis. A heat map produced by the heatmap function from R allows us to visualize the differential action of these genes between the two groups of patients. Note how the different software modules can be integrated to provide a very rich data-analysis environment. Figure 2 shows clearly that these two groups can be distinguished in terms of gene expression. We can carry out many other tests, for example, whether genes encoded on a particular chromosome (or perhaps on a specific strand of a chromosome) are over-represented amongst those selected by moderated t -test. Many of these questions are normally addressed in terms of a hypergeometric distribution, but they can also be thought of as two-way or multi-way tables, and alternate statistical tests (all readily available in R) can be applied to the resulting data. We turn our attention briefly to the use of the Gene Ontology (GO) annotation in conjunction with these data. We first identify the set of unique LocusLink identifiers among our selected Affymetrix probes. The function GOHyperG is found in the GOstats package. It carries out a hypergeometric test for an overabundance of genes in our selected list of genes for each term in the GO graph that is induced by these genes (Figure 3 ). The smallest p -value found was 1.1e-8 and it corresponds to the term, "MHC class II receptor activity". We see that six of the 12 genes with this GO annotation have been selected. Had we used a slightly less conservative gene selection method then the number of selected genes in this GO annotation would have been even higher. Reproducing the above results for any other species or chip for which an annotation package was available would require almost no changes to the code. The analyst need only substitute the references to the data package, hgu95av2 , with those for their array and the basic principles and code are unchanged. Similarly, substitution of other algorithms or statistical tests is possible as the data analyst has access to the full and complete source code. All tools are modifiable at the source level to suit local requirements. Conclusions We have detailed the approach to software development taken by the Bioconductor project. Bioconductor has been operational for about three years now and in that time it has become a prominent software project for CBB. We argue that the success of the project is due to many factors. These include the choice of R as the main development language, the adoption of standard practices of software design and a belief that the creation of software infrastructure is an important and essential component of a successful project of this size. The group dynamic has also been an important factor in the success of Bioconductor. A willingness to work together, to see that cooperation and coordination in software development yields substantial benefits for the developers and the users and encouraging others to join and contribute to the project are also major factors in our success. To date the project provides the following resources: an online repository for obtaining software, data and metadata, papers, and training materials; a development team that coordinates the discussion of software strategies and development; a user community that provides software testing, suggested improvements and self-help; more than 80 software packages, hundreds of metadata packages and a number of experimental data packages. At this point it is worth considering the future. While many of the packages we have developed have been aimed at particular problems, there have been others that were designed to support future developments. And that future seems very interesting. Many of the new problems we are encountering in CBB are not easily addressed by technology transfer, but rather require new statistical methods and software tools. We hope that we can encourage more statisticians to become involved in this area of research and to orient themselves and their research to the mixture of methodology and software development that is necessary in this field. In conclusion we would like to note that the Bioconductor Project has many developers, not all of whom are authors of this paper, and all have their own objectives and goals. The views presented here are not intended to be comprehensive nor prescriptive but rather to present our collective experiences and the authors' shared goals. In a very simplified version these can be summarized in the view that coordinated cooperative software development is the appropriate mechanism for fostering good research in CBB.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545600.xml

539292

Passive immunotherapy against Aβ in aged APP-transgenic mice reverses cognitive deficits and depletes parenchymal amyloid deposits in spite of increased vascular amyloid and microhemorrhage

Background Anti-Aβ immunotherapy in transgenic mice reduces both diffuse and compact amyloid deposits, improves memory function and clears early-stage phospho-tau aggregates. As most Alzheimer disease cases occur well past midlife, the current study examined adoptive transfer of anti-Aβ antibodies to 19- and 23-month old APP-transgenic mice. Methods We investigated the effects of weekly anti-Aβ antibody treatment on radial-arm water-maze performance, parenchymal and vascular amyloid loads, and the presence of microhemorrhage in the brain. 19-month-old mice were treated for 1, 2 or 3 months while 23-month-old mice were treated for 5 months. Only the 23-month-old mice were subject to radial-arm water-maze testing. Results After 3 months of weekly injections, this passive immunization protocol completely reversed learning and memory deficits in these mice, a benefit that was undiminished after 5 months of treatment. Dramatic reductions of diffuse Aβ immunostaining and parenchymal Congophilic amyloid deposits were observed after five months, indicating that even well-established amyloid deposits are susceptible to immunotherapy. However, cerebral amyloid angiopathy increased substantially with immunotherapy, and some deposits were associated with microhemorrhage. Reanalysis of results collected from an earlier time-course study demonstrated that these increases in vascular deposits were dependent on the duration of immunotherapy. Conclusions The cognitive benefits of passive immunotherapy persist in spite of the presence of vascular amyloid and small hemorrhages. These data suggest that clinical trials evaluating such treatments will require precautions to minimize potential adverse events associated with microhemorrhage.

Background Alzheimer's disease is characterized not only by the presence of parenchymal amyloid deposits and intracellular tangles but also by the presence of amyloid deposits in the vasculature, a condition referred to as cerebral amyloid angiopathy (CAA). The CAA observed in both Alzheimer's disease patients [ 1 ] and some of the transgenic mouse models [ 2 ] is primarily composed of the shorter form of amyloid beta (Aβ), Aβ 1–40 , while the majority of amyloid deposits in the parenchyma are composed of Aβ 1–42 , although the compact amyloid deposits also contain Aβ 1–40 . Anti-Aβ immunotherapy has been considered as a potential treatment for Alzheimer's disease for some time [ 3 , 4 ]. Active immunization with a vaccine including Aβ 1–42 fibrils progressed to human clinical trials where its administration was suspended due to meningoencephalitits in a subset of patients [ 5 ]. To date there have been pathology reports on two patients who participated in the trial and subsequently died [ 6 , 7 ]. Both reports note that while the numbers of parenchymal amyloid deposits appeared lower than expected in these cases, the CAA in these patients did not appear outside the normal range for Alzheimer's disease. In addition, one report mentioned multiple cortical hemorrhages and the presence of hemosiderin around the CAA vessels [ 7 ]. Given the adverse reactions to the active immunization, the irreversibility of such procedures and the variable antibody response to vaccines in older individuals [ 8 ], passive immunization against the Aβ peptide emerged as an alternative immunotherapeutic strategy. Studies in young and middle aged APP-transgenic mice have reported significant amyloid reductions with passive immunization [ 9 - 11 ]. Such treatments also demonstrate rapid improvements of memory function in APP-transgenic mice, sometimes without detectable reductions in amyloid [ 12 - 14 ]. Most recently, intracranial administration of anti-Aβ antibodies has been shown to not only remove Aβ but also clear, early-stage, hyperphosphorylated-tau aggregates [ 15 ]. Importantly, in the only prior study evaluating adoptive antibody transfer in older APP-transgenic mice, Pfeifer et al . [ 16 ] reported a doubling of cerebral microhemorrhages associated with significant reductions in amyloid burden after administration of an N-terminal specific anti-Aβ antibody. Materials and Methods Experiment design Mice derived from APP Tg2576 mice were obtained from our breeding program at University of South Florida started in 1996 [ 17 ]. For the 5-month treatment study, 13 APP-transgenic mice, aged 23 months, were assigned to one of two groups. The first group received weekly intraperitoneal anti-Aβ antibody injections (antibody 2286; mouse-monoclonal anti-human Aβ 28–40 IgG1; Rinat Neurosciences, Palo Alto, CA) for a period of 5 months ( n = 6). The second group received weekly intraperitoneal anti-AMN antibody (2906; mouse-monoclonal anti- Drosophila amnesiac protein IgG1; Rinat Neurosciences, Palo Alto, CA) injections for a period of 5 months ( n = 7). Seven nontransgenic mice were also assigned to one of two groups. The first group received weekly intraperitoneal anti-Aβ antibody injections for a period of 5 months ( n = 4). The second group received weekly intraperitoneal anti-AMN antibody injections for a period of 5 months ( n = 3). For the time course study of 1-, 2- or 3-month treatment, 22 APP-transgenic mice aged 19 months were assigned to one of four experimental groups, as described previously [ 14 ]. The first three groups received weekly intraperitoneal anti-Aβ antibody injections for 3 months, 2 months or 1 month, ending when all mice were 22 months of age. The fourth group received weekly intraperitoneal anti-AMN antibody injections for 3 months. Behavioral analysis Following 3 and 5 months of treatment, the mice from the 5-month study were subjected to a two-day radial-arm water-maze paradigm. The apparatus was a 6-arm maze as described previously [ 18 ]. On day one, 15 trials were run in three blocks of 5. A cohort of 4 mice were run sequentially for each block (i.e., each of 4 mice get trial one, then the same mice get trial two, etc.). After each 5-trial block, a second cohort of mice was run permitting an extended rest period before mice were exposed to the second block of 5 trials. The goal arm was different for each mouse in a cohort to minimize odor cues. The start arm was varied for each trial, with the goal arm remaining constant for a given individual for both days. For the first 11 trials, the platform was alternately visible then hidden (hidden for the last 4 trials). On day two, the mice were run in exactly the same manner as day one except that the platform was hidden forall trials. The number of errors (incorrect arm entries) was measured in a one-minute time frame. As standard practice, mice failing to make an arm choice in 20 seconds are assigned one error, but no mice in this study had to be assigned an error in this manner. The same individual administered the antibody treatments and placed mice in the radial-arm water maze. Due to the numbers of mice in the study the researcher was unaware of treatment group identity of each mouse. Also, the dependent measures in the radial-arm water-maze task are quantitative, not evaluative, so the potential for tester bias is reduced. In order to minimize the influence of individual trial variability, each mouse's errors for 3 consecutive trials were averaged producing 5 data points for each day, which were analyzed statistically by ANOVA using StatView (SAS Institute Inc., NC). Tissue preparation and histology On the day of sacrifice mice were weighed, overdosed with 100 mg/kg Nembutal (Abbott laboratories, North Chicago, IL), and then intracardially perfused with 25 mL of 0.9% sodium chloride. Brains were rapidly removed, and the left half of the brain was immersion fixed for 24 h in freshly prepared 4% paraformaldehyde in 100 mM KPO 4 (pH 7.2) for histopathology. The hemi-brains were then incubated for 24 h in 10%, 20% and 30% sucrose sequentially for cyroprotection. Horizontal sections of 25 μ thickness were collected using a sliding microtome and stored at 4°C in Dulbecco's phosphate-buffered saline with sodium azide (pH 7.2) to prevent microbial growth. A series of 8 equally spaced tissue sections 600 μ apart were randomly selected spanning the entire brain and stained using free-floating immunohistochemistry for total Aβ (rabbit polyclonal anti-pan Aβ; Biosource, Camarillo, CA, 1:10,000) as previously described [ 2 , 14 ]. A second series of tissue sections 600 μm apart were stained using 0.2% Congo red in NaCl-saturated 80% ethanol. Another set of sections were also mounted and stained for hemosiderin using 2% potassium ferrocyanide in 2% hydrochloric acid for 15 min, followed by a counterstain in a 1% neutral red solution for 10 min. Quantification of Congo red staining and Aβ immunohistochemistry was performed using the Image-Pro Plus (Media Cybernetics, Silver Spring, MD) to analyze the percent area occupied by positive stain. One region of the frontal cortex and three regions of the hippocampus were analyzed (to ensure that there was no regional bias in the hippocampal values). The initial analysis of Congo red was performed to give a total value. A second analysis was performed after manually editing out all of the parenchymal amyloid deposits to yield a percent area restricted to vascular Congo red staining. To estimate the parenchymal area of Congo red, we subtracted the vascular amyloid values from the total percentage. For the hemosiderin stain the numbers of Prussian blue-positive sites were counted on all sections and the average number of sites per section calculated. Looking at the sections at a low magnification we were able to observe a qualitative differences between animals; however, the percent area was so low that many fields contained no positive stain. Eight equally spaced sections were examined and the number of positive profiles was determined and averaged to a per-section value. To assess possible treatment-related differences, the values for each treatment group were analyzed by one-way ANOVA followed by Fisher's LSD means comparisons. Results Reversal of cognitive deficits by passive amyloid immunotherapy The radial-arm water-maze task detects spatial learning and memory deficits in transgenic mouse models [ 18 , 19 ]. We treated 23-month-old mice for 5 months with anti-Aβ antibody 2286 or control antibody 2906 (against a Drosophila -specific protein) and tested them for spatial navigation learning in a two-day version of the radial-arm water maze after 3 months of treatment and, using a new platform location, again after 5 months of treatment. At both testing times we found that APP-transgenic mice treated with the control antibody failed to learn platform location over two days of testing and were significantly impaired compared to the nontransgenic mice treated with either antibody (Fig. 1 ). However, APP-transgenic mice administered the anti-Aβ antibodies demonstrated a complete reversal of the impairment observed in the control-treated APP-transgenic mice, ending day two with a mean performance near 0.5 errors per trial (Fig. 1 ). Although learning at the later time point, when the mice were 28 months of age, may have been slightly slower for all groups, there was no impairment of the anti-Aβ antibody-treated APP. Figure 1 Spatial learning deficits in APP-transgenic mice were reversed following 3 and 5 months of immunization. Mice were tested in a two-day version of the radial-arm water maze. Solid lines represent APP-transgenic mice while dashed lines represent nontransgenic mice. Open symbols indicate anti-AMN, control-antibody treatment (○: APP-transgenic, control antibody; △: nontransgenic, control antibody) while closed symbols indicate anti-Aβ antibody treatment (●: APP-transgenic, Aβ antibody; ▲: nontransgenic, Aβ antibody). Panel A shows mean number of errors made over the two-day trial period following 3 months of immunization. Each data point is the average of 3 trials. Panel B shows the mean number of errors made over the 2-day trial period following 5 months of immunization. For both graphs * indicates p < 0.05, ** indicates p < 0.001 when the APP-transgenic mice receiving control antibody are compared with the remaining groups. Passive amyloid immunotherapy clears parenchymal Aβ deposits, but increases vascular amyloid In a prior experiment examining the effects of passive anti-Aβ immunotherapy for 1, 2 or 3 months in APP-transgenic mice killed at 21 months of age [ 14 ], we found a time-dependent reduction of both Aβ immunostaining of diffuse and fibrillar deposits and Congo-red staining of fibrillar amyloid deposits. In the current study we found a similar reduction in both Aβ immunostaining (Table 1 ) and total Congo-red staining (Fig. 2A , left panel; p < 0.001 frontal cortex and p < 0.01 hippocampus) after 5 months of immunotherapy. We noted that the bulk of what remained was vascular amyloid. We then separately analyzed vascular and parenchymal deposits which revealed a near 90% reduction in parenchymal deposits ( p < 0.001) but a 3–4 fold elevation of vascular Congo-red staining ( p < 0.0001; Fig. 2A , center and right panels, respectively). We also separately analyzed vascular and parenchymal Congo-red staining on mice from our earlier study [ 14 ], treated passively for 1, 2 or 3 months with anti-Aβ or control antibody, and found a similar result. There was a graded reduction in overall Congo-red staining nearing 75% as duration of antibody exposure increased (as reported previously; Fig. 2B ). However, when separated into vascular Congo-red deposits and parenchymal deposits, there was an antibody-exposure-time-dependent increase in vascular deposition in both hippocampus and frontal cortex (Fig. 2C ; p < 0.05 frontal cortex and hippocampus) and a corresponding nearly 90% decrease in parenchymal deposits (Fig. 2D ; p < 0.001 in frontal cortex and hippocampus). Table 1 Total Aβ load is significantly reduced following 5 months of anti-Aβ antibody treatment. Percent area occupied by positive immunohistochemical stain for Aβ is shown ± standard error of the mean for both the frontal cortex and hippocampus. Also shown is the percent reduction of Aβ observed following anti-Aβ antibody treatment Region % area positive for Aβ: control treated % area positive for Aβ: anti-Aβ treated % reduction following anti-Aβ antibody treatment Frontal Cortex 34.855 ± 2.265 9.681 ± 0.754 72 Hippocampus 23.994 ± 0.985 8.212 ± 0.596 66 Figure 2 Passive immunization with anti-Aβ antibodies decreases total and parenchymal amyloid loads while increasing vascular amyloid in frontal cortex and hippocampus of APP-transgenic mice. Panel A shows total amyloid load measured with Congo red, vascular amyloid load and parenchymal amyloid load from APP-transgenic mice administered control IgG (C) or anti-Aβ IgG (Aβ) for a period of 5 months. Panels B-D show total amyloid load (Panel B), vascular amyloid load (Panel C) and parenchymal amyloid load (Panel D) from APP-transgenic mice administered control IgG for 3 months (Cont IgG) or anti-Aβ IgG for a period of 1, 2, or 3 months (Anti-Aβ IgG). For all panels, the solid bar and solid line represent values from the frontal cortex, while the open bar and dashed line represent values from the hippocampus. ** p < 0.01. These differences were readily observed examining micrographs of sections from these mice. Mice treated with control antibodies revealed occasional cortical vascular amyloid deposits (22 months, Fig. 3A , 28 months, Fig. 3C ), while mice administered anti-Aβ antibodies had increased amounts of vascular amyloid staining (3-month treatment, Fig 3B ; 5-month treatment, Fig 3D ). Those vessels containing amyloid following treatment with anti-Aβ antibody also exhibited apparent increases in microglial activation as measured by CD45 expression (Fig. 3F ) compared to mice treated with control antibody (Fig. 3E ). Unfortunately, the shifting numbers and sizes of vascular and parenchymal deposits caused by the antibody therapy greatly complicated measurement of microglial activation per vascular deposit area so that this apparent increase in staining intensity could not be quantified accurately. Figure 3 Increased Congo red staining of blood vessels following anti-Aβ antibody administration is associated with activated microglia. Panels A and B are from the frontal cortex of 22-month-old APP-transgenic mice immunized for 3 months with either control antibody (3A) or anti-Aβ antibody (3B). Panels C and D are from the frontal cortex of 28-month-old APP-transgenic mice immunized for 5 months with either control antibody (3C) or anti-Aβ antibody (3D). Panels E and F show a high-magnification image of CD45 immunohistochemistry (black) counterstained with Congo red (red) from 28-month-old APP-transgenic mice immunized for 5 months with either control antibody (Panel E) or anti-Aβ antibody (Panel F). Panels A-D, magnification = 100X. Scale bar in Panel B = 50 μ for panels A-D. Panels E-F, magnification = 200X. Scale bar in Panel E = 25 μm for panels E-F. Passive amyloid immunotherapy causes increased microhemorrhage We used the Prussian blue histological stain to label hemosiderin, a ferric oxide material produced in the breakdown of hemoglobin. Extravenous blood in the brain leads to microglial phagocytosis of the erythrocytes and breakdown of the hemoglobin within them. These ferric oxide-containing microglia are thus markers of past hemorrhage. In untreated, aged APP-transgenic mice we observed very few profiles positive for Prussian-blue staining in the frontal cortex (section counterstained with neutral red; Fig. 4A ). However, following anti-Aβ antibody treatment for 5 months we observed an increase in the number of Prussian-blue profiles in the frontal cortex, which were readily detectable at a low magnification in the microscope (Fig. 4B ). In the absence of anti-Aβ treatment, or even when treated with antibody for one month, most vessels did not stain with Prussian blue, and could be identified only using the red counterstain (Fig. 4C ). However, even with 3 months of anti-Aβ antibody treatment we observed frequent vessels with associated Prussian-blue staining (Fig 4D ). Using adjacent sections stained for Congo red, we confirmed that all vessels showing microhemorrhage contained amyloid (Figs. 4E and 4F ; we were unable to double-label Prussian blue-stained sections with either Congo red or thioflavine-S). However, only a minority of vessels containing amyloid demonstrated hemorrhage. Figure 4 Microhemorrhage associated with CAA following systemic administration of anti-Aβ antibodies. Panels A and B are low magnification images of the frontal cortex of APP-transgenic mice receiving either control antibodies (Panel A) or anti-Aβ antibodies (Panel B) for a period of 5 months. Panels C and D show representative images of amyloid containing vessels stained for Prussian blue (blue), counterstained with neutral red (red), from APP-transgenic mice receiving either control antibodies (Panel C) or anti-Aβ antibodies (Panel D) for a period of 3 months. Panel E shows a blood vessel in the frontal cortex stained for Prussian blue (blue), counterstained with neutral red, from an APP transgenic mouse administered anti-Aβ antibodies for 5 months. Panel F shows the same blood vessel on an adjacent section stained for Congo red, indicating that the blood vessel does in fact contain amyloid. Scale bar panel A = 120 μm for panels A-B. Scale bar panel C = 25 μm for panels C-D. Scale bar in panel F = 25 μm for panels E-F. When we counted the number of Prussian blue-positive profiles in those animals receiving control antibody there was an average of one profile per every two sections (Fig. 5 ) and this number remained the same in both control groups (aged 22 or 28 months). Following treatment with anti-Aβ antibody for a period of two months we observed a striking increase in Prussian-blue staining, approximately five times that observed in either the control group or the mice immunized for one month (Fig. 5 , p < 0.001). Following this initial increase in Prussian-blue staining, we observed a linear increase in staining associated with increasing duration of anti-Aβ antibody treatment (Fig 5 ). Five months of anti-Aβ antibody treatment demonstrated a six-fold increase in Prussian-blue staining when compared the control groups (Fig. 5 ). Figure 5 Number of Prussian blue-positive profiles increases with duration of anti-Aβ antibody exposure. The graph shows quantification of the average number of Prussian blue-positive profiles per section from mice administered control IgG for 3 or 5 months (Cont) or anti-Aβ IgG for 1, 2, 3 or 5 months (anti-Aβ). ** p < 0.01. Discussion Earlier studies with vaccines against the Aβ peptide demonstrated protection from the learning and memory deficits associated with amyloid accumulation in APP-transgenic mice [ 14 , 19 ]. Passive immunization protocols with anti-Aβ antibodies also produced cognitive benefits, in some cases even in the absence of significant reduction in amyloid burden [ 12 , 13 ]. Our recent work found that 3 months of anti-Aβ treatment of 18-month-old APP-transgenic mice improved spontaneous alternation performance on the Y-maze [ 14 ]. In the present work we confirmed that passive anti-amyloid immunotherapy can reverse spatial learning deficits in APP-transgenic mice and that this benefit of immunotherapy is retained, even in aged mice (26 and 28 months old at testing) with long-established amyloid pathology. Additionally, we describe a more rapid means of testing spatial reference memory to reveal learning and memory deficits in APP-transgenic mice. This two-day version of the radial arm water maze included greater spacing of individual trials (mice spent time in their home cage after every trial), combined with less spacing of aggregate trials (fifteen trials per day rather than four or five) to facilitate learning of platform location in the nontransgenic mice, with a clear absence of learning in the age-matched transgenic mice. A substantial reduction in total Congophilic amyloid deposits was observed in old APP-transgenic mice treated with anti-Aβ antibodies for 2 or more months. This measurement of total Congo-red staining included both parenchymal and vascular amyloid staining. When we analyzed the sections for only vascular amyloid (CAA) we found that this measure was significantly increased following 2, 3 and 5 months of anti-Aβ antibody treatment. The remaining parenchymal amyloid load was almost completely eliminated with this antibody approach. Clearly, because total amyloid load was significantly reduced not all amyloid was shifted into the vessels; but, it appears that at least some of the Congophilic material was redistributed to the vasculature. At the present time the mechanism for this redistribution is unclear. However, one possibility is that the microglia associated with the antibody-opsonized amyloid, either by phagocytosis or surface binding, and transported the material to the vasculature, possibly in an attempt to expel it. We and others have shown evidence for microglial involvement in the removal of amyloid using both intracranial anti-Aβ antibody injections [ 11 , 21 ] and systemically administered anti-Aβ antibody treatment [ 14 ], as well as ex vivo studies [ 10 , 22 ]. Here we also report our impression that microglia surrounding CAA vessels in immunized mice expressed more CD45 than control transgenic mice. This increased expression could be due to either increased expression in the same number of microglial cells or an increased number of microglial cells in these animals. It is feasible that this microglial activation was simply in reaction to the presence of increased amyloid in the blood vessels. However, it is equally likely that microglia activated by the opsonized material migrated to the vessels for disposal of the amyloid. Cerebral amyloid angiopathy (CAA) is defined as the deposition of congophilic material in meningeal and cerebral arteries and arterioles (capillaries and veins can also show CAA but less frequently), and it occurs to some extent in nearly all Alzheimer's disease patients [ 23 ]. Severe CAA, affecting about 15% of cases, can be associated with both infarction and hemorrhagic injury [ 24 , 25 ]. It has also been shown that the severity of CAA can be directly linked to the severity of dementia in Alzheimer's disease patients [ 26 ]. In the current study we found a significantly increased number of microhemorrhages in the brain as detected by Prussian-blue staining, associated with the increase in CAA following passive immunization. Another transgenic mouse model of amyloid deposition, the APP23 mice, have been shown to deposit amyloid in both brain parenchyma and blood vessels and show a CAA associated increase in spontaneous cerebral hemorrhages [ 27 ]. Moreover, Pfeifer et al . [ 16 ] showed that these spontaneous hemorrhages were significantly increased following 5 months of passive immunization of 21-month-old APP23 mice using an anti-Aβ antibody with an N-terminal epitope, similar to those typically developed in active immunization with vaccines [ 4 , 28 , 29 ]. When young mice (6 months of age) were immunized following the same protocol, no hemorrhages were observed. More recently, DeMattos et al . [ 30 ] showed that passive immunization with an N-terminal antibody (3D6: directed against amino acids 1–5 of Aβ) of PDAPP transgenic mice also resulted in significantly increased microhemorrhage. They were unable to detect increased microhemorrhage with a mid-domain antibody (266: directed against amino acids 13–28 of Aβ). Notably, antibody 266 fails to bind Aβ deposited in CAA vessels or amyloid plaques [ 31 ]. Importantly, Ferrer et al . [ 7 ] noted the presence of CAA and microhemorrhage in the brain of one patient that participated in the Aβ-vaccine trial, even though the parenchymal amyloid appeared lower than expected. Also, Nicoll et al . [ 6 ] noted that CAA appeared unaffected in the brain of another patient that participated in the Aβ-vaccine trial. It remains to be determined whether these observations regarding increased CAA and microhemorrhage in transgenic mice are relevant to trials of passive immunotherapy in humans. It should be noted that, in spite of extending the period of immunotherapy to 5 months, there was no discernable loss of the cognitive benefits of immunotherapy in the transgenic mice, all of whom showed increased microhemorrhage. While the observation that antibody 266 does not result in vascular leakage encourages testing of this idiotype, data from the Zurich cohort of the Aβ vaccine trial argue that brain-reactive antibodies may be important for cognitive benefits [ 32 ]. Conclusions Our opinion is that these results suggest that passive immunotherapy against Aβ should proceed with appropriate precautions taken to minimize the risk of hemorrhage (e.g., by excluding patients taking anticoagulants) and instituting measures to detect such hemorrhages if they do occur, irrespective of the antibody specificity or proclivity for microhemorrhage in aged APP-transgenic mice. List of abbreviations Aβ : Amyloid-beta. APP: Amyloid precursor protein CAA: Cerebral amyloid angiopathy. IgG1: Immunoglobulin G type 1. Competing interests The authors declare that they have no competing interests. Authors' contributions DMW treated the mice, performed the behavioral analysis, processed the tissue and performed pathological analyses, and drafted the manuscript. ARojiani evaluated slides and provided expert opinion regarding CAA and microhemorrhage. ARosenthal and SS developed, produced and purified the antibodies used in the studies. MJF performed DNA extraction and PCR for genotyping of the mice. MNG oversees the breeding colony generating mice for the studies, collected samples from the mice and assisted in editing the manuscript. DM conceived the design of the study, guided data interpretation and assisted in editing the manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539292.xml

526371

An unusual case of an ulcerative colitis flare resulting in disseminated intravascular coagulopathy and a bladder hematoma: a case report

Background Disorders of coagulation have long been associated with inflammatory bowel disease. Children, as well as adults, with both active and inactive ulcerative colitis have been found to have abnormal coagulation and fibrinolysis. Disseminated intravascular coagulation arises from an overwhelming of the haemostatic regulatory mechanisms leading to an excessive generation of thrombin and a failure of the normal inhibitory pathways to prevent systemic effects of this enzyme. Ulcerative colitis has been associated with disseminated intravascular coagulation in conjunction with septicemia, toxic megacolon and surgery. Case presentation A fourteen-year-old boy with a history of poorly controlled ulcerative colitis presented with nonbilious emesis, hematochezia, and hematuria. Laboratory workup revealed disseminated intravascular coagulation. He was placed on triple antibiotics therapy. An infectious workup came back negative. A computerized tomography (CT) scan of the abdomen revealed a marked thickening and irregularity of the bladder wall as well as wall thickening of the rectosigmoid, ascending, transverse, and descending colon. Patient's clinical status remained stable despite a worsening of laboratory values associated with disseminated intravascular coagulation. Patient was begun on high dose intravenous steroids with improvement of the disseminated intravascular coagulation laboratory values within 12 hours and resolution of disseminated intravascular coagulopathy within 4 days. A thorough infectious workup revealed no other causes to his disseminated intravascular coagulation. Conclusions The spectrum of hypercoagulable states associated with ulcerative colitis varies from mild to severe. Although disseminated intravascular coagulation associated with ulcerative colitis is usually related to septicemia, toxic megacolon or surgery, we present a case of an ulcerative colitis flare resulting in disseminated intravascular coagulation and a bladder hematoma.

Background A wide variety of disorders are associated with the development of disseminated intravascular coagulation (DIC). Initiation usually involves mechanical tissue injury and or endothelial cell activation and injury. DIC arises from an overwhelming of the haemostatic regulatory mechanisms leading to an excessive generation of thrombin and a failure of the normal inhibitory pathways to prevent systemic effects of the enzyme leading to DIC [ 1 ]. Ulcerative colitis has been associated with DIC. In previously reported cases, DIC has arisen from active disease in conjunction with septicemia, toxic megacolon or surgery [ 2 - 5 ]. The authors report a pediatric case of DIC associated with a colitis flare resulting in a bladder hematoma. Case presentation A 14-year-old boy with a diagnosis of ulcerative colitis based on colonic histology, serology and a normal barium study of his small bowels was admitted with a five-day history of nonbilious vomiting and bloody diarrhea. Additional symptoms included recent onset hematuria, and low-grade fevers to 100.4 C over the prior four days. He had also sustained a 25 lb weight loss in the last six months, indicating a lack of disease control. As an outpatient, his maintenance therapy included mesalamine (1 gram three times a day), and mercaptopurine (75 mg once per day). In addition, he had been started on prednisone approximately 7 weeks prior for treatment of an ulcerative colitis flare. His current dose of prednisone was 10 mg once a day. Soon after symptoms begun, he had been placed on ciprofloxacin as treatment for a presumptive flare. Physical exam showed he was afebrile, with a heart rate of 130 beats per minute, respiratory 16 breaths per minute and blood pressure 115/67 mmHg. He was alert although with a sallow appearance. Abdominal exam revealed a soft nontender nondistended abdomen. Rectal showed normal external exam with grossly bloody stool. Initial blood work showed hemoglobin of 12.3, a normal white blood cell count, normal differential and normal platelet count with a mildly elevated prothrombin time of 16.2 with an international normalized ratio (INR) of 1.2. Urine analysis showed a specific gravity of 1.035, 3+blood, +ketones and > 100 RBC per high powered field and 0–5 WBC per high power field. Abdominal ultrasound revealed irregular shaped bladder wall. Patient was placed on intravenous fluids (IV) as well as metronidazole (IV). Blood and urine cultures were sent for analysis. Stool was sent for culture and for Clostridium difficile toxin analysis. Serial repeat lab works the following day revealed a dropping hemoglobin (7.4 g/dL) and platelet count (64 K/mm 3 ) increasing PT/PTT (21.3/47 seconds) with an INR of 1.8. Blood smear showed moderate amount of elliptocytes, schistocytes, microcytes and fragmented red blood cells. Initial DIC panel revealed an elevated D-dimer of 4.9 mcg/mL with a normal thrombin time and fibrinogen. Thrombin time subsequently increased to > 120 seconds. D-dimers increased to 10.3 mcg/mL. A computerized tomography (CT) scan of the abdomen revealed a marked thickening and irregularity of the bladder wall as well as wall thickening of the rectosigmoid, ascending, transverse, and descending colon (Figure 1 ). Urology was consulted and felt that this represented a submucosal hematoma. Patient was begun on broad-spectrum antibiotics because of concerns regarding possible bacteremia and a worsening DIC laboratory picture. Blood, stool and urine cultures returned negative. Viral cultures and monoclonal antibody staining for adenovirus detection in the urine was negative. Despite a worsening in the DIC panel, the patient remained clinically unchanged. IV steroids were begun approximately 36 hours into patient's hospital stay. Patient had a stabilization of PT/PTT/INR/thrombin time and D-dimer, and a subsequent normalization of labs over the following 4-day period ( Figure 2 , 3 , 4 , 5 , 6 , 7 , 8 ). Patient's diarrhea and hematuria resolved as well. Colonscopy revealed chronic colitis consistent with ulcerative colitis. Cystoscopy revealed a fibrin clot consistent with submucosal hematoma. Patient was discharged from the hospital on a steroid taper, and remains in remission to date. Conclusions Disorders of coagulation have long been associated with inflammatory bowel disease [ 6 - 11 ]. Children, as well as adults, with both active and inactive ulcerative colitis have been found to have abnormal coagulation and fibrinolysis[ 11 ]. It is unclear whether this is a direct or indirect result of inflammatory bowel disease. Although hypocoagulable states have been noted in the literature, most studies indicate an associated hypercoagulable state. There appears to be an increase in thrombin-anti-thrombin complex and a decrease in antithrombin III activity, which causes an increase in thrombin generation[ 10 , 12 , 13 ]. Other studies have demonstrated an increase in fibrinogen content, increase Factor VIII, and Factor IX activity, platelet count and aggregation rate[ 9 , 12 ]. These hypercoagulable abnormalities return towards normal with therapy in direct correlation with sedimentation rate and clinical disease activity [ 12 ], but can still show mild abnormalities despite clinical remission[ 14 ]. The hypercoagulable state in ulcerative colitis is associated thromboembolic events; although uncommon, deep vein thrombosis, pulmonary embolisms and stroke have been associated with ulcerative colitis[ 6 , 15 - 18 ]. Disseminated intravascular coagulopathy is a rare occurrence in inflammatory bowel disease. When it occurs, it is usually associated with other co-founding problems such as septicemia, toxic megacolon or surgery. Presented is a case of DIC associated solely with an ulcerative colitis flare resulting in a bladder hematoma. We presume that the occurrence of DIC in this patient resulted from an acute flare on top of a chronic unremitting course of ulcerative colitis. A thorough infectious work-up of this patient did not reveal any infectious etiology that would have predisposed him to develop DIC. The presumed cause of the DIC was damage to the endothelial wall of the colonic blood vessels, which exposed blood to excessive amounts of tissue factor. This in turn led to the excessive generation of thrombin and a failure of the normal coagulation inhibitory pathways. By treating the ulcerative colitis flare, we decreased the intestinal inflammation and thereby decreased the endothelial cell damage. This, theoretically, resolved the DIC. Patient's clinical symptoms and laboratory values normalized after treatment with intravenous steroids, completely resolving the disseminated intravascular coagulopathy. Competing interests The authors declare that they have no competing interests. Authors' contributions DLS drafted the manuscript. KM and DC participated in the manuscript preparation. All authors approved the final manuscript. Figure 1 Abdominal CT revealing a marked thickening and irregularity of the bladder wall consistent with bladder hematoma. Figure 2 Graphic illustration of C-reactive protein throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Figure 3 Graphic illustration of hemoglobin throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Figure 4 Graphic illustration of platelets throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Figure 5 Graphic illustration of prothrombin time throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Figure 6 Graphic illustration of international normalized ratio throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Figure 7 Graphic illustration of partial thromboplastin time C-reactive protein throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Figure 8 Graphic illustration of D-dimer throughout hospitalization: Day 1 (admission date) – Day 9 (day of discharge). Patient received intravenous steroids at approximately 36 hours into hospitalization. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526371.xml

539286

HTLV-1 and -2 envelope SU subdomains and critical determinants in receptor binding

Background Human T-cell leukemia virus (HTLV) -1 and -2 are deltaretroviruses that infect a wide range of cells. Glut1, the major vertebrate glucose transporter, has been shown to be the HTLV Env receptor. While it is well established that the extracellular surface component (SU) of the HTLV envelope glycoprotein (Env) harbors all of the determinants of interaction with the receptor, identification of SU subdomains that are necessary and sufficient for interaction with the receptor, as well as critical amino acids therein, remain to be precisely defined. Although highly divergent in the rest of their genomes, HTLV and murine leukemia virus (MLV) Env appear to be related and based on homologous motifs between the HTLV and MLV SU, we derived chimeric HTLV/MLV Env and soluble HTLV-1 and -2 truncated amino terminal SU subdomains. Results Using these SU constructs, we found that the 183 and 178 amino terminal residues of the HTLV-1 and -2 Env, respectively, were sufficient to efficiently bind target cells of different species. Binding resulted from bona fide interaction with the HTLV receptor as isolated SU subdomains specifically interfered with HTLV Env-mediated binding, cell fusion, and cell-free as well as cell-to-cell infection. Therefore, the HTLV receptor-binding domain (RBD) lies in the amino terminus of the SU, immediately upstream of a central immunodominant proline rich region (Env residues 180 to 205), that we show to be dispensible for receptor-binding and interference. Moreover, we identified a highly conserved tyrosine residue at position 114 of HTLV-1 Env, Tyr 114 , as critical for receptor-binding and subsequent interference to cell-to-cell fusion and infection. Finally, we observed that residues in the vicinity of Tyr 114 have lesser impact on receptor binding and had various efficiency in interference to post-binding events. Conclusions The first 160 residues of the HTLV-1 and -2 mature cleaved SU fold as autonomous domains that contain all the determinants required for binding the HTLV receptor.

Background Human T-cell leukemia virus type 1 (HTLV-1) has been found primarily in CD4+ and CD8+ T-lymphocytes in vivo [ 1 - 3 ], whereas CD8+ T-lymphocytes are thought to be the in vivo reservoir of HTLV-2 [ 4 ]. However, the in vitro tropism of HTLV-1 and -2, as determined using HTLV envelope-pseudotyped virions or envelope-induced cell fusion assays, appears to be ubiquitous [ 5 - 7 ]. Indeed, we recently showed that Glut1, the ubiquitous vertebrate glucose transporter, serves as a receptor for HTLV-1 and -2 envelope glycoprotein (Env) [ 8 ]. While the precise organization and properties of the receptor-interacting Env domains has not been reported, we found that the amino terminal two-thirds of the HTLV-1 extracellular surface component (SU) are sufficient to confer HTLV-1 tropism to an ecotropic Friend murine leukemia virus (F-MLV) Env [ 9 ]. A cell fusion interference assay performed with this HTLV/F-MLV Env chimera and the parental Env confirmed that this 215 amino acid Env domain, harbors HTLV-1 receptor-binding determinants [ 9 ]. The corresponding domain in MLV Env SU – located upstream of a conserved K/R L L T/N L V Q motif in the SU of the HTLV-1 and F-MLV Env [ 9 , 10 ] – is well characterized and comprises two main functional regions: an amino terminal sequence harboring the receptor-binding determinants, VRA, VRB and VRC [ 11 - 13 ], and a proline-rich region (PRR), starting at the first proline residue of the GPRVPIGP sequence [ 11 , 14 ] and flanked by two highly conserved GXDP [ 15 ] and CXXC [ 16 ] motifs (Figure 1 ). In the ecotropic and amphotropic (Ampho) MLV Env, the PRR is a putative hinge region implicated in conformational changes, triggered after receptor binding, and subsequent fusion [ 17 , 18 ]. In the central region of the HTLV SU, a short sequence (Env residues 180 to 205) harbors high proline content and could be a homologue of the MLV PRR. Figure 1 Homologous modular domains in HTLV and MLV envelopes. Friend-MLV (F-MLV) Env and HTLV-1 Env are schematically represented as open and solid boxes, respectively. Boxes represent, from left to right, the signal peptide which comprises the first 34 and 20 amino acid residues of F-MLV and HTLV Env, respectively, the extracellular surface component (SU) and the transmembrane component (TM) including the carboxy terminal R peptide in F-MLV, which is cleaved in the mature Env glycoprotein [64, 65]. Env landmark positions are indicated and the MLV proline-rich regions (PRR) and the HTLV SU PRR homologue (PRRH) are delineated by vertical lines within the SU at the positions indicated by solid arrowheads. The PRR and PRRH start at the first proline (P) residue downstream of the conserved GXDP motif. Env sequences represented in the figure are obtained from F-MLV strain 57 (accession number CAA26561); P-MLV, F-MCF polytropic MLV (AAA46483); X-MLV, NZB xenotropic MLV (AAA46531); A-MLV, amphotropic MLV strain 4070A (AAA46515); HTLV-2 (NP_041006); and HTLV-1, MT2 strain (VCLJMT). Residue numbering starts from the first methionine of the Env signal peptides. Proline residues and homologous motifs are noted in bold. Amino acid sequence alignments were performed using the Clustal program in the Megalign alignment software package (DNAStar) with manual adjustments. Several studies using synthetic peptides and neutralizing antibodies against the HTLV Env have shown that determinants within this proline rich region homologue (PRRH) are involved in interference to Env-mediated syncytium formation [ 19 - 21 ]. The PRRH had been thought to encode the receptor-binding domain, as based on cell-to-cell fusion assays [ 19 , 22 - 24 ]. However, although PRRH synthetic peptides can block HTLV Env-mediated syncytia formation, they have no effect on HTLV SU binding [ 25 ] and infection [ 26 ]. Indeed, we and others have shown that Env receptor binding per se , as well as interference to receptor-binding, cell-to-cell fusion, syncytium formation, and infection involve several distinct cell surface-associated parameters [ 27 - 29 ]. In the present report, we produced soluble forms of wild-type and mutant HTLV-1 and 2 SU amino terminal subdomains and tested their receptor-binding abilities. We also tested their ability to specifically interfere with HTLV Env cell surface binding, Env-mediated cell-to-cell fusion, and retroviral infection. By testing these essential parameters of Env-mediated dissemination, we delineated the Env receptor-binding domain (RBD) to the first 160 residues of the mature HTLV-1 and -2 SU, excluding the PRRH, and we identified a conserved tyrosine residue at position 114 of HTLV-1 Env as a critical determinant for HTLV Env receptor binding. Results Motif conservation and similar modular organization of HTLV and MLV SU, and identification of a proline-rich region homologue (PRRH) in the HTLV SU As shown in Figure 1 , our alignment of the MLV and HTLV SU reveals several notable motif conservations outlining a similar modular organization of the MLV SU and HTLV SU. A (K/R)LL(T/N)LVQ motif, highly conserved between the F-MLV and HTLV-1 SU, is located immediately downstream of the PRR and its PRRH counterpart, respectively. Another highly conserved motif between MLV and HTLV, GXDP, is found immediately upstream of the PRR/PRRH (Figure 1 ). These two motifs compelled us to notice the PRRH, between the PSQ and KLLTLVQ sequences in HTLV-1, and between the PTQ and KILKFIQ sequences in HTLV-2 (Figure 1 ). As counted from the first and last proline in the delineated sequence, the PRRH has a proline content of 30.8% and 30.4% for HTLV-1 and -2, respectively. This is slightly lower than the 35.3%, 36%, 36%, and 35.6% proline content for the ecotropic, polytropic, xenotropic, and amphotropic MLV Env, respectively (Figure 1 ). The presence of a PRRH in the HTLV SU appeared to be characteristic of their MLV-like modular organization, since HTLV SU average proline content outside of the PRRH does not exceed 11%. Functional, soluble HTLV Env-receptor binding determinants MLV SU receptor binding determinants are all located upstream of the PRR [ 11 , 30 ]. To test whether the HTLV Env receptor binding determinants are also located upstream of the potential PRRH, we constructed a chimeric Env and several soluble HTLV-1 and -2 SU amino terminal subdomains. The chimeric HTLV/MLV Env, H1 183 FEnv, comprises the 183 amino terminal residues of the HTLV-1 SU ending with the PSQL residues fused to the PIGP sequence of the F-MLV PRR (Figure 2A ). In this Env chimera the receptor-binding domain (first 269 residues) of the F-MLV Env was replaced with the potentially corresponding domain of the HTLV-1 Env SU (Figure 2A ). The chimeric H1 183 FEnv construct – which lacks the HTLV PRRH but has the MLV PRR – was properly expressed in transfected cells and was revealed on immunoblots with an anti-MLV SU polyclonal antibody (Figure 3A ). Accordingly, an anti-HTLV-1 monoclonal antibody raised against a PRRH epitope did not bind this chimeric Env (data not shown). Figure 2 Schematic representation of HTLV/MLV Env chimeras and HTLV SU amino terminal subdomains. Env landmark positions are indicated and SU landmark sequences and positions are indicated by arrowheads. Open arrowheads indicate the position of construct borders. (A) HTLV/MLV Env chimeras. The H1 215 FEnv and H1 183 FEnv HTLV/MLV Env chimeras were obtained by replacing the 329 and 269 amino terminal residues of the F-MLV Env (open boxes) with the amino terminal 215 and 183 amino acid residues of the HTLV-1 Env (solid boxes), respectively. The H1 215 FEnv chimera, previously described and formerly designated HHproFc [9], has been renamed here for sake of nomenclature homogeneity. (B) Soluble HTLV-1 (H1) and HTLV-2 (H2) SU amino terminal subdomains, H1 215 SU, H2 211 SU, H1 179 SU, and H2 178 SU were constructed as fusion proteins with a carboxy terminal hemagglutinin (HA) or rabbit immunoglobulin Fc (rFc) tag. All amino acid residue numbering starts from the first methionine of the HTLV-1 or -2 Env signal peptide, the amino terminal 20 and 21 aa residues, respectively. Figure 3 Intracellular expression of HTLV-1 Env chimeras and soluble SU subdomains. Cell extracts (A, B) or culture supernatants (C) were prepared from 293T cells transfected with either full length Env (A) or soluble SU subdomains (B, C) expression vectors as depicted in figure 2. Membranes were probed with either (A) an anti-MLV SU antiserum to detect F-MLV and H1 183 FEnv uncleaved Env precursor proteins (F-MLV Prgp85 and H1 183 Fenv Pr, respectively) indicated by arrowheads, and cleaved SU (F-MLV SUgp70 and H1 183 FEnv SU, respectively) indicated by circles, or (B, C) an anti-rabbit IgG antiserum to detect carboxy terminal rFc-tagged soluble subdomains, including the Ampho-MLV SU subdomain (A 397 SU). HTLV-1 and -2 SU amino terminal subdomains with or without their respective PRRH were constructed as fusion proteins with either an influenza hemagglutinin (HA) or rabbit immunoglobulin Fc (rFc) carboxy terminal tag (Figure 2B ). The H1 215 SU and H2 211 SU subdomains comprise the first 215 and 211 residues, counting from the first methionine in the signal peptide through the KLLTLVQ of HTLV-1 and KILKFIQ of HTLV-2 Env, respectively (Figure 2B ). The H1 179 SU and H2 178 SU, comprising the amino terminal 179 and 178 amino acids of the HTLV-1 and -2 Env, respectively, exclude the PRRH sequence (Figure 2B ). Cell lysates and cell culture supernatants were analyzed to evaluate intracellular expression and secretion of functional SU amino terminal domains in transfected-cell cultures, respectively. H1 215 SU and H2 211 SU, containing the PRRH sequence, and H2 178 SU lacking this PRRH were all efficiently expressed in transfected cells (Figure 3B ). It is noteworthy, however, that recovery of tagged H1 179 SU molecules was largely inefficient because the vast majority of this protein was cleaved (data not shown). In contrast, no significant cleavage was observed with the other soluble domains released in the medium (not shown) (Figure 3C ). As expected for immunoadhesins, H1 215 SU, H2 211 SU, and H2 178 SU rFc-tagged domains were detected as dimers under non-reducing conditions (not shown). Immunoblots of cell extracts revealed two forms of intracellular H1 215 SU and H2 211 SU (Figure 3B ); this was likely due to variable glycosylation of these subdomains. However, a single secreted, soluble form of each of these amino terminal subdomains was detected in cell culture supernatants (Figure 3C ). A truncated Ampho-MLV SU-rFc fusion protein that comprises the amino terminal 397 residues of the Ampho-MLV Env fused to a carboxy terminal rFc tag was constructed (A 397 SU) and used as a heterologous control. A single form of this truncated SU was efficiently expressed in transfected cells (Figure 3B ), and abundantly secreted in cell culture medium (Figure 3C ). HTLV-1 and -2 SU subdomains with HTLV receptor binding properties The amino terminal subdomains were tested for their ability to bind to HTLV receptor-presenting cells by flow cytometry. Using this cell surface binding assay, all of the soluble HTLV SU subdomains bound to the A23 hamster fibroblast cell line (Figure 4 ) as well as to all other cell lines tested, including 293T (human kidney fibroblasts), NIH3T3 and NIH3T3TK - (murine fibroblasts) [ 29 ], HeLa (human ovarian carcinoma cells), D17 (canine fibroblast), Jurkat (suspension human T cell line), activated primary human T cells, and numerous other cell lines and primary cell types that are thought to express the HTLV receptor. As expected from our previous work [ 31 ], none of these soluble HTLV SU subdomains showed detectable binding on resting T lymphocytes. Notably, binding of the HTLV SU to these cells occurred whether they formed or not syncytia in the presence of HTLV Env [ 29 ] and data not shown). Binding by H2 178 SU was similar to H2 211 SU, demonstrating that the first 158 residues of the mature HTLV-2 SU, without the 20 amino acids of the amino terminal signal peptide, are sufficient for cell surface binding, and therefore that the PRRH is not required for receptor binding (Figure 4A ). Figure 4 HTLV-1 and -2 SU subdomains interfere with HTLV Env SU cell surface binding. (A) Conditioned medium from control 293T cells (open histograms) or from 293T cells expressing soluble rFc-tagged HTLV-1 H1 215 SU, HTLV-2 H2 211 SU and H2 178 SU, or Ampho-MLV A 397 SU subdomains (filled histograms), were incubated with A23 hamster cells for 30' at 37°C and binding was assessed by flow cytometry following addition of a secondary FITC-conjugated anti rabbit IgG antibody. Similar results were obtained in binding assays performed using all cell lines described in the text. (B) To assess binding interference, target 293T cells were transfected with the indicated Env construct and subsequently incubated with the HA-tagged H2 178 SU domain (filled histograms). Binding was detected by FACS following incubation with an anti HA 12CA5 mouse mAb and a FITC-conjugated anti mouse IgG antibody. Open histograms represent background levels of fluorescence. SU constructs are schematically represented below each graph by solid (HTLV), open (F-MLV) or grey (Ampho-MLV) boxes. To determine whether cell surface binding of these soluble SU domains corresponded to bona fide binding to the HTLV receptor, we performed an Env-specific binding interference assay. In this assay, transfection of the above described chimeric Env and SU subdomains into 293T cells resulted in interference to cell surface binding by the soluble HA-tagged H2 178 SU subdomain (Figure 4B ). Indeed, nearly complete interference was observed when cells were transfected with the amino terminal subdomain constructs, in the presence and absence of PRRH sequences (H1 215 SU and H2 211 SU versus H1 183 FEnv and H2 178 SU) (Figure 4B ). This effect was specific as HTLV SU binding was not inhibited by a heterologous A 397 SU domain (Figure 4B ). Therefore, we showed that the first 163 and 158 residues, with a cleaved signal peptide, of the mature HTLV-1 and HTLV-2 SU, respectively, contained the entire HTLV Env RBD. These data also showed that HTLV-1 and 2 cross-interfered, consistent with the fact that they recognize the same cell surface receptor for infection [ 8 , 32 ]. Interference to HTLV Env-mediated cell-to-cell fusion by HTLV SU amino terminal subdomains Viral envelope interference occurs when cell surface receptors are occupied by receptor-interacting Env components [ 33 - 35 ]. Since interference to the different Env-mediated functions involves distinct components [ 27 - 29 ], we also tested the abilities of the H1 183 FEnv and the HTLV SU amino terminal subdomains to interfere with HTLV Env-mediated cell fusion. Interference to cell fusion was measured using a quantitative HTLV envelope cell fusion interference assay (CFIA), as previously described [ 9 ]. HTLV-1 Env-induced cell fusion was significantly diminished upon expression of the H1 215 SU subdomain in target cells, 12% ± 2% of control fusion ( P < 0.001), consistent with previous observations using the H1 215 FEnv chimera [ 9 ]. Significant interference to cell fusion was also observed with the H1 183 FEnv chimera, which lacked a PRRH, down to 26% ± 4% of control fusion ( P < 0.001) (Figure 5 ). The corresponding HTLV-2 SU subdomains produced a nearly identical cell fusion interference profile: interference by the H2 211 SU isolated domain, in which the PRRH was maintained, resulted in 15% ± 3% of control cell fusion levels, while the H2 178 SU subdomain, lacking the HTLV PRRH, inhibited HTLV-1 Env-induced cell fusion to 24% ± 6% of control levels ( P < 0.001) (Figure 5 ). It is noteworthy that similar data were obtained when comparing cell fusion interference by H1 215 FEnv and H1 183 FEnv. These effects were specific to HTLV SU amino terminal domains as A 397 SU did not interfere with HTLV-1 Env-mediated cell fusion (83% ± 11% of control fusion) (Figure 5 ). Furthermore, no interference was observed when these truncated HTLV SU fragments and chimeric Env were tested against heterologous, fusogenic control Env such as AΔR Env, FΔR, XenoΔR and VSVG (data not shown). Altogether, these results confirmed our findings that receptor-binding determinants are present within the first 183 and 178 amino acids of the HTLV-1 and -2 Env, respectively. They also indicated that the PRRH (H1 215 SU and H2 211 SU), although unnecessary for receptor binding, modulates the efficiency of interference to HTLV Env-induced cell-to-cell fusion ( P < 0.03). Figure 5 HTLV-1 and -2 SU subdomains interfere with HTLV Env-mediated cell fusion. Cell-to-cell fusion assays were performed by cocultivating fusogenic HTLV-1 Env-expressing cells with target cells expressing the Env derivatives indicated and schematically represented below each histogram. HTLV-1 Env-mediated cell fusion in the presence of target cells transfected with empty vector (Mock) yielded 200 to 1000 blue foci in 4 independent experiments and these levels were defined as 100% cell fusion. Cell fusion levels in the presence of HLTV SU mutants or the A 397 SU control Ampho-MLV SU subdomain is shown as percent of control. Mean fusion percentages were determined from three to four independent experiments. Error bars represent the standard error of the mean. Interference to HTLV Env-mediated infection by HTLV SU amino terminal subdomains Interference, as described above, was based on the inhibition of cell-to-cell fusion induced by fusogenic Env expressed in the absence of other viral proteins. We further evaluated the abilities of the Env chimeras and soluble subdomains to specifically interfere with HTLV Env-mediated infection. HTLV Env-pseudotyped MLV virions, MLV(HTLV), were produced to infect 293T target cells. Because these recombinant cell-free virions are not competent for replication, this viral pseudotype infection assay tests a single round of infection, and does not measure replication and subsequent exponential viral dissemination. Therefore, relative infection values are expressed in linear rather than logarithmic scales. Infection of mock-transfected target cells, devoid of interfering Env domains, resulted in a mean infection value of 9905 ± 1117 infectious units per ml (iu/ml), and this was taken as 100% control infection (Figure 6 ). Similar values, 8803 ± 1871 iu/ml or 89% ± 19% of control infection, were obtained upon infection of target cells expressing a heterologous SU subdomain, A 397 SU (Figure 6 ). Expression of the H1 183 FEnv and H1 215 FEnv chimeric Env in target cells significantly reduced MLV(HTLV) infection to 324 ± 98 iu/ml, 3.3% ± 1% of control infection, and to 307 ± 129 iu/ml, 3.1% ± 1.3% of control infection, respectively (Figure 6 and data not shown). Similarly, the H2 178 SU and H2 211 SU subdomains diminished MLV(HTLV) infection to 191 ± 56 iu/ml and 215 ± 122 iu/ml, 1.9% ± 0.6% and 2.2% ± 1.3% of control infection, respectively (Figure 6 ). The specificity of interference to infection by HTLV Env constructs was assessed by their lack of interference abilities toward Ampho-MLV Env-pseudotyped virions, MLV(Ampho) (data not shown). Thus, for both HTLV-1 and -2, the amino terminal domain upstream of the PRRH was sufficient for specific interference to HTLV Env-mediated infection. Furthermore, in contrast to the cell fusion interference assays described above, the PRRH did not detectably influence MLV(HTLV) infection. Figure 6 HTLV-1 and -2 SU subdomains interfere with infection by HTLV envelope-pseudotyped virions. 293T cells (5 × 10 5 ) expressing the indicated interfering Env derivatives were infected with cell-free HTLV-2 Env-pseudotyped virions MLV(HTLV) carrying a LacZ reporter gene. Infected cells were detected 2 days later by X-gal staining. Infection values are represented as percent of control infection, i.e., relative to infection of mock (pCDNA3.1) transfected target cells, calculated as infectious units per ml of virus containing supernatant (i.u./ml). Data are representative of at least three independent experiments performed in duplicate. Error bars represent the standard error of the mean. Because HTLV dissemination appears to occur mostly via cell-to-cell contact, we also tested envelope interference to infection by HTLV-1 SU amino terminal domains using a cell-to-cell transmission interference assay. In this assay, cells harboring interfering chimeric Env and soluble subdomains were cocultured with cells producing MLV(HTLV) virions. Transfection of either chimeric Env or soluble subdomains into HeLa target cells decreased MLV(HTLV) infection to levels similar to those observed in the cell fusion interference assay presented in figure 5 (data not shown). Identification of residues within the HTLV SU amino terminal domain that modulate receptor binding and HTLV Env-mediated interference Two key residues contained in the HTLV SU RBD and conserved between HTLV-1 and -2, arginine 94 (Arg 94 ) and serine 101 (Ser 101 ) for HTLV-1 Env which correspond to Arg 90 and Ser 97 in HTLV-2 Env, have been shown to alter cell-to-cell fusion and infection when mutated [ 36 , 37 ]. To determine whether mutations of these residues had an effect on receptor binding, we generated H1 215 SU subdomains with either Arg 94 or Ser 101 mutated to Ala, yielding the mutant H1(R94A)SU and H1(S101A)SU subdomains, respectively. We also evaluated mutations of Asp 106 , mutant H1(D106A)SU, and Tyr 114 , mutant H1(Y114A)SU, both residues found to be highly conserved between all human and simian T cell leukemia viruses (unpublished observations). Surprisingly, cell surface binding profiles of H1(R94A)SU and H1(S101A)SU mutants were not significantly altered when compared to binding by the parental H1 215 SU, whereas the H1(D106A)SU mutant presented reduced binding to HTLV receptor-bearing cells and the H1(Y114A)SU mutant showed a nearly complete abrogation of cell surface binding (Figure 7A ). Loss of binding observed with the two latter mutants was not due to decreased soluble SU fragment production, as assessed by immunoblotting of transfected-cell culture media (Figure 7A ). Moreover, equivalent binding profiles were obtained when the same mutations were introduced into the HTLV-2 soluble RBD H2 178 SU (data not shown). Altogether, these experiments demonstrated that Tyr 114 , and to a lesser extent Asp 106 , are key residues involved in HTLV Env receptor binding. Figure 7 HTLV-1 SU amino terminal domain mutants. (A) H1 215 SU constructs were generated with the following SU amino terminal point mutations; R94A, S101A, D106A and Y114A. The abilities of these soluble H1 215 SU constructs to bind 293T cells were assessed by flow cytometry (gray histograms). The levels of expression of the various soluble SU subdomains are shown under each histogram. The abilities of the H1 215 SU mutants to interfere with (B) HTLV Env-induced cell fusion and (C) MLV(HTLV) pseudotype infection was assayed as described in Figs. 5 and 6. Data are representative of at least three independent experiments performed in duplicate. Error bars represent the standard error of the mean. We next tested the abilities of these mutants to interfere with HTLV Env-mediated cell fusion and infection, using the assays described above. As mentioned above, all wild-type and mutant HTLV SU subdomains were produced and secreted with a similar efficiency (Figure 7A ). Expression of the H1(D106A)SU and H1(Y114A)SU mutants, with decreased capacities to bind the HTLV receptor, correlated with decreased interference to HTLV Env-mediated cell fusion and infection. Indeed, H1(Y114A)SU, which had nearly undetectable level of binding, showed the lowest levels of interference and thus allowed the highest levels of HTLV Env-mediated cell fusion and infection (56% ± 16% and 46% ± 10%, respectively) (Figure 7 ). Nevertheless, levels of fusion and infection were lower than that observed when the heterologous A 397 SU was used as a negative control of interference (83% ± 11% and 89% ± 19% for cell fusion and infection, respectively). Thus, overexpression of mutant HTLV SU fragments with highly decreased receptor binding abilities can still exert, albeit to a significantly lesser extent, interference to HTLV Env-mediated cell fusion and infection. We found that similar levels of interference to HTLV Env-mediated cell fusion and infection were observed when either the parental H1 215 SU or the mutant H1(S101A)SU were expressed in target cells (Figure 7B and 7C ). This is consistent with the capacity of this mutant to bind target cells at levels similar to that of wild type H1 215 SU. However, interference to HTLV Env-mediated cell fusion and infection did not always correlate with cell surface binding profiles. While the H1(R94A)SU mutant inhibited cell fusion and infection, its effects were significantly lower than those of the wild-type H1 215 SU (56% ± 8% and 32% ± 2.3%, respectively) (Figure 7B,7C ). Thus, although neither Arg 94 nor Ser 101 of the HTLV-1 SU appears to play a direct role in binding, Arg 94 modulates HTLV Env-mediated fusion and infection (Figure 7 ), likely via post-binding effects rather than binding per se . In conclusion, Tyr114 appeared as the main determinant identified so far for HTLV Env binding, whereas the effects previously described with Arg 94 and Ser 101 are most likely associated with post-binding events. Discussion Here, we report the generation of MLV Env with chimeric HTLV/MLV SU and truncated HTLV-1 and -2 amino terminal SU subdomains that can be expressed in and secreted from eukaryotic cell lines in functional, soluble form. Using these constructs, we demonstrated that the amino terminal 163 and 158 residues (i.e., expunged of their Env signal peptide) of the mature HTLV-1 and -2 Env SU, respectively, were sufficient to exert both HTLV receptor binding and efficient interference to diverse HTLV Env-mediated functions, including binding, cell-to-cell fusion and cell-free as well as cell-to-cell infection. Although the PRRH sequence comprising amino acid residues 180 to 215 of the HTLV-1 Env and 176 to 211 of the HTLV-2 Env was previously thought to be a receptor binding site, our data preclude a major role for this region in the binding properties described above. Indeed, whereas a synthetic peptide composed of amino acids 197 to 216 and located within the HTLV-1 PRRH, has been reported to interfere with HTLV Env-induced syncytia formation [ 22 ], this peptide was later shown to compete neither with receptor binding of the entire HTLV-1 Env SU [ 38 ], nor with infection [ 26 ]. It is therefore likely that the effects reported for PRRH-derived peptides, as measured by syncytia formation, are solely due to post-receptor binding events. However, we identified Tyr 114 of the HTLV-1 Env, which corresponds to Tyr 110 of the HTLV-2 Env, as a key residue in HTLV Env binding and for all the aforementioned HTLV Env-mediated functional assays. We could not detect binding of H1(Y114A)SU by flow cytometry, while this mutant exerted residual, albeit significantly decreased, interference to HTLV Env-mediated cell fusion and infection. Altered folding outside of the binding domain per se , rather than direct alteration of the receptor-binding site, could also account for the lack of binding of this mutant. However, we favor the latter hypothesis, since the H1(Y114A)SU mutant was properly folded and transported to the plasma membrane and secreted in the medium as efficiently as wild type RBD, thus arguing against gross misfolding of this mutant. Accordingly, Tyr 114 appears to be conserved in all known human and simian T cell leukemia viruses strains, which share the same receptor. The receptor-binding site in MLV RBD is composed of a combination of several cysteine loops located upstream of the PRR [ 11 , 39 ] which is linked to a conserved anti-parallel β core [ 13 ]. The isolation of an F-MLV SU amino terminal subdomain allowed crystallization of MLV RBD and the modeling of the RBD cysteine loop arrangement [ 13 ]. The precise organization of cysteine loops, likely to harbor the receptor binding determinants, within the HTLV SU amino terminus remains to be established. Nevertheless, the identification of Tyr 114 as a key HTLV-1 RBD residue points at this determinant as a very likely receptor-binding core. This, together with previous works relying on syncytia formation and cell-to-cell transmission [ 36 , 37 ], will help to distinguish between bona fide receptor binding determinants and determinants involved at a post-binding level. Another recently identified determinant, the Pro-His-Gln SU motif conserved among gammaretroviruses such as MLV and feline leukemia viruses (FeLV), has been determined to play a major role in viral entry during post-binding events [ 40 ]. The mechanism of this effect involves a direct interaction of MLV SU soluble forms with Env attached SU carboxy terminus [ 41 - 46 ]. This interaction between the SU amino and carboxy termini leads to the T cell-restricted tropism of a natural isolate of FeLV, FeLV T, in which the SU Pro-His-Gln motif is mutated. Indeed, FeLV T is restricted in cat to T cells because they naturally express an endogenous soluble FeLV RBD-related factor called FeLIX that trans-complements the lack of the SU Pro-His-Gln motif in the FeLV T Env and restores its post-binding defect [ 47 ]. Despite the HTLV-1 and F-MLV SU homologous modular organization and the assignment of several common motifs between the two latter SU, no obvious Pro-His-Gln motif homologue is present in the HTLV SU amino terminus. Whether a FeLIX-like molecule that interacts with HTLV Env exists in human T cells remains to be addressed. Furthermore, the fact that the Pro-His-Gln has been shown to play a major role in transactivation of viral infection in several gammaretroviruses which are efficiently infectious as cell-free virions [ 42 , 44 , 48 ], raises the question whether the apparent lack of such a motif in the HTLV simple oncovirus-like SU is linked to the relative inefficiency of HTLV Env-mediated infection by cell-free virions. The HTLV SU subdomains described here should prove to be valuable in addressing such questions. The recent identification of Glut1, the ubiquitous glucose transporter of vertebrates [ 49 ], as a receptor for HTLV Env [ 8 ] adds an additional similarity between the Env of HTLV, a deltaretrovirus, and that of gammaretroviruses. All these virus Env recognize multimembrane-spanning metabolite transporters [ 50 , 51 ]. This and the common modular organization of the HTLV and MLV SU raise questions regarding the origin of the HTLV Env. It has previously been reported that envelopes of invertebrate retroviruses may have been "captured" from other viruses [ 52 - 54 ]. As HTLV and MLV have strongly divergent overall genomic organizations, "envelope capture" from related ancestor genes might account for the close relationship between the Env of these phylogenetically distant viruses [ 10 ]. Conclusions We have generated truncated domains of the HTLV Env amino terminus, upstream of residues 183 and 178 of the HTLV-1 and -2 Env, respectively, that were sufficient to bind target cells of different species through interaction with the HTLV Env receptor. We also identified a tyrosine at position 114 and 110 in HTLV-1 and -2 Env, respectively, as a key determinant for this binding. In addition to their use for further exploration of the mechanisms involved in HTLV entry, the tagged HTLV-1 and -2 RBD subdomains described here are novel tools for the detection of Glut1 cell surface expression and intracellular trafficking. Indeed, we tracked intracellular expression of EGFP-tagged HTLV SU subdomains by time-lapse microscopy, and found that they are preferentially routed toward cell-cell contact areas (unpublished observations), where Glut1 is particularly abundant [ 55 ] and our unpublished observations). Furthermore, those HTLV SU derivatives could be of particular importance in view of the key roles played by Glut1 in various biological processes, including T cell survival and activation [ 31 , 56 ], tumor genesis [ 57 , 58 ], and neuronal activity [ 59 ]. Interestingly, soluble HTLV SU subdomains inhibit Glut1-mediated glucose transport, and accordingly, expression of mutants with diminished receptor binding ability resulted in less pronounced inhibition [ 8 ] and data not shown). Thus, these HTLV SU derivatives could also be used as glucose transport inhibitors. These data demonstrate the potential for the novel and broad utility of these reagents in the study of HTLV infection as well as biological processes involving glucose transport and metabolism. Materials and methods Construction of chimeric Env and HTLV-1 and -2 SU subdomains To exchange the PRR and PRRH regions, we introduced an allelic Mfe I restriction site in the HTLV-1 and F-MLV Env. Introduction of this site in F-MLV resulted in the substitution of a glutamine and leucine (QL) dipeptide for the parental arginine and valine (RV) residues of the GPRVPIGP motif, at the start of the MLV Env PRR. Introduction of the MfeI site in the PSQL motif of the HTLV-1 SU maintained the parental QL residues, at the start of the HTLV Env PRRH. By exchanging domains at the Mfe I sites, we derived the H1 183 FEnv chimera containing the amino terminal 183 residues of the HTLV Env followed by the F-MLV PRR. In this chimera, the PSQL/PIGP hybrid sequence is generated at the exchange border, and the PRRH of HTLV is replaced by the F-MLV PRR (Figure 2A ). In contrast, the entire PRRH of HTLV-1 is present in the H1 215 FEnv chimera – this Env chimera has been previously described and designated HHproFc [ 9 ]. The H1 183 FEnv and H1 215 FEnv chimeras, as well as the parental HTLV-1 and F-MLV Env, were inserted in an allelic fashion into the previously described pCEL retroviral Env expression vector [ 60 ]. The HTLV-2 Env expression vector, pCSIX/H2, was constructed by inserting the HindIII – EcoRI fragment from pHTE-2 (a gift from M-C Dokhelar) encompassing the HTLV-2 env gene, the pX region and the 3' LTR into pCSI (CMV promoter, SV-40 intron) [ 61 ] at the HindIII and EcoRI restriction sites. The H1 215 SU, H2 211 SU, H1 179 SU, and H2 178 SU subdomains, corresponding to the HTLV-1 and -2 SU amino terminus with and without their respective PRRH, were generated by PCR and subcloned into the pCSI expression vector as fusion proteins harboring a carboxy terminal rFc or HA tag (Figure 2B ). The H1(R94A)SU, H1(S101A)SU, H1(D106A)SU, and H1(Y114A)SU substitution mutants were generated by oligonucleotide-directed PCR mutagenesis on the H1 215 SU vector and subcloned into the pCSI expression vector. All PCR-generated DNA fragments were sequenced using an ABI Prism 310 sequencer. Cloning details are available upon request. Protein expression and immunoblots Approximately 5 × 10 5 293T cells per 35 mm well were transfected with 5 μg of vectors using a calcium-phosphate-Hepes buffered saline (HBS) transfection protocol. Transfection medium was replaced with 3 ml of fresh culture medium twenty hours post-transfection. Forty-eight hours post-transfection cell culture medium (supernatant) was recovered and filtered through a 0.45 μm pore-size membrane to remove cell debris. Twenty μl were directly analyzed by SDS-PAGE (15% polyacrylamide gel), and the rest was aliquoted and stored at -20°C for later use in binding assays (see below). Cell extracts were collected 48 h post-transfection in 1 ml of cell lysis buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.1% sodium dodecyl sulfate [SDS], 1% Nonidet P-40, 0.5% deoxycholate, and a cocktail of mammalian protease inhibitors [Sigma]) and clarified by two successive centrifugations at 13,000 rpm for 10 min at 4°C in a microcentrifuge. Approximately 20 μl of each extract, adjusted after normalization for protein concentration using the Bradford assay (Sigma), were subjected to electrophoresis on SDS-15% acrylamide gels, followed by transfer onto nitrocellulose (Protran; Schleicher & Schuell). Membranes were blocked in phosphate-buffered saline (PBS) containing 5% powdered milk and 0.5% Tween 20, probed with a 1:1000 dilution of a goat anti-RLV gp70 polyclonal antibody (Viromed) followed by a horseradish peroxidase-conjugated anti-goat immunoglobulin (for detection of chimeric Env), or goat anti-rabbit-IgG-horseradish peroxidase-conjugated immunoglobulins (for detection of rFc-tagged SU subdomains). Immunoblots were subsequently washed three times with PBS-0.1% Tween 20 and revealed by chemiluminescence (ECL+, Amersham). Binding and binding interference assays Binding assays were performed as previously described [ 31 ]. Briefly, 5 × 10 5 target cells were detached with a PBS-EDTA solution, collected by centrifugation, incubated for 30' at 37°C with 300 μl of rabbit Fc-tagged soluble HTLV-1, HTLV-2, or Ampho-MLV truncated SU, washed, labeled with an anti-rabbit-IgG FITC-conjugated antibody, and analyzed on a FACSCalibur (Becton Dickinson). Data analysis was performed using the CellQuest software (Becton Dickinson). For interference studies, 293T cells were transfected with 4 μg of Env or Env SU subdomain expression vectors (carboxy terminal rFc-tagged forms) using the calcium-phosphate-HBS method. Under these conditions, transfection efficiencies ranged from approximately 80 to 90% of the target cells. Twenty-four and 48 hours post-transfection, cells were collected and transfected 293T cells expressing the different interfering HTLV or Ampho-MLV domains were incubated with a challenging HA-tagged soluble HTLV-2 SU amino terminal subdomain (H2 178 SU-HA). Cells were stained using a primary 12CA5 anti HA antibody followed by an anti-mouse-IgG FITC-conjugated antibody before detection by flow cytometry. Envelope interference to cell fusion assay Briefly, the HTLV/MLV Env chimera, H1 183 FEnv, was used to interfere with challenging HTLV Env. The interfering non-fusogenic H1 183 FEnv and truncated HTLV SU subdomains were transiently transfected into HeLaCD4LTRLacZ, a cell line highly susceptible to HTLV Env-induced fusion that contains a stably integrated Tat-dependent LacZ expression vector [ 62 ]. These transfectants were cocultured with Tat-expressing NIH3T3(TK-) cells (NIH3T3(TK-)Tat) that were transiently transfected with the challenging HTLV Env. The NIH3T3(TK-)Tat cell line is resistant to HTLV-Env-induced syncytia formation, despite its ability to express the HTLV receptor and to bind HTLV Env, and thus can be used to precisely monitor fusion of the HeLaCD4LTRLacZ target cells [ 9 , 29 ]. H1 183 FEnv Env and truncated HTLV SU subdomains plasmid DNA (2 to 3 μg) was transfected into HeLaCD4LTRLacZ cells, while challenging, fusogenic HTLV-1 Env plasmid (1 μg) was transfected into NIH3T3(TK-)Tat. The interfering Env or SU subdomain-presenting cells were detached 24 hours post-transfection and 1–2 × 10 5 cells were cocultured for 24 hours with 1–2 × 10 5 challenging HTLV-1 Env-presenting NIH3T3(TK-)Tat cells. Subsequently, the cocultured cells were fixed and stained for β-galactosidase expression as described previously [ 60 ]. Transfection efficiencies of the HeLaCD4LTRLacZ target cells were approximately 50%. Mock transfections were performed with similar amounts of control plasmid DNAs. Env interference was measured by the decreased number of blue foci and was expressed as percent blue foci of control fusion (mock-transfected target cells). Data are represented as mean interference (± standard deviation), and statistical significance of interference levels was determined using a pairwise Student's t test. Envelope interference to infection assay MLV(Ampho) and MLV(HTLV) pseudotyped virions were produced after transfection of 10 6 293T cells with 5 μg pCSI/Ampho or pCSIX/H2, respectively, 5 μg pCL/Gag-Pol [ 29 ] and 10 μg of pCLMFG-LacZ [ 63 ], using a calcium-phosphate-HBS transfection protocol. Supernatants were recovered 48 hours post transfection and filtered through 0.45 μm pore-size membrane to remove cell debris, and stored at -80°C. The pCLMFG-LacZ plasmid is a retroviral expression vector that provides a packageable RNA coding for the LacZ gene marker. pCSI/Ampho is an expression vector encoding the Ampho-MLV Env, and the HTLV-2 Env expression vector, pCSIX/H2, is described above. Virion-containing supernatants were used to infect target 293T cells expressing the chimeric Env or HTLV RBD subdomains. Transfection efficiencies of target 293T cells were >80% in all experiments. Infections were performed 36–48 hours post-transfection on cultures grown in 12 well plates (Costar) at 37°C, medium was changed 24 hours later, and confluent cell monolayers were fixed, stained for β-galactosidase activity before counting blue foci. Interference to infection was determined by infecting transfected target cells with approximately 100 and 1000 iu. Infection was evaluated as described above, and the number of LacZ-positive blue colonies counted was normalized by multiplying by the appropriate dilution factor. The resulting infection values were analyzed as iu/ml of virus containing supernatant. Subsequently the relative infection levels in cells expressing the HTLV SU domains were compared to those of mock transfected cells and were expressed as percentages of control infection (% control). List of abbreviations used HTLV Human T-cell leukemia virus SU envelope extracellular surface component Env envelope glycoprotein MLV murine leukemia virus F-MLV Friend-MLV RBD receptor-binding domain PRR proline-rich region PRRH proline rich region homologue Ampho amphotropic HA influenza hemagglutinin rFc rabbit immunoglobulin constant fragment A 397 SU Ampho-MLV Env fused to a carboxy terminal rFc tag CFIA cell fusion interference assay iu/ml infectious units per ml Arg 94 arginine 94 Ser 101 serine 101 Tyr 114 tyrosine 114 FeLV feline leukemia viruses HBS Hepes buffered saline PBS phosphate-buffered saline SDS sodium dodecyl sulfate Competing interests The authors declare that they have no competing interests. Authors' contributions FJK designed and realized or supervised most of the experiments and co-wrote the manuscript. NM participated to some molecular constructions, set up, realized and analyzed most binding assays and FACS analyses and participated to the redaction of the manuscript. ENG set up and performed the cell-to-cell transmission assay and performed the corresponding experiments, CV constructed some of the RBD point mutants and tested them, MS initiated the project, co-participated in the design of the study, co-coordinated its realization and co-wrote the manuscript, and JLB realized some of the molecular constructs, performed some of the experiments, co-participated in the design of the study, co-coordinated its realization and co-wrote the manuscript. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539286.xml

554095

The role of NGOs in global health research for development

Background Global health research is essential for development. A major issue is the inequitable distribution of research efforts and funds directed towards populations suffering the world's greatest health problems. This imbalance is fostering major attempts at redirecting research to the health problems of low and middle income countries. Following the creation of the Coalition for Global Health Research – Canada (CGHRC) in 2001, the Canadian Society for International Health (CSIH) decided to review the role of non-governmental organizations (NGOs) in global health research. This paper highlights some of the prevalent thinking and is intended to encourage new thinking on how NGOs can further this role. Approach This paper was prepared by members of the Research Committee of the CSIH, with input from other members of the Society. Persons working in various international NGOs participated in individual interviews or group discussions on their involvement in different types of research activities. Case studies illustrate the roles of NGOs in global health research, their perceived strengths and weaknesses, and the constraints and opportunities to build capacity and develop partnerships for research. Highlights NGOs are contributing at all stages of the research cycle, fostering the relevance and effectiveness of the research, priority setting, and knowledge translation to action. They have a key role in stewardship (promoting and advocating for relevant global health research), resource mobilization for research, the generation, utilization and management of knowledge, and capacity development. Yet, typically, the involvement of NGOs in research is downstream from knowledge production and it usually takes the form of a partnership with universities or dedicated research agencies. Conclusion There is a need to more effectively include NGOs in all aspects of health research in order to maximize the potential benefits of research. NGOs, moreover, can and should play an instrumental role in coalitions for global health research, such as the CGHRC. With a renewed sense of purpose and a common goal, NGOs and their partners intend to make strong and lasting inroads into reducing the disease burden of the world's most affected populations through effective research action.

" Each country needs to be able to generate knowledge relevant to its own situation, to allow it to determine its particular health problems, appraise the measures available for dealing with them, and choose the actions likely to produce the greatest improvement in health. This should not be seen as the exclusive preserve of universities or research councils, but equally of health/public services, non-governmental organizations, etc." [ 1 ]. 1 Introduction Non-governmental organizations (NGOs) have been defined by the World Bank as 'private organizations that pursue activities to relieve suffering, promote the interests of the poor, protect the environment, provide basic social services, or undertake community development'. NGO activities can be local, national or international. NGOs have contributed to the development of communities around the world and are important partners of many governments – while remaining independent from governments. According to the Human Development Report [ 2 ], there were in 2002 over 37,000 NGOs in the world, a growth of 19.3% from 1990. Their purposes differ but overall two categories dominate: economic development and infrastructure (26%) and research (23%) . NGOs are generally regarded as valued partners in health research for development, research being viewed as a broad process involving not only the production of knowledge, but also up-stream and down-stream activities needed for its relevance and effectiveness, such as priority setting and knowledge translation. NGOs have made and continue to make substantive contributions through supporting relevant and effective research. In her address at the First Steering Committee Meeting of the International Conference on Health Research for Development in 1999, the (then) Director General of the World Health Organization (WHO), Dr. Gro Harlem Brundtland, voiced her appreciation of NGOs as a partner with WHO in health research [ 3 ]. There are several views on what is meant by global health and global health research. In its simplest form, global health is population health on a global scale, and global health research is research which addresses the health of human populations around the globe. Global health also refers to 'inherently global health issues', that is, health-determining phenomena that transcend national borders and political jurisdictions, such as globalization and climate change. In setting global health research priorities, both the burden of disease and inherently global issues should be considered [ 4 , 5 ]. The vision of health research as proposed by the Commission on Health Research for Development [ 6 ] is a systems approach driven by equity, focused on country needs and priorities, and within an interactive regional and global framework. This paper will address global health as it was defined in a Canadian consultation paper on global health research held in 2001 , that is, the health of individuals and societies in less developed, less resourced, poorer nations and regions of the world. A major global health research issue is the inequitable distribution of research efforts and funds directed towards populations suffering the world's greatest health problems. This situation has been referred to as the 10/90 gap because only a meager 10% of all health research funding is being used to address 90% of the world's burden of disease, suffered primarily in developing countries [ 7 ]. Because of this imbalance, there have been major attempts at redirecting research efforts and funds to the health problems of low and middle income countries. One of the roles of health research is to ensure that the measures proposed to break out of the vicious cycle of ill health and poverty are based, as far as possible, on evidence, so that the resources available to finance these measures are used in the most efficient and effective way possible [ 8 ]. There are many different types of health research. At the 6 th Global Forum on Health Research, held in Arusha, Tanzania in November 2002, Dr. Gerald Keusch, Director of the Fogarty International Center, listed the scope of health research as including: fundamental discovery research, pathogenesis research, epidemiology research, clinical research, product development research, translational and adaptational research, operational research, health services research, policy research and research on health systems [ 9 ]. NGOs involved in health research have primarily undertaken operational and action research, but many have also participated in other types of research such as epidemiological research, social science research, product development research, translational research, health services research, and policy research. The purpose of this paper is to document the role that NGOs have played in global health research and to highlight the need to expand this role. This paper is also intended as a tool to stimulate research activity in NGOs and to advocate for increased NGO involvement in global health research. Following a brief review on the central role of global health research in development, the roles of NGOs at different stages within the research process are discussed and illustrated with a few examples. Key challenges are also identified. The last part of the paper identifies future needs for strengthening the role of NGOs in global health research. 2 Global Health Research and Development While research means different things to different people, it may best be defined as 'a knowledge loop' from generation of knowledge to its effective use [ 10 ]. Indeed, there has been a progressive paradigm shift from narrow 'research' to broader 'knowledge creation and management' [ 11 ]. This broad definition is consistent with that of the Organization for Economic Co-operation and Development (OECD) [ 12 ] which states that "research and experimental development comprise creative work undertaken on a systematic basis in order to increase the stock of knowledge, including knowledge of man, culture and society, and the use of this stock of knowledge to devise new application". Research is recognized as a fundamental ingredient for action [ 13 , 14 ], and it is essential for development because it informs policies and programs; it also guides the development of human resources in these and related domains (see Figure 1 ). However, the links among research, policy-making, programming and training, with advocacy constantly in the background, need to be strengthened. It is being increasingly recognized that investments in health research can be economic and social investments [ 15 ]. In a WHO discussion paper on knowledge for better health, the emphasis is on research as an investment rather than a cost, on the need to turn research into action, and on the vital part of the civil society ( World report on knowledge for better health 2004). Figure 1 The relationship between research and development 2.1 Global health research priorities The call to shift health research priorities from problems of industrialized countries to those affecting populations in developing countries is not new. In 1990, concerns regarding the inequitable distribution of research efforts were first raised in the Report to the Commission on Health Research for Development [ 6 ]. Since then, progress has been made to try to correct this gap, and to build capacity in the countries of greatest need. The 2002 WHO World Health Report [ 16 ] focuses on risks that contribute to the global burden of disease and death, both in developing and developed countries. Dollar expenditures on health research today, however, remain markedly inequitable in terms of populations served and disease burden addressed. Pneumonia, diarrheal diseases, tuberculosis and malaria, when combined, have been estimated to account for more than 20% of the disease burden in the world (mostly in developing countries), yet they receive less than 1% of the total public and private funds which are devoted to health research. The 10/90 gap is as wide as ever [ 7 ]. 2.2 Milestones in global health research and development Several important initiatives have been undertaken to address the global health research agenda. They have been fostered by individuals and groups from local, national and international bodies who shared a common vision in advocating for health research directed towards the low and middle income countries. 2.2.1 Commission on Health Research for Development The Commission on Health Research for Development declared in 1990 that "For the most vulnerable people, the benefits of research offer a potential for change that has gone largely untapped" [ 6 ]. The Commission highlighted several obstacles in undertaking this research, and among others: 1) the insufficient (worldwide) funding of health research directed towards health problems of people in developing countries; 2) the inefficient application of resources; 3) the neglect of major health problems; 4) the lack of individual and institutional health research capacity; 5) the lack of technology transfer; and 6) fragmentation and competition among research initiatives. The challenge to remedy this situation was set down and ultimately led to the establishment of the Council for Health Research in Development (COHRED) in 1993. COHRED works in partnership with WHO, the World Bank and other organizations to strengthen the role of health research at the country level. Over the years, COHRED has assisted increasing numbers of countries in the exploration and implementation of essential national health research (ENHR) strategies. Networks were created to facilitate national level activities in Africa, Asia, and the Commonwealth Caribbean. For example, AFRO-NETS, the 'African Networks for Health Research and Development', was established in 1997 to facilitate exchange of information among different networks active in this type of research in English-speaking Africa, and to facilitate collaboration in the fields of capacity building, planning and research. Regional and global working groups and projects were established which allowed experiences with ENHR to be shared. Several communication strategies were utilized, including quarterly newsletters, websites and other publications to share experiences and lessons learned. A framework for capacity development, a critical component of ENHR, was established through partnerships and like-minded networks and organizations. The book, ' Forging Links for Health: Perspectives from the Council on Health Research for Development ", [ 14 ] and the discussion paper ' Health Research for Development: The Continuing Challenge [ 1 ] review what has happened in the intervening years since the Commission on Health Research for Development made its first major recommendations in 1990. Several questions remain unanswered: • To what extent have the recommendations been implemented? • Have the recommendations made a real difference in the lives of the countries that carry 90% of the disease burden? • Has 'Essential National Health Research' worked? • What is the current situation with regard to health research for development? • Where and how do we proceed from here? The 2000 International Conference on Health Research for Development provided COHRED and several partner organizations with an opportunity to review and reflect on their experience with health research, its impact on health and equity and to devise a global strategy for the first years of the coming millennium [ 14 ]. 2.2.2 Global Forum for Health Research The Global Forum for Health Research, created in 1998 as a response to the Report of the WHO ad hoc Committee on Health Research Relating to Future Intervention Options [ 17 ], has provided a forum for stakeholders to review global health research priorities, promote ongoing analysis of the international health research situation and facilitate coalition building to support its central objective to help correct the 10/90 gap. The Global Forum is managed by a council of 20 members representing government policymakers, multilateral and bilateral agencies, foundations, international NGOs, women's associations, research institutions, and the private sector. It holds funding competitions on targeted global health topics and awards research grants to applicants from low and middle income countries. Its most recent report [ 18 ] emphasized the need for action by combined efforts of the public and private sectors. It also recognized the role of NGOs as a partner in contributing to these efforts. 2.2.3 Canadian Coalition for Global Health Research In November 2001, four Canadian federal agencies, Canadian International Development Agency (CIDA), International Development Research Centre (IDRC), Health Canada, and Canadian Institutes of Health Research (CIHR) signed a Memorandum of Understanding to support national consultation regarding Canada's role in global health research. This marked the first time in Canadian history that Canada's two overseas development agencies, Health Canada and Canada's major federal health research funding agency have collaborated to address global health research. The Canadian Coalition for Global Health Research (CCGHR) is developing into a network of health researchers, funding agencies, NGOs, and other stakeholders committed to support the pursuit of effective global health research by ensuring that all these groups work together as effectively as possible with researchers in developing countries. This collaborative approach serves as a framework for future research projects in the area of global health, with each organization bringing its own specific area of expertise to the table. It aims to improve the effectiveness of development assistance and to increase the sustainable health gains per dollar of Canadian funds invested in research. 3 Key Roles of NGOs in Global Health Research Inequities in health are caused by a number of determinants, including the use of or access to health care facilities. Research which addresses these issues requires an intersectoral approach, involving trans-disciplinary teams and methodologies. Building trans-disciplinary teams requires commitment from the research community to seek out colleagues from other disciplines, from the funding agencies to appreciate innovative initiatives, from the community at large as partners and contributors, and from the policy arena to develop strategies for intersectoral policies and programs which may well have the lead outside ministries of health. Indeed, working outside government altogether may well be a solid and sustainable strategy. Understanding and engaging the broader community on these issues comes naturally to communities unrestricted by bureaucratic boundaries. This is where NGOs excel. NGOs have contributed to all different stages of the research cycle (see Figure 2 ), namely in advocacy, priority setting, capacity building, resource mobilization, sharing and utilization of research findings, and networking. Traditionally, many NGOs which have undertaken activities that address health issues in resource-poor settings are service-oriented NGOs and concentrate their efforts on implementing "action" programs. This type of NGO finds it difficult to identify resources that would allow them to conduct research. While there are NGOs involved in actually conducting research, for most the focus is usually evaluation. Links with the research community are often weak. Other NGOs undertake innovative field-based experimental research. The effectiveness of these initiatives is often learned by trial and error. Unfortunately, while this enhances effective and efficient implementation in the field, research results are only infrequently analyzed appropriately. There are also barriers to dissemination or sharing of research results to a wider audience (eg. other districts within the same country) and to different audiences (eg. to other researchers, research institutions, etc.). Typically, NGO involvement in research is more downstream of knowledge production and it usually takes the form of a partnership with more traditionally-oriented research organizations such as universities or dedicated research agencies. There is a need to include NGOs in the reconceptualization of global health research to ensure completion of the cycle from generation of knowledge to its effective use. Figure 2 Research Process We describe the key roles of NGOs below, using, as a framework, the categories of primary functions of health research systems as recently identified by Butler [ 1 ]. 3.1 Stewardship One of the strengths of NGOs has been as advocates for the populations they serve. Health research can make NGOs become more effective advocates. Governments depend on health research for needs assessments, formulation of policy options, implementation of interventions and evaluation of action plans. Empowered citizens and NGOs can demand accountability of the government. They can also encourage international donors to focus on the health priorities of countries and thus facilitate a check and balance mechanism for good governance. Good governance is needed to improve collaboration and cooperation at the international, national and regional levels in order to tackle inequity. High scientific standards are fundamental components of effective health governance, particularly as they relate to health research systems. The role of research in mobilizing and supporting NGOs, particularly around issues of inequities, is important. NGOs can provide stewardship in terms of the promotion and advocacy for relevant research, shaping research priorities, and the setting and interpretation of ethical frameworks for research. NGOs can often play a more powerful role using the results of research than can the research community itself. Mobilizing communities, utilizing mechanisms for advocacy and acting as an interface between the research community and its wider community will enhance a sense of strong governance and stewardship. 3.1.1 Promotion and advocacy for relevant global health research There is widespread agreement that health research is not sufficiently valued by many societies as a critical input to human and socioeconomic development. The result is often an environment that is neither conducive to, nor supportive of, research. A culture is necessary that recognizes the value of research and one which builds a supportive environment for research [ 19 ]. There is a need not just to allocate funds for research, but also to allocate these funds to areas of research that would have the greatest or maximum social benefit. Advocacy for relevant research, that is, the type of research that will make a difference in terms of equity, health, well-being and development of people, is an important role for NGOs [ 20 ]. Not only can NGOs identify researchable topics, but they can also stimulate demand for relevant research. However, the existing power structure in the research arena often works against NGOs because of a narrow view of research as merely producing new knowledge, with limited consideration of upstream operations (identification of research needs, questions, and priorities), downstream actions (knowledge management, dissemination and translation), and the advocacy efforts required to connect research with policies, programs and training. Historically, the influence of the biomedical researchers' lobby has been the strongest with regard to agenda-setting and fundraising. Behavioral scientists and social health researchers generally have much weaker potential to influence resource allocation, agenda-setting and policy formulation. Partnerships could be strengthened and supported between NGOs and social science researchers in resource-poor countries to improve influence potential, as the social sector issues that tend to be most relevant to human populations are also of utmost importance to NGOs. Creating a favorable environment for "relevant" research requires a health system that is supportive and provides financing opportunities. It also requires the existence of a culture of "evidence-generating and evidence-based research". There must be a healthy relationship between communities, researchers and policy makers. Networks to share experiences, lessons learned and policy impact can be enhanced by partnerships with NGOs. A disproportionately large number of people living in developing countries suffer large disease burdens. Promoting research and development on neglected diseases or issues of global health significance may contribute to bridging the 90/10 research gap, by stimulating research by public or civil society organizations on issues that do not represent marketable research, and are therefore neglected by the private sector. There is a role for NGOs in advocating for more research on these neglected topics (see under 4.1, example of initiative for neglected disease drugs, Médecins sans Frontières [MSF]). Health research needs to generate knowledge that will facilitate the identification of choices and options to reinforce equity-based policies and programs. In doing so, it also needs to address the difficulties of collecting data that are of primary importance when inequities are discussed. The essential function that data serve will allow tracking and monitoring of resources for research and for improving opportunities for those researchers in more disadvantaged countries. NGOs often have access to information that will highlight inequities and the determinants of inequities. Similarly, NGOs can advocate for formative and evaluative research on programs that address major health problems, but which are generally a low priority for funding agencies. In doing so, they can contribute to making data available for evidence-based decision-making in policy and program planning. Food system-based approaches to reducing micronutrient deficiencies and malnutrition in general are one of these under-researched areas. 3.1.2 Shaping research priorities NGOs are well-placed to foster public participation in decisions about health research, as they are close to communities. They can provide the mechanisms by which such public participation is ensured in decision-making processes. Significant progress has been made over the last decade in health research priority-setting for the implementation of ENHR at the country level. Among the lessons learned, it appears that community involvement is in most cases an unresolved issue [ 21 ]. What is certain is that, critically at the priority-setting stage of the research cycle, the community must be involved, and NGOs may be instrumental in achieving this. Defining the research that needs to be done requires the input of civil society and NGOs as much at the beginning as at the end, in terms of dissemination, communication and action. 3.1.3 Setting and interpreting ethical frameworks NGOs assume a range of roles in research, but a thread that runs through all these is their representation and advocacy for the vulnerable. Broad research roles are described in greater detail in other sections of this paper. This section focuses on the role of NGOs in shaping and interpreting ethical frameworks [ 22 - 25 ], that is, the incorporation of ethical principles in their research partnerships with other organizations. As researchers or research partners, NGOs have a responsibility to ensure that ethical issues are addressed in both the design and conduct of the research. There are distinctive challenges in conducting health research in developing countries, namely to fulfill moral duties of justice and respect in the face of poverty, lack of resources and the potential for exploitation. The Nuffield Council on Bioethics [ 26 ] designed an ethical framework for health research in developing countries based on the duty to alleviate suffering, to show respect for persons, to be sensitive to cultural differences, and to not exploit the vulnerable. As NGO research is often conducted among the most vulnerable populations, where power relations are tipped in favor of researchers and those who are literate and eloquent, issues of informed consent and participants' understanding of it and the research, as well as participants having access to the benefits of research, are of special concern. Particularly when research is conducted by first world researchers in resource-limited settings, NGOs who partner in this research at times need to recommend and advocate for reviews from local research and ethics committees, as well as those from industrialized countries. Where relevant, they may also encourage the development of independent national ethics committees and national ethical guidelines, taking account of existing international guidelines [ 22 - 25 ]. This process may involve interpreting cultural ethical frameworks and beliefs, for instance, culturally appropriate means of obtaining informed consent from research participants. In addition, NGOs can make sure that the development of local expertise in health research is an integral component of research proposals. As watchdogs, NGOs actively seek breaches of ethics and hold researchers to account when the principles of respect for persons, beneficence and justice are not upheld, a role they are well positioned to assume given their understanding of and links to marginalized groups. Watchdogs, as they uncover ethical breaches that may be defined by culture or power relations, have assisted in shaping ethical frameworks to better address ethics when research is conducted among vulnerable groups. In the communities where NGOs work, they can act as community partner members of and witnesses to research. In this role they can assist with, for example, interpreting research objectives to participants to ensure that consent is informed and the rights of subjects are respected. They may provide researchers with enumerators or local information to expedite the data collection process. NGOs can also monitor the long-term outcomes arising from research, and make sure that the participants benefit from successful intervention. As knowledge translators, NGOs interpret the knowledge generated by research to their constituents, a key role in working towards the vulnerable having access to the benefits of research that could improve their lives. This may be research conducted in these communities or globally. 3.2 Mobilizing resources for research While current levels of financial resources are not sufficient to adequately respond to the demonstrated need for health research, there are many sources of "funds" for health research. Some are monetary contributions and some are in-kind contributions. NGOs can provide not only direct funding for projects (albeit in a limited manner) but, and perhaps equally important, they can provide valuable in-kind funding. Thus, personnel or materials developed by NGOs can be used in health research projects at little or no cost. Some NGOs are directly involved in the administration of research grants. Others may be the fiduciary agent for a grant to a research organization that is exploring an issue related to an NGO program. However, most are organizations that work with communities. A major role is therefore to identify resource gaps using networks to link communities, health providers and managers, and funding agencies in a meaningful way so that financing can appropriately be directed to targeted health issues. NGOs may also contribute by identifying other potential sources of funding, for instance, in the local private sector. 3.3 Knowledge generation Knowledge can be acquired in various ways, by many methods, and by different types of people; there are different cultures of enquiry. Because of their typical 'grass-roots' experience, several NGOs are able to access indigenous knowledge and specific information, which may be less attainable for other types of organizations. This type of knowledge might be very useful when pooled with knowledge acquired by others; in this way, a more comprehensive analysis can occur. NGOs can be particularly adept in conducting formative research (baseline studies, needs assessment), in operational or action research and in process and impact evaluation. This type of research is particularly relevant for setting priorities, for informing intervention, as well as for identifying further research needs. Although knowledge generation is generally not a primary NGO activity, there may be specific 'knowledge generation' research niches for NGOs. For instance, as suggested by the Canadian Council for International Cooperation (CCIC) and actually carried out by a few NGOs, " There is a need for NGOs to be more involved in policy research even in Canada " (Interview with B. Tomlinson, CCIC). Figure 3 illustrates the research cycle in the narrow sense of knowledge generation. This cycle applies whatever the research type, and whether the research is conducted by an NGO or an academic institution. Figure 3 The research (knowledge generation) cycle (adapted from McKenzie [36]) 3.4 Utilization and management of knowledge While asserting that the production of knowledge is the primary function of research, and that levels of knowledge have increased considerably, a discussion paper for the International Conference on Health Research [ 1 ] also recognizes that the ability to draw from research in terms of lessons learned, application to interventions, and programming and policies which support the overarching goal of equity, is often lacking. Inadequacies include the inability of developing countries to access pertinent international research literature and knowledge bases (either as contributors or users), the inability to access new information technologies, and the inability to ensure closer links among the research community, health service managers and health policy makers. The effective use of research findings and their dissemination is an increasingly important public health policy concern. In 1995, an international research conference was held in Vancouver, Canada, on dissemination research. This type of research is similar to what is now called 'translational research' , that is, the conversion of research findings from basic, clinical or epidemiological environmental health science research into information, resources, or tools that can be applied by health care providers and community residents to improve public health outcomes in at-risk neighborhoods. NGOs are frequently at the interface of applied research and policy-making, at least at the administrative level, and their potential input into research utilization for policy-making needs to be valued. Research can make a substantive contribution in at least three phases of the policy-making process: agenda-setting, policy formulation, and implementation [ 27 ]. It is widely recognized that health research is underutilized in policy-making. The generation of new knowledge is highly valued, but its translation and use does not appear to be valued as much [ 28 ], which may partly explain why application of newer knowledge is often a weak link in the research cycle. Factors potentially enhancing utilization can be identified by exploration of priority-setting, activities of the health system at the interface between research and policy-making, and the role of recipients, or "receptors", of health research [ 27 ]. There are several models of research utilization in policy-making, but interactive or exchange models may be more conducive to the effective use of research than unilateral models because they bring researchers and decision-makers closer together [ 10 , 27 ]. NGOs often play a critical role in interpreting the evidence and translating its relevance for local communities. Inevitably the level of involvement by the community depends on relevance and opportunity for action and advocacy. Assessing and evaluating opportunities for advocacy and action occur as NGOs work with communities on these issues. Effective involvement of the community and its participation is a "matter of reciprocity and continuing dialogue in which participation takes different forms and influences change in several directions" [ 14 ]. Once the evidence has been analyzed and assimilated, NGOs can serve as intermediaries in delivering feedback to communities and in the planning, implementing and monitoring of new interventions, policies or other actions which might have been proposed. The knowledge and information acquired by NGOs can be unique and offer added insight into new ideas for future health research. This is, in part, because of the extensive interrelationships NGOs have forged with different communities, organizations, the private sector and governments, among others, often over decades of dedicated work. Additionally, NGOs are in a good position to test the ability of research findings to be scaled up in a 'real world' environment. According to Lavis et al [ 10 ], while the "knowledge loop" needs to be completed, that is, from knowledge production to knowledge-based decision-making through knowledge transfer or brokering, not all research organizations should become involved in knowledge transfer; if they do, the knowledge pyramid may be shaky. Innovations stemming from research are at the base of the pyramid, and actionable messages are at the top. Individual studies and synthesis of research knowledge are the intermediate layers. Lavis et al contend that it may not be relevant to transfer knowledge from individual studies, but rather, from bodies of cumulative research knowledge, and that knowledge transfer brokers are needed for this purpose. This model of specialized roles is probably more relevant at the macro level and in industrialized countries. In resource-poor countries, polyvalent organizations such as NGOs have a key role in sharing, translating and implementing research findings at the community and country level. They provide channels for the use of research results at the community level, as they are closest to the communities themselves. For that very reason, they may also feel more compelled to complete the research cycle, including application of the findings. Third World Network , for instance, an independent non-profit international network of organizations and individuals involved in issues relating to development, conducts and disseminates research to help organizations around the world participate in and influence international economic and social policy. NGOs may also be involved in testing pilot models of intervention and in their subsequent scaling-up. 3.5 Capacity development The preliminary examination of the functions performed by the some 125 organizations involved in a significant way in health research reveals that while knowledge generation is a concern shared by most, research capacity strengthening receives relatively little attention [ 1 ]. One weakness or inattention in research capacity strengthening activities, for example, has been the lack of a recognized career path for local health researchers which has resulted in diverting promising researchers to other careers or to other countries. The development and retention of research capacity remains a challenge in many countries [ 29 ]. Quality control and assurance requires skills and structures which support these objectives. Skills such as leadership, advocacy, networking and communication are important and need to be built through capacity development. Research management is also a skill which needs to be strengthened and a skill that will improve the quality, appropriateness and timeliness of research and its dissemination. NGOs in the North and in resource-poor countries often have the capacity for facilitating training and for sharing the lessons learned in needed skills. Partnership with NGOs in such capacity-building needs to be valued and reinforced. The Canadian Society for International Health (CSIH) and the Canadian Public Health Association (CPHA) have participated in capacity-building activities in many countries and continue to share their experiences and lessons learned. Support for such sharing and building capacity makes sense and should be facilitated by donor agencies. WHO, through its creation of a Department of Research Policy and Cooperation within the cluster of Evidence and Information for Policy, has defined as one of its objectives: "the development of initiatives aimed at strengthening research capacity in the developing world with the ultimate aim of enshrining research as a foundation for policy". A number of other international initiatives have also attempted to address some of these capacity issues: the International Health Policy Program (IHPP), the Applied Research on Child Health (ARCH) project, the Swiss Commission for Research Partnership with Developing Countries (KAPE) and, in Canada, the IDRC. Since 1970, IDRC has been providing financial and technical assistance to academic institutions, government agencies and NGOs in developing countries, as a means of promoting sustainable and practical development and strengthening indigenous research capacity. IDRC's experience provides important and valuable lessons about implementing applied research in partnership with NGOs [ 30 ], as summarized in the table 1 . Table 1 Lessons learned from research in partnership with NGOs: IDRC experience First, applied research should have a practical application, reinforce knowledge and skills, and introduce and promote innovative, effective strategies and approaches for improving human health and well-being. Not only should research results be for local application, they should also be shared and adapted to other venues and contexts. Second, efforts need to be made to build knowledge and understanding about the benefits accruing from applied research. NGOs, by their very nature, are action-oriented. Applied research is often perceived as of limited use to their ends, an esoteric, academic exercise of limited value to the immediate needs of the poor and disadvantaged. Time and effort need to be invested in nurturing an understanding within the academic community of the value of applied research within the context of development efforts. Third, applied research should be used to develop and strengthen local research capabilities. NGOs do not, as a rule, possess the internal capacity and skills to design and conduct applied research studies. Attention should be paid to assisting NGOs in making contact with qualified researchers, and increasing NGO knowledge and skills to negotiate the terms of reference for applied research studies. This cannot be achieved simply through providing information about applied research methodologies or organizing a single workshop. Trust has to be developed between the NGO and academic communities, as a means of reinforcing linkages between them and building upon and using their comparative strengths, characteristics and areas of expertise to design and conduct applied research. Fourth, local communities should be involved in the design and implementation of applied research activities. The local people need to understand the purpose of the proposed research, provide input and advice about its design and conduct, and be actively involved in the application and dissemination of research results. Without the active participation of the community, the utility and eventual application of the research results will be of little value. Source : [ 30 ] The CPHA, through the CIDA-funded initiative Canada's International Immunization Program – Phase 2 (CIIP2), dedicated 5% of the program's budget to applied research. Part of this funding was used to strengthen primary health care in developing countries through the NGOs that implemented the immunization and primary health care activities through the auspices of CIIP2. NGOs who wish to become more involved in research generally recognize the need for extramural training and support. Partnering with universities and research institutions may provide such training opportunities. Additionally, there are international institutions such as INTRAC (International NGO Training and Research Centre) that are specifically geared towards meeting the challenges and needs of NGOs in research. Those NGOs that are part of international networks can draw from the body of research conducted elsewhere. NGOs may also provide substantive input into research training, be it by grounding research methods in reality so that research is more applicable, or by providing research sites and questions for academia and graduate students. NGOs may also be in a good position to identify young scientists and promising investigators in host countries. Stimulating the demand for research by user groups, rather than supply-driven research, is one of the three strategies identified by Harrison & Neufeld [ 31 ] for capacity-building for essential national health research. NGOs and communities as user groups could be the target of capacity-building efforts. 4 NGO involvement in health research There is a lack of accessible and centralized information on NGO involvement in health research, although the CPHA CIIP2 applied research publication lists over 20 examples of NGO-related applied research carried out in the 1990s. The examples given below are based on discussions with a limited number of Canadian and international NGOs: CARE, World Vision Canada (WV), CECI (Centre d'étude et de coopération internationale), Inter Pares, HKI (Helen Keller International), and CCIC. In the case of AMREF (Africa Medical Research Foundation), ADI (Alzheimer's Disease International), Médecins sans Frontières (MSF) and RITC (Research for International Tobacco Control), most of the information was obtained from their websites and related publications and documents. The interviews and discussions covered the specifics of the implication of the NGO in health research, lessons learned through the experience, and respondents' perceptions on the role of NGOs in global health research, and on the strengths and weaknesses of their organization in this regard. These selected NGOs provide insight into some of the critical issues facing NGO involvement in global health research. It should be kept in mind that this selection is small and not meant to be representative. Nonetheless, all of these NGOs are involved, directly or indirectly, in global health research, and they are all Canadian or present in Canada. 4.1 NGOs and their involvement in global health research: illustration cases The interviews covered a broad range of cases, from NGOs little involved in research to those actually conducting independent research. The types of involvement are briefly described below. A salient observation is that what is considered as research by different NGOs is, for the most part, unclear and highly variable. This suggests the need for NGOs to develop common views on what is research, the various types of research, and the components of the research process. The interviews also revealed that while some NGOs are reluctant to be involved in research, others are eager to strengthen their capacity to do so. CECI has long been involved in health research, although it is reluctant to call this 'research'. A major activity is the undertaking of baseline studies that typically include an assessment of the health and nutritional status of populations. The data are used to orient or reorient programs, and to inform communities. In Cambodia, for instance, it conducted an initial assessment for a project aimed at improving the livelihood of rural poor in two sectors: health/nutrition, and agriculture marketing (CECI and Cambodia Researchers for Development: Improving Livelihood of the Cambodian Rural Poor: Strategies in Health, Nutrition and Agricultural Commodity Marketing, 2001). One interesting aspect of its recent work is the 'policy feedback' that it conducts in its large projects. The intent of the analysis is to clearly identify the lessons learned, and to discuss these with decision-makers and technical officers. This may be considered as part of 'knowledge translation' and it can be a particularly useful approach in advancing policies and programs. While CECI is also involved in health projects that do not include research even in a broad sense, it conducts research in areas that are indirectly related to health. For instance, in the IDRC-funded project intended to alleviate poverty in Burkina Faso, Viet Nam and Nepal, it collaborates with local universities and research institutions for the research and training components, notably on adapting the assessment of poverty to the specific context. CCIC and its member NGOs are involved in international policy research. For instance, Trade-Related Aspects of Intellectual Property Rights (TRIPS) agreements have implications on access to drugs. In the reorientation of CIDA for improved aid effectiveness, there are obvious health implications, including how to respond to health plans as defined by health ministries, and assist with poverty reduction strategies. CCIC sees research on policies as a critical role of NGOs, and considers that NGOs should be more involved in the policy debate both in Canada and globally. World Vision (WV) Canada is active in research, particularly (but not only) in the framework of its MICAH projects (Micronutrients and Health in Africa) funded by CIDA. It primarily conducts formative and evaluation research (see table 2 for report of findings in Sénégal, published jointly with CIDA). Although it has PhD or MSc level personnel in each of its technical units, it does not have in-house research expertise per se ; it partners with research institutions, in the field and in Canada. It does not have the capacity to analyze all the data that it collects and therefore it collaborates with academic institutions in Canada. Graduate students can use the data for their theses. WV officers may also sit on graduate students' supervisory or examining committees. The primary use of the research findings is to reorient programs and inform the community. As programs may have to change their operations as a result of such research, the exercise may, at times, be regarded as threatening. Table 2 Final evaluation report, World Vision Canada, Micronutrient-for-Health Project in Sénégal (2002) The objectives of the project initiated in 4 districts in 1997 were to reduce micronutrient malnutrition among women and children, to reduce the incidence of illnesses affecting micronutrient status, and to strengthen local capacity for controlling micronutrient malnutrition. The baseline study revealed a high rate of (iron deficiency) anemia in pregnancy (49%), of low retinol (vitamin A) levels in breastmilk (57%), and of low serum retinol concentrations among preschool-age children. Iodine deficiency was widespread, with 20% of school-age children showing severely low urinary iodine levels. A similar survey was conducted after 4 years of project activities, and included control zones in each district. The final evaluation showed an almost complete elimination of vitamin A deficiency in the project areas, which was primarily attributable to the high coverage of vitamin A supplementation of under-fives and postpartum women. Household use of iodized salt increased from 6% to 14%. Anemia remained high among pregnant women (44%), however, in spite of the iron-folate supplementation scheme. The rate of intestinal parasites declined, but the project did not have an impact on diarrhea. The MICAH project had a positive impact in strengthening the national vitamin A policy of Sénégal. The evaluation report was published by the project and widely disseminated. The survey findings and recommendations were fed into the design of an up-scaling phase of the project, with more emphasis on the reduction of anemia among women. CARE is directly involved in research, and its involvement covers the whole process from conceptualization of the research question to data management and dissemination of research results. Some offices have staff whose role is specifically research-related, but this varies. They also work with partners. CARE has even been contracted by some donors to conduct research. The research is primarily qualitative, including participatory approaches, as well as operations and action research. CARE also conducts surveys, situation analyses and policy reviews. It receives funding for research from bilateral and multilateral agencies, and from large organizations such as Family Health International and the Population Council. Helen Keller International (HKI) is a technical assistance NGO that is also directly involved in research as part of its mandate. It addresses the causes of preventable blindness. It also provides rehabilitation services to blind people, and helps reduce micronutrient malnutrition which can cause blindness and death in children. It is involved in most stages of the research cycle, focusing on operations and action research. HKI's focus on blindness and micronutrients is a strength in that its research is more focused than that of other NGOs involved in health and nutrition. Its funds for research come from different sources. A research component may be built into programs, some operational research is conducted with funds for surveillance, or funds are provided for R&D specifically (eg. for FRAT studies [Fortification Rapid Assessment Technique]) and for the development of tools to assess the quality of nutrition interventions leading to adoption of relevant strategies (in Mozambique, Burkina, Mali and Niger). HKI has in-house expertise in research. There are several full-time research positions. In addition, it works with research partners at the local level, as well as with universities in Canada and USA. Inter Pares was created in 1975 to support NGOs from the South and to provide international development education in Canada. Inter Pares uses its own funds to conduct social research on political and economic issues, primarily action research. For instance, it carried out collaborative research with NGOs in the Philippines and of Bangladesh on family planning policy, and in Africa it has carried out research on economic issues. With Forum Afrique Canada , for instance, it is studying Canadian government trade and aid policy after G-8. It has in-house research expertise, particularly in sociology, although there is no research position as such. It usually works with partners, as it is a small NGO. Inter Pares uses research findings mainly for education and advocacy. AMREF has been active since 1957 in the field of applied health research and has an extensive bibliography documenting research results in the form of peer-reviewed publications, theses, manuals, reports, abstracts and conference presentations. The focus of AMREF's research activities has been primarily in the operational and applied domains. Many have addressed the important disease burden caused by communicable diseases such as malaria (see table 3 ) and schistosomiasis, but others have addressed organizational issues such as health information systems and technological issues like field diagnostics. Table 3 Example of an AMREF research study listed in its extensive bibliography In 1995, D'Allessandro et al [ 32 ] published a study which compared the efficacy of insecticide-treated and untreated bednets in preventing malaria in children living in the Gambia. The survey included 2300 children between the ages of 1 and 4 years; 1500 from villages who had received insecticide-treated bednets within their primary health care and 800 from villages which had not received treated bednets. It was found that the greatest benefit, in terms of reduced malaria morbidity, was observed in children who slept regularly under treated bednets. Measurable benefits were also accrued in children who slept regularly under untreated bednets, compared to children who did not use bednets at all. The conclusion of this study was that educational campaigns might well promote even the use of untreated nets because of the additional health benefits, while ultimately aiming at coverage with insecticide-treated bednets. Alzheimer 's Disease International (ADI), an NGO affiliated with WHO, specifically provides support for research among its numerous activities. In particular, it supports the research work of the 10/66 Dementia Research Group (the 10/66 refers to the dementia research gap, in which 'less than one-tenth of all population-based research into dementia is directed towards the two-thirds or more of cases living in developing parts of the world [ 31 ]). The vision of ADI is that research not only generates awareness, but is the basis for policy which, subsequently, can provide the impetus for development of appropriate services for affected persons. The 10/66 Dementia Research Group divides its research activities into pilot studies, qualitative studies, intervention studies and population-based studies. This group has published a consensus statement [ 33 ] and a methods paper [ 34 ], and members are now publishing research results (see table 4 ). This NGO's 10/66 Dementia Research Group has regional networks in India and South Asia, Latin America and the Caribbean, China and South East Asia, Africa and Russia, Eastern and South Eastern Europe which are coordinated by Dr. Martin Prince of the Institute of Psychiatry in London, England. Table 4 Example of an Alzheimer's Disease International- supported health research study The results of a population-based study undertaken in Kerala, India to evaluate a community dementia case-finding program was published by Shaji and collaborators in 2002 [ 35 ]. Their aim had been to validate a training program where local community health workers (CHWs) were trained to identify possible cases of dementia. The training program consisted of 2 ½ hours of formal instruction. Workers' diagnoses were then confirmed by an experienced psychiatrist. The 19 CHWs identified 51 possible cases among 1979 persons aged 60 and older in their communities. There was expert confirmation for 33 of these cases (65%). Although the remaining 18 did not have dementia, 13 did in fact suffer from other psychiatric illnesses and only 5 had no psychiatric diagnosis at all. The conclusion was that CHWs can play an important role in identifying cases of dementia in a community setting. Médecins sans Frontières (MSF) was the first NGO to both provide emergency medical assistance and publicly bear witness to the plight of the populations they served. MSF is at the forefront of emergency health care as well as care for populations suffering from endemic diseases and neglect. MSF has undertaken an initiative on drugs for neglected infectious disease which combines advocacy, research and capacity development, and networking. In contrast with private sector research, it is need-driven rather than profit-driven. Five pilot projects are currently underway focusing on capacity building and technology transfer. This initiative started with a review of pharmaceutical research and development outcomes over the last 25 years and of current private and public initiatives. Highlights of the findings and conclusions, published in Lancet [ 36 ], are provided in table 5 . Table 5 Analysis of trends, drug research and development for tropical diseases, MSF (2002) The extensive review revealed that of 1393 new chemical entities marketed between 1975 and 1999, only 16 were for tropical diseases and tuberculosis. All new drugs for neglected diseases represented a clear therapeutic benefit, and all are included in the WHO Essential Drug List, which indicates the importance of new drugs for neglected diseases. In contrast, over the same period, two out of three new drugs offered little advantage over existing ones. There was no indication that drug development for neglected diseases would significantly improve in the near future, however. Private-public partnerships, or else, incentives to encourage private investment towards the development of new cost-effective drugs may help overcome this limitation. For the most neglected tropical diseases which may not account for a large share of the global burden of disease, a new approach is needed. The feasibility of an international not-for-profit network that would focus on the most neglected diseases is being tested in the on-going pilot projects. Research for International Tobacco Control (RITC) is an International Secretariat based at IDRC headquarters (Ottawa) that funds multidisciplinary tobacco control research projects in developing countries. Its mission is to create a strong research, funding and knowledge base for the development of effective tobacco control policies and programs, through a combination of research, dissemination, strengthening of capacity and coordination. RITC concentrates on research on psycho-social correlates of tobacco use. It provides support to research projects conducted by NGOs, such as, the Youth and Tobacco Survey conducted in Russia by the Russian Public Health Association, with the technical assistance of the CPHA (table 6 ). Table 6 Youth and Tobacco Survey, Russian Public Health Association (RPHA), 1999 This study is part of the Global Youth and Tobacco Survey (GYTS) undertaken in several countries around the world. The survey in Russia was designed to provide prevalence data on tobacco use among adolescents in school (13–16 years), and to better understand and assess students' knowledge, attitudes and practices related to tobacco. Information pertained to, for instance, age of initiation of tobacco use, perceived health risks and social benefits, extent of peer and advertising pressure, perception of the tobacco-related curriculum, and likelihood that tobacco users will quit. The survey raised awareness on the issue of smoking and youth. Several recommendations were made by the Association to Parliament on the basis of survey findings, including the adoption of legislation to limit tobacco advertising, to reduce the tar and nicotine content of cigarettes, and to have an impressive warning labeling on packages. Seminars and conferences on the survey results and their implications were held. The Association has prepared a report: "Tobacco or Health in Russia". Additionally, the President of the RPHA, as a result of the GYTS survey and the ground-breaking leadership role the RPHA played in tobacco control among youth in Russia, was a member of the delegation from the Russian Federation to the WHO Framework Convention on Tobacco Control (FCTC) – hence an example of the translation of applied research results into policy action. CPHA has also provided technical and financial support through its various international initiatives funded by CIDA to its public health association partners to carry out of the GYTS in Burkina Faso, Niger, Haiti, and Cuba; and in partnership with Institutes of Public Health in Bosnia & Herzegovina, Serbia & Montenegro, and in the UN-administered province of Kosovo. The results from the surveys (carried out in collaboration with the Centers for Disease Control and Prevention [CDC] of USA) are being used to develop tobacco control policy and youth-focused smoking prevention and cessation programs. The experience of CSIH in global health research is described under 5.2.1. 4.2 NGOs' perceived strengths and weaknesses in research The following summary of perceived NGO strengths (and weaknesses) in health research is based primarily on the data from individual and group interviews that were specifically conducted as inputs to this discussion paper. It is a common view that NGOs are in a good position to participate in health research because of their knowledge of, and their presence in, local communities. Furthermore, their involvement increases the relevance of research to communities. " NGOs give a human face to research, and they are in a good position also to build on indigenous capacity " (Interview with S. Baker, HKI Africa). Additionally, they may be more compelled to complete the research cycle and apply the findings. Their involvement in research is perceived as a motivation to use the research to design, develop and respond to circumstances affecting development. Evaluation research, in which they are frequently involved, tells them whether or not they have an impact. Since they are closely connected with communities, they have the ability to see the application of their research results. For this reason, NGO-initiated research is often more likely to be translated into practice in a timely manner as it is almost always directly related to practice. The NGO structure brings concreteness and a style which is guided by values and beliefs with an action orientation. Another major strength of some NGOs is their international networks which give them access to technical information and support. Finally, NGO values of ethics, solidarity and dialogue are important for health research to contribute to reducing inequities and for empowerment. 4.3 Constraints to greater NGO involvement in health research 4.3.1 NGO views on research and its congruence with their mandate Many reasons that account for the reluctance of NGOs to become (more) involved in research activities pertain to NGO perceptions on research (Center for Advanced Studies of International Development. Symposium on NGO/Academic Linkages. East Lansing: Michigan State University; April 16–17, 1993; Edwards M, Griffiths M: Terms of Reference for the DSA [Development Studies Association] Workshop on the Academic Practitioner Interface. London: DSA, 1994). In the past, research was an academia-driven and based, elitist and theoretical exercise, the results of which are of little use to NGOs and the communities with which they work. Traditional research strategies and approaches were seen as top-down, non-participative and controlled by external actors. Research activities were regarded as requiring special technical expertise, much time and effort, access to professional journals and research literature, and substantial human and financial resources, characteristics not typically found in NGOs. Finally, the scientific rigor demanded by researchers was believed to be difficult to achieve in field-based situations, where unpredictability and subjectivity are the norm. What NGOs perceive as research and as their role in this respect varies widely. For instance, there is some hesitation and even reluctance in including baseline studies or project evaluation under 'research'. Another obstacle is the fact that some NGOs that raise funds from the public are afraid to go against the expectations of the donors if the money is reallocated for research, and particularly for policy research in Canada: " Donors do not want to hear that NGOs are doing research as they are implementation organizations " (Interview with C. MacDonald, World Vision Canada). 4.3.2 Lack of training opportunities, funding, time and motivation Among other barriers, interview respondents mentioned lack of training opportunities, lack of funding owing to their (limited) mandate, priorities of funding agencies, and time constraints. Because of lack of training or of specialized researchers, NGOs may not be in a position to conduct top quality research, and scientific rigor may be lacking in certain instances. Lack of access to scientific literature when in the field can also be a major shortfall. It is often difficult to secure research funding from certain donors. Many Canadian NGOs rely heavily on negotiated contracts with CIDA, which leaves little time and place for research. However, CIDA does fund some research (in a broad sense), particularly formative and summative evaluations. These are encouraging trends, but the aim should be for bilateral agencies to openly fund some research, like in the UK and some Scandinavian countries. Lack of interest or of a clear view of the whole research process can also be considered as impediments. As several NGOs do not see research as part of their mandate, they may not be willing to get involved in research: " NGOs do not have a research mandate, and therefore we do not foresee developing research expertise in-house. Linking for instance with universities is feasible for development-driven research " (Interview with R. Hazel, CECI). NGOs may have to change their structures and priorities in order to support autonomous research. 4.3.3 Scale and type of NGO research Because NGO research is often conducted on a small scale and is usually of a qualitative nature, it often goes unrecognized by governments, and even by research organizations and funding agencies, which tend to favour large scale quantitative research. NGOs interested in pursuing a research profile require a type of mentorship in terms of standard performance indicators in the research domain. For example, publication has traditionally not been a strength and much NGO research does not reach beyond the gray literature or report level. Because scientific publications are an important means of transfer or dissemination of research results, NGO capacity to publish their findings needs to be strengthened. 4.3.4 Weak links with the international research community There is not enough networking and collaboration between NGOs and the international research community, including academia. This has traditionally been due to a dichotomy in the interests of NGOs and the academic community, in that NGOs are more oriented towards a development agenda, while academics tend towards special research interests. 5 Future needs In light of the above issues and concerns, and in order to foster greater interest and participation of NGOs in research, the barriers of lack of interest, lack of funds, lack of training and lack of recognition, among others, need to be addressed. We discuss some strategies below. 5.1 Opportunities to build NGO capacity in research, in Canada and overseas Substantial global health academic capacity has developed within NGOs both in the U.S. and the U.K. For example, Family Health International (FHI) has integrated research, training, and development capacity on an evidence-based foundation. It also embodies many of the competitive aspects of private sector-led initiatives that can allow creativity and innovation. As emphasized by Harrison and Neufeld [ 31 ], however, capacity building efforts for health research have been of most benefit to industrialized countries. In order to ensure that less developed countries are the principal beneficiaries, they recommend, as part of a three-pronged strategy, the nurturing and support of multi-stakeholder problem-oriented learning, and research networks which include NGOs. The other components of the strategy are research investments that explicitly reduce the high cost of knowledge translation in developing countries, and the stimulation of demand-driven research. A peer-learning process is among the strategies for NGO capacity building in research. NGOs can draw on expertise already developed in research-based NGOs. The learning process should be shared with NGOs from the North and from the South. Additionally, NGOs should consider taking the initiative in organizing scientific activities (seminars, workshops, symposia) on global health research topics, which could serve as a catalyst in bringing together different stakeholders. 5.2 Building partnerships and alliances 5.2.1 Creating and facilitating networks that support global health research The creation of networks which have the common goal of supporting global health research is one way to strengthen partnerships and to consolidate valuable resources from each partner. Leadership and governance issues are necessary hurdles which can be overcome by focusing on the ultimate gains in terms of supporting and conducting successful research activities. NGOs can assist in the establishment and functioning of these networks, particularly by providing stable infrastructure support. One of the greatest challenges is in making the network function effectively through different leadership turnovers in the different partner organizations. CSIH has had global health research as part of its mandate since its formation. CSIH is an active member of the Canadian Coalition for Global Health Research (CCGHR), and a key challenge in this capacity is to develop a strong foundation of understanding and mutual respect amongst all players in global health research, including NGOs. CSIH experience in global health research is described in table 7 . Table 7 The experience of CSIH in global health research In the early years, CSIH partnered with the Canadian University Consortium for Health In Development, which represented all the major universities in Canada and their partners in research and development. This partnership represented a strong and vital part of CSIH's operations. Following the decrease in funding for such a partnership, there was a decision to disband the Consortium and establish a network of universities and colleges that would promote and support academic and research interests within the Society. This network, which was formally given the status of a Division for a few years, has been and continues to be functional but not as a strong advocacy unit. This was largely due to the fact that funding for the network was cut by CIDA in 2000. Nevertheless the network is an important source of technical support for CSIH in its projects and advice. Following the Thailand meeting in October 2000, Canadians were challenged to explore the role that they could play in diminishing the 10/90 Gap in Global Health Research funding available to Low and Middle Income Countries (LMIC). To this end an interest group was formed, of which CSIH was part in order to carry on the momentum of Thailand and future explorations. Key people met with decision makers during the spring and summer of 2000 and September 11, 2001, marked the inaugural workshop in Vancouver to discuss global health research and the 10/90 Gap. CSIH was one of two NGOs who attended. Following that meeting, CSIH was invited to participate in a new Canadian Coalition for Global Health Research (CCGHR). CSIH was active in suggesting that the concept of a coalition was a way to emphasize the role of advocacy and action that is necessary for global health research initiatives to be successful. As of October 2001, the Coalition included two NGOs (CSIH and CPHA) who were part of a lobby to expand the mandate of CIHR (Canadian Institutes of Health Research) to include global health research in more than one Institute. To this end, the Institutes of Gender and of Aboriginal Health joined the Institute of Population Health in realizing its global health mandate. The Global Health Research Initiative memorandum of understanding (MOU) was signed at the 2001 Canadian Conference for International Health (hosted by CSIH). The amendment to MOU as a result of negotiations between CSIH and CIHR included NGOs as one of the important players. The first formal retreat for the coalition was held in August 2002. CSIH was formally named as a member of the Coalition Steering Group. The Working Group on the Role of NGOs in Research was affirmed as separate from the Advocacy Working Group. CSIH agreed to take the lead to collaborate with other key NGOs to develop a paper and case studies. CSIH as part of CCGHR lobbied in the spring and summer of 2002 to the G-8 for the inclusion of a commitment to global health research within NEPAD (New Partnership for Africa's Development). Support for global health research in Africa was announced and funds were set aside for this new initiative. The first Annual Meeting of CCGHR was held at the Canadian Conference on International Health (CCIH) in October 2002. The Working Groups reported at that meeting and CSIH announced the formation of a Research Committee and invited its members to participate. The Executive Director drafted an outline of a background paper on the Role of NGOs in Global Health Research and presented to the plenary session of the annual CCGHR meeting for comment and feed-back. An ad hoc Working Group on Research was formed to draft the background paper with a view that it will be a position paper for CSIH and provide a background working paper for the Coalition. In the autumn of 2002, the first request for proposals for global health research grants was released. Despite the fact that NGOs were named as important partners, they were not invited to be part of the review panel for this round. It was noted as a deficiency in the review of the process by CIHR. To date, CSIH has been an active and welcomed participant of all key meeting of CCGHR Steering Committee meetings. CSIH remains actively engaged in working groups on Governance to determine options for institutionalization of the CCGHR. In collaboration with CIHR and IDRC, CSIH is actively planning the Second Annual Global Health Research Meetings at CCIH and the integration of significant research and development content in the conference. 5.2.2 Partnerships with universities and other research institutions Partnerships with universities and other research institutions is one means of strengthening the research capacity of NGOs, and also of academia. NGOs and research organizations each have a unique 'value-added' contribution to make to global health research and therefore, partnering among them amplifies their individual strengths. Such partnering may be a real challenge for NGOs, however, as their institutional culture is so different. NGOs may be invited by universities to partner, but plans are often already laid out, so that the NGOs may only be involved in executing the plans. What NGOs want is to be part of the research process from the start. An interesting initiative to document and promote South-Canada health research partnerships is currently underway [ 37 ]. NGOs are open to partnerships with academia, but the goal has to be development-oriented. Experience suggests that it is often difficult to reconcile the academic and development framework, for instance when integrating MSc or PhD students in development projects. Nonetheless, integrating graduate students in NGO projects should be a good strategy for balanced and equal partnerships. As less than 1% of university-based health research in Canada is directed towards the problems of global health according to a Canadian Consultation on Global Health Research held in 2001 , the prospects of university-NGO research links are constrained by funding. Nevertheless, the recent Global Health Research Initiative of the coalition of Canadian institutions funding global health research is promising as it opens new avenues for research collaboration between the North and the South, and hopefully also between universities and NGOs. During the 1990s, there were several attempts to bridge the NGO/academic gap with respect to health research in developing countries. Save the Children UK, Oxfam and some US-based institutions supported workshops and symposia that aimed at bringing together representatives from both communities as a means of building links and forging partnerships in support of increasing the scale, scope and relevancy of health research in developing countries (Edwards M, Griffiths M: Terms of Reference for the DSA [Development Studies Association] Workshop on the Academic Practitioner Interface. London: DSA, 1994). CPHA, through the CIDA-funded Canada's International Immunization Program – Phase 2 (CIIP2), supported applied research carried out by Laval University, Université de Montreal and University of British Columbia. At the time, these were quite innovative approaches to applied health research, linking the universities with local NGOs. In 1995 CPHA also organized a Symposium on NGO/University Linkages for Health Research [ 30 ]. One mechanism to expand the use of research generated by NGOs is to improve the linkage between NGOs and universities. Each complements the other in the area of health research. NGOs offer proximity to people and situations, reality-based and context-specific research environments, opportunities to develop and assess innovative strategies and research methods, a means of disseminating and popularizing the results of research projects, and credibility outside of academia. Universities and other research organizations offer expertise in research design and application, an environment for reflection, access to and knowledge about most recent literature, a tradition of scientific rigor and interest in new, innovative research methods and approaches, and a high degree of credibility. Academics can also provide guidance and advice on how to prepare research proposals and to carry out research studies, guidance in the preparation of reports and publishing of research results, and training for NGO staff in research methods. The participants of the CPHA Symposium on NGO/University Linkages for Health Research in Developing Countries [ 30 ] identified several mechanisms that could help bridge the gap between NGOs and universities as a means of facilitating future collaborative research initiatives. There must be first and foremost a real willingness on the part of both parties to modify their attitudes about the role and capabilities that each can offer. Mechanisms to achieve this end include conferences and seminars, newsletters, and the use of e-mail and the Internet. Another suggestion called for the use of "field-friendly" research methodologies. It was also recommended that, although the objective is not to transform NGOs into research institutes, they should receive more training in research methods and proposal development. It was suggested that exchanges take place wherein university researchers use sabbatical leave to work with NGOs and NGO personnel be seconded to universities to provide a field perspective. Additionally, research results need to be disseminated quickly and in a format that ensures maximum access by those in the field who are to apply the knowledge generated. Otherwise, research creates expectations within the NGO community and study population that remain unsatisfied. The development of innovative North-South research partnerships is the focus of a working paper prepared for the CCGHR by Neufeld et al [ 37 ]. As emphasized in this document, such partnerships are not an end in themselves, but rather, they are to contribute to sustainable health research systems and to health development. Principles of research partnerships, and a useful model to assess these, are proposed. Although types of partnerships are not specifically detailed, it is implied that NGOs are important research partners. Finally, as mentioned earlier, stronger partnerships between NGOs and social science researchers in particular should be sought in resource-poor countries. Lessons learned from these partnerships, in the areas of action and indeed policy and legislation (for example, in the tobacco or environment fields) show how evidence can be transformed into action with the right partnerships between researchers and NGOs. 5.2.3 An NGO Network for Global Health Research? NGOs may benefit in various ways from developing a global health research network. First, many NGOs already operate at national and international levels and understand the challenges of coordination and communication which this entails. There is a need to identify overarching principles of NGO contribution to health research. Second, NGOs must be both proactive and interactive within the framework of the health research agenda. Roles must be clearly understood by each partner. In any case, advocacy for relevant research and use of results would be a critical function of the network (see training modules on advocacy [ 38 ]). Such networks may enhance the ability of NGOs to partner with other research stakeholders in multisectoral coalitions, and even to initiate partnerships with research organizations. In the framework of an NGO network of this sort, the following discrete activities could be envisaged by the lead NGO: • To invite NGOs to post on a selected website success stories, as well as their experience/opinions/needs/priority research issues, using a template adapted from the one developed for this purpose in the UK • To organize workshops for NGOs who are, or who wish to become, involved in research, with research organizations where deemed appropriate. The purpose would be: • To link these NGOs in order for them to interact on research issues; • To share lessons learned and success stories of research involvement of NGOs and their partners • To enhance understanding of, and collaboration with, potential research partners; • To set-up a core group of NGOs involved in global health research to convey NGO views to global health research fora and organizations. The following are a few key questions that could be addressed by an NGO network: 1. How can NGOs contribute to the framing of the research questions if we were to support the necessary equity-based research for improving the overall performance of the health system? 2. How do we balance this with the necessity for research which documents and monitors sustained and emerging inequities which may have a greater impact on the health and well being of individuals than the health (care) system will ever have? 3. How can we ensure that NGOs influence research priorities so that they are reflective and evaluative of overseas development assistance (ODA) direction and priorities such as national poverty reduction strategies. For example what is the impact of PRSPs on equity and health? How will researchers monitor this? What could be the potential role of NGOs in partnership with researchers to begin to monitor and evaluate this new direction in overall aid policy? 4. How can NGOs be best represented within the international research community? The new millennium offers many challenges. In order to maximize the potential benefits of health research, all partners including NGOs must share a common vision and recognize and appreciate the strengths of each. Participation in health research needs to be a coordinated effort. One key challenge will be to establish better communication among all partners in health research. This can only be achieved by a willingness to share in leadership, ownership and in the conduct of health research activities. Another key challenge will be to explore ways in which funding for health research can be strengthened. Leveraging must be seen as a strategic tool of NGOs to maximize dollars allocated to health research. Conclusion Several NGOs have had impressive track records in global health research. Other NGOs have expressed an interest in becoming more involved in global health research. Their contribution to more equitable, ethical, relevant and effective research is crucial and needs to be strengthened. Research has to be regarded as a broad loop system rather than restricted narrowly to the production of knowledge. This is particularly critical for global health research whose primary goal should be to improve health and its determinants in low and middle income countries. NGOs principal roles in the process pertain to shaping research priorities, advocacy for more relevant research, translating and using research findings, in addition to generating new knowledge in areas where they may have a comparative advantage, notably qualitative, social, action, evaluative, and policy research. NGO partnerships with research organizations should be seen as means of a mutual enhancement of health research capacity and contribution to development. NGOs should be instrumental in building with other stakeholders coalitions for global health research with the aim of closing the 10/90 health research gap. List of Abbreviations ADI Alzheimer's Disease International AFRO-NETS African Networks for Health Research and Development AIDS Acquired immunodeficiency syndrome AMREF African Medical and Research Foundation ARCH Applied Research on Child Health (ARCH) Project CCIC Canadian Council for International Cooperation CCISD Centre de coopération internationale en santé et développement CCIH Canadian Conference on International Health (hosted by CSIH) CECI Centre canadien d'étude et de coopération internationale CCGHR Canadian Coalition for Global Health Research CDC Centers for Disease Control and Prevention CHW Community health worker CIDA Canadian International Development Agency CIHR Canadian Institutes of Health Research CIIP2 Canada's International Immunization Program – Phase 2 COHRED Council on Health Research for Development CPHA Canadian Public Health Association CSIH Canadian Society for International Health DFID Department for International Development (UK) DSA Development Studies Association ENHR Essential National Health Research FCTC Framework Convention on Tobacco Control (WHO) FHI Family Health International FRAT Fortification Rapid Assessment Technique GYTS Global Youth and Tobacco Survey (WHO and CDC) HKI Helen Keller International IDRC International Development Research Centre IHPP International Health Policy Program INTRAC International NGO Training and Research Centre KFPE Commission for Research Partnerships with Developing Countries LMIC Low and middle-income countries MICAH Micronutrients and Health in Africa (WV project) MSF Médecins sans frontières NEPAD New Partnership for Africa's Development NGO Non-governmental organization OECD Organization for Economic Co-Operation and Development ODA Overseas Development Assistance PRSP Poverty Reduction Strategy Papers RITC Research for International Tobacco Control RPHA Russian Public Health Association SHARED Scientists for Health and Research for Development TRIPS Trade-Related Aspects of Intellectual Property Rights UNDP United Nations Development Programme WHO World Health Organization WV World Vision (Canada) Authors' contributions HD designed the outline of the paper, conducted interviews with NGO representatives, wrote the first complete draft, and coordinated the review process within the CSIH. JHR drafted some sections and provided comments on the successive versions of the papers. MM provided international NGO and a field based perspectives to the paper, in addition to conducting group discussions with NGO personnel. LJ designed the figures, and edited and formatted the text. TG contributed to the conceptualisation and content of the manuscript. She wrote several sections and edited others, and she provided case studies.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC554095.xml

539293

Metabolic scaling: consensus or controversy?

Background The relationship between body mass (M) and standard metabolic rate (B) among living organisms remains controversial, though it is widely accepted that in many cases B is approximately proportional to the three-quarters power of M. Results The biological significance of the straight-line plots obtained over wide ranges of species when B is plotted against log M remains a matter of debate. In this article we review the values ascribed to the gradients of such graphs (typically 0.75, according to the majority view), and we assess various attempts to explain the allometric power-law phenomenon, placing emphasis on the most recent publications. Conclusion Although many of the models that have been advanced have significant attractions, none can be accepted without serious reservations, and the possibility that no one model can fit all cases has to be more seriously entertained.

Introduction: Kleiber and metabolic scaling In 1932, Kleiber published a paper in an obscure journal [ 1 ] showing that standard metabolic rates among mammals varied with the three-quarters power of body mass: the so-called "elephant to mouse curve", termed "Kleiber's law" in this review. Since that date, this and similar allometric scaling phenomena have been widely and often intensively investigated. These investigations have generated continuing debates. At least three broad issues remain contentious, each compounded on the one hand by the problem of obtaining valid data (in particular, finding procedures by which reliable and reproducible measures of standard metabolic rate can be obtained, especially in poikilotherms) and on the other by statistical considerations (in particular, the validity of fitting scattered points to a straight line on a semi-logarithmic plot). The first issue is disagreement as to whether any consistent relationship obtains between standard metabolic rate and body mass. Moreover, those who acknowledge such a relationship hold divergent opinions about its range of application. Is it valid only for limited numbers of taxa, or is it universal? Since the 1960s there has been a measure of consensus: a consistent allometric scaling relationship does exist, at least among homoiotherms. Nevertheless, not all biologists agree, and scepticism is widespread, particularly about the alleged universality of Kleiber's law. Second, assuming that some version of Kleiber's law (a consistent metabolic scaling relationship) applies to at least some taxa, there are disagreements about the gradient of the semi-log plot. That is, if B = aM b , where B = standard metabolic rate, M = body mass, and a and b are constants, what is the value of b ? Kleiber [ 1 ] and many subsequent investigators claimed that b = 0.75, and on this matter too a measure of consensus has obtained since the 1960s. Once again, however, not all biologists agree. A significant minority of investigators hold that b = 0.67; and other values have been suggested, at least for some organisms. Third, assuming a consistent scaling relationship and an agreed value of b , how is Kleiber's law to be interpreted mechanistically? What is its physical or biological basis? For those who claim that b = 0.67, this issue is simple: standard metabolic rate depends on the organism's surface to volume ratio. But for proponents of the majority view, that b = 0.75, the issue is not simple at all. Many interpretations have been proposed, and since several of these are of recent coinage and seem to be mutually incompatible, a critical comparative review seems timely. Kleiber's initial paper [ 1 ] found support within a decade. The allometric scaling relationship B = aM b (B = standard metabolic rate, M = body mass, a and b are constants and b is taken to be approximately 0.75), was inferred by other investigators during the 1930s [ 2 , 3 ]. Relevant data have been reviewed periodically since then (e.g. [ 4 - 15 ]) and recent developments have rekindled interest in the field. Many biological variables other than standard metabolic rate also reportedly fit quarter-power scalings (relationships of the kind V = kM b , where V is the variable in question, k is a constant and b = n/4; n = 3 for metabolic rate). Examples include lifespans, growth rates, densities of trees in forests, and numbers of species in ecosystems (see e.g. [ 9 ]). Some commentators infer that Kleiber's law is, or points to, a universal biological principle, which they have sought to uncover. Others doubt this, not least because it is unclear how (for example) tree densities can be consequences of metabolic scaling or can have the same mechanistic basis. This article focuses on the metabolic rate literature, mentioning other variables only in passing, because most debates in the field have arisen from metabolic rate measurements. Variations in the value of b Most debates about the value of b assume some version of Kleiber's law: i.e. that a single allometric scaling relationship fits metabolic rates over a wide range of organisms. However, as noted in the introduction, there are dissenters. Everyone acknowledges considerable variation both within and among taxa, no matter whether b = 0.75, 0.67, or some other number. The question is whether these variations are deviations from a general law, or whether there is no such law. Conflicting opinions on this fundamental point recall the traditional philosophical difference between physicists and biologists: the former are inclined to see abstract mathematical generalities in any set of numerical data, the latter to see concrete particulars. All recent attempts to explain Kleiber's law by "universal" models have involved physicists and mathematicians; the sceptics are predominantly biologists. Dodds et al. [ 16 ] re-examined published scaling data from Kleiber's original paper onwards and concluded that the consensus ( b = 0.75) was not statistically supported. Feldman [ 17 ] found no evidence for any wide-ranging allometric power law in biology and dismissed all attempts to explain scaling relationships by physical or mathematical principles. Atanasov and Dimitrov [ 18 ] found evidence that b ranges from around 0.67 to more than 0.9 over all major animal groups, the values perhaps reflecting complexity of organisation; single values such as 0.75 emerge only as averages over each group. Other investigators have been less sceptical; publications by Enquist and Niklas [ 19 , 20 ] give particularly impressive support to the generality of Kleiber's law because Niklas was previously among the doubters. Whatever one's position, it is indisputable that the Kleiber relationship has many exceptions, even among mammals. Bartels [ 21 ] showed that some mammals, such as shrews, have B values well above those expected from the Kleiber curve. Andersen [ 22 ] discussed the high B values for whales and seals and attributed them to the cold-water habitat. Nevertheless, Kleiber's law has been extended beyond placental mammals to birds and marsupials. Birds have generally higher a values than placental mammals and marsupials have lower ones, but the 0.75-power relationship is still inferred by many investigators (e.g. [ 4 ]). McNab [ 13 ] accepted Kleiber's law as a general approximation but emphasized species variations, which he attributed to differences in diet, habitat and physiological adaptation. Elgar and Harvey [ 23 ] also found variability among groups of species but reasoned that standard metabolic rates vary taxonomically rather than with temperature regulation, food intake or activity. Economos [ 24 ] was also critical of McNab, at least in respect of mammals. It is difficult to define "standard metabolic rate" in poikilotherms; ambient temperature, time since last meal and other variables markedly affect measurements [ 9 , 13 , 25 ]. A heterogeneous array of poikilotherm data [ 5 ] revealed an "average" b value of roughly 0.75. There were wide divergences in some taxa; notwithstanding these, Hemmingsen [ 4 , 5 ] argued that over all animals, plants and protists, metabolic rate scales as the 0.75-power of body mass. More recently published data [ 26 , 27 ] support this conclusion for a wide range of organisms and body masses. However, a careful re-evaluation of Hemmingsen's data by Prothero [ 28 ] cast further doubt on the applicability of Kleiber's law to unicellular organisms. Scepticism persists, mostly on the grounds of the intrinsic variability of the data, which is too often underestimated because it is disguised in the customary logarithmic plots and is seldom subjected to adequate statistical analysis [ 11 , 29 ]. However, this too has been debated; a suitable choice of procedures for estimating parameters might eliminate inconsistencies and discrepancies from the data, giving more credence to the belief that b = 0.75 over a wide range of taxa [ 30 ]. In the following section we shall examine some of the more divergent data in more detail. In short, there is a clear but by no means total consensus that (i) Kleiber's law is widely (even universally) applicable in biology, (ii) b is approximately 0.75. Variability in the data is generally admitted, so the consensus – and the claim that Kleiber's law manifests a general biological principle – can legitimately be doubted. The mass transfer model [ 31 ] Some of the doubts about the consensus are powerfully supported by studies on small aquatic organisms. Reviewing a large literature on metabolic rates in aquatic invertebrates and algae, Patterson [ 31 ] deployed chemical engineering principles to explain why the b values ranged from about 0.3 to 1.2 in these taxa (his Table 1 provides an excellent summary). Assuming that the delivery of nutrients to each organism entails diffusion through a boundary layer, Patterson showed how water movements and organism size might affect such delivery and hence determine metabolic rate. Using simple geometrical models of organisms (plates, cylinders and spheres), he derived b values ranging from 0.31 to 1.25, more or less consistent with the experimental values. Patterson plotted two dimensionless numbers against each other, viz . Sherwood number, Sh = h m W/D, where h m = mass transfer coefficient, W = characteristic dimension of organism and D = diffusivity; and Reynolds number (a function of organism size), Re = ρUW/μ, where ρ = density, U = water flow speed and μ = coefficient of viscosity. The graphs, which had the form Sh = c.Re d , where d = 0.5 for ideal laminar flow and 0.8 for turbulent flow (c is a constant of proportionality), revealed the relative importance of diffusion and mass transfer (convective movement) in the supply of materials. Patterson was able to derive an expression for h m , and was thus able to relate the supply of materials to body mass. The two main attractions of this model are (1) good agreement with a wide range of data and (2) derivation from basic physical principles without ad hoc biological or other assumptions. Patterson's approach has implicit support in the literature: Coulson [ 32 ] used chemical engineering principles to argue that mammalian metabolic rates are supply-limited, but he did not develop the argument in mathematical detail. However, Patterson's model has drawbacks. First, it is hard to see how his reasoning can be generalised to other taxa, notwithstanding Coulson's proposal (discussed in a later section). Second, by focusing on diffusion and convective mass transfer, he ignored active processes in the uptake of materials, which are likely to dominate in many organisms. Third, he assumed that metabolism in general is supply-limited; in homoiotherms at least, it is more nearly demand-limited under resting conditions, though even this is an oversimplification. The Patterson model has not been given much attention by other investigators in the field and perhaps it deserves more consideration. Despite its inherent limitations (it is exclusively concerned with small aquatic eukaryotes) it is a potentially fruitful contribution to biophysics. Scaling of metabolic rate with surface-to-mass ratio Several workers accept the reality of allometric scaling but question the value b = 0.75, which a consensus of physiologists has accepted since the 1960s. Many of these sceptics claim that the "true" value of b is 0.66 or 0.67 because the principal determinant of metabolic scaling is the surface-to-volume ratio of the organism; hence, assuming constant body density, the surface-to-mass ratio. The first study to suggest this explanation for the mass dependence of B is attributed to Rubner [ 33 ], who studied metabolic rates in various breeds of dog. Heusner [ 34 ] reported that b is approximately 0.67 for any single mammalian species and suggested that the interspecies value of 0.75 is a statistical artefact. Feldman and McMahon [ 35 ] disagreed, but Heusner sustained his position in subsequent articles. For instance, reviewing a substantial body of published data [ 36 ], he argued that metabolic rate data for small and large mammals lie on parallel regression lines, each with a gradient of approximately 0.67 but with different intercepts (i.e. values of a , termed the "specific mass coefficients"). Hayssen and Lacy [ 37 ] found b = 0.65 for small mammals and b = 0.86 for large ones, again suggesting that b = 0.75 is a cross-species "average" with no biological significance; but it is questionable whether their data were measurements of standard metabolic rate in all cases. McNab [ 13 ] reported lower values: 0.60 and 0.75, respectively. Heusner [ 36 ] reasoned that if a few large mammals are added to a sample of predominantly small ones, a single regression line for all the data might have a gradient around 0.75. This, however, is misleading, as the following paragraphs will argue. According to Heusner, the ratio B/M 0.67 is a mass-independent measure of standard metabolism. Variations indicate the effects of factors other than body mass. Other workers broadly share Heusner's opinion (see e.g. [ 12 ] for review and [ 38 ] for a good recent exemplar). Bartels [ 21 ] found a value of 0.66 for mammals; Bennett and Harvey [ 39 ] reported 0.67 for birds. Of course, if B varies as M 0.67 , the interesting problem is not the index ( b ) in the Kleiber equation but the allegedly constant relationship between specific mass coefficient ( a ) and body size. This point was developed by Wieser [ 40 ], who distinguished the ontogeny of metabolism, which comprises several phases but follows the surface rule (M 0.67 ) overall, from the phylogeny of metabolism, which concerns the mass coefficients ( a ). Following Heusner's argument, Wieser [ 40 ] wrote the allometric power law in the form B = a n M 0.66 and deduced that the specific mass coefficient a n = aM 0.09 . Here, a is an interspecific mass coefficient (3.34 w in mammals if M is in kg). Another difficulty with this type of explanation lies in the calculation of body surface area; the Meeh coefficient, k, where surface area = kM 0.67 , is difficult to measure unequivocally but is generally taken as ~10 (see [ 3 ]). Yet another possible difficulty was identified by Butler et al. [ 41 ], who questioned Heusner's dimensional analysis argument and concluded that no version of Kleiber's law (i.e. no value of b that is constant over a range of species) could be substantiated by his approach. The claim that b = 0.67 remains a minority view. Those who accept it are faced with the twin difficulties of (i) establishing that their estimates of surface area are correct and (ii) explaining why, in Wieser's notation, a n = aM 0.09 . Moreover, even if such arguments as Heusner's are valid for homoiotherms, it is hard to justify their extrapolation to poikilothermic animals, plants and unicellular organisms, all of which are held by consensus to fit Kleiber's law (but see the two preceding sections). Why should temperature fluxes across the body surface be the main determinants of metabolic rate in poikilotherms, particularly microorganisms? Even in mammals, maintenance of body temperature might not be the main contributor to energy turnover at rest (see later). Contrary to the view of Dodds et al. [ 16 ], therefore, b = 0.67 cannot be treated as a "null hypothesis". Throughout the remainder of this article, the consensus position will be assumed: Kleiber's law is valid for a wide range of organisms, and b = 0.75. This assumption is made tacitly and provisionally and does not imply dismissal of the foregoing sceptical arguments; but a field can only be reviewed coherently from the consensus point of view. McMahon's model [ 42 ] A vertical column displaced by a sufficiently large lateral force buckles elastically. The critical length of column, l cr , = k(E/ρ) 1/3 d 2/3 , where d = column diameter, E = Young's modulus and ρ = density. If E and ρ are constant then l cr 3 = cd 2 , where c is a constant of proportionality. McMahon [ 42 ] applied this reasoning to bone dimensions for stationary quadrupeds. In a running quadruped the limbs support bending rather than buckling loads but the vertebral column receives an end thrust that generates a buckling load. It follows that all bone proportions change in the same way with animal size. The mass of a limb, w l , = αld 2 , where α is a constant. If w l is proportional to M, as it generally must be, then M = βld 2 , where β is another constant. Hence (given the above relationship between l and d) M is proportional to l 4 , implying that l is proportional to w l 1/4 ; hence d is proportional to w l 3/8 , or M 3/8 . Empirical support for this relationship appeared in [ 43 ]. McMahon [ 42 ] also applied this argument to muscles. The work done by a contracting muscle, W, is proportional to σAΔl, where σ is tensile strength, A is the cross-sectional area and Δl is the length change during contraction. The power developed, W/t (t = time), is therefore σAΔl/Δt. Since σ and Δl/Δt are roughly constant and independent of species, W/t varies with A; and since A is proportional to d 2 , W/t it is proportional to d 2 , and therefore to (M 3/8 ) 2 = M 3/4 . If this deduction applies to any skeletal muscle (as seems plausible), then it applies to the entire set of metabolic variables supplying the muscular system with nutrients and oxygen. Hence, B varies as M 3/4 . A broadly comparable but simpler argument was advanced by Nevill [ 44 ]; large mammals have proportionately more muscle mass than smaller ones. If the contribution of the muscle to B (which Nevill assumes is proportional to M) is partialled out, then the residual B is proportional to M 2/3 . Nevill's paper is seldom cited. One difficulty with McMahon's model is that little of the energy turnover under conditions of standard metabolic rate measurement entails muscle contraction. The model might still be valid if maximum metabolic rate followed the same allometric scaling law as B; this has been widely believed, and Taylor et al. [ 45 ] adduced evidence for it. However, recent detailed studies [ 46 - 48 ] indicate that maximum metabolic rate in birds and mammals scales as M 0.88 , not M 0.75 , although there are disagreements about whether aerobic capacity determines the allometry of maximum metabolic rate [ 48 , 49 ]. Weibel [ 50 ] presented a large set of data to this effect. (On the other hand, there are reports that in birds the index decreases rather than increases with increasing metabolic output, e.g. [ 58 ].) Another drawback of the McMahon model is that it cannot apply to organisms without muscles, such as protists. This perhaps explains why McMahon's elegant deduction has been largely ignored in recent debates about Kleiber's law. The Economos model [ 51 ] An increased gravitational field increases energy metabolism in animals [ 52 , 53 ]. Work against gravity is proportional to M 1.0 . If maintenance metabolism were related to surface area (proportional to M 0.67 ) then a combination of the two effects, surface-to-mass ratio and work against gravity, might explain the observed M 3/4 relationship. This model [ 51 ] is difficult to assess: it is not clear why the two proposed factors, surface area dependence and gravitational loading, should combine for all animals (and other taxa) in just the right proportions to generate a 0.75-power dependence on body mass. To take just one example, aquatic microbes are more affected by Brownian motion than by gravity, so why should they show the same balance between surface-to-mass ratio and gravitational effects as mice or elephants? Pace et al. [ 54 ] suggested that the Economos model could be critically tested under conditions of weightlessness in space. No corroboration (or refutation) by studies on astronauts has been reported. Allometric scaling in cells and tissues Before more recent models purporting to explain Kleiber's law are discussed, some comments are needed on scaling of metabolism at the organ, tissue and cell levels. Belief that the Kleiber relationship can be explained in terms of the inherent properties of the cells dates from the 1930s [ 3 , 55 ] and persists (e.g. [ 56 , 57 ]. Standard metabolic rate (B) is usually measured as oxygen consumption rate, which correlates with nutrient utilization [ 9 , 15 ] and rates of excretion of nitrogenous and other wastes [ 2 ]; so research in the field has been dominated by respiratory studies. Lung volume, trachaeal volume, vital capacity and tidal volume all scale as M but respiratory frequency varies as M -0.31 , ventilation rate as M 0.77 and oxygen consumption rate as M 0.72 [ 58 - 60 ]. All mammals extract a similar percentage of oxygen (~3%) from respired air [ 9 ]. The significance of "pulmonary diffusion capacity" has been debated; it scales as M 1.0 so it is disproportionate in bigger animals [ 17 , 61 - 65 ]. Stahl [ 60 ] described the scaling of cardiovascular and haematological data. Blood haemoglobin concentration is the same for all mammals except those adapted to high altitudes. Blood volume is ~6–7% of body volume for all mammals except aquatic ones. Erythrocyte volume varies with species but bears no obvious relationship to M. The oxygen affinity of haemoglobin varies with body size, being lower in smaller mammals, which unload oxygen to their tissues more rapidly. Capillary density is more or less constant in mammals with bodies larger than a rat's, though it is greater in the smallest mammals [ 65 ]. The heart accounts for ~0.6% of body mass in all mammals [ 66 ]. Heart rate scales as M -0.25 , cardiac output as M 0.81 (60) and circulation time as M 0.25 . The energy cost of supplying the body with 1 ml of oxygen is similar for all mammals [ 15 ]. Standard metabolic rate has two main components: service functions, e.g. the operation of heart and lungs; and cellular maintenance functions, e.g. protein and nucleic acid turnover (e.g. [ 67 ]). Krebs [ 68 ] elucidated this second component by studying tissue slices; his investigation has since been extended. Oxygen consumption per kg decreases with increasing M in all tissues, but tissues do not all scale identically. Horse brain and kidney have half the oxygen consumption rates of mouse brain and kidney but the difference between these species in respect of liver, lung and spleen is 4-fold [ 68 - 70 ]. Metabolic rate in liver scales as M 0.63 ; for some organs the exponent is closer to 1.0; the sum of oxygen consumption rates over all tissues gives – approximately – the expected 0.75 index [ 71 ]. The difficulty of recalculating B from tissue-slice data is considerable, so the Martin and Fuhrman calculation [ 71 ] has wide confidence limits. Spaargen [ 72 ] suggested that tissues that use little oxygen constitute different percentages of body mass in large and small mammals, leading to a distortion of the surface law (B = M 2/3 ), which would otherwise be valid. More recently, however, Wang et al. [ 73 ] repeated the Martin and Fuhrman calculation using improved data, and found impressive support for the consensus B = M 3/4 . Cells of any one histological type are size-invariant among mammals but allometric scaling is reported at the cellular level; e.g. the metabolic rate of isolated hepatocytes scales as M -0.18 [ 74 ]. Numbers of mitochondria per gram of liver (or per hepatocyte), however, scale as M -0.1 [ 75 , 76 ]. The apparent discrepancy between these values might be illusory ( cf. [ 77 ]), or it might indicate a greater proton leak in mitochondria from livers of smaller animals [ 78 ] or allometry in redox slip [ 79 ]. Also, larger animals have smaller inner mitochondrial membrane surface areas (the scaling is M -0.1 ) and different fatty acid compositions [ 71 ]. The discrepancy between the scalings of hepatocyte and whole-body metabolism is probably explained by the decrease in liver mass, which scales as M 0.82 [ 75 , 80 ]. Combining liver mass with hepatocyte oxygen consumption, the derived scaling for liver metabolism is M 0.82 .M -0.18 = M 0.64 , consistent with the experimental tissue-slice data (M 0.63 ; see above). Combining liver mass with mitochondrial number per hepatocyte gives a similar value [ 77 ]. Cytochrome c and cytochrome oxidase contents scale roughly as M 0.75 [ 81 - 85 ]. The allometric scaling of mitochondrial inner membrane area, and the body-size-related differences in unsaturated fatty acid content, remain unexplained. Isolated mammalian cells reportedly attain the same mitochondrial numbers and activities after several generations in culture, irrespective of the tissue of origin or the organism's body mass [ 86 - 88 ]. If allometric scaling is lost at the cellular level after several generations in vitro , then presumably mitochondrial densities, inner membrane areas and cytochrome levels somehow become "normalized". This is a readily testable prediction [see [ 89 ]], but it does not appear to have been subjected to critical experiments. If it is corroborated there will be interesting mechanisms to investigate. The main conclusions from this section are: (a) different organs make different contributions to the scaling of whole-organism metabolic rates; (b) differences at the cellular level make relatively small contributions to scaling at the organ level; (c) these differences at cellular level might disappear altogether after several generations in culture. The most striking conclusion is (b). It implies that allometric scaling of metabolic rate does not after all, for the most part, reside in cellular function but at higher levels of physiological organisation. If this is the case, then the alleged applicability of Kleiber's law to unicellular organisms is called into question. Resource-flow models Coulson's flow model [ 42 ] was mentioned earlier. It relates tissue or organ oxygen consumption rates to circulation times, i.e. to the rate of supply of oxygen and nutrients, and these scale as M 0.25 (see previous section). Coulson's approach contrasts with traditional biochemical measurements: the principal variable is not the concentration of a resource but the supply rate ; metabolic activity depends on encounter frequency not concentration . This perspective merits further development, particularly by extension to the cell internum [ 89 - 93 ]. Obviously, it is within the cell that the reactant molecules are passed over the catalysts; and the flow rate increases with the cell's metabolic activity, as Hochachka [ 93 ] cogently described. However, flow theories advanced to explain Kleiber's law have not followed this line of argument. Banavar et al. [ 94 , 95 ] and Dreyer and co-workers [ 27 , 96 ] have shown that the Kleiber relationship can be deduced from the geometries of transport networks, without reference to fluid dynamics. Broadly, these authors argue that as a supply network with local connectivity branches from a single source (in a mammalian circulatory system, the heart is the source), the number of sites supplied by the network increases. Natural selection has optimized the efficiency of supply. A general relationship can be derived between body size and flow rate in the network: delivery rates per unit mass of tissue vary with the quarter-power of body size (M), implying the validity of Kleiber's law. The most detailed account of this argument [ 95 ] begins with the reasonable assumption that M scales with L D , where L is the physical length of the organism and D is its dimensionality. It proceeds with a theorem: the sum of flows through all parts of the network, F, is proportional to the (dimensionless) length multiplied by the metabolic rate. A quantity measuring the total flow of metabolites per unit mass of organism is then defined: r 1 = F/M. r 1 (which has units of inverse time) measures the dependence of the network's geometry on body mass, so it indicates the energy cost of metabolite delivery. Another parameter, r 2 , measures the metabolite demand by the tissues: r 2 = the dimensionless length of the "service volume" (the amount of tissue that consumes one unit of metabolite per unit time). It is then deduced that B is proportional to (Mr 1 /r 2 ) D/(D+1) . Provided that r 1 and r 2 change proportionately – i.e. supply always matches demand – then for a three-dimensional organism, Kleiber's law follows. According to Banavar et al. [ 94 ], deviations from Kleiber's law indicate inefficiency or some physiological compensation process. This model has been criticized [ 97 ] because the assumed network does not resemble (e.g.) the mammalian circulatory system, where only terminal nodes (capillaries), not all nodes (as the model implies), are metabolite exchange sites. Also, the model seems to predict that r 1 /r 2 will decrease as B rises from standard to maximal; but the best data suggest the opposite trend (see earlier discussion: [ 46 - 48 ]). Banavar et al. do not explicitly allow for differences among tissue types, which are considerable (see above), except perhaps in terms of rather implausible variations among r 1 /r 2 ratios. On the other hand, the model is simple and flexible and it reflects recent developments in the physics of networks. If it could be applied to flow at the cellular level, it might accord with the requirements discussed at the beginning of this section; though it is difficult to see how this can be achieved. Rau [ 98 ] also advanced a fluid-flow model, but his conception is physical not geometrical. Assuming Pouseille flow through an array of similar tubes, such as capillaries, and a roughly constant flow speed, Rau used scaling arguments to derive the relationship t = kM 1/4 , where t is the transport time and k is a constant. If the fluid transport rate (essentially the reciprocal of t) is proportional to B/M, Kleiber's law follows. However, Rau's model appears to assume that because metabolic rate is energy per unit time, it can be equated with the product of fluid volume flow rate and pressure (since energy is equal to pressure times volume). This assumption, which appears to be based exclusively on dimensional analysis, is fallacious. Four-dimensional models Blum [ 99 ] observed that the "volume" of an n-dimensional sphere of radius r is V = π n/2 r n /Γ(n/2 + 1), and that A = dV/dr = nπ n/2 r n-1 /Γ(n/2 + 1). Here, Γ(n) is the gamma-function such that Γ(n + 1) = n n , Γ(2) = 1 and Γ(3/2) = π 1/2 /2. Suppose two objects have "volumes" V 1 and V 2 and "areas" A 1 and A 2 . From the foregoing, A 1 /A 2 = (V 1 /V 2 ) (n-1)/n ; so if n = 4, a 3/4-power relationship between "volumes" (hence, masses?) emerges from a familiar mathematical principle. Might Kleiber's law therefore follow from a four-dimensional description of organisms? Speakman [ 100 ] pointed out that if n = 4, then A is volume (it has three dimensions) and V is hypervolume, the biological significance of which is obscure. However, West et al. [ 88 , 101 , 102 ] have indeed proposed a four-dimensional model to explain the Kleiber relationship, and considerable claims have been made for their account. This model addresses the supply of materials (particularly oxygen) through space-filling fractal networks of branching tubes. It assumes that as a result of natural selection, organisms maximize their use of resources. The initial account [ 101 ] assumed that energy dissipation is minimised at all branch-points in the network and that the terminal branches are size-invariant (for instance, blood capillaries are the same lengths and diameters in mice and elephants). Kleiber's law and analogous scalings were deduced from these assumptions. In particular, the three-quarters-power exponent was shown to be inherent in the geometry of a branching network that preserves total cross-sectional area at each branch point. The circulatory systems of large animals such as mammals are not exactly area-preserving, but West et al. [ 101 ] reasoned that this objection could be circumvented by considering the pulsatile flow generated in the larger arteries by the action of the heart. A second, simpler account [ 102 ] developed the model from a geometrical basis. The crucial feature of the branching network is the size-invariance of the terminal units. The effective exchange area, a , is a function of the element lengths at each level of the hierarchy, but one of these, the terminal one (l 0 ), is invariant. Writing Φ as a dimensionless function of the (dimensionless) ratio l 1 /l 2 leads to a (l 0 , l 1 , l 2 ,...) = l 1 2 Φ(l 0 /l 1 , l 2 /l 1 ...) Introducing a scaling factor, λ, leads to a (l 0 , l 1 , l 2 ,...) = λ 2 l 1 2 Φ(l 0 /λ l 1 , l 2 /l 1 ...) which is not proportional to λ 2 because l 0 is fixed. The dependence of Φ on λ is not known a priori , but it can be parameterized as Φ(l 0 /λ l 1 , l 2 /l 1 ...) = λ ε Φ(l 0 /l 1 , l 2 /l 1 ...), where ε is between 0 and 1. This power law reflects the fractal character of the network's hierarchical organization. Similar reasoning is applied to body volume, hence body mass, and the following expression for the exchange surface area is derived:- a = kM r , r = (2 + ε)/(3 + ε + ζ), where k is a constant and ζ (0 < ζ < 1) is an arbitrary exponent of length, just as ε is an arbitrary exponent of area. If natural selection has acted to maximize the scaling of a , then ε must tend to 1 and ζ to 0. This gives r = 0.75. If a limits the supply of oxygen and nutrients, and hence determines standard metabolic rate, then B is proportional to a and Kleiber's law follows. The model has several attractions: it derives from well-established physical principles, invokes natural selection and is mathematically impeccable. It implies that cells and organelles transport materials internally along space-filling fractal networks rather than by "diffusion", which seems correct [ 83 , 85 , 86 , 103 ]. The self-similarity of these transport networks is emphasized particularly in [ 88 ]. The dimensionalities of effective exchange surfaces, a , are predicted to be closer to 3 than 2; empirically, the microscopic convolutions of surfaces such as the mammalian intestinal mucosa are well known. The mass of the smallest possible mammal is deduced and shown to be close to the mass of the shrew. Other approaches to exchange networks, assuming minimum energy expenditure and scale-invariance, have led to similar models [ 104 ]. The model can be adapted, with no loss of rigour, to new data: Gillooly et al . [ 105 ] showed that the fractal supply network principle can be combined with simple Boltzmann kinetics to explain the effects of both body mass and temperature on metabolic rates. Since mass and temperature are the primary determinants of many physiological and ecological parameters, this work suggests that the model [ 88 ] could revolutionize biology. This is an impressive range of successes. However, West and his co-workers make claims that are less compelling. The observation that cytochrome oxidase catalytic rates fit the same allometric curve as whole-organism metabolic rates is claimed as corroboration. However, cytochrome oxidase is not an organism, or a cell: it does not have a metabolic rate. It is also debatable whether mitochondria can be said to have "metabolic rates". (In contrast, Hochachka and Somero [ 106 ] noted that oxygen turnover in the whole biosphere can be fitted to the same curve; but they recognized this as "a contingent fact with no biological significance".) Also, the explanation derived by West and his colleagues for the alleged body-mass-invariance of the metabolic rates of cultured cells (see earlier) is mathematically neat, but it leads to no experimentally testable predictions, and the heterogeneous data sources cited in this context make the explicandum itself unconvincing. Finally, the model is said to explain the quarter-power scalings of a wide range of biological variables other than metabolic rate, including population densities of trees [ 19 ] and carnivorous animals [ 107 ], plant growth rates, vascular network structure and maturation times [ 18 , 108 ], and life-spans [ 88 ]. It is not clear why any of these variables should depend on the fractal geometries of space-filling supply networks, still less on metabolic rates; though there is widespread interest in the application of scaling laws in ecology, for instance in modelling biodiversity [ 109 ] and food webs [ 110 ]. Moreover, there are definite flaws in the model:- (1) If West et al. were correct, maximal and standard metabolic rates should both scale as M 0.75 . The weight of evidence suggests that maximum rate in homoiotherms scales as M 0.88 (see earlier discussion [ 46 - 49 ] and following section). (2) During maximal energy output by an organism, the supply of material is likely to be limiting. For example, in mammals, muscle contraction is responsible for most of the energy turnover at maximum output and it is generally believed that the rate is limited by oxygen supply (if anaerobic capacity is ignored). However, under standard metabolic rate conditions, energy demand is generally more significant, i.e. for the service and cellular maintenance functions mentioned previously. Therefore, it is not clear why the geometry and physics of the supply system should predict the allometric scaling of standard rather than maximal metabolic rate. ("Supply" and "demand" under conditions of maximal aerobic metabolism are complex terms because many physiological steps are involved. The extent to which each step limits the maximum metabolic rate might be quantifiable by a suitable extension of metabolic control analysis [ 111 ]; this remains an active research area to which West et al. scarcely refer.) (3) The mathematical derivations given in [ 101 ] are idealisations, but they do not seem to allow for large deviations from b = 0.75. However, there are often wide differences among empirical b values, as discussed earlier; these were addressed in, for example, [ 18 ] and [ 31 ]. Also, the model does not account, or allow, for the differences in allometric scaling among mammalian tissues and organs [ 66 , 73 , 80 ]. (4) West et al. accept that some of their proposed hierarchical supply networks might be "virtual" (as in mitochondria) rather than explicit (as in mammalian blood circulation), but it is not clear why such networks must always have the same geometry. For instance, why should the intracellular network discussed by Hochachka [ 93 ] show area-preserving branching? There is no evidence that it does. Moreover, the "flow" of reductants through mitochondria presumably takes place in the plane of the inner membrane, which has one dimension fewer than (say) the mammalian circulatory system, so even if mitochondria can be said to have "metabolic rates", the 0.75-power law cannot apply here; yet, allegedly, it does apply. These difficulties show that the West et al. model, despite its impressive economy, elegance, consistency and range, cannot be accepted unreservedly in its present form. The very generality, or "universality", of this model has made it suspect for some biologists [ 25 ]. The implication that it reveals a long-suspected universal biological principle implicit in Kleiber's law has ensured its attraction for others [ 14 ]. The model of Darveau and co-workers [ 112 ] This group elaborated a multi-cause rather than a single-cause account of allometric scaling. Their "allometric cascade" model holds that each step in the physiological and biochemical pathways involved in ATP biosynthesis and utilization has its own scaling behaviour and makes its own contribution (defined by a control coefficient between 0 and 1) to the whole-organism metabolic rate. Thus, many linked steps rather than a single overarching principle account for Kleiber's law. This idea is inherently plausible, and the model is attractive because it draws upon recent advances in metabolic control analysis in biochemistry [ 111 ] and physiology [ 113 ]. It emphasises that standard metabolic rate is determined by energy demand, not supply; and it predicts an exponent for maximal metabolic rate in mammals between 0.8 and 0.9, rather than 0.75, which agrees with experimental findings [ 46 - 49 ] and the data cited by Weibel [ 50 ]. Implicitly – though the authors do not emphasize this – it seems capable of explaining b values that are far from 0.75 ( cf [ 31 ]). It is hardly surprising, therefore, that many responses to the Darveau et al. model have been positive [e.g. [ 114 ]]. However, Darveau et al. made no attempt to explain why the values of b are typically around 0.75, as West et al. and others have done. The model is phenomenological, not physical and mathematical; their equations are not derived from any fundamental principle(s). Moreover, their data cover only some three orders of magnitude of body mass, whereas many studies have involved much wider ranges. This might make their overall b values misleading [ 103 ] or, alternatively, more credible [ 18 ]. When their equations are applied to a mass range of eight orders of magnitude, different b values are obtained, not necessarily consistent with published data; but on the other hand, the published data might not be correct. In the first published account of this model [ 112 ] the mathematical argument was flawed. The basic equation was given in the form B = aΣc i M b(i) , where a is a constant coefficient, c I is the control coefficient of the i th step in the cascade and b(i) is the exponent of the i th step. By definition, the sum of all the c I values is unity. Darveau et al. did not derive this equation; they stated it. They also stated that the overall exponent, the b term in the Kleiber equation, is a weighted average of all the individual b(i) values, the weighting being determined by the relevant control coefficients. It has been suggested that this leads to untenable inferences. For example, since the units of B and a are fixed, the units of c I must depend on those of b(i); but by definition, both b(i) and c I must be dimensionless. Also, according to the basic equation, the contribution made by each step to the overall metabolic rate depends on the units in which body mass is measured. If this criticism is valid then it is impossible to evaluate the model as it stands, because any attempt to align its predictions with experimental data would be meaningless. Another reservation about this model is that it does not purport to apply to all taxa, as the West et al. model does; it relates only to metazoa, and in particular to homoiotherms. However, most of the relevant data in the literature concern homoiotherms. A subsequent publication from this group [ 115 ] re-stated the basic equation in the form B = aΣc I (M/m) b(i) , where the constant a is described as the "characteristic metabolic rate" of an animal with characteristic body mass m . This eliminates the problem of mass units, because the mass term has been rendered dimensionless; and it is mathematically simple to express control coefficients in dimensionless form. The revised equation might therefore be immune to some of the criticisms levelled at its predecessor. However, some of the earlier reservations remain: the equation remains phenomenological, not physical or geometrical; and the restriction in its range of application is explicit. Nevertheless, these considerations by no means invalidate the model. Indeed, it is supported by data from experiments in exercise physiology [ 116 ]. The models of Darveau et al . [ 112 , 115 ], Banavar et al. [ 94 , 95 ] and West et al. [ 88 , 102 ] all have attractive features; but they all have flaws, and they cannot be reconciled with one another. If the positive contributions to biology that these models represent could be further developed, and their defects eliminated, could they be harmonized? If so, the advancement of our understanding would be considerable. Conclusions Several explanatory or quasi-explanatory models have been proposed for the allometric scaling of metabolic rate with body mass. Most of them have significant attractions, particularly the most recent ones, but none of them can be unreservedly accepted. The variability of experimental data leaves room for doubt that Kleiber's law is universally or even widely applicable in biology [ 17 , 117 ], yet most workers in the field presume that it is. Even if such doubts are set aside, no model has yet addressed every relevant issue. For example, the biochemical reasons for the allometric scalings of mitochondrial inner membrane areas and unsaturated fatty acid contents, and the direct proportionality of "pulmonary diffusion capacity" to body mass, remain unexplained. Despite the continuing controversy in the field, the consensus remains, and practical use has been made of Kleiber's law, for example in making numerical predictions of anatomical and physiological parameters for veterinary applications [ 118 ]. Perhaps the last word should be given to Bokma [ 119 ], whose most recent paper explores the power-scaling of metabolic rate to body mass ( b ) on an intra-specific basis from a total of 113 species. He came to the conclusion that there was no single universal value of b . This evidence alone must make us more sceptical of there being some unifying law involved that demands that b holds close to 0.75. There is clearly no consensus otherwise Nature , Science and the Proceedings of the National Academy of Sciences USA would cease to publish so regularly many of the articles to which we have referred. The subject is not only unresolved, but remains very much within the general interest of biologists. Kleiber's law remains a fascinating mystery; possibly a delusion, possibly a widespread or even ubiquitous biological phenomenon for which no entirely satisfactory account has yet been offered. Recent developments, though mutually conflicting as they stand, have the potential to lead to new insights and to uncover one or more general biological principles that will have a profound impact on our understanding of the living world.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539293.xml

549563

The Darlington and Northallerton Long Term Asthma Study: pulmonary function

Background The Darlington and Northallerton Asthma Study is an observational cohort study started in 1983. At that time little was published about long term outcome in asthma and the contribution of change in reversible disease or airway remodelling to any excess deterioration in function. The study design included regular review of overall and fixed function lung. We report the trends over fifteen years. Methods All asthmatics attending secondary care in 1983, 1988 and 1993 were recruited. Pulmonary function was recorded at attendance and potential best function estimated according to protocol. Rate of decline was calculated over each 5-year period and by linear regression analysis in those seen every time. The influence of potential explanatory variables on this decline was explored. Results 1724 satisfactory 5-year measurements were obtained in 912 subjects and in 200 subjects on all occasions. Overall rate of decline (ml/year (95%CI)) calculated from 5-year periods was FEV1 ♂41.0 (34.7–47.3), ♀28.9 (23.2–34.6) and best FVC ♂63.1 (55.1–71.2)ml/year, ♀45.8 (40.0–51.6).The principal association was with age. A dominant cubic factor suggested fluctuations in the rate of change in middle life with less rapid decline in youth and more rapid decline in the elderly. Rapid decline was possibly associated with short duration. Treatment step did not predict rate of deterioration. Conclusions Function declined non-linearly and more rapidly than predicted from normal subjects. It reports for the first time a cubic relationship between age and pulmonary function. This should be taken into account when interpreting other articles reporting change in function over time.

Background It is recognised that the average decline in pulmonary function is greater in asthmatic subjects than in the general population [ 1 , 2 ]. This might be due to deterioration in potentially reversible disease [ 3 ] or the development of persistent obstruction following airway remodelling [ 4 ]. The published longitudinal studies do not differentiate between the two possible mechanisms. The Darlington and Northallerton Long Term Asthma Study was started in 1983 when little was known about long term outlook in asthmatics. Its objective was to observe mortality and decline in pulmonary function in asthmatics sufficiently severe to be referred for a hospital opinion. Decline in best achievable function is proposed as a measure of the airway remodelling. We accept that this decline might also be associated with Chronic Obstructive Pulmonary Disease (COPD). This label implies a physiological diagnosis, but the condition, like asthma, is better defined in terms of the underlying inflammation [ 5 ]. The diagnoses are therefore not necessarily mutually exclusive. Changes in best function are reported without prejudice to the underlying type of inflammation whether asthmatic, COPD, or both. We wished this to be a population study as far as possible, so we invited all subjects satisfying a broad definition of asthma referred to a single-handed respiratory physician in a well defined geographical area to participate. Very few refused, but patients managed entirely in general practice were necessarily excluded. Best function, assessed according to a defined protocol [ 6 ] implicitly accepted in published guidelines [ 7 ], and potential explanatory variables were recorded prospectively at each visit. Smokers were not specifically excluded, but smoking habit was taken into account when the diagnosis was in doubt. Subjects with severe established fixed obstruction were excluded. Thus there was bias against the inclusion of asthmatic smokers destined to develop fixed obstruction rapidly, so we do not present any analyses confined to smokers. Although it was not until the mid 1980's that the use of prophylactic inhaled corticosteroids became standard practice, the majority of our subjects were maintained on inhaled steroids throughout the period, but the dose intensified [ 8 ]. Therefore only the change in dose could not account for any secular trends in decline in function. In this paper we explore the influence of demographic and other factors on change in function as observed over five year intervals. Methods These, described in detail elsewhere [ 8 - 11 ] are summarised and where directly relevant expanded below. Subjects All asthmatics currently attending secondary care clinics in the Darlington and Northallerton Health Districts were eligible for recruitment in 1983, 1988 and 1993, and reviewed in 1988, 1993 and 1998. Asthma was diagnosed clinically and confirmed by reversibility of FEV1 or peak flow of at least 15% on more than one occasion, either spontaneously, or in response to bronchodilator at any time since referral [ 9 ]. Subjects were only eligible for the study if they had been observed for at least one year before entry, and if not stable when first reviewed, entry was deferred by three months in an attempt to achieve stability. Socio-demographic variables were recorded as previously reported [ 10 ]. The first group of this dynamic cohort study was recruited during the calendar year 1983 but subsequent review and recruitment was between the 1 st April and the 31 st March in subsequent 12-month periods[ 11 ]. Allowance for the extra 3 months of the first interval has been made in calculating rate of change. Subjects are included in this report provided they had two technically satisfactory measurements of actual FEV1 or two estimates of best FEV1 or best FVC according to protocol. Social and demographic variables The following variables recorded included: age, gender, height, duration of asthma, atopy, childhood asthma, smoking habit and lifetime amount smoked, social class. Duration of asthma was determined from the first onset or from relapse after a symptom free interval of at least five years, if this was applicable. Atopy was determined by at least one of the following skin tests resulting in a diameter at least 3 mm greater than control: D Pterynissimus, cat, grass pollen, A fumigatus. Childhood asthma was defined as a childhood history of recurrent lower respiratory tract symptoms with wheeze, in the absence of a localised structural damage such as bronchiectasis. It was sub-divided into those with (Gap Asthmatics) and without (Continuous since Childhood) a symptom free gap of at least five years. Never smokers were those smoking less than one cigarette a day for one year. Ex-smokers had been abstinent for at least three months at the time of examination. The lifetime amount smoked was determined from the average consumption (expressed in packs of 20 cigarettes) multiplied by the duration of smoking to give a figure in pack-years. Therapeutic regimen This was characterised by the use of long acting beta agonists and the corticosteroid step: none, low dose inhaled (<800 micrograms per day), intermediate dose inhaled (800–1000 micrograms per day), high dose inhaled (>1000 micrograms per day), oral 1–9 mgs per day, oral ≥ 10 mgs per day, unstable (treatment not mutually agreed as satisfactory over the last three months). Pulmonary function Actual function Actual function was that recorded at attendance. Best function This was estimated according to the published protocol [ 6 ]. The notes were searched from January 1 st of the previous year, and the highest value, including the after-bronchodilator reading at attendance, was accepted as best subject to the following. (a) If >80% predicted (Cotes [ 12 ]); after-bronchodilator (b) If 70–80% predicted; after-bronchodilator and stable on mutually agreed satisfactory preventative treatment with twice daily recording of peak flow for one week (c) If <70% of predicted; after-bronchodilator with formal trial of corticosteroids (prednisolone 30 mgs for at least five days with stability of peak flow for at least 48 hours) If the above was not already satisfied, best function was immediately established according to the protocol. Measurement Spirometry was performed using a rolling wedge spirometer (Vitalograph Limited) and peak flow with the mini peak flow meter (Clement Clark Limited). Actual function was measured opportunistically at the clinic in current attenders and at a special clinic for those who had been discharged. One of three research fellows, one of two research nurses or CKC were responsible for checking current pulmonary tests in the clinical records and performing further tests when required by protocol for best function. Ethical approval This was obtained from the Darlington, Northallerton and South Durham Ethics Committees at various stages of the study. Statistical analysis The principal independent variable was change during each 5-year period so each individual contributed up to 3 observations to the analysis. For the 200 subjects with observations on all four occasions a secondary analysis could be based on decline over 15 years, and for each of these subjects the mean rate of decline was estimated from the four observations using linear regression analysis. Best and actual/best peak flow measured independently of the spirometric outcome variables were chosen as the functional potential predictors of outcome. This approach was taken to avoid the mathematical relationship between baseline level and change that applies if the spirometric variables are used as predictors of outcome. Many of these potential predictive variables are correlated with each other and there are arguments for and against models examining the effect of socio-economic variables with or without allowance for the respiratory function of the patients and their age. We present a relevant selection from the large range of unrestricted and hierarchical models that we constructed. We allowed for the inevitable regression to the mean associated with outcome measures subject to appreciable variability. Initially the starting value of an outcome functional variable was regressed against a full set of potential predictors of this function in order to generate a set of residuals indicating the extent to which an individual starting measurement is higher or lower than would be expected. Then the change in function was regressed against the residual. The new set of residuals indicates the extent to which the apparent rate of change is affected by regression to the mean enabling an adjusted rate of change to be calculated. This adjusted change in function was then taken as the dependent variable in subsequent analyses. These included the univariate effect of potential risk factors on the adjusted change in function. In the multivariate analyses we omitted variables which were consistently non-significant, but all others are retained in the presentation. Models are shown with and without the inclusion of age. Univariate analysis showed that there were cubic relationships between change in function and age and duration. The cubic terms are retained in the multivariate presentation. We also examined interactions between predictor variables where there was some a priori reason for believing that such an interaction was plausible but none was demonstrated. Results General Of 1138 subjects recruited, 155 died before the first review, 49 were lost to follow up and 22 did not have satisfactory spirometric tests. The remaining 912 had at least one paired result of actual or best FEV1 or best FVC and 200 had satisfactory assessments of best FVC on all four occasions. Demographic details at entry are given in Table 1 . There were no relevant differences in social factors between the subjects observed on all four occasions and the rest. Current smokers represented approximately 13% of the population in 1983/88 and 10% in 1993/98 with a mean tobacco load of ♂29.2 (ex-smokers 27.6) and ♀21.4 (ex-smokers 13.2) pack years. Ever-smokers were significantly older than never-smokers (53.9 v 44.9 years p < 0.001), but allowing for age, the atopic status of ever-smokers and never-smokers was the same. The proportion of subjects stable on inhaled or oral steroids rose from 66.5% in 1983 to 82.1% in 1988 with no increase thereafter. Higher doses (> = 800 micrograms) of inhaled steroids increased from 30.8% in 1988 to 41.6% in 1998. Table 1 Details of subjects at entry Male Female n 457 455 Age (years)(sd) 50.4(15.4) 49.3(16.3) Duration of asthma (years)(sd) 17.0(16.2) 18.6(15.4) Atopic n(%) 258(56.5) 227(49.9) Asthma n(%) childhood 124(27.1) 131(28.8) gap 53(11.6) 42(9.2) adult 280(61.3) 282(63.3) Social Class n(%) 1–2 131(28.8) 139(30.5) 3 168(36.9) 180(39.6) 4–5 156(34.3) 136(29.9) Smoking Habit n(%) never 147(32.2) 247(54.3) ever 239(52.3) 161(35.4) current 71(15.5) 47(10.3) Pulmonary Function(sd) Actual FEV1 l. 2.50(1.09) 1.97(0.79) Best FEV1 l. 2.60(1.08) 2.08(0.82) Best FVC l. 4.17(1.25) 3.05(0.84) Best PEF l/min 465.7(113.7) 382.0(85.1) Actual/Best PEF % 84.4(13.7) 83.2(15.9) Change in function The mean annual decline in function calculated using all five-year intervals (Table 2 ) was greater than expected when compared with predicted values and when expressed as %predicted [ 12 ] there were no consistent differences in the rates of decline of male and female never-smokers. As the confidence intervals suggest, standard deviations were large indicating a wide distribution of changes in different individuals. There was no secular trend in outcome in successive calendar periods. Figure 1 shows changes over 5, 10 and 15 year periods after entry against predicted [ 12 ] (improved >7.5%, no change, deteriorated >7.5%). Actual FEV1 improved in approximately one quarter and best FVC in rather less than 20% of subjects. The proportion of subjects showing deterioration in FEV1 >7.5% (35%) did not increase with time after entry, but excess decline in FVC was observed in more subjects after 15 year's observation (58%) than after five (37%) (χ 2 for trend, 59.0 (p < 0.001)). Table 2 Annual Decline in Pulmonary Function calculated from the mean of all 5-year paired observations, and by linear regression over 15 years in the 200 subjects with all four observations Annual Decline n Actual FEV1 l. Best FEV1 l. Best FVC l. Males Over 5 Years All Subjects 776 41.0 (34.7–47.3) 42.9 (37.5–48.3) 63.1 (55.7–71.2) Never smokers 256 34.2 (22.2–46.2) 33.4 (22.7–44.1) 49.2 (35.6–62.8) Over 15 Years 85 46.7 (38.7–54.7) - 64.7 (54.6–74.8) Females Over 5 Years All Subjects 848 28.9 (23.2–34.6) 34.4 (29.8–39.0) 45.8 (40.0–51.6) Never smokers 375 28.7 (22.4–38.8) 30.6 (22.2–35.2) 45.2 (36.4–54.0) Over 15 Years 115 24.7 (14.8–31.2) - 44.6 (37.9–51.3) Figure 1 The proportion of all subjects and never-smokers showing decline (>7.5%), no change, or improvement (>7.5%) in function against predicted over 5, 10 and 15 years. For the 200 subjects with complete observations over 15 years the mean rate of decline was similar to that observed over five year periods, but the standard deviation of the rates of decline was half that of the 5-year estimates, reflecting greater accuracy from multiple measurements. Even when calculating change this way several individuals improved or showed excessive loss in function against predicted [ 12 ]. Associations with entry variables As there was no secular trend in the change of function over the five year periods, the date of observation is not considered in the analyses below. All subjects: 5-year intervals Univariate analysis The univariate relationships between the potential explanatory variables and change in function after allowance for regression to the mean are summarised in Table 3 . These are derived from all available pairs of observations at five year intervals. The strong associations with age are not linear. This is demonstrated in fig 2 which shows the fitted plots of the cubic equations for the unadjusted rate of decline for all three measures. The maximum decline was in the mid 40's for all three variables (actual FEV1 44 ml/yr; best FEV1 56 ml /yr; best FVC 70 ml/yr). During the eighth decade rate of decline in function recovered towards the published predicted values [ 12 ] to 27, 29 and 49 ml/yr respectively. Table 3 Rate of loss of Function Univariate Coefficients (ml per year) (after allowance for regression to the mean) Actual FEV1 Best FEV1 Best FVC Estimate 95% CI Estimate 95% CI Estimate 95% CI Gender (male v female) 7.7 0.8 + ,16.1 6.2 2.5 + ,14.9 12.2 2.5,21.9 Age at entry (per decade) (difference from age 50) Linear 5.6 + 11.4 + ,0.2 10.1 + 15.9 + , 4.8 + 2.1 + 8.7 + ,4.5 quadratic 4.3 + 6.0 + , 2.5 + 5.0 + 6.8 + , 3.3 + 4.5 + 6.5 + , 2.6 + cubic 1.3 0.3,2.3 2.0 1.0,2.9 1.4 0.3,2.6 Duration of asthma at entry (per decade) (difference from duration 20 yrs) Linear 3.1 + 7.4 + ,1.1 0.5 3.8 + ,4.8 0.1 4.8 + ,5.0 quadratic 3.6 0.6,6.6 2.7 0.5 + ,5.8 5.6 2.2,9.0 cubic 0.9 + 1.6 + , 0.1 + 0.9 + 1.7 + , 0.1 + 1.4 + 2.3 + , 0.6 + Atopic 5.7 + 14.2 + ,2.8 0.2 8.5 + ,8.9 3.0 + 12.7 + ,6.7 Childhood asthma (versus no childhood asthma) Gap 3.4 + 16.7 + ,9.9 1.4 12.2 + ,15.0 4.0 + 19.4 + ,11.3 Yes 20.8 + 30.6 + , 10.9 + 10.6 + 20.9 + , 0.4 + 22.0 + 33.3 + , 10.6 + Social Class (versus classes 1 and 2) Three 6.3 + 16.5 + ,3.8 11.2 + 21.7 + , 0.8 + 9.2 + 20.8 + ,2.3 Four and Five 10.2 + 21.0 + ,0.7 14.5 + 25.6 + , 8.3 + 8.6 + 21.0 + ,3.9 Amount smoked (per 10 pack years) 1.5 0.6 + ,3.7 0.2 2.1 + ,2.4 2.9 0.4,5.4 Best PEF(per 10% deficit) 3.3 + 5.3 + , 1.2 + 2.8 + 4.8 + , 0.7 + 0.4 + 2.7 + ,2.0 Actual/Best PEF(10% deficit) 12.7 + 15.6 + , 9.8 + 2.8 + 6.2 + ,0.6 0.6 + 4.1 + ,2.8 (+) Indicates relative gain in function Figure 2 The relationship between age and annual change in function, observed over 5-year periods There were no statistically significant relationships with atopy. Longer duration of asthma was significantly associated with favourable outcome for all variables. In all cases a cubic relationship between loss of function and duration suggests high initial rates of loss (actual FEV1 55 ml/yr; best FEV1 51 ml/yr; best FVC 74 ml/yr at one year), with improvement to a plateau at around 20 years duration (30 ml/yr, 37 ml/yr and 45 ml/yr). Childhood asthma was significantly associated with a relatively favourable outcome, though this benefit was only seen in those for whom asthma had been continuous. The outcome of the 'gap asthmatics' was similar to those with adult onset. Higher social class was associated with worse outcome, significantly so for best FEV1. Low initial function and worse control as assessed by actual/best PEF both predicted less loss in actual FEV1. Although there were no significant associations with current smoking status, there was significantly greater loss in best FVC with amount smoked. Multivariate analysis As childhood asthma, duration of asthma and age at entry are potentially confounded and were significant in some of the univariate analyses, we performed a series of multivariate analyses progressively including each in turn. Tables 4 , 5 , 6 show the results after the inclusion of duration and then age. Male gender was significantly associated with greater decline in best FEV1 and, in contrast to the univariate analysis, best FVC. After the inclusion of duration, childhood asthma was still associated with favourable outcome in both actual FEV1 and best FVC, but it was displaced by age in all the models. Duration remained significantly associated with less change in best FVC even after allowance for age. In both models, actual FEV1 declined less with lower entry actual/best PEF. Best FEV1 declined more rapidly in those with a high initial best PEF. When age was not included in the model, membership of social classes 1 and 2 remained associated with unfavourable outcome for best FEV1. Table 4 Multivariate Coefficients (sd) for Rate of loss of Actual FEV1 (after allowance for regression to the mean) Age Excluded Age Included Coefficient p Coefficient p Gender (male v female) 5.5 (4.8) 0.25 5.7 (4.7) 0.23 Asthma since childhood 13.5 + (5.1) 0.008 4.9 + (5.8) 0.40 Social Classes 1 & 2 (v. classes 4 & 5) 4.8 (5.4) 0.65 1.8 (5.4) 0.88 Social Class 3 (v. classes 4 & 5) 1.1 (5.1) 0.6 + (5.0) Amount smoked (per 10 pack years) 1.8 (1.2) 0.13 1.0 (1.2) 0.39 Best PEF(per 10% deficit) 1.2 + (1.2) 0.33 0.9 + (1.2) 0.41 Actual/Best PEF(10% deficit) 12.0 + (1.6) <0.0001 13.2 + (1.6) <0.0001 Age at entry (per decade) (difference from age 50) Linear 3.5 + (3.1) quadratic 4.9 + (0.9) cubic 1.2 (0.6) 0.02 Duration of asthma at entry (per decade) (difference from duration 20 yrs) Linear 1.3 (2.8) 2.0 + (2.4) quadratic 3.2 (1.5) 0.9 (0.5) cubic 0.85 + (0.38) 0.03 2.5 + (0.39) 0.52 (+) Indicates relative gain in function Table 5 Multivariate Coefficients for loss of Best FEV1 (after allowance for regression to the mean) Age Excluded Age Included Coefficient (sd) p Coefficient (sd) p Gender (male v female) 11.3 (5.1) 0.03 11.9 (5.4) 0.02 Asthma since childhood 5.1 + (5.3) 0.34 1.1 + (6.0) 0.86 Social Classes 1 & 2 (v. classes 4 & 5) 13. (5.7) 0.05 10.7 (5.6) 0.11 Social Class 3 (v. classes 4 & 5) 2.9 (5.3) 1.4 (5.2) Amount smoked (per 10 pack years) 0.7 (1.3) 0.60 0.5 (1.3) 0.71 Best PEF(per 10% deficit) 3.4 + (1.3) 0.008 3.1 + (1,2) 0.01 Actual/Best PEF(10% deficit) 0.7 + (1.8) 0.71 1.4 + (1.8) 0.44 Age at entry (per decade) (difference from age 50) Linear 8.4 + (3.2) quadratic 5.1 + (0.9) cubic 1.8 (0.5) 0.0003 Duration of asthma at entry (per decade) (difference from duration 20 yrs) Linear 3.1 (2.4) 0.2 (2.4) quadratic 2.5 (1.6) 0.4 (1.6) cubic 0.91 + (0.39) 0.02 0.29 + (0.40) 0.46 (+) Indicates relative gain in function Table 6 Multivariate Coefficients for Rate of loss of best FVC (after allowance for regression to the mean) Age Excluded Age Included Coefficient (sd) p Coefficient (sd) p Gender (male v female) 10.8 (5.6) 0.05 11.4 (5.6) 0.04 Asthma since childhood 16.0 + (6.0) 0.007 4.7 + (6.9) 0.49 Social Classes 1 & 2 (v. classes 4 & 5) 10.7 (6.4) 0.13 7.5 (6.4) 0.29 Social Class 3 (v. classes 4 & 5) 0.2 + (6.0) 1.4 + (5.9) Amount smoked (per 10 pack years) 1.9 (1.4) 0.18 0.9 (1.4) 0.52 Best PEF(per 10% deficit) 0.6 + (1,.4) 0.65 0.6 + (1.9) 0.67 Actual/Best PEF(10% deficit) 1.3 (1.8) 0.47 0.1 (1.9) 0.94 Age at entry (per decade) (difference from age 50) Linear 2.1 + (3.7) quadratic 4.0 + (1.0) cubic 1.4 (0.6) 0.02 Duration of asthma at entry (per decade) (difference from duration 20 yrs) Linear 4.0 (2.8) 0.8 (2.8) quadratic 5.2 (1.7) 3.2 (1.8) cubic 1.46 + (0.44) 0.001 0.97 + (0.46) 0.03 (+) Indicates relative gain in function Height (not tabulated) was consistently directly associated with decline at a significance level of the order of 15% in univariate and multivariate analyses. In the latter analyses, height suppressed the associations with gender and social class. There were no significant associations with instability of regimen, nor with steroid step or the regular prescription of bronchodilators. None of the models considered made any relevant difference to the shape or gradients of the curves shown in fig 2 . 200 Subjects: decline estimated using all four observations Generally similar univariate associations were demonstrated. However on multivariate analysis, amount smoked remained in the FVC model (coefficient 13.3 ml/year per 10 pack years P = 0.020), and the effect of gender was stronger in both models (actual FEV1 23.8, actual FVC 19.8 ml/year, (P < 0.001). Height was again consistently directly associated with decrease in function at a significance level of the order of 15%. When height was included to allow for its association with gender, the difference in rate of decline in FEV1 between males and females was substantially reduced to 12.5 ml per year (p = 0.010) and was no longer significant for best FVC (p = 0.14). Discussion The hypothesis that persistent airway obstruction may develop in asthmatics implies that persistent airway obstruction and reversible obstruction associated with asthma are not mutually exclusive diagnoses However if COPD is regarded as a syndrome; this implies a non-asthmatic inflammatory process [ 5 ]. As the two types of inflammation need not be mutually exclusive, some asthmatics might have both types raising the possibility of an interaction between the two processes in some individuals, 'The Dutch Hypothesis' [ 13 ]. We have already shown that on cross sectional analysis green sputum, which is a feature of severity in COPD [ 14 ], is associated with diminished best function independent of smoking habit [ 15 ]. We are satisfied that our subjects satisfied the clinical criteria for asthma and had the relevant inflammatory process. Of course, the study is not of the entire asthmatic population as some individuals will remain undiagnosed while others will have been managed entirely in general practice, but all those referred to hospital and willing were entered into the study. We believe that the referral threshold was relatively low, and very few patients were referred outwith the local area. The subjects were recruited by a single handed hospital physician, with no other selection. There were no differences in the demography or function of those entering at different times [ 11 ]. Most were stable on maintenance corticosteroids [ 8 ]. The rate of decline of actual FEV1 was similar to that reported by Ulrik [ 1 ] (38 ml/year) and Lange [ 2 ] (50 ml/year). This study demonstrates for the first time that decline in best after bronchodilator function, the definitive physiological measurement for COPD [ 16 ], is similar to that of actual function. Actual/best PEF tended to improve over the period [ 8 ], so measurement of actual FEV1 potentially under-estimates any decline in best FEV1, but in practice loss of actual FEV1was not relevantly different from that of best FEV1. As in the previous studies [ 1 , 2 ] the rate of decline of actual FEV1 is greater than that suggested by reference equations [ 12 , 17 ] which are derived from cross-sectional data. Values derived from longitudinal data may differ from cross-sectional observations for a number of reasons [ 18 , 19 ]. These include a cohort effect and, with lung function, loss of height with ageing which will mask decline in the cross sectional tables. We allowed for loss of height by using height at the start of each 5-year period. After discounting this there was little difference between the genders, the principal association with change in function being the age of the subjects. Longitudinal studies suggest increased loss of FEV1 in elderly normal subjects [ 20 - 22 ], but the rate of decline in our study is greater at all ages than that reported from the general population [ 20 ]. Our results suggest that that a cubic model involving age is appropriate for all of the three measures of respiratory function that we have analysed. Although the pattern in a particular individual cannot be determined from these computed curves, they strongly suggest that decline is not linear. The interpretation of the pattern of change is critically dependent on its normal course. The critical points on the curve include the age at which maximum life-time function is achieved and the possibility of a plateau, before deterioration starts The age when personal best is achieved is variously estimated between the early twenties and the late thirties [ 19 ], and may well vary from person to person. In looking for a plateau phase Robbins et al [ 23 ] demonstrated both positive and negative slopes in different individuals. Any decline seen where improvement or no change is anticipated must be excessive and so the slow decline that we observed in the third decade may be an under-representation of the true loss of function. It is more difficult to explain the faster decline in early compared to late middle age. As the entry criteria of this study were a diagnosis of asthma with no attempt to exclude co-existent COPD, it is possible that this reflects a period of life when the effects of the inflammatory process associated with COPD are particularly apparent. In subjects where asthma and COPD co-exist, COPD might contribute to presentation in these subjects, and decline in function might be particularly rapid especially if there were interaction between the two inflammatory processes. The above is compatible with the apparently unsustainable rates of decline observed early in COPD as in the Euroscop study [ 24 ]. We confirmed that decline accelerates after in late life in these asthmatics as it does in normal subjects [[ 20 , 21 ], and [ 22 ]]. Outcome in successive periods is necessarily confounded by the effects of management and attrition. Attrition is particularly relevant to the analysis of change in function, as the most powerful predictor of survival in these subjects is best function [ 11 ]. Higher dose of inhaled corticosteroids at the start of an observation period was not associated with better outcome within the same period. This is unsurprising because a higher dose reflects more severe disease. More surprisingly, there was also no evidence that the higher average dose in successive periods was associated with a favourable secular trend in observed, or maximum, function. Nevertheless in contrast, actual/best peak flow improved [ 8 ] and standardised mortality declined with time [ 11 ]. The latter was reduced twofold although subjects on higher therapeutic steps remained more likely to die even after allowance for other risk factors including best function. It may be that clinicians are able to recognise subjects of poor prognosis irrespective of function, but are more successful in reducing mortality than loss of function by adjustment of therapy. None of these considerations necessarily imply that the use of inhaled corticosteroids is ineffective in terms of function. As there were, of necessity, no controls, any effect of treatment, dictated by clinical need, is impossible to confirm. Improvement in functional outcome might have been confounded by reduced mortality in those with severe disease, or the dose response curves for reduction of mortality might be very different from those for prevention of airway remodelling. We depended primarily on patient recall in making a diagnosis of childhood asthma. Although we had a low threshold for accepting the diagnosis, presumably recall would be more consistent when childhood symptoms were severe. This might produce a bias in favour of demonstrating associations. Nevertheless the univariate associations with change in function were stronger in those who claimed that they had had persistent symptoms since childhood, than in those who recalled childhood asthma after a gap of at least five years. The outcome in gap asthmatics was similar to those with adult onset, but surprisingly subjects with symptoms persistent since childhood showed a more favourable trend. This appears to be inconsistent with the long established observation that childhood asthma compromises adolescent and early adult lung function and that this is related to persistent symptoms [ 25 , 26 ]. However duration is probably the critical factor. The history of asthma is likely to be long in adults with symptoms persistent from childhood and rapid decline is shown to be associated with short duration. Although these subjects did have relatively poor lung function at entry to the study [ 10 ], probably reflecting their function on reaching maturity, it does not necessarily follow that retarded development will be succeeded by an excessive rate of decline later in life. It may be that at any age whether in childhood or adulthood the first decade of the disease is critical in determining loss of function. This might not be true in those with pure COPD, and explain why our findings are contrary to the 'horseracing effect' (the horse that runs fastest continues to extend its lead) as described by Fletcher in his classical population studies [ 27 ]. There the comparison was with normal subjects rather than those with established airway disease with differing lengths of history. The effect of smoking appeared small, but there were few current smokers and the tobacco load was light. The study was not designed to observe the effects of tobacco smoking in asthma; the separate analysis of non-smokers was intended to describe the decline in the function of asthmatics unencumbered by the effects of cigarette smoking. It is inevitable that we have underestimated the potential association between smoking and decline in function in asthma and so do not suggest that the effect of smoking in asthmatics in general is unimportant. The association between low initial actual/best function, implying poor control, and apparently favourable outcome is highly likely to reflect response to treatment. There was a strong relationship between low social class and poor control in the 1983 entry [ 10 ]. This may account for some of the paradoxical benefit of lower social class, possibly even hiding a real disadvantage. Conclusions We present the decline in function in a dynamic cohort of adult asthmatics observed over fifteen years. The majority were treated with inhaled corticosteroids throughout the period. As there are no internal asthmatic or normal controls our study cannot determine definitively whether there is excess decline in the pulmonary function of adults managed conventionally with inhaled cortico-steroids. It does suggest, however, that the dose of inhaled steroids may not be critical over the recommended range. Our study confirms that the pattern of decline in actual and best ventilatory function is similar. This is important when comparing this study with epidemiological exercises where actual rather than best function has been measured. Furthermore these original findings in respect of the cubic effect of age should be taken into account when interpreting other articles reporting age effects on function, particularly where analysis of cross-sectional observations may imply that decline in the function is linear. Competing interests The author(s) declare that they have no competing interests. Authors' contributions CKC conceived the study and was responsible for design of data forms, recruitment and all the clinical aspects. RJP undertook the analysis of data. Both authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549563.xml

529297

Assessment of the dynamics of atrial signals and local atrial period series during atrial fibrillation: effects of isoproterenol administration

Background The autonomic nervous system (ANS) plays an important role in the genesis and maintenance of atrial fibrillation (AF), but quantification of its electrophysiologic effects is extremely complex and difficult. Aim of the study was to evaluate the capability of linear and non-linear indexes to capture the fine changing dynamics of atrial signals and local atrial period (LAP) series during adrenergic activation induced by isoproterenol (a sympathomimetic drug) infusion. Methods Nine patients with paroxysmal or persistent AF (aged 60 ± 6) underwent electrophysiological study in which isoproterenol was administered to patients. Atrial electrograms were acquired during i) sinus rhythm (SR); ii) sinus rhythm during isoproterenol (SRISO) administration; iii) atrial fibrillation (AF) and iv) atrial fibrillation during isoproterenol (AFISO) administration. The level of organization between two electrograms was assessed by the synchronization index (S), whereas the degree of recurrence of a pattern in a signal was defined by the regularity index (R). In addition, the level of predictability (LP) and regularity of LAP series were computed. Results LAP series analysis shows a reduction of both LP and R index during isoproterenol infusion in SR and AF (R SR = 0.75 ± 0.07 R SRISO = 0.69 ± 0.10, p < 0.0001; R AF = 0.31 ± 0.08 R AFISO = 0.26 ± 0.09, p < 0.0001; LP SR = 99.99 ± 0.001 LP SRISO = 99.97 ± 0.03, p < 0.0001; LP AF = 69.46 ± 21.55 LP AFISO = 55 ± 24.75; p < 0.0001). Electrograms analysis shows R index reductions both in SR (R SR = 0.49 ± 0.08 R SRISO = 0.46 ± 0.09 p < 0.0001) and in AF (R AF = 0.29 ± 0.09 R AFISO = 0.28 ± 0.08 n.s.). Conclusions The proposed parameters succeeded in discriminating the subtle changes due to isoproterenol infusion during both the rhythms especially when considering LAP series analysis. The reduced value of analyzed parameters after isoproterenol administration could reflect an important pro-arrhythmic influence of adrenergic activation on favoring maintenance of AF.

Background Atrial Fibrillation (AF) results from multiple, rapidly changing and spatially disorganized activation wavelets sweeping across the surface of the atria [ 1 ]. Among factors contributing to genesis and / or maintenance of circulating wavelets, Autonomic Nervous System (ANS) seems to play a major pro-arrhythmic role [ 2 ]. The arrhythmogenic influence of sympathetic and vagal mechanisms has been documented in several clinical and experimental studies [ 3 , 4 ]. In men, ablation of the major parasympathetic pathways to the atria drastically reduced vagally mediated atrial fibrillation [ 4 ]. It has also been reported that sympathetic stimulation by shortening atrial refractory periods, may increase vulnerability to atrial fibrillation in different experimental models [ 5 ]. The shortening of action potential duration and flattening of the restitution slope to cycle length changes induced by adrenergic activation, are two of the mechanisms favoring spiral wave induction and restraining spiral wave break-up [ 6 ]. Changes in action potential may also contribute to the perpetuation of atrial fibrillation [ 7 ]. In normal hearts, both vagal and sympathetic mechanisms have been associated with paroxysmal atrial fibrillation (PAF) initiation. Most of PAF episodes observed in patients with structural heart disease are triggered by sympathetic activation and vagal withdrawal [ 8 ]. Spectral analysis of heart rate variability before PAF episodes has further clarified the pro-arrhythmic role of the autonomic nervous system [ 9 , 10 ]. Bettoni [ 9 ] observed a primary increase in adrenergic drive occurring over at least 20 minutes before onset of PAF episodes followed by a shift towards a vagal predominance immediately before arrhythmia onset. Other authors described an increase in sympathetic modulation of sinus node (or a loss of vagal modulation) before PAF onset in the majority of patients [ 10 - 12 ]. More recently Lombardi [ 7 ] reported that signs of sympathetic activation characterized up to 70% of PAF episode onset, whereas in the remaining ones a vagal predominance was detectable. An increase in vagal modulation can also promote the stability of AF [ 13 ]. Even if AF has been classically described as a random process, a few studies have recently documented, using various signal processing methods, the existence of some determinism underlying AF. Linear analysis techniques documented relationships between intra-atrial recordings using both time-domain methods [ 14 ] and spectral-domain approaches [ 15 , 16 ], while the presence of non-linear patterns have been also recognized [ 17 , 18 ]. By using linear and non-linear indexes we have recently assessed [ 19 ] the dynamics of intra-atrial signal and local atrial period (LAP) series during different AF episodes. In particular, regularity (R) and synchronization (S) indexes [ 20 ], based on the estimation of the corrected conditional entropy and the corrected cross-conditional entropy respectively, were used to describe the dynamics in intra-atrial signals, whereas the LAP series were investigated using regularity and the level of predictability (LP). These parameters were suitable to describe the fine changing characteristic of atrial signals and LAP series [ 19 ] when passing from different atrial rhythms classified according to the Wells' criteria [ 21 ]. In the present paper, we evaluated whether changes in adrenergic control mechanisms could influence determinisms and dynamics of atrial signals and exploited the capability of linear and non-linear parameters (R and S indexes for intra-atrial signals, R and LP indexes for LAP series) to capture them. Adrenergic activation was mimicked by isoproterenol infusion. The effects of this sympathomimetic drug was evaluated in a small group of patients with a history of PAF during sinus rhythm and atrial fibrillation: four experimental conditions were analyzed (sinus rhythm (SR), sinus rhythm during isoproterenol administration (SRISO), atrial fibrillation (AF) and atrial fibrillation during isoproterenol administration (AFISO)). Experimental protocol Patient population Nine patients (8 males and 1 female; mean age 60 ± 6 years) selected to sustain a left atrial ablation with encirclement of the pulmonary veins by transeptal approach were included in the study. All subjects were suffering from atrial fibrillation (AF) and were non responsive to anti-arrhythmic therapy (pharmacological therapy and electrical cardioversion). Paroxysmal and persistent AF episode were present in, respectively, 5 and 4 subjects. A history of AF was present for an interval ranging from 2 months to 10 years. The mean left ventricular ejection fraction was > 40% in all patients; the mean left atrial diameter was 37 ± 3 mm in 7 patients and 51 ± 8 mm in 2. Structural heart disease was present in 4 patients. Reported symptoms included palpitations (6 subjects), fatigue after effort (9 subjects) and syncope (2 subjects). Arterial hypertension was the most common comorbidity in our study group (4 patients). All the patients were in anti-arrhythmic drug wash-out at the time of the study. Flecainide, propafenone, metoprolol, cordarone and methyldigoxin were ceased ≥ 5 half-lives before ablation. Transoesophageal echocardiography was performed the day before the procedure to exclude atrial thrombus. The Medical Ethical Committee approved this study and all subjects gave their written consent. Study design We investigated the effect of adrenergic activation induced by infusion of isoproterenol on atrial electrical activity. The electrophysiological procedure was performed in the Electophysiology Laboratory of the "Istituto Clinico Sant'Ambrogio" of Milan, Italy. Intracavitary electrocardiograms were recorded during the ablation procedure in which arrhythmic foci inside the pulmonary veins of the left atrium were electrically isolated. The research project protocol included an intracavitary recording of multiple atrial electrograms during sinus rhythm and after induction of atrial fibrillation. In both experimental conditions, the recording was repeated during intravenous infusion of isoproterenol (0.01–0.02 mcg/kg/min) tiered to determine a 30% increase of heart rate. The four clinical experimental conditions were defined as sinus rhythm (SR), sinus rhythm during isoproterenol administration (SRISO), atrial fibrillation (AF) and atrial fibrillation during isoproterenol administration (AFISO). Details on the four epochs of the study can be found in Figure 1 . Figure 1 Timing and sequences of the experimental protocol epochs. The experimental protocol included four recording periods during: I) sinus rhythm (SR); II) sinus rhythm during isoproterenol infusion (SRISO); III) atrial fibrillation (AF); IV) atrial fibrillation during isoproterenol infusion (AFISO). The recordings during SR lasted at least 5 minutes (range 5 – 8 minutes), those during AF 90 seconds on average (range 60 – 120 seconds). The recordings during infusion of isoproterenol started after the drug had determined a 30% increase of heart rate. Induction of AF started 15 minutes after the end of SRISO, to guarantee the correct isoproterenol wash-out. In all the patients, AF inducibility was obtained at twice diastolic threshold by burst atrial pacing (5'-second bursts at an output of 20 mA) from the mid coronary sinus beginning at a cycle length of 250 ms and reducing by 10 ms intervals until atrial refractoriness. All the nine patients were inducible. AF was considered inducible if it persisted for more than 1 minute. If AF terminated after less than 1 minute, induction was repeated until a maximum of 3 times. If AF became sustained (lasting > 10 minutes), ablation was performed after external DC cardioversion. All our patients underwent the procedure in spontaneous SR. The duration of registration was an important parameter for the reliability of the analysis, because an insufficient number of atrial potentials (less than 250 – 300) could give errors in the estimation of conditioned probability. Therefore at least 5 minutes (range 5 – 8 minutes) of sinus rhythm and 90 seconds (range 60 – 120 seconds) of atrial fibrillation were registered. The electrophysiological study was carried out using a deflectable 20 pole St Jude catheter (length 95 cm, 7 F, interelectrode spacing 2 – 10 mm), a deflectable decapolar catheter with a distal ring configuration, Lasso-Cordis Biosense Webster catheter (length 115 cm, 7 F, interelectrode spacing 2 – 5 mm) and 4 mm distal electrode catheter, Medtronic Sprinklr , with irrigated tip (for ablation, length 115 cm, 7 F, interelectrode spacing 2 – 5 mm). The St Jude catheter was in contact with the right atrial wall and inserted in the coronary sinus below the left atrium. The Lasso-Cordis Biosense Webster catheter and the Medtronic Sprinklr catheter were positioned in the superior pulmonary veins at the inside of the atrium. For the purpose of this study, one surface ECG tracing and nine intracavitary atrial electrograms were stored on digital memory for subsequent off line analysis. In all patients, electrograms labeled 2 – 3 – 4 – 5 corresponded, respectively, to the superior, middle, middle inferior and inferior wall of the right atrium; electrogram 6 to coronary sinus ostium; electrograms 7 – 8 – 9 indirectly corresponded to the inferior and the left wall of the left atrium whereas electrogram 10 to the left superior pulmonary vein. Electrograms 2 – 9 were recorded with 20 pole St Jude catheter, electrogram 10 with Lasso-Cordis Biosense Webster catheter. Methods and Data analysis Regularity Conditional entropy ( CE ) may be used to estimate a regularity index, defined as the degree of recurrence of a pattern in a signal. CE represents the amount of information carried by the most recent sample x ( i ) of a normalized realization of x when its past L - 1 samples are known. CE is defined as [ 22 ]: where p ( x L - 1 ) represents the probability of the pattern x L - 1 ( i - 1) of length L - 1 ( x L - 1 ( i - 1) = { x ( i - 1),..., x ( i - L + 1)}) and p ( x ( i ) / x L - 1 ) the conditional probability of the sample x ( i ) given the pattern x L - 1 . In (1) the first summation is extended to all the possible x L - 1 patterns, the second one is extended to all the different L th samples of the pattern x L ( i ) ( x L ( i ) = { x ( i ), x L - 1 ( i - 1)}). CE is maximum if x is complex and unpredictable and it reaches zero as soon as a new sample can be exactly predicted from the previous L -1 ones. Using CE over short data series can cause an unreliable estimate of CE ( CÊ ) : when the conditioning pattern x L - 1 ( i - 1) is found only once in the series x (i.e. p ( x ( i ) / x L - 1 ) = 1), CÊ decreases to zero with L . As a consequence both periodic and completely unpredictable signals exhibit CÊ equal to zero when L increases. Therefore the corrected conditional entropy ( CCE ) must be introduced to perform a reliable measure over short data series: CCE ( L ) = CÊ ( L ) + perc ( L )· Ê (1)     (2) where perc ( L ) is the percentage of length L patterns found only one time in the data set and Ê (1) is the estimate of Shannon entropy of the process x. perc ( L )· Ê (1) represents the corrective term that compensates the null information associated to the pattern found only once and it increases with L , while CÊ ( L ) decreases with L . The minimum value of the CCE is the best estimate of CE and it's taken as an index of complexity: the larger the index, the less predictable the processes. The CCE is normalized by the Shannon entropy of the process in order to derive an index independent of the different probability distribution of the processes, thus obtaining: An index of regularity (the opposite of complexity) may be defined as: R x = 1 - min( NCCE ( L ))     (4) R x tends to zero if x is a fully unpredictable process, it tends to one if x is a periodic signal and it assumes intermediate values for those processes that can be partially predicted by the knowledge of the past samples [ 20 ]. Synchronization The cross-conditional entropy is introduced to define an index of synchronization, related to the repetition of a complex pattern involving two signals. Given two normalized signals, the cross-conditional entropy of x given a pattern y is defined as [ 20 ]: where p ( y L - 1 ) represents the probability of the pattern y L - 1 ( i ) and p ( x ( i ) / y L - 1 ) the conditional probability of the sample x ( i ) given the pattern y L - 1 . CE x/y represents the amount of information carried by the most recent sample of the signal x when L - 1 past samples of y are known. Over short data series, this definition suffers from the same limitations as conditional entropy, so analogously corrective terms and normalization are introduced. The uncoupling function ( UF ) is defined as: UF x,y ( L ) = min( NCCE y/x ( L ), NCCE x/y ( L ))     (6) in order to measure the amount of information carried by one signal that can't be derived from the knowledge of past samples of the other signal. In this way both causal directions are tested and it is taken the one that leads to the best prediction. For every length L pattern, UF chooses as input the signal that can be the best predictor of the other one. Besides, the joint pattern does not take into account past samples of the output signal to prevent to have a high coupling strength only because one signal has a large index of regularity. The minimum of UF is taken as an index of uncoupling between x and y , therefore an index of synchronization (the opposite of uncoupling) can be defined as: S x,y = 1 - min( UF x,y ( L ))     (7) and it quantifies the maximum amount of information exchanged between the two signals. S x,y tends to zero if the two processes are uncoupled, it tends to one if they are perfectly synchronized and it assumes intermediate values when the two signals are able to exchange a certain amount of information [ 20 ]. Level of predictability A discrete time series x ( n ) can be modeled as the output of an autoregressive model of p order where n is the discrete-time index, the a k are the model coefficients and w ( n ) is a Gaussian white noise process of variance feeding the model. The actual sample differs from its model prediction, thus generating the prediction error An index of the level of predictability ( LP ) may be defined as follows LP = (1 - σ e / σ x )     (10) where σ e is the standard deviation of e ( n ) and σ x is the standard deviation of the process x . LP measures the percentage of power which may be predicted by the autoregressive model. In the case of a purely random signal ( σ e is quite close to σ x ) LP tends to zero, while in the case of a linearly predictable signal ( σ e tends to zero) the index tends to one and it assumes intermediate values for those processes that may be partially predicted from the model. Signal pre-processing All signals were appropriately recorded and digitized to a 1000 Hz sampling rate at 16-bit resolution. All bipolar electrograms were band-pass filtered (40 – 250 Hz) to remove baseline shift and high frequency noise. In order to cancel the possible effects of ventricular interference (affecting especially recordings during sinus rhythm), the averaged ventricular interference complex was computed and subtracted from each atrial signal [ 16 ]. In details, from the surface ECG, the occurrences of QRS were determined, and a template of the ventricular interference in each atrial electrogram was constructed by signal-averaging windows of 140 ms around each QRS (windows were positioned 40 ms before and 100 ms after the R wave). The template was then subtracted from the atrial signals at each occurrence of QRS. After detecting time-instants of local atrial depolarization using a derivative / threshold algorithm, the detected depolarizations were visually scored and missed / erroneous detections were corrected by an expert operator using an interactive software. Then, the local atrial period (LAP) series were derived as the sequence of temporal distances between two consecutive local atrial activations [ 19 ] (the procedure is shown in Figure 2 ). Figure 2 Extraction of local atrial period (LAP) series from intra-atrial signal. (a) Example of a recorded electrogram during sinus rhythm before pre-processing; (b)-(c) the related LAP series, obtained as the sequence of temporal distances between two consecutive local atrial activations, after detecting time-instants of local atrial depolarization using a derivative / threshold algorithm. a.u. = arbitrary unit In analogy with previous studies [ 14 , 23 ], after canceling the ventricular interference, the absolute value of the output of the band-pass was low-pass filtered (50 Hz) and then sub-sampled (100 Hz) principally to reduce signal length and computation time. Considering the number of recorded signals and patients, we analyzed 79 recordings in SR and in SRISO and 77 recordings in AF and in AFISO. Twelve recordings were disregarded for the low quality of electrograms. For each recording, atrial signals were divided into six-second segments and then analyzed. Regularity index was estimated for each six-second segment in each recording site and for each patient, while synchronization index was estimated for each pair of close recording sites (interelectrode distance equal to one). Statistical analysis The statistical analysis was carried out using Student's t -test for paired data, comparing each rhythm before and after isoproterenol administration, and between organized (SR) and not organized rhythm (AF). Results Atrial signals Figure 3 shows an example of the distribution of regularity index (R) computed in one patient and in a single recording site (electrogram 8) during the four experimental conditions. The R values, computed in the six-second segments, are superimposed to their mean value. A significant reduction (p < 0.001) of the index passing from sinus rhythm to AF was detectable. Comparing the results obtained from the same rhythm with and without isoproterenol, a reduction of R was observed after drug administration in both sinus rhythm and atrial fibrillation. In this particular case, the decrease during sinus rhythm was statistically significant (p < 0.001). In particular, considering all patients recording sites, we observed 59 reductions (42 with p < 0.05) over 79 recordings passing from SR to SRISO and 40 (17 with p < 0.05) over 77 passing from AF to AFISO. This result reflects the global tendency of entire dataset, as illustrated in Table 1 , where the mean value obtained from all patients recording sites is showed, underlining a statistically significant reduction passing both from SR to AF and from SR to SRISO. Figure 3 Values of the R index in a single subject. Example of regularity (R) index computed for a patient in the electrogram 8 during the four phases of the analysis (SR, SRISO, AF, AFISO). Performance of the R index in the various six-second segments (circle) is superimposed to its mean value. A significant reduction of the R index can be observed passing both from sinus rhythm to atrial fibrillation and from sinus rhythm to sinus rhythm after isoproterenol administration. *p < 0.001 Table 1 Mean ± SD values of the proposed indexes in the four analyzed phases SR SRISO AF AFISO AS R 0.49 ± 0.08 0.46 ± 0.09 † 0.29 ± 0.09* 0.28 ± 0.08 S 0.28 ± 0.02 0.28 ± 0.03 0.20 ± 0.06* 0.20 ± 0.06 LAP R 0.75 ± 0.07 0.69 ± 0.10 † 0.31 ± 0.08* 0.26 ± 0.09 † LP 99.99 ± 0.001 99.97 ± 0.03 † 69.46 ± 21.55* 55 ± 24.75 † Mean ± SD values of the proposed indexes in the four analyzed phases: sinus rhythm (SR), sinus rhythm during isoproterenol infusion (SRISO), atrial fibrillation (AF) and atrial fibrillation during isoproterenol infusion (AFISO) for Atrial Signals (AS) and LAP series. † p < 0.0001 the comparison of a same rhythm before and during isoproterenol infusion. The t -test comparing organized (SR) to not organized (AF) rhythms always results significant (* p < 0.0001). Concerning the synchronization index (S), a significant decrease was observed only when comparing sinus rhythm to atrial fibrillation (Table 1 ). Evaluating results separately for every recording pair, we observed 35 decreases (7 with p < 0.05) over 69 values passing from SR to SRISO and 32 (9 with p < 0.05) over 66 passing from AF to AFISO. Local atrial period LAP series were analyzed using the level of predictability and the regularity index. Figure 4 illustrates an example of the LAP series during SR, SRISO, AF, AFISO and the corresponding NCCE function. The regularity ( R = 1 - min( NCCE )) decreases visibly passing from sinus rhythm to atrial fibrillation. A decrease in both SR and AF after isoproterenol administration is also observed. The mean values showed in Table 1 are obtained as all patients mean and they underline an analogous tendency. In particular, regularity reductions are found in 7 patients after isoproterenol administration during both sinus rhythm and atrial fibrillation (5 with p < 0.05 passing from SR to SRISO; 2 passing from AF to AFISO). Figure 4 LAP series in the four analyzed phases and the corresponding NCCE functions. Example of LAP series during (a) SR, (b) SRISO, (c) AF, (d) AFISO; (e)-(h) the corresponding NCCE functions (solid lines) depicted as sum of two terms: the decreasing one (CE, dotted line) and the increasing one (the corrective term, dash-dotted line). A clear increase in the minimum value (*) can be observed passing from SR to SRISO to AF to AFISO; therefore the R index = 1 - min( NCCE ) (see text for more details) decreases passing from SR to SRISO to AF to AFISO. Figure 5 shows the local atrial period series during SR, SRISO, AF, AFISO and the corresponding prediction errors. The prediction error increases passing from SR to SRISO to AF to AFISO. All patients mean reflects this tendency as shown in Table 1 . In particular, considering single patients, reductions of level of predictability is observed in 8 of them after isoproterenol administration during both sinus rhythm and atrial fibrillation; the reductions are statistically significant (p < 0.05) in 6 patients passing from SR to SRISO, and only in 4 passing from AF to AFISO. Figure 5 LAP series in the four analyzed phases and the corresponding prediction errors. Example of LAP series during (a) SR, (b) SRISO, (c) AF, (d) AFISO; (e)-(h) the corresponding prediction errors. A prediction error increase can be observed passing from SR to SRISO to AF to AFISO, showing the inability of the autoregressive model to predict the LAP series as the rhythm becomes less organized. This is equivalent to a reduction of the LP index = (1 - σ e / σ x ) passing from SR to SRISO to AF to AFISO. Discussion The autonomic nervous system plays an important role in the genesis and maintenance of atrial fibrillation, but characterization and quantification of its pro-arrhytmic effects are extremely complex and therefore difficult to define. Aim of this study was to evaluate the capability of linear and non-linear parameters to capture the fine changing in the dynamics of atrial signals and LAP series during adrenergic activation induced by the injection of a sympathomimetic drug. The existence of determinism and of an underlining order during AF has recently been shown. In particular both linear [ 14 - 16 ] and non-linear [ 17 , 18 ] patterns have been recognized. In the present study, where a relative small population is considered, we observed a reduction of spatial organization after isoproterenol administration both in sinus rhythm and in atrial fibrillation. In particular analysis of LAP series showed a significant decrease of both LP and R indexes within the same rhythm after isoproterenol administration (see Table 1 ). This reduction could be related to an increase of atrial wave fronts disorganization. In fact in previous studies [ 19 ], both indexes were demonstrated to decrease passing from SR to AF-I and AF-II Wells' classes. Therefore it can be argued that the reduction observed after isoproterenol infusion may be a sign of a high disorganization induced by the drug: atrial activation patterns become less periodic, less predictable and less regular. In fact, in agreement with our previous findings [ 19 ], the higher is the regularity and predictability of sequence of atrial activation, the fewer are the circulating 'mother' wavelets according to Jalife's model [ 24 ]. This finding is in agreement with known effects of sympathetic activation at atrial level [ 25 ] and may provide additional insights to the understanding of the pro-arrhythmic role of ANS in patients with AF. In addition, as previously reported [ 19 ], a marked reduction of R and LP indexes was observed passing from SR to AF. Concerning results obtained from atrial signals, both synchronization and regularity indexes showed a marked reduction passing from SR to AF well in keeping with previous findings [ 19 ] that documented the capability of these indexes to discriminate between organized and not-organized rhythms. However the two indexes were not able to evidence any differences after isoproterenol infusion. Only the R index was found significantly decreased in SR after drug administration. Nevertheless, the parameters revealed a tendency toward organization reduction after isoproterenol administration in both rhythms. In particular, responses to isoproterenol were more homogenous during sinus rhythm than during AF. This is maybe due to the fact that patients with a clinical history of AF, could already present alteration in atrial electrical properties likely to be involved in their predisposition to develop AF. Accordingly isoproterenol effects on dynamics of atrial signals are more evident in an organized rhythm (SR) than in AF, where an already disorganized rhythm can not be further fragmented by drug infusion. During AF in fact it has been suggested [ 1 ] that several wavefronts of electrical activity propagate through the atria in an irregular manner; this activity may partly obscure isoproterenol effects. Nevertheless, it has also been reported that atrial electrical activity may vary not only in relation to arrhythmia duration but also in relation to the structural characteristics of the atria [ 26 ]. In conclusion, the proposed set of linear and non-linear parameters is able to capture subtle changes in atrial dynamics during AF and drug infusion. These indexes could be employed to provide new insight into the mechanisms leading to initiation and maintenance of AF episodes.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529297.xml

514499

Stress echocardiography in heart failure

Echocardiography has the ability to noninvasively explore hemodynamic variables during pharmacologic or exercise stress test in patients with heart failure. In this review, we detail some important potential applications of stress echocardiography in patients with heart failure. In patients with coronary artery disease and chronic LV dysfunction, dobutamine stress echocardiography is able to distinguish between viable and fibrotic tissue to make adequate clinical decisions. Exercise testing, in combination with echocardiographic monitoring, is a method of obtaining accurate information in the assessment of functional capacity and prognosis. Functional mitral regurgitation is a common finding in patients with dilated and ischaemic cardiomyopathy and stress echocardiography in the form of exercise or pharmacologic protocols can be useful to evaluate the behaviour of mitral regurgitation. It is clinical useful to search the presence of contractile reserve in non ischemic dilated cardiomyopathy such as to screen or monitor the presence of latent myocardial dysfunction in patients who had exposure to cardiotoxic agents. Moreover, in patients with suspected diastolic heart failure and normal systolic function, exercise echocardiography could be able to demonstrate the existence of such dysfunction and determine that it is sufficient to limit exercise tolerance. Finally, in the aortic stenosis dobutamine echocardiography can distinguish severe from non-severe stenosis in patients with low transvalvular gradients and depressed left ventricular function.

Background The identification of viable hibernating myocardium in patients with coronary artery disease and chronic left ventricular (LV) dysfunction is, up to today, the most common use of stress echocardiography in patients with heart failure. However, to search viable myocardium or the presence of contractile reserve is only one of plugs of the physiopathologic puzzle in a failing heart (Figure 1 and 2 ). If we consider the ability of echocardiography to provide valuable haemodynamic information accurately and non-invasively, it is ideally suited for application during stress testing to objectively assess other physiopathologic components of heart failure. These include the study of exercise physiology, the presence and the behaviour of concomitant mitral regurgitation (MR), the prediction of response to resynchronization therapy etc. Figure 1 Physiopathologic components of systolic heart failure that can be potentially explored with stress echocardiography. Figure 2 Physiopathologic components of diastolic heart failure assessable with stress echo. Therefore, the present review will detail some important potential applications of stress echocardiography in patients with heart failure in the evaluation of the different clinical and physiopathologic aspects of heart failure syndrome. Systolic heart failure Searching the myocardial viability The most common cause of heart failure in the Western world is coronary artery disease, accounting for up to 60% of cases [ 1 ]. In patients with coronary artery disease and chronic LV dysfunction, it is crucial to distinguish between viable and fibrotic tissue to make adequate clinical decisions. Noncontractile but viable myocardium may correspond to different states that are important but difficult to distinguish, i.e., ischemia, stunning, nontransmural infarction, or hibernation and in individual patients these pictures may coexist [ 2 ]. After brief episodes of coronary occlusion and reflow a reversible global LV dysfunction can occur. This phenomenon was called myocardial stunning [ 3 ]. It is characterized as prolonged mechanical dysfunction after coronary reflow despite resumption of normal perfusion and lack of permanent tissue damage. Stunning seems to result from alterations in contractile proteins in response to sublethal ischemic insults. This phenomenon can occur in several settings, including after acute reperfused myocardial infarction and after CABG. In humans, the return of functional recovery may require days to weeks [ 4 ]. Hence, diagnostic methods to distinguish stunning from necrosis are particularly relevant for clinical investigation and management in patients with acute, severe LV dysfunction or cardiogenic shock after revascularization. Persistent wall motion abnormalities can be observed by echocardiography at a time when chest pain, ST segment deviation, and regional perfusion had recovered. The presence of contractile reserve during dobutamine infusion identifies the stunning but viable myocardium from myocardial necrosis. The term " Hibernating myocardium " was first termed by Rahimtoola to indicate the state of reversible dysfunctional myocardium, which was considered to be the result of a state of persistently impaired myocardial function at rest, caused by reduced coronary blood flow, and which could be partially or completely restored to normal either by improving blood flow or reducing oxygen demand [ 5 ]. Echocardiography can detect viable myocardium during infusion of drugs which have ability to elicit an enhanced contractile response by recruiting contractile proteins. At least two drugs have these proprieties: the dobutamine, a synthetic β1 agonist with additional α1- and β2-stimulating properties and the enoximone that inhibits cyclic adenosine monophosphate-specific phosphosdiesterase [ 6 , 7 ]. Routinely, the dobutamine is the most common stressor used, whereas the enoximone is particularly useful in patients on beta-blocker therapy [ 7 , 8 ]. The mechanism by which dobutamine stimulation elicits a contractile response in hypoperfused dysfunctional segments without precipitating ischemia has been demonstrated by Sun et al. [ 6 ]. By using positron emission tomography and echocardiography, they demonstrated that the improvement in contractile function during dobutamine infusion was associated with a concomitant increase in myocardial blood flow. The increase in myocardial blood flow occurs because there is persistent, albeit reduced, coronary flow reserve distal to a stenosis which dobutamine may exploit. Another mechanism whereby contractile response may be elicited during dobutamine infusion is through its peripheral vasodilator effect, which causes reduction in LV end-systolic wall stress by reducing afterload [ 6 ]. Moreover, dipyridamole echocardiography (up to 0.84 mg/kg over 10 minutes) can identify regions with myocardial viability [ 9 ]. Dipyridamole leads to transiently increased coronary flow, which leads to improved contractility in viable myocardium [ 9 ]. A small study comparing dipyridamole with dobutamine revealed 93% concordance [ 10 ]. Combined dipyridamole-dobutamine (low-dose dipyridamole followed by low-dose dobutamine) has also been proposed and found to recruit a contractile reserve in some asynergic segments that were nonresponders after dobutamine or dipyridamole alone [ 11 ]. An initial evaluation of end diastolic wall thickness of akinetic segments with resting echocardiography can be used as an initial screening technique for assessment of viability. Indeed, akinetic regions with an end diastolic wall thickness <6 mm do not contain viable myocardium and do not improve in function after revascularization [ 12 ]. However, in segments with a thickness ≥ 6 mm, additional testing is needed because approximately 40% of these regions do not contain viable myocardium and will not improve after revascularization [ 12 ]. Therefore, myocardial thinning should not be equated with the lack of myocardial viability, and in some patients, these regions can improve in contractile function after revascularization [ 13 ]. The detection of subendocardial infarcts became clinically relevant because the quantification of non-viable myocardium in addition to viable myocardium in that region of LV is important in predicting contractile improvement following revascularization. Thus, the ratio of viable to total myocardium (viable plus non-viable) in the dysfunctional region was more accurate that absolute amount of viable myocardium alone in predicting functional improvement [ 13 ]. Unfortunately, currently available techniques, such as single photon emission computed tomography, dobutamine stress echocardiography and positron emission tomography are still insufficient to provide a comprehensive assessment including the evaluation of subendocardial infraction with respect to magnetic resonance imaging [ 14 ]. During stress echocardiography is possible to observe four response patterns based on regional wall function: normal, ischemic, viable and necrotic. In the normal response, a segment is normokinetic at rest and normal or hyperkinetic during stress. In the ischemic response, a segment worsens its function during stress from normokinesis to dyssynergy. In the necrotic response, a segment akinesia remains akinetic during stress. In the viability response, a segment with resting dysfunction improves during stress. During pharmacologic stress, a viable response at low dose can be followed by ischemic response at high dose (biphasic response). This biphasic response is suggestive of viability and ischemia, with jeopardized myocardium fed by a critically stenosed coronary artery [ 15 ]. A resting akinesia which becomes dyskinesia during stress reflects a purely passive mechanical phenomenon and should not be considered a true active ischemia. The overall sensitivity and specificity of dobutamine echocardiography for predicting recovery of regional function after revascularization was 84% and 81% respectively [ 16 ]. In a study by Afridi et al., the best predictive value for recovery of function after revascularization was most often noted in patients demonstrating an ischemic response during low and high doses of dobutamine infusion [ 17 ]. On the other hand, sustained improvement of regional function during dobutamine infusion was a poor marker of recovery function. Sensitivity of dobutamine echocardiography may be affected by several factors such as the severe reduction of myocardial blood flow that can preclude the contractile response, the premature interruption of dobutamine infusion, resting tachycardia that may renders the myocardium ischemic and dobutamine can only augment ischemia [ 16 ]. On the contrary, the specificity may be affected by the tethering effect, the injured subendocardial portion of myocardium that does not respond to dobutamine when the infarction is confined to subendocardium, and also specificity may be reduced in myocardial regions that do not develop an ischemic response [ 16 ]. The main clinical issue to search the myocardial viability is that patients with evidence of hibernating myocardium who do not undergo revascularization have poor prognosis with high incidence of cardiac events [ 18 ]. In contrast, evidence of viable myocardium in patients undergoing successful revascularization is associated with longer survival and improvement of both symptoms and LV function [ 19 ]. However, the presence of myocardial viability is only relevant in patients with severely depressed LV function and has a prognostic impact only when a significant amount of viable myocardium is present. Therefore, the final end point of searching the myocardial viability is to predict the recovery of global myocardial function after revascularization. At this purpose there is a relation between improvement in left ventricular ejection fraction (LVEF) and the number of segments with contractile reserve, indicating that extent of jeopardized but viable myocardium determine the magnitude of improvement of LV function after revascularization. Usually a level of ≥ 4 viable segments, which corresponds an improvement in wall motion score index >0.25 (about 20% of left ventricle), is advised as a cutoff value to predict improvement of LVEF [ 20 ]. However, despite the presence of substantial viability, in some patients LVEF does not improve after revascularization because not only the amount of dysfunctional but viable tissue but also LV remodelling and enlargement determines the improvement in function following revascularization [ 21 ]. Thus, patients with a high end systolic volume (≥ 140 ml) due to LV remodelling have a low likelihood of improvement of global function [ 21 ] (Figure 3 ). Figure 3 A schematic flow chart for searching segmental and global systolic function in chronic ischemic LV dysfunction. Assessing the functional capacity In most patients with chronic heart failure, symptoms are not present at rest but become limiting with exercise. Despite this, the major measures used to characterise the symptoms, the severity, the mechanisms and the prognosis of heart failure are obtained at rest. Exercise testing, in combination with echocardiographic monitoring, may be an attractive and practical method of obtaining accurate information which can aid in the diagnosis of heart failure as well as the assessment of functional limitation and prognosis. Exercise rather than dobutamine is the stressor of choice to evaluate functional capacity due to the possibility to combine echocardiographic variables with common parameters available during physiologic exercise. Symptom limited exercise testing can be undertaken using either treadmill or bicycle exercise protocols. Available data about the safety of exercise testing in patients with significant heart failure suggest a very low incidence of serious adverse events such as arrythmias or hypothension. The echocardiographic monitoring during exercise testing may have an additional value overt the conventional parameters assessed during exercise testing such as functional capacity, symptoms and peak oxygen uptake that become part of the final interpretation. Indeed, several haemodynamic parameters can be noninvasively obtained with echocardiography such as LVEF at rest and during stress deriving the contractile reserve, the behaviour of mitral valve function, the pulmonary artery pressure, the right ventricular function, the diastolic function (Table 1 ). In this way, it is possible to observe the variation of the monitored variables and to correlate these with the appearance of symptoms, i.e. impairment of global contractile function followed by increase in pulmonary pressure with dyspnoea. The critical level to define the presence of contractile reserve is generally defined as an increase of at least 5% (in absolute terms) in the global LVEF [ 22 ] (Figure 4 ). The change in the systolic pulmonary artery pressure (sPAP) at rest and during exercise is among others, the most frequently utilized echocardiographic variable. It can reliably be estimated by adding the right atrial pressure derived from the tricuspid reguritation jet velocity [ 23 ]. The right atrial pressure can be estimated at rest by the response of inferior vena cava to deep inspiration and assumed to be constant throughout exercise. Sometimes the use of echocardiographic contrast agents such as agitated saline solution may help to enhance Doppler signals. Pulmonary hypertension determined by echocardiography has been defined as a peak of sPAP >30 mmHg at rest and >45 mmHg during exercise [ 24 ]. Right ventricular dysfunction predicts impaired exercise capacity and decreased survival in patients with both moderate and advanced heart failure [ 25 ]. There are several clinically validated methods to detect right ventricular dysfunction. Tricuspid annular plane systolic excursion (TAPSE) visualized from the apical four-chamber view is an easy measure and can be used a surrogate of right ventricular function. A TAPSE value of 14 mm or less means the presence of right ventricular dysfunction and is a significant adverse prognostic indicator [ 26 ]. More recently, the evalution of tricuspid systolic annular tissue Doppler velocity has been introduced as index of right ventricular function and a value less than 10.8 cm/s indicates patients with abnormal right ventricular function [ 27 ]. Table 1 Potential parameters obtainable during exercise echocardiography. Common variables during exercise test Additional echocardiographic variables during exercise test Duration of exercise Contractile reserve Peak VO2 Mitral valve function Anaerobic threshold Pulmonary systolic pressure Oxygen pulse Right ventricular function VO2 workload ratio Diastolic function O2 saturation Figure 4 Echocardiographic apical four-chamber images (end-systolic frames) from two patients with and without contractile reserve. Moreover, these evaluations are useful not only for the diagnosis but also for predicting the outcome of patients overt the symptoms of heart failure. Indeed, some patients with marked reduction in myocardial contractility at rest, but with good residual contractile reserve, have a favourable exercise capacity and prognosis, whereas patients with mild symptoms and similar degree of abnormal myocardial contractility at rest, but without contractile reserve, have poor outcome [ 28 ]. Although a VO2max <14 ml/kg/min is well known as a measure for deciding on eligibility for cardiac transplantation, it has been clearly shown that there is no absolute threshold for adverse prognosis and that VO2max uptake should be considered as a continuous variable. In term of discriminating survivors from non survivors, it appears that VO2max <10 ml/kg/min definitively defines high risk, while a value >18 ml/kg/min defines low risk; those values in between may represent a grey zone. Thus stress echocardiography yields the greatest incremental prognostic value in patients with intermediate values of VO2max (10–14 ml/Kg/min) and helps to further stratify the risk of patients with intermediate (Table 2 ) [ 29 ]. Table 2 Additive prognostic value of stress echo in patients with intermediate values of VO2max (10–14 ml/Kg/min). Risk Low (5–10% year) High (≥ 25–30% year) Exercise capacity ≥ 8–10 min <8 min Contractile reserve yes no Pulmonary hypertension <45 mmHg >45 mmHg Right ventricular dysfunction no yes Mitral regurgitation ↓ or = ↑↑ Looking at the behaviour of mitral valve MR is a common finding in heart failure patients. In patients with dilated and ischaemic cardiomyopathy, the MR is typically functional and reflects geometric distortions of LV chamber, which displaces the normal valve and subvalvar closing mechanisms. This functional MR is a consequence of adverse LV remodelling and increased sphericity of the chamber. It is typically dynamic and a marker of adverse prognosis. The 5-year survival of heart failure patients with functional MR ranges from 39.9% to 48.7% depending on the degree of MR [ 30 ]. Stress echocardiography in the form of exercise or pharmacologic protocols can be useful in the assessment of MR. Exercise echocardiography is usually preferred due to the possibility to reproduce physiological setting. Even though supine bike protocol allows to obtain good image acquisition, upright bicycle or treadmill protocols are more frequently utilized in the practical setting. Treadmill exercises can be performed using the standard protocols such as Bruce or modified Bruce, while gradual increase in the bike workload of 20–25 W every 2–3 minutes is often applied until the patients achieves either the target heart rate or develops symptoms of fatigue or shortness of breath. Sometimes, pharmacologic stress is used with dobutamine protocol at low or intermediate doses infusion. In the collection of echocardiographic data should be included: the MR jet to evaluate the MR jet area and the vena contracta width, the velocity time integrals (mitral and aortic) to calculate the regugitant volume and the effective regurgitant orifice area (EROA), the LV volumes to assess the myocardial contractility and the tricuspid regurgitant jet velocity to measure the sPAP that is an useful index of the haemodynamic burden of MR. Exercise echocardiography can play several roles in the assessment of the behaviour of mitral valve in heart failure patients. First , in symptomatic patients with LV dysfunction and a clinical picture suspicious for new or worsening MR, but not evident at resting echo examination, exercise echocardiography can demonstrate a worsening of MR which helps to correlate this scenario with the patient's symptoms (Figure 5 ) [ 31 ]. Second , LV contractility, in presence of MR, can impair or improve during exercise with consequent modification of MR. Patients with presence of contractile reserve show a decrease in MR [ 32 ], whereas generally a fall in stroke volume is associated with an increase in mitral regurgitant volume during isometric exercise, which increases systemic resistances and thereby afterload [ 33 ]. These observations support the concept of presence of a vicious circle between LV function and behaviour of MR. Therefore, to study these patients with exercise echocardiography may be important for assessing the response of MR to medical therapy and for the following prognostic implications. Indeed, third , in patients with ischemic MR and LV dysfunction, quantitative assessment of exercise-induced changes in the degree of MR provides independent prognostic information. Significant exercise-induced increases in MR (increase in ERO ≥ 13 mm 2 ) unmask patients at high risk of poor outcome. The cardiac mortality rate of medically treated patients with dynamic MR during exercise is 39% at 20 months which represents excess mortality in patients in functional class II or III (Figure 6 ) [ 34 ]. Figure 5 Apical four-chamber view at rest and during exercise in patients with ischemic mitral regurgitation showing a large exercise-induced increase in mitral regurgitation. SPAP = Systolic pulmonary pressure. Figure 6 Relationship between contractile reserve, mitral regurgitation and pulmonary pressure and its contribution in defining the prognosis in patients with functional mitral regurgitation. MR = mitral regurgitation; sPAP = systolic pulmonary pressure. Finally, dobutamine protocol has a different role in the contest of ischemic MR. Generally, in this setting it is used to evaluate the behaviour of MR in relation with the presence or absence of myocardial viability (Figure 7 ). Dobutamine infusion has the ability to decrease MR volume due to a reduction of afterload and mitral orifice size that may occur as a result of the vasodilatory and inotropic effects of dobutamine [ 35 , 36 ]. Therefore, if during dobutamine protocol we find myocardial viability and a concomitant reduction of MR, these results should be interpreted with caution because we cannot assume a direct effect of the presence of myocardial viability on the MR. Thus, the complex interplay between haemodynamic effects of dobutamine, myocardial viability and behaviour of MR has to be taken in mind during clinical management of patient with LV dysfunction and ischemic MR, i.e. revascularization alone versus revascularization plus mitral valve surgery. Figure 7 Targets and effects of dobutamine stress echo in patients with mitral regurgitation and chronic ischemic left ventricular dysfunction. EROA = Effective regurgitant orifice area. Evaluating the contractile reserve beyond hibernating myocardium It is commonly believed that the assessment of contractile reserve is only confined and clinically useful to search the myocardial viability in patients with LV dysfunction and coronary artery disease. Growing published data suggest the utility in searching the presence of contractile reserve in non ischemic dilated cardiomyopathy (DCM). While in the ischemic cardiomyopathy the search of myocardial viability is focused to find the presence of reversible segmental myocardial dysfunction and its possible effect on global systolic LV recovery after revascularization, in DCM the primary end point is to evaluate the presence of residual global contractile reserve. Both dobutamine and exercise testing have been used in the study patients with DCM, but there is a clear predominance for the use of dobutamine test. The doses of dobutamine utilized vary from investigators, but safety in its use in this population has been documented in doses as high as 40 μg/kg per minute. In the interpretation of results both wall motion score index and the LV volume to derive LVEF must be calculated. LV systolic function at the time of diagnosis has been proposed to be the strongest predictor of survival in DCM, but now the presence of contractile reserve recognised by dobutamine echocardiography seems to be the best marker of good prognosis in patients with severe LV dysfunction at rest [ 37 , 38 ]. Patients with significant improvement in their wall motion score index and LVEF during dobutamine infusion have a better survival rate and increase in the LVEF during follow-up period [ 37 ]. The data extracted from dobutamine study can be used as an adjunct or alternative to predict VO2max and exercise capacity of patients with heart failure, especially when the patients fall into the gray zone of VO2max (10–14 ml/kg/min) or when there is limitation to ambulation [ 29 ]. Moreover, the response to dobutamine infusion predicts the improvement in LVEF with beta-blocker therapy in patients with advanced heart failure. Patients with contractile reserve experienced a greater improvement in LVEF with beta-blocker by biologically augmenting myocyte a chamber contractility [ 39 ]. Whereas, in the absence of contractile reserve (when myocytes have been replaced by fibrosis), ventricular function cannot improve by this biological mechanism because there are not enough contractile units and the sympatholytic effects of beta-blocker prevail [ 39 ]. However, the clinical use of dobutamine stress echocardiography in patients with chronic heart failure may be limited by a substantial proportion of patients already receiving beta-blocker therapy at time of evaluation. In these patients enoximone echocardiography might be a valid alternative to low-dose dobutamine for evaluating contractile reserve showing a more potent and a better safety profile than dobutamine [ 8 ]. Stress echocardiography may also help in the identification of patients in the initial phase of cardiomyopathy overt normal resting echocardiographic parameters. Both dobutamine and exercise have to be used to screen for the presence of latent myocardial dysfunction in patients who had exposure to cardiotoxic agents [ 40 ]. Diastolic heart failure The prevalence of diastolic heart failure in the community is now to be at least as high as that reported in previous studies of hospitalised patients; almost half of all patients with heart failure have diastolic heart failure [ 41 ]. The term asymptomatic diastolic dysfunction is used to refer to an asymptomatic patient with a normal LVEF and abnormal echo-Doppler pattern of LV filling; this is often seen, for example, in patients with hypertensive heart disease. If such patients exhibit symptoms of effort intolerance and dyspnoea, especially if there are evidence of venous congestion and edema, the term diastolic heart failure can be used [ 42 ]. Resting echocardiography is most useful in the assessment of LV size, LVEF and the use of Doppler-derived indices of diastolic function has impact on the identification of diastolic dysfunction. However, to determine whether an abnormality of diastolic function is the cause of the patient's symptoms, we need to demonstrate the existence of such dysfunction and determine that it is sufficient to limit exercise tolerance. Therefore, the stress echocardiography, in particular exercise echocardiography could be useful in dyspnoeic patients with apparently normal LV function to unmask the presence of diastolic dysfunction (signs of elevated LV filling pressure) during exercise as cause of symptoms. Patients with diastolic heart failure, as well as those with diastolic dysfunction and little or no congestion, exhibit exercise intolerance for several reasons. First, an elevated LV diastolic and pulmonary venous pressure during exercise causes reduction in lung compliance, which increases work of breathing and evokes the symptom of dyspnoea [ 42 ]. Second, a substantial number of patients who have LV hypertrophy, high relative wall thickness and small end diastolic volume exhibit a low stroke volume and a depressed cardiac output [ 43 ]. These hearts exhibit a limited ability to utilize the Frank-Starling mechanism during exercise. Such limited preload reserve, specially if coupled with the chronotropic incompetence limits the cardiac output during exercise [ 44 ]. Third, elevated LV diastolic and pulmonary venous pressures in patients with normal LVEF are directly related to abnormalities in the diastolic proprieties of the ventricle. This is not to say contractile function is entirely normal, but subtle and latent abnormalities of contractile function could be present in many patients, in whom, however, diastolic dysfunction is the dominant feature [ 42 ]. All these aspects can be assessed during exercise echocardiography (Table 3 ). In particular the assessment of diastolic function during exercise has been shown to be feasible [ 45 ]. Combining transmitral flow velocity with annular velocity obtained at level of the mitral annulus with tissue Doppler (E/E') has been proposed as a tool for assessing LV filling pressures that combines the influence of transmitral driving pressure and myocardial relaxation [ 46 ]. Patients with rest E/E' >15 can be classified as having elevated filling pressure. A rest E/E' <8 suggests normal filling pressure and a range of 8 to 15 represents a gray zone. E and E' velocities increased significantly after exercise. In normal subjects because of proportional increases of both velocities, the E/E' ratio do not change significantly with exercise; this observation can be taken as a normal diastolic response during exercise [ 45 ]. Indeed, if E/E' ratio increases up to 15 we can suppose a pathological increase of LV filling pressure during exercise. This evaluation must be associated to the assessment of cardiac output and sPAP during exercise with appearance of symptoms. Finally, with the evaluation of systolic LV function during exercise it is possible to discover the portion of patients with concomitant latent myocardial dysfunction but predominant diastolic abnormality (Figure 8 ). Table 3 Useful echocardiographic parameters to evaluate diastolic function during exercise test in patients with suspected diastolic heart failure. Transmitral Doppler indices E/E' ratio Cardiac output Preload reserve Contractile reserve Pulmonary systolic artery pressure Figure 8 Schematic diagnostic algorithm in patients with suspected diastolic heart failure. LV = Left ventricle; LVEF = Left ventricular ejection fraction. Aortic stenosis with left ventricular dysfunction Stress echocardiography with dobutamine infusion is particularly useful in clinical decision making in patients with aortic stenosis with LV dysfunction and low transvalvular gradients. In this group of patients, an important clinical question rises: is the low gradient a consequence of low cardiac output due to a severe aortic stenosis which has led to LV dysfunction or is the low gradient a consequence of LV dysfunction is unrelated to aortic stenosis and the aortic stenosis is an incidental finding? It is well known that the transvalvular gradients are flow-depentent parameters so that they are influenced by LV function. The aortic valve area can be accurately determined by Doppler echocardiography with continuity equation and that correlate well with Gorlin formula [ 47 ]. However, it has been shown that valve areas calculated by the Gorlin formula is flow-dependent and usually increase with flow, probably due to the flow dependence of the empirical constant C of the Gorlin formula, which represents the ratio of effective to anatomical orifice area. Burwash et al., with dobutamine stress-echocardiography, demonstrate a flow-dependent increase in actual orifice aortic valvular area calculated with continuity equation [ 48 ]. Therefore, the assessment of valve area does not solve the diagnostic dilemma in these patients, because we cannot distinguish between severe fixed from flow-dependent (relative) aortic stenosis. Thus, it is important to perform pharmacological manoeuvres to increase cardiac output so that valve area can be calculated at higher flow rate. Dobutamine stress echocardiography until 20 γ/kg/min with concomitant evaluation of cardiac output, aortic valve area and gradients, is a useful and reliable test to distinguish between severe fixed from relative aortic stenosis in presence of low gradient and LV dysfunction. On the basis on test results, it is possible to distinguish 3 groups of patients [ 49 ] (Figure 9 ): 1. Patients with an improvement of contractile function but no significant increase in valve area and an increase of transvalvular gradients. These patients have severe fixed aortic stenosis and are good candidate for surgery with an acceptable peri-operative surgical risk. 2. Patients with contractile reserve with an increase of aortic valve area without substantial increase in transvalvular gradients. These patients have a non critical aortic stenosis and the LV dysfunction is not related to the aortic stenosis and should be treated conservatively. 3. Finally, patients without contractile reserve and no modification of valve area and transvalvular gradients. These patients represent an ambiguous group, because can represent patients with end-stage severe aortic stenosis with severe LV dysfunction or patients with severe LV dysfunction without contractile reserve and incidental aortic stenosis. However, this group has very poor prognosis. Figure 9 Possible results during dobutamine stress echocardiography in presence of aortic stenosis, low cardiac output and low transvalvular gradients. AVA: Aortic valve area; CO: Cardiac output. When interpreting the results of a dobutamine study in these patients to rule out or confirm definitively the presence of a severe fixed aortic stenosis, it is advisable to use an absolute cut-off value of the aortic valve area at peak of dobutamine >1 cm 2 rather than an increase of ≥ 0.3 cm 2 from baseline alone [ 49 , 50 ]. Conclusions Beyond the identification of viable hibernating myocardium, stress echocardiography is particular useful in patients with systolic and diastolic heart failure to assess the different physiopathologic component of heart failure syndrome and can aid to an appropriate clinical decision making. List of abbreviations LV : left ventricular MR: mitral regurgitation LVEF: left ventricular ejection fraction DCM: dilated cardiomyopathy TAPSE : Tricuspid annular plane systolic excursion sPAP : Systolic pulmonary artery pressure Competing interests The manuscript is not under consideration elsewhere and the data presented have not been previously published. All authors have read and approved the manuscript. No financial support was received for this study. The content of this manuscript is not associated with any financial interest or other relations that could lead to a conflict of interest. Author's contribution Concerning the authorship, the listed authors have contributed as follows to the manuscript: EA and MP: 1) conception, design, analysis and interpretation of data, 2) drafting of the manuscript and 3) final approval of the manuscript MO and AM: 1) critical revision of the manuscript for important intellectual content, and 3) final approval of the manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514499.xml

539279

Proportional odds ratio model for comparison of diagnostic tests in meta-analysis

Background Consider a meta-analysis where a 'head-to-head' comparison of diagnostic tests for a disease of interest is intended. Assume there are two or more tests available for the disease, where each test has been studied in one or more papers. Some of the papers may have studied more than one test, hence the results are not independent. Also the collection of tests studied may change from one paper to the other, hence incomplete matched groups. Methods We propose a model, the proportional odds ratio (POR) model, which makes no assumptions about the shape of OR p , a baseline function capturing the way OR changes across papers. The POR model does not assume homogeneity of ORs, but merely specifies a relationship between the ORs of the two tests. One may expand the domain of the POR model to cover dependent studies, multiple outcomes, multiple thresholds, multi-category or continuous tests, and individual-level data. Results In the paper we demonstrate how to formulate the model for a few real examples, and how to use widely available or popular statistical software (like SAS, R or S-Plus, and Stata) to fit the models, and estimate the discrimination accuracy of tests. Furthermore, we provide code for converting ORs into other measures of test performance like predictive values, post-test probabilities, and likelihood ratios, under mild conditions. Also we provide code to convert numerical results into graphical ones, like forest plots, heterogeneous ROC curves, and post test probability difference graphs. Conclusions The flexibility of POR model, coupled with ease with which it can be estimated in familiar software, suits the daily practice of meta-analysis and improves clinical decision-making.

Background A diagnostic test, in its simple form, tries to detect presence of a particular condition (disease) in a sample. Usually there are several studies where performance of the diagnostic test is measured by some statistic. One may want to combine such studies to get a good picture of performance of the test, a meta-analysis. Also, for a particular disease there may be several diagnostic tests invented, where each of the tests is subject of one or more studies. One may also want to combine all such studies to see how the competing tests are performing with respect to each other, and choose the best for clinical practice. To pool several studies and estimate a summary statistic some assumptions are made. One such assumption is that differences seen between individual study results are due to chance (sampling variation). Equivalently, this means all study results are reflecting the same "true" effect [ 1 ]. However, meta-analysis of studies for some diagnostic tests show that this assumption, in some cases, is not empirically supported. In other words, there is more variation between the studies that could be explained by random chance alone, the so-called "conflicting reports". One solution is to relax the assumption that every study is pointing to the same value. In other words, one accepts explicitly that different studies may correctly give "different" values for performance of the same test. For example, sensitivity and specificity are a pair of statistics that together measure the performance of a diagnostic test. One may want to compute an average sensitivity and an average specificity for the test across the studies, hence pooling the studies together. Instead, one may choose to extract odds ratio (OR) from each paper (as test performance measure), and then estimate the average OR across the studies. The advantage is that widely different sensitivities (and specificities) can point to the same OR. This means one is relaxing the assumption that all the studies are pointing to the same sensitivity and specificity, and accepts that different studies are reporting "truly different" sensitivity and specificity, and that the between-study variation of them is not due to random noise alone, but because of difference in choice of decision threshold (the cutoff value to dichotomize the results). Therefore the major advantage of OR, and its corresponding receiver-operating-characteristic (ROC) curve, is that it provides measures of diagnostic accuracy unconfounded by decision criteria [ 2 ]. An additional problem when pooling sensitivities and specificities separately is that it usually underestimates the test performance [[ 3 ], p.670]. The above process may be used once more to relax the assumption that every study is pointing to the same OR, thus relaxing the "OR-homogeneity" assumption. In other words, in some cases, the remaining variation between studies, after utilizing OR as the summary performance measure, is still too much to be attributed to random noise. This suggests OR may vary from study to study. Therefore one explicitly assumes different studies are measuring different ORs, and that they are not pointing to the same OR. This difference in test performance across studies may be due to differences in study design, patient population, case difficulty, type of equipment, abilities of raters, and dependence of OR on threshold chosen [ 4 ]. Nelson [ 5 ] explains generating ROC curves that allow for the possibility of "inconstant discrimination accuracy", a heterogeneous ROC curve (HetROC). This means the ROC curve represents different ORs at different points. This contrasts with the fact that the homogeneous-ROC is completely characterized by one single OR. There are a few implementations of the heterogeneous ROC. One may classify them into two groups. The first group is exemplified by Tosteson and Begg [ 6 ]. They show how to use ordinal regression with two equations that correspond to location and scale. The latent scale binary logistic regression of Rutter and Gatsonis [ 4 ] belong to this group. The second group contains implementations of Kardaun and Kardaun [ 7 ], and Moses et al [ 8 ]. Moses et al explain a method to plot such heterogeneous ROC curve under some parametric assumptions, and they call it summary ROC (SROC). When comparing two (or more) diagnostic tests, where each study reports results on more than one test, the performance statistics (in the study results) are correlated. Then standard errors computed by SROC are invalid. Toledano and Gatsonis [ 9 ] use the ordinal regression model, and account for the dependency of measurements by generalized estimating equations (GEE). However, to fit the model they suggest using a FORTRAN code. We propose a regression model that accommodates more general heterogeneous ROC curves than SROC. The model accommodates complex missing patterns, and accounts for correlated results [ 10 ]. Furthermore, we show how to implement the model using widely available statistical software packages. The model relaxes OR-homogeneity assumption. In the model, when comparing two (or more) tests, each test has its own trend of ORs across studies, while the trends of two tests are (assumed to be) proportional to each other, the "proportional odds ratio" assumption. We alleviate dilemma of choosing weighting schemes such that do not bias the estimates [[ 11 ], p.123], by fitting the POR model to 2-by-2 tables. The model assumes a binomial distribution that is more realistic than a Gaussian used by some implementations of HetROC. Also, it is fairly easy to fit the model to (original) patient level data (if available). Besides accounting better for between-study variation, we show how to use the POR model to "explain why" such variation exists. This potentially gives valuable insights and may have direct clinical applications. It may help define as to when, where, how, and on what patient population to use which test, to optimize performance. We show how to use "deviation" contrast, in parameterization of categorical variables, to relax the restriction that a summary measure may be reported only if the respective interaction terms in the model are insignificant. This is similar to using grand mean in a "factor effects" ANOVA model (compared to "cell means" ANOVA model). We show how to use nonparametric smoothers, instead of parametric functions of true positive rate (TPR) and/or false positive rate (FPR), to generate heterogeneous ROC for a single diagnostic test across several studies. Our proposed POR model assumes the shape of the heterogeneous ROC curve is the same from one test to the other, but they differ in their locations in the ROC space. This assumption facilitates the comparison of the tests. However, one may want to relax the POR assumption, where each test is allowed to have a heterogeneous ROC curve with a different shape. One may implement such generalized comparison of the competing diagnostic tests by a mixed effects model. This may improve generalizability of meta-analysis results to all (unobserved) studies. Also, a mixed effects model may take care of remaining between-study variation better. Methods Average difference in performances To compare two diagnostic tests i and j, we want to estimate the difference in their performance. However, in reality such difference may vary from one paper (study) to the other. Therefore Δ i,j,p = PERF i,p - PERF j,p , where the difference Δ depends on paper index p, where PERF i,p is observed performance of test i in paper p. To simplify notation, assume that a single number measures performance of each test in each paper. We relax this assumption later, allowing for the distinction between the two types of mistakes (FNR and FPR, or equivalently TPR and FPR). We decompose the differences (1) Δ i,j,p = PERF i,p - PERF j,p = δ i,j + δ i,j,p , where δ i,j is the 'average' difference between the two tests, and δ i,j,p is deviation of the observed difference within paper p from the average δ i,j . The δ i,j is an estimator for the difference between performance of the two tests. Note by using deviation parameterization (similar to an ANOVA model) [[ 12 ], pp.51 & 45] we explicitly accept and account for the fact that the observed difference varies from one paper to the other, while estimating the 'average' difference. This is similar to a random-effects approach where a random distribution is assumed for the Δ i,j,p and then the mean parameter for the distribution is estimated. In other words, one does not need to assume 'homogeneous' difference of the two tests across all the papers, and then estimate the 'common' difference [ 13 ]. The observed test performance, PERF, may be measured in several different scales, such as paired measures sensitivity and specificity, positive and negative predictive values, likelihood ratios, post test odds, and post test probabilities for normal and abnormal test results; as well as single measures such as accuracy, risk or rate ratio or difference, Youden's index, area under ROC curve, and odds ratio (OR). When using OR as the performance measure, the marginal logistic regression model (2) logit(Result pt ) = β 0 + β 1 * Disease pt + β 2 * PaperID pt + β 3 * Disease pt * PaperID pt + β 4 * TestID pt + β7 * Disease pt * TestID pt + β 6 * TestID pt * PaperID pt + β 7 * Disease pt * TestID pt * PaperID pt implements the decomposition of the performance. Model (2) is fitted to the (repeated measures) grouped binary data, where the 2-by-2 tables of gold-standard versus test results are extracted from each published paper. In the model (2) Result is an integer-valued variable for positive test result (depending on software choice, for grouped binary data, usually Result is replaced by number of positive test results over the total sample size, for each group); Disease is an indicator for actual presence of disease, ascertained by the gold standard; PaperID is a categorical variable for papers included in the meta-analysis; and TestID is a categorical variable for tests included. Regression coefficients β 2 to β 7 can be vector valued, meaning having several components, so the corresponding categorical variables should be represented by suitable number of indicator variables in the model. Indexes p and t signify paper p and test t. They define the repeated measures structure of the data [ 10 ]. Note model (2) fits the general case where there are two or more tests available for the disease, where each test has been studied in one or more papers. Some of the papers may have studied more than one test; hence the results are not independent. Also the collection of tests studied may change from one paper to the other, hence incomplete matched groups. From model (2) one can show that LOR pt = β 1 + β 3 * PaperID pt + β 5 * TestID pt + β 7 * TestID pt * PaperID pt and therefore the difference between performance of two tests i and j, measured by LOR, is LOR pi - LOR pj = β 5 * TestID pi - β 5 * TestID pj + β 7 * TestID pi * PaperID pi - β 7 * TestID pj * PaperID pj where we identify δ i,j of the decomposition model (1) with the β 5 * TestID pi - β 5 * TestID pj , and identify δ i,j,p with β 7 * TestID pi * PaperID pi - β 7 * TestID pj * PaperID pj . If there is an obvious and generally accepted diagnostic test that can serve as a reference category (RefCat) to which other tests can be compared, then a "simple" parameterization for tests is sufficient, However, usually it is not the case. When there is no perceived referent test to which the other tests are to be compared, a "deviation from means" coding is preferred for the tests. Using the deviation parameterization for both TestID and PaperID in the model (2), one can show that β 5 * TestID pt is the average deviation of the LOR of test t from the overall LOR (the β 1 ), where the overall LOR is the average over all tests and all papers. Therefore β 5 * TestID pt of model (2) will be equivalent to the δ i,j of the decomposition model (1), and β 7 * TestID pt * PaperID pt equivalent to δ i,j,p . Proportional odds ratio model Model (2) expands each study to its original sample size, and uses patients as primary analysis units. Compared to a random-effects model where papers are the primary analysis units, it has more degrees of freedom. However, in a real case, not every test is studied in every paper. Rather majority of tests are not studied in each paper. Therefore the data structure of tests-by-papers is incomplete with many unmeasured cells. The three-way interaction model (2) may become over-parameterized. One may want to drop the term β 6 * Disease pt * TestID pt * PaperID pt . Then for the reduced model (3) logit(Result pt ) = β 0 + β 1 * Disease pt + β 2 * PaperID pt + β 3 * Disease pt * PaperID pt + β 4 * TestID pt + β 5 * Disease pt * TestID pt we have LOR pt = β 1 + β 3 * PaperID pt + β 5 * TestID pt , where the paper and test effects are completely separate. We call this reduced model the Proportional Odds Ratio (POR) model, where the ratio of odds ratios of two tests is assumed to be constant across papers, while odds ratio of each test is allowed to vary across the papers. Note the difference with the proportional odds model where ratio of odds is assumed to be constant [ 14 ]. In the POR model (4) OR pt = OR p * , t = 1 , 2 , ..., k , p = 1 , 2 , ..., m where t is an index for the k diagnostic tests, and p is an index representing the m papers included in the analysis. OR p is a function capturing the way OR changes across papers. Then to compare two diagnostic tests i and j OR pi / OR pj = where the ratio of the two ORs depends only on the difference between the effect estimates of the two tests, and is independent of the underlying OR p across the papers. Thus the model makes no assumptions about the shape of OR p (and in particular homogeneity of ORs) but merely specifies a relationship between the ORs of the two tests. One may want to replace the PaperID variable with a smooth function of FPR or TPR, such as natural restricted cubic splines. There are two potential advantages. This may preserve some degrees of freedom, where one can spend by adding covariates to the model to measure their potential effects on the performance of the diagnostic tests. Thus one would be able to explain why performance of the same test varies across papers. Also, this allows plotting a ROC curve where the OR is not constant across the curve, a flexible ROC (HetROC) curve. (5) logit(Result pt ) = β 0 + β 1 * Disease pt + β 2 * S(FPR pt ) + β 3 * Disease pt * S(FPR pt ) + β 4 * TestID pt + β 5 * Disease pt * TestID pt + β 6 * X pt + β 5 * Disease pt * X pt To test the POR assumption one may use model (2) where the three-way interaction of Disease and TestID with PaperID is included. However, in majority of real datasets this would mean an over-parameterized model. Graphics can be used for a qualitative checking of the POR assumption. For instance, the y-axis can be LOR, while the x-axis is paper number. To produce such plot, it may be better to have the papers ordered in some sense. One choice is to compute an unweighted average of (observed) ORs of all the tests the paper studied, and use it as the OR of that paper. Then sort the papers based on such ORs. The OR of a test may vary from one paper to the other (with no restriction), but the POR assumption is that the ratio of ORs of two tests remains the same from one paper to another. If one shows ORs of a test across papers by a smooth curve, then one expects that the two curves of the two tests are proportional to each other. In the log-OR scale, this means the vertical distance of the two curves remains the same across the x-axis. To compute the observed LOR for a test in a paper one may need to add some value (like 1/2) to the cell counts, if some cell counts are zero. However, this could introduce some bias to the estimates. Among the approaches for modeling repeated-measures data, we use generalized estimating equations to estimate the marginal logistic regression [ 15 ]. Software is widely available for estimation of parameters of a marginal POR model. These include SAS (genmod procedure), R (function geese), and STATA (command xtgee), with R being freely available open source software [ 16 ]. One may use a non-linear mixed effects modeling approach on the cell-count data for estimation of parameters of the POR model. The Paper effect is declared as random, and interaction of the random effect with Disease is included in the model, as indicated in model (2). However, such mixed effects non-linear models are hard to converge, especially for datasets where there are many papers studying only one or a small number of the included tests (such as the dataset presented as example in this paper). If the convergence is good, it may be possible to fit a mixed model with the interaction of Disease, Test, and the Paper random effect. Such model relaxes the POR assumption, besides relaxing the assumption of OR-homogeneity. In other words, one can use the model to quantitatively test the POR assumption. One should understand that the interpretation of LOR estimate from a marginal model is of a population-average, while that of a mixed model is a conditional-average. Therefore there is a slight difference in their meaning. Expanding the proportional odds ratio model One may use the frameworks of the generalized linear models (GLM) and the generalized estimating equations (GEE) to extend the POR model and apply it to different scenarios. By using suitable GLM link function and random component [[ 17 ], p.72], one may fit the POR model to multi-category diagnostic tests, like baseline-category logits, cumulative logits, adjacent-categories and continuation-ratio logits [[ 17 ], chapter 8]. A loglinear 'Proportional Performance' (PP) regression may be fitted to the cell counts, treating them as Poisson. Also, one may fit the PP model to the LORs directly, assuming a Gaussian random component with an identity link function. Comparing GEE estimates by fitting the model to 2-by-2 tables versus GEE estimates of the model fitted directly on LOR versus a Mixed model fitted on LOR, usually statistical power decreases across the three. Also, there is issue of incorporation of sample sizes that differ across studies. Note some nuisance parameters, like coefficients of all main effects and the intercept, won't need to be estimated as they are no longer present in the model fitted directly on LORs. One may avoid dichotomizing results of the diagnostic test by using the 'likelihood ratio' as the performance measure, and fitting a PP model to such continuous outcome. For a scenario where performance of a single test has been measured multiple times within the same study, for example with different diagnostic calibrations (multiple thresholds), the POR estimated by the GEE incorporates data dependencies. When there is a multi-layer and/or nested clustering of repeated measures, software to fit a mixed-effects POR model may be more available than an equivalent GEE POR. When POR is implemented by a logistic regression on 2-by-2 tables, it uses a grouped binary data structure. It takes a minimal effort to fit the same logistic model to the "ungrouped" binary data, the so-called "individual level" data. Methods of meta-analysis that allow for different outcomes (and different numbers of outcomes) to be measured per study, such as that of Gleser and Olkin [ 18 ], or DuMouchel [ 19 ], may be used to implement the POR model. This would prevent conducting parallel meta-analyses that is usually less efficient. Results Deep vein thrombosis To demonstrate how to fit the POR model, we use a recent meta-analysis of diagnostic tests for deep vein thrombosis (DVT) by Heim et al. [ 20 ]. In this meta-analysis there are 23 papers and 21 tests, comprising 483 potential performance measurements, while only 66 are actually observed, thus 86% of cells are not measured. We fitted the reduced marginal logistic regression model (3). Table 1 shows the parameter estimates for Test effects. SAS code to estimate the parameters is provided [see additional file 1 ].Data files are provided in Additional file 2 . Table 1 Parameter estimates for test effects Coefficient Test Deviation* 95% Confidence Limits p value** β 5 † 1 Asserachrom 0.524 0.2293, 0.8186 0.0005 2 Auto Dimertest 0.222 -0.1466, 0.5912 0.2376 3 BC D-Dimer -0.993 -2.4195, 0.4333 0.1724 4 D-Dimer test 0.225 0.1, 0.3494 0.0004 5 Dimertest -2.092 -2.3392, -1.8439 <.0001 6 Dimertest EIA -0.929 -1.1756, -0.6825 <.0001 7 Dimertest GOLD EIA -0.193 -0.4784, 0.0935 0.1871 8 Dimertest II -0.731 -0.9774, -0.4843 <.0001 9 Enzygnost 0.399 0.1209, 0.6766 0.0049 10 Fibrinostika 0.857 0.6865, 1.0266 <.0001 11 IL Test 0.809 0.0914, 1.5256 0.0271 12 Instant I.A. 0.558 0.216, 0.9006 0.0014 13 Liatest -0.143 -0.3375, 0.0511 0.1486 14 LPIA 0.182 -0.0354, 0.3997 0.1007 15 Minutex -0.323 -0.8394, 0.193 0.2197 16 Nephelotex 0.654 0.4325, 0.8745 <.0001 17 NycoCard -0.797 -1.0434, -0.5506 <.0001 18 SimpliRED 0.393 0.1467, 0.6398 0.0018 19 Tinaquant 0.703 0.0113, 1.3948 0.0464 20 Turbiquant -0.328 -1.6596, 1.0032 0.629 21 VIDAS 1.004 0.365, 1.6424 0.0021 β 1 Overall LOR 2.489 2.4175, 2.5606 < .0001*** * estimate of deviation from overall LOR ** p-value for null hypothesis of Deviation = 0 ***p-value for null hypothesis of LOR = 0 † LOR(Result pt ) = β 1 + β 3 * PaperID pt + β 5 * TestID pt Since we have used deviation contrast for the variables, estimate of β 1 is the "overall mean" for the log-OR. This is similar to an ANOVA analysis where the overall mean is estimated by the model. Therefore the average OR is equal to exp(2.489) = 12.049. Components of β 5 estimate deviation of LOR of each test from the overall LOR. Software gives estimates of SEs, plus confidence intervals and p-values, so inference is straightforward. A forest plot may be used to present the results of the modeling in a graphical way. This may connect better with clinically oriented audience. In Figure 1 we have sorted the 21 tests based on their LOR estimate. Figure 1 Comparing performance of each diagnostic test to the overall LOR The horizontal axis is log-OR, representing test performance. The dashed vertical line shows overall mean LOR. For each diagnostic test the solid square shows the LOR, while the horizontal line shows the corresponding 95% CI. If the horizontal line does not intersect the vertical line, the test is significantly different from the overall mean LOR. Note that the CIs in the plot are computed by adding the overall LOR to the CI for the deviation effect of each particular test. This ignores the variability of the overall LOR estimate. One can estimate the LOR of a test and its CI more accurately by some extra computations, or by fitting a slightly modified model. A method is illustrated and implemented [see additional file 1 ]. However, the gain in accuracy was small in this particular example. The model also estimates paper effects. However, one may not be interested in those primarily. One can translate LOR to other measures of test performance. There are numerous types of these measures. We provide code to convert the LOR estimated by the POR model to such measures. Note that majority of them, unlike LOR, are in pairs. This means in order to compare two tests, one needs to use two numbers to represent each single test. For example, sensitivity-specificity is a pair. If a test has a higher sensitivity than the other test, while having a lower specificity, it is not immediately clear which test is better. Also, note that some performance measures are independent of disease prevalence, while others depend on prevalence. This means the same test would perform differently for populations with different disease prevalence. Note in order to compute some of the performance measures, one needs to assume a prevalence and sensitivity or specificity. We assumed a disease prevalence of 40%, and a specificity of 90%, for Table 2 , as the tests are mainly used for ruling out the DVT. Table 2 Other performance measures for the 21 diagnostic tests of DVT Diagnostic Test DOR Prev. Spec. Sens. AUC PPV NPV LRAT LRNT PTO PTOAT PTONT PTPAT PTPNT 1 Asserachrom 20.3 0.4 0.9 0.693 0.888 0.822 0.815 6.933 0.341 0.667 4.622 0.227 0.822 0.185 2 Auto Dimertest 15.0 0.4 0.9 0.626 0.864 0.807 0.783 6.258 0.416 0.667 4.172 0.277 0.807 0.217 3 BC D-Dimer 4.5 0.4 0.9 0.332 0.732 0.688 0.669 3.315 0.743 0.667 2.210 0.495 0.688 0.331 4 D-Dimer test 15.1 0.4 0.9 0.626 0.865 0.807 0.783 6.263 0.415 0.667 4.175 0.277 0.807 0.217 5 Dimertest 1.5 0.4 0.9 0.142 0.566 0.486 0.611 1.419 0.953 0.667 0.946 0.636 0.486 0.389 6 Dimertest EIA 4.8 0.4 0.9 0.346 0.741 0.697 0.674 3.459 0.727 0.667 2.306 0.485 0.697 0.326 7 Dimertest GOLD EIA 9.9 0.4 0.9 0.525 0.826 0.778 0.740 5.248 0.528 0.667 3.499 0.352 0.778 0.260 8 Dimertest II 5.8 0.4 0.9 0.392 0.766 0.723 0.689 3.920 0.676 0.667 2.613 0.450 0.723 0.311 9 Enzygnost 18.0 0.4 0.9 0.666 0.879 0.816 0.802 6.661 0.371 0.667 4.440 0.247 0.816 0.198 10 Fibrinostika 28.4 0.4 0.9 0.759 0.910 0.835 0.849 7.592 0.268 0.667 5.061 0.178 0.835 0.151 11 IL Test 27.0 0.4 0.9 0.750 0.907 0.833 0.844 7.503 0.277 0.667 5.002 0.185 0.833 0.156 12 Instant I.A. 21.1 0.4 0.9 0.701 0.890 0.824 0.818 7.006 0.333 0.667 4.671 0.222 0.824 0.182 13 Liatest 10.4 0.4 0.9 0.537 0.831 0.782 0.745 5.371 0.514 0.667 3.581 0.343 0.782 0.255 14 LPIA 14.5 0.4 0.9 0.616 0.861 0.804 0.779 6.163 0.426 0.667 4.109 0.284 0.804 0.221 15 Minutex 8.7 0.4 0.9 0.492 0.813 0.766 0.727 4.921 0.564 0.667 3.281 0.376 0.766 0.273 16 Nephelotex 23.2 0.4 0.9 0.720 0.897 0.828 0.828 7.202 0.311 0.667 4.801 0.207 0.828 0.172 17 NycoCard 5.4 0.4 0.9 0.376 0.758 0.715 0.684 3.763 0.693 0.667 2.509 0.462 0.715 0.316 18 SimpliRED 17.9 0.4 0.9 0.665 0.878 0.816 0.801 6.648 0.372 0.667 4.432 0.248 0.816 0.199 19 Tinaquant 24.3 0.4 0.9 0.730 0.900 0.830 0.833 7.300 0.300 0.667 4.867 0.200 0.830 0.167 20 Turbiquant 8.7 0.4 0.9 0.491 0.812 0.766 0.726 4.909 0.566 0.667 3.273 0.377 0.766 0.274 21 VIDAS 32.9 0.4 0.9 0.785 0.918 0.840 0.863 7.851 0.239 0.667 5.234 0.159 0.840 0.137 DOR = Diagnostic Odds Ratio Prev. = Prevalence Spec. = Specificity Sens. = Sensitivity AUC = Area Under Curve (assuming homogeneous OR) PPV = Positive Predictive Value NPV = Negative Predictive Value LRAT = Likelihood Ratio For Abnormal Test LRNT = Likelihood Ratio For Normal Test PTO = Pre Test Odds PTOAT = Post Test Odds Of Abnormal Test PTONT = Post Test Odds Of Normal Test PTPAT = Post Test Probability Of Abnormal Test PTPNT = Post Test Probability Of Normal Test We suggest graphs to compare tests when using such "prevalence-dependent paired performance measures" [ 21 ]. In Figure 2 we have used a pair of measures, 'probability of disease given a normal test result' and 'probability of disease given an abnormal test result', the dashed red curve and the dot-and-dash blue curve respectively. Figure 2 Post-test probability difference for diagnostic test VIDAS The way one may read the graph is that, given a particular population with a known prevalence of disease like 40%, we perform the diagnostic test on a person picked randomly from the population. If the test turns normal, the probability the person has disease decreases from the average 40% to about 4% (draw a vertical line from point 0.4 on x-axis to the dashed red curve, then draw a horizontal line from the curve to the y-axis). If the test turns abnormal, the probability the person is diseased increases from 40% to about 57%. The dotted green diagonal line represents a test no better than flipping a coin, an uninformative test. The farther the two curves from the diagonal line, the more informative the test is. In other words, the test performs better. One can summarize the two curves of a test in a single curve, by computing the vertical distance between the two. The solid black curve in the figure is such "difference" curve. It seems this particular test is performing the best in populations with disease prevalence of around 75%. One can use the difference curve to compare several tests, and study effect of prevalence on the way the tests compare to each other. In Figure 3 two tests VIDAS and D-Dimer from the DVT example are compared. From the model estimates we know that both tests perform better than average. And that VIDAS performs better than D-Dimer. Figure 3 Comparing post-test probability difference for VIDAS – D-Dimer The black solid curve is comparing the two tests. For populations with low disease prevalence (around 17%), the D-Dimer is performing better than VIDAS. However, when the prevalence is higher (around 90%), VIDAS is preferred. Simultaneous confidence bands around the comparison curve would make formal inference possible. Random effects A nonlinear mixed effects POR model fitted to cell counts of the DVT dataset does not converge satisfactorily. We fitted the mixed model to a subset of the data where only two tests and seven papers are included, Table 3 . For codes refer to the additional file 1 . Table 3 Data structure for two diagnostic tests Test Paper Instant I.A. NycoCard 3 Elias, A. 1996 (171) X X 8 Legnani, C. 1997 (81) X X 11 Leroyer, C. 1997 (448) X 12 Scarano, L. 1997 (126) X X 13 van der Graaf, F. 2000 (99) X X 21 Wijns, W. 1998 (74) X 22 Kharia, HS. 1998 (79) X TOTAL 6 5 Five of the seven papers have studied both the tests. Result of SAS Proc NLMixed still is sensitive to initial values of parameters. The three-way interaction term of disease, test, and paper in the mixed model (where POR is not assumed) is insignificant, Table 4 . A POR assumption for the two tests may be acceptable. Table 4 Comparing parameter estimates from three models POR-relaxed Mixed * POR Mixed POR Marginal overall LOR 1.389 (0.993, 1.786) 0.868 (0.568, 1.169) 2.593 (2.522, 2.664) Test (NycoCard) -0.903 (-1.811, 0.006) -0.93 (-1.104, -0.755) -0.561 (-0.829, -0.293) Test*Paper 0.016 (-1.556, 1.588) --- --- * logit(Result) = β 0 + β 1 * Disease + β 2 * PaperID + β 3 * Disease * PaperID + β 4 * TestID + β 5 * Disease * TestID + β 6 * Disease * TestID * PaperID The estimate of overall LOR from both the POR-mixed model and POR-marginal model are significantly different from zero. However, the mixed model estimate of LOR is much smaller than the marginal one. For non-linear models, the marginal model describes the population parameter, while the mixed model describes an individual's [[ 15 ], p.135]. The estimate of deviation of test (NycoCard) from the overall LOR is closer in the two models. Plus the marginal estimate is closer to 0 than the mixed estimate. One expects coefficient estimates of mixed model being closer to zero, compared to the fixed model, while the mixed model CI's being wider. Meta-analysis of a single test: the baseline OR p function Sometimes one may be interested in constructing the ROC curve for the diagnostic test. A homogeneous ROC curve assumes the performance of the test (as measured by LOR) is the same across the whole range of specificity. However, this assumption may be relaxed in a HetROC. We fitted a simplified version of model (5) for test SimpliRED, logit(Result pt ) = β 0 + β 1 * Disease pt + β 2 * S(FPR pt ) + β 3 * Disease pt * S(FPR pt ) where index t is fixed, and then used estimates of the coefficients to plot the corresponding HetROC, Figure 4 . Figure 4 Heterogeneous ROC curve for diagnostic test SimpliRED The eleven papers that studied test SimpliRED are shown by circles where the area is proportional to the sample size of the study. The black dashed curve is ROC curve assuming homogeneous-OR. The red solid curve relaxes the assumption, hence a heterogeneous ROC curve. The amount of smoothing of the curve can be controlled by the "degree-of-freedom" DF parameter. Here we have used a DF of 2. Codes to make such plots are presented in the additional file 1 . Model checking Checking the POR assumption, model (2) may be used to reject significance of the three-way interaction term. However, the dataset gathered for the DVT meta-analysis is such that no single paper covers all the tests. Moreover, out of 21, there are 7 tests that have been studied in only one paper. For Figure 5 we chose tests that have been studied in at least 5 of the 23 papers. There are 5 such tests. Note that even for such "popular" tests, out of 10 pairwise comparisons, 3 are based on only one paper (so no way to test POR). Four comparisons are based on 4 papers, one based on 3 papers, and the remaining two comparisons are based on 2 papers. Figure 5 Observed log-odds-ratios of each diagnostic test We sorted the papers, the x-axis, based on average LOR within that paper. We fitted Lowess smooth lines to the observed LORs of each test separately. Figure 5 shows the smooth curves are relatively parallel. Note the range of LORs of a single test. The LORs vary considerably from one paper to the other. Indeed the homogeneity-of-ORs assumption is violated in four of the five tests. Also, to verify how good the model fits the data, one may use an observed-versus-fitted plot. Plots or lists of standardized residuals may be helpful finding papers or tests that are not fitted well. This may provide a starting point for further investigation. Discussion A comparison of the relative accuracy of several diagnostic tests should ideally be based on applying all the tests to each of the patients or randomly assigning tests to patients in each primary study. Obtaining diagnostic accuracy information for different tests from different primary studies is a weak design [ 3 ]. Comparison of the accuracy of two or more tests within each primary study is more valid than comparison of the accuracy of two or more tests between primary studies [ 22 ]. Although a head-to-head comparison of diagnostic tests provides more valid results, there are real-world practical questions that meta-analysis provides an answer that is more timely and efficient than a single big study [ 23 ]. Meta-analysis can potentially provide better understanding by examining the variability in estimates, hence the validity versus generalizability (applicability). Also, there may be tests that have never been studied simultaneously in a single study, hence meta-analysis can "reconstruct" such a study of diagnostic tests. Relaxing the assumption of OR homogeneity In meta-analysis of two (or more) diagnostic tests, where attention is mainly on the difference between performances of two tests, having a homogeneous estimate of performance of each single test is of secondary importance, and it may be treated as nuisance. The POR model assumes differences between LORs of two tests are the same across all papers, but does not assume the OR of a test is the same in every paper. Hence no need for homogeneity of OR of a test across papers that reported it, but shifting the assumption one level higher to POR. Common versus average effect size The POR model uses "deviation from means" parameterization. Then one does not need to drop the interactions coefficient β 3 in the model logit(Result) = β 0 + β 1 * Disease + β 2 * PaperID + β 3 * Disease * PaperID , to interpret β 1 , the overall LOR. This means the POR model explicitly accepts that performance of the diagnostic test varies across the papers, but at the same time estimates its mean value. McClish explains if a test for OR homogeneity shows heterogeneity, there may be no 'common' measure to report, but still there is an 'average' measure one can report. [ 13 ] Advantages of using 2-by-2 tables We demonstrated how to fit the POR model to the cell counts, rather than to the OR values. This, we believe, has several advantages. 1. One does not need assuming normality of some summary measure. This results in binomial distributional assumption that is more realistic. 2. Also, different study sample sizes are incorporated into the POR model without faulty bias-introducing weighting schemes, as shown by Mosteller & Chalmers [ 25 ]. And extension of the POR model to individual level patient data is much easier. 3. The effective sample size for a meta-analysis by a random model is the number of papers included, which is usually quite small. There is a great danger for overfitting. And the number of explanatory variables one could include in the model is very restricted. Since we use the grouped binary data structure, the patients are the effective sample size, hence much bigger degrees of freedom. The way the random-effects model is usually implemented is by extracting OR from each paper, and assuming LOR being normally distributed. Then the distinction between the two types of mistakes (FNR and FPR, or equivalently TPR and FPR) is lost, since one enters the LOR as datapoints into the model. The bivariate model by Houwelingen et al [ 26 ] tries to fix this, by entering two datapoints into the model for each test from each paper. A fourth advantage of fitting the POR model to the cell counts is that the two types of mistakes are included in the model. Consider the logistic regression logit(Result) = β 0 + β 1 * Disease + β 2 * PaperID . Then we have log(true positive/false negative) = β 0 + β 1 + β 2 * PaperID . Substituting a value for the covariate (here PaperID) such as a modal or average value, and using the model estimates for the betas, one gets the log-odds. Then one exponentiates it to get the TP/FN, call it Q. Now it is easy to verify that sensitivity = Q/(1+Q). Likewise we have log(false positive/true negative) = β 0 + β 2 * PaperID , that we call = log(W). Then specificity = 1/(1+W). Also, one can apply separate weights to the log(true positive/false negative) and log(false positive/true negative), to balance the true positive and false positive rates for decision making in a particular clinical practice. When collecting papers from biomedical literature for meta-analysis of a few diagnostic tests, it is hard to come up with a complete square dataset, where every paper has included all the tests of interest. Usually the dataset contains missing values, and a case-wise deletion of papers with missing tests means a lot of data is thrown away. A method of analysis that can utilize incomplete matched groups may be helpful. The POR model allows complex missing patterns in data structure. Convergence of marginal POR model seems much better than non-linear mixed model, when fitted to cell counts of incomplete matched groups. This is an advantage for using GEE to estimate POR. The fact that one can use popular free or commercial software to fit the proposed models, facilitates incorporation of the POR modeling in the practice of meta-analysis. Unwanted heterogeneity versus valuable variability The POR model utilizes the variation in the observed performance of a test across papers. Explaining when and how the performance of the test changes, and finding the influential factors, is an important step in advancing science. In other words, rather than calling it 'heterogeneity', treated as 'unwanted' and unfortunate, one calls it 'variability' and utilizes the observed variability to estimate and explain when and how to use the agent or the test in order to optimize their effects. Victor [ 32 ] emphasizes that results of a meta-analysis can only be interpreted if existing heterogeneities can be adequately explained by methodological heterogeneities. The POR model estimates effect of potential predictors on between-study variation, hence trying to 'explain' why such variation exists. The POR model incorporates risk of events in the control group via a predictor, such as observed prevalence, hence a 'control rate regression'. [ 26 ] ROC curve Although implementing the HetROC means that one accepts the diagnostic test performs differently in different FPRs along the ROC curve, in some implementations of HetROC, such as method of summary ROC, one compares tests by a single point of their respective ROCs. This is not optimal. (The Q test of the SROC method is a single point test, where that point on the ROC may not be the point for a specific cost-benefit case.) In such method although one produces a complete SROC, but one does not use it in comparing the diagnostic tests. In the POR model, one uses LOR as the measure for diagnostic discrimination accuracy, and builds statistical test based on the LOR-ratio, hence the test corresponds to whole ROCs (of general form). The ROC graph was designed in the context of the theory of signal detectability [ 27 , 28 ]. ROC can be generated in two ways, by assuming probability distribution functions (PDFs) for the two populations of 'diseased' and 'healthy', or by algebraic formulas [ 29 ]. Nelson claims the (algebraic) ROC framework is more general than the signal detection theory (and its PDF-based ROC) [ 5 ]. The location-scale regression models implement ROC via PDFs, while the method of summary-ROC uses algebraic approach. The POR model uses a hybrid approach. While POR may be implemented by logistic regression, the smoothing covariate resembles the algebraic method. Unlike location-scale regression models that use two equations, POR uses one equation, hence it is easier to fit by usual statistical packages. One may use a five-parameter logistic to implement the HetROC. However, the model cannot be linearized, then according to McCullagh [ 14 ] it won't have good statistical properties. The POR model not only relaxes assumption of Var1/Var2 = 1, where Var1 and Var2 are variances of the two underlying distributions for the two populations, but even monotonicity of ROC. Hence the model can be used to represent both asymmetric ROCs and non-regular ROCs (singular detection). In building HetROC curve, the POR model accommodates more general heterogeneous ROCs than SROC, because it uses nonparametric smoother instead of arbitrary parametric functions used in SROC method. When in the POR model the smoother covariate is replaced by log{TPR*FPR/ [(1-TPR)*(1-FPR)]}, a HetROC similar to SROC of Moses et al is produced. When one uses a smooth function of FPR in the POR model, it is equivalent to using a function of outcome as predictor. This resembles a 'transition model'. Ogilvie and Creelman [ 30 ] claim that for estimating parameters of a best fitting curve going through observed points in the ROC space, least squares is not good since both axes are dependent variables and subject to error. They claim maximum likelihood is a preferred method of estimation. Crouchley and Davies [ 31 ] warn that, although GEE is fairly robust, it becomes inconsistent if any of the covariates are endogenous, like a previous or related outcome or baseline outcome. They claim a mixed model is better for studying microlevel dynamics. We have observed that the smooth HetROC curve may become decreasing at right end, due to some outlier points. Using less smoothing in the splines may be a solution. When there is only one diagnostic test, and one is mainly interested in pooling several studies of the same test, the POR model estimates effect sizes that are more generalizable. By using the smoother (instead of PaperID), one fits a sub-saturated model that allows inclusion of other covariates, hence it is possible to estimate effect of study level factors on performance and explain the heterogeneity. Also it does not assume any a priori shape of the ROC, including monotonicity. Plus, it enables graphing of the HetROC. It does not need omission of interaction terms to estimate the overall performance, and it does not need assumption of OR homogeneity. If several performance measurements of the same test is done in a single study, like evaluating the same test with different diagnostic calibrations, the POR model provides more accurate estimates, by incorporating the dependence structure of the data. Random effects When there is heterogeneity between a few studies for the same diagnostic test, one solution to absorb the extra between-study variation is to use a random/mixed effects model. However, Greenland [ 33 ] cautions when working with random effect models: 1. if adding random effect changes the inference substantially, it may indicate large heterogeneity, needing to be explained; 2. specific distributional forms for random effects have no empiric, epidemiologic, or biologic justification. So check its assumptions; 3. the summary statistic from random-effect model has no population-specific interpretation. It represents the mean of a distribution that generates effects. Random models estimate unit specific coefficients while marginal models estimate population averages. The choice between unit-specific versus population-average estimates will depend on the specific research questions that are of interest. If one were primarily interested in how a change in a covariate affect a particular individual cluster's mean, one would use the unit-specific model. If one were interested in how change in covariate can be expected to affect the overall population mean, one would use the population-average model. The difference between "unit-specific" models and "population-average" models arises only in the case of a nonlinear link function. In essence random-effect model exchanges questionable homogeneity assumption for a fictitious random distribution of effects. Advantage of a random model is that SE and CI reflect unaccounted-for sources of variation, and its drawback is that simplicity of interpretation is lost. When residual heterogeneity is small, fixed and random should give same conclusions. Inference about the fixed effects (in a mixed model) would apply to an entire population of cases defined by random effect, while the same coefficient from a fixed model apply only to particular units in the data set. Crouchley and Davies [ 31 ] explain one of the drawbacks of their random model is that it rapidly becomes over-parameterized, and also may encounter multiple optima. Follow-ups We suggest these follow-ups: 1. the POR model has been implemented both by marginal and mixed models. It would be useful to implement a marginalized mixed POR model; 2. in clinical practice, usually a group of diagnostic tests is performed on an individual, for a particular disease. Some of these tests are requested simultaneously and some in sequence. It would be useful, and practically important, to extend the POR model such that it incorporates such sequence of testing and a priori results; 3. the utility of POR model may be extended to meta-analysis of therapeutics. Competing interests The author(s) declare that they have no competing interests. Authors' contributions MSS conceived of the model, and participated in its design and implementation. JS participated in implementation of the model and performing of the example analysis. Both authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here: Supplementary Material Additional File 1 In this file we present sample codes for a few of the models presented in the paper. The estimation mostly has been done in SAS, while the graphing (and some model-fitting) has been done in R. Click here for file Additional File 2 This zipped file contains 8 data files, in the .csv (comma separated value) and .xls (MS Excel) formats. They are to be used with the SAS and R codes we presented in the Appendix [additional file 1]. Five files are for the SAS codes presented in the Appendix. The file names are "data5.xls", "data5_t12&17.xls", "u125.xls", "data5_t18.xls", "data6.xls". Three files are for the R codes presented in the Appendix. The file names are "obsVSfit.csv", "dataNewExcerpt2.csv", and "data6_lor2.csv". Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539279.xml

548669

clc is co-expressed with clf or cntfr in developing mouse muscles

Background The ciliary neurotrophic factor (CNTF) receptor is composed of two signalling receptor chains, gp130 and the leukaemia inhibitory factor receptor, associated with a non-signalling CNTF binding receptor α component (CNTFR). This tripartite receptor has been shown to play important roles in the development of motor neurons, but the identity of the relevant ligand(s) is still not clearly established. Recently, we have identified two new ligands for the CNTF receptor complex. These are heterodimeric cytokines composed of cardiotrophin-like cytokine (CLC) associated either with the soluble receptor subunit cytokine-like factor-1 (CLF) or the soluble form of the binding receptor itself (sCNTFR). Results Here we show that, during development, clc is expressed in lung, kidney, vibrissae, tooth, epithelia and muscles during the period of development corresponding to when motoneuron loss is observed in mice lacking a functional CNTF receptor. In addition, we demonstrate that it is co-expressed at the single cell level with clf and cntfr , supporting the idea that CLC might be co-secreted with either CLF or sCNTFR. Conclusion This expression pattern is in favor of CLC, associated either with CLF or sCNTFR, being an important player in the signal triggered by the CNTF receptor being required for motoneuron development.

Background CLC (cardiotrophin-like cytokine) shares homology with CNTF (ciliary neurotrophic factor) and CT-1 (cardiotrophin-1) and requires co-expression with either CLF (cytokine-like factor-1) or the soluble form of the CNTFR to be secreted [ 1 , 2 ]. The CLC-CLF heterodimer displays activities only on cells expressing a functional CNTF receptor [ 1 ] and therefore CLC is likely to be part of the developmentally important second ligand for CNTFR. The existence of such a second ligand has been suggested by the phenotype of mice lacking any of the three receptor subunits comprising the functional CNTF receptor complex (LIFRβ, gp130 and CNTFR) which exhibit significant reductions in motoneuron number [ 3 - 5 ] whereas CNTF-deficient mice have no motoneuron loss during development [ 6 ]. There are however two prerequisites for CLC to play a major role in motoneuron development: 1) CLC must be expressed in the environment of motoneurons during development. 2) As it cannot be secreted alone, it must be co-expressed with either CLF or sCNTFR, in the same cell. Results and Discussion Developmental expression of clc Since the expression of clc has only been studied in adult mouse tissues [ 7 ], we first examined the expression of genes encoding CLC or its co-secreted proteins, CLF and CNTFR in various embryonic tissues using reverse transcription and quantitative real-time polymerase chain reaction (RT-PCR). In all tissues tested from E16.5 and E18.5 (Table 1 ), the level of expression of clc is very low when compared with that of clf or cntfr . The highest level of clc expression was observed in the muzzle, a very heterogeneous region containing different positive tissues, as described below. Clc expression is also observed in lung, kidney, brain and skeletal muscles such as the tongue or limb muscles. Table 1 RT-PCR analysis of clc , clf and cntfr expression a clc clf cntfr E16.5 E18.5 E16.5 E18.5 E16.5 E18.5 Skeletal muscle 0.113 ± 0.02 0.936 ± 0.12 0.232 ± 0.003 24.84 ± 2.13 90.25 ± 2.71 6.72 ± 0.76 Heart NS c NS NS NS NS NS Tongue 3.12 ± 0.12 1.2 ± 0.09 495 ± 17.1 77.1 ± 4.64 28.3 ± 10.9 2.6 ± 0.83 Muzzle 11.4 ± 0.75 9.25 ± 0.79 1050 ± 65.3 524 ± 85.8 425 ± 47.5 57.9 ± 4.02 Lung 4.08 ± 0.65 16.4 ± 0.48 2290 ± 490 2240 ± 184 43.8 ± 15.6 ND b Kidney 5.55 ± 0.12 6.75 ± 1.15 61.4 ± 4.59 ND 51.8 ± 9.49 19.8 ± 3.12 Liver 0.191 ± 0.02 0.129 ± 0.03 0.071 ± 0.001 0.034 ± 0.09 0.047 ± 0.002 0.038 ± 0.03 Brain 1.04 ± 0.12 0.317 ± 0.09 150 ± 3.39 74.3 ± 16.4 216 ± 2.01 236 ± 16.1 Spinal cord ND 0.183 ± 0.01 0.123 ± 0.01 61.2 ± 6.25 6.13 ± 0.03 109 ± 27.4 a Expression of clc , clf and cntfr was determined using reverse transcription and quantitative real-time PCR as detailed in Experimental Procedures and expressed as fM of cDNA/μg total RNA. b not determined c not significant To further assess the potential involvement of CLC in the development of motoneurons, we performed in situ hybridization experiments to determine the pattern of expression of clc in the environment of developing motoneurons and compare it with the expression of both clf and cntfr . Motoneuron death occurs between E14.5 and E18.5 in mice lacking in the ability to produce a functional CNTF receptor complex [ 5 ], suggesting that expression of CNTFR and its relevant ligands is critical between these timepoints. We therefore studied clc mRNA expression levels at E16.5. Clc is expressed in muscles along the whole rostro-caudal axis, at the brachial level (Fig. 1A ) as well as at the lumbar level (Fig. 1E and [ 8 ]. It is also expressed in the tongue (Fig. 1C ) like clf (Fig. 1D ). The identity of muscle cells (Fig. 1G ) was confirmed by double staining performed on transgenic mice with the nlacZ reporter gene under the control of the muscle-specific MLC promoter [ 9 ]. All clc -positive muscle fibers also stained positive for clf (Fig. 1E , 1F , 1G , 1H and [ 8 ]). clc expression was not detected in certain clf -positive muscles however, such those around the vibrissae (Fig. 1I and 1J ). Since the level of clc expression is generally low, this could reflect the limited sensitivity of the in situ hybridization technique used. To determine the onset of clc and clf expression in the muscles, the motoneuron targets, we performed in situ hybridizations at different stages. Clc and clf are expressed, although at low levels, as soon as the muscles develop and are clearly observed at E14.5 (Fig. 1K and 1L ). Figure 1 In E16.5 mouse embryos clc is expressed in muscles. Cryostat sections (A, E-L) or vibratome sections (B-D) from E16.5 (A-J) or E14.5 (K, L) were hybridized to clc (A, C, E, G, I and K) or clf (D, F, H, J and L). The control clc sense probe gave rise to very faint staining (B). Transverse section through the forelimb (A, K and L), the hindlimb (E and F) and saggital sections through the tongue (C and D) showing expression of clc and clf in muscles. The identity of the clc -positive cells such as muscle fibers was confirmed by double staining and compared to clf -positive cells. In situ hybridization using Dig-labeled probes for clc and clf (cytoplasmic blue staining) was performed on sections through shoulder muscles (G, H) or vibrissae (I, J) from E16.5 MLC nlacZ mice, which express the nlacZ reporter gene under the control of a muscle-specific MLC promoter. Subsequently, the sections were processed for immunohistochemical detection of β-galactosidase (nuclear brown staining). Arrows indicate double-labeled cells. Transverse section through the muzzle (I and J) shows that vibrissae (arrowheads) are positive for clc and clf whereas only clf is detected in muscles (asterisks) surrounding vibrissae. c, cartilage; m, muscle. Scale bars are 200 μm in A, E and F, 25 μm in G and H and 100 μm in I-L. Clc is also expressed in several organs in which reciprocal epithelial-mesenchymal interactions are essential, such as the developing vibrissae (Fig. 1I and 2I ), tooth, kidney, and lung. In the kidney, clc is expressed in the comma-shaped body (Fig. 2A ). Strikingly, CLF and CNTFR are expressed in different structures, clf being synthesized in the tips of the ureteric (Fig. 2B ) and cntfr being synthesized by mesenchyma cells surrounding these structures (Fig. 2C ). In the lung, both clc and cnftr are expressed faintly in distal airway epithelium whereas clf is strongly expressed in distal and proximal epithelia (Fig. 2D , 2E and 2F ). Sections through molar tooth germs (Fig. 2G and 2H ) show that clf is expressed in both the mesenchyma surrounding the dental follicle which gives rise to alveolar bone and the inner enamel epithelium whereas clc is expressed only in the former. Clc and clf are also co-expressed in the epithelium bordering the mandibles and the lips although clf is also expressed in mesenchyma (Fig. 2I and 2J ). Together these results are in agreement with the expression pattern described for both clf [ 10 ] and cntfr [ 11 ]. Figure 2 Clc is expressed in epithelia Transverse sections from E16.5 mouse embryos were hybridized to clc (A, D, G and I), clf (B, E, H and J) or cntfr (C and F). Sections through the kidney (A-C) show that clc is expressed in developing nephrons (arrows), clf in ureteric tips (arrowheads) and cntfr in nephrogenic mesenchyme. Sections through the lung (D-F) show that whereas clf is strongly expressed in both distal (arrowheads) and proximal (arrows) epithelia, clc and cntfr are weakly expressed in distal epithelium. Boxed areas are shown in higher magnification in the corner of each panel. Sections through molar tooth germs (G, H) show that mesenchyma (arrows) surrounding the dental follicle is positive for both clc and clf and that the inner enamel epithelium (arrowheads) expresses only clf . Coronal sections through muzzle (I, J) show that both clc and clf are expressed in the epithelium bordering the mandibles and in between the lips and mandibles (arrow) as well as in follicles of vibrissae (arrowheads); in addition, clf is expressed in mesenchyma (asterisks). a, pulmonary artery; dp, dental papilla; de, dental epithelium; oc, oral cavity; uli, upper lip; lli, lower lip. Bars: 100 μm in A-H, 200 μm in I and J. Co-expression of clc , clf and cntfr in the developing muscle In transfected cells CLC requires either CLF or sCNTFR to be secreted [ 1 , 2 ]. This cooperative effect requires the expression of genes for both factors in the same cell. To ascertain whether a single muscle cell can express at least CLC and CLF or CLC and sCNTFR, we studied co-expression on hind-limb muscle sections. We performed double in situ hybridization of clc and clf and of clc and cntfr . Most muscle cells expressed both clc , (revealed using NBT/BCIP; Fig. 3A, C ) and clf or cntfr (Fig. 3B, D ; revealed using Fast Red). Co-expression was observed at the single cell level demonstrating that in vivo CLC could be co-secreted either with CLF or sCNTFR. Figure 3 Double-labeling detects co-expression of clc and clf or clc and cntfr in individual muscle cells. Single sections of E16.5 muscles were hybridized with two probes. Dig-labeled clc (A-D) and Fluo-labeled clf (A, B) or cntfr (C, D). Anti-Dig antibodies were applied first and stained using NBT/BCIP to reveal cells expressing clc (A, C). Anti-Fluo antibodies were then applied and detected using Fast red to reveal cells expressing clf (B) and cntfr (D), after removal of the first red reaction product. Most muscle cells express clc and clf or clc and cntfr (examples indicated by arrows). Bars: 100 μm. Conclusions Clc is expressed in developing muscles during the period of motoneuron loss in mice lacking a functional CNTF receptor and it is co-expressed with both CLF and CNTFR. This expression pattern is in favor of the hypothesis that CLC is an important player in the signal triggered by the CNTF receptor and that is required for motoneuron development. In addition, our results show that in the kidney, clc is expressed in cells neighboring those expressing clf or cntfr but it is not co-expressed with these genes suggesting either the possible existence of an additional protein capable of inducing secretion of CLC or that CLC is not secreted in these cells and therefore not functional. Because genetic deletion of cntf fails to perturb neuronal development before birth, we can hypothesize some functional redundancies in vivo that will require the analysis of double or triple knockout mice for CNTFR ligands to clarify their respective involvement in mouse neural development. Methods RT and real time PCR Total RNA was extracted using Trizol reagent (Invitrogen) from E16.5 or E18.5 mouse tissues according to the manufacturer's instructions. Complementary cDNA was synthesised from 2 μg of RNA by random hexamer priming using MMLV reverse transcriptase (Promega). Quantitative PCR was performed using a capillary real-time LightCycler (Roche Diagnostics), and the data analysed using "Fit Point Method" (Roche Diagnostics). For comparison of gene expression levels, all quantifications were normalized to endogenous gapdh to account for variability in the initial concentration of RNA and for differences in the efficiency of the reverse transcription reactions. The following primers were designed to amplify mouse clc : 5'-GCTACCTGGAGCATCAACT-3', 5'-GGTGACTGTACGCCTCATAG-3'; clf : 5'-CAGTCAGGAGACAATCTGGT-3', 5'-ACGTGAGATCCTTCATGTTC-3'; cntfr : 5'-CTACATCCCCAATACCTACA-3', 5'-GTGAATTCGTCAAAGGTGAT-3'; gapdh : 5'-TGCGACTTCAACAGCAACTC-3', 5'-CTTGCTCAGTGTCCTTGCTG-3'. Results are expressed in fmole of cDNA/μgRNA. Probes Plasmid cDNA clones were linearized and transcribed with T7 or T3 polymerase using digoxigenin (Dig) or fluorescein (Fluo)labeling reagents (Roche Diagnostics). Probes were used at a concentration of 500 ng/ml. The cntfr clone was as previously described [ 12 ] and the mouse clf [ 13 ] and clc probes corresponded to the isolated cDNAs. In situ hybridization In situ hybridization was performed as described previously [ 14 ] on 20 μm-thick frozen transverse cryostat sections prepared from mouse embryos fixed with 4% paraformaldehyde in PBS, and cryopreserved in 15% sucrose in PBS before embedding in OCT compound (Miles). Alternatively, 100 μm-thick vibratome sections were prepared from fixed embryos embedded in glutaraldehyde/gelatin. After hybridization overnight at 70°C with Dig-labeled riboprobes, the slides were washed twice in 1X SSC, 50% formamide at 70°C for 30 min and blocked in the presence of 4% blocking reagent (Roche Diagnostics) and 20% inactivated sheep serum. The slides were then incubated with anti-Dig-alkaline-phosphatase (AP)-conjugated antibody (1/5000, Roche Diagnostics), washed and revealed by NBT/BCIP staining. In order to confirm that muscle fibers, per se , express clc and clf , double in situ hybridization / immunohistochemistry was carried out as described [ 15 ] on sections from E16.5 MLC nlacZ mice, which express the nlacZ reporter gene under the control of a muscle-specific myosin light chain promoter. After in situ hybridization, slides were rinsed in PBT (PBS, 0.1% Triton), and sections were successively incubated for 1 h with blocking solution containing 2% BSA, 2% heat-inactivated donkey serum in PBT and then overnight at 4°C with rabbit anti-β-galactosidasel (1/1000, Cappel). After three washes in PBT, slides were incubated 1 h at RT with a biotin donkey anti-mouse secondary antibody. Slides were then washed in PBS, and TBS (50 mM Tris-HCl, 0.15 M NaCl, pH 7.6), and incubated for 30 min at RT in ABC streptavidin/HRP in TBS. Staining was revealed with DAB (D4293, Sigma) in the presence of H 2 O 2 . Double in situ hybridization was performed as described previously [ 14 ]. Briefly, Dig- and Fluo-labeled probes were mixed in hybridization buffer and applied to sections. After hybridization at 70°C overnight and washing at 65°C, the first probe was revealed using a 1:2000 dilution of anti-Fluo-alkaline phosphatase (AP)- conjugate (Roche Diagnostics) and Fast Red (Sigma) as a substrate. Sections were photographed at this stage. After AP inactivation with 0.1 M glycine, pH 2.2, the second probe was revealed using a 1:5000 dilution of anti-Dig-AP and NBT/BCIP staining. Fast Red precipitates were then removed by incubating the slides in increasing concentrations of ethanol culminating in two final incubations in 100% ethanol for 10 min before cleaning with Histoclear and mounting with Eukitt (VWR, Strasbourg, France). Photomicrographs of the NBT/BCIP results were then taken for comparison with those showing the Fast Red results on the same sections. Competing interests The author(s) declare that they have no competing interests. Authors' contributions BB performed in situ hybridizations whereas DD and HG performed RT-PCR analyses. GE and JFG provided the clc and clf probes before publication. OL participated in the experimental design and coordination of the research. All authors read and approved the final manuscript.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548669.xml

539245

Current practices in spatial analysis of cancer data: data characteristics and data sources for geographic studies of cancer

The use of spatially referenced data in cancer studies is gaining in prominence, fueled by the development and availability of spatial analytic tools and the broadening recognition of the linkages between geography and health. We provide an overview of some of the unique characteristics of spatial data, followed by an account of the major types and sources of data used in the spatial analysis of cancer, including data from cancer registries, population data, health surveys, environmental data, and remote sensing data. We cite numerous examples of recent studies that have used these data, with a focus on etiological research.

Introduction Understanding the spatial patterns of diseases in a population can provide insight as to their causes and controls. Indeed, this notion is at the very root of the field of epidemiology [ 1 ]. The recent explosion in data gathering, linkage and analysis capabilities fostered by computing technology, particularly geographic information systems (GIS), has greatly improved the ability to measure and assess these patterns. Large and complex georeferenced data sets are now readily available through Spatial Data Clearinghouses, facilitating analyses by researchers unaffiliated with the government agencies that have historically controlled data access. Meanwhile, increasingly sophisticated statistical tools have evolved to keep pace with the increased data availability and computing power. The purpose of this article is to provide an overview of spatial data and its relevance to population-based cancer surveillance and research in the United States as of 2004. We begin by discussing a number of the distinctive characteristics of spatial data, which can sometimes hinder efforts to understand cancer etiology. We then proceed to describe the kinds of data sets that are available, accompanied by a survey of some applications using these data. Finally, we discuss several ongoing efforts to provide central repositories of geospatial data. Given the vast scope of cancer research taking place worldwide, our survey is necessarily partial, and we have chosen to emphasize etiology over other research themes with spatial dimensions, such as patterns of treatment or access to care [ 2 ]. Qualities of spatial data Spatial data refer to data with locational attributes. Most commonly, locations are given in Cartesian coordinates referenced to the earth's surface. These coordinates may describe points, lines, areas or volumes. This need not be the only spatial framework; "relative spaces" may be defined in which distance is defined in terms of some other attribute, such as sociodemographic similarly or connectedness along transportation networks [ 3 , 4 ]. Spatial data have special qualities that require specialized statistical techniques and modeling approaches. A complete discussion of these special qualities is well beyond the scope of this article, but here we describe a number of the more compelling and recurring themes. For a focused discussion on the limitations on analysis that these data characteristics impose, see the companion piece to this article, "Current Practices in Spatial Analysis of Cancer Data: Flies in the Ointment, Or, The Limitations of Spatial Analysis" [ 5 ]. Individual humans represent the basic unit of spatial analysis in cancer research. Individuals are categorized as either having or not having a disease or attribute of a disease, and are assigned coordinates corresponding to the location of their place of residence, a technique known as geocoding. As with all measurements, geocoding involves some error. A growing body of literature is exploring the nature of this error and its potential to bias epidemiologic studies [ 6 ]. Among the topics that have been investigated are systematic problems with geographic reference files [ 7 ], the ramifications of different geocoding algorithms [ 8 ], positional accuracy [ 9 ] and how to handle non-residential addresses, such as rented post office boxes [ 10 ]. Assigning individuals to their place of residence also poses problems, although this is usually the only locational information that is available. Often the goal of a geographic analysis is to identify a common environmental exposure in a population, but exposures that are occupational or recreational may not necessarily reveal themselves in a residential analysis. Also, given the long latency period for many cancers and the mobility of the American population, the relevant exposure may be associated with a prior address. Difficult-to-measure behavioral risk factors such as smoking and diet often further confound attempts at geographic analysis. Owing to confidentiality restrictions, researchers outside of central cancer registries typically do not have access to address-level data. In such instances, case data are aggregated by some functional or political unit such as census tract, county or ZIP code. Even when the case data are geocoded, population data must be aggregated, at least to the level of the census block, which is the smallest unit for which any population information is available. Knowing that there are four women with breast cancer living on the same street is not sufficient, by itself, to draw conclusions about whether the street displays an unusual incidence pattern; one must also know the number and ages of women without breast cancer on the same street. Short of conducting one's own thorough door-to-door census, this question cannot be answered, except by aggregating the street segments into blocks. When additional variables, such as measures of income or education, or also of interest, then still larger analytical units must be chosen. The necessity for aggregating spatial data raises a whole set of analytic issues regarding the extent to which the act of aggregating introduces error and bias. It is theoretically possible to achieve dramatically different, even contradictory results, simply as a consequence of aggregating the data in a different fashion [ 11 ]. This is true not only for aggregations at different spatial scales, but also different aggregations at the same scale. Geographers have termed this the "modifiable areal unit problem" [ 12 ]. A special case of the modifiable areal unit problem is the ecological inference problem, which specifically refers to the lack of congruity between associations found in aggregated and individual-level data. In practice, well-chosen scales and groupings can minimize the modifiable areal unit problem and allow reasonable consistency between aggregated and individual-level results [ 13 ]. There will always be exceptions to this, however, as evidenced by the many studies attempting to relate low-level indoor radon concentrations with lung cancer incidence. Individual-level studies have repeatedly found a positive correlation, while area-level studies have found a negative correlation at low radon levels [ 14 - 16 ]. Despite a general appreciation of how these discrepant results represent an example of aggregation bias, there is still active debate over what these results say about low-level radon risk [ 17 , 18 ]. Often analyses need to be performed on data that was collected at different spatial scales, such as a study using cancer cases aggregated by ZIP code and modeled air pollutant data at the census tract level. The resulting scale-translation problem is a recurring one that has inspired many independent solutions, and is known variously as areal interpolation, the polygon overlay problem, and the problem of inference with spatially misaligned data, among other terms [ 19 ]. The most naïve solution to this problem is to assume that each measured value is homogeneous within each spatial unit. Under this assumption, using our example, a ZIP code that is coincident with four census tracts would be broken into four polygons, each having the same cancer rate but different air pollutant values. More sophisticated cartographic overlay techniques have been developed that involve using covariate information to infer variation within spatial units. To date, these techniques have been primarily applied toward estimating population surfaces rather than cancer or other disease rate surfaces [ 20 , 21 ]. Hierarchical Bayesian and multi-level logit models have also shown promise [ 22 - 25 ]. Spatial autocorrelation is another distinctive quality of spatial data that requires the use of specialized analytic methods. Spatial autocorrelation is the tendency for nearby observations to have correlated attribute values. For most data sets involving the distribution of human populations and their characteristics, spatial autocorrelation is positive, meaning that neighboring individuals tend to have similar characteristics. Understanding the characteristics and qualities of spatial autocorrelation is essential to adequately model and interpret geographic patterns. For example, it is not appropriate to perform ordinary least squares regression on spatial data, because the presence of spatial autocorrelation means that the observations are not independent. Performing such a regression generally results in downwardly biased estimations of variance, which yields overstated levels of significance. In general, spatially autocorrelated data is less informative in a model than uncorrelated data. There is an ample literature on assessing and properly accounting for spatial autocorrelation in geographic analysis [ 26 , 27 ]. A final critically important characteristic of spatial data is spatial nonstationarity, or the tendency for relationships between and among variables to vary by geographic location [ 28 ]. First-order or strong stationarity refers to the degree to which measured values vary spatially, while second-order or weak stationarity refers to the degree to which the uncertainties in these measured values vary spatially. So-called global statistics ignore nonstationarity, suggesting that relationships across space are constant. The simple linear equation that has traditionally been used to express the relationship between rainfall and altitude is a well-known example. Local statistics, in contrast, take nonstationarity into account, at least first-order nonstationarity. Brunsdon et al. [ 29 ] used the technique of geographically weighted regression to demonstrate that both the slope and intercept of the rainfall-altitude equation vary considerably in space. The range and breadth of local statistics has seen rapid growth in recent years [ 27 , 30 ]. Local statistics are less adept at accounting for second-order nonstationarity. Indeed, many of these methods require the assumption of constant variance across space. Because of the uneven distribution of human populations, this assumption is seldom met for health data. Specifically, disease rates in areas with smaller numbers of cases are more variable than those in areas with larger numbers of cases, a property that has also been termed "variance instability" [ 31 ]. Variance instability is particularly pervasive on maps, since it is extremely difficult to design a map that is not visually biased toward either sparsely populated or densely populated areas [ 32 , 33 ]. A simple example is the tendency for rural counties to contain disproportionate numbers of unusually high or unusually low disease rates and thus visually dominate a choropleth map. The problem is compounded by the tendency of such counties to be large in size; for these reasons, maps of United States counties are often visually dominated by such states as Idaho, Nevada and Wyoming. Efforts to include information about data uncertainty have shown promise, but have not seen widespread use [ 34 ]. One common way of addressing this problem is to produce smoothed maps, whereby the rate for a given area is influenced by the rates of neighboring areas. There are many algorithms available to accomplish this [ 35 ], ranging from conceptually straightforward spatial filters [ 36 ] to computationally-intensive Bayesian approaches [ 37 , 38 ]. Properly accounting for second-order spatial nonstationarity in maps and models remains an active research area. Types and sources of data In this section we described the primary types and sources of data most frequently used in the geographic analysis of cancer, along with examples of their application. These are summarized in Table 1 . Table 1 Sources of Cancer Registry Data Dataset name Source Agency URL Geographic Resolution SEER*Stat, Cancer Mortality Maps and Graphs, State Cancer Profiles National Cancer Institute County Florida Cancer Data System University of Miami School of Medicine County Cancer Incidence and Mortality Rates in Kentucky Kentucky Cancer Registry County New York Cancer Incidence by ZIP code NYS Department of Health ZIP code 1. Cancer registries A cancer registry is a data collection system that tracks cancer cases that have been diagnosed or treated in a specific institution or geographic area. Cancer registries typically collect information from medical records provided by hospitals, doctors, other care facilities, medical laboratories, and/or insurers. Data collected by cancer registries is stored under secure conditions so as to protect confidentiality. Historically, observed geographic differences in cancer incidence have been of great interest in trying to understand more about factors which may influence risk of these diseases. Such differences have served as the basis for studies of migrant populations and acculturation differences in migrant groups. They have been possible because cancer is one of the few chronic diseases for which high quality population-based disease surveillance systems have been in place for many years in many countries of the world. Cancer registry data has been widely applied toward the production of cancer atlases [ 39 ], studies analyzing the spatial distribution of particular cancer sites [ 40 ], and studies assessing spatial clustering [ 41 ]. Most recently, cancer studies have been undertaken which build on the combined resources of cancer registry data and increasingly available GIS tools. Because address at diagnosis is available for most registry cases it can be geocoded and integrated in a GIS with social and environmental attribute information available at a variety of geographic scales. Examples of such approaches include studies of childhood cancer which examine rate differences in areas of low versus intense agricultural pesticide use [ 42 ], heavy traffic patterns [ 43 ], or high air pollution [ 44 ]. Alternatively, cancer registry data can serve to identify population-based cases for studies using case-control or cohort designs, which can in turn be integrated into a GIS for area attribute data. Examples of this approach include case-control studies of childhood leukemia and traffic patterns [ 45 - 48 ]. and a studies of breast cancer incidence associated with residence in high pesticide use areas in a large case-control study [ 49 , 50 ]. and in a large cohort study [ 51 ]. For these types of studies, cancer registry data offer both a number of strengths and limitations. Primary strengths include the comprehensiveness of geographic coverage, detailed information on disease subgroups, and rich covariable information on demographic characteristics for each newly diagnosed case of cancer. Because registry data are abstracted from medical records and reflect information for a snapshot in time, primary limitations include the lack of historical information on various factors of potential interest including residential mobility and relevant personal behaviors. Cancer registries typically collect information on the residential address for individuals newly diagnosed with cancer at the time of that diagnosis. Since this is the locational information which serves as the basis for national and international statistics on area cancer rates, it is also useful for looking at area characteristics associated with rate differences, although inferences about etiologic associations are limited for these long latency diseases, and even more so for residentially mobile populations. The Surveillance, Epidemiology and End Results (SEER) program of the National Cancer Institute (NCI) offers county-level incidence data for its member registries, which cover part or all of eight states, through its SEER*Stat software. Because it provides direct access to individual cancer records, users must first sign a data access agreement. County-level mortality data for the entire United States, collected and maintained by the National Center for Health Statistics (NCHS), is also accessible through SEER*Stat. These data include all causes of death, not just cancer deaths. Selected county-level cancer data may also be accessed through the NCI's Cancer Mortality Maps and Graphs and State Cancer Profiles web sites. The latter was launched in 2003 and contains a host of innovative statistical graphics. Many individual state registries also offer additional geographically referenced data. For example, the Florida Cancer Data System web site allows users to generate a variety of county- and facility-level tables and county-level maps on demand. The Kentucky Cancer Registry also offers a county-level mapping application. New York State offers a limited set of ZIP code level data for the four most common cancer types in the mid-1990s. Currently, county-level cancer incidence data is not available nationally. 2. Population data The United States Census Bureau is the principal source of data on the entire population; most countries have comparable agencies. Since cancer rates are calculated by dividing the number of cases by the number of people at risk, census data is frequently referred to as "denominator data". Census data are readily available in electronic format through the Census Bureau web site, . Data are available in three basic formats. American FactFinder is a web-based application that allows users to drill down through geographic levels to find data tables of interest. It is most useful for data queries that are well-focused. Data may also be downloaded through an ftp server. This method obtains raw text files that require computer code to be written before the data can be easily accessed or manipulated. This method is most useful for users with large data needs who are in possession of some database programming skills. The third approach is to purchase DVDs from the Census Bureau's Customer Service center. The DVDs allow data output in many spreadsheet and database formats, facilitating the ability for users to process and analyze the data. There are also a large number of third-party vendors who offer similar services [ 52 ]. The four primary data files emanating from the 2000 census are named Summary File 1 through Summary File 4 (SF1–SF4). SF1 contains population counts by age, sex, race and ethnicity and basic housing characteristic information for the entire population, to the block level. SF2 contains similar information, detailed for ethnic subgroups, American Indian and Alaska Native tribes, and multiple-race individuals. These data are suppressed when the total number of individuals in a given geographic unit totals fewer than 100. SF3 contains detailed housing, demographic, and socioeconomic data to the census block group or census tract level, based on a long form that was sent to one in six households. Census block groups have an optimal population size of 1,500 and census tracts have an optimal population size of 4,000, though in practice populations vary widely. SF4 contains the same information as SF3 for detailed race and ethnic groups, with the same suppression rule as SF2. In addition to these four primary data files, the Census Bureau also provides digital cartographic boundary files for political entities in the country, as well as approximations of postal code boundaries known as ZIP code tabulation areas (ZCTAs). The Census Bureau also conducts the American Community Survey (ACS), an ongoing survey designed to reach 3 million households each year nationwide. The goal of this survey is to allow the publication of detailed demographic and socioeconomic information more often than once a decade. Data for geographic units totaling more than 65,000 people will be released annually, while data for smaller geographic units will be based on either a three or five year moving average. It will replace the census long form, which will not be administered in 2010. There will undoubtedly be a challenging adjustment period as public health researchers begin to use ACS data. At present, the level of information available for intercensal time points is quite limited, and derives from Census Bureau estimates at the state or county level. These estimates are used in the calculation of cancer rates by federal and state agencies, although some research has shown that they are not especially reliable, particularly county-level estimates for specific race groups [ 53 ]. Various private vendors publish intercensal estimates for areas smaller than counties, though it is impossible to verify their accuracy. Since many vendors use the Census Bureau estimates as controls (for example, vendor estimates of ZIP code populations in a county must add to the Census Bureau estimate for that county), vendor estimates necessarily suffer from the same limitations as the Census Bureau estimates. Finally, some state governments publish their own population estimates. Generally, these estimates are thought to represent improvements over the Census Bureau estimates because of higher levels of local knowledge and a broader use of data sources. We are unaware of any independent efforts to evaluate these claims, however. Examples include the population estimates and projections published by the California Department of Finance, and those by the Epidemiology Program of the Cancer Research Center of Hawaii. The latter population estimates were developed in response to a concern that the Native Hawaiian population was substantially undercounted in previous censuses, and are used by the NCI in calculating national cancer rates. The 2000 census allowed respondents to select more than one race, although cancer data are only beginning to be collected in this manner. As a result, population data from 2000 must be "bridged" back to the earlier single-race categories to allow comparisons with earlier data. NCHS developed a sophisticated bridging algorithm taking into account age, sex, distribution of single-race groups within counties, and other covariates [ 54 ]. This algorithm is reflected in the 1991–2003 population projections and estimates that are published on the NCI web site and included in their statistical software. The Census Bureau itself uses a simpler algorithm in its estimates, allocating equal proportions of each multiple-race combination to the constituent single races [ 55 ]. Given the multiplicity of population estimates and methods for calculating them that are available, it is important to be aware of the sources of these data, and how they may influence the confidence associated with a particular research result. This is especially true for small-area analyses, where uncertainties are highest. In addition to the issues noted above, it is important to realize that even the decennial census counts are not as accurate as popularly believed. The census represents an attempt to enumerate the population as of a single date, but invariably some people are missed or double-counted. These undercounts and overcounts are differential by race, socioeconomic status, and geographic area, potentially biasing cancer rates [ 56 , 57 ]. Countless epidemiologic and geographic studies make use of census data in some capacity, including most studies that report cancer rates for geographic areas. It is also quite common to use census data where individual-level data are not available, particularly for indicators of socioeconomic status [ 58 - 60 ], educational attainment [ 61 ] and housing characteristics [ 7 ]. Table 2 summarizes the population data sources described in this section. Table 2 Sources of Population Data Dataset name Source Agency URL Geographic Resolution 2000 Census Summary Files 1–4 US Census Bureau Census Tract, Block Group or Block (varies by data element) American Community Survey US Census Bureau Areas with populations >65,000 E-1 City/County Population Estimates, with Annual Percent Change California Department of Finance City/County US Population Data, 1969–2001 National Cancer Institute County 3. Surveys In addition to the Census Bureau as a primary source of sociodemographic attribute data, special survey data can provide valuable information on these characteristics for population groups in some areas. Perhaps one of the best known such surveys is the CDC-sponsored Behavioral Risk Factor Surveillance System (BRFSS), which is touted as the "world's largest telephone survey". Designed in the 1980s to track trends in behavioral risk factors at the state level, this ongoing system of national surveys also provides subarea and subgroup information within some of the larger states. Some researchers have estimated county-level behavioral risk factor prevalence by combining the statewide BRFSS data with county-level demographic data [ 62 , 63 ]. A mapping application to view BRFSS response data at the state and metropolitan level is also available . Another well-known national survey is the NCHS's National Health and Nutrition Examination Survey (NHANES), which has been in place since 1960 and combines questionnaire information with a national physical examination and biomonitoring program. NCHS also sponsors a National Health Care Survey (NHCS), a National Health Interview Survey (NHIS), a National Immunization Survey (NIS), and a National Survey of Family Growth (NSFG). Similarly designed large-scale efforts to track temporal and area differences for targeted health behaviors within a state include California's Tobacco Survey, Women's Health Survey, and Health Information Survey (Table 3 ). Table 3 Sources of survey data. Survey data recorded at the ZIP code level are designed to give valid estimates of risk factor distributions at the State level. Dataset name Source Agency URL Geographic Resolution Behavioral Risk Factors Surveillance Survey (BRFSS) Centers for Disease Control ZIP code National Health and Nutrition Examination Survey (NHANES), National Health Care Survey (NHCS), National Health Interview Survey (NHIS), National Immunization Survey (NIS), National Survey of Family Growth (NSFG). National Center for Health Statistics Metropolitan Statistical Area, National Region California Tobacco Survey California Department of Health Services ZIP code California Women's Health Survey California Department of Health Services ZIP code California Health Information Survey UCLA Center for Health Policy Research ZIP code Although population survey data has not been extensively incorporated into GIS studies to date, these resources may in the future provide some opportunity to characterize regional differences in behavioral risk profiles targeted for specific health outcomes. 4. Environmental data Over the past several decades there has been a large increase in the availability of spatially registered environmental data in the United States and other countries. Much of these data have been collected as a result of environmental regulations or government-funded research efforts. Examples of US programs to collect spatial data on concentrations or releases of pollutants in the environment include the United States Geological Survey (USGS) National Assessment of Water Quality program (NAWQA) , the Environmental Protection Agency (EPA) National Air Toxics Assessment database , and EPA's Toxic Release Inventory program . EPA has organized environmental data in an umbrella database called Envirofacts Data Warehouse . Some states have extensive efforts to collect additional environmental data. An example is California's Pesticide Use Reporting program ) that requires reporting of all agricultural pesticide use at the level of Public Land Survey System sections (a unit approximately one square mile in area). There are several issues to consider in using these data for assigning "exposure" in epidemiologic studies. Monitoring data collected for regulatory purposes should be carefully evaluated for its usefulness for estimating individual exposures. The fate and transport of the chemicals in the environment should also be considered. Simple proximity measures to sites of chemical releases may not adequately describe the transport of the chemical in the environment. The likely route of exposure should be considered along with the biological plausibility for an association between the exposure and disease under study. Finally, much of the environmental monitoring data was collected within the past decade and reconstructing exposure over longer periods more relevant to cancer incidence will be challenging. Environmental databases have begun to be used in epidemiology studies of cancer to determine if disease mortality or incidence rates are higher in areas with specific environmental exposures (i.e., ecologic study designs) or as a means of classifying individuals with respect to their potential exposure in an analytic epidemiologic study design (i.e., case-control, cohort studies). With few exceptions, the residence location is used as the geographic location for assigning exposure. Below we provide an overview of the various types of spatially registered exposure data and include examples of their use in epidemiologic studies of cancer. a. Water quality data The US EPA is responsible for regulating public drinking water supplies. A water supply is regulated if it has 5 or more connections or serves at least 25 people. Routine monitoring is required for a variety of contaminants and naturally occurring elements including disinfection by-products, arsenic, nitrate, certain pesticides and volatile organic chemicals. States are required to report violations of the Maximum Contaminant Levels (MCL) to EPA. Since 1996, EPA has been required to maintain a National Contaminant Occurrence Database (NCOD) using occurrence data for both regulated and unregulated contaminants in public water systems. The majority of historical public water supply measurement data, however, reside with the states. Some states record the latitude and longitude of the locations where the water samples were taken (location in the distribution system, point of entry to the distribution system, or water source location). The location information is typically not publicly available but may be available to researchers with appropriate approvals. The water quality data are reported by utility and to be useful for epidemiologic studies a linkage to the towns served must be established. In larger metropolitan areas multiple utilities may serve a city or, conversely, one utility may serve multiple towns and subdivisions. Therefore, establishing an accurate linkage between the study participant's addresses and water utilities is essential to avoid misclassification of exposure. Long-term exposure metrics can be calculated when a lifetime water source history is collected. Examples of studies using public supply water quality monitoring data include studies of disinfection by-products [ 64 - 66 ]., nitrate [ 67 , 68 ]., radionuclides [ 69 , 70 ]., and arsenic [ 71 , 72 ]. Contaminants such as disinfection by-products and volatile organic compounds vary in concentration across a public supply distribution system. GIS-based modeling efforts have been used to improve estimates of exposure at individual residences [ 73 , 74 ]. In contrast to public water supplies, private domestic wells are not regulated and there are no monitoring requirements, although well owners may be required to provide some water quality information upon the sale of a property in some states. Some states have conducted representative surveys of private well water quality [ 75 ]. A nationwide survey was conducted by EPA in 1988–1990 [ 76 , 77 ]. The US Centers for Disease Control (CDC) conducted a survey of coliform bacteria, nitrate, and atrazine in private wells in nine Midwestern States . The paucity of historical water quality data for private wells limits the exposure assessment for epidemiologic studies of cancer in this population. The USGS NAWQA program has been collecting information on nutrients, pesticides, volatile organic compounds, radionuclides, and major ions in more than 50 river basins and aquifers since 1991. All of the measurement data include spatial attributes. Because the goal of this research effort is to understand ambient water quality (not necessarily the same as drinking water quality) these data may not be of direct use in epidemiologic studies. However, the NAWQA data may be useful in modeling efforts to estimate contaminant levels in private wells. EPA also maintains two data management systems containing water quality information collected by federal, state, and private groups for surface and ground waters in all 50 states. The Legacy Data Center (LDC) is an archived database with data dating from the early 20th century up to the end of 1998. STORET contains data collected beginning in 1999, along with older data documented data from the LDC. Table 4 summarizes the sources of water quality data. Table 4 Sources of Water Quality data Database name Source Agency URL Geographic Resolution National Contaminant Occurrence Database EPA Public water utility National Water Quality Assessment (NAWQA) Data Warehouse USGS Latitude and longitude Legacy Data Center/STORET EPA Latitude and longitude b. Air pollutants The EPA collects and processes monitoring data from states on six criteria air pollutants (carbon monoxide, nitrogen dioxide, ozone, sulfur dioxide, particulate matter [PM10 and PM2.5], lead) and hazardous air pollutants, of which 188 have been identified. The hazardous air pollutants (HAP), also known as air toxics, are those for which there is some evidence of an increased risk for cancer or adverse reproductive outcomes. Routine monitoring of HAPs is not required and the monitoring data that exists is sparsely distributed compared with the criteria air pollutants. The data are maintained in the Air Quality Systems database. EPA compiles HAP emissions from stationary sources (points and areas) and mobile sources in a National Toxics Inventory (NTI) database (now combined with the National Emissions Trends data in the National Emissions Inventory database), which is updated at three-year intervals. To do the updates, EPA obtains emissions inventories from state environmental agencies and supplemental data from other sources, including the Toxic Release Inventory. The first nationwide inventory was in 1996. The spatial scale of the emissions data varies by type of source. Location information for point sources emissions is available, whereas area-source emissions are estimated at the county level. Using a dispersion model EPA has estimated the annual average HAP concentrations for each census tract in the contiguous US [ 78 ]. These datasets are summarized in Table 5 . Table 5 Sources of Air Quality Data Dataset name Source Agency URL Geographic Resolution Air Quality System database EPA Monitoring stations (latitude, longitude) National Emissions Inventory EPA Varies (point locations, county level) Air pollutant monitoring data has been used in studies of lung cancer, which have generally employed some type of dispersion model to estimate exposure for metropolitan areas or census tracts [ 79 - 81 ]. Recently the modeled concentrations of HAP have been used to evaluate childhood cancer incidence [ 44 ]. Other studies have also evaluated traffic density and childhood cancer incidence [ 43 ]. c. Agricultural Pesticides In the United States the U.S. Department of Agriculture (USDA) is the main federal agency responsible for collecting information on pesticide use on crops and livestock. The availability of historical agricultural pesticide use data in the US has been reviewed [ 82 ]. The first comprehensive survey of pesticide use on crops occurred in 1964 [ 83 ] and periodic surveys were conducted thereafter through the 1970s. These early surveys only provided national or regional estimates of crop-specific use for individual pesticides. From 1986 onwards, the USDA surveys produced state-specific estimates of pesticide use on field crops in the major producing states and from 1990 onwards, biannual state-specific estimates of pesticide use on fruits and vegetables were also available. Several states have collected their own pesticide use information but most data collection efforts have been recent. Oregon enacted legislation requiring reporting of agricultural pesticide use beginning in 2002; however, insufficient funding was provided for additional years. State pesticide use data are most comprehensive for California, which has had some type of mandatory reporting for agricultural pesticides since the 1950s, currently overseen by the California Department of Pesticide Regulation. Beginning in 1969, information about restricted-use pesticides was made public. In 1990, a new law required growers to report all pesticide use on crops on a monthly basis, including the pesticide name and manufacturer, crop treated, the public land survey section where the pesticide was applied, the date and time of application, number of acres treated, method of application, and application rates. The availability of this detailed pesticide use data at the spatial scale of a section led to the development of methods to link the use data to cancer incidence data [ 84 ] for use in an ecologic study of childhood cancer at the census tract level [ 42 ]. The California data have also been used in a case-control study of pancreas cancer [ 85 ], cohort study of breast cancer [ 51 ], and an as-yet unpublished case-control study of childhood cancer. Methods have also been developed to estimate potential pesticide exposure at residences by linking pesticide use data to crop maps [ 86 , 87 ]. Pesticide "exposure" is assigned to homes that have crop fields within distances that reflect likely pesticide drift. Table 6 summarizes the sources of pesticide data. Table 6 Sources of Pesticide Data Dataset name Source Agency URL Geographic Resolution Agricultural Chemical Use USDA State California Pesticide Use Reporting database California Department of Pesticide Regulation Public Land Survey Section (approximately one square mile) d. Industrial releases and hazardous waste The Emergency Planning and Community Right to Know Act of 1986 in the United States requires certain industries to report to EPA annually their releases and waste management activities involving specific toxic chemicals. The data are available to the public in a database called the Toxics Release Inventory (TRI). Manufacturing, metal mining, coal mining, and electric generating facilities must report the estimated mass of toxic chemicals released into the environment (air, water, land, or underground injection), treated on-site, or shipped off-site for further waste treatment. Reporting is required only for facilities that meet certain minimum criteria in terms of the pounds of toxic chemical produced or processed; persistent chemicals that bioaccumulate are subject to lower minimum reporting requirements. The regulations do not require environmental monitoring, so much of the data are estimates of releases. Location information is reported by the business and is not verified by EPA. Some of the strengths and limitations of these data for environmental health studies has been described [ 88 , 89 ]. Canada also requires reporting of emissions of chemicals rated by the International Agency for Research on Cancer as likely, probable, and possible human carcinogens for 64 industrial sectors [ 90 ]. These data form part of the Canadian Environmental Quality Database, which also contains a national inventory of municipal waste disposal sites, municipal drinking water data, air quality data, and historical industrial location and productivity data [ 91 ]. A large multi-province case-control study of 18 cancer sites was conducted with the aim of linking residential histories by postal code to the environmental database for cancer surveillance. To date, one analysis of residential proximity to 7 types of heavy industries and risk of non-Hodgkin lymphoma (NHL) has been published. Residential proximity within 3.2 km of copper smelters and <0.8 km of sulfite pulp mills was associated with an increased risk of NHL [ 92 ] after adjusting for employment in the industries evaluated. Earlier case-control studies of NHL [ 93 ] and leukemia [ 94 ] found elevated risks for residing close to industrial sites but these studies relied on a self-reported assessment of the distance of the residence from industrial facilities which may be subject to recall bias. The EPA maintains information on the location of waste handlers, waste treatment facilities and waste sites that are regulated under the Resource and Conservation Recovery Act (RCRA) and the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA), also known as the Superfund law in the RCRAInfo database available through the Envirofacts Data Warehouse. Information on the location of companies issued permits to discharge waste into rivers is maintained in the Permit Compliance System database (also available through Envirofacts). These data sources are summarized in Table 7 . Table 7 Sources of Hazardous Waste Data Dataset name Source Agency URL Geographic Resolution Toxics Release Inventory EPA Latitude, longitude HazDat ATSDR Latitude, longitude RCRAInfo EPA Latitude, longitude Permit Compliance System (PCS) The U.S. Agency for Toxic Substances and Disease Registry (ATSDR) was established by Congress in 1980 under CERCLA. Since 1986, ATSDR has been required to conduct a public health assessment at each of the sites on the EPA National Priorities List, waste sites deemed to be the most hazardous. The aim of these evaluations is evaluate exposure to hazardous substances and health effects among the population living in vicinity of the site [ 95 ]. The location of the sites and information on specific contaminants by the type of media (soil, air, water) in which they were measured are available from the ATSDR HazDat database web site. Limitations of these monitoring data for cancer studies include the limited historical measurement data. A few studies have evaluated cancer incidence among those potentially exposed to hazardous waste sites [ 96 ] or municipal waste sites and incinerators [ 97 , 98 ]. The reconstruction of historical exposure to releases from industries and waste sites is difficult for studies of cancers of long latency. A few studies have evaluated proximity and residence duration near sites. Long duration of residence within one-half mile of a chemical plant manufacturing PCBs was positively correlated with blood serum PCB concentrations [ 99 ]. However, none of the epidemiologic studies to date determined whether proximity resulted in meaningful exposure to chemicals from the sites. Confounding by socioeconomic status should also be evaluated because manufacturing and waste facilities are more likely to be located in neighborhoods of lower socioeconomic status [ 100 ] and socioeconomic status is associated with the incidence of some cancers. 5. Remote sensing/aerial imaging Remotely sensed data include images of the earth and our atmosphere obtained by satellites or aircraft. The usefulness of the information depends largely on the technology used to obtain the imagery and the additional processing that has been done to georeference the data. The USGS Earth Resources Observation Systems Data Center (EDC) is the major U.S. storehouse of these data. Aerial photography has been available since the early part of the twentieth century. Digital Orthophoto Quadrangles (DOQs) which are digital images of aerial photos which combine the image characteristics of a photo with the georeferenced qualities of a map are available through EDC from 1987 through the present. DOQs are available in black and white, natural color, or color-infrared images and have 1-meter ground resolution. Satellite imagery useful for land cover characterization includes the multispectral Landsat imagery available as early as 1972. USGS has created historical land use and land cover data derived from 1970s and 1980s aerial photography (the Land Use and Land Cover Data). A national land cover datasets (NLCD) derived from Landsat multispectral imagery for 1992 is available. The Multi-resolution Land Characteristics (MRLC) national dataset which represents land cover in 2000 is currently being developed. Table 8 summarizes these data sources. Applications of these data to studies of cancer have included mapping residences on crop maps to estimate their probable exposure to agricultural pesticides [ 49 , 87 , 101 ]. Table 8 Sources of Remote Sensing Data Dataset name Source Agency URL Geographic Resolution Digital orthophoto quadrangles USGS 1:12,000 Satellite imagery USGS 1 meter to 1 km National Landcover Dataset (NLCD) 1992 USGS 30 meters Multi-resolution Land Characteristics (MRLC) 2000 Centralized geospatial data availability The data sources we have described are available from a multitude of federal and state agencies. The National Cancer Institute's Geographic Information Systems web site offers links to many of these sources, as well as links to freely available geographical tools and resources. There have also been several initiatives to try and compile spatial data into a shared, centralized information system [ 102 ]. Such centralized systems offer the promise of standardized data coding systems, file formats and geographic boundary definitions. They also facilitate the sharing of metadata, or descriptive information about the data. The leader in this endeavor has been the Federal Geographic Data Committee . The FGDC is a consortium of federal agencies with the charge of developing the National Spatial Data Infrastructure (NSDI), a set of technologies, policies, standards and procedures that facilitate the creation and sharing of geospatial data. Among the achievements of the FGDC is the establishment of the National Spatial Data Clearinghouse, a central catalog of links to geospatial data and metadata. In 2003, an enhanced web portal was launched to further facilitate access to this data. Many states have echoed the national clearinghouse with clearinghouses of their own. The New York GIS Clearinghouse , for example, boasts over 400 member institutions providing links to thousands of datasets. The cancer data collection community has yet to fully engage this resource. As of January 2004, no cancer incidence or mortality data was available through the national clearinghouse. The keyword "cancer" provided only a link to the Environmental Defense Scorecard, a web site from which various environmental data sets can be accessed, particularly those published by the EPA . Most of the very limited data in the "human health and disease" category accessible through the web portal consisted of hospital and other health facility locations for a smattering of states. In some cases, the steps required to make cancer data available through the national clearinghouse would be modest. For example, the NCI's mortality data, geographic boundary files, and associated metadata used in its Cancer Mortality Maps and Graphs web site are easily accessed and downloaded, and only minor modifications would be required to make them compliant with FGDC standards. The DataWeb is another centralized online data resource, consisting of a network of online data libraries created in a collaboration between the CDC and the US Census Bureau. The libraries consist of both microdata and aggregate data in numerous categories. Available health data includes NHANES and NHIS survey data and county-level mortality. Information in DataWeb is accessed through DataFerret, an application that prepares data sets for the user to download. It allows users to select a "databasket" of variables and then recode those variables as needed. Users develop and customize data tables and may download them to their desktop in a variety of common formats. Conclusion In this article we have surveyed the distinctive characteristics of spatial data, along with commonly available sources of data relevant to etiologic cancer research. Spatial analysis is invaluable for data exploration, identification of geographic patterns, generation of new hypotheses, and providing supporting evidence about existing hypotheses. A geographic perspective will be increasingly relevant as GIS software, spatial analytic methods, and the availability and quality of geographically referenced data continues to improve.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539245.xml

521185

MicroRNA Is a Major Regulator

Since their discovery a decade ago, microRNAs (miRNAs) have emerged as major regulators of gene expression in eukaryotes of all kinds. Only 20 to 40 nucleotides long, a miRNA binds to a specific target sequence within a much longer messenger RNA (mRNA), inhibiting its translation and thus controlling expression of the corresponding gene even after the DNA itself has been read. Within the human genome, there are about 250 genes that code for miRNAs. Each miRNA has the potential to bind to many different transcripts. Variations in miRNA sequence dictate the gene transcripts to which each will bind most strongly. It has become clear that miRNAs play a critical role in controlling gene expression, for example, in larval developmental transitions and neuronal development in the worm Caenorhabditis elegans , growth control and apoptosis in the fruitfly Drosophila melanogaster , hematopoietic differentiation in mammals, and leaf development, flower development, and embryogenesis in the plant Arabidopsis thaliana . Despite their significance, the full range of genes miRNAs target is unknown, as is the best method for discovering them. In a new study, Debora Marks, Chris Sander, and colleagues describe an algorithm for determining the targets of miRNAs, and show they include more than 10% of all human genes. The algorithm uses three factors to evaluate whether a potential target site is likely to actually be regulated by miRNA. First, the target site must have some degree of sequence complementarity to one or more of the known miRNAs. Second, the strength with which the predicted target and its miRNA bind together, which can be calculated from the sequence and other structural factors, must be higher than some threshold. Finally, evolutionary conservation—the presence of the target–miRNA pair in different organisms—is factored in, because the likelihood that the target and miRNA actually pair in vivo is greater if the pair is found in multiple types of organisms. Using these principles, and the specific weighting they assigned to each factor, Marks and colleagues identified 2,273 genes in humans, rats, and mice that are likely targets for miRNA regulation. This is probably an underestimate of the total, since the researchers required each candidate gene to have at least two miRNA target sites. The authors identified another 2,128 genes with only one target site, but note that the false-positive rate here is likely to be high. Whatever the final number, the implication is that several thousand of our approximately 30,000 genes are under the control of miRNAs. Of special interest is that these putative targets include many genes known to be associated with the fragile X mental retardation protein, a crucial but still poorly understood player in mRNA regulation, whose absence leads to a type of mental retardation called fragile X syndrome. microRNA gene networks The researchers' findings also reinforce several emerging principles of miRNA-based regulation. First, it is widespread among multicellular eukaryotes, and sequences are surprisingly conserved. Of the 78 known miRNAs in Drosophila , 28 have close relations in mammals. Second, an individual miRNA may regulate multiple genes—Marks and colleagues found that the average miRNA interacts with seven distinct mRNAs, with a range from 0 to 268. Third, the genes regulated by a single miRNA may be functionally related, such as components of the protein degradation system or specific signal transduction pathways. Fourth, single genes may be regulated by multiple miRNAs—the gene that encodes amyloid precursor protein, for example, has at least eight miRNA sites—suggesting that expression may be combinatorially controlled by numerous cellular influences. These results provide resources for a host of experiments to elucidate the mechanism of miRNA action, which is not well understood. Several of the identified mammalian miRNA–target pairs have near-perfect matching sequences. In both plants (where miRNAs were first discovered) and animals, such matches are associated with degradation of the mRNA. The authors fully recognize that their algorithm, called miRanda, is not the last word in miRNA target identification. In order to improve both the search for targets and the algorithm itself, they are making the algorithm and full sets of results in vertebrates available free to other researchers ( www.microrna.org ), who can modify its parameters as experimental results and new models dictate.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC521185.xml

526417

Genome wide analysis of common and specific stress responses in adult drosophila melanogaster

Background During their life, multicellular organisms are challenged with oxidative stress. It is generated by several reactive oxygen species (ROS), may limit lifespan and has been related to several human diseases. ROS can generate a wide variety of defects in many cellular components and thus the response of the organism challenged with oxidative stress may share some features with other stress responses. Conversely, in spite of recent progress, a complete functional analysis of the transcriptional responses to different oxidative stresses in model organisms is still missing. In addition, the functional significance of observed transcriptional changes is still elusive. Results We used oligonucleotide microarrays to address the specificities of transcriptional responses of adult Drosophila to different stresses induced by paraquat and H 2 O 2 , two oxidative stressors, and by tunicamycin which induces an endoplasmic reticulum (ER) stress. Both specific and common responses to the three stressors were observed and whole genome functional analysis identified several important classes of stress responsive genes. Within some functional classes, we observed that isozymes do not all behave similarly, which may reflect unsuspected functional specificities. Moreover, genetic experiments performed on a subset of lines bearing mutations in genes identified in microarray experiments showed that a significant number of these mutations may affect resistance of adult Drosophila to oxidative stress. Conclusions A long term common stress response to paraquat- or H 2 O 2 -induced oxidative stresses and ER stress is observed for a significant number of genes. Besides this common response, the unexpected complexity of the stress responses to oxidative and ER stresses in Drosophila, suggest significant specificities in protective properties between genes associated to the same functional classes. According to our functional analysis, a large part of the genome may play a role in protective mechanisms against oxidative stress in Drosophila.

Background Cells are frequently submitted to exogenous or endogenous stresses. In aerobic cells, reactive oxygen species (ROS), produced by respiration and other biological processes, are a major source of endogenous stress. These ROS include the superoxide radical (O 2 • - ), hydrogen peroxide (H 2 O 2 ) and the highly reactive hydroxyl radical (OH•). Increased endogenous production, exposition to exogenous sources of ROS or reduction in antioxidant defense capacity cause molecular damages such as alterations in proteins, lipids and DNA, and may lead to cell death. Oxidative stress is believed to limit the lifespan of multicellular organisms [ 1 , 2 ] and oxidative lesions have been implicated in several human cardiovascular and neurodegenerative diseases [ 3 ]. A better understanding of the in vivo responses to oxidative stress is thus of major fundamental and practical importance. Much data describing the action of ROS and their derivatives in cultured cells are now available. For instance, ROS have been shown to activate signal transducing components, like p53 or members of the NF-κB pathway, resulting either in increased antioxidative protection or in activation of apoptotic pathways [ 2 ]. Nevertheless, a comprehensive integrated picture of the in vivo cellular responses of metazoans to oxidative stress is still not available. Genetic data suggest that different protection mechanisms are involved in vivo according to the type of ROS that induces the stress [ 4 ]. In addition, a wide diversity of macromolecules may undergo oxidative damage and induce secondary cellular stresses. These secondary effects of ROS could be similar to the alterations in macromolecules observed in other stress conditions, such as heat stress, endoplasmic reticulum (ER) stress (induced by accumulation of misfolded proteins), or UV-induced DNA damage. The relative importance of these common and specific responses to oxidative and other cellular stresses still has to be determined. In the yeast Saccharomyces cerevisiae , microarray experiments have shown that similar transcriptional responses are observed in a large number of different environmental conditions, including oxidative stress induced by H 2 O 2 or menadione [ 5 , 6 ]. According to the authors, this common environmental stress response (CER) may reflect the need for yeast to adapt quickly to rapidly changing external conditions. Similar transient variations of protein levels were also observed in proteomic experiments and highlighted the existence of an H 2 O 2 stimulon [ 7 ]. It is not clear whether such common stress responses exist in long-living multicellular organisms since, unlike unicellular organisms, their cells are probably submitted to slower and smaller variations of the extracellular medium; furthermore, the survival of just one cell is not generally crucial for the survival of the organism as a whole. Considering its powerful genetic and genomic tools, Drosophila melanogaster is a relevant model to address the question of the specificities of in vivo responses to various stresses in multicellular organisms and to identify novel genes that play a protective role. Nevertheless; there are some limitations to such studies on living flies. Firstly, adult flies are mainly composed of post mitotic cells; thus data obtained with flies may be relevant for comparison with the stress response of post mitotic tissues in other organisms (for instance mammals' neurons) but could be less useful to address the question of stress response in dividing cells. Secondly, limitations also arise from ROS-generating compounds delivery which in Drosophila is usually performed through food ingestion. This method severely limits kinetic studies of acute stress responses on flies, since, within a few hours, large fluctuations in the quantity of ingested food are observed in batches of flies transferred to a new medium. To overcome this problem some experiments were performed on starved flies. The major issue with such an approach is that the observed transcriptional changes could result from the starvation stimulus as well as from the effect of the studied compound. Therefore such experiments may in fact characterize the interference of two different stress responses with different induction times and kinetics rather than a bona fide oxidative stress response. A previous microarray study, performed with such a strategy on 4500 Drosophila genes, analyzed changes in transcription induced by paraquat, which produces superoxide radical (O 2 • - ) intracellularly: 5.2% of the genes (n = 236) were found to be stress responsive. Kinetic analysis revealed that transcriptional modifications lead to the establishment of a more or less stable new expression profile 12 hours after stress induction, thus indicating the existence of a long term stress response (LTSR) [ 8 ]. This stability probably reflects the late response to paraquat-induced stress but variations before 12 h are more difficult to interpret since they may also arise from the stress effects of starvation. This analysis also did not address the question of the specific responses to different ROS and could not distinguish between specific responses to oxidative stress and responses common to other cellular stresses. Furthermore, since the arrays covered only 30% of the estimated total number of Drosophila genes, this study was also limited for functional statistical analysis and detailed analysis of functional classes involved in stress response. From previous considerations, we chose to focus on comparisons of the long term transcriptional response (LTSR) of adult flies 24 h after ingestion of different stress-generating compounds. These responses may be representative of those of postmitotic tissue exposed to physiological chronic stress conditions (even if the level of stress is certainly higher in our experiments). Thus we investigated, at a full genome wide level, the transcriptional LTSR in adult Drosophila submitted during 24 h to three types of stresses: paraquat or H 2 O 2 -induced oxidative stresses and tunicamycin-induced ER stress. This latter drug inhibits N-linked glycosylation, thus leading to an accumulation of misfolded proteins in the ER (referred to as ER stress) which is known to elicit a specific response: the Unfolded Protein Response (UPR) [ 9 ]. We show in this paper that some of the transcriptional changes observed for these three stress conditions are similar, thus defining a class of multiple stress responsive genes. Nevertheless, in addition to this common long term stress response (CLTSR), many genes are transcriptionally regulated in a stress-specific manner. A statistical analysis identified classes of molecular functions or cellular processes over-represented inside clusters of genes undergoing transcriptional changes. Unexpectedly, both up and down regulations were observed for members of the same functional class. This may reveal novel functional specificities for these genes. In addition, we investigated whether genes that display significant transcriptional variations play a functional role in oxidative stress resistance. We present data suggesting that this could be the case for a large number of the stress-responsive genes identified in our study, which emphasizes the polygenicity of the stress responses, at both a molecular and a functional level. Methods Stocks All the lines tested for paraquat stress resistance were collected from the Bloomington stock center. To minimize genetic background effects, when the mutation was linked to a w + transposon insertion, the line was outcrossed for 4 generations against a w - isogenic strain of Canton S background before stress experiments. For homozygous lethal mutations we analyzed flies heterozygous for the mutation issued from a cross between the mutant line and the same w Canton S strain. Stress resistance tests and collection of fly tissues We used 50 ml vials containing 1 ml of a solid medium composed of 1.3% low melting agarose, 1% sucrose and either 1% H 2 O 2 , 5 or 15 mM paraquat, 12 μM tunicamycin (all from Sigma) or no toxic compound (control tubes). These compounds were incorporated at 45°C to avoid loss of activity. 3 day old males were placed in groups of 30 in these vials and maintained at 26°C with a 12:12 light-darkness alternation. In survival tests, dead flies were counted twice a day until the end of the experiment. In each experiment at least 3 vials of 30 male flies were used. To test mutant lines we performed three independent experiments in order to minimize false positive detection. Survival data were submitted to a log-rank analysis to detect statistically significant survival differences between mutant and w Canton S flies. We considered that a mutation had a significant effect on survival under oxidative stress when the mean of log 10 (p-log-rank) for the three experiments was lower than -3 and at least two experiments had p-log-rank < 0.001. For microarray experiments, for each condition, 200 Canton S males were kept 24 hours on the corresponding medium and then frozen in liquid nitrogen for subsequent RNA extraction. Independent batches of males from separate experiments were used for replicate experiments. All fly manipulations were performed at the same stages of the 12:12 light cycle to prevent any undesirable effects from circadian variations. Sample preparation and data analysis We analyzed 4 samples of control flies and 3 (paraquat 15 mM) or 2 (all other conditions) samples of stressed flies. Total RNAs were purified by three rounds of Trizol reagent (GIBCO/BRL) extraction before precipitation. cDNA were synthesized from 10 μg total RNA aliquots and biotin-labelled cRNA targets synthesized using the BioArray high yield RNA transcript-labelling kit (Enzo Biochem) according to the manufacturer's instructions. Hybridizations on Drosophila Genome Arrays (Affymetrix) and subsequent washing were performed on a GeneChip Fluidics Station according to the manufacturer's instructions before scanning on a GeneArray scanner. Data extraction was performed first by the MAS5 Affymetrix program which provides absolute values (AV) and detection p-values (DP) for each probe set. These data were loaded into an Access database for subsequent analysis. We retained only the experimental points that presented a mean value greater than 0.1 for the DPs of all the different samples of at least 1 of our 4 experimental conditions. This reduced the number of probe sets further analyzed from 14028 to 8976. The AV data from each microarray were then normalized against the AV mean value of the 4 control samples by a quantile method which performs optimally [ 10 ]. Since a large number of flies (~300) were used for each RNA sample hybridized to a microarray, variations in signal arising from individual transcription differences are greatly reduced. This is reflected in the high values of correlation coefficients between microarrays corresponding to the same experimental condition (data not shown). The normalized values were used for further comparison of each of the stress condition samples with the 4 control samples and for statistical validations of the variations using the SAM program [ 11 ]. For this analysis, we used a fold change threshold value of 1.5 and a mean FDR (false detection rate) lower than 10%. 1368 independent probe sets that fulfilled these conditions for at least one type of stress were retained for further analysis. A hierarchical divisive clustering of the data of these probe sets was performed using the SOTA [ 12 ] implementation available at . For each probe set, the ratios for all the combinations between stress conditions AVs and reference AVs were computed and Ln2 transformed. The SOTA algorithm used on this dataset with linear correlation distance with 0 offset, 1000 cycles and 1.01 variability threshold parameters, led to the detection of 19 clusters. Functional analysis Information from the Gene Ontology (GO) database (December 2003, [ 13 ]) was combined with the Affymetrix data through the THEA program to investigate which classes are over- or under-represented in the dataset of stress responsive genes. Briefly, according to the Gene Ontology hierarchical structure, each probe set was assigned, when possible, to its original annotation and to the associated parent annotations. The number of probe sets for the different GO terms was computed for groups of probe sets defined according to different criteria (such as whole microarray probe sets, detected probe sets or probe sets belonging to a given cluster). For each GO term G, the distribution between the group D of all the detected probe sets (N G D probe sets issued from a total of N D , probability P G = N G D /N D ) and a group C of particular interest, such as a cluster (N G C probe sets issued from a total of N C ) are compared. The hypothesis of equal repartition between these two groups would predict that, inside the N C probe sets of group C, N C *P G probe sets should be associated to the GO term. We computed the p-value P N for the null hypothesis of no association between the two distributions, with a binomial distribution with N C tries, a probability P G and N G C successes. Threshold values for P N helped to define the GO terms over- or under-represented in the group C. Quantitative real time RT-PCR analyses Experiments were performed as described [ 14 ] with 2 μg of the total RNA samples used for microarrays (control, P15, P5, H1 and T12). Primers were designed to generate an amplicon of about 100 nucleotides and their sequences are described below (Forward/Reverse primers): FBgn0015039: TATGCTCTTCAACCTACTGCTGC/TAGGCGTAAAATTGAATCCACTC FBgn0010383: GACGCTGAACGGATATGGCAT/ATGTAGGTCATCCCGAACTGTC FBgn0015035: CAACTCTGAATTTGGCTCTCATCC/AGCGGGTTTCTCCTCCTCAA FBgn0034334: GAAGCCGGATATGTTACGCAAG/TTCACCAGATAGCCGATGATG FBgn0038024: CCTCAACAAGTACCCGAATGTG/TACTCCCTTCAGTTCCACGGC RP49: CCGCTTCAAGGGACAGTATCTG/CACGTTGTGCACCAGGAACTT Annealing temperature was 62°C except for RP49 and FBgn0034334 transcript level quantifications, for which it was 60°C. We normalized samples by comparison with the levels of the RP49 housekeeping gene. Levels of transcripts under various stress conditions are compared with the transcript level observed in control flies. Results Transcriptome variations in adult Drosophila are strongly dependent on the type of stress to which they are submitted We wished to compare the transcriptomes of flies submitted to continuous stresses induced by ingestion of paraquat, H 2 O 2 or tunicamycin at concentrations leading to similar effects on viability. Survival curves were obtained for 3 day old male flies raised on media containing different concentrations of these drugs (Fig. 1 ). Concentrations of 1% H 2 O 2 , 5 mM paraquat and 12 μM tunicamycin had similar effects on the survival of flies and were chosen for further studies. A paraquat concentration of 15 mM was also used for comparison with previous studies [ 8 ]. Figure 1 Lifespan reduction in Drosophila submitted to paraquat-, H 2 O 2 - and tunicamycin-induced stress 3-5 day-old Canton S wild type males were placed at t = 0 by groups of 30 in vials containing 15mM paraquat (P15: ●), 5 mM paraquat (P5: ◆), 1% H 2 O 2 (H1: △) or 12 μM tunicamycin (T12: □). Dead flies were counted twice a day to determine survival. 3 vials of 30 individuals were used for each condition. When no toxic compound had been incorporated in the medium, more than 90% survival was observed at t = 120 h (not shown). Similar average lifespan was observed for P5, H1 and T12 around t = 80 h while it was significantly reduced in P15. Arrow indicates the time (t = 24 h) at which flies were collected for RNA extraction. Note the 20% lethality observed at this time for P15 condition. Dead flies were discarded before RNA extraction. RNA were obtained from separate experiments with 3 day old male flies reared at 26°C with a 12:12 hours light and dark (LD) alternation, on media containing no drug (4 reference samples), 15 mM paraquat (P15: 3 samples), 5 mM paraquat (P5: 2 samples), 1% H 2 O 2 (H1: 2 samples) or 12 μM tunicamycin (T12: 2 samples). Thus a minimum of 8 pairwise comparisons were made for each condition which ensured good statistical significance, as confirmed by quantitative PCR experiments (see below). Stresses were induced 24 h before collection of flies, which occurred at the same time (9 h) of the 12:12 hours light/dark cycle to eliminate the effect of circadian variations. Hybridizations were performed on Affymetrix GeneChips and the data processed as described in the Material and Methods. The statistical significance of transcriptional variations was assessed using the SAM program with a threshold of 1.5 [ 11 ]. A good correlation was observed between our P15 results and previous studies with the same stress conditions [ 8 ]: among the 246 stress responsive ESTs of Zou et al. , 201 were associated to a detectable probe set on our chip, 56% of which were selected by SAM analysis and 72% of which displayed a fold change greater than 1.3 (not shown). The remaining discrepancies may arise from differences in statistical selections, in analyzed tissues (thorax and abdomen in [ 8 ], whole flies in this study) or in genotype: compared to the w 1118 flies used in [ 8 ], our wild type Canton S flies were more resistant to paraquat 15 mM (mortality of 20% vs 54% at 24 h) and presented an increased medium lifespan (48 vs 35 days at 26°C, on standard medium). Among the 8976 probe sets significantly detected in adult flies (see Material and Methods), 1111 were up or downregulated with P15 treatment, this number being reduced to 608, 72 and 221 for P5, H1 and T12 treatments respectively. Thus, even with similar effects on flies survival, the fraction of the genome detected as stress responsive on microarrays was highly dependent on the nature of the stress, varying about ten times from 7% (P5) to 0.7% (H1). This first analysis defined a total of 1368 probe sets and 1343 genes which are induced or repressed at least in one stress condition. They were used for further analysis. Common and specific responses to different stress We plotted transcriptional variation correlations for the different oxidative stress conditions (Fig. 2 ). We observed a high degree of correlation between the two paraquat experiments (correlation coefficient c = 0.86, Fig. 2a ). The slope of the linear regression curve, however, was 1.14 which indicates that variations in transcription induced by paraquat may be dose-dependent for most genes in D. melanogaster . Lower correlations were observed for the linear regressions between P5 ratios and either H1 ratios (c = 0.64, slope = 0.40, Fig. 2b ), or T12 ratios (c = 0.43, not shown). Figure 2 Correlations between P5 and P15 or H1 microarray measurements For each of the 1368 probe sets selected in the SAM analysis, the mean Ln2 ratios between the absolute values (AV) for stress and reference conditions were compared in two dimensional plots. Bold lines are the linear regression curves for the two comparisons, the thin lines correspond to a complete correlation for eye guidance. A good correlation is observed between P5 and P15 with a slope of 1.14, while it is much weaker between P5 and H1 (slope of 0.4). Clustering analysis provided further information about the specificity of stress responses. We chose an unsupervised divisive clustering method (SOTA [ 12 ]) to analyze the data and we checked that other methods such as Self Organizing Maps [ 15 ] yielded similar results (not shown). The SOTA analysis predicted 19 clusters. The complete list of the 1368 probe sets with their cluster assignment is provided as Tab.S1 (Additional file 1 ) in supplementary data. In Tab.1, (Additional file 8 ) we present the average log-ratios in each stress condition for the 19 identified clusters. These data confirm the high correlation between the results for P15 and P5 and the general tendency toward smaller variations for P5. However, the genes included in cluster 7 exhibit a more severe repression in the P5 condition than in the P15 condition which may reflect a differential transcriptional response as a function of oxidant concentration. Notably, in clusters 5, 6, 7, 9, 10, 13, 16 and 18 which regroup 642 probe sets, significant variations for H1 were observed in the same direction than for P5 or P15. This suggests that, for a large number of genes, both oxidative compounds induce similar transcriptional responses. Therefore, the fact that the number of probe sets validated by the SAM procedure as being significatively affected in the H1 condition is smaller than in the paraquat conditions may be a consequence of a similar but weaker effect of H 2 O 2 on the transcriptome rather than fundamental differences in the responses to the two oxidants. What is the specificity of the oxidative stress responses induced by paraquat or H 2 O 2 compared to the ER stress response induced by tunicamycin? The 19 clusters from Tab.1 (Additional file 8 ) can be regrouped into 7 large classes of genes: Classes A and B contain genes respectively downregulated and upregulated in both oxidative stress and ER stress conditions. Inside these two large groups, 237 genes included in clusters 9, 10 and 13 are regulated in a similar fashion in all four stress conditions. Genes from classes C and D (48% of stress responsive probes) are respectively downregulated and upregulated by oxidative but not ER stress. Conversely, class F genes are upregulated in ER stress but not in oxidative stresses. In the atypical classes E and G, opposite variations are observed for the two types of stress: genes of class G are upregulated by ER stress but downregulated in oxidative stress while genes of class E display an opposite behavior. Overall, our data emphasize both specificities and similarities in these stress responses: the classes A and B (238 and 276 probe sets, respectively) which include genes displaying similar responses to both oxidative and ER stresses, represent a sizeable fraction (38%) of the stress responsive probes. In contrast, genes that vary in opposite directions, included in the classes E (104 probe sets) and G (60 probe sets), represent a smaller part of those stress responsive probes (12%). Classes of stress responsive genes Using the Gene Ontology annotation [ 13 ] we identified the molecular functions that are over- or under-represented among all the 1368 stress-responsive probesets compared to the distribution of functions identified for the complete set of 8976 detectable probesets (see Material and Methods). The analysis was first performed independently for each set of genes validated by the SAM procedure for each stress Table 2a to c (Additional file 9 ). A similar analysis for biological processes is given in supplementary Table S2 (Additional file 2 ). The most prominent functional classes over-represented in the paraquat sets are the peptidases (including peptidases which are part of the proteasome complex), the peptidase inhibitors, the glutathione transferases (GT) and oxidoreductase enzymes or electron transporters, including the P450 cytochromes. These classes could all be involved in the detoxifying processes that follow oxidative stress and are discussed in more detail below. In addition, lipases and more prominently the triacylglycerol lipases, also over-represented, may contribute to the regeneration of membranes after oxidative damage. Most of these features seem to be part of a general stress response since triacylglycerol lipases, peptidases with chymotrypsin or trypsin activity and GTs are also over-represented in the H 2 O 2 and tunicamycin specific sets of genes. The transaminases, the cyclohydrolases, the oxidoreductases and the hydroxymethyltransferases define the signature of functional classes over-represented in the two types of oxidative stresses. In contrast, proteins which bind to iron ions or monooxygenases are specifically over-represented in the paraquat set. As expected, the ER set presents features that are distinct from oxidative stress responses, that is the over-representation of hydrolases acting on glycosyl compounds, UDP-glucuronosyltransferases and tRNA ligases. This last class suggests that modifications of the translation rate may be an in vivo response to ER stress. Besides these ER stress-specific classes, peptidases with elastase activity and epoxide hydrolases are over-represented in both paraquat- and tunicamycin-induced stresses. Interestingly, this last class of proteins is involved in the metabolism of juvenile hormone which has been shown to be involved in heat stress response [ 16 ]. We then performed a similar analysis for the groups of genes identified in the clustering process. To increase the statistical significance of the analysis, we used the 7 groups A to G instead of the 19 initial clusters. This analysis, given as Tables S3 (Additional file 3 ) and S4 (Additional file 4 ) of supplementary data, allowed us to identify molecular function and process signatures in some clusters. For instance, for the genes repressed for oxidative and ER stress conditions (group A) specific over-representations are observed for alkaline phosphatases, diazepam binding proteins and proteins involved in acyl-CoA metabolism. Signatures of group B (genes upregulated for oxidative and ER stresses) include proteins involved in response to abiotic stimuli, including GTs and glutathione peroxydases, and tRNA ligases. This last feature may indicate that the organism reacts to sustained stress by an increase of protein synthesis. Nevertheless, genes involved in proteins biosynthesis are surprisingly under-represented among the stress responsive genes. Retinoid binding proteins and transporters are specifically over-represented in group C. Surprisingly, in this group of genes, a large number of peptidases are present along with a strong proportion of protease inhibitors. Signatures for group D (genes upregulated under oxidative but not ER stress) include chaperones associated with the heat shock response, glutamate synthases and proteins involved in ATP-dependent proteolysis. Glutamate synthases, together with the upregulated genes Ahcy13 and Eip55E , may be required to increase the pool of glutathione, a major actor in redox regulation and phase II detoxification [ 17 ]. Additional signatures for group D include two other processes, inosinate (IMP) biosynthesis and amino acid biosynthesis. Interestingly, we found under-representation of their parent processes (closer to the root of the ontology), namely nucleic acid metabolism and protein biosynthesis. Finally, in the ER stress specific groups F and G, the disulfide isomerase proteins and the glucuronosyltransferases, known to play an important role in the UPR following ER stress in yeast, are over-represented together with proteins involved in lipid metabolism. Overall, our data suggest that oxidative and ER stress induce comparable transcriptional modifications of a significant number of genes known to be involved in a limited number of functional classes. Gene-specific stress responses inside functional classes In contrast to previous work limited to partial analysis of the genome, the use of whole genome Affymetrix chips allowed us to investigate the specificity of transcriptional responses for genes associated with a given functional class. The thioredoxin system plays a major role in oxidative stress defense and needs to be better functionally characterized. In Drosophila, the peroxiredoxin proteins show thiol-dependent peroxidase activity and use thioredoxin, but not glutathione, as a source of reducing power. Indeed, Drosophila lacks glutathione reductase [ 18 ] and its function is apparently substituted by thioredoxin reductase. Interestingly, we observed significant differences in the transcriptional behavior of the members of the thioredoxin system when flies were submitted to paraquat stress. The thioredoxin class (GO:0030508) counts 7 members with either a sequence matching perfectly the consensus catalytic site WCGPCK ( CG4193 , CG3864 , Txl/CG5495 and CG1141 ) or with one mismatch ( CG8993 , CG13473 and CG3719 ). Only the Txl gene is significantly overexpressed over the 1.5 fold threshold, the other genes presenting no change or a weaker overexpression ( Trx-2 ). This strongly argues for a specificity of these thioredoxins in the defense process with an important role for the Txl gene. Similarly, among the five genes presenting a thioredoxin peroxydase activity (GO:0008379), only two ( CG12013 and CG1633 ) are overexpressed, the others ( CG12174 , CG5826 and CG6888 ) being unaffected in the studied conditions. Among the related genes only the peroxyredoxin CG11765 is overexpressed, while the glutathione peroxydase-like CG15116 , very similar to the thioredoxin peroxydase CG12013 , is significantly repressed. These specificities strengthen the concept of a functional diversification of these proteins in spite of their common ability to confer resistance to oxidants in Drosophila cells [ 19 ]. When the organism is challenged to oxidative stress, in addition to performing direct enzymatic detoxification of toxic compounds, it must also limit the appearance of the most toxic species. Therefore, since free iron catalyses the production of the highly toxic hydroxyl radical (OH•) from H 2 O 2 by the Fenton reaction, its concentration must be tightly controlled. Transferrin and ferritin proteins play a major role in this control [ 20 ]. Furthermore, variations in iron concentration may modify gene expression in the cell through the iron regulatory proteins Irp that bind to the iron responsive elements (IRE) located in their target genes UTRs. Under paraquat stress, we observed a coordinated and specific response of genes used in regulation of free iron concentration and iron-regulated response: the two ferritin subunits and the iron regulatory protein 1B ( irp1B ) are overexpressed, while the transferrin 1 ( tsf1 ) gene is severely repressed. Nevertheless, neither the irp1A nor the tsf2 and tsf3 genes show any significant transcriptional change. This suggests that each isoform of these families plays a specific role in iron homeostasis in the organism. More complex specificities can be observed in larger functional classes. The glutathione transferases (GTs; GO:0004364) play important roles the detoxification process after genotoxic stresses [ 21 ]. As expected, a large number of them (16/34) are overexpressed after paraquat-induced oxidative stress Table 3a (Additional file 10 ) but 4 are underexpressed under the same conditions. One of these GT repressed by paraquat (FBgn0034334) is also severely repressed by H 2 O 2 -induced stress. Moreover, among the 16 GTs overexpressed in paraquat-induced stresses, 7 are overexpressed and 3 underexpressed in ER-stressed flies, while 6 show no other significant transcriptional variation. Interestingly, all the GTs overexpressed in both paraquat and tunicamycin experiments are also slightly induced in H 2 O 2 -stressed flies. Overall, our data suggest that both "generalist" GTs that are able to protect the organism against various stresses and more specialized GTs, required only for protection against well defined stresses, coexist inside the cell. A similar conclusion can be drawn for the P450 cytochromes (GO:0015034). Among 58 detectable P450 cytochromes, 12 are underexpressed and 12 overexpressed during paraquat stress, 4 of these latter being also upregulated in tunicamycin-stressed flies (Tab.3b, Additional file 10 ). We observed a general tendency of these paraquat-inducible P450 cytochromes to be also overexpressed in H 2 O 2 -stressed flies. One cytochrome gene (FBgn0015035) displays peculiar behavior since it is induced by paraquat but strongly repressed by H 2 O 2 . Another gene (FBgn0015039) is induced specifically by tunicamicyn. Quantitative RT-PCR experiments confirmed the specificities observed on microarrays (Fig. 3 ). Figure 3 Comparison of transcript level variations detected with microarrays and with quantitative real-time PCR (Q-RT-PCR) Transcript levels were analyzed for genes encoding three P450 cytochromes (FBgn0015039, FBgn0010383 and FBgn0015035) and two glutathione transferases (FBgn0034334 and FBgn0010041). The Ln2 ratios between the transcript levels under stress conditions (P15, P5, H1 and T12) and the reference condition, obtained with Q-RT-PCR (white bars) and microarray analysis (black bars), are indicated for each gene. Error bars: standard errors. The complete data for the peptidases class (GO:0008233) analysis – given as Tab.S5 (Additional file 5 ) of supplementary data- provides a striking feature: most of the 131 peptidases selected by the SAM analysis (among 361 that were detectable) are downregulated by either both paraquat-induced oxidative stress and ER stress (54 peptidases) or paraquat-induced stress only (41 peptidases); nevertheless, a small number (36) of them are upregulated by paraquat. Closer examination of these latter genes revealed that 22 are proteasome endopeptidases. Further analysis of the proteins belonging to the proteasome complex, (GO:0000502) (which also contains proteasome regulatory proteins) shows that 33 out of 45 detectable proteasome constituents (73%) are likely upregulated by paraquat treatment (Tab. 3c, Additional file 10 ), both 19S and 20S subunits being coordinately regulated. Interestingly, the induction level is clearly correlated to the dose of paraquat used. Moreover, this induction is very specific since it is not observed in H1 or T12 conditions for any of these genes. The functional significance of this observation needs to be addressed in Drosophila strains mutant for proteasome subunits, challenged with paraquat, H 2 O 2 or tunicamycin stresses. Many genes transcriptionally affected by oxidative stress modulate oxidative stress resistance When a fly experiences an oxidative stress we can expect that the subsequent transcriptional modifications may arise from several mechanisms. Firstly, the organism can mount a protective response, for instance by inducing proteins which will reduce adverse consequences of the toxic compound. Only a few functional classes (such as GTs, electron transporters, chaperones) identified in our functional analysis of stress-regulated genes can be clearly associated to such known protective mechanisms from oxidative stress (Tab. 2, Additional file 9 ). Secondly the toxic drug itself may induce transcriptional changes which could play a role in its toxicity. The relative part of these protective or toxic responses to oxidative stress is unknown. We thus investigated whether genes detected in our microarray analysis could be involved in oxidative stress protection against paraquat or in its induced toxicity. We addressed this issue using a genetic approach, taking advantage of the availability of numerous strains bearing mutations in genes detected in the microarray paraquat set. Twenty nine such lines were recovered from public stock centers and adult flies were analyzed for their survival after transfer to a medium containing 10 mM paraquat. Most of the mutations used arise from P elements insertion in the 5' regulatory region of the genes which are expected to induce partial or complete loss of function mutations. Indeed, as shown in table 4 , most of them have been characterized as either lethal recessive mutations or hypomorphic loss of function mutations and, in some cases, do not complement a deficiency. Particular attention was paid to ensure that the genetic background was controlled in these experiments and stringent statistical conditions were used for the data analysis (see material and methods). Several conclusions can be drawn from these genetic experiments. a) First, as shown in Fig. 4a , under these conditions, a high proportion of the 29 tested strains present statistically significant survival differences from the w Canton S reference strain. Indeed, the results of our experiments show that 13 mutant lines out of 29 tested (45%) are either significantly more resistant (6 lines) or more sensitive (7 lines) to paraquat than their wild-type counterparts (Tab. 4, Additional file 11 and Fig. 4b ). This ratio is at least 10 times higher that what is expected from previous genetic screens (see discussion) and suggest a strong relationship between transcriptional stress response and functional in vivo susceptibility to oxidative stress. b) For the genes studied there is no clear correlation between the observed induction or repression under paraquat treatment and the effect of the mutation on the paraquat resistance or sensitivity phenotypes (Tab.4, Additional file 11 ). This suggests that, in the steady stress conditions used, both deleterious and protective gene regulations are taking place. c) Our genetic data point out the large functional diversity of genes that are able to modulate the oxidative stress resistance in vivo : ion channel ( Sh ), thioredoxin reductase ( Trxr-1 ), fatty acid elongase ( Baldspot ), phosphatase ( aay ) and phosphatase regulator ( CG9238 ), transcription factor ( Xbp1 ) and peptidase ( Acer ). Interestingly, among these 13 mutants, only 2 were previously known to be associated to oxidative stress resistance ( Sh and Trxr-1 ) and most of them had no known function in adult flies. Coupling between microarray and genetic experiments is thus a powerful way to extend our knowledge on the biological function of Drosophila genes without biased hypothesis and to provide some clues on the function of mammalian homologues. Figure 4 Resistance to paraquat-induced stress of flies mutant for genes identified in microarray experiments a) 29 Drosophila lines bearing mutations in genes identified in our microarray experiments as being stress-responsive were recovered from public stock centers. When the mutation was linked to a w + transposon insertion these lines were outcrossed with a w + Canton S reference line. 3–6 day old male flies were then tested for their resistance to oxidative stress 68 h after transfer to 10 mM paraquat medium. Tested flies were either homozygous (notation #i/#i in the X axis) for viable mutations or heterozygous (notation #i/ w ) for lethal mutations (in this case they are issued from a cross with w + Canton S females). For simplicity, identification of lines (#i) refers to the Bloomington stock number and the genotype of the line is provided in Tab. 4. We present in this Figure the results of one of three independent experiments that we used for the complete statistical analysis presented in Table 4. Compared to male flies issued from a cross between w - males and Canton S females (noted w/+, dark bar), significant differences in resistance or sensitivity to paraquat can be observed for a large number of the lines tested. Error bars: standard error. b) Example of survival curves on 10 mM paraquat-containing medium of some mutant male flies. Flies heterozygous for a lethal mutation in the Angiotensin converting enzyme related ( Acer ) gene are sensitive to paraquat, while flies homozygous for an insertion in the gene CG9238 are clearly more resistant to paraquat than w /+ control flies. Neither of these genes was previously suspected to play a role in oxidative stress resistance. For instance, we found that the Dgp-1 gene is induced in flies challenged with paraquat stress and that its disruption leads to stress resistance. The Dgp-1 protein is strongly similar to the mammalian GTPBP1 protein which presents a GTP binding domain and strong similarity with the elongation factor Ef-Tu [ 22 ]. Interestingly, expression of GTPBB1 is enhanced by gamma interferon in a monocytic cell line, suggesting that this protein in involved in host defense mechanisms. Nevertheless, no phenotype was observed in mice disrupted for this gene, maybe because of compensation by a gene of the same family [ 22 ]. Our data provide evidence that, in flies, Dgp-1, the GTPB1 homologue, is indeed involved in protective mechanisms against stress. The similarity with EF-Tu suggests that this protection might be linked to a downregulation of protein synthesis. In agreement to this hypothesis, it is noticeable that mutants for the translation negative regulator Thor present a significant sensitivity to paraquat stress (confidence index -2,4 in Tab. 4, Additional file 11 and Fig. 4a ) and has been shown to be sensitive to bacterial infection [ 23 ]. Discussion In this paper we present the characterization of the in vivo transcriptional responses of adult Drosophila males submitted to four different continuous stresses : paraquat (two conditions), H 2 O 2 or tunicamycin. Experiments on yeast submitted to several types of stress including oxidative stress have shown that fast transient responses occurring during the first three hours are followed by stable long term (>12 hours) changes [ 5 , 6 ]. Similarly, previous experiments on paraquat-induced stress in Drosophila have shown sustained long term changes in transcript levels which are more or less stable 12 hours after stress induction [ 8 ]. Since, as discussed previously, there are clear technical limitations to short term kinetic studies on Drosophila submitted to ingestion driven stress, we focused our efforts on the observation of these long term stress responses (LTSR) and performed our transcriptome analysis 24 hours after stress induction. At this time point, more than 95% of flies were alive for P5, H1 and T12 treatments, while 19% of lethality was observed in the P15 experiments. In addition, during the next 24 hours, in all conditions, less than 30% of the animals died. We thus expect that any secondary effects linked to the level of lethality are minimal in our experiments. In agreement with this assumption we noticed that in the experiments of Zou et al. similar results were obtained when the transcriptome was analyzed 12 hours (when lethality was negligible) or 24 hours after ingestion of 15 mM paraquat. Furthermore, when functional analysis was performed, we were unable to detect significant differences in the signature of the genes detected in the P5 and P15 experiments, which should be the case if the level of lethality plays an important role for gene transcription. We thus conclude that the secondary effects linked to the levels of lethality in the Zou et al. experiments and in our work do not significantly affect the transcriptome and that the variations observed are primarily due to the stresses experienced by the flies. Our data present clear evidence of a common long-term stress response (CLTSR) in transcription of Drosophila genes: at least 237 genes contained in clusters 9, 10 and 13 show similar changes in transcription for the three stressors studied. This number could be a minimum estimation of the extent of the CLTSR, since it is mainly limited by the weaker transcriptional variations observed in the H 2 O 2 -induced flies. We think that this may be due to a smaller number of cells experiencing stress when flies ingest H 2 O 2 . Additional data for comparison with various stress responses (immune stress [ 24 ], starvation [ 25 ] and, during the submission of this work, hyperoxia and aging [ 26 ]) are presented in Supplementary text T1 (Additional file 7 ) and Table S6 (Additional file 6 ) and confirm the existence of a core of similar transcriptional responses between these stresses. The CLTSR shows certain similarities with the common environmental response (CER) described in yeast [ 5 , 6 ]: in both cases heat-shock genes, genes involved in the detoxification processes, or associated with fatty acid metabolism and DNA repair show similar changes in all the stress conditions studied. Nevertheless, there are also obvious differences between these two responses. For instance, in contrast to what occurs in CER, no large scale coordinated transcriptional changes for genes involved in translation inhibition or energy production were detected in CLTSR. This may reflect the fact that, in our experiments, the CLTSR corresponds to a long-term adaptation of the stressed Drosophila cells, while the variations observed in yeast are transient (of course we cannot exclude long term post-transcriptional modifications in the translation apparatus and the metabolic pathways activities of stressed flies). Alternatively, these data may reflect differences in the adaptation of dividing cells (yeast) and post-mitotic cells (Drosophila) to stress conditions. For instance, in the latter case, upregulation of the iron responsive protein 1b gene may lead to translational downregulation of the succinate dehydrogenase gene through an IRE [ 27 ] and hence modulate energy production as in the yeast, but in a different way. Additionally, in Drosophila, translation repression may also be involved in stress response but relying on a small subset of genes (which would then not have been detected with our functional analysis). Interestingly, in support to this hypothesis, we found that the translational repressor Thor is induced under stress conditions and that mutations in this gene confer a slight but significant sensitivity to paraquat-induced stress. However, our finding that tRNA ligases are upregulated in oxidative and ER stress may indicate a requirement for increased protein synthesis under sustained stress conditions. Kinetic studies using another oxidative stress paradigm are needed to clarify this point. In view of our results, it would be also interesting to investigate possible variations in stress response in mammalian tissues either mitoticaly active or quiescent. Besides their similarities, the LTSRs also display marked differences. One of the most striking specific expressions is displayed by the genes encoding for the proteasome subunits. These proteins belong to the two large complexes 19S (regulatory complex) and 20S (proteolytically active complex) which, together, form the 26S proteasome [ 28 ]. Most of them (73%) are specifically induced by paraquat- but none by H 2 O 2 - or tunicamycin-induced stresses. It is also noticeable that, in contrast to proteasome constituents, ubiquitin protein ligases are under-represented among paraquat responsive genes. The 20S proteasome, inactive in its native form, is able to specifically degrade oxidized proteins in vitro and in vivo, and has been considered to be the main actor in this process [ 29 ]. Nevertheless, it has been recently proposed that, while the 20S proteasome is active during oxidative stress and limits the accumulation of oxidized proteins, the 26S, inactive in presence of ROS, "cleans" the cell in the following recovery process, eliminating thereby the accumulated altered proteins [ 30 ]. This seems to be a very important aspect of oxidative stress defense since oxidization of proteins can result in protein fragmentation and partial unfolding, and induce the formation of cytotoxic insoluble aggregates, a process that is known to be implicated in an increasing number of human pathologies [ 31 , 32 ]. The observed coordinated upregulation of genes encoding both 19S and 20S proteasome subunits when Drosophila cells are submitted to continuous paraquat stress strongly suggests that both complexes are indeed important in vivo for oxidized proteins degradation. We observed no such induction of proteasome components in H 2 O 2 -stressed Drosophila. This result is coherent with previous studies shoving that the proteasome subunit are not transcriptionally regulated in cultured mammalian cells treated with H 2 O 2 [ 33 ]. However this is surprising since it has been shown, in mammalian cells, that the proteasome is in fact involved in the degradation of misfolded glycoproteins as well as oxidized proteins after H 2 O 2 treatment [ 34 ]. Recent data in lens epithelial cells showed that H 2 O 2 induces an increase in proteasome activity and E1 ubiquiting activation enzyme levels without any increase in E1 mRNA levels [ 35 ]. In view of our data, we propose that two different strategies are used in D. melanogaster to deal with oxidative challenge and increase proteasome activity: one response, induced by H 2 O 2 , would rely on post-transcriptional mechanisms as shown in mammalian cells; while the other response, induced by paraquat, would rely on coordinated increase of transcription of the proteasome genes of both 19S and 20S subunits. A number of functional classes are clearly over-represented among the genes involved in the LTSRs. The analysis of these specific functional classes revealed an important heterogeneity of stress-specific responses among their members. For instance, we have shown that only a subset of genes potentially involved in the thioredoxin pathway are upregulated during paraquat stress. Whether the remaining genes are involved in an earlier phase of the stress response, in a subset of tissues or in other processes unrelated to stress protection needs to be addressed. Interestingly, in agreement with this last hypothesis, one of these genes, Jafrac2, which codes for a thioredoxin peroxidase, has been recently assigned an unexpected role in caspase-regulated cell death [ 36 ]. The P450 cytochromes and the glutathione transferases also display striking stress-specific responses. For the GTs, 3 genes are downregulated by tunicamycin and 4 by paraquat, while 6 are upregulated by paraquat and 7 by both drugs. When we tried to correlate this information with GT classifications [ 21 ] we found that the latter group contained almost exclusively δ-type GTs (Table 3a). This suggests that this insect-specific class, unlike other Drosophila GTs, may have acquired a broad-spectrum detoxifying function which is required to counteract both oxidative and tunicamycin-induced cellular damages and/or that these GTs molecular targets are altered in both types of stress. One important issue is whether our findings are representative of long term transcriptional responses in Drosophila submitted to real physiological chronic stresses. Indeed, the stress levels experienced by flies in this work are probably much higher than those experienced in real life. Nevertheless, the tight correlation that we observe between P5 and P15 experiments demonstrates that most of the genes undergoing transcriptional changes at a high concentration of paraquat display similar changes (although at a reduced level) when the concentration is threefold lower. This suggests that many genes identified in this study may also be induced in low intensity chronic stress. A striking feature of our results is the large number of genes not previously associated with stress response which show transcriptional changes under paraquat-induced oxidative stress conditions. We investigated the biological validity of these observations in a genetic study of mutations in some of these genes. Since our microarray data suggest that the stress responses may be highly polygenic (with at least 10% of the genome involved), we took a particular care to ensure that there was a controlled genetic background in these experiments. We found that 45% of the mutations tested were associated with either resistance or sensitivity to paraquat, which confirms this idea of a highly polygenic process. It should be stressed that, since many of the tests were performed on heterozygous flies, the proportion of genes functionally involved in oxidative stress resistance may be higher. Extrapolation of the results obtained with this small subset of 29 genes to the 1107 genes found to be regulated by paraquat, suggests that some 500 genes may modulate paraquat sensitivity in vivo . This contrasts with two previous genetic screens to detect paraquat hypersensitive mutants, which concluded that only a few genes are involved in paraquat hypersensitivity [ 37 , 38 ]. These studies however analyzed only EMS viable mutations on the X, 2nd and 3rd chromosome. They would thus have missed any lethal mutations that could confer a sensitivity phenotype to heterozygous flies by gene dosage reduction. In fact, when we performed a P{ w + ; UAS}- based screen we found that a large proportion of P-element insertions may confer H 2 O 2 or paraquat resistance or sensitivity ([ 14 ] and Girardot et al. unpublished) in agreement with the results presented here. If all the transcriptional responses to a stress were protective for the organism; we would expect a clear correlation between the direction of the transcriptional response of the genes studied and the effect of their mutations on stress resistance. A significant result of our experiments is that we could not find such a correlation. It thus appears that the transcriptional responses to oxidative stress may be either protective or deleterious for the flies. The simplest explanation for this result is that, besides the protective responses mounted by the organism cells (for instance in inducing detoxifying proteins), the paraquat also induces transcriptional changes that play a role in its toxicity. In mammalian cells, several transcription factors may be regulated by oxidative stress, either by direct modification by the ROS or through signaling pathways, and have either pro- (Jun, p53) or anti-apoptotic effects (NF-κB, HSF1) [ 2 ]. In addition, the choice between survival and apoptosis may depend on the intensity of the stress and on the cell type, as it has been clearly demonstrated in the case of p53 [ 39 ]. Signaling pathways which activate these factors are strongly conserved between mammals and Drosophila and it is conceivable that, like in mammalian cells, their activation in flies by oxidative stress may induce complex transcriptional responses of both pro-survival and deleterious factors. In this case the integration of these complex responses at the level of the organism will determine the final outcome (protective or deleterious) and, eventually, in the case of a transient stress of limited intensity, the return to an unstressed equilibrium state. Thus the protective or deleterious role of a stress responsive gene cannot be predicted simply but should be uncovered systematically by genetic studies. Interestingly, in our genetic experiments, halving the dosage of the Xbp1 gene resulted in increased sensitivity of flies to paraquat-induced stress. Xbp1 is known to be involved in ER stress response in mammals [ 40 ]. It has been shown that it is regulated by processing of its mRNA by the C-terminal endonuclease Ire1. Conversely, we observed no transcriptional change of Xbp1 in Drosophila challenged with tunicamycin but it is overexpressed in oxidative stress conditions. Our in vivo genetic study suggests that this regulation is functionally relevant to oxidative stress protection in Drosophila. Thus Xbp1 may protect against different stress conditions through different modes of regulation (transcriptional or post-transcriptional regulation). In agreement to the conservation of this mechanism between flies and mammals, it has been shown recently that, in a mammalian dopaminergic cell line, Xbp1 is induced by the parkinsonian mimetic 6-hydroxydopamine which is known to induce oxidative stress [ 41 ]. Another gene that affects the flies stress resistance in vivo is Acer . This gene encodes one of two Drosophila proteins homologous to the mammalian angiotensin converting enzyme (ACE) gene family. Controversial findings have linked Ace to stress resistance and aging ([ 42 ] and references therein). Acer is more similar to the mammalian gene Ace2 . It has been recently shown that both Acer and Ace2 are essential regulators of heart function [ 43 ]. Interestingly, complete targeted disruption of Ace2 in mice results in increased angiotensin II levels and upregulation of hypoxia-induced genes. In Drosophila, the targets of Acer are not known and complete loss of function of the gene results in embryonic lethality. We found that halving the dosage of Acer in adult flies results in increased sensitivity to paraquat stress. Considering the mammalian data, one hypothesis to explain this result is that heart cells of Acer /+ flies may already experience a mild hypoxic stress which sensitizes them to the additional paraquat-induced oxidative stress. Targeted expression of Acer in Drosophila heart cells may help to test this hypothesis. In this genetic study, based on a small subset from the genes found to be regulated by stress in our microarray experiments, we identified genes with no previously known function as in vivo modulators of oxidative stress resistance. Since genomic programs steadily increase the number of transposon targeted genes it will become easier to perform this kind of genetic analysis to increase our knowledge of integrated mechanisms of stress resistance in Drosophila. In conclusion, our data confirm that full genome scanning by microarray experiments and analysis of multiple experimental conditions constitutes a powerful tool to uncover potentially significant biological features that can be subsequently confirmed by genetic experiments. Supplementary Material Additional file 1 For each of the 1368 probe sets identified as stress responsive in our data analysis, we calculated and reported in this table, for each stress condition, the mean ratio <Stress condition> ≤ (AV stress i / AV ref j )> i,j where AV stress i and AV ref j correspond to the average value measured for the i th sample in the stress condition and the j th sample respectively in the reference condition. To facilitate visual inspection, we used a color code (red corresponding to upregulation, green to downregulation) with thresholds corresponding to fold changes of 1.8 (dark colors), 1.5 (medium) and 1.25 (light). The standard error for each measurement is given in parenthesis. For each probe set, the mean detection p-value from MAS5 analysis of reference samples is reported in column 4 and cluster assignment in column 9. . Click here for file Additional file 2 For stresses induced by a) paraquat (5mM and 15mM experiments), b) H 2 O 2 or c) tunicamycin we analyzed the distribution in biological processes (as defined by the Gene Ontology (GO) database) of the genes selected by the SAM analysis (responsive genes) and compared it to the same distribution for all the genes significantly detected on our microarrays (analysed genes). We report here the significantly over- or under-represented (P < 0.005) biological process and the number of analysed and responsive genes found inside these classes, for the different stress conditions. The p-value P associated to the null hypothesis of no association with a binomial distribution hypothesis is given for each class, (only classes with P < 0.005 were retained). For clarity of the figure some redundant branches of the tree were removed. Color codes for the classes: dark blue: classes present in the 3 stress responses; yellow: classes present in the two oxidative stress responses; green: classes present in paraquat and tunicamycin stress responses; light cyan: classes present in H 2 O 2 and tunicamycin stress responses. 34 Color code for statistical analysis: orange: underrepresented class, blue: over-represented class. . Click here for file Additional file 3 For over or under-represented molecular functions classes we report here the number of analyzed (column 3) and responsive genes found inside the 7 groups of clusters A to G (columns 4 to 10, see text for details on the definition of these groups). A schematic response to oxidative and ER stress of the genes included in these groups is given in the first two lines. The number of genes inside each group is given in line 3. A color code identifies cases when the number of genes differs statistically (p<0.005) from a random distribution: orange: under-represented class, blue: over-represented class. Click here for file Additional file 4 For over or under-represented biological process classes we report here the number of analyzed (column 3) and responsive genes found inside the 7 groups of clusters A to G (columns 4 to 10, see text for details on the definition of these groups). A schematic response to oxidative and ER stress of the genes included in these groups is given in the first two lines. The number of genes inside each group is given in line 3. A color code identifies cases when the number of genes differs statistically (p<0.005)from a random distribution: orange: under-represented class, blue: over-represented class. Click here for file Additional file 5 Stress response for the peptidases Click here for file Additional file 6 List of stress responsive genes detected in aging and hyperoxia, immune or starvation stress experiments were compared with our data. Lines 1 to 3 indicate the number of genes found to be repressed (-), induced (+) or either (total) in the different experiments. Lines 5 to 7 indicate the number of genes in each of these categories found among our 1397 35 stress responsive genes (classes A to F) and the corresponding percentage from the initial number. In lines 8, 9 (respectively 10, 11) the same analysis is reported for genes included in the A (respectively B) classes defined as common stress responsive classes in our analysis. O 2 , old, infection, starvation : expression data from the different experiments compared to our data. All : List of 26 genes which are responsive to at least 4 stresses in these independent experiments Click here for file Additional file 7 Relationships to other stresses Click here for file Additional file 8 Table 1: Stress response characteristics of clusterized genes.The 1368 probe sets retained after statistical analysis were submitted to a divisive clustering algorithm (SOTA) which predicted 19 clusters. For each probe set k inside a cluster we calculated, for each stress condition, the mean ratio R k = <Ln2 (AV stress i / AV ref j )> i,j where AV stress i and AV ref j denote the average value measured for the i th sample in the stress condition and the j th sample respectively in the reference condition. The mean of the R k values provides a measurement of the mean intensity of variation for the genes inside a cluster, which is reported in this table. To facilitate visual inspection, we used a color code (red colors corresponding to upregulation, green colors to downregulation) with thresholds corresponding to fold changes of 1.8 (dark colors), 1.5 (medium) and 1.25 (light). The number N of probe sets in each cluster is also reported. From these values we identified groups of clusters (named from A to G) which present close behavior and were used for statistical functional analysis. Clusters corresponding to the common long term stress response (CLTSR) are outlined in red. Click here for file Additional file 9 Table 2: Functional analysis of stress responsive genes.For stresses induced by a) paraquat (5mM and 15mM experiments), b) H2O2 or c) tunicamycin we analyzed the distribution in functional classes (as defined by the Gene Ontology (GO) database) of the genes selected by the SAM analysis (responsive genes) and compared it to the same distribution for all the genes significantly detected on our microarrays (analysed genes). We report here the significantly over- or under-represented (P<0.005) molecular functions and the number of analysed and responsive genes found inside these classes, for the different stress conditions. The p-value P associated to the null hypothesis of no association with a binomial distribution hypothesis is given for each class, (only classes with P<0.005 were retained). For clarity of the figure some redundant branches of the tree were removed. Color codes for the classes: dark blue: classes present in the 3 stress responses; yellow: classes present in the two oxidative stress responses; green: classes present in paraquat and tunicamycin stress responses; light cyan: classes present in H 2 O 2 and tunicamycin stress responses. Color code for statistical analysis: orange: under-represented class, blue: over-represented class. Click here for file Additional file 10 Table 3: Analysis of stress responses for members of some functional classes.From the 1368 stress responsive probe sets we extracted the subsets associated with genes annotated in the GO database as a) glutathione transferases (GO:0004364), b) P450 cytochromes (from list at http://p450.antibes.inra.fr/) and c) proteasome component (GO:0004299). For each probe set k within one of these subsets, we calculated and reported in this table, for each stress condition, the mean ratio <Stress condition >k =<(AVstressi / AVref j)>i,j where AVstressi and AVrefj correspond to the average value measured for the ith sample in the stress condition and the jth sample respectively in the reference condition. To facilitate visual inspection, we used a color code (red corresponding to upregulation, green to downregulation) with thresholds corresponding to fold changes of 1.8 (dark colors), 1.5 (medium) and 1.25 (light). The standard error for each measurement is given in parenthesis. For each probe set, the mean detection p-value from MAS5 analysis of reference samples is reported in column 3 and cluster assignment in column 8. In column 9 additional information is reported for each class: in a) we indicate the GT class deduced from sequence comparison with human and mouse GTs and from [44] (D: delta, O: omega, T: theta, T2: distantly related to theta, Z: zeta); in b) the name of the genes are reported; in c) we indicate the proteasome subunit to which the genes defined in column 2 belong. Note that in c) a large number of genes not retained by SAM analysis (without cluster number) seem to be upregulated in P15 condition. Genes used for comparison between microarray and quantitative RT-PCR (Fig. 3) are outlined in bold character. Click here for file Additional file 11 Table 4: Analysis of mutant flies' resistance to paraquat-induced oxidative stress. Paraquat resistance of 29 mutant lines was assayed in three independent experiments as described in Fig. 4. The survival data were submitted to a log-rank statistical analysis by comparison with w /+ reference flies. The results are presented in this table. Column 1 contains the tested genotype (same conventions as in Fig. 4a : Bloomington line numbers). The corresponding genotypes are described in column 6. The symbol for the gene affected is reported in column 2. Information from FlyBase about the allele used in this study is given in column 7 with a one character code: A: amorph; H: hypomorph; N: non complementation of deficiency; L: letal; R: recessive mutation; 5: insertion in the 5' regulatory region, 5'UTR or intron; C: insertion in the coding region. Column 8 indicates whether the tested line was outcrossed or not before the test. Column 4 is the result of a log-rank analysis of the second survival experiments shown in figure 4. A confidence index which refers to the mean of log10 (p-log-rank) for the three experiments is given in column 5. We considered that a strain had a significant effect on survival under oxidative stress conditions when this confidence index was lower than -3 and at least two experiments presented p-log-rank < 0.001. Under these stringent conditions 13 genotypes are shown to confer resistance (R) or sensitivity (S) to paraquat as indicated in column 3. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC526417.xml

212700

Developmental Origins and Evolution of Buchnera Host Cells

When it comes to exploiting a niche, endosymbionts take the prize. In endosymbiosis, one organism—the endosymbiont —invades the cells of another, in some cases taking up residence in a way that actually benefits the host. Bacteria are particularly adept at making themselves indispensable by insinuating themselves into some fundamental aspect of an organism's biology. The endosymbiotic hypothesis proposes that this is how certain eukaryotic organelles evolved from endosymbiotic bacteria. Insights into the mechanisms governing endosymbiosis will help biologists understand how this mutually beneficial relationship evolved and provide clues to one of the fundamental questions in biology: How did the eukaryotic cell evolve? Over 10% of insect species rely on endosymbionts for their development and survival. In this issue, David Stern and colleagues look at one of the most studied pairs, the pea aphid and Buchnera aphidicola , and discover clues to the molecular foundation of their shared fate. ( Buchnera , which can no longer survive outside its host cell, is thought to produce essential amino acids that the aphid cannot get on its own.) While it is known that Buchnera are transferred from clusters of bacteriocytes in the mother to the adjacent early-stage embryo, it has been unclear how the bacteriocytes develop. Previous studies of the bacteria's genome have failed to explain the genetic basis of Buchnera 's ability to invade aphid cells. Consequently, Stern and colleagues have focused on the bacteriocytes, the specialized insect cells that house Buchnera , shedding light on the development of these cells as well as on the evolutionary adaptations in the aphid that made the bacteriocytes hospitable to Buchnera . The researchers show that bacteriocytes differentiate and proliferate independently of Buchnera 's presence in the cell, and they identify three aphid transcription factors (proteins that regulate gene expression) that are expressed in three distinct stages during early-bacteriocyte development in the aphid embryo. The first protein is expressed just before Buchnera enters the embryo; a second, as the bacteria invades; and a third, after the transfer is nearly complete. A second wave of the same transcription factors occurs at a later stage in aphid embryo development and increases the population of bacteriocytes. This two-step specification of bacteriocytes, which occurs in related Buchnera -carrying aphid species, appears to be an evolutionarily conserved feature of aphids. It even occurs in an aphid species that once had a Buchnera endosymbiont and now has a yeast-like symbiont that lives outside the bacteriocytes. But this process is not observed in males of another aphid species that do not carry Buchnera . While traces of the first transcription factor activated in bacteriocytes are evident, the characteristic gene-expression pattern is not, and the aphids have no mature bacteriocytes. While it seems that the aphid has evolved new domains of expression in the bacteriocyte for these transcription factors—none of these transcription factors is expressed at a similar stage in other insects—the researchers cannot yet say whether these genes direct the specification of bacteriocytes. Still, these transcription factors are likely to play important roles in the bacteriocyte, suggesting that the union of aphids and Buchnera involved significant adaptations by the host. Aphid host of Buchnera endosymbionts

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC212700.xml

539053

Imaging Lymph Nodes with Nanoparticles

Accurate staging of cancers is one of the most important parts of the work up of patients for both prediction of prognosis and determination of the most appropriate treatment. And an essential part of this work up is assessing whether or not there has been lymphatic spread. Current methods include surgical removal of nodes for examination and various types of imaging, ranging from ultrasound to newer technologies such as magnetic resonance imaging (MRI). All these methods have problems; some are very invasive, others are very time consuming, and none are completely reliable. 3-D image of lymph node after automated analysis However in one of the more exciting crossovers from chemistry into medicine, researchers have developed nanoparticles to improve the diagnostic accuracy of MRI. The nanoparticles contain a central superparamagnetic iron oxide core and are covered by dextran, imparting long circulation times and biocompatibility. When injected intravenously, the nanoparticles localize to lymphoid tissue, and are internalized into macrophages. There is then a decrease in signal intensity on T2- and T2*-weighted images, and when metastases are present there is a recognizably abnormal pattern on MRI scans. In a previous paper published in the New England Journal of Medicine , Ralph Weissleder and colleagues described using these nanoparticles to assess lymphoid spread in patients with prostate cancer. Now, in a paper published in this month's PLoS Medicine , they have gone further by extending the analysis to patients with different types of cancer, and producing an algorithm that allows semiautomation of the procedure. The authors developed the algorithm in a training group of 36 patients and then validated it in a group of 34 patients. The results are encouraging: the analysis showed a sensitivity of 98% (95% confidence interval, 88%–99%) and a specificity of 92% (95% confidence interval, 87%–96%). The advantages of automating this procedure are substantial, not least because it can remove the problem of different observers assessing data differently. And what is more, once the data have been collected and assessed it is possible to reconstruct a virtual picture of the patient's lymph nodes, thus potentially allowing accurate surgical removal of the nodes.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539053.xml

539251

Negative and positive childhood experiences across developmental periods in psychiatric patients with different diagnoses – an explorative study

Background A high frequency of childhood abuse has often been reported in adult psychiatric patients. The present survey explores the relationship between psychiatric diagnoses and positive and negative life events during childhood and adulthood in psychiatric samples. Methods A total of 192 patients with diagnoses of alcohol-related disorders (n = 45), schizophrenic disorders (n = 52), affective disorders (n = 54), and personality disorders (n = 41) completed a 42-item self-rating scale (Traumatic Antecedents Questionnaire, TAQ). The TAQ assesses personal positive experiences (competence and safety) and negative experiences (neglect, separation, secrets, emotional, physical and sexual abuse, trauma witnessing, other traumas, and alcohol and drugs abuse) during four developmental periods, beginning from early childhood to adulthood. Patients were recruited from four Psychiatric hospitals in Germany, Switzerland, and Romania; 63 subjects without any history of mental illness served as controls. Results The amount of positive experiences did not differ significantly among groups, except for safety scores that were lower in patients with personality disorders as compared to the other groups. On the other side, negative experiences appeared more frequently in patients than in controls. Emotional neglect and abuse were reported in patients more frequently than physical and sexual abuse, with negative experiences encountered more often in late childhood and adolescence than in early childhood. The patients with alcohol-related and personality disorders reported more negative events than the ones with schizophrenic and affective disorders. Conclusions The present findings add evidence to the relationship between retrospectively reported childhood experiences and psychiatric diagnoses, and emphasize the fact that a) emotional neglect and abuse are the most prominent negative experiences, b) adolescence is a more 'sensitive' period for negative experiences as compared to early childhood, and c) a high amount of reported emotional and physical abuse occurs in patients with alcohol-related and personality disorders respectively.

Background It is difficult to assess the impact of childhood traumatic events on the psychiatric disorders in adulthood, as neither prospective research studies, nor experimental approaches are possible. Nevertheless, an increasing number of retrospective reports suggest that psychiatric disorders may be related to childhood psychological traumas such as neglect, physical or emotional abuse [ 1 - 6 ]. In particular, significant correlations between the severity of psychiatric symptoms and that of stressful and traumatic experiences during childhood were found [ 7 - 12 ]. Reports of physical and sexual abuse in childhood are more frequent in psychiatric patients than in the healthy population [ 13 - 16 ]; among these are patients diagnosed with affective disorders [ 17 - 19 ], somatization disorders [ 20 - 22 ], borderline personality disorders [ 3 , 7 , 23 - 25 ], substance-related disorders [ 26 - 28 ], and schizophrenic disorders [ 29 - 31 ]. Specifically, several studies have documented high rates of trauma in individuals with severe mental illness [ 32 ]. For a sample of schizophrenic women, Friedmann and Harisson (1984) reported that 60% of them had suffered childhood sexual abuse [ 33 ]. Abused patients displayed more pronounced symptoms such as hallucinations [ 34 , 35 ] and delusions [ 36 ]. Any conclusion to such reports, however, must be drawn by taking into consideration that the validity of childhood memories, particularly in psychiatric patients, may be questioned, as the range of childhood traumas indexed in these studies is generally limited, and often only childhood sexual abuse is targeted. Moreover, the observed relationships are correlational in nature, and do not justify the conclusion that childhood trauma favors the development of psychiatric disorders. Antecedents of developing psychopathology may also provoke certain parental behavior. Also, a third variable, such as social conditions, may have caused both childhood abuse and later pathological development. Another notable finding is that the prevalence rates of antecedent traumatic events vary considerably across studies. This may be due to different definitions of abuse which include more detailed [ 13 ] or more global [ 23 ] descriptions. Furthermore, the amount of psychosocial elements such as neglect, family disturbance, the nature of preexisting and subsequent attachment patterns, special competencies, etc., is difficult to be assessed or taken into account. Only a limited number of studies [ 37 , 38 ] have so far included control groups, allowing one to compare self-reports of abusive sexual experiences during childhood in psychiatric patients to those in the healthy population. There is also a lack of research studies that assess these issues within different cultural backgrounds. The present study sets out to evaluate reported positive and negative life events from early childhood to adulthood in psychiatric patients. We addressed some of the above-mentioned problems by examining abuse histories across a range of several psychiatric diagnoses within a controlled cross-national design. We sought to examine whether (a) negative life experiences are positively associated with psychiatric diagnoses in adulthood, and (b) early childhood and adolescence were 'sensitive periods', that is, whether psychiatric diagnoses were more closely related to negative experience in these developmental periods. The present study includes a German/Swiss and a Romanian psychiatric group, in order to determine whether reports vary between cultural backgrounds. Methods Subjects Patients were recruited from four Psychiatric Hospitals within two different cultural settings, Switzerland/Southern Germany versus the Moldavia region in Romania: the Center for Psychiatry Reichenau and the Center for Psychiatry Weissenau in Germany, the Psychiatric Hospital Münsterlingen in Switzerland, and the Psychiatric Hospital "Socola", Jassy in Romania. A total of 192 psychiatric inpatients (98 German and 94 Romanian psychiatric patients, range 18–78 years) filled in the questionnaire. Sixty-three control subjects without any history of psychiatric diagnosis were recruited from the clinical staff and the university employees (Konstanz in Germany, Jassy in Romania) as controls (38 Germans and 25 Romanians). The control subjects have been simply inquired whether they had any stationary hospitalization in the psychiatry; no further assessments have been done. After a full explanation of the study, written informed consent was obtained from all subjects. By considering the clinician-made diagnoses which were written down from the medical files available in the psychiatric clinics the patients were recruited from, the patients were distributed in four diagnostic groups: alcohol-related disorders (n = 45), schizophrenic disorders (n = 52), affective disorders (n = 54), and personality disorders (n = 41). At all psychiatric clinics in Germany/Switzerland and Romania the diagnoses were made according to the ICD-10 criteria. Within our patient groups, the following lifetime mental disorders were assessed by using the ICD-10: alcohol-related disorders (dependence syndrome, psychotic and unspecified mental disorders due to the use of alcohol), schizophrenic disorders (paranoid schizophrenia, schizoaffective disorder, and undifferentiated schizophrenia), affective disorders (bipolar depressive disorders, recurrent depressive disorder, cyclothymia, and dysthymia), and personality disorders (borderline, schizoid, paranoid, histrionic, dissocial, and dependent personality disorder respectively). A few patients within our sample were diagnosed with comorbid symptoms: 4 patients with affective disorders had symptoms of substance abuse and 7 of them had anxiety symptoms; also, within the schizophrenic disorders group, 2 patients had symptoms of alcohol abuse and 6 had depressive symptoms. There were also patients who received two diagnoses: one of which was a personality disorder (i.e., 5 patients with affective disorders, 7 with alcohol-related disorders, and 2 with schizophrenic disorders). In these cases, we considered the other diagnosis for the distribution into the diagnostic groups. Table 1 summarizes the demographical characteristics of all subjects. The patient groups were similar with respect to the psychiatric history. There were differences among groups concerning gender distribution, age, and education. The gender-distribution differences among groups were due to the high number of women within the control and the affective disorders groups. With regard to the noted age differences, the patients with affective and alcohol-related disorders respectively had higher mean age as compared to all the other groups. The education differences among groups are only due to the lower educational level in patients with alcohol-related disorders. Romanian patients with affective disorders had a longer psychiatric history than the German/Swiss ones [t(51) = 2.3, p < 0.05]. Regarding gender distribution and the average duration of education, the German/Swiss and Romanian diagnostic groups were similar. The German/Swiss controls were significantly older than the Romanian ones [t(43) = 4.4, p < 0.001] and the German/Swiss patients with alcohol-related disorders were significantly younger than the Romanian ones [t(61) = 3.5, p < 0.001]. Table 1 Demographic characteristics of the control and of the patient groups 1, 2 Alcohol Related Disorders Schizophrenic Disorders Affective Disorders Personality Disorders Controls G/S R G/S R G/S R G/S R G/S R Analysis N N N N N χ 2 p Gender 12 <.05 Female 6 10 10 11 14 19 11 6 23 15 Male 14 15 18 13 10 11 15 9 15 10 Mean ± SD Mean ± SD Mean ± SD Mean ± SD Mean ± SD F p Age 31 ± 9 45 ± 12 34 ± 8 36 ± 10 40 ± 12 44 ± 8 34 ± 8 32 ± 11 38 ± 13 28 ± 7 6 <.001 Education 2 ± 1 2 ± 1 3 ± 1 3 ± 1 3 ± 1 3 ± 1 2 ± 1 3 ± 1 3 ± 1 3 ± 1 4 <.01 Psychiatric history (yrs) 6 ± 1 7 ± 10 8 ± 9 11 ± 10 4 ± 6 9 ± 9 4 ± 5 6 ± 7 - - 2 n.s. 1 Abbreviations: G/S: German/Swiss; R: Romanian; Education: 0 = no education, 1 = school for the mentally challenged, 3 = middle school, 4 = high school; 2 Areas where the values are written with bold characters indicate significant differences between German/Swiss and Romanian subjects within the diagnostic groups (i.e., alcohol-related disorders, schizophrenic disorders, affective disorders, personality disorders and control group respectively). Material Life experiences were assessed with the Traumatic Antecedents Questionnaire (TAQ) [ 40 ]. The TAQ is a 42-item self-rating questionnaire, which covers 11 subscales enquiring into the severity of positive (i.e., competence and safety) and negative experiences (i.e., neglect, separation, secrets, emotional abuse, physical abuse, sexual abuse, witnessing, other traumas, and alcohol and drugs) during four developmental periods (ages 0–6, 7–12 13–18, and ≥ 19). Each subscale includes 2–6 items. Each item requires the occurrence of a certain type of experience for each of the different age periods. The subjects were asked to score on a frequency/intensity scale the degree to which it describes their experience: 0 ("never or not at all"), 1 ("rarely or a little bit"), 2 ("occasionally or moderately"), 3 ("often or very much"), and DK ("don't know"). In a subsequent step, the average scores were calculated within each developmental period for each of the 11 subscales. The procedure we used was the following: first, the "don't know" responses were noted in a non-numerical manner, by using asterisks (*) to indicate missing values and these values were counted as 0; secondly, the response scores were added up and the sum was divided by the total number of items within the subscale in that age period for which there were numerical scores. By using this procedure, we excluded "don't know" responses from the total scores calculation. Data analysis Comparisons of demographic data were made with analysis of variance (ANOVA) and with two-tailed unpaired t-tests for continuous variables. Chi-square analysis was used to compare nominal data. The differences between groups were evaluated individually for each TAQ scale by repeated-measures ANOVA with the cultural background (German/Swiss versus Romanian), psychiatric status (alcohol-related disorders, schizophrenic disorders, affective disorders, personality disorders or controls), and gender (female versus male) as between-subjects factors, and developmental period (4 periods) as within-subjects factor. The probability level for rejecting the null hypothesis was set at P < 0.05. Post-hoc comparisons were carried out to evaluate main effects and interactions using Bonferroni/Dunn tests. A principal components analysis was also applied to the entire sample in order to identify those factors, which could account for individual variability across the eleven scales of the TAQ. The principal components were derived by using varimax rotation to orthogonalize solutions. Results Positive experiences Table 2 lists group mean scores on each of the two positive experiences scales. The patients generally exhibited lower mean scores on reported positive experiences as compared to the controls. The reported level of competence did not differ between diagnostic groups [F(4,198) = 0.8, n.s] or cultural samples [F(1,198) = 0.1, n.s]. A main effect of developmental period [F(3,594) = 25.7, P < 0.001] was explained by the increase of competence from early childhood to adolescence (P < 0.05), and by the decrease of competence in adulthood as compared to adolescence (P < 0.05). Table 2 Mean scores of positive experiences across developmental periods among all groups Positive Experiences and Age at Onset Alcohol Related Disorders Schizophrenic Disorders Affective Disorders Personality Disorders Control Group Analysis Mean ± SD Mean ± SD Mean ± SD Mean ± SD Mean ± SD F p Early Childhood (0–6) Competence 1.5 ± 0.9 1.9 ± 0.9 1.7 ± 1.0 1.7 ± 0.9 1.7 ± 0.1 0.7 n.s Safety 1.4 ± 0.8 1.7 ± 0.8 1.6 ± 0.8 1.2 ± 0.8 1.6 ± 0.7 3.2 <.05 Latency (7–12) Competence 2.1 ± 0.8 2.0 ± 0.8 2.1 ± 0.7 1.9 ± 0.8 2.1 ± 0.9 0.2 n.s Safety 1.8 ± 0.8 1.8 ± 0.8 1.9 ± 0.7 1.3 ± 0.7 1.9 ± 0.7 5.1 <.001 Adolescence (13–18) Competence 2.1 ± 0.9 2.1 ± 0.7 2.2 ± 0.6 2.1 ± 0.8 2.3 ± 0.8 0.6 n.s Safety 1.9 ± 0.8 1.7 ± 0.8 1.8 ± 0.8 1.3 ± 0.8 2.0 ± 0.8 4.0 <.01 Adulthood (≥19) Competence 1.9 ± 1.0 1.8 ± 0.8 2.1 ± 0.8 2.0 ± 0.8 2.2 ± 0.8 1.3 n.s Safety 1.7 ± 0.8 1.7 ± 0.8 1.7 ± 0.8 1.5 ± 0.7 2.1 ± 0.7 4.9 <.001 For both cultural samples, patients with personality disorders reported lower values on the safety subscale than any of the other groups [F(4,215) = 4.5, P < 0.01]. Post hoc tests showed that patients with personality disorders (P < 0.001) and those with affective disorders (P < 0.01) reported less such experiences as compared to the controls. The reported level of safety increased through adolescence [F(3,645) = 11.1, P < 0.001], but the interaction of the developmental period with the psychiatric status revealed a decrease in safety accounts from the age of 13–18 years towards adulthood in all patient groups [F(12,645) = 2.8, P < 0.001]. Negative experiences Table 3 shows the mean scores of traumatic experiences for all patient groups and for the control group across developmental periods. Negative experiences were more frequent in patients than in controls as indicated by significant main effects of the psychiatric status on each of the nine subscales. In addition, there was an important increase of amount of reported negative experiences across developmental periods. Table 3 Mean scores of negative experiences among all groups Negative Experiences and Age at Onset Alcohol Related Disorders Schizophrenic Disorders Affective Disorders Personality Disorders Control Group Analysis Mean ± SD Mean ± SD Mean ± SD Mean ± SD Mean ± SD F p Early Childhood (0–6) Neglect 0.6 ± 0.5 0.7 ± 0.7 0.5 ± 0.5 0.8 ± 0.6 0.2 ± 0.3 8.1 <.001 Separation 0.5 ± 0.6 0.4 ± 0.5 0.4 ± 0.6 0.6 ± 0.8 0.2 ± 0.4 3.9 <.01 Secrets 0.9 ± 0.9 1.2 ± 1.1 1.0 ± 0.9 1.4 ± 1.1 0.6 ± 0.7 4.8 <.01 Emotional Abuse 0.7 ± 0.6 1.0 ± 1.0 0.7 ± 0.8 1.2 ± 0.9 0.3 ± 0.5 8.2 <.001 Physical Abuse 0.4 ± 0.6 0.4 ± 0.5 0.3 ± 0.5 0.7 ± 0.7 0.1 ± 0.3 5.7 <.001 Sexual Abuse 0.0 ± 0.3 0.2 ± 0.6 0.1 ± 0.3 0.3 ± 0.6 0.1 ± 0.4 2.0 n.s. Trauma Witnessing 0.4 ± 0.5 0.4 ± 0.5 0.5 ± 0.6 0.7 ± 0.9 0.1 ± 0.2 7.6 <.001 Other Traumas 0.4 ± 0.5 0.3 ± 0.4 0.2 ± 0.4 0.4 ± 0.6 0.2 ± 0.3 3.1 <.05 Alcohol/Drug Abuse 0.4 ± 0.8 0.3 ± 0.6 0.5 ± 0.7 0.5 ± 0.7 0.1 ± 0.2 4.0 <.01 Latency (7–12) Neglect 0.7 ± 0.6 0.8 ± 0.7 0.7 ± 0.5 1.1 ± 0.8 0.5 ± 0.5 5.0 <.001 Separation 0.6 ± 0.7 0.5 ± 0.6 0.6 ± 0.7 0.9 ± 0.8 0.5 ± 0.6 2.8 <.05 Secrets 1.0 ± 0.9 1.2 ± 0.1 1.1 ± 0.9 1.6 ± 1.1 0.8 ± 0.8 4.2 <.01 Emotional Abuse 0.8 ± 0.6 1.1 ± 0.9 1.0 ± 0.8 1.4 ± 0.9 0.7 ± 0.8 4.6 <.01 Physical Abuse 0.5 ± 0.7 0.6 ± 0.7 0.6 ± 0.6 0.9 ± 0.8 0.5 ± 0.6 2.5 <.05 Sexual Abuse 0.1 ± 0.3 0.2 ± 0.5 0.1 ± 0.2 0.5 ± 0.8 0.1 ± 0.4 5.3 <.001 Trauma Witnessing 0.5 ± 0.6 0.5 ± 0.5 0.7 ± 0.6 0.9 ± 0.8 0.4 ± 0.5 5.5 <.001 Other Traumas 0.4 ± 0.5 0.4 ± 0.4 0.4 ± 0.5 0.5 ± 0.6 0.3 ± 0.4 1.4 n.s. Alcohol/Drug Abuse 0.5 ± 0.7 0.3 ± 0.6 0.6 ± 0.7 0.6 ± 0.8 0.2 ± 0.4 4.1 <.01 Adolescence (13–18) Neglect 1.0 ± 0.6 0.9 ± 0.7 1.0 ± 0.5 1.2 ± 0.7 0.8 ± 0.6 2.5 <.05 Separation 0.9 ± 0.8 0.8 ± 0.8 0.9 ± 0.7 0.9 ± 0.7 0.6 ± 0.7 2.4 <.05 Secrets 1.1 ± 0.9 1.3 ± 1.0 1.1 ± 0.9 1.5 ± 1.0 0.7 ± 0.8 4.7 <.01 Emotional Abuse 0.8 ± 0.6 1.3 ± 0.9 1.1 ± 0.7 1.4 ± 0.9 0.8 ± 0.8 4.6 <.01 Physical Abuse 0.8 ± 0.7 0.5 ± 0.6 0.5 ± 0.6 1.0 ± 0.9 0.5 ± 0.6 5.0 <.001 Sexual Abuse 0.2 ± 0.4 0.2 ± 0.4 0.2 ± 0.4 0.5 ± 0.8 0.1 ± 0.3 3.2 <.05 Trauma Witnessing 0.6 ± 0.6 0.5 ± 0.5 0.7 ± 0.6 1.0 ± 0.8 0.4 ± 0.5 5.9 <.001 Other Traumas 0.6 ± 0.6 0.5 ± 0.5 0.4 ± 0.5 0.6 ± 0.5 0.3 ± 0.3 3.4 <.05 Alcohol/Drug Abuse 1.1 ± 1.0 0.6 ± 0.8 0.7 ± 0.9 0.9 ± 0.8 0.4 ± 0.7 5.3 <.001 Adulthood (19≥) Neglect 1.4 ± 0.8 1.3 ± 0.9 1.2 ± 0.7 1.2 ± 0.8 0.9 ± 0.7 3.0 <.01 Separation 1.6 ± 0.7 1.1 ± 0.9 1.5 ± 0.8 1.2 ± 0.9 1.0 ± 0.7 4.3 <.01 Secrets 1.1 ± 1.0 1.3 ± 1.1 1.2 ± 0.9 1.6 ± 0.9 0.5 ± 0.7 8.1 <.001 Emotional Abuse 0.9 ± 0.7 1.2 ± 0.9 1.1 ± 0.9 1.4 ± 0.9 0.6 ± 0.6 6.8 <.001 Physical Abuse 1.0 ± 0.9 1.0 ± 0.8 0.9 ± 0.9 1.0 ± 0.9 0.4 ± 0.6 5.1 <.001 Sexual Abuse 0.3 ± 0.6 0.6 ± 0.8 0.4 ± 0.7 0.5 ± 0.7 0.2 ± 0.4 2.8 <.05 Trauma Witnessing 0.8 ± 0.6 0.7 ± 0.7 1.0 ± 0.8 1.0 ± 0.9 0.5 ± 0.4 4.3 <.01 Other Traumas 1.2 ± 0.6 1.0 ± 0.7 1.1 ± 0.8 1.0 ± 0.6 0.4 ± 0.4 14.8 <.001 Alcohol/Drug Abuse 2.1 ± 0.8 0.8 ± 0.9 1.1 ± 0.9 1.0 ± 0.9 0.3 ± 0.6 29.3 <.001 With respect to the experiences of neglect , the psychiatric patients reported higher rates than the controls [F(4,214) = 5.7, P < 0.001], the post hoc tests revealing that all patient groups reported more such experiences as compared to the controls: patients with personality disorders (P < 0.001), alcohol-related disorders (P < 0.05), schizophrenic disorders (P < 0.05), and affective disorders (P < 0.05). There was an increase of the amount of reported neglect experiences across developmental periods [F(3,642) = 91.5, P < 0.001]. Across developmental periods there were significant effects of the psychiatric status [F(12,642) = 3.2, P < 0.001]: the post hoc tests showed that patients with personality disorders (P < 0.001) and with alcohol-related disorders (P < 0.01) reported a highly significant increase of the amount of neglect experiences across developmental periods as compared to the controls (Figure 1 ). Figure 1 Mean neglect score across developmental periods among all groups. The psychiatric patients reported higher rates than the controls [F(4,214) = 5.7, P < 0.001]. There was an increase of the amount of reported neglect experiences across developmental periods [F(3,642) = 91.5, P < 0.001]. Error bars stand for standard error of the mean. Asterisks indicate significant main effects of the psychiatric status. Irrespective of the psychiatric status and developmental periods, Romanian subjects generally reported a higher amount of neglect experiences, as shown by the main effect of the cultural background [F(1,214) = 6.4, P < 0.05]. Romanian patients, particularly those with schizophrenic disorders, reported a higher incidence of neglect experiences than their German counterparts (P < 0.01), as revealed by the interaction between the psychiatric status and cultural background [F(4,214) = 5.6, P < 0.001]. As indicated by the interaction between the developmental period and the cultural background, the mean scores of neglect experiences were higher in the Romanian sample as compared to the German/Swiss one for the earliest (0–6 years) period [F(3,642) = 5.3, P < 0.001]. Separation Patients, particularly those with alcohol-related disorders, personality disorders, and affective disorders reported more often separation experiences than controls [F(4,227) = 3.3, P < 0.01, P < 0.01 for post-hocs]. Mean scores on separation increased with age, and were highest in adulthood [F(3,681) = 103.0, P < 0.001]. Secrets Higher patient mean scores were confirmed by the main effect of the psychiatric status [F(4,182) = 6.8, P < 0.001], especially for those with personality disorders (P < 0.001) and with schizophrenic disorders (P < 0.001), as revealed by post-hoc tests. There was also an indication of sensitivity to the cultural background [F(1,182) = 7.5, P < 0.01], as there was an increase in these scores in the Romanian sample, irrespective of the diagnosis and developmental period. Emotional abuse was more frequently reported by patients than by controls [F(4,194) = 7.0, P < 0.001] and more frequently by patients with personality disorders (P < 0.01) and by schizophrenic patients (P < 0.05) than the ones with a history of alcohol-related disorders (Figure 2 ). A main effect of the developmental period [F(3,582) = 24.0, P < 0.001] was explained by an increase of the reported emotional abuse from early childhood to adolescence (P <0.05), and a decrease in adulthood (P < 0.05) were noted. Similar to the case of the neglect experiences, the Romanian sample scored also higher than the German/Swiss sample, mainly for the earliest (0–6 yr.) period, as revealed by the interaction between the development period and the cultural background [F(3,582) = 5.4, P < 0.01]. Figure 2 Mean emotional abuse score across developmental periods among all groups. Emotional abuse was more frequently reported by patients than by controls [F(4,194) = 7.0, p < 0.001]. A main effect of the developmental period [F(3,582) = 24.0, P < 0.001] was explained by an increase of the reported emotional abuse from early childhood to adolescence, and a decrease in adulthood were noted. Error bars stand for standard error of the mean. Asterisks indicate significant main effects of the psychiatric status. Irrespective of the developmental period, physical abuse was more often reported by patients with personality disorders [F(4,202) = 5.7, P < 0.001] (Figure 3 ). Post-hoc comparisons also revealed higher rates of physical abuse reports among patients with alcohol-related disorders (P < 0.01) and with schizophrenic disorders (P < 0.05) than among controls. The reports of physical abuse generally increased across developmental periods, with adulthood as the most susceptible period of such reports [F(3, 606) = 35.1, P < 0.001]. The interaction of the developmental period with the psychiatric status showed that this increase in physical abuse reports across developmental periods was mainly to be remarked in patients [F(12,606) = 3.0, P < 0.001]. Figure 3 Mean physical abuse score across developmental periods among all groups. Physical abuse was more often reported by patients with personality disorders [F(4,202) = 5.7, P < 0.001]. The reports of physical abuse generally increased across developmental periods, with adulthood as the most susceptible period of such reports [F(3, 606) = 35.1, P < 0.001]. Error bars stand for standard error of the mean. Asterisks indicate significant main effects of the psychiatric status. Sexual abuse (Figure 4 ) was primarily reported by patients, and not by controls [F(4,205) = 5.2, P < 0.001], and particularly by patients with personality disorders (P < 0.001). Higher rates of sexual abuse were reported among patients with alcohol-related disorders (P < 0.01), with schizophrenic disorders (P < 0.05), and with affective disorders (P < 0.05) than among controls as shown by post-hoc tests. If sexual abuse was experienced, it occurred particularly in later developmental periods [F(3,615) = 20.4, P < 0.001]. Sexual abuse was more often experienced by female patients [F(1,205) = 10.0, P < 0.001] after puberty [F(3, 615) = 10.0, P < 0.001], as revealed by the interaction between the developmental period and gender. We also found a 3-way interaction between the developmental period, psychiatric status and cultural background [F(12,615) = 2.6, P < 0.01). The interaction between the developmental period and cultural background revealed that Romanian but not German/Swiss schizophrenics reported more frequently sexual abuse particularly in adulthood [F(3,615) = 5.0, P < 0.01]. Figure 4 Mean sexual abuse score across developmental periods among all groups. Sexual abuse was primarily reported by patients, and not by controls, and particularly by patients with personality disorders [F(4,205) = 5.2, P < 0.001]. If sexual abuse was experienced, it occurred particularly in later developmental periods [F(3,615) = 20.4, P < 0.001]. Error bars stand for standard error of the mean. Asterisks indicate significant main effects of the psychiatric status. Trauma witnessing was reported most often by patients with personality disorders as compared to all other groups [F(4,209) = 8.0, P < 0.001]. Post-hoc tests showed that patients with affective disorders (P < 0.01) and with alcohol-related disorders (P < 0.05) also reported more experiences of trauma witnessing than the controls. Irrespective of the diagnosis, Romanian patients, but not controls, reported higher mean scores on this variable and more often than their German/Swiss counterparts [F(1,209) = 17.0, P < 0.001]. The interaction between the cultural background and developmental period indicated in the Romanian sample an increase of trauma witnessing in adulthood [F(3,627) = 8.0, P < 0.001]. Other traumas Similar to the pattern of trauma witnessing, all patients reported a greater number of traumatic events than the control group [F(4,211) = 8.0, P < 0.001]: alcohol-related disorders (P < 0.001), personality disorders (P < 0.001), schizophrenic disorders (P < 0.01), and affective disorders (P < 0.01), as explained by post-hocs. An increase in the amount of other traumas reports across the developmental periods with highest values in adulthood for all patient groups [F(12, 633) = 7.0, P < 0.001] was also revealed. Alcohol and drug abuse As previously expected, patients treated for alcohol-related disorders reported more alcohol and drug abuse than all the other groups [F(4,213) = 12.3, P < 0.001]. The post-hoc tests showed that the other patient groups also reported more alcohol and drug abuse when compared to the control group: affective disorders (P < 0.001), personality disorders, (P < 0.001) and schizophrenic disorders (P < 0.05). As also anticipated, abuse increased across developmental periods until adulthood [F(3,639) = 110.0, P < 0.001], particularly in patients with alcohol-related disorders, as revealed by the interaction between the developmental period and psychiatric status [F(12,639) = 14.3, P < 0.001]. Irrespective of the diagnosis, Romanian patients, but not controls, showed higher mean scores on reporting alcohol and drug abuse than their German/Swiss counterparts [F(4,213) = 3.4, P < 0.01], as shown by the interaction between the psychiatric status and cultural background. Compared to the German group, alcohol and drug abuse in the Romanian sample was higher, particularly in adulthood [F(3,639) = 9.8, P < 0.001]. Interrelations A principal components factor analysis was performed to explore interrelationships among TAQ subscales. The results of this analysis indicated that the most appropriate solution involved five factors that jointly accounted for 56.2% of the total variance in the dataset. Table 4 summarizes the results of the varimax rotation for the five-factor solution. The first factor showed high positive loadings on physical abuse , sexual abuse , trauma witnessing , and other traumas , obviously explains the traumatic experiences. The second factor showed high positive loadings on competence and safety , apparently accounting for variance attributed to positive experiences. The third factor showed high positive loadings on the first year of illness with alcohol and drug abuse . The fourth factor, consisting of separation , evidently explains disruptions of attachment. The fifth factor, which included secrets and emotional abuse , appeared to account for family chaos. Thus, the structure of the study instrument was well reproduced for the present sample, which included different psychiatric diagnoses and different cultural backgrounds. Table 4 Varimax solution with five factors for negative and positive childhood experiences across developmental periods in psychiatric patients with different diagnoses 1 Factor Loading 2 Variables FACTOR 1: Traumatic Experiences 3 FACTOR 2: Positive Experiences 4 FACTOR 3: Vulnerability to Alcohol Abuse 5 FACTOR 4: Disruptions of Attachment 6 FACTOR 5: Family Chaos 7 Competence -0.0 0.8 0.2 0.1 0.1 Safety 0.2 0.8 -0.0 0.0 -0.2 Neglect 0.2 -0.3 0.1 -0.0 0.3 Separation 0.4 0.2 -0.0 0.8 0.1 Secrets -0.0 -0.1 -0.0 0.1 0.7 Emotional Abuse 0.2 0.0 0.6 -0.0 0.6 Physical Abuse 0.7 -0.0 -0.0 0.0 0.0 Sexual Abuse 0.4 0.0 0.0 -0.6 0.1 Witnessing 0.6 0.0 -0.0 0.0 0.2 Other Traumas 0.6 0.2 0.1 0.2 0.1 Alcohol & Drug Abuse 0.4 -0.2 0.5 0.0 -0.3 First Year of Illness -0.1 0.1 0.8 -0.1 0.2 1 Total percent of variance = 56.2% 2 Shaded areas indicate specific domains of the TAQ contributing to each factor 3 Eigenvalue = 4.88; percent of variance = 40.7% 4 Eigenvalue = 1.476; percent of variance = 12.3% 5 Eigenvalue = 1.013; percent of variance = 8.4% 6 Eigenvalue = 0.835; percent of variance = 7.0% 7 Eigenvalue = 0.815; percent of variance = 6.8% Discussion The study aimed at exploring whether psychiatric diagnoses, e.g. alcohol-related disorders, schizophrenic disorders, affective disorders, and personality disorders are related to retrospectively reported positive and negative life events across developmental periods, and if so, whether special developmental periods are characterized by more negative experiences than others. Our findings demonstrate a strong association between reports of traumatic events and certain psychiatric disorders. In other studies, negative experiences were reported by individuals with diagnoses such as affective disorders [ 18 , 41 ] and schizophrenic disorders [ 42 , 43 ], but these experiences were less common and cumulatively less severe. Negative experiences were particularly prominent in patients with personality disorders [ 24 , 25 , 44 ] and in patients with substance-related disorders [ 26 , 45 , 46 ]. Negative experiences were reported more often in late childhood and adolescence than in early childhood and adulthood. Previous studies indicated that the earlier onset of abuse was associated with greater severity and longer duration of mental problems [ 2 , 10 , 45 , 47 ]. If the present findings are consistent with some prior studies [ 5 , 9 , 16 ] in that they indicate a relationship between physical and sexual abuse and psychiatric disorders, they do not support the view expressed by Van der Kolk et al. about early abuse at an early stage of development [ 48 ]. The current investigation showed that many psychiatric patients had terrible histories of childhood physical and/or sexual abuse. This finding was marginally significant for the childhood sexual abuse histories and must therefore be interpreted with caution. However, one should keep in mind that self-report questionnaires depend heavily upon conscious retrieval capacity for autobiographic events. It is conceivable that in the current group of patients, early abuse events were less remembered as compared to abuse events experienced later in childhood. An advantage of the TAQ used in the present study is the assessment of negative experiences during both childhood and adulthood, while most of the other studies have so far focused primarily on the impact of childhood abuse, except Cascardi et al. [ 49 ] and Goodman et al. [ 32 ]. Another advantage of the TAQ is that it addresses the issue of neglect [ 50 ]. Given the sample of patients with different psychiatric diagnoses, this replicates Van der Kolk's et al. notion that patients who experience neglect early in their lives develop serious problems with affect regulation [ 51 ]. The present data add to the evidence, suggesting that neglect, emotional and physical abuse are experienced by many psychiatric patients [ 52 , 53 ]. This implies that although childhood traumas may contribute to a mental disorder in adulthood, the lack of secure attachments maintains it. Although emotional neglect has received less attention, perceived emotional rejection by parents has been associated with alcohol abuse [ 54 ] and delinquency [ 55 ] during adolescence and adulthood. Early emotional injuries could possibly trigger vulnerability to noxious experiences. Furthermore, experiences of parental loss or separation were prominent in adulthood especially for the patients with alcohol-related disorders and with affective disorders. The high incidence of such negative experiences during this period in the patients with alcohol-related disorders could be, at the same time, a direct consequence of the behavioral deviance of these individuals and contribute to the maintenance of alcohol abuse. Limitations of the study The present data has to be considered in the light of several possible limitations. First, the information obtained by self-report and without external evidence could be less reliable and valid, especially if we take into account the sensitive nature of this research. Herman and Schatzow, however, provide empirical support for the validity of abused patients' self-reports as well [ 56 ]. They found that when corroborating evidence is sought, the majority of women are able to obtain confirmation of abuse. No independent corroborating evidence was sought for any self-reported case of childhood negative experiences. Therefore, the validity of abuse reports cannot be assured. Recall may be biased, but there is no evidence that psychiatric patients are more likely to lie about or imagine child abuse [ 57 , 58 ]. There is some evidence, however, that "patients are biased to underreport abuse histories rather than to over report them" [ 59 ]. There were some "don't know" subject answers regarding abuse/neglect experiences, most of them in the early childhood. Most probably, the patients had difficulties recalling experiences that occurred at a very young age rather than trying to evade giving a positive answer. Furthermore, another methodological limitation in this study is that measuring neglect/emotional abuse in early childhood is particularly difficult as the awareness of it necessitates the development of a degree of differentiation and autonomy, which is seldom the case with psychiatric patients. Both individual interviews and self-report questionnaire methods present higher figures than chart reviews do, indicating that patients usually do not spontaneously offer such information to their therapists. When offered, the information is not reliably documented [ 57 ]. However, the data from our ongoing study in patients with personality disorders suggest that reports on events in general and physical abuse events in particular are highly stable across two measurement periods of time separated by 24 months. We also note that our sample consisted of psychiatric inpatients, and thus may not be representative of the broader population of patients with these disorders. The clinical validity of the TAQ has also been criticized [ 60 ]. The questionnaire is meant to be an applied clinically oriented measure, which has not yet been proved to be a psychometrically sound research instrument. This issue should be addressed in future studies using both convergent and divergent instruments. Romanian patients diagnosed with schizophrenic disorders differed significantly, with respect to the number of negative events, as compared to their German counterparts. One factor accounting for this difference might be the stressful environment during the former Ceausescu regime in Romania. During this 25-year period violations of human rights, terror, and corruption prevailed [ 61 , 62 ]. This result may also be due to the different diagnostic procedures used by Romanian and German/Swiss clinicians. Reports of higher rates of psychotic-like or specifically schizophrenic symptoms do not necessarily imply a diagnosis of schizophrenia. Once abuse is identified, a change of diagnosis, from schizophrenia to PTSD, is often made, with significant advantages for the individuals [ 30 ]. Conclusions The present study demonstrates an association between negative life events in childhood and psychiatric diagnoses in adult life, which is in line with a number of other studies [ 6 , 63 ]. Unlike previous reports [ 3 , 64 ], we found that psychiatric patients were more likely to report more negative life events during late childhood and adolescence rather than during early childhood and adulthood. These conclusions corroborate with one of the central hypothesis of life-span psychotraumatology, that is, adolescence is an extremely critical phase in the development of later psychopathology [ 65 , 66 ]. However, in line with findings offered by earlier controlled studies [ 37 , 38 ], psychiatric patients were more likely to report higher rates of negative life events during childhood than controls did. Although one cannot assume a direct causal relationship between childhood abuse and adult psychopathology from the present data, the present study provides further preliminary and explorative evidence for the high load of negative life events in psychiatric patients. An advantage of this study is the examination of the abuse histories across a range of four psychiatric diagnoses within a controlled comparison design. Our findings are important and clinically highly relevant for further etiological research of causal and maintenance factors of psychiatric symptomatic, as well as for the research on the treatment of these conditions. The special value of the study lies in its cross-national comparison from a clinical psychological point of view including a highly underresearched country like Romania. More attention should be paid to the sad situation of the patients in Romania who are often under inadequate pharmacological and insufficient psychotherapeutic treatment, as well as under inappropriate hospitalization conditions. Further research should concentrate on the epidemiology and developmental psychopathology of psychiatric populations in other countries than the usually researched ones. Generally, reports of traumatic experiences during the whole lifespan should be more carefully considered in the clinical diagnosis process and in the development of treatment programs for the psychiatric patients. Competing interests The author(s) declare that they have no competing interests. Authors' contributions ES carried out the study in Germany, performed the statistical analysis and drafted the manuscript. DB carried out the study in Romania and drafted the manuscript. BR conceived of the study and drafted the manuscript. FN participated in the design of the study. MS participated in the design of the study. KS participated in the coordination of the study. KH participated in the coordination of the study. TE conceived of the study and drafted the manuscript. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539251.xml

549205

Quantitative evaluation and modeling of two-dimensional neovascular network complexity: the surface fractal dimension

Background Modeling the complex development and growth of tumor angiogenesis using mathematics and biological data is a burgeoning area of cancer research. Architectural complexity is the main feature of every anatomical system, including organs, tissues, cells and sub-cellular entities. The vascular system is a complex network whose geometrical characteristics cannot be properly defined using the principles of Euclidean geometry, which is only capable of interpreting regular and smooth objects that are almost impossible to find in Nature. However, fractal geometry is a more powerful means of quantifying the spatial complexity of real objects. Methods This paper introduces the surface fractal dimension (D s ) as a numerical index of the two-dimensional (2-D) geometrical complexity of tumor vascular networks, and their behavior during computer-simulated changes in vessel density and distribution. Results We show that D s significantly depends on the number of vessels and their pattern of distribution. This demonstrates that the quantitative evaluation of the 2-D geometrical complexity of tumor vascular systems can be useful not only to measure its complex architecture, but also to model its development and growth. Conclusions Studying the fractal properties of neovascularity induces reflections upon the real significance of the complex form of branched anatomical structures, in an attempt to define more appropriate methods of describing them quantitatively. This knowledge can be used to predict the aggressiveness of malignant tumors and design compounds that can halt the process of angiogenesis and influence tumor growth.

Background The term "angiogenesis" defines the fundamental process of the development and growth of new blood vessels from the pre-existing vasculature, and is essential for reproduction, development and wound repair [ 1 ]. Under these conditions, it is highly regulated: i.e. "turned on" for brief periods of time (days) and then completely inhibited. The cyclic nature of the microvascular bed in the corpus luteum provides a unique experimental model for examining the discrete physiological steps of angiogenesis in the life cycle of endothelial cells which, together with pericytes (supportive vascular smooth muscle cells), carry all of the genetic information necessary to form tubes , branches and entire capillary networks . However, many human diseases (including solid tumors) are driven by persistently up-regulated angiogenesis [ 1 ]. In some non-malignant processes, such as pyogenic granuloma or keloid formation [ 2 ], angiogenesis is prolonged but still self-limited ; however, this is not true of tumor angiogenesis which, once begun, continues indefinitely until the entire tumor is eradicated or the host dies. Without blood vessels, tumors cannot grow beyond a critical size (1–2 mm) or metastasize to another organ. Angiogenesis is one of the most complex dynamic processes in biology, and is highly regulated by a balance of pro- and anti-angiogenic molecules. It is now widely accepted that the "angiogenic switch" is "off" when the effects of pro-angiogenic molecules is balanced by that of anti-angiogenic molecules, and "on" when the net balance is tipped in favor of angiogenesis [ 1 , 3 ]. Pro- and anti-angiogenic molecules can be secreted from cancer cells, endothelial cells, stromal cells, blood, and the extra-cellular matrix [ 4 , 5 ], the relative contributions of which are likely to change with tumor type and site, as well as with tumor growth, regression and relapse [ 1 ]. Although considerable advances have been made in our molecular and cellular knowledge of the promotion , mediation and inhibition of angiogenesis, very little is known about its underlying complex dynamics . Vasculature and more generally tubular organs develop in a wide variety of ways involving many cell processes [ 6 - 8 ]. In mathematical terms, angiogenesis is a non-linear dynamic system that is discontinuous in space and time , but advances through qualitatively different states . The word state defines the configuration pattern of the system at any given moment, and a dynamic system can be represented as a set of different states and a number of transitions from one state to another over a certain time interval [ 9 , 10 ]. At least seven critical steps have so far been identified in the sequence of angiogenic events on the basis of sprout formation: a) endothelial cells are activated by an angiogenic stimulus; b) the endothelial cells secrete proteases to degrade the basement membrane and extra-cellular matrix; c) a capillary sprout is formed as a result of directed endothelial cell migration, d) grows by means of cell mitoses and migration, and e) forms a lumen and a new basement membrane; f) two sprouts come together to form a capillary loop; and g) second-generation capillary sprouts begin to form [ 1 , 11 , 12 ] (Fig. 1 ). The progression of these states generates a complex ramified structure that irregularly fills the surrounding environment (Fig. 2 ). The main feature of the newly generated vasculature is the structural diversity of the vessel sizes, shapes and connecting patterns. Tumor vessels are structurally and functionally abnormal [ 1 , 3 ]: unlike normal vessels, they are highly disorganized, tortuous and dilated, and have uneven diameters, and excessive branching and shunts. This may be mainly due to the heterogeneous distribution of angiogenic regulators, such as vascular-endothelial growth factor (VEGF), basic fibroblastic growth factor (bFGF) and angiopoietin [ 5 , 13 ], leading to chaotic tumor blood flow, and hypoxic and acidic tumoral regions [ 5 , 14 - 16 ]. Moreover, although it is commonly believed that the endothelial cells making-up tumor vessels are genetically stable, diploid cells (and thus different from genetically unstable neoplastic cells), tumor vasculature seems to be much more unpredictable [ 17 ]. These conditions all reduce the effectiveness of treatments, modulate the production of pro- and anti-angiogenic molecules, and select a subset of more aggressive cancer cells with higher metastatic potential [ 1 ]. A large number of clinical trials of anti-angiogenic therapies are being conducted throughout the world, but investigators are still concerned about how to achieve the maximum benefit from them and how to monitor patient response. There are currently no markers of the net angiogenic activity of a tumor that can help investigators to design specific anti-angiogenic treatment strategies [ 5 , 18 ], but it is reasonable to resume that the quantification of various aspects of tumor vasculature may provide an indication of angiogenic activity. One often-quantified element of tumor vasculature is microvessel density (MVD), which is used to allow a histological assessment of tumor angiogenesis. The results of studies carried out over the last decade have suggested the value of using tumor MVD as a prognostic index in a wide variety of solid cancers, and it has also recently been assumed that MVD may reveal the degree of angiogenic activity in a tumor. On the basis of these assumptions, the quantification of MVD is thought to be a surrogate marker of the efficacy of anti-angiogenic agents as well as a means of assessing which patients are good candidates for anti-angiogenic therapy. However, MVD has a number of substantial limitations, mainly due to the complex biology characterizing tumor vasculature [ 17 ], and the highly irregular geometry that the vascular system assumes in real space , which cannot be measured using the principles of Euclidean geometry because it is only capable of interpreting regular and smooth objects that are almost impossible to find in Nature. However, quantitative descriptors of its geometrical complexity can be usefully abstracted from the fractal geometry introduced by Benoit Mandelbrot in 1975 [ 20 , 21 ]. We here discuss the surface fractal dimension (D S ) as a quantitative index of the 2-D geometrical complexity of vascular networks and their behavior during computer-simulated changes in vessel density and distribution . Geometrical properties of a vascular network The human vascular system can be geometrically depicted as a complex fractal network of vessels that irregularly branch with a systematic reduction in their length and diameter [ 19 ]. Fractal objects are mainly characterized by four properties: a) the irregularity of their shape; b) the self-similarity of their structure; c) their non-integer or fractal dimension ; and d) scaling , which means that the measured properties depends on the scale at which they are measured [ 22 ]. One particular feature of fractal objects is that the schemas defining them are continuously repeated at decreasing orders of magnitude, and so the form of their component parts is similar to that of the whole [ 20 , 21 ]: this property is called self-similarity . Unlike geometrical self-similarity , which only concerns mathematical fractal objects in which every smaller piece is an exact duplicate of the whole (e.g. Koch's snowflake curve, Sierpinski's triangle and Menger's sponge), statistical self-similarity concerns all complex anatomical systems, including tumor vasculature. The smaller pieces constituting anatomical entities are rarely identical copies of the whole, but more frequently "similar" to it and, in such systems, the statistical properties of the pieces are proportional to the statistical properties of the whole [ 23 ]. Dimension is a numerical attribute of an object that does not depend on its process of generation, and has been defined in two ways. The first is the topological or Euclidean dimension (Fig. 3 ), which assigns an integer to every point or set of points in Euclidean space ( E ): 0 to a point (defined as that which has no part); 1 to a straight line (defined as a length without thickness), 2 to a plane surface (defined as having length and thickness, but no depth); and 3 to three-dimensional figures (a volume defined by length, thickness and depth). The second was introduced by the mathematicians Felix Hausdorff and Abram S. Besicovitch, who attributed a real number to every natural object in E lying between the topological dimensions 0 and 3 (Fig. 3 ). Benoit Mandelbrot uses the symbol D γ to indicates the topological dimension, and the symbol D to indicate that of Hausdorff-Besicovitch (also called the fractal dimension ). The D γ and D of all Euclidean figures are coincident ( D γ = D ), but this is not true of fractal objects in which D is always > D γ . As no anatomical entity corresponds to a regular Euclidean figure, their dimension is always expressed by a non-integer number falling between two integer topological dimensions. In our case (Fig. 2 ), the vascular network has a dimension lying between 2 (plane surface) and 3 (volume), and any two-dimensional section of a vascular system (as in the case of a histological section) has a dimension lying between 0 (the dimension of a single isolated point) and 2 when the sectioned vessels entirely fill a plane surface (Figs. 3 and 4 ). Anatomical structures are also hierarchical systems that operate at different spatial and temporal scales , and different patterns can change, appear or disappear depending on the scale of magnification [ 22 ]. A fundamental characteristic is that the process operating at a given scale cannot be important at higher or lower scales [ 23 ]. The irregularity and self-similarity underlying scale changes are the main attributes of the architectural complexity of both normal and pathological biological entities [ 22 - 26 ]. In other words, the shape of a self-similar object does not change when the scale of measure changes because every part of it is similar to the original object; however, the magnitude and other geometrical parameters (e.g. the outline perimeter) of an irregular object differ when inspected at increasing resolutions that reveal an increasing number of details [ 25 ]. Over the last decade, accumulating experimental evidence has shown that the fractal patterns or self-similar structures of biological tissues can only be observed within the scaling window of an experimentally established measure of length ε 1 - ε 2 (Fig. 4 ), within which experimental data sets follow a straight line with a slope (1-D) : i.e. the fractal dimension remains invariant at different magnifications [ 20 - 27 ]. Methods Computer-aided modeling of two-dimensional vascular tree complexity We have developed a computer model to simulate the geometrical complexity of a histological two-dimensional section of a tumor vascular tree that automatically generates an unlimited number of images with a changeable density of vessels irregularly distributed on a planar surface. In order to simplify the model, we considered all of the vessels as rounded, unconnected objects of equal magnitude (Fig. 5 ). As the parameters of a model must be as few as possible and it is necessary to reduce mathematical complexity [ 28 - 30 ], we included only two variables: a) the number of vessels; and b) their distribution in the surrounding environment. The vessel distribution patterns were randomly generated using different time-dependent seeds for random number function generation. One thousand images were automatically generated for each vessel density (from five to 50 vessels, with the number being increased by five in each group), and their D S were estimated using the box-counting method [ 22 ]. D S was automatically estimated using the equation: where ε is the side-length of the box, and N ( ε ) the smallest number of boxes of side ε required to completely contain the irregular object (Fig. 4 ). As the zero limit cannot be applied to biological images, D S was estimated by means of the equation: (2) D S = d where d is the slope of the graph of log [N(ε)] against log (1/ε) , in a fixed range of side-lengths ( ε 1 - ε 2 ) empirically evaluated by visualization [ 20 - 29 ]. Statistical analysis All of the data are expressed as mean values ± standard deviation, and the results were analysed using the Statistica software package (StatSoft Inc. Tulsa, USA). Unvaried analysis was performed by means of the Student t as required for parametric variables. p values of less than 0.05 were considered statistically significant. Results The computer-aided simulations showed that different D S values can be obtained for images with the same vessel density (Fig. 6 ). As the only variable in these images is the vessel distribution pattern, D S depends on the irregular arrangement of the vessels in the surrounding environment. D S also significantly increased ( p <0.05) when higher vessel densities were considered in the system (Fig. 6 ) because of the greater space filled by the vascular component (as shown in Fig. 3 ); the increased density of the vessels reduces the variability in their space-filling properties, and thus the standard deviation (Fig. 6 ). Discussion and conclusions One of the most important and distinctive characteristics of biological systems is the complexity of their shape ( geometrical or spatial complexity) and functions ( behavioral complexity). Complexity is a real quality of organized biological matter that is mainly manifested in the living world as diversity and organization . No two anatomical systems are exactly alike because of the enormous variability not only between the different members of a population, but also between the component parts of an organism. The word complexity has long been used descriptively in order to describe, for example, a large number of genes or cellular interconnections [ 33 ], but complexity can also reside in the structure of a system ( i.e. an intricate architecture or the existence of many different component parts with varying interactions) or its non-linear functions ( i.e. physiological rhythms are rarely strictly periodic but fluctuate irregularly over time) [ 34 ]. The vascular system is a complex network consisting of branched tubes of different sizes that are irregularly settled in the surrounding environment [ 6 , 7 ]. This geometrical characteristic highlights the complexity of its generating process in space and time , and greatly biases any quantitative method that tends to idealize it as a smooth and regular Euclidean object. However, both normal and tumor vasculature can more properly be considered fractal objects because of their irregular shape ( spatial conformation ), self-similar structure, non-integer dimension and dependence on the scale of observation ( scaling effect ) [ 19 , 35 - 37 ]. We here discuss the estimate of D S as a quantitative index of the 2-D spatial complexity of the vascular tree, in order to provide a closer-to-reality measure of this complex anatomical entity (Figs. 3 and 4 ). The theory underlying D S was abstracted from fractal geometry, which is also called the geometry of irregularity [ 20 , 21 ]. The concept of spatial conformation has played a fundamental role in the study of biological macromolecules in chemistry (particularly biochemistry) since the early 1950s. However, it has only been introduced in the science of morphology as theoretical morphology , which studies extant organismal forms (complex structures of interdependent and subordinate elements whose relationships and properties are largely determined by their function in the whole) as a subset of the range of theoretically possible morphologies [ 32 ]. The significance of D S also comes from the fact that, like any other complex biological system, the vascular tree cannot be correctly quantified by measuring its individual properties (i.e. micro-vessel density, MVD). D S is a parameter that depends on: a) the number of vessels; b) the spatial relationships between the vascular components; and c) the interactions between the vascular components and the surrounding environment. In other words, its estimate is "ecologically" important because it provides a quantitative index of the "habitat structure". As computer models are crucial for scientific procedures, and the modeling process itself represents the hypothetical-deductive approach in science [ 30 - 32 ], we developed a simple computer-aided model capable of generating an unlimited number of 2-D images of a simulated vascular network. The model was simplified by using a minimum amount of mathematical complexity and only two variables: the number of vessels and their pattern of distribution. A total of 10,000 images showing a different number of unconnected vessels irregularly distributed on a planar surface were automatically generated (Fig. 5 ) and, interestingly, it was found that D S increased with the number of vessels making up the system (Fig. 6 ); furthermore, its value changed when the same number of vessels were differently distributed in the surrounding environment. In other words, it is plausible that an equal number of vessels may have different space-filling properties depending on their distribution pattern. These results suggest the usefulness of this model when comparing real vasculature configurations in order to explore the morphological variability that can be produced in nature, as it is now well known that aberrant vascular architectures in tumors may affect the uniform delivery of specific drugs to all cancer cells [ 15 ]. The model also suggests that: a) D s can be an estimate of the 2-D geometrical complexity of the vascular system. As 2-D vascular complexity depends on the number of vessels and their distribution pattern, the use of MVD quantification alone to measure the angiogenic dependence of a tumor is strongly biased because the number of vessels does not reflect the number of tumor cells that can be supported by a vessel. Moreover the metabolic needs of cancer cells vary with the tissue of origin and change with tumor progression [ 18 ]. b) D S depends on the degree of vessel contiguity and continuity . These two geometrical properties determine what is called the intercapillary distance , and are not only involved in the spatial complexity of tumor vasculature, but also reflect the inviolable demand of a growing tumor for sufficient levels of nutrition and oxygen exchange. Inter-capillary distances are locally defined by the net balance between pro- and anti-angiogenic molecules in each microtissue region, as well as by non-angiogenic factors such as the oxygen and nutrient consumption rates of tumor cells. In normal tissue, vessel density fairly accurately reflects cell metabolic demands because evolutionary pressures have led to close and efficient coupling between vascular supply and metabolic needs. In tumors, the close coupling between vascular density and oxygen or nutrient consumption ( i.e. the environment) may be loosened [ 18 ], thus altering not only the number of vessels but also the whole vascular architecture [ 15 , 38 ]. c) D S falls between 0 (corresponding to the Euclidean dimension of a point) and 2 (the dimension of a plane). The more D S tends towards 2, the more the analyzed vascular configuration tends to fill a 2-D space and the greater its geometrical complexity. In conclusion, the present study indicates that the complex geometry of tumor vasculature and its well-known biological characteristics [ 18 ] mean that vascular network cannot be measured on the basis of MVD estimates alone. These findings also support the findings of various authors who have shown the uselessness of MVD as a predictor of anti-angiogenic treatment efficacy or for stratifying patients in therapeutic trials [ 14 , 39 - 41 ]. Scientific knowledge develops through the evolution of new concepts, and this process is usually driven by new methodologies that provide previously unavailable observation. The potential broad applicability of the proposed quantitative index makes it possible to explore the range of the morphological variability of vasculature that can be produced in nature, thus increasing its diagnostic importance in cancer research. Abbreviations Ds, Surface fractal dimension; 2-D, two-dimensional; VEGF, Vascular-endothelial growth factor; bFGF, basic fibroblastic growth factor; MVD, microvessel density. Competing interests The author(s) declare that they have no competing interests. Authors' contributions FG conceived, coordinated and designed the study and drafted the manuscript; CR, PC, BF, EEF, EC, MCI participated in designing the study and drafting the manuscript. All of the authors have read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549205.xml

387280

Phage Display Libraries Identify T Cells

Doctors and researchers often look for the rapid proliferation of T cell populations, key defensive players in the immune system, as a telltale sign that the body is working hard to fend off a foreign threat. Every one of these circulating white blood cells carries a T cell receptor (TCR) that binds to a specific protein, or antigen, when displayed on the surface of a cell. A match between TCR and displayed antigen results in the cell's death and the subsequent expansion of T cell clones, all programmed to recognize the original offending protein. Some TCRs bind and expand in response to pathogenic antigens, such as viral or bacterial proteins. But T cells can also react and proliferate inappropriately in response to the body's own proteins, leading to destructive autoimmune diseases such as multiple sclerosis, which is characterized by immune system attacks on nervous tissue. Self-recognizing TCRs, however, can also target and destroy tumors—though full activation of these T cells is inconsistent and poorly understood. Peptide display Identifying the particular antigen behind an exploding population of T cells is invaluable for finding the source of autoimmune diseases and studying immune responses to cancer. But it's a laborious and time-consuming process, as researchers are faced with the prospect of sifting through millions upon millions of possible matches between TCRs and their prospective antigen epitopes—the part of the antigenic molecule to which the receptor binds. Now, as they report in this issue of PLoS Biology , Frances Crawford and colleagues have developed a novel method for rapidly identifying TCR mimotopes—peptide sequences similar or identical to epitopes that also elicit the immune response—which can be used to determine the antigen of a given T cell population. Working backwards, the team started off with two different T cell clones that had been previously selected for with a known antigen—a peptide called p3K. One clone was derived from mice genetically engineered to have broadly reactive T cells; the other, a conventional clone, was much more sensitive to the precise molecular structure of p3K. Crawford and colleagues then created a “peptide library” comprising more than 30,000 baculoviruses (viruses that selectively target insect cells), each one carrying a slightly different version of the p3K gene, varied in regions of the peptide known to be important for TCR binding. These p3K genes were embedded within a major histocompatibility complex (MHC) gene—a type of cell surface protein that holds displayed antigens and is also important for proper TCR recognition. The team then unleashed their virus library onto insect cells that, once infected, began to produce the specific peptide–MHC complexes encoded on the viral DNA. The insect cells then shuttled these proteins to their surfaces, resulting in a vast array of cells that each displayed a unique variant of the p3K–MHC complex. This “display library” was then incubated with fluorescently labeled TCRs from the two different clones. By observing and isolating the insect cells that lit up, the researchers could see which of the thousands of cells displaying peptide–MHC possessed a mimotope capable of binding a TCR. Because the genetic information about the displayed complex was still stored within the virus-infected cell, the researchers could determine the full peptide sequence responsible for the identified mimotopes. Confirming the effectiveness of their method, the results of the fluorescence experiments echoed the authors' original characterizations about the two populations of T cells. The broadly reactive TCR bound to several different uniquely displayed complexes; it had 20 mimotopes. The conventional TCR, however, bound only to one peptide–MHC complex, an almost perfect match to the original p3K peptide. Though this study was based on a known antigen and epitope (which allowed verification of the method), the baculovirus display library technique described here could easily be used on T cell populations with unknown antigens. With such a tool, researchers could, for example, identify the antigens connected with tumor-fighting T cells and, through inoculation, possibly induce the production of similar T cells in cancer patients who lack them.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC387280.xml

549211

Life-threatening ventricular arrhythmia recognition by nonlinear descriptor

Background Ventricular tachycardia (VT) and ventricular fibrillation (VF) are ventricular cardiac arrhythmia that could be catastrophic and life threatening. Correct and timely detection of VT or VF can save lives. Methods In this paper, a multiscale-based non-linear descriptor, the Hurst index, is proposed to characterize the ECG episode, so that VT and VF can be recognized as different from normal sinus rhythm (NSR) in the descriptor domain. Results This newly proposed technique was tested using MIT-BIH malignant ventricular arrhythmia database. The relationship between the ECG episode length and the corresponding recognition performance was studied. The experiments demonstrated good performance of the proposed descriptor. An accuracy rate as high as 100% was obtained for VT/VF to be recognized from NSR; for VT and VF to be recognized from each other, the recognition accuracy varies from 84.24% to 100%. In addition, the results were compared favorably against those obtained using Complexity measure. Conclusions There is strong potential for using the Hurst index for malignant ventricular arrhythmia recognition in clinical applications.

Introduction If a life-threatening ventricular tachycardia (VT) or ventricular fibrillation (VF) is detected promptly, a high energy electrical shock can be delivered to the heart, in an attempt to return the heart to a normal sinus rhythm (NSR). If a normal sinus rhythm is misinterpreted as VT or VF, leading to delivering of an unnecessary shock, it can damage the heart, causing fatal consequences to the patient. Therefore, correct and prompt detection of VT or VF is of great importance. However, the detection of these life-threatening cardiac arrhythmia is difficult because the waveform and frequency distribution of these life-threatening arrhythmia changes with the prolonged duration [ 1 ]. Furthermore, practical problems such as poor contact, movement, interference, etc, can produce artifacts that mimic these rhythms [ 2 ]. Till now, many linear techniques for VT/VF detection have been developed, such as the probability density function method [ 3 ], rate and irregularity analysis [ 4 ], analysis of peaks in the short-term autocorrelation function [ 5 ], sequential hypothesis testing algorithm [ 6 , 7 ], correlation waveform analysis [ 8 ], four fast template matching algorithms [ 9 ], VF-filter method [ 2 , 10 ], spectral analysis [ 1 ], and time-frequency analysis [ 11 ]. However, these methods exhibit disadvantages, some being too difficult to implement and compute for automated external defibrillators (AED's) and implantable cardioverter defibrillators (ICD's), and some only successful in limited cases. For example, the linear techniques [ 5 , 11 ] using the features of amplitude or frequency have shown their limits, since the amplitude of ECG signal decreases as the VF duration increases, and the frequency distribution changes with prolonged VF duration. Therefore, more sophisticated signal processing techniques are needed to fully describe and characterize VT and VF and facilitate the development of new detection schemes with high correct detection rate, or equivalently, with low false-positive and false-negative performance statistics. Recent studies [ 12 , 13 ] have shown that the cardiac dynamics are complex and non-linear. Even if they could be described by a set of differential equations, they would be of high dimensionality. Normally, each heart beat is initiated by a stimulus from pacemaker cells in the SA node in the right atrium. The activation wave then spreads through the atria to the AV junction. Following activation of the AV junction, the cardiac impulse spreads to the ventricular myocardium through a specialized network, the His-Purkinje system. This branching structure of the conduction system is a self-similar tree with finely scaled details on a microscopic level. The spread of the depolarization wave is represented by the QRS complex in ECG. Spectral analysis of the waveform reveals a broadband of frequencies. To explain the inverse power-law spectrum, West has conjectured that the repetitive branches of the His-Purkinje system represent a fractal set in which each generation of the self-similar tree imposes greater detail onto the system [ 14 ]. The effect of the finely branching fractal network is to subtly decorrelate the individual pulses that superpose to form the QRS complex. The distribution in path lengths resulting from the fractal nature of the branches give rise to a distribution of decorrelation time. Some methods developed based on the theory of non-linear dynamics have been highlighted for the analysis of the signals generated from non-linear system [ 15 ]. Due to the complex and non-linear dynamical behavior of the cardiac conduction system, non-linear dynamics or non-linear mathematical models are considered to be suitable tools for the analysis of ECG signals. Non-linear techniques have been proven to be major cornerstones for understanding the ECG signals [ 13 , 16 , 17 ]. Some non-linear techniques [ 18 - 20 ] have been developed for life-threatening ventricular arrhythmia recognition. However, there are still many problems requiring solution. The computational demands for most of the existing algorithms are considerably high and a long ECG episode duration is needed. In order to strike a balance between lower computational burden and reliable recognition performance, a non-linear descriptor, the Hurst index, is proposed as a new tool in this study for recognition of the life-threatening ventricular arrhythmia. The Hurst index is defined in the multiscale domain as a feature to quantify the non-linear dynamical behavior (such as, self-similarity, roughness and irregularity) of the ECG signal for detecting the life-threatening ventricular arrhythmia. ECG episodes with VT and VF from MIT-BIH malignant arrhythmia database [ 21 ] are tested for cardiac abnormality recognition. The data also included some NSR signals to check on the validity of the algorithm. Experimental results are compared with those obtained by a typically used non-linear technique, the Complexity measure, which has been shown to perform well for life-threatening ventricular arrhythmia recognition [ 20 ]. In this paper, the complexity measure is Zheng's complexity measure without exception. Detailed description of Zheng's complexity measure technique can be find in [ 20 ]. The present paper is organized as follows. Mathematical background on the proposed non-linear descriptor is given in Section. Methodology for the recognition of ventricular arrhythmia is described in Section. Section covers the experimental results and discussions. Lastly, a conclusion of the proposed study is given in Section. Multiscale-based non-linear descriptor Multiscale analysis is a useful framework for many signal processing tasks. Wavelet transform is a good tool for multiscale analysis, which allows the expansion of a signal from the time domain into the time-frequency domain. In this paper, the Hurst index, defined in multiscale space, is proposed for the characterization of ECG episodes. The Hurst index, H , is a single scalar parameter describing the fractal Brownian motion (fBm) model, which is a useful model for nonstationary stochastic self-similar processes with long term dependencies over wide ranges of frequencies [ 22 ]. fBm is an extension of the ordinary Brownian motion, and is a zero-mean Gaussian nonstationary stochastic process B H ( t ), t ∈ ℝ, 0 < H < 1, [ 23 ]. Self-similarity is inherent to the fBm structure. The fractal dimension D is a commonly used parameter for measuring self-similarity. The relationship between the fractal dimension, D , and the Hurst index H is: D = S - H , where S is the topology dimension. For a one-dimensional signal, S = 2; for a two-dimensional image, S = 3 [ 24 ]. The fBm model has following features: • It is non-stationary, which necessitates some time-dependent analysis. E ( B H ( t ) B H ( s )) = σ 2 /2(| t | 2 H + | s | 2 H - | t - s | 2 H )     (1) where E ( · ) represents the expectation operator, σ is the standard deviation, t is a time variable, s is a time lag variable. Based on Equation (1), the variance of fBm, is computed as var ( B H ( t )) = σ 2 | t | 2 H . • It is self-similar, which necessitates some scale-dependent analysis. { B H ( at )} ≜ a H B H ( t ), a ∈ ℝ + (2) where ℝ + is the set of positive real numbers. ≜ means equality in distribution, which means that the fBm has stationary increments, and the probability properties of the process B H ( t + s ) - B H ( t ) only depend on the lag variable s . The scalar index H of fBm is related to the complexity and roughness of fBm samples. Consider a discrete orthogonal wavelet decomposition of a given fBm, B H ( t ). For any given resolution 2 J , the wavelet mean-square representation of fBm is: Computing the corresponding wavelet coefficients amounts to evaluating the following approximate coefficients a j [ n ] and detail coefficients d j [ n ]: where φ ( t ) is the corresponding smooth function of wavelet ψ ( t ). Flandrin et al. in [ 22 ] have deduced the following theorem: When normalized according to Wavelet coefficients of fBm give rise to: where V ψ ( H ) is constant, which depends on both the chosen wavelet and the fBm index H . It follows the power-law behavior of the wavelet coefficients' variance: log 2 ( var ( d j [ n ])) = (2 H + 1) j + constant (9) Therefore, the fBm index H (and hence the associated fractal dimension D = 2 - H ) can be easily obtained from the slope of this variance plotted as a function of scale in a log-log plot. Life-threatening ventricular arrhythmia recognition by Hurst index For each testing ECG episode, the following steps are performed: • Perform wavelet decomposition and computation of its detail coefficients at different scales. • Compute the Hurst index H according to Equation (9). • Detect the life-threatening ventricular arrhythmia in the feature space of H . In this study, the wavelet used is a quadratic spline wavelet with compact support and one vanishing moment. It is a first derivative of a smooth function [ 25 ], whose discrete Fourier transform is: The low-pass and high-pass filters L ( ω ) and G ( ω ) are respectively: The dyadic wavelet transform (WT) of a digital signal f ( n ) can be calculated with Mallat's algorithm [ 26 ] as follows: where is a smoothing operator. is the wavelet transform of digital signal f ( n ). l k | k ∈ Z and g k | k ∈ Z are coefficients of a low-pass filter L ( ω ) and a high-pass filter G ( ω ), respectively, and, L ( ω ) = Σ k ∈ Z l k e - ikω , G ( ω ) = Σ k ∈ Zgk e - ikω . Based on the frequency analysis of the ECG characteristic waves [ 27 ], scale 2 j ( j = 1 to 4) are selected. For each experimental episode, its wavelet transform coefficient sets d 1 , d 2 , d 3 and d 4 corresponding to different scales 2 1 , 2 2 , 2 3 , 2 4 are computed. The Hurst index H is then computed according to Equation (9). Smaller Hurst index corresponds to larger fractal dimension and more irregular signal. Comparative Experimental Results and Discussions Description of the test data The database used in this study is the MIT-BIH malignant ventricular arrhythmia database [ 21 ] with a sample frequency of 250 Hz . Typical waveforms of VT and VF as well as NSR are shown in Figure 1 to 3 . Selected ECG episodes with different lengths are tested for evaluating the performance of the life-threatening ventricular arrhythmia recognition using the Hurst index. Each ECG episode is characterized by the Hurst index H , computed by Equation (9). The statistical distribution of the Hurst indexes for characterizing different types of episodes is studied so that VT and VF can be recognized in the feature domain of the Hurst index. Recognition performance is measured by Sensitivity ( SE ), Specificity ( SP ) and Accuracy ( ACR ). They are defined as: Sensitivity = ; Specificity = ; Accuracy = . Where TP is true positive, the abnormal case being correctly recognized as abnormal one; FN is false negative, the abnormal case being wrongly recognized as normal one; TN is true negative, the normal case being correctly recognized as normal one; and FP is false positive, the normal case being wrongly recognized as abnormal one. Lastly, results are compared with that of Complexity measure technique. Figure 1 Typical life-threatening ECG waveform of NSR Figure 2 Typical life-threatening ECG waveform of VT Figure 3 Typical life-threatening ECG waveform of VF In this study, about 5076 ECG episodes are tested for performance evaluation of life-threatening ventricular arrhythmia recognition using the proposed Hurst index. Among them, 2588 cases are NSR episodes, 1390 cases are VT episodes, and 1098 are VF episodes. In order to explore the effect of the time series lengths on the recognition performance using the proposed Hurst index, analyzing was conducted using different lengths of ECG episodes from 1 sec to 5.5 sec with a difference of 0.5 sec . For each length, the whole dataset was randomly divided into two equal parts for training and testing, respectively. From a clinical point of view, it is essential to recognize and diagnose malignant ventricular arrhythmia as soon as possible. This calls for detection with as short a length of the time series as possible. The statistical results, viz, the means and standard deviations for characterizing NSR, VT and VF episodes using the Hurst index are given in Table 1 . As a comparison, the results by the complexity measure technique, are given in Table 2 . Graphical descriptions of the results listed in Tables 1 and 2 are shown in Figure 4 and 5 respectively. Table 1 Statistical results of Hurst index for episode characterization Episode Length Hurst index NSR VT VF Mean SD Mean SD Mean SD 1 sec 0.6099 0.0981 0.8117 0.0775 0.8567 0.0579 1.5 sec 0.6206 0.0805 0.8269 0.0671 0.8597 0.0501 2 sec 0.6317 0.0619 0.8373 0.0558 0.8618 0.0438 2.5 sec 0.6349 0.0549 0.8398 0.0509 0.8682 0.0419 3 sec 0.6389 0.0458 0.8445 0.0409 0.8766 0.0399 3.5 sec 0.6389 0.0458 0.8445 0.0409 0.8766 0.0399 4 sec 0.6395 0.0436 0.8452 0.0403 0.8794 0.0395 4.5 sec 0.6398 0.04 0.8455 0.0397 0.8797 0.0392 5 sec 0.6399 0.035 0.8458 0.0391 0.8799 0.0387 5.5 sec 0.6399 0.035 0.8458 0.0388 0.8799 0.0386 Table 2 Statistical results of Hurst index for episode characterization Episode Length Complexity measure NSR VT VF Mean SD Mean SD Mean SD 1 sec 0.1674 0.0433 0.2775 0.0428 0.2798 0.0498 1.5 sec 0.1476 0.0403 0.2562 0.0428 0.2601 0.0498 2 sec 0.1319 0.037 0.2413 0.0335 0.2454 0.0432 2.5 sec 0.1245 0.0366 0.2311 0.0335 0.239 0.0432 3 sec 0.1192 0.0363 0.2229 0.0349 0.2351 0.037 3.5 sec 0.1129 0.0348 0.2168 0.0349 0.2298 0.037 4 sec 0.1095 0.0332 0.2149 0.0342 0.2242 0.0343 4.5 sec 0.1071 0.0321 0.2136 0.0342 0.2205 0.0343 5 sec 0.1056 0.0315 0.2129 0.0342 0.2187 0.0341 5.5 sec 0.1056 0.0313 0.2129 0.0339 0.2187 0.0341 Figure 4 The mean and standard deviation values for characterizing NSR, VT and VF episodes using the Hurst index Figure 5 The mean and standard deviation values for characterizing NSR, VT and VF episodes using the Complexity measure From the results shown in Figure 4 and 5 , the following observation can be made. • As the episode length increases, the mean of Hurst index for every type of rhythm basically increases and tends to approach a relatively stable value, while the standard deviation decreases gradually. • For a particular episode length, from NSR to VT then to VF, the corresponding Hurst index increases gradually. The increase from NSR to VT is more than the increase from VT to VF. • As the episode length increases, the mean of Complexity measure for every type of rhythm basically decreases and tends to approach a relatively stable value, while the standard deviation decreases gradually. • For a particular episode length, from NSR to VT then to VF, both the Hurst index and the Complexity measure increase gradually, in which, the increase from NSR to VF is far more than the increase from VT to VF. • The mean values of Hurst index vary slower than those of Complexity measure as the episode length increases from 1 sec to 5.5 sec . It is concluded that the Hurst index is more stable than the Complexity measure with respect to episode lengths. Using the Hurst index for VT or VF recognition from NSR with different episode lengths, there is no false detection, meaning that the VT/VF can be totally correctly recognized from NSR without exception. For the Complexity measure, when the length of ECG episode is longer than 1 sec , it has as good performance as the Hurst index; when the length of the ECG episode is 1 sec , there is 6 false negatives and 27 false positives; when the length of the ECG episode is 1.5 sec , there is 1 false negatives and 5 false positives. The statistical values of SE , SP and ACR for VT/VF recognition from NSR using the Hurst index are all 100%. Hence, the Hurst index can be used to detect VT and VT earlier. As for VF differentiation from VT, the statistical values of SE , SP and ACR for different episode lengths using the Hurst index and the Complexity measure, are shown in Table 3 . The computational time of the Hurst index and the Complexity measure for different ECG episode length are presented in Table 4 . From Table 3 , the following conclusions can be obtained: Table 3 Statistical values of SE , SP and ACR for VF differentiation from VT Episode Length Hurst index Complexity measure SE SP ACR SE SP ACR 1 sec 0.8351 0.8482 0.8424 0.8242 0.8302 0.8275 1.5 sec 0.8780 0.8698 0.8734 0.8689 0.8597 0.8637 2 sec 0.9080 0.8834 0.8942 0.9007 0.8798 0.8890 2.5 sec 0.9408 0.9158 0.9268 0.9381 0.9194 0.9277 3 sec 0.9608 0.9439 0.9513 0.9654 0.9489 0.9562 3.5 sec 0.9754 0.9669 0.9707 0.9818 0.9734 0.9771 4 sec 0.9854 0.9849 0.9851 0.9918 0.9885 0.9899 4.5 sec 0.9936 0.9914 0.9924 1 0.9986 0.9992 5 sec 1 0.9978 0.9988 1 1 1 5.5 sec 1 1 1 1 1 1 Table 4 Computation time comparison in seconds Length of episode Hurst index Complexity measure Length of episode Hurst index Complexity measure 1 sec 0.0546 0.0654 1.5 sec 0.0697 0.0824 2 sec 0.0794 0.1143 2.5 sec 0.0933 0.1538 3 sec 0.1168 0.2176 3.5 sec 0.1401 0.2991 4 sec 0.1885 0.4003 4.5 sec 0.2407 0.609 5 sec 0.2803 0.6833 5.5 sec 0.3122 0.7792 • The performance on differentiating VT and VF is worse than the performance of VT/VF recognition from NSR, for both the Hurst index and the Complexity measure. • The recognition performance by either descriptors improves as the length of ECG episode increases. • When the length of ECG episode is less than or equal to 2 sec , the recognition performance for the Hurst index is better. When the length of ECG episode is longer than 2 sec and less than 5 sec , the recognition performance for the Complexity measure is better. When the length of ECG episode is longer than 5 sec , VT and VF can be 100% differentiated with either descriptor, the recognition performance for both descriptors are same. According to Table 4 , the computational time for the Hurst index is less than that for the Complexity measure. These two algorithms are programmed using MATLAB 5.3 running on a SUN SPARC-333 MHz workstation. The computational burden for the Hurst index is O ( N log 2 N ), while the computational burden for the complexity is O ( N 2 ), where N is the length of ECG episode. It is noted that with more powerful computer programming in C, the computational speed will be further improved. Time is an important factor for saving lives in clinical situations, therefore, algorithm with less computational burden is obviously preferred. In addition, using short ECG episode length is preferred for earlier detection of arrhythmia (such as VT/VF). Based on the experimental results, it is observed that the Hurst index has a better potential for clinical adaptation than the Complexity measure. Conclusions In this paper, a new technique based on multiscale analysis and non-linear dynamics was presented for VT and VF recognition. Hurst index defined across multiscale was proposed for characterizing ECG episode so that life-threatening arrhythmia can be recognized. Furthermore, upon applying to the MIT-BIH malignant ventricular arrhythmia database, the performance for malignant arrhythmia recognition using Hurst index was compared with that using Zheng's complexity measure. The Hurst index requires less computation and is more reliable in detecting VT and VF with short ECG episode. There is strong potential for using the Hurst index for malignant ventricular arrhythmia recognition in clinical applications. Authors' contributions SY conceived the study, performed data analysis and drafted the manuscript. CKL and KSM guided the study, helped the analysis and interpretation of the results, and critically reviewed the manuscript. All authors read and approved the final script.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC549211.xml

548133

Effect of bradykinin on nitric oxide production, urea synthesis and viability of rat hepatocyte cultures

Background It is well known that cytotoxic factors, such as lipopolysaccharides, derange nitrogen metabolism in hepatocytes and nitric oxide (NO) is involved among the other factors regulating this metabolic pathway. Hepatocytes have been shown to express large levels of NO following exposure to endotoxins, such as bacterial lipopolysaccharide and/or cytokines, such as tumour necrosis factor-α (TNFα), interleukin-1. The control role of arginine in both urea and NO biosynthesis is well known, when NO is synthesized from arginine, by the NOS reaction, citrulline is produced. Thus, the urea cycle is bypassed by the NOS reaction. Many authors demonstrated in other cellular types, like cardiomyocytes, that bradykinin caused the increase in reactive oxygen species (ROS) generation. The simultaneous increase of NO and ROS levels could cause peroxynitrite synthesis, inducing damage and reducing cell viability. The aim of this research is to study the effect of bradykinin, a proinflammatory mediator, on cell viability and on urea production in cultures of rat hepatocytes. Results Hepatocytes were treated with bradykinin, that stimulates nitric oxide synthase (NOS). NO release was determined using 4,5 diaminofluorescein diacetate (DAF-2DA), as fluorescent indicator of NO. Addition of the NOS inhibitor, N g -nitro-L-arginine methyl ester (L-NAME), to the culture medium inhibited the increase of NO production. Exposure of hepatocytes to bradykinin 0,1 mM for 2 hours resulted in a significant decrease of urea synthesis. Cell viability, instead, showed a significant decrease 24 hours after the end of bradykinin treatment as determined by 3-(4,5-dimethyl-2-thiazolyl)-2,5diphenyl-2H-tetrazolium (MTT) assay. L-NAME addition recovered urea production and cell viability at control values. Conclusion The findings suggest that the cell toxicity, after bradykinin treatment, effectively depends upon exposure to increased NO levels and the effects are prevented by L-NAME. The results show also that the increased NO synthesis induces a reduced urea production, that is another index of cell damage.

Background It is well known that cytotoxic factors, such as lipopolysaccharides, derange nitrogen metabolism in hepatocytes and nitric oxide (NO) is involved among the other factors regulating this metabolic pathway [ 1 ]. NO is a free radical that is involved in many cellular events. In the biological systems NO has an halflife long lasting few seconds. It is an oxidation intermediate, therefore is both an oxidant and a reducing agent of metabolic products. Its biosynthesis is mainly performed by converting L-arginine to L-citrulline. L-arginine analogues, such as N g -nitro-L-arginine methyl ester (L-NAME), act as false substrates and are selective inhibitors of NO synthesis. NO synthase (NOS) is either a constitutive or inducible enzyme. The endothelial isoform (e-NOS) and the neuronal isoform (n-NOS) are constitutive. The inducible form of the enzyme (i-NOS), has the main property to be not regulated by intracellular calcium concentration and Ca 2+ -calmodulin complex, unlike the constitutive form [ 2 ]. It is known that iNOS is expressed by many cell types including macrophages, smooth muscle cells and hepatocytes [ 3 ]. Hepatocytes have been shown to express large levels of NO following exposure to endotoxins, such as bacterial lipopolysaccharide and/or cytokines, such as tumour necrosis factor-α (TNFα), interleukin-1 [ 4 , 5 ]. NO may posses both cytoprotective and cytotoxic properties, depending on the amount and the isoform of NOS by which it is produced [ 6 ]. NO generally mediates beneficial responses, but becomes deleterious when coexistence with enhanced superoxide formation leads to the synthesis of peroxynitrite, a potent oxidant and nitrating agent [ 7 ]. According to this hypothesis, authors studied the effect of bradykinin, a proinflammatory mediator kinin, on cell viability and on urea production in cultures of rat hepatocytes. Kinins exert numerous physiological and pathological actions; they partecipate in vascular and cellular events that accompany the inflammatory processes. In pathological states, kinins are thought to be implicated in inflammatory diseases and in haemorrhagic and endotoxic shock [ 8 ]. To demonstrate the decrease of cell viability and urea production by bradykinin, the authors studied its effects on NO production. The measurements of NO release from hepatocytes were investigated by using a NO-specific fluorescence indicator, 4,5 diaminofluorescein diacetate (DAF-2DA) [ 9 ]. Results Effect of bradykinin treatment on NO production The amounts of released NO were measured using DAF-2DA, that specifically reacts with the oxidized form of NO, producing the fluorescent triazolofluorescein [ 9 ]. NO determination was performed after 2 hours of incubation in the presence of bradykinin (0.01 mM and 0.1 mM). As shown in figure 1 the treatment with 0.01 mM bradykinin did not produce NO increase compared to control, but 0.1 mM bradykinin increased significantly the NO release. In contrast no appreciable NO release was observed during the same period in hepatocytes cultured with 0.1 mM bradykinin and 1.68 mM L-NAME. Effect of bradykinin treatment on urea production To evaluate urea synthesis after bradykinin treatment, the hepatocytes were treated with 1 mM NH 4 Cl for 2 h. Figure 2 shows that only the treatment with 0.1 mM bradykinin significantly decreased urea production and that the treatment with 0.1 mM bradykinin and 1.68 mM L-NAME did not produce a significant urea level decrease in comparison to control. Effect of bradykinin treatment on cell viability To determine the effects of bradykinin on cell viability, the hepatocytes were exposed to bradykinin (either 0.01 mM or 0.1 mM) for an incubation time of 2 hours. In one experimental series, the cell viability was determined by MTT test after 2 hours of incubation. In a second one, culture medium containing bradykinin was removed and replaced with the same fresh medium at 2 hours after the addition of bradykinin, and then cell viability was measured 24 hours after the end of bradykinin treatment. The MTT test after 2 hours of incubation does not indicate any significant viability difference in treated hepatocyte cultures in comparison to control (figure 3A ). By MTT test after 24 h (figure 3B ), a significant lowering of viability is observed in bradykinin 0.1 mM treated hepatocytes in comparison to control. The decrease was significantly reduced by the simultaneous treatment with L-NAME 1.68 mM even if always significantly lower than in control. Cell viability was validated by Trypan blue exclusion test (Table 1 ). Discussion The role of NO as mediator of hepatic injury after endotoxic shock remains controversial [ 16 ]. Increased NO production in response to cytokines has been demonstrated in cultured hepatocytes [ 17 ]. Laskin et al. [ 18 ] demonstrated that the induction of acute endotoxemia, caused an increase in NO production in the liver. This was associated with expression of inducible nitric oxide synthase (iNOS) messenger m-RNA in hepatocytes. Also our data showed an increase of NO production after 2 hours treatment of culture with 0.1 mM bradykinin in an arginine supplemented medium, as substrate for the synthesis of NO. The simultaneous treatment with L-NAME, a known inhibitor of NOS, blocked the increase of NO production. In this work we analyzed the urea synthesis after bradykinin treatment. Urea synthesis was decreased after 2 hours treatment with bradykinin 0.1 mM and the simultaneous treatment with L-NAME leaves urea biosynthesis unaltered. These data can be attributed to the control role of arginine in both urea and NO biosynthesis. When NO is synthesized from arginine, by the NOS reaction, cytrulline, an intermediate of urea cycle, is produced. Thus, the urea cycle is bypassed by the NOS reaction [ 1 ]. Whether NO exerts cytotoxic or cytoprotective action remains unclear [ 6 ]. We also found a significant decrease of viability, at long term, in hepatocytes subjected to bradykinin treatment. The simultaneous treatment of hepatocytes with L-NAME improves cell viability even if control levels are not restored. The data show that the increased NO production plays a role in liver damage induction, that follows the proinflammatory mediator treatment. The hepatocellular injury attributed to NO may be due either to its direct cytotoxicity or its reaction with superoxide to produce the toxic nitrogen metabolite peroxynitrite [ 19 ]. Oldenburg et al. [ 20 ], demonstrated in other cell types, like cardiomyocytes, that bradykinin caused the increase in reactive oxygen species (ROS) generation. At last, our results show that the increased NO synthesis induces a reduced urea production, that is an index of cell damage. The simultaneous treatment of liver cell cultures with L-NAME decreases NO levels and sustains overall biosynthesis activities and cell viability. Conclusions In summary, we conclude that 0.1 mM bradykinin treatment induces an increase of NO levels and reduction of urea synthesis in the hepatocytes. This increased NO production mediates, after 24 hours, cell toxicity as shown by MTT test. In contrast, the administration of the NOS inhibitor L-NAME protects against cell damage and increases urea levels, suggesting that NO plays a key role in the bradykinin-induced liver damage. Methods Materials Unless otherwise specified, all chemicals were obtained from Sigma (St. Louis, MO, USA). Isolation and culture of rat hepatocytes Hepatocytes were isolated from male rats, Wistar strain, (180 to 200 gbw), by a modification of the method of Seglen [ 10 ]. All procedures on the animals were performed according to the CEE directive n. 86/609 on animal experimentation. Rats were anesthetized with diethylether, the pre-perfusion of the liver in situ was performed at a rate of 20–30 ml/min with Ca 2+ -free Hanks balanced salt solution. The liver was then excised and the digestion was carried out by adding 0.05% (w/v) collagenase (type IV) in Hanks balanced salt solution supplemented with CaCl 2 ·H 2 O (0.0186 g/L) at a flux rate of 40 ml/min. At this point liver was transferred to a square plate containing 100 ml of RPMI 1640 medium supplemented with 200 mM L-glutamine, 20 ml/L essential amino acid solution and 10 ml/L non-essential amino acid solution, 1% antibiotic antimycotic stabilized solution and 100 μM L-arginine (incomplete medium). The cells were dispersed by gentle distruption with a stainless steel comb. After filtration through 200 μm Nytal mesh, parenchymal cells (hepatocytes) were separated from nonparenchymal cells (endothelial cells, Kupffer cells and stellate cells) by centrifugation at 50 g in Eppendorf Centrifuge 5810R at 4°C for 2 minutes and then washed twice in washing buffer [ 11 ]. Then the cells were resuspended in the same medium and filtered through 63 μm Nytal mesh. The viability of the cells was more than 80%, as estimated by trypan blue dye exclusion test [ 12 ]. After cell counting the cells were diluited at a concentration of 5 × 10 5 cells/ml with incomplete medium supplemented with 2% fetal calf serum, 0.1 U/ml insulin and 10 -6 M dexamethasone (complete medium). The hepatocytes were then plated in 24 well-plates coated with rat tail collagen at the final cell density of 2.5 × 10 5 cells per well and incubated at 37°C in an humidified atmosphere of 5% CO 2 and 95% air. After 6 hours incubation, the medium was changed and replaced with incomplete medium to remove dead cells. To verify the isolation method efficiency, the acid fosfatase activity per mg of proteins was evaluated. According to literature data, the specific activity of acid fosfatase in nonparenchimal cells is 1,7 folds the same activity in parenchimal cells [ 13 ]. Treatment After 24 hours of culture the hepatocytes were exposed either to bradykinin (0.01 mM and 0.1 mM) or bradykinin 0.1 mM supplemented with L-NAME 1.68 mM [ 14 ]. Determination of NO from hepatocytes DAF-2DA (Alexis Biochemicals, Lausen, Switzerland) was dissolved in DMSO (1 mg/0.45 ml) and diluted to 10 μM in phosphate buffer (0.1 M, pH 7.4). Then the cells were either incubated in phosphate buffer containing 10 μM DAF-2DA, bradykinin (0.01 mM and 0.1 mM) and bradykinin 0.1 mM supplemented with L-NAME 1.68 mM. After 2 hours of incubation in this reaction mixture, the fluorescence from the reaction of DAF-2DA with NO released from hepatocytes was measured with Perkin-Elmer MPF-44B Spectrofluorimeter calibrated for excitation at 495 nm and emission at 515 nm. Results were expressed as a percentage of the fluorescence of the samples in comparison to control. Determination of urea synthesis To determine the effects of bradykinin on urea production, cells were treated either with bradykinin (0.01 mM and 0.1 mM) and cotreated with bradykinin 0.1 mM and L-NAME 1.68 mM. At the same media 1 mM NH 4 Cl was added. After 2 hours urea levels in the media were measured by spectrophotometric method using Urea Color 2 Kit (Sclavo Diagnostics, Siena, Italia) measuring absorbance at 600 nm and blank sample with the same NH 4 Cl final concentration was used. Urea synthesis was calculated as ng urea per cell per hour. Determination of cell viability Cell viability was determined by MTT test method [ 15 ] and confirmed by Trypan blue exclusion test [ 12 ]. MTT (5 mg/ml) was dissolved in RPMI-1640 without phenol red. The solution is filtered through a 0.2 μm filter and stored at 2–8°C for frequent use. To determine the effects of bradykinin on cell viability, cells were either treated with bradykinin (0.01 mM and 0.1 mM) and cotreated with bradykinin 0.1 mM and L-NAME 1.68 mM for a 2 h period. After that cells were used either immediately or after an additional 24 h incubation period in incomplete medium. For the determination of cell viability, the medium has been discarded and MTT solution was added and incubated for 3 hours. At the end of the incubation period the MTT solution was removed and the cells and dye cristals were dissolved by adding dimethylsulfoxide (DMSO). Absorbance was measured at 570 nm in a Shimadzu UV-2100 Spectrophotometer and the results were expressed as a percentage of the absorbance of the samples in comparison to control. Statistical analysis At least four independent determinations of each parameter were compared to control using Student's T-test. Differences were considered significant when p < 0.05 was obtained. Authors' contributions SS: Fluorimetric analysis and overall statistical analysis of data MG: Director of research MS: Spectrophotometric analysis CR: Primary hepatocyte cultures and characterization All authors read and approved the final manuscript

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548133.xml

406512

Effects of fire and fire intensity on the germination and establishment of Acacia karroo, Acacia nilotica, Acacia luederitzii and Dichrostachys cinerea in the field

Background While fire has been used in some instances to control the increase of woody plants, it has also been reported that fire may cause an increase in certain fire-tolerant Acacia tree species. This study investigated germination of Acacia karroo , A. luederitzii and Dichrostachys cinerea , thought to be increasing in density, as well as the historically successful encroaching woody species, A. nilotica, in savanna grassland, Hluhluwe-iMfolozi Park, South Africa. A. karroo is thought to be replacing A. nilotica as the dominant microphyllous species in the park. We tested the hypothesis that observed increases in certain woody plants in a savanna were related to seed germination and seedling establishment. Germination is compared among species for burnt and unburnt seeds on burnt and unburnt plots at three different locations for both hot and cool fires. Results Acacia karroo showed higher germination ( A. karroo 5.1%, A. nilotica 1.5% and A. luederitzii 5.0%) levels and better establishment ( A. karroo 4.9%, A. nilotica 0.4% and A. luederitzii 0.4%). Seeds of the shrub Dichrostachys cinerea did not germinate in the field after fire and it is thought that some other germination cue is needed. On average, burning of A. karroo , A. nilotica and A. luederitzii seeds did not affect germination. There was a significant difference in the germination of burnt seeds on burnt sites (4.5%) and burnt seeds on unburnt plots (2.5%). Similarly, unburnt seeds on unburnt sites germinated better (4.9%) than unburnt seeds on burnt sites (2.8%). Conclusion We conclude that a combination of factors may be responsible for the success of A. karroo and that fires may not be hot enough or may occur at the wrong time of year to control A. karroo establishment in HiP. Although germination and establishment of A. karroo was higher than for A. nilotica a competitive advantage after fire could not be shown.

Background The increasing density in the woody component of savannas has been widely reported [ 1 - 5 ] with special mention being made of Acacia karroo Hayne [ 6 , 7 ] and A. nilotica (L.) Willd. Ex Del. subsp. kraussiana (Benth.) Brenan, [ 8 , 9 ]. in some areas, as major contributors to the phenomenon. In Hluhluwe-iMfolozi Park Dichrostachys cinerea (L.) Wight & Arn. and A. luederitzii Engl. var. retinens (Sim) Ross & Brenan are also thought to contribute to this phenomenon. In hard seeded legumes dormancy is broken by rupturing part of the seed coat. The rupturing of the seed coat may be induced by heat from fire [ 10 ] enabling water to enter the seed and start the process of germination. Many studies have confirmed a release of legume seeds from dormancy after fire [ 10 - 17 ]. Fire temperature or intensity also has an effect on the germination of seeds [ 17 , 18 ] and low intensity fires may not be enough to break dormancy of hard-seeded legumes [ 19 ]. In other cases lower fire temperatures are preferable for germination with an increase in fire temperature causing seed mortality [ 18 ]. While some studies report that a decrease in grass cover favours the establishment of woody seedlings due to reduced competition [ 20 , 21 ], others [ 6 , 22 ] challenge these findings. These differences may however, be a result of species reacting differently to fire or competition. Some Acacia species are shade intolerant resulting in decreased seedling establishment in shady areas [ 20 , 23 , 24 ]. Other Acacia species have been found to be tolerant of low light conditions and may even experience increased seedling survival [ 6 ]. The frequency of fires may affect the direction of change in woody plant density [ 5 ]. While it has been suggested that fire may increase Acacia densities [ 10 ], it is also used to clear acacias from grassland [ 25 ]. This contradictory situation in the literature concerning the effect of fire necessitates further research, as it is clear that continuous use of incorrect burning practices may have disastrous consequences. This study investigated the direct (heat) and indirect (grass removal) effects of fire on seed germination and seedling establishment of A. nilotica , A. karroo , A. luederitzii and Dichrostachys cinerea in Hluhluwe-iMfolozi Park (HiP), where an increase in woody plant density over the past 40 years has been reported [ 26 - 28 ]. It has also been reported that A. karroo is apparently replacing A. nilotica as the dominant microphyllous element [ 27 , 28 ]. This study reports on the effects of burning, fire intensity and burning of sites on germination; burning, fire intensity, burning of sites and grass length (shade) on seedling establishment and specific species responses to treatments (treatment species interactions). Results Germination None of the seeds of D. cinerea germinated in the field and it was therefore excluded from the model for the field experiment. Testing for differences among treatments was based on the maximum number of seedlings for each species at each location over the 31-week period (Figure 2 ). A description of the factors used in both the germination and establishment models is given in Table 1 . Figure 2 Mean number of germinated seeds recorded over a 31-week period at three different locations in HiP for a) Acacia karroo , b) Acacia luederitzii and c) Acacia nilotica . Table 1 Descriptions of factors used in the models and number of seeds used for each factor. Germination Establishment Factor/Description Total number of seeds Number not germinated Number germinated Percent germinated Total number of seeds Number not established Number established Percent established Total 4073 3923 150 3.68 4062 3966 96 2.36 Location Seme 1348 1287 61 4.53 1337 1302 35 2.62 Nombali 1364 1300 64 4.69 1364 1316 48 3.52 Le Dube 1361 1336 25 1.84 1361 1348 13 0.96 Species A. karroo 1786 1695 91 5.10 1788 1701 87 4.87 A. luederitzii 720 684 36 5.00 707 704 3 0.42 A. nilotica 1567 1544 23 1.47 1567 1561 6 0.38 Burnt or unburnt burnt 2021 1950 71 3.51 2030 1985 45 2.22 unburnt 2052 1973 79 3.85 2032 1981 51 2.51 Tall or short grass tall (>0.1 m) 2039 1961 78 3.83 2041 1993 48 2.35 short 2034 1962 72 3.54 2021 1973 48 2.38 Site burnt or unburnt burnt 2052 1977 75 3.65 2052 2003 49 2.39 unburnt 2021 1946 75 3.71 2010 1963 47 2.34 The ratio of the model deviance to the degrees of freedom was small (0.29) indicating that the model was a good fit. Location and species were the only main effects significantly affecting germination (Table 2 ). Acacia karroo had the highest germination of all species (Table 1 ). Table 2 Statistics indicating significance of the factors and interactions on germination. Significant factors are in bold. Factor df Log-likelihood Chi-Square Wald Stat. P Location 2 -587.555 13.915 11.547 0.003 Species 2 -597.790 34.386 25.394 0.000 Burnt status 1 -582.073 2.951 2.822 0.093 Grass length 1 -580.622 0.050 0.050 0.822 Site burn status 1 -580.608 0.021 0.021 0.885 Location*species 4 -582.584 3.974 3.827 0.430 Location*burn status 2 -582.929 4.664 4.373 0.112 Location*grass length 2 -586.296 11.397 10.812 0.004 Location*site burn status 2 -580.703 0.212 0.211 0.900 Species*burn status 2 -581.173 1.151 1.145 0.564 Species*grass length 2 -581.019 0.843 0.837 0.658 Species*site burn status 2 -583.309 5.424 5.166 0.076 Burn status*grass length 1 -580.767 0.340 0.341 0.559 Burn status*site burn status 1 -585.060 8.926 8.656 0.003 Grass length*site burn status 1 -587.530 13.866 13.082 0.000 Interaction terms that had a significant effect on germination were, location × grass length, burn status × site burn status and grass length × site burn status (Table 2 ). Germination of burnt seeds in burnt sites (4.5%) was significantly higher than that of burnt seeds in unburnt sites (2.5%). Similarly, unburnt seeds in unburnt sites had a higher germination percentage (4.9%) than unburnt seeds in burnt sites (2.8%). The estimated odds of germination and their associated probabilities for the factors and their interactions are given in Additional file 1 . The odds ratios for significant effects were calculated. Thus a comparison between A. karroo and A. nilotica with regards to seeds germinating was made, where Thus the odds of germinating are four times more for A. karroo than for A. nilotica . Similarly A. nilotica was four times less likely to germinate than A. luederitzii while A. karroo and A. luederitzii had the same odds of germinating. Differences in germination among species for the various treatments are given in Table 3 . Table 3 A comparison of germination among species for the different levels of the main factors. A. karroo A. luederitzii A. nilotica Factor/description n Total count Not germ germ % erm Total count Not germ germ %germ Total count Not germ germ % germ Location*Species Seme 48 591 558 33 5.91 240 224 16 7.14 517 505 12 2.38 Nombali 48 596 551 45 8.17 240 227 13 5.73 528 522 6 1.15 Le Dube 48 599 586 13 2.22 240 233 7 3 522 517 5 0.97 Burnt or unburnt*Species burnt 72 886 839 47 5.6 360 344 16 4.65 775 767 8 1.04 unburnt 72 900 856 44 5.14 360 340 20 5.88 792 777 15 1.93 Tall or short grass*Species tall 72 895 851 44 5.17 360 340 20 5.88 784 770 14 1.82 short 72 891 844 47 5.57 360 344 16 4.65 783 774 9 1.16 Site burnt or unburnt*Species yes 72 900 854 46 5.39 360 338 22 6.51 792 785 7 0.89 no 72 886 841 45 5.35 360 346 14 4.05 775 759 16 2.11 There was 2.3 times less germination at Le Dube than at Nombali and 2.6 times less at Le Dube than at Seme. Germinations were 1.2 times more likely at Seme than at Nombali. Seedling establishment The ratio of the model deviance to the degrees of freedom was small (0.17) indicating that the model fitted the data well. Location and species were the only main effects significantly affecting establishment in the field (Table 4 & Figure 3 ). Acacia karroo showed significantly higher percentage establishment than any of the other species ( Additional file 2 , Table 5 & Figure 3 ). Table 4 Statistics indicating significance of factors and interactions on establishment. Significant factors are indicated in bold. Factor df Log-likelihood Chi-Square p Location 2 -443.238 22.292 <0.001 Species 2 -395.199 96.079 <0.001 Burnt status 1 -395.050 0.297 0.586 Grass length 1 -395.049 0.002 0.962 Site burn status 1 -395.040 0.018 0.894 Location*species 4 -391.756 6.568 0.161 Location*burn status 1 -380.850 21.812 <0.001 Location*grass length 2 -373.542 14.617 <0.001 Location*site burn status 2 -367.865 11.353 0.003 Species*burn status 2 -367.468 0.795 0.672 Species*grass length 2 -367.344 0.248 0.884 Species*site burn status 2 -367.180 0.329 0.848 Burn status*grass length 1 -366.723 0.913 0.339 Burn status*site burn status 1 -360.267 12.913 <0.001 Grass length*site burn status 1 -351.784 16.965 <0.001 Figure 3 Predicted mean establishment for the significant main effects of a) species and b) location. Vertical error bars show 95% confidence limits. Table 5 A comparison of establishment among species for the different levels of the main factors A. karroo A. luederitzii A. nilotica Factor/Description n Total count Not estab estab % estab Total count Not estab estab % estab Total count Not estab estab %estab Location*Species Le Dube 48 599 590 9 1.53 240 239 1 0.42 522 519 3 0.58 Nombali 48 598 553 45 8.14 238 237 1 0.42 528 526 2 0.38 Seme 48 591 558 33 5.91 229 228 1 0.44 517 516 1 0.19 Burnt or unburnt*Species Burnt 72 886 843 43 5.1 347 346 1 0.29 797 796 1 0.13 Unburnt 72 902 858 44 5.13 360 358 2 0.56 770 765 5 0.65 Tall or short grass*Species Tall 72 897 854 43 5.04 360 358 2 0.56 784 781 3 0.38 Short 72 891 847 44 5.19 347 346 1 0.29 783 780 3 0.38 Site burnt or unburnt*Species Yes 72 900 855 45 5.26 360 359 1 0.28 792 789 3 0.38 No 72 888 846 42 4.96 347 345 2 0.58 775 772 3 0.39 Interaction terms, location × burn status, location × grass length, location × site burn status, burn status × site burn status and grass length × site burn status had a significant effect on establishment (Table 4 ) (Figure 4 ). Figure 4 Predicted mean establishment for significant interactions of site burn status and a) location, b) seed burn status and c) grass length. The solid line represents unburnt sites and the dotted line burnt sites. Vertical error bars show 95% confidence limits. Additional file 2 gives the estimated odds of non-establishment and their associated probabilities for the factors and their interactions. The odds ratios for significant effects were calculated and are given (see Additional file 3 ). Acacia karroo was 16.2 times more likely to establish than A. nilotica. Similarly A. luederitzii was 1.4 times more likely to establish than A. nilotica while A. karroo had 11.2 times more chance of establishing than A. luederitzii. Species differences in establishment for the various treatments are given in Table 5 . The odds of establishment were 8046.2 times less at Le Dube than at Nombali and 5850.5 times less at Le Dube than at Seme. 1.4 times more seedlings were likely to establish at Nombali than at Seme. Discussion The lack of germination of D. cinerea in the field suggests that some disturbance other than fire is needed to cause a release from dormancy and commence germination. Germination of all species in the field was low. As the seeds relied on cotyledons for food, soil moisture may have been a limiting factor. As rainfall was not recorded, this should be kept in mind when interpreting the results. Five point one percent of A. karroo seeds germinated, which was higher than the other two species. Story [ 29 ] found similar levels of germination for A. karroo, with 6.6% of seeds germinating under natural conditions in the field. He also found that A. karroo germination was erratic, with germinations still being recorded after 423 days. This was similar to what was found in this study, with the number of A. karroo seedlings still increasing until the end of the experiment. Acacia nilotica also showed dormancy with sporadic germination events over the 31-week period. Acacia luederitzii did not show dormancy with most germinations taking place in the first 3 weeks of the experiment. Acacia nilotica has a thick seed coat, which could account for it's poor level of germination. One would predict increased germination of burnt seeds due to a breaking of dormancy [ 18 ], but this was not the case. A possible explanation is that the temperature of the fires in this study, though not measured, might not have been sufficient to break dormancy in this species. Some Acacia species are temperature specific, suggesting a temperature threshold for germination [ 18 , 20 ]. This is unlikely in this case as Radford et al . [ 30 ] found A. nilotica seeds to be highly vulnerable to fire with a 80% mortality of seeds on the soil surface. The current study, however, found no difference in germination between burnt and unburnt seed or seeds burnt at different temperatures. This finding is inconsistent with the recent study by Kanz [ 20 ] who found increased seed germination in low fires compared to the control as well as that of Okello and Young [ 31 ] who found increased germination of unburnt seeds. Auld & O'Connell [ 18 ] had similar results to that of Kanz [ 20 ] with strong germination responses to heat. Location had a significant effect on germination with Le Dube having very low germination overall and Seme having the most germinations. Germination at Nombali and Seme were similar. Site-specific effects may be attributed to various factors such as microclimate or soil type. Sites may also have different water infiltration rates and runoff, which may result in differences in germination levels. Okello and Young [ 31 ], however, found that soil type did not affect germination or establishment of Acacia drepanolobium in Kenya. The current study did not find a difference in the number of seedlings in burnt and unburnt patches. While neither burning of seeds nor burning of sites had any effect on germination, the interaction factor proved significant with unburnt seeds showing increased germination in unburnt sites as did burnt seeds in burnt sites. Kanz [ 20 ] also found greater seedling emergence of unburnt seeds in unburnt areas. This might be a result of burnt seeds imbibing faster than unburnt seeds, possibly making them more susceptible to rot. Burnt seeds would therefore show poorer germination in unburnt areas due to increased moisture retention. Similarly, unburnt seeds would require more moisture to imbibe, resulting in decreased germination in burnt areas due to decreased moisture in these open areas. Whilst more seeds germinated in short grass at both Le Dube and Nombali, those at the short-grass site (Seme) had higher levels of germination in tall grass sites. The short grass site at Seme is a white rhinoceros ( Ceratotherium simum ) grazing lawn with very short grass, which may lead to seeds losing moisture through more direct sunlight. This suggests a similar pattern to the seed burn × site burn interaction. The tall grass site at Seme had higher germination than any of the other tall or short grass sites. This may be due to possible site-specific effects mentioned earlier. There was also an interaction between grass length and site burn with seeds in burnt, short grass showing higher germination than those in burnt, tall grass and unburnt sites showing higher germination in tall grass. As half of the seeds on a burnt or unburnt site were burnt themselves, it is possible that this interaction is due to temperature sensitivity in seeds. Burning in tall grass (hotter fires) may be detrimental to the germination of seeds [ 18 ] while cooler fires may be sufficient to break dormancy and cause germination. Higher germinations in unburnt tall grass areas suggest a shade effect. This is not certain, as the effects of shade and grass competition were not separated in this study. Acacia karroo has however been reported as having an increased ability to survive in shade with recruitment of seedlings being dependent on moisture availability [ 6 ]. Tall grass species may retain more moisture than short grass species, affording seeds a better opportunity for germination. No species factor interactions were observed suggesting that though species had different germination levels, they did not respond differently to the treatments. The same factors and interactions found to be significant influences on germination were found to influence establishment. This was expected as increased germination for these treatments would result in better establishment. The interaction patterns for most of the treatments, however, were different to those of the germination model. Owing to the low levels of germination, interspecific and intraspecific competition was thought to play a minor role in seedling establishment. Le Dube again had the least seedlings at 31 weeks while Nombali had the best establishment. Seme, which had the highest level of germination, had establishment levels somewhere between that of the other two sites. It is again suggested that this may be due to soil or rainfall factors. Forty-five out of forty-eight seedlings established at Nombali and thirty-three out of thirty-five at Seme were A. karroo seedlings. This species is known to be dependent on moisture availability for survival [ 6 ] and these two sites might have better water retaining ability than Le Dube. At week 31, 87 A. karroo seedlings had established as opposed to six of A. nilotica and three of A. luederitzii. The high germination, but poor survival of A. luederitzii suggests that the absence of this species in the Hluhluwe section of HiP is not due to seed limitation or germinability, but possibly due to environmental factors decreasing its ability to establish. The differences in seedling survival between species are consistent with those reported by Kanz [ 20 ] who found higher seedling survival for A. karroo than A. nilotica . The location × grass length interaction revealed the same patterns as for germination with regards to Nombali and Seme with Seme showing better establishment in tall grass and Nombali showing better establishment in short grass. There was no difference between establishment on tall and short grass at Le Dube. The short grass site at Nombali had the highest number of seedlings surviving at week 31. The grass length × site burn interaction displayed the same patterns as for the germination model, but this was not the case for the seed burn status × site burn status interaction. While unburnt seeds still did well on unburnt sites, burnt and unburnt seeds showed decreased establishment on burnt sites suggesting that, as a result of increased irradiance, burnt (open) sites may not hold sufficient moisture for seedlings to survive. The interaction effects found to be significant for establishment only, both suggest the importance of fire temperature. Location × seed burn status and location × site burn status could both relate to the different grass lengths, and thus specific fire temperatures, at the three sites. Temperature sensitivity in Acacia species have been reported elsewhere [ 11 , 14 , 17 , 20 ]. Kanz [ 20 ] found increased survival and growth in burnt areas. In this study, Nombali was the only location to have higher establishment on burnt sites, while Seme had increased establishment on unburnt sites and Le Dube very little establishment overall. In general, however, this study found no difference in establishment in burnt and unburnt areas. Chirara, Frost & Gwarazimba [ 7 ] found that intensity of grass defoliation does not affect seedling establishment of A. karroo during the first year. Similarly, there was no difference in establishment of A. karroo in burnt or unburnt and tall or short grass sites. Smith & Goodman [ 32 ] reported that A. nilotica seedlings, however, almost exclusively occurred away from canopy cover, suggesting an inability to establish in shaded environments. Acacia tortilis also showed a greater proportion of established seedlings in open than shaded areas [ 23 ]. We did not find a difference in establishment of A. nilotica in tall and short grass, but its establishment was so low that no real prediction can be made. Conclusions Seedling establishment of A. karroo is strongly moisture dependent [ 6 ] and one would expect that A. karroo is more likely to invade moist rather than semi-arid grassland. This suggests that Hluhluwe Game Reserve, being an area with moist grassland, would be more prone to invasion by A. karroo . It has also been reported that A. karroo has the ability to withstand fire [ 17 ]. A combination of these factors may contribute to the success of A. karroo in the field and may be the reason for A. karroo 's success over A. nilotica as the most important encroaching Acacia species in HiP at present. The literature does, however, suggest that high intensity fires may result in seed mortality [ 18 , 20 ]. It has, however, been reported that A. karroo seedlings survive fires from as little as 12 months of age [ 29 ]. Therefore, if fires are not hot enough to kill the seeds allowing them to germinate and seedlings to establish, management burns in the following year may not be useful in its attempt to control the establishment of this species. Back fires have higher fire intensities than head fires [ 20 ]. We therefore suggest that backfires be used during management burns and that fire frequency be increased in suitable areas in an attempt to slow down the rate of encroachment by A. karroo. It has been reported that spring burns are the most effective ([ 33 ] in [ 29 ]) and this should be taken into account. Methods Study site The study was done in HiP, KwaZulu-Natal, South Africa (28°00' – 28°26' S, 31°43' – 32°09'E). HiP is a 960 km 2 fenced protected area comprising the former Hluhluwe and iMfolozi Game Reserves, and the corridor of land that links the areas. The park has a moderate coastal climate, ranges in altitude from 60 – 750 m above sea level [ 34 ] and has a summer rainfall ranging between 760 and 1250 mm per annum. Hluhluwe Game Reserve has a mean annual rainfall of 990 mm, while iMfolozi Game Reserve has a mean annual rainfall of 720 mm [ 34 ]. Periodic fluctuations in above or below average annual rainfall occur, resulting in wet and dry spells of approximately nine years [ 35 ]. The range in average monthly temperature is between 13 and 33°C [ 36 ]. Most of Hluhluwe Game Reserve is found on rocks of the Ecca and Beaufort series with some basalt in the east [ 37 ]. King [ 37 ] identified seven geological formations: (1) the Granite-Gneiss base, (2) the Table Mountain sandstone, (3) the Dwyka tillite, (4) The Ecca and Beaufort series, (5) the Stormberg series, (6) fault breccias and (7) recent deposits. The main soils types associated with the Ecca and Beaufort series are Swartland and Sterkspruit, while areas of Shortlands, Milkwood and Bonheim series are found in association with the dolerite regions [ 34 ]. They also report that shallow Mispah soils occur extensively in the reserve. The vegetation in the park has been described as bushveld – savannah comprising five broad vegetation types [ 38 ]. The thickets are wooded groups of similar-sized, small (usually less than three metres) trees of mainly one species that grows densely to the exclusion of other species. The thornveld consists of scattered thorn trees on grassland with deciduous, broad-leaved trees standing out above the thorn trees while the woodlands are densely wooded areas of tall trees that may contain many different, mainly broadleaved species. The well drained, shallow soils of the rocky outcrops support scattered trees of various sizes, while the termite mounds are nutrient rich patches sustaining dense clumps of trees that form small, wooded islands [ 38 ]. Locally the reserve is described as Natal Lowveld Bushveld and falls within the savanna biome [ 39 ]. The field experiment took place in the Hluhluwe and Corridor sections of the HiP. Acacia luederitzii occurs in large numbers in certain areas of the iMfolozi part of the reserve but is mostly absent from the Hluhluwe and Corridor sections. Acacia nilotica, A. karroo and D. cinerea are found throughout the park. As opposed to the scattered trees found in iMfolozi, A. nilotica covers extensive areas of Hluhluwe and the Corridor and is usually found below the 300 m contour [ 34 ]. Whateley & Porter [ 34 ] described an A. karroo – D. cinerea induced thicket throughout the area, but particularly in the Corridor and Hluhluwe Reserves. Acacia luederitzii seeds used in this study were therefore collected in iMfolozi Game Reserve while those of the other species were collected in Hluhluwe. Germination The effect of fire, fire intensity and burning of sites on the germination of seeds of A. nilotica , A. karroo , A. leuderitzii and D. cinerea was tested in a field experiment. Seeds of all species were collected between May and August 2000. Parasitized seeds were extracted. Prior to planned management burns, six groups of seeds were placed in tall grass (taller than 0.10 m) and six in short grass (shorter than 0.10 m) at three locations (Nombali, Seme and Le Dube). Tall grass produces hotter fires than short grass due to increased fuel load, which increases available heat energy [ 40 ]. Sites were cleared of existing pods/ seeds prior to the experiment and as podding season was over, no uncontrolled additions are expected to have occurred. Dichrostachys cinerea seeds were only put out at Seme and Nombali. Each group contained 22 A. nilotica , 25 A. karroo , 10 A. leuderitzii and 10 D. cinerea seeds. Seeds were placed on the soil surface a day before each of the burns (Nombali two days before). This is considered the natural situation for the seeds with soil stored seed banks being virtually non-existent [ 41 ]. Seme and Le Dube were burnt on 2 October and Nombali on 30 September 2000 shortly before the start of spring rains and natural seed release. After the burns, three of the groups of burnt seeds were removed from the tall and short grass and placed on unburnt tall and short grass sites at the same location respectively. Three groups of unburnt seeds were then added to each of the tall and short grass sites. A 13 mm mesh cage with 18 cm × 18 cm × 18 cm sides was used to protect each group of seeds and any germinated seedlings from rodent and herbivore predation. Cages were placed at half metre intervals and seeds placed on the soil surface in a group in the middle of each cage Seeds were considered to be germinating when a root started showing. A diagrammatical representation of the experiment is given in Figure 1 . Germination was recorded at 1, 3, 5, 7, 9, 11, 14, 17, 20, 23, 27 and 31 weeks. The experiment ended in May 2001. Figure 1 Diagrammatical representation of the experimental design used to test the effect of fire on seed germination and establishment. Arrows indicate movement of seeds between burnt/unburnt tall/short grass plots. We thus applied 96 possible seed treatment combinations for investigating factors affecting germination in the field (4 species × 2 burn treatments × 3 locations × 2 location burn treatments × 2 fire intensities). Seedling establishment To test the effect of fire, fire intensity, burning of sites and grass length (shade) on seedling establishment of A. nilotica , A. karroo , A. leuderitzii and D. cinerea , data as on week 31 of the field experiment described above were used. Seedlings were considered to be established when they were rooted in the ground and the cotyledons replaced with leaves. Establishment was based on the total number of seeds. Data analysis The "STATISTICA ® " [ 42 ] Generalized Linear Model (GLZ) module was used to construct linear logistic models for germination and establishment proportions as response variables for the field experiment. As data were recorded as presence (1) or absence (0) of seedlings, a binomial distribution was assumed [ 43 ]. In both cases, main effects and second order interactions were included in the model. The logit model may therefore be written as follows: where = the log of variable 1 and 2 at different levels of the factors as given below λ' = the overall mean effect of the categories = the effect of the j th species ( j = A. karroo , A. luederitzii , A. nilotica , D. cinerea ) = the effect of the k th location ( k = Le Dube, Nombali, Seme) = the effect of the l th seed burn status ( l = burnt, unburnt) = the effect of the m th grass length ( m = short, tall) = the effect of the n th site burn status ( n = burnt, unburnt) = the interaction effect between the j th species and the k th location = the interaction effect between the m th grass length and the n th site burn status. The logit model may be written as a generalized linear model as follows: where , , , , , , , , , and are parameters to be estimated from the data and B, C, D, E and F refer to the explanatory variables species, location, burn status, grass length and site burnt status respectively. The estimated parameters for the GLZ were used to obtain the estimated parameters for the logit model. The estimated parameters of the odds were calculated for each factor or combination of factors (including the intercept) as the exponent of the estimated parameters of the logit model. The estimated odds of germination under any condition were then calculated as the product of the estimated parameter of the odds of the intercept (estimated geometric mean odds) and the factor or combination of factors in question. The odds of germination for significant treatment combinations were compared. The predicted number of seeds germinating and seedlings establishing as calculated with the model based on presence/absence data, were seen as being appropriate for interpretation as summaries of the data. Thus, differences in the predicted mean number of seeds germinating and seedlings establishing (given as a fraction of the total number of seeds) were illustrated graphically for each significant treatment combination. Authors' contributions MW designed the experiment, participated in fieldwork, performed the statistical analysis and drafted the document. MJS participated in fieldwork, the coordination of the study and drafting of the document. JJM supervised the work and assisted in the drafting of the document. All authors read and approved the final manuscript. Supplementary Material Additional File 1 The parameters of the logit model and odds, estimated odds of germination and the ratio of germination to non-germination for the factors included in the model for germination of certain Acacia seeds in HiP. Gives parameters of the logit model and estimated odds of germination for the various levels of the factors used. Click here for file Additional File 2 The parameters of the logit model and odds, estimated odds of establishment and the ratio of establishment to non-establishment for the factors included in the model for establishment of certain Acacia species in HiP. Gives parameters of the logit model and estimated odds of establishment for the various levels of the factors used. Click here for file Additional File 3 Odds ratios for all significant interactions of the establishment model. Compares the odds of establishment between different levels of the factors used. Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC406512.xml

539047

Finding Cures for Tropical Diseases: Is Open Source an Answer?

The Tropical Disease Initiative will be a Web-based, community- wide effort where scientists from the public and private sectors join together to discover new treatments

Only about 1% of newly developed drugs are for tropical diseases, such as African sleeping sickness, dengue fever, and leishmaniasis [1] . While patent incentives and commercial pharmaceutical houses have made Western health care the envy of the world, the commercial model only works if companies can sell enough patented products to cover their research and development (R&D) costs. The model fails in the developing world, where few patients can afford to pay patented prices for drugs. It is easy and correct to say that Western governments could solve this problem by paying existing institutions to focus on cures for tropical diseases. But sadly, there does not appear to be enough political will for this to happen. In any case, grants and patent incentives were never designed with tropical diseases in mind. Two main kinds of proposals have been suggested for tackling the problem. The first is to ask sponsors—governments and charities—to subsidize developing-country purchases at a guaranteed price [ 2 , 3 , 4 ]. The second involves charities creating nonprofit venture-capital firms (“Virtual Pharmas”), which look for promising drug candidates and then push drug development through contracts with corporate partners. In this article, we discuss the limitations of these two approaches and suggest a third, “open source,” approach to drug development, called the Tropical Diseases Initiative (TDI). We envisage TDI as a decentralized, Web-based, community-wide effort where scientists from laboratories, universities, institutes, and corporations could work together for a common cause (see www.tropicaldisease.org ). Why Open Source? The idea behind asking sponsors to subsidize developing country purchases at a guaranteed price is that this will prop up drug prices and restore incentives for developing new drugs [ 2 , 3 , 4 ]. In other words, it is a way of fixing the patent problem. However, subsidies have an important weakness: it is almost impossible to determine correctly how large the subsidy should be. In principle, the most cost-effective solution is to set a subsidy that just covers expected R&D costs. But how large is that? R&D costs are very poorly known, with the published estimates quoting uncertainties exceeding $100 to $500 million per drug. If the subsidy is set too low, companies cannot cover their R&D costs and nothing will happen. Set the subsidy too high, and the sponsor's costs skyrocket. To date, no sponsor has tried to implement these proposals. In the “Virtual Pharma” approach, governments and philanthropies fund organizations that identify and help support the most promising private and academic research. Examples include the Institute for One World Health ( www.iowh.org ), a not-for-profit pharmaceutical company funded mainly through private sources and the Gates Foundation, and the Drugs for Neglected Diseases Initiative ( www.dndi.org ), a public sector not-for-profit organization designed to mobilize resources for R&D on new drugs for neglected diseases. Virtual Pharmas have clearly started to bear fruit, and are responsible for most candidate treatments for tropical diseases currently under development. For example, the Drugs for Neglected Diseases Initiative has a portfolio of nine projects spread out across the drug development pipeline for the treatment of leishmaniasis, sleeping sickness, Chagas disease, and malaria [6] . But Virtual Pharmas face three important problems. The first is similar to the problem faced by subsidy proposals: guessing private-sector R&D costs. One needs to understand what a product costs in order to negotiate the best possible price—and guessing wrong is likely to be expensive. Second, Virtual Pharma's development pipelines will run dry without more upstream research. Research has been particularly weak in exploiting genomic insights [7] . Third, tropical disease research is badly underfunded. For this reason, Virtual Pharma cannot succeed without rigid cost containment. We believe that a new, community-wide consortium, the Tropical Disease Initiative, can help solve these problems. Its success would help keep Virtual Pharma's R&D pipeline full. Furthermore, it would use open-source licenses to keep its discoveries freely available to researchers and—eventually—manufacturers. As we explain below, well-designed open-source licenses are the key to containing Virtual Pharmas' R&D costs. While we expect the final choice of license to be made by TDI's members, the guiding principle should be to pick whatever license lets developing country patients derive the most benefit from TDI's work. Possible choices are shown in Box 1 . Box 1. Possible Licenses for TDI Discoveries A public-domain license that permits anyone to use the information for any purpose. Licenses similar to the Creative Commons Attribution License ( http://creativecommons.org/licenses/by/2.0 ) that permit anyone to use the information for any purpose, provided proper attribution is given. Licenses such as the General Public License ( www.opensource.org/licenses/gpl-license.php ) that prohibit users from seeking intellectual property rights. Licenses that permit commercial companies to obtain and exploit patents outside the developing world. These would allow Virtual Pharma to stretch its own R&D funds by letting corporate partners sell patented products to ecotourists, governments, and other consumers living in the industrialized world. How It Works To date, open-source methods have made little headway beyond software [8] . However, computing and computational biology are converging. In the same way that programmers find bugs and write patches, biologists look for proteins (“targets”) and select chemicals (“drug candidates”) that bind to them and affect their behavior in desirable ways. In both cases, research consists of finding and fixing tiny problems hidden in an ocean of code. What would open-source drug discovery look like? As with current software collaborations, we propose a Web site where volunteers use a variety of computer programs, databases, and computing hardware ( Figure 1 ). Individual pages would host tasks like searching for new protein targets, finding chemicals to attack known targets, and posting data from related chemistry and biology experiments. Volunteers could use chat rooms and bulletin boards to announce discoveries and debate future research directions. Over time, the most dedicated and proficient volunteers would become leaders. Figure 1 The TDI Model of Online Collaboration Ten years ago, TDI would not have been feasible. The difference today is the vastly greater size and variety of chemical, biological, and medical databases; new software; and more powerful computers. Researchers can now identify promising protein targets and small sets of chemicals, including good lead compounds, using computation alone. For example, a SARS protein similar to mRNA cap-1 methyltransferases—a class of proteins with available inhibitors—was recently identified by scanning proteins encoded by the SARS genome against proteins of known structure [9] . This discovery provides an important new target for future experimental validation and iterative lead optimization. More generally, existing projects such as the University of California at San Francisco's Tropical Disease Research Unit (San Francisco, California, United States) show that even relatively modest computing, chemistry, and biology resources can deliver compounds suitable for clinical trials [10] . Increases in computing power and improved computational tools will make these methods even more powerful in the future. Just as they do today, Virtual Pharmas would choose the best candidates. The difference is that open-source drugs could not be patented in developing countries. This would not stop Virtual Pharma from developing promising discoveries. (S. Nwaka, V. Hale, personal communications). Importantly, TDI would be a great boost to the efforts of Virtual Pharmas, because it would help to contain the costs of discovering, developing, and manufacturing drugs. Cost Containment TDI would contain costs in three important ways. First, TDI would ask volunteers to donate their time (and any patentable discoveries) to the collaboration. Instead of financial incentives, TDI would offer volunteers non-monetary rewards, such as ideological satisfaction, the acquisition of new skills, enhancement of professional reputation, and the ability to advertise one's skills to potential employers. Software collaborations have demonstrated that these incentives are a good way to attract and motivate programmers [11] . Similar incentives should work equally well for biologists, chemists, and other scientists. Second, we have already pointed out that existing proposals have difficulty containing costs. The root cause is patents. Normally, society relies on competition to keep prices low. Patents—by design—short-circuit competition by giving the owners the legal right to prevent others from using (or even developing) their invention. TDI, on the other hand, would restore competition by making drug candidates available to anyone who wanted to develop them. We expect sponsors to exploit this advantage by signing development contracts with whichever company offers the lowest bid. Such competitive bidding is a powerful way to contain costs, and is also a good way to develop drugs. Virtual Pharma has extensive experience supervising contract research. Third, the absence of patents would continue to keep prices low once drugs reached the market. The generic drug industry shows what happens once drug makers are allowed to compete. US drugs frequently fall to about one-third their original price when patents expire [12] . Intellectual Property Rights Would universities and corporations really let their people volunteer? Won't they insist on intellectual property rights? The practical answer is that sensible managers do not care about intellectual property rights unless they expect to earn a profit. This explains why sophisticated university licensing offices seldom bother to interfere with open-source software projects that are not commercially valuable [13] . The same logic would apply to open-source drug discovery. We would hope that life sciences companies would make a similar calculation. But permitting employees to participate is only the beginning. We think that universities and companies will also donate the data, research tools, and other resources needed to make TDI even stronger. The reason, once again, is that they have little to lose. The value of their intellectual property depends almost entirely on US and European diseases. For this reason, it costs very little to share their information with tropical disease researchers. In fact, drug companies already do this [14] . TDI's main challenge will be to show donors that an open-source project can keep members from diverting donated information back into the commercially lucrative diseases that affect patients in the West. Finally, there are precedents for private companies developing drugs off patent. During the 1950s, March of Dimes (see www.marchofdimes.com ) developed polio vaccines without any patents at all [15] . It then signed guaranteed purchase contracts with any drug maker willing to develop commercial-scale production methods. The incentive may not have been conventional, but it worked. And why not? The contracts made good business sense: contract profits may have been small compared to the profits on patented drugs, but so was the risk. Fifty years later, contract research still makes sense. Generic drug companies, developing world drug manufacturers, contract research organizations, and biotech firms have all said that they would consider contracts to develop open-source drug candidates. (M. Spino, S. Sharma, F. Hijek, and D. Francis, personal communications). Next Steps So far, we have described a shoestring operation that exists mainly on the Web. Except for computer time, budgets would be more or less the same as existing software collaborations. Computing would be expensive but manageable. Today's biologists routinely scrounge resources from university machines or borrow time on home computers [ 16 , 17 ]. This Web-centric approach would be a good start, but not a complete solution. Computational biology works best when it can interact with experimental chemistry and biology. Nevertheless, a low-budget computational approach is probably enough to generate new science, suggest ideas for follow-up experiments, and make new drug candidates available under licenses designed to yield maximum benefit to the developing world. In practice, an open-source drug discovery effort is likely to include modest experiments. Many academic scientists control discretionary resources and, in some cases, tropical disease grants. Furthermore, good science generates its own funding. We expect experimentalists to turn the collaboration's Web pages into grant proposals. That said, TDI's volunteers will be most productive if sponsors back them. Charities could support open-source drug discovery by making wet chemistry and biology experiments a top priority. Corporations could also help by donating funds, laboratory time, or previously unpublished results. One low cost/high value option would be for companies that have already tried a particular research direction to warn TDI if the collaboration was about to investigate a known dead end. (R. Altman, personal communication) Conclusion Open-source drug discovery is feasible—that is, no known scientific or economic barrier bars the way. But what are the risks? Experience with software collaborations highlights the main social and economic challenges. First, the project will have to find and motivate volunteers. Based on existing software collaborations, we estimate a required minimum “critical mass” of a few dozen active members. Second, modest chemistry and biology experiments will be needed to increase the chances for success. Resources of several hundred thousand dollars per year—mostly in the form of in-kind donations of databases, laboratory access, and computing time—would make open-source drug discovery much more powerful. By most standards, such risks are real but acceptable. The largest uncertainties are scientific. Can a volunteer effort based on computational biology and modest experiments produce the high-quality drug candidates that Virtual Pharma needs? A successful program must (1) make a significant contribution toward supplying the genomic insights that tropical disease research needs to move forward, and (2) make useful drug candidates available for development and production under open-source licenses. Open-source drug discovery looks feasible. The only way to be sure is to do the experiment—and we invite you to join us. To learn more about TDI or to volunteer, go to http://www.tropicaldisease.org

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC539047.xml

546042

Aging and Death in E. coli

As human beings, aging and death are an inevitable part of our lives. As we pass through each decade, the concrete signs of aging—greying hair, aches and pains, the gradual failure of one organ system after another—and the realization that we are mortal increasingly occupies our thoughts. All other multicellular animals and plants also show clear signs of aging, as do some single-celled organisms. In the yeast Saccharomyces cerevisiae (baker's yeast), for example, the function of individual cells gradually declines with time, and each yeast cell has a finite life span. In organisms like this, it has been proposed that reproduction by asymmetric division is a prerequisite for aging. In other words, for a unicellular organism to age, when it divides, it must give rise to a “parent” cell and a smaller offspring cell (as in yeast), which then has to go through a juvenile phase of growth or differentiation before it divides. At each cell division, the parent cell becomes older until it reaches its natural life span and dies. A growing microcolony of E. coli But what about organisms that produce two apparently identical cells when they divide? Do such organisms age? The assumption has been for some years that cells that divide symmetrically do not age and are functionally immortal. Eric Stewart and colleagues have now tested this idea by analyzing repeated cycles of reproduction in Escherichia coli , a bacteria that reproduces without a juvenile phase and with an apparently symmetric division. E. coli is a rod-shaped organism that reproduces by dividing in the middle. Each resultant cell inherits an old end or pole and a new pole, which is made during the division. The new and the old pole contain slightly different components, so although they look the same, they are physiologically asymmetrical. At the next division, one cell inherits the old pole again (plus a brand new pole), while the other cell inherits, a not-quite-so-old pole and a new pole. Thus, Stewart and co-workers reasoned, an age in divisions can be assigned to each pole and hence to each cell. The researchers used automated time-lapse microscopy to follow all the cell divisions in 94 colonies, each grown from a single fluorescently labeled E. coli cell. In all, the researchers built up a lineage for 35,049 cells in terms of which pole—old or new—each cell had inherited at each division during its history. They found that the cells inheriting old poles had a reduced growth rate, decreased rate of offspring formation, and increased risk of dying compared with the cells inheriting new poles. Thus, although the cells produced when E. coli divide look identical, they are functionally asymmetric, and the “old pole” cell is effectively an aging parent repeatedly producing rejuvenated offspring. Stewart and his colleagues conclude that no life strategy is immune to the effects of aging and suggest that this may be because immortality is too costly or is mechanistically impossible. This may be bad news for people who had hoped that advances in science might eventually lead to human immortality. Nevertheless, E. coli should now provide an excellent genetic platform for the study of the fundamental mechanisms of cellular aging and so could provide information that might ameliorate some of the unpleasantness of the human aging process.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC546042.xml

529268

A high prevalence of cumulative trauma disorders in Iranian instrumentalists

Background Cumulative trauma disorders (CTDs) are common in musicians and their prevalence has been the subject of a number of studies in most western countries. Such studies are scarce in developing countries despite the possibility that CTDs may have a different prevalence in these countries, especially when considering traditional musical instruments and different methods of playing. Although not formally studied before, according to our experience the prevalence of CTDs seemed to be high among Iranian instrumentalists. We proposed this study to determine the prevalence of CTDs in amateur music students playing one of the two traditional Iranian instruments: Daf and Setar. Methods In a prospective cross sectional study, we interviewed and examined the students of three music training centers in Iran. Seventy eight instrumentalists, who were playing Daf or Setar and twelve students who had not started playing yet were regarded as case and control groups respectively. Some of them also underwent electrodiagnostic studies. Results Forty-seven percent (17 of 36) of the Setar players and 57% (24 of 42) of the Daf players and fifty-three percent (41 of 78) of the instrumentalists as a whole had CTDs. None of them had carpal tunnel syndrome. Conclusions Our study revealed that the prevalence of CTDs in Iranian instrumentalists was unusually high. In addition to age, other variables may be contributory. This needs to be further studied.

Background Cumulative trauma disorders, also called repetitive stress injuries, overuse syndromes or repetitive motion injuries [ 1 - 3 ] are common in musicians [ 4 , 5 ] and are caused by repetitive motions. Nerve entrapments, stress fractures, tendonitis, bursitis and muscle strains have been labeled in this category [ 1 , 5 ]. To date no study has been performed about the prevalence of cumulative trauma disorders (CTDs) in players of Iranian instruments. According to our experience, it seemed to be unusually high when compared with related prevalence in nonprofessional players of classical instruments as reported by Fry [ 4 ]. This study was performed to determine the prevalence of cumulative trauma disorders in amateur music students playing two traditional Iranian instruments: Daf and Setar. Daf is a percussion musical instrument that has a circular wooden frame covered with goat skin with or without metal discs around its edge. To play Daf the player shakes it and hits it with both hands (fig. 1 ). Setar is a string musical instrument that has 4 strings. It is played by the index of right hand (fig. 2 ). In comparison to classic musical instruments, Setar resembles the Guitar, but Daf doesn't have any similar equivalent. Methods In a prospective cross sectional study, we interviewed and examined the students of three music training centers, numbering 94. Twelve students who were at their first sessions and hadn't begun to play were selected as control group. Age, sex and duration of playing (date of starting and daily playing time) as well as vocational and avocational risk factors for developing CTDs were recorded after a direct interview. Then the students were referred to a physician who did not know whether the student belonged to the case or control group. He then evaluated their upper limbs and necks. Specific attention was paid to pain, paresthesia, sensory changes, tenderness, range of motion, muscle power and muscle stretch reflexes. In addition, Phalen, Tinel and carpal compression tests [ 6 ] were performed to detect the presence of carpal tunnel syndrome (CTS); the most common neuropathy reported in instrumentalists [ 5 ]. Since the standard diagnostic test for CTS is electrodiagnostic study [ 7 ], all of the students were asked to attend our center for electrodiagnostic studies. In all of the participants, antidromic median sensory nerve action potential (SNAP) was obtained from the third digit at both 7 and 14 cm. Then the split times and amplitudes were compared. Also distal latency for the motor median nerve was obtained. We also compared the wrist versus midpalm compound muscle action potential (CMAP) amplitudes [ 8 - 10 ]. Electromyographic investigation was not performed. The data were analyzed by SPSS software using Chi square and Fisher's exact tests. Results Ninety four students were included in this study. Four students were excluded from the study, two because of a history of musculoskeletal pain before attending the music center, one because of playing two instruments and one, serving as a typist. Twelve of the students who had not started to play were assigned to the control group and the remaining 78 students; 42 in Daf and 36 in Setar groups; were considered as case group (table 1 ). Mean age of students in the case group was 21.2 years (SD: 3.8) including 47 females and 31 males. Mean duration of instrument playing in this group was 7.9 months (SD: 5.4). Mean age of students in control group was 25.2 years (SD: 9.2). This group consisted of 9 females and 3 males. Mean Duration of daily playing in Setar students was more than Daf students (1.6 Vs 1.5 hours) which was not statistically significant (P value = 0.8). Mean duration of daily playing in male and female students was 1.8 hours and 1.4 hours respectively. Which was not statistically significant (P value = 0.64). Forty-one students in case group (53% of the total of 78) had musculoskeletal pain and there was a significant correlation between playing Daf and Setar and development of musculoskeletal symptoms. The prevalence of pain among females was twice as much as males but the difference was not statistically significant (P value = 0.12) (table 2 ). The prevalence of musculoskeletal pain in Daf players was more than Setar players (57%vs 47%, P value = 0.38); again, this difference was not significant (table 2 ). Regarding the location of pain, hand was the most common site; it was painful in 65% of cases (table 3 ). Twenty six students, all from the case group attended our electrodiagnostic center, none of them had carpal tunnel syndrome. None of the students in control group had problems in their exams and none of them attended for electrodiagnostic studies. Discussion A large number of amateur Daf and Setar players with a history of playing of less than 1 year (7.9 months) and almost 1.5 hours a day had musculoskeletal pain (that is considered a form of CTDs). The prevalence of pain in this group was much greater than students in tertiary music schools who train for some years for 6 hours a day (53% vs. 9.3–21%, respectively) and almost equals professional orchestra players (73–75%) [ 5 ]. In a group of instrumentalists (Guitarists, Harpists, Pianists etc.) Bejjani et al found a 77.5% prevalence of upper extremity disorders serious enough to impair the performance or to cause the musician to stop playing at least temporarily [ 11 ]. similarly, it is possible that some of the students also had quit playing before they had chance to enter our study (case selection bias) this might have caused an underestimation of the prevalence of the CTDs observed in this study. So it may be reasonably concluded that 53% is the minimum prevalence of CTDs in the studied group. How can we explain this high prevalence? The most important cause of CTDs is repetitive motions and in fact multiplication of duration and intensity of exercise [ 4 ]. Since both the duration and the intensity of exercise in these players were much less than that of professional music students or music trainees, other factors should be considered. According to Fry [ 4 ], other important factors that predispose to CTDs are genetics and student technique. Since the students were taught in certified centers and the music teachers were satisfied with the students' techniques, we assumed that playing method was not of primary concern. The particular instrument has been shown to be a risk factor for developing CTDs [ 5 , 14 ]. On the other hand, Setar is not heavier than guitar, nor does its playing need awkward positions, so the instrument in itself may not explain this high prevalence. Another risk factor is age [ 12 ]. It has been shown that adults who start playing, may be more vulnerable to developing CTDs. The extent to which this may have affected the results of our study is not clear so we additionally proposed that the studied group might have been inherently susceptible to develop CTDs because of some genetic factors such as joint hypermobility. This hypothesis can explain, at least in part, the wide range (9–49%) [ 5 - 12 ] of the prevalence of CTDs in music students by different studies. However, genetic analyses and larger studies are needed for validating this hypothesis. As mentioned, we did not find any case of symptomatic carpal tunnel syndrome or nerve conduction abnormalities suggesting subclinical median neuropathy at the wrist, implying that CTS is a more advanced form of cumulative trauma disorders when compared to musculotendinous unit CTDs. Hand was the most common painful site in this study (65%); a finding which is in concert with other published studies (41–54 %)[ 4 , 11 ]. There are two other findings in the current study left to be explained: First: Daf is played being held using both hands but Setar is being held like Guitar. So it can be postulated that playing Daf is more harmful than Setar and we expected more CTDs in Daf players. Although, the prevalence of CTDs was higher in Daf players, the difference was not significant (p value = 0.38). A significant difference may be found with a large scale study. Second: it has been known that CTDs are more common in females [ 5 , 12 , 13 ]. In our study, we also found a higher occurrence of CTDs among females but the study failed to reveal a statistically significant difference (p value = 0.12), perhaps because of small sample size. Conclusions Our study revealed that the prevalence of CTDs in Iranian instrumentalists was abnormally high. This is an unusual finding that can't be fully explained by the difference in the instruments (classical versus traditional), playing method or intensity of the exercise. Other susceptibility factors such as age at the starting of playing or genetic predisposition may be contributory. Larger studies focusing on individual characteristics and genetic analyses are needed to delineate other important factors. Competing interests The authors declare that they have no competing interests. Abbreviations CTD: cumulative trauma disorders CTS: carpal tunnel syndrome Authors' contributions SS: suggesting the proposal, examining the volunteers, writing the paper. BK & SMJS: examining the volunteers AB: great help in writing the paper and statistical analysis PJ: statistical analysis Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC529268.xml

514894

Towards the development of a DNA-sequence based approach to serotyping of Salmonella enterica

Background The fliC and fljB genes in Salmonella code for the phase 1 (H1) and phase 2 (H2) flagellin respectively, the rfb cluster encodes the majority of enzymes for polysaccharide (O) antigen biosynthesis, together they determine the antigenic profile by which Salmonella are identified. Sequencing and characterisation of fliC was performed in the development of a molecular serotyping technique. Results FliC sequencing of 106 strains revealed two groups; the g-complex included those exhibiting "g" or "m,t" antigenic factors, and the non-g strains which formed a second more diverse group. Variation in fliC was characterised and sero-specific motifs identified. Furthermore, it was possible to identify differences in certain H antigens that are not detected by traditional serotyping. A rapid short sequencing assay was developed to target serotype-specific sequence motifs in fliC . The assay was evaluated for identification of H1 antigens with a panel of 55 strains. Conclusion FliC sequences were obtained for more than 100 strains comprising 29 different H1 alleles. Unique pyrosequencing profiles corresponding to the H1 component of the serotype were generated reproducibly for the 23 alleles represented in the evaluation panel. Short read sequence assays can now be used to identify fliC alleles in approximately 97% of the 50 medically most important Salmonella in England and Wales. Capability for high throughput testing and automation give these assays considerable advantages over traditional methods.

Background Salmonella express flagellar (H), polysaccharide (O) and capsular (Vi) antigens which determine strain pathogenicity and therefore variation of these antigens has formed the basis for Salmonella serotyping. The Kauffmann-White scheme, first published in 1929, divides Salmonella into more than 2500 serotypes according to their antigenic formulae. Within these, 46 O antigen groups are recognised by Salmonella serotyping. O antigen synthesis and assembly is encoded by the rfb gene cluster which typically contains 12 open reading frames, and ranges in size between serotypes, from approximately 8 kbp to 23 kbp. The variation of O antigens is not due to individual gene sequence variation, but rather to different sets of genes [ 1 ]. Approximately 20,000 repeating flagellin proteins polymerise to form the flagellar filament. The ends of the protein are conserved and responsible for the hairpin shape of the subunit while variation in the central region generates the antigenic diversity. Most serotypes exhibit diphasic flagellar antigen expression by alternately expressing two genes, fliC (phase 1) and fljB (phase 2) which encode flagellins of different antigenicity. Salmonella serotyping methods recognise 63 distinct phase 1 flagellar antigenic factors and 37 phase 2 flagellar antigenic factors although the latter are not always present. Some antigenic factors, denoted by square brackets in formulae, may be present or absent without affecting serotype designation. Serotyping methods are stable, reproducible and have high typeability, yet there are several drawbacks, particularly the dependence on availability of antisera considering the ethics, cost and quality control measures necessary to maintain such a supply. Pulsed-field gel electrophoresis (PFGE) [ 2 , 3 ] is currently the bench-mark for molecular subtyping of Salmonella , however it is best used in combination with plasmid profiling and ribotyping for strain discrimination for epidemiological purposes [ 4 ]. Other approaches include fluorescent amplified fragment length polymorphism (FAFLP) [ 5 ] and multi-locus enzyme electrophoresis (MLEE) [ 6 ] which sample genomic DNA and provide a view of genetic diversity between strains and partially group some serotypes, but on the whole do not group or identify serotypes. Multi-locus sequence typing (MLST) has been used to discriminate between Salmonella strains by sampling variation in a set of housekeeping genes which precludes antigen encoding genes [ 7 ]. In 1993, Luk et al [ 8 ] published a length heterogeneity PCR (LH-PCR)-based method that targeted genes only associated with particular O antigens (A, B, C2 and D), while a more recent study by Fitzgerald et al [ 9 ] developed a serotype specific PCR assay targeting a single O serotype (O:6,14). Several studies have used a molecular approach to discriminate between particular flagellar serotypes (9, 11–12). FliC fragment restriction patterns using a dual enzyme combination allowed differentiation of flagellar types b, i, d, j, l,v, and z 10 but r and e,h nor [f],g,m, [p], g,p, and g,m,s could be separated using this technique [ 10 ]. Hong used restriction fragment patterns of fliC and fljB for serotyping of poultry Salmonella but could not distinguish S . Enteritidis from S . Gallinarum and S . Dublin [ 11 ]. Design of a multiplex-polymerase chain reaction (multiplex-PCR) to identify 1,2, 1,5, 1,6, 1,7, 1,w, e,n,x and e,n,z 15 second-phase antigens has been reported [ 12 ]. Peters and Threlfall reported fliC restriction fragment length polymorphism (RFLP) profiles were not specific enough to differentiate between certain serotypes [ 13 ]. To date no studies have attempted a universal molecular serotyping approach. Relevant publicly available sequence data is incomplete, as is epitope mapping information about specific serotypes, therefore approaches are currently being explored to characterise the expressed antigen or the encoding genes as an alternative to traditional serotyping. For fliC , evidence from antibody binding studies suggests that sequences of ~300 nucleotides of the central variable region of flagellin correlate with serotype [ 14 ] and differences in amino acid sequence can be associated with differences in antigenic specificity. Comparative sequencing has distinguished some salmonella serotypes or biotypes [ 15 - 18 ]. Previous studies have provided full gene sequence for 19 phase 1 flagellar types. The need for a robust single molecular technology to discriminate different serotypes is clear, however sequence data representing all 63 recognised phase 1 flagellar types is incomplete. The aim of this study was to generate full gene sequences for representatives of the majority of phase 1 flagellar serotypes with a view to identifying serotype-specific motifs. These were then used to design a short sequence- or single nucleotide polymorphism-(SNP) based assay targeting characteristic motifs using pyrosequencing. This technology is based on sequencing by synthesis; four nucleotides are added step-wise to a primer-template mix. Incorporation of a nucleotide i.e. extension of the DNA strand, leads to an enzymatic reaction resulting in a light flash. A pyrogram is produced from which the template DNA sequence is deduced. The assay was validated on a panel of 55 strains to initiate a DNA sequence based approach for serotyping Salmonella enterica . Results and discussion Alignment of 106 fliC sequences generated in this study and 32 phase 1 flagellin sequences previously published (see Methods section), representing 35 phase 1 flagellar serotypes revealed a clear division of sequences into two groups. Representative sequences are aligned in Additional file 1 . A tree indicating the relatedness of these sequences generated from translated DNA sequence supported this division with a 100% bootstrap value (Figure 1 ). Sequences encoding phase 1 flagellar antigens exhibiting antigenic factors "g" or "m,t" are referred to as members of the g-complex and the fliC sequences of this group clustered exclusively with the non-motile strains Gallinarum and Pullorum on the tree (Cluster I, Figure 1 ). The level of amino acid sequence homology within Cluster I sequences was 90.05%. Sequences not encoding the antigenic factors "g" or "m,t", formed the second group of sequences (Cluster II), referred to here as the non-g complex. Lower levels (80.3 %) of amino acid similarity were observed within Cluster II. Sero-specific polymorphisms were identified within the central variable region where consensus sequences of Cluster I and Cluster II diverged, between amino acid positions 160 – 407 (based on amino acid numbering system of the sequenced strain of S . Typhimurium (AE008787) represented here as sequence type Typhimurium_a). Figure 1 Protein distance tree of H1 antigens. The tree displays the inferred amino acid sequence distances between full H1 antigens. The label displays the H1 antigen, and the Salmonella serotype or subgroup from which the sequence was obtained. Esherichia coli fliC - H7 sequence was used to root the tree. Bootstrap values are displayed at major nodes. Sequences labelled with _a, _b or _c indicate an H1 allele found to be encoded by multiple sequences (Additional file 2). Salmonella fliC sequences were conserved at their termini and variable in the central region between serotypes [ 16 , 18 ] and clustered according to allele. Amino acid and nucleotide positions described here-in are with reference to the sequenced strain LT2. It was apparent from the alignment of sequences generated in this study that two assays were required, one encompassing Cluster I strains and one for Cluster II. Multiple alignments were created for each cluster and regions of the fliC gene containing sero-specific polymorphisms were identified at nucleotide positions 917 – 933 and 739–749 in Cluster I and Cluster II respectively (Figures 3 and 4 ). PCR primers were designed to amplify the target region in each sequence (see below). One multiplex PCR was developed for each group containing a mixture of specific primers. All primers designed for short sequence assays in this study are shown in Additional file 3 and the testing algorithm is shown in Figure 5 . Figure 3 Sequence motifs at target g. The assay for g-complex strains detected 17 bp of sequence commencing at nucleotide position 917. Fifteen sequence types were identified and differentiated between H1 serotypes. Figure 4 Sequence motifs at target non-g. The assay for non-g complex strains detected 9 bp of sequence commencing at nucleotide position 739. Sixteen sequence types were identified among non-g complex strains tested. Figure 5 Algorithm for identification of unknown isolates. Times given are approximate for 96 samples using methods described. Summary of fliC sequence variation within the g-complex All polymorphisms within the g-complex sequences analysed are displayed in Figure 2 The target region (highlighted) was selected because it conferred multiple sero-specific amino acid substitutions and was variable at the DNA level. In the 17 bp nucleotide sequence assayed, 15 sequence types were identified (Figure 3 ). This region was assayed against the test panel of 17 Salmonella strains belonging to the g-complex and was able to exclusively identify sequence motifs corresponding to phase 1 flagellar serotypes. The serotypes not differentiated by this assay ([f],g,m, [p], g,m, g,m,s and g,m, [p],s or non-motile Gallinarum) were known from full sequencing to be identical at the target region. Figure 2 Amino acid polymorphisms among fliC of g-complex strains. Alignment displaying polymorphic codons only of g-complex fliC genes. Codon numbering is based on LT2 sequence, a slash is used where codons fall between LT2 codons in alignment. Highlighted area indicates region analysed in g-complex assay. Amino acid differences between g-complex strains identified by full sequencing The following polymorphisms located in fliC of the g-complex are likely to be involved in specific epitope formation: two amino acid sequence types were observed in 25 fliC-[f],g,m, [p] sequences obtained from Salmonella enterica serovar Enteritidis strains. Twenty-three S . Enteritidis strains demonstrated complete conservation in their DNA sequence (B16, B18, JTCM02 and 20 phage type 4 strains (Enteritidis_b)). The sequence of B17 was congruent with published S . Enteritidis (M84980) (Enteritidis_a), and exhibited a single amino acid (Ser>Gly at 302) substitution compared to sequence type Enteritidis_b. Published S . Othmarschen (U06455) fliC-g,m , [t] sequence inferred the same amino acid sequence as Enteritidis_a but exhibited a silent mutation at the DNA level. As the fliC sequence for these two serotypes was identical it was apparent that the sequence included here represented an S . Othmarschen strain in which the t factor was absent. Published S . Gallinarum sequences demonstrated 100% DNA homology to Enteritidis_b except for a SNP encoding a stop codon in M84975. S . Pullorum and S . Gallinarum are non-motile as they do not express flagella. Antisera to the g factor antigen react strongly with induced-motility S . Pullorum culture, indicating that g epitopes are expressed in these cells [ 19 ]. This correlates with our sequence data as S. Pullorum clusters with g,m sequences (Figure 1 ). Biotype-specific polymorphisms for S . Pullorum were observed at amino acid position 91 and 323. Molecular identification of S . Pullorum and S . Gallinarum would be of considerable benefit as standard serotyping cannot differentiate these two serotypes. FliC-g,q was differentiated from all other g-complex sequences by an Asp>Gly serotype-specific polymorphism observed at position 284 for S . Moscow. A Thr>Ala substitution at residue 304 conferred by a single nucleotide polymorphism (SNP) was identified between sequences of g,m and g,p, congruent with a previous report [ 20 ], and forms the basis for differentiation of these two serotypes. DNA polymorphisms, but no inferred amino acid substitutions, were observed between strains exhibiting g,m,s and g,m, [p],s. The p factor was not coded for by the fliC sequences of these strains. S . Essen fliC-g,m was distinct from other g and m coding sequences by an Asp>Asn substitution at 283. fliC-g,p,s could be differentiated from fliC-g,p by a Thr>Ala substitution at 254. A motif of two amino acids at positions 302 and 307 was common to S . Derby, S . Agona, S . Adelaide, and S . Berta which exhibit phase 1 flagellar antigenic factors "f" and "g". This motif was exclusive to these serotypes. DNA sequence variation at corresponding positions allowed S . Derby and S . Agona to be distinguished from S . Adelaide and S . Berta. FliC-g,z 51 ; and fliC-m,t with fliC-g,m,t each form distinct clusters (Figure 1 ). Summary of fliC sequence variation within the non-g complex Sequence conservation within alleles that did not encode g or m,t antigenic factors was demonstrated by 97.8 – 99.1% homology and 80.35% homology was measured in the complex. The high level of variability between alleles in this group did not allow association of specific amino acids to epitope formation that was possible with the g-complex sequences. The quantity and distribution of polymorphic bases observed in this group (specified below) meant that there was a choice of regions that could be used for differentiation. Following testing of four possible regions, the region encompassing amino acids 248–250 was selected for use in the final non-g assay. Each serotype had a unique motif at the target region except fliC-l,v and fliC-l,z 13 which shared a sequence type (Figure 4 ). Some amino acid sequences were not identical within non-g alleles, including i, r, d, e,h, a and z 4 ,z 23 ( Additional file 1 ). A previous study of fliC-i sequences reported no variation in a 260 bp region among seven Typhimurium strains [ 17 ]. Six full S . Typhimurium fliC s and a fragment spanning nucleotides 434–1090, corresponding to amino acids 159 – 400, of a further 20 S . Typhimurium strains were sequenced. Three distinct DNA sequences which resulted in translated differences in the expressed peptides were observed within the serotype. Sequence type "Typhimurium_a" was detected in 18 strains, identical to the sequenced strain LT2. Sequence type "Typhimurium_b" was detected in four strains and was differentiated by a SNP at 768, conferring a 256 Glu>Lys substitution. Sequence type "Typhimurium_c" conferred a Glu>Lys substitution and an Ala>Thr substitution at 263 and was found in two strains: 571896 and 571913. Strains 571896 and 571913 were phage type DT104 however, other strains tested did not conform to recognised phage typing patterns so no assured correlation could be made with phage type or other phenotype. S . Choleraesuis sequence ( fliC-c ) differed from that published (AF159459) at one nucleotide, conferring amino acid substitution of Thr >Ser at codon 99. FliC sequences of nine S . Heidelberg strains were identical, consistent with the results of a previous report [ 18 ]. The published sequence for fliC-r of S . Rubislaw (X04505) differed from S . Heidelberg at three amino acids. The S . Muenchen sequence determined in this study differs in twelve amino acids to the published S . Muenchen (X03395), and differed in 25 amino acids from the S . Duisberg sequence in this study. S . Anatum, S . Newport and S . Saintpaul exhibit factors e,h in their phase 1 flagellar. Amino acid sequence was conserved in two strains of S . Saintpaul but distinct for each serotype due to four amino acid substitutions at codons 192, 213, 238, 356. S . Brandenburg and S . Panama exhibit l,v in the phase 1 antigen, no inferred amino acid differences were detected. FliC-l,v sequences clustered with fliC-l,z 13 (Figure 1 ). FliC from three strains exhibiting the z 4 antigenic factor in phase 1 flagellar were sequenced. Cluster analysis grouped these sequences together in the non-g group although they contain regions of sequence similar to g-complex strains (amino acid positions 96 – 164). Z 4 ,z 24 is distinct from z 4 ,z 23 and z 4 ,z 23 sequences varied within the serotype at seven amino acid positions: 235, 237, 239, 242, 253, 351 and 369. The complex mosaic nature of fliC is evident from analysis of amino acid alignment of sequences in particular strains from subgroups in the SARC collection (see Materials and Methods). Molecular serotyping assays By comparison of amino acid sequences coding for antigens of the different serotypes, sero-specific motifs were identified. Individual regions of fliC were selected for the g-group and non-g group to provide unique sequence for as many serotypes as possible, while keeping the assay simple to perform and analyse. Two multiplex PCRs were developed for the production of fliC amplicon of g-complex strains and fliC amplicon of non-g strains. Sero-specific motifs in each amplicon were consequently identified by sequencing-by-synthesis. G-complex assay Fifteen sequence types were identified in the 17 bp of nucleotide sequence assayed (Figure 3 ). Twenty-seven strains were tested and each produced a recognised sequence motif which differentiated between serotypes. Serotypes would be fully resolved through the detection of further polymorphisms, for example g, [s],t and g,t can be separated through additional detection of a A>G change at nucleotide position 777 conferring amino acid Ser>Gly substitution specific to g,t. Non-g assay Fourteen sequence types were identified in the 9 bp of nucleotide sequence assayed (Figure 4 ). Thirty strains were tested, each producing a recognised sequence motif allowing separation of serotypes. Serotypes l,v and l,z 13 gave the same motif at the target region but could be separated by nucleotide substitution A>G at position 783 conferring a Thr>Ala change. The stability of the targeted polymorphisms in Salmonella phase 1 flagellar antigens was demonstrated through testing on a panel of 55 isolates. The SNP responsible for the antigenic difference between serotypes g,m and g,p was within the target region and so could be differentiated by the assay. The amino acid substitution that separated fliC-g,p,u was also encoded within the sequence assayed. Antigens i, r, c, d, b, e,h, k, a, z 41 , z, z 10 , z 4 ,z 23 , z 4 z 24 , g,q, g,m,p, g,p,u, [f],g,t, g,z 51 and biotype S . Pullorum gave unique motifs, l,v and l,z 13 shared a motif. Some serotypes for which certain factors may be present or absent (denoted by square brackets in antigenic formulae) were not separated from similar serotypes: [f],g,m, [p], g,m and g,m, [p],s; [f],g,m, [p] and g,m, [t]; g, [s],t and g,t although these could be separated by other DNA polymorphisms as discussed. Two motifs were observed for k, each specific to S . Thompson and IIIb. Two motifs were observed for d, specific to S . Duisberg and S . Muenchen / S . Schwarzengrund. Published sequence data for fliC-m,t , from serotypes S . Banana, S . Oranienburg and S . Pensacola were included in assay design. The polymorphic region targeted by the assay is predicted to differentiate m,t sequences from other g-complex antigens, and also differentiate S . Pensacola from S . Banana and S . Oranienberg. Strains exhibiting factors m,t were not available for testing. Conclusions A high level of sequence homology between fliC genes of g-complex strains was observed. Data produced for this study is congruent with a previous report of g-complex sequences [ 16 ]. The genetic basis between distinct antigens in this group of sequences can be a single amino acid substitution. Specific motifs could be identified as the genetic basis for particular antigenic differences and hence their involvement in epitope formation and stability among strains inferred. Full gene sequences were distinct for each antigen analysed in this study. Furthermore, analysis of multiple representatives revealed that some antigens were encoded for by multiple sequences. In these cases DNA sequence based methods are more discriminatory than traditional serotyping methods which do not recognise these as distinct antigens. Assays were designed such that an unknown strain could be identified in respect of its phase 1 flagellar antigen in two steps. The specific PCR acted as the first level of identification and the resultant amplicon was used for the pyrosequencing assay. A positive PCR indicated which of two Pyrosequencing assays to apply. Each assay was uniform in that only one mix of pyrosequencing primers and one dispensation order was needed. All the strains tested were successfully amplified by PCR. As some analyses have been performed on unpublished data, exhaustive testing of the assay will be performed to confirm specificity and typeability of all recognised serotypes. Molecular serotyping will incorporate the desirable properties of serology (typeability, reproducibility, epidemiological significance) together with the advantages of DNA analysis (ability to automate, labour saving, serum independent). Antisera production and associated quality control measures would be unnecessary for a DNA sequence based method. Time-consuming flagellar phase reversal to identify both flagellar antigens is not necessary at the genetic level. Other advantages include reduced labour costs, rapid results in comparison to traditional serotyping methods. DNA sequence data is highly portable and easy to interpret. The method described was easily automated by use of the vacuum preparation tool for the DNA strand separation step and could be further automated by use of robotics for PCR set-up. Result output included a pyrogram, raw text and confidence level; automation of data analysis could be achieved by use of a computer script to screen at a set confidence level and cross-check results against a database of recognised motifs. With the capability to identify approximately 97% of phase 1 flagellar antigens from medically important Salmonella strains occurring in England and Wales, the assay can be used now as an economic screen of unknown isolates and alleviate the burden on routine serotyping work. A scheme including the phase 1 flagellar assay and complementary assays for phase 2 flagellar and polysaccharide antigens is currently being piloted and based on incidence data of the top 50 serotypes from 2003, it is anticipated that the scheme will provide a complete molecular serotype for around 80% of isolates and confident prediction of 76% of the remainder. Future work Alternative sero-specific polymorphisms identified in this study could be exploited by similar assays to allow further separation when antigens did not give unique pyrograms. The alliance of the fliC assay to a fljB and rfb assay would allow the full antigenic formulae of Salmonella serotypes to be determined. Common phase 2 flagellar antigens will be selected for sequencing and together with published data will be analysed for sero-specific motifs and a short sequence assay designed with the approach described in this study. In 1993, Luk et a l [ 8 ] outlined a simple length heterogeneity PCR for identification of Salmonella major serogroups A, B, C2, and D. They based their PCR on the presence/absence of genes or sequence polymorphisms within shared genes. Essentially, only serogroups A and D possess a gene to synthesise tyvelose but serogroup A genes carry an early stop codon and do not produce the sugar itself. Only groups B and C2 possess a gene to synthesise abequose but the sequences are distinct. We have also designed a preliminary pyrosequencing assay to distinguish these serogroups based on amplification and short sequences of these genes (data not shown). In summary, epitopes are conformational and it is difficult to determine which amino acids would interact from a linear sequence. However, in the g-complex sequences some amino acid changes could be identified as responsible for differences in antigenic factors because variation was minimal. There is no common factor among the non-g antigens and the sequences are much more heterogenous; there are too many substitutions to draw conclusions about epitope specific sequences. Epitope mapping could be used to further investigate epitopes responsible for antigenic specificity. Methods Bacterial strains Strains exhibiting the different phase 1 flagellar antigenic factors were selected from Salmonella Reference Collections A, B and C obtained from the University of Calgary. Multiple isolates of S . Enteritidis phage type 4, and S . Typhimurium phage type DT104 plus a panel of serotyped strains were gratefully received from the Salmonella Reference Laboratory, Health Protection Agency, Colindale ( Additional file 2 ). DNA preparation, PCR and sequencing MagNA Pure instrument and Total Nucleic Acid Extraction Kit 1 (Roche, East Sussex). PCR reactions contained 1X PCR buffer, 20 pmoles of FL_START2, 20 pmoles rFSa1 [ 21 ], 1 U Taq polymerase, 0.25 mM of each dNTP, 4 mM MgCl 2 (Sigma-Aldrich, Dorset). PCR amplification of the fliC gene was performed with an 9700 GeneAmp PCR System (Applied Biosystems, Cheshire): 35 cycles of 95°C for 60 sec, 50°C for 60 sec, 72°C for 30 sec followed by a 7 min final extension at 72°C. PCR products were purified with Qiaquick spin columns (Qiagen Ltd, West Sussex) and quantitated by gel electrophoresis using Ready-to-Run pre-cast gels (Amersham Biosciences, Buckinghamshire). Fifty to one-hundred nanograms of the purified PCR product was used for cycle sequencing, with specific primers ( Additional file 3 ) and the CEQ DTCS dye terminator kit (Beckman Coulter, Buckinghamshire). Excess dNTPs were removed from sequencing reactions using GenClean, a 96-well plate format gel filtration system (Genetix Ltd, Hampshire). Sequencing reactions were run on a CEQ 8000XL capillary sequencer (Beckman Coulter). Primers were designed on generated sequence aided by Eprimer3 [ 22 ] in a primer walking approach to complete sequencing of the full gene. Sequences generated have been submitted to GenBank (Accession numbers AY649696-AY6497242). Sequence analysis Data were analysed and assembled using SeqMan, a component of the DNA Star software package. Multiple alignments were created using BioEdit (Tom Hall, North Carolina State University). Phylogeny inference package Phylip (Joe Felsenstein, University of Washington) was used to compute a distance matrix from protein sequences and build trees illustrating the relatedness of fliC sequences. Some previously published sequences were included ( Additional file 3 ). Polymorphisms postulated to be serotype specific were identified from the alignments of full fliC sequences; in-house programme MOP-UPs [ 23 ] identified motifs and designed primers to user-specified groups of sequences in the alignment (Anthony Underwood, Health Protection Agency, London). Assay design Two multiplex PCRs were designed to amplify polymorphic regions of both g- and non-g complex fliC sequences. Amplicon sizes were approximately 316 bp for g-complex strains, 170 bp – 250 bp, (size varied according to serotype) from non-g strains. The order of nucleotide dispensation was tailored to enable the first two dispensations to act as negative and positive controls. DNA preparation and Pyrosequencing A 1 μl loop of cells was boiled in 100 μl of sterile distilled water for 10 min at 95°C. One microlitre of lysate was used for PCR. Two PCRs were run in parallel to amplify fragments of the fliC gene. Fifty microlitres of PCR product was prepared containing 1 U Taq polymerase, 0.25 mM of each dNTP, 4 mM MgCl 2 . PCR for amplification of g-complex strains used three forward primers: 100 pmol GPYRO-A; 12.5 pmol GPYRO-B; and 12.5 pmol GPYRO-C; and 125 picomoles of reverse biotinylated primer G-REV. PCR for amplification of non-g strains used 14 forward primers (NON-G-PYRO-A, NON-G-PYRO-B etc.) in equal concentrations. Thirteen biotinylated reverse primers (NON-G-REV-A, NON-G-REV-B etc.) were used in equal concentrations. In each 50 μl reaction 125 pmol of mixed forward primer and 125 pmol mixed reverse primer was used. Primer sequences are detailed in Additional file 3 . Thermocycling was performed with an Applied Biosystems 9700 GeneAmp PCR System using a touch-down programme: initial denaturation step of 94°C for 2 min; followed by 17 cycles of 94°C for 20 sec, 66°C (-1°C per cycle) for 30 sec, 72°C for 30 sec; followed by 20 cycles of 94°C for 20 sec, 54°C for 30 sec, 72°C for 30 sec. The excess primers were removed using a filter plate and vacuum system (Genetix Ltd, Hampshire) before visualising the PCR products on the Ready-To-Run agarose system (as previously). Biotinylated single-stranded DNA was immobilized on streptavidin-coated sepharose beads (Amersham Biosciences, Buckinghamshire) with binding buffer. The mixture was agitated at 1400 rpm for 10 min at room temperature. Single stranded DNA bound to beads was isolated from the mixture using a series of wash steps for 5 seconds each in turn, 70% ethanol, 0.2M NaOH and washing buffer. Ninety-six samples were prepared in 2 minutes by automation of strand separation step using a vacuum preparation tool (Pyrosequencing AB, Uppsala, Sweden). A combination of pyrosequencing primers was used for each assay; 20 primers for the non-g complex assay, and three for the g-complex assay ( Additional file 3 ) into which DNA was eluted. Pyrosequencing primers were annealed to single-stranded DNA on the beads by heating to 80°C for 2 minutes and allowed to cool slowly. Single stranded binding protein, enzyme mix, substrate mix and dNTPs (Pyrosequencing) were added sequentially by the instrument according to the programmed dispensation order. Authors' contributions CM carried out the sequencing, constructed the multiple alignments and designed the assays. TP carried out the serotyping. TP advised on Salmonella serotyping and provided strains. CA CM SG TP and JL participated in the design of the study. CA conceived of and coordinated the study. All authors read and approved the final manuscript. Supplementary Material Additional File 1 Amino acid alignment. Amino acid alignment of 106 fliC gene sequences representing 32 H1 alleles. Sequences labelled with _a, _b or _c indicate an H1 allele encoded by multiple sequences (Additional file 2). Codon numbering is in reference to the sequence of Typhimurium_a which represents sequenced strain LT2. Click here for file Additional File 3 Primers used for PCR and Pyrosequencing. Orientation of the primer is represented by F (forward) or R (reverse) and approximate position is given as nucleotide distance from 5' end of fliC *These primers were also used as pyrosequencing primers Click here for file Additional File 2 Sequences used in this study. Sequences labelled with _a, _b or _c indicate an H1 allele encoded by multiple sequences Click here for file

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC514894.xml

548696

Long-term survival rates of laryngeal cancer patients treated by radiation and surgery, radiation alone, and surgery alone : studied by lognormal and Kaplan-Meier survival methods

Background Validation of the use of the lognormal model for predicting long-term survival rates using short-term follow-up data. Methods 907 cases of laryngeal cancer were treated from 1973–1977 by radiation and surgery (248), radiation alone (345), and surgery alone (314), in registries of Connecticut and Metropolitan Detroit of the SEER database, with known survival status up to 1999. Phase 1 of this study used the minimum chi-square test to assess the goodness of fit of the survival times of those who died with disease to a lognormal distribution. Phase 2 used the maximum likelihood method to estimate long-term survival rates using short-term follow-up data. In order to validate the lognormal model, the estimated long-term cancer-specific survival rates (CSSR) were compared with the values calculated by the Kaplan-Meier (KM) method using long-term data. Results The 25-year CSSR were predicted to be 72%, 68% and 65% for treatments by radiation and surgery, by radiation alone, and by surgery alone respectively, using short-term follow-up data by the lognormal model. Corresponding results calculated by the KM method were: 72+/-3%, 68+/-3% and 66+/-4% respectively. Conclusions The lognormal model was validated for the prediction of the long-term survival rates of laryngeal cancer patients treated by these different methods. The lognormal model may become a useful tool in research on outcomes.

Background Literature review [ 1 ] indicated that local control, laryngeal preservation, and survival rates of larynegeal cancer patients were similar after transoral laser resection, open partial laryngectomy, and radiotherapy. Open partial laryngectomy was reserved for patients with locally recurrent tumors. There are still some unanswered questions. Will radiation combined with surgery give a better result than single modality treatment alone? Will treatment results from the community centers follow published data from prestigious centers? After radiotherapy, radio-resistant cells theoretically may take some time to grow before recurrence. Short-term data may not reflect long-term local control and survival rates. We attempt to address these questions in the present study. The lognormal distribution is defined as the distribution of a random variable whose logarithm is normally distributed. The purpose of this study is to validate the use of the lognormal model [ 2 - 4 ] by estimating the long-term survival from short-term follow-up data of laryngeal cancer treated by three different treatment methods: radiation and surgery, radiation alone, and surgery alone. We have previously validated the application of the lognormal model for small cell lung cancer [ 5 ], glottic laryngeal cancer [ 6 ], prostate cancer [ 7 ] and breast cancer [ 8 ]. This model may be useful for randomized clinical trials because it allows the prediction of long-term survival rates several years earlier than is possible by using the standard actuarial life table/Kaplan-Meier method of calculation [ 9 ]. The idea that long-term survival rates can be estimated from short-term follow-up data is attractive because this method shortens the delay in further research to improve cancer treatment. The validation of the lognormal model has two phases. Phase 1 tests the goodness of fit to a lognormal distribution of the survival times of those cancer patients who died with disease. Phase 2 attempts to verify the lognormal model, which uses short-term follow-up data to predict long-term survival rates. These survival rates are then compared with values calculated by the Kaplan-Meier life table method from available long-term data. The second phase has been difficult to implement because of the general lack of large number of patients with sufficiently long follow-up information. With the SEER database [ 10 ], the validation of the lognormal model is now possible. Methods We analyzed a total of 907 cases of laryngeal cancer treated by three treatment methods: radiation and surgery (248 patients), radiation alone (345 patients), and surgery alone (314 patients), registered in two registries, Connecticut and Metropolitan Detroit, from 1973–1977 with known survival status up to 1999 extracted from the SEER database. Fourty-seven patients with unknown survival time and 165 patients with missing treatment methods were excluded. A table of the patient characteristics was listed in Table 1 . The cause-specific survival time was defined as the interval from the date of diagnosis to the date of death from laryngeal cancer or the date of last follow-up for censoring purposes, if the patient was alive and still being followed at the time of analysis. Patients who died of other causes were also censored at date of death. Phase 1 – Test of goodness of fit for lognormality Testing for lognormality was done separately for laryngeal cancer patients who died with disease (as distinct from those who died of an intercurrent disease) treated by: radiation and surgery, radiation alone, and surgery alone. For Phase 1, a minimum chi-square method was used to estimate the standard deviation S and mean M of the log 10 (survival time) of the distribution of patients who died of the cancer. The minimum chi-square test was run by a Microsoft Excel program. A range of S-values and M-values were tested to reduce the chi-square values to a minimum. In order to determine if the observed survival times for a given cohort are lognormally distributed, the results are given in terms of probability levels of significance P for the chi-square estimates which correspond to a minimization of chi-square and hence a maximization of P. The test statistic of the minimum chi-square test was minimized by varying the parameters and the P-value gave the significance of the test. The class intervals were in the powers of 2 in months of the survival time, such as 0–2, >2–4, >4–8, >8–16, and so on. The number of cases in each interval should not be less than 5. The null hypothesis being tested is H 0 : that there is no difference between the observed survival times and the expected survival times calculated from a lognormal distribution with a specified S and M. If P < 0.05 the null hypothesis is rejected. Phase 2 – Validation of the lognormal model A second computer program for the Phase 2 of the study was also run using Microsoft Excel to estimate the cured fraction C by using a maximum likelihood method described by Boag [ 4 ]. Using the lognormal model, the standard deviation (S) was fixed, and only the two remaining parameters, the mean (M) and the cured fraction among all patients (C), were kept floating. Multiple iterations converged to a stable solution for C. The cause-specific survival rate (CSSR) at time τ = [C + (1-C) × Q] × 100%, where Q = the integral of the lognormal distribution between the limits τ and infinity, τ = long-term survival time, C was calculated by the maximum likelihood method. The five-year cohort extended from 1973 to the end of 1977. Prediction of the long-term CSSR were made after one year follow-up, i.e. at the end of 1978, for the laryngeal cancer patients with three different treatments: radiation and surgery, radiation alone, and surgery alone. The predicted CSSR were then validated by comparing with the results calculated by the Kaplan-Meier method using the actual follow-up data available up to 1999. Log-rank tests were performed for the three different treatment groups over the whole time period in order to see if there was a difference in CSSR between treatments. Results Using data from 1973–1977, the minimum chi-square tests verified that the survival times of the patients who died of laryngeal cancer and who were treated by three treatments followed three different lognormal distributions. At minimum chi-square, the S and M values, the numbers of patients N and the P values of the minimum chi-square tests for the three different treatments are listed in Table 2 . As in a prospective trial at interim analysis, Phase 2 was performed for each of the five-year cohort periods after one year of follow-up. Phase 2 predicted, using the maximum likelihood method, the long-term 10-, 15-, 20-, and 25-year CSSR. An S value was selected for the maximum likelihood method, so that the prediction curve obtained by the lognormal model would best fit the Kaplan-Meier graph at one year after the five-year cohort period. Hence at the time of prediction by the lognormal model, the long-term data were not known yet. Table 3 lists the results for each treatment arm and their comparison with calculation obtained by the Kaplan-Meier method. The predictions were validated by comparing them with the Kaplan-Meier calculation using available data up to 1999. The 25-year CSSR were predicted by the lognormal model after only short term follow-up to be 72%, 68% and 65% for treatments by radiation and surgery, radiation alone, and surgery alone respectively. The 25-year CSSR were found to be 72+/-3% (one standard error), 68+/-3% and 66+/-4% respectively by Kaplan-Meier method. Long-term survival rates at other years, e.g. 10-, 15-, and 20-year CSSR, were all within one standard error compared with the Kaplan-Meier calculations. There were no statistically significant differences between the CSSR of the three different treatment groups for all stages combined (p = 0.35 by log-rank test for the three treatments). Figures 1 , 2 and 3 show the comparisons of the three different treatments: radiation and surgery, radiation alone, and surgery alone respectively for laryngeal cancer in the Kaplan-Meier graph at the year 1999 compared with the lognormal model prediction curve which could be obtained at 1978. The SEER database provides localized and regional disease stages, instead of T1N0 or T2N0 stages, etc. The lognormal model prediction of the 25-year CSSR for 454 patients with localized stage disease were 84%, 75%, and 80% for treatments by radiation and surgery, radiation alone, and surgery alone respectively. The 25-year CSSR calculated by the Kaplan-Meier method were 83+/-4%, 75+/-4%, and 77+/-6% respectively (p = 0.08, log-rank test for the three treatments). For 286 patients with regional stage disease, the 25-year CSSR were 62%, 58%, and 52% respectively, and compared with Kaplan-Meier method were 63+/-6%, 58+/-7%, and 46+/-6% respectively (p = 0.76, log-rank test for the three treatments). Discussion The current study demonstrates the application of the lognormal model to a population based study. The lognormal model is being applied in a manner that can be applied to prospective trials in practice. Our previous publication about the application of the lognormal model was for different stages of laryngeal cancer patients treated by one radiation oncologist, Dr. M. Lederman in the Royal Marsden Hospital, United Kingdom [ 6 ]. The current study shows that the lognormal model is applicable in different scenarios. Detroit and Connecticut registries were chosen because they have the earliest data starting from 1973 and they include a large population of both white and black patients. Generally cause-specific death rates underestimate the mortality associated with a diagnosis of the specific cancer, because some patients died of other causes[ 11 ]. Gamel et al .[ 12 ] found that the follow-up time should be one standard deviation beyond the mean of the survival time so as to obtain stable results. In the current study, five-year cohorts extended from 1973 to the end of 1977 were used. Stable results for prediction of the long-term CSSR were obtained after one year of follow-up. In this study, for localized stage disease 5-year CSSR were 91%, 83%, 95% for treatments by radiation and surgery, radiation alone, and surgery alone respectively with Kaplan-Meier method. Jones et al .[ 13 ] found that the 5-year tumor-specific survival for those treated by radiation was 87% and for those treated by surgery was 77% (p = 0.1022). Both radiation and surgery are equally effective for treating early stage laryngeal carcinoma. In the current study, there were marginal difference in the results of the three treatments for localized stage patients. Jorgensen et al .[ 14 ] found that among patients with T1 glottic carcinomas the 5-year locoregional control rate was 88%, i.e. 88% of patients were cured by radiotherapy alone. The 5-year disease-specific survival (DSS) was 99%, i.e. salvage surgery added approximately 11% to the survival of T1 glottic patients. Only 4% (12/312) of T1 glottic patients underwent laryngectomies. Locoregional control among T2 glottic patients was 67% and the DSS 88%, and 18% (41/233) of patients underwent laryngectomies. The corresponding results among T3 glottic patients were 30% and 59%, about 50% of patients underwent laryngectomies. For T3 glottic carcinomas, initial surgery did not produce better survival rates. Franchin et al .[ 15 ] studied T1 and T2 glottic carcinoma. The 5-year and 10-year overall survival rates were 83% and 63.5%, respectively. The overall 10-year local control rate for patients with T1 and T2 glottic carcinoma was 89%. In this study, the treatment results were marginally different for localized stage (p = 0.08, log-rank test for the three treatments) and the treatment results were similar for regional stage (p = 0.76, log-rank test for the three treatments). For localized stage combination radiation and surgery may have more certainty of disease control and hence long-term survival benefit. The treatment results were similar for regional stage because the patients were diagnosed late and disease control may be more difficult. The predicted survivals were within one standard error of the Kaplan-Meier estimations for both localized and regional stages. It shows that the prediction method can work for both good and poor prognosis cases. The practical value of this study is that this lognormal model may be used for prediction of the results of prospective trials earlier than the Kaplan Meier method. This lognormal model may become a useful tool in research about outcomes. Use of this lognormal model could result in more rapid advances in cancer treatment and have the potential benefit of a reduction of cost in cancer research. Competing interests The author(s) declare that they have no competing interests. Authors' constributions PT: Data analysis and writing of the manuscript. EY, RS: Critical appraisal of the manuscript. JT: Data analysis and critical appraisal of the manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548696.xml

548682

Daily antibiotic cost of nosocomial infections in a Turkish university hospital

Background Many studies associated nosocomial infections with increased hospital costs due to extra days in hospital, staff time, extra investigations and drug treatment. The cost of antibiotic treatment for these infections represents a significant part of hospital expenditure. This prospective observational study was designed to determine the daily antibiotic cost of nosocomial infections per infected adult patient in Akdeniz University Hospital. Methods All adult patients admitted to the ICUs between January 1, 2000, and June 30, 2003 who had only one nosocomial infection during their stay were included in the study. Infection sites and pathogens, antimicrobial treatment of patient and it's cost were recorded. Daily antibiotic costs were calculated per infected patient. Results Among the 8460 study patients, 817 (16.6%) developed 1407 episodes of nosocomial infection. Two hundred thirty three (2.7%) presented with only one nosocomial infection. Mean daily antibiotic cost was $89.64. Daily antibiotic cost was $99.02 for pneumonia, $94.32 for bloodstream infection, $94.31 for surgical site infection, $52.37 for urinary tract infection, and $162.35 for the other infections per patient. The treatment of Pseudomonas aeruginosa infections was the most expensive infection treated. Piperacillin-tazobactam and amikacin were the most prescribed antibiotics, and meropenem was the most expensive drug for treatment of the nosocomial infections in the ICU. Conclusions Daily antibiotic cost of nosocomial infections is an important part of extra costs that should be reduced providing rational antibiotic usage in hospitals.

Background Nosocomial infections are frequent complications of hospitalization and also an important public health problem in developing countries, as well as in developed ones. The socioeconomic impact, ie, prolongation of hospitalization, mortality, and cost of these infections adversely effects patients and nations' economic well-being [ 1 ]. The cost of nosocomial infections includes increased length of hospital stay, staff time, laboratory cultures of pathogens and antimicrobial treatment [ 2 - 4 ]. Although, cost of antimicrobial treatment is an important part of health expenditure, data on this subject are extremely limited in Turkey. The aim of our study was to determine daily antibiotic cost of nosocomial infection per infected patient in a university hospital. Methods The hospital setting Akdeniz University Hospital is a 600-bed tertiary referral centre in Antalya, Turkey, treating 27 000 patients per year. The study was conducted in six adult medical and surgical intensive care units (ICUs) with a total of 51 ICU beds. Neonatal ICU was not included in the study. Since 1993, the institutional policies of hospital infection control have been implemented by infection control team. Definitions and study population In our hospital, routine prospective, active surveillance of nosocomial infections in all ICUs is performed by one infection control nurse, supervised by an infection control physician. Nosocomial infections are defined using the Centers for Disease Control and Prevention criteria [ 5 , 6 ]. We do not follow patients for signs of infection after discharge unless they are readmitted to the hospital. Between January 1, 2000 and June 30, 2003, all inpatients hospitalized in one of the adult ICUs were included in this study. Data on antimicrobial treatment were recorded for patients aged 15 or above presenting with only one nosocomial infection. For all patients included in the study, the following were recorded: age, sex, infection site, microbiologic data, antimicrobial therapy and antibiotic cost. Measurement of costs In our ICUs, all antimicrobial prescriptions are recommended by an infectious disease consultant. Antimicrobial agents prescribed only for therapeutic indications were recorded. The daily antibiotic cost was calculated in US dollars based on June 2003 prices of antimicrobial agents provided by the hospital pharmacy. The daily antibiotic cost per infected patient was calculated by the multiplication of box price and number of daily doses that was used for that infection. Two costs were calculated for each antimicrobial; the minimal cost (min) was based on the lowest recommended parenteral daily dose and the maximal cost (max) was based on the highest recommended parenteral daily dose. Results Between January 1, 2000, and June 30, 2003, a total of 8460 patients were admitted to the adult ICUs. Overall, 817 patients developed 1407 episodes of nosocomial infections, accounting for an infection rate of 16,6%. Among them, 233 patients (mean age:50,1; sex ratio female:male 0,49) had only one nosocomial infection. Mean daily antibiotic cost was found $89,64 per infected patient ($8,56 to $359,28). Among the sites of nosocomial infections, urinary tract infections had the lowest daily antibiotic cost per infected patient (Table 1 ). The mean daily antibiotic cost for pneumonia was the highest of all sites, but patients with bloodstream infection reached the highest range of daily cost ($31,31 to $359,28). In addition, mean daily antibiotic cost was found $162,35 for seventeen other infections including postoperative meningitis, mediastinitis, and empyema. Table 1 Daily antibiotic costs according to the sites of nosocomial infections. Site of nosocomial infections Number of patients Range of daily cost per infected (US$) Mean (US$) Pneumonia 111 10,02–250,74 99,02 Bloodstream infections 28 31,31–359,28 94,32 Surgical site infections 11 17,12–204,74 94,31 Urinary tract infections 66 8,56–228,68 52,37 Out of 233 patients, 206 patients had microbiologically documented infections, 177 patients were infected by a single pathogen while 29 patients were diagnosed as having polymicrobial infections (Table 2 ). Pseudomonas aeruginosa was the most prevalent bacteria followed by Klebsiella spp. and Acinetobacter spp.. P. aeruginosa infections had the highest overall daily antibiotic cost per infected patient than other pathogens. Among 21 Staphylococcus aureus strains 15 (72%) were resistant to methicillin. The overall daily antibiotic cost of methicillin resistant S. aureus (MRSA) infections was two to three times higher than infections with susceptible strains. However, median daily cost per infected patient for MRSA infections were lower than susceptible strain infections. Table 2 Daily antibiotic cost according to the pathogens. Pathogens No. Overall daily cost (US$) Range of daily cost per pathogen (US$) Median (US$) Pseudomonas aeruginosa 60 5567,06 17,98–204,74 100,04 Acinetobacter spp 36 4951,06 17,98–359,28 92,47 Stenotrophomonas maltophilia 4 310,99 31,31–100,04 89,82 Klebsiella spp 37 3104,22 10,02–179,64 79,6 Enterobacter spp 10 961,79 60,42–139,08 77,28 Escherichia coli 23 1652,57 10,02–179,64 60,41 Proteus spp 3 49,04 13,24–49,04 49,04 Staphylococcus aureus Methicillin-susceptible (MS) 6 476,22 10,02–142,2 74,52 Methicillin-resistance (MR) 15 1188,1 35,88–142,2 71,1 CoNS* MS-CoNS 1 49,64 16,34–49,64 49,64 MR-CoNS 3 148,92 17,52–49,64 49,64 Enterococcus spp 16 1141,25 32,29–142,2 49,64 Candida spp 21 235,52 8,56–38,64 8,56 No pathogen identified 27 3166,75 10,02–359,28 114,6 * Coagulase negative Staphylococcus Among the 350 antibiotic prescriptions for nosocomial infections, piperacillin-tazobactam and amikacin were the most prescribed antibiotics (Table 3 ). Carbapenems especially meropenem were the most expensive drugs. Table 3 Daily cost of antimicrobial agents for nosocomial infections. Antimicrobial agents Number of prescriptions Daily cost per infected (US$) Min Max Betalactams Ampicillin-sulbactam 20 23,92 71,76 Piperacillin-tazobactam 48 85,95 114,6 Ticarcillin-clavulanate 4 14,9 22,35 Carbapenems Imipenem 31 100,04 150,06 Meropenem 23 179,64 359,28 Cephalosporins Cefepime 26 38,64 77,28 Ceftazidime 18 35,82 71,64 Cefoperazone-sulbactam 16 40,28 80,56 Cefazoline 10 10,02 20,04 Ceftriaxone 3 19,7 39,4 Aminoglycosides* Amikacin 52 - 7,96 Netilmicin 21 - 24,48 Tobramycin 3 - 18,56 Fluoroquinolones Ciprofloxacin 18 49,04 98,08 Ofloxacin 2 25,74 51,48 Levofloxacin 1 41,07 82,14 Glycopeptides Vancomycin 13 49,64 87,27 Teicoplanin 12 71,1 142,2 Other antibiotics Clarithromycin 4 11,57 23,14 Trimethoprim-sulphamethoxazole 2 4,48 8,96 Metronidazole 2 5,51 10,04 Antifungal agents Fluconazole 21 8,56 34,24 Total 350 8,56 359,28 * dosage given as once-daily. Disscussion Cost is an important factor which determines the physician's choice of medication to treat patients in spesific stiuations. In this study, we tried to demonstrate the daily cost of antimicrobial treatment of nosocomial infections according to site of infection, pathogen and antimicrobial agent. In different studies, economical analysis regarding costs attributable to nosocomial infections has been evaluated and reported between $1018 to 2280 per infected patient [ 7 - 9 ]. Jarvis et al reported that the estimated average costs of nosocomial infections were $558 to 593 for each urinary tract infection, $2734 for each surgical site infection, $3061 to 40000 for each bloodstream infection, and $4947 for each pneumonia [ 1 ]. Daily cost of antimicrobial treatment has been reported to be a significant extra cost attributable to nosocomial infections. In this study, we found an average daily antibiotic cost of $89,64 per nosocomial infection. It is clear that cost of overall antibiotic treatment for a period of approximately 10–15 days is $900 to $1350. Prolongation of hospital stay has been the major extra cost attributable to nosocomial infections in many reports [ 2 - 4 ], but in comparative case-control study from our country, Yalcin et al. [ 8 ] found that cost of antibiotic therapy of $1190 per infected patient, accounted for about 75% of the total extra cost. This finding may be due to the high prices of antibiotics in Turkey. To calculate the true costs of antibiotic therapy, hidden costs arising from intravenous administration, labor, serum antibiotic assay, monitoring hematological and biochemical indices and adverse effects of antibiotics must be considered [ 10 ]. The present study does not include these relevant "hidden costs" that could substantially modify the total cost of an antibiotic treatment. Although, hidden costs were not calculated, an average daily antibiotic cost of a single nosocomial infection is found to be markedly high in our hospital. This result is within the limits reported by other large economic studies, suggesting that our data is comparable to those found in other countries and with other assessment methods. In a French prevalence survey, Astagneau et al.[ 11 ] reported an average daily antibiotic cost between FF 520 to 1085 (about $86 to $160) per nosocomial infection. French et al.[ 12 ] and Haley et al.[ 13 ] reported an average cost of antibiotic treatment of $190 and between $72 to $128 per nosocomial infection, respectively. In Turkey, Yalcin et al.[ 14 ] found that daily antibiotic cost of nosocomial infections was $70 per patient. The daily antibiotic cost varies markedly according to site of infection. Our study has demonstrated that pneumonia and bloodstream infections were associated with the highest daily antibiotic costs as reported in other studies [ 11 , 13 , 14 ]. Surgical site infections had also high daily antibiotic cost in our study. In their case-control study, Coello et al. reported that antibiotic therapy for surgical patients was the second most significant contributor to cost [ 15 ]. In the present study, nosocomially infected patients that had only one nosocomial infection were considered for analysis. Clearly, antimicrobial treatment of patients with multiple nosocomial infections might be much more expensive. P. aeruginosa infections had the highest daily antibiotic cost followed by other non-fermentative bacilli. Infections caused by P. aeruginosa are difficult to treat because of its virulence and relatively limited choice of effective antimicrobial agents, so, these infections often require combination therapy. Emergence of resistance in P. aeruginosa has been associated with increased morbidity, mortality, and costs [ 16 ]. On the other hand, although the overall antibiotic cost of MRSA infections was higher than infections with susceptible strains, the daily antibiotic cost per infected patient with MRSA was lower with susceptible strain infections. MRSA infections are treated by glycopeptides which cost less than beta-lactams in our country. Astagneau et al reported that the daily antibiotic cost of multi-resistant bacterial infections such as multi-resistant P. aeruginosa infections, was 20% higher than susceptible infections, but the daily antibiotic cost per infected patient for MRSA infections was not higher than for susceptible strain infections [ 11 ]. Expensive antibiotics, such as piperacillin-tazobactam, carbapenems, cefepime, ciprofloxacin, teicoplanin were prescribed more commonly than the cheaper agents such as ampicillin-sulbactam, ceftriaxone or ofloxacin in our ICUs. These expensive antibiotics were mainly prescribed for resistant and severe gram-negative nosocomial infections, such as ventilator-associated pneumonia and postneurosurgical meningitis in ICU. Physicians may be forced to choose empirical antibiotic therapy with broad spectrum antimicrobials by increasing bacterial multi-resistance. In conclusion, mean daily antibiotic cost was found $89,64 per nosocomial infection in our ICUs and nosocomial pneumonia had the highest daily antibiotic cost per infected patient. It is clear that cost of antibiotic therapy of nosocomial infections is an important part of extra cost attributable to nosocomial infection. Approximately one third of nosocomial infections are preventable by full implementation of the current infection control guideline recommendations [ 17 ]. Each institution should develope empirical antibiotic guidelines according to its own local nosocomial infections data. Infection control measures, such as education of health care workers regarding antimicrobial agents and resistance; isolation of patients infected with multi-resistant organisms, should be implemented to reduce infections and expensive antibiotic prescriptions. Competing interests The author(s) declare that they have no competing interests. Authors' conributions DI collated and analyzed the data, participated in the study design and was principal writer of manuscript. ANY conceived the study. GO carried out the laboratory studies. RS, FG, OT and LM participated in the patient management. All authors read and approved the final manuscript. Pre-publication history The pre-publication history for this paper can be accessed here:

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC548682.xml

545211

Meningitis and Climate in West Africa

Many different things combine to cause epidemics of disease. Among these factors are the characteristics of the infecting organism, the resistance of the host, and, as is increasingly realized, climatic conditions. El Niño, the best known climatic disturbance, is caused by a warming of the Pacific Ocean, which then affects the climate globally. Previous work has suggested that this recurring phenomenon can have a profound effect on the incidence of many diseases, including dengue, malaria, and diarrheal diseases. In a paper in this month's PLoS Medicine , Sultan and colleagues from a climate research institute and an infectious diseases center in France looked at the relation between climate and meningitis outbreaks in Mali in West Africa, a region that every year between February and May sees devastating epidemics of meningococcal meningitis affecting up to 200,000 people. The most important recurring climatic event in this region is a dry wind, known as the Harmattan, that blows throughout the winter, causing a drop in humidity and the production of vast quantities of dust. What the authors found was that over the years 1994–2002, the week of the onset of the yearly meningitis epidemic came at around the same time as the peak of one measure of the wind—the sixth week of the year. As Pascual and Dobson say in their Perspective article on this study, “Sultan and colleagues' study is exceptional in that it illustrates a clear relationship between an external environmental variable and the initiation of disease outbreaks.” How do climatic changes influence disease? In some cases, such as the role of flooding in spreading a waterborne disease, the causes are perhaps obvious, but why should a dry wind affect disease incidence? Previous works have suggested that the climate can work in a number of ways, by influencing the life cycle of both disease vectors and the disease-causing organism, and, as here perhaps, by affecting the resistance of the host. Sultan and colleagues speculate that the drying effects of the wind on the mucous membranes could increase the chances of the organism getting established in the human host. Whatever the causes, one very useful feature of climate is that, once the patterns are understood, they can often be predicted. A way of predicting these meningitis epidemics could be enormously useful. Sultan and colleagues looked at only a few years, but if these findings are confirmed over a longer time period, they could make preparing for an epidemic much more efficient.

/Users/keerthanasridhar/biomedlm/data/PMC000xxxxxx/PMC545211.xml